THE DEVELOPMENT AND APPLICATION OF A 
FREIGHT TRANSPORT FLOW MODEL FOR SOUTH 
AFRICA 
 
 
 
 
 
Jan Hendrik Havenga 
 
 
 
 
 
 
Dissertation presented for the degree of Doctor of Philosophy (Logistics 
Management) at  
Stellenbosch University 
 
Department of Logistics:  Faculty of Economic and  
Management Sciences 
 
 
 
 
Promoter  
Prof. W.J. Pienaar 
December 2007
Declaration: 
 
I, the undersigned, hereby declare that the work contained in this dissertation 
is my own original work and that I have not previously in its entirety or in part 
submitted it at any university for a degree. 
..........................                                        .......................... 
          Signature              Date 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
TABLE OF CONTENTS 
 
List of Tables v 
List of Figures vi 
CHAPTER 1 – DISSERTATION SUMMARY ............................................................. 1 
CHAPTER 2 - PROBLEM STATEMENT  AND OVERALL APPROACH .................. 7 
2.1 Introduction and background 7 
2.2 Definitions of logistics and measurement 12 
2.2.1 Definition of measurement................................................................................. 12 
2.2.2 A definition of logistics ....................................................................................... 14 
2.3 Global trends 15 
2.3.1 Industrial development context .......................................................................... 15 
2.3.2 An underlying typology....................................................................................... 16 
2.3.3 Global trends...................................................................................................... 19 
2.4 The state of global macro-logistics measurement 29 
2.4.1 Typological differences ...................................................................................... 30 
2.4.2 Infrastructure requirements................................................................................ 31 
2.4.3 Performance of logistics as a cost to the economy ........................................... 32 
2.5 The South African situation 35 
2.5.1 Transnet events ................................................................................................. 35 
2.5.2 The increasing scarcity of information ............................................................... 36 
2.5.3 Social factors ..................................................................................................... 37 
2.5.4 The quest for measurement............................................................................... 39 
2.6 Proposed approach 41 
2.6.1 Methodology ...................................................................................................... 41 
2.6.2 The transport flow model ................................................................................... 42 
2.6.3 The commodity flow model ................................................................................ 43 
2.6.4 The logistics cost model .................................................................................... 43 
2.7 Economic and historical base for continuance 43 
CHAPTER 3 - MACRO-ECONOMIC LITERATURE STUDY ................................... 44 
3.1 Introduction and background 44 
3.2 The quest for a macro context  
3.2.1 The macro context of science............................................................................ 44 
3.2.2 Macro-economics............................................................................................... 47 
3.2.3 Macro-measurement.......................................................................................... 48 
3.3 The systemic relationship between the various roles of logistics 52 
3.4 Economic systems and the role of logistics 55 
3.4.1 Macro-economics, clusters and cross-cutting factors........................................ 55 
 i
3.4.2 Economic clusters.............................................................................................. 57 
3.4.3 Cross-cutting factors defined ............................................................................. 63 
3.5 From macro-economic imperative to measurement 64 
CHAPTER 4 - LITERATURE STUDY AND HISTORICAL CONTEXT OF MACRO 
TRANSPORT LOGISTICS MEASUREMENT HISTORY IN SOUTH AFRICA ........ 66 
4.1 Introduction 66 
4.2 Infrastructure development during the primary economy phase 66 
4.3 Production and operations management during the industrial economy phase 68 
4.3.1 Introduction ........................................................................................................ 68 
4.3.2 Verburgh’s 1958 contribution............................................................................. 73 
4.4 Transport and distribution logistics during the market economy phase 82 
4.4.1 Introduction ........................................................................................................ 82 
4.4.2 Verburgh’s 1968 contribution............................................................................. 84 
4.4.3 Smith’s contribution............................................................................................ 88 
4.4.4 Hamilton’s first study.......................................................................................... 94 
4.4.5 Hamilton’s second study.................................................................................... 98 
4.5 Integrated trading platforms during the establishment of electronic communities 103 
4.5.1 Introduction ...................................................................................................... 103 
4.5.2 Freight Transport Data Bank ........................................................................... 107 
4.5.3 Land freight transport statistical review ........................................................... 111 
4.6 Globalised systems during the globally networked economy (2000-2020) and the 
expectations around global bio-engineering 113 
4.7 Lessons learned 115 
CHAPTER 5 - RESEARCH OBJECTIVES AND METHODOLOGY ...................... 116 
5.1 Introduction 116 
5.2 Problems identified in past research 117 
5.2.1 Methodology .................................................................................................... 117 
5.2.2 Repeatability .................................................................................................... 118 
5.2.3 Scope............................................................................................................... 118 
5.2.4 Context............................................................................................................. 119 
5.2.5 Micro-economic detail...................................................................................... 120 
5.2.6 Systemic relationships ..................................................................................... 121 
5.2.7 Global links and benchmarking........................................................................ 122 
5.2.8 Memories of the future..................................................................................... 122 
5.2.9 Scalability......................................................................................................... 123 
5.3 Overarching methodology 123 
CHAPTER 6 - NATIONAL FREIGHT FLOW MODEL............................................ 124 
6.1 Introduction and background 124 
 ii
6.2 The structure of the national freight flow model 124 
6.2.1 Rail freight flows............................................................................................... 125 
6.2.2 Road freight flows ............................................................................................ 125 
6.3 Calculation of road tonnages 126 
6.4 Calculation of rail tonnages 132 
6.5 List of corridors, metropolitan and rural areas 132 
6.6 The national freight flow model 134 
CHAPTER 7 - NATIONAL FREIGHT FLOW MODEL RESULTS AND STRATEGIC 
INTERPRETATION................................................................................................. 135 
7.1 Introduction 135 
7.2 Total supply of, and demand for, surface freight transport 135 
7.3 Typologies and the current distribution of freight 141 
7.4 The current position of rail 147 
7.5 Trend in freight movements over the typologies 151 
7.6 Trends in movement over the various corridors 154 
7.7 Trends in movement in the various metropolitan areas 155 
7.8 Trends in movement in the various rural areas 156 
7.9 Developing transport performance measures 157 
7.10 Infrastructure planning 159 
CHAPTER 8 - COMMODITY FLOW MODEL METHODOLOGY ........................... 160 
8.1 Introduction and background 160 
8.2 Supply and demand per commodity on a geographical basis 161 
8.2.1 Inter-sectoral tons moved ................................................................................ 161 
8.2.2 Estimation of freight on a regional basis.......................................................... 167 
8.3 Estimation of flows per commodity 170 
8.3.1 Modelled flows ................................................................................................. 170 
8.3.2 Enhancing modelled flows based on the Pareto process................................ 176 
8.4 Long-term forecast of the South Africa economy – the validation process 177 
8.4.1 Broad methodological processes..................................................................... 177 
8.4.2 The BER’s macro-economic forecasting model .............................................. 178 
8.5 The final model output 179 
CHAPTER 9 - COMMODITY FLOW MODEL RESULTS AND STRATEGIC 
INTERPRETATION................................................................................................. 180 
9.1 Introduction 180 
9.2 Analysis of the economy by weight 180 
9.3 Commodity growth 185 
9.4 Corridor demand 187 
9.5 The indirect advantages of commodity flow views 193 
 iii
9.6 Understanding break bulk commodities – the case of processed foods 196 
9.7 Conclusion 197 
CHAPTER 10 - LOGISTICS COST PERSPECTIVES............................................ 198 
10.1 Introduction 198 
10.2 The development of the model 199 
10.3 The global state of logistics 200 
10.4 Results for South Africa 202 
10.5 A cost case study for South Africa – taking macro-logistics cost measurement to 
the next level 205 
10.5.1 Finding tariffs ................................................................................................... 206 
10.5.2 Developing a model for a possible macro saving ............................................ 208 
10.5.3 Applying the model .......................................................................................... 210 
10.6 Conclusion 215 
CHAPTER 11- THE CONTRIBUTION OF THIS RESEARCH AND FUTURE 
CHALLENGES ....................................................................................................... 217 
11.1 Introduction 217 
11.2 The contribution of this research 219 
11.3 Development and growth objective of the economy – how to implement change 221 
11.4 Distribution of wealth objective of the community – the national development 
application of this work 222 
11.5 Sustainability of the environment 225 
11.6 Measurement for future direction – a new model for sustainable measurement 229 
11.7 Final thoughts on the macro perspective 233 
11.8 Conclusion 234 
LIST OF SOURCES................................................................................................ 236 
 
 
 
 
 
 
 
 
 
 
 
 
 iv
List of Tables 
Table 1: Alford-Bangs's standard inventory carrying charges table........................................ 33 
Table 2: Logistics cost comparisons for various economies ................................................... 34 
Table 3: Overarching economic classification......................................................................... 61 
Table 4: Detailed sub-sectors of the extraction and beneficiation sectors of the economy .... 62 
Table 5: Phases of logistics development ............................................................................... 65 
Table 6: Modal market share in 1957...................................................................................... 80 
Table 7: Distribution of road transport for reward within and beyond 50 km zones in 1958 ... 81 
Table 8: Verburgh's forecasts for rail transport in tons (‘000) vs. actual data......................... 86 
Table 9: Verburgh's forecasts for economic production in tons (‘000) vs. actual data............ 87 
Table 10: Modal market share................................................................................................. 89 
Table 11: Modal market share in percentage.......................................................................... 89 
Table 12: Average transport distance ..................................................................................... 89 
Table 13: Van der Veer's market share figures of 1981.......................................................... 93 
Table 14: Comparison of Van der Veer's 1981 figures with those of Verburgh ...................... 93 
Table 15: Hamilton's modal market share figures compared to Verburgh and Smith............. 96 
Table 16: Fleet performance calculations based on Hamilton's 1981 data............................. 96 
Table 17: Modal market share in 1985 (Hamilton) .................................................................. 99 
Table 18: Hamilton's modal market share % of 1985............................................................ 100 
Table 19: Hamilton's average transport distance per mode.................................................. 100 
Table 20: Economic performance related to transport input ................................................. 101 
Table 21: Pretorius' modal market share 1985-1990 ............................................................ 108 
Table 22: Undercounting of rail performance in Pretorius' sample ....................................... 109 
Table 23: Modal market share from 1957 - 1990 based on all available sources................. 109 
Table 24: Modal share - Hamilton's first study ignored and Pretorius’ data normalised ....... 110 
Table 25: Shipment frequency of the transportable GDP ..................................................... 111 
Table 26: Commodity transport income for reward - Statistics South Africa 2003 data........ 112 
Table 27: Number of CTO stations per year analysed.......................................................... 126 
Table 28: Average gross ton per truck type as calculated from RFA data............................ 129 
Table 29: Tare based on vehicle concepts published by the Road Freight Association....... 129 
Table 30: Corridors................................................................................................................ 133 
Table 31: Rural and metropolitan areas ................................................................................ 134 
Table 32: Surface freight supply in South Africa between 1993 and 2004 (million tons)...... 136 
Table 33: Surface freight market share between 1993 and 2004 ......................................... 137 
Table 34: Survey comparison: All surveys ............................................................................ 137 
Table 35: Market share comparison: All surveys (based on tons) ........................................ 138 
Table 36: GDP growth rate comparison ................................................................................ 139 
Table 37: Correlation coefficient comparison for all surveys................................................. 140 
Table 38: Summary of typology characteristics..................................................................... 146 
Table 39: Tonnage growth over various typologies (million tons) ......................................... 152 
Table 40: Percentage market share within typology ............................................................. 153 
Table 41: Corridor growth from 1993 to 2004 (million tons).................................................. 154 
Table 42: Metropolitan freight transport growth between 1993 and 2004 (million tons) ....... 156 
Table 43: Rural freight transport growth between 1993 and 2004........................................ 157 
Table 44 : Regional descriptions ........................................................................................... 168 
Table 45: Decay functions for various commodities.............................................................. 174 
Table 46: Number of times each ton produced is observed on typologies ........................... 182 
Table 47: Impact of a cumulative set of intermodal facilities................................................. 193 
Table 48: Conversion factors of logistics............................................................................... 203 
Table 49: Comparable typology costs in cent per ton-kilometer ........................................... 206 
Table 50: Volume and tariff cases for case study ................................................................. 209 
Table 51: Gauteng to Durban model result (rand millions, except first data row) ................. 211 
Table 52: Gauteng to Cape Town model result (rand millions, except first data row)) ......... 212 
Table 53: Cost saving for 75% switch scenario with price increase options......................... 213 
Table 54: Transport costs for a ton of containerised cargo................................................... 224 
 
 
 v
List of Figures 
 
Figure 1: The three pillars of sustainable development............................................................. 8 
Figure 2: From sustainable development to logistics requirements........................................ 10 
Figure 3: Abridged illustration of the growth in economic platforms........................................ 16 
Figure 4: Physical support architectures of economic platforms............................................. 18 
Figure 5: Global trends relating to the underlying typology..................................................... 20 
Figure 6: Factors influencing customers' preference for global logistics services .................. 28 
Figure 7: From firm-level interests to macro trans-national interests...................................... 49 
Figure 8: The development of clusters .................................................................................... 59 
Figure 9: The economic value chain according to the standard industrial classification......... 61 
Figure 10: Logistics phases and surveys ................................................................................ 66 
Figure 11: Verburgh's 1958 illustration.................................................................................... 74 
Figure 12: Rail transport as percentage of transportable GDP in tons (Indexed 1937=100).. 77 
Figure 13: Jones's analysis of traffic (note uneven distribution in ton and income)................ 91 
Figure 14:  Rail transport as percentage of transportable GDP (indexed 1937=100)............. 92 
Figure 15: Extrapolation of Verburgh's original 1957 data sets (indexed 1937=100) ........... 101 
Figure 16: Research trade-offs.............................................................................................. 118 
Figure 17: The role of decision making in the strategic process ........................................... 120 
Figure 18: Scenarios of the future ......................................................................................... 122 
Figure 19: Average daily truck traffic N1 Cape Town-Johannesburg (bi-directional) – 2004 130 
Figure 20: Average daily truck traffic N1 Cape Town-Johannesburg (bi-directional) – 2005 131 
Figure 21: Average daily truck traffic N3 Durban-Johannesburg (bi-directional) – 2004 ...... 131 
Figure 22: Average daily truck traffic N3 Durban-Johannesburg (bi-directional) – 2005 ...... 132 
Figure 23: South African indicators as a percentage of world indicators .............................. 141 
Figure 24: Freight typologies (ton-kilometers and percentage share)................................... 143 
Figure 25: Schematic of primary traffic.................................................................................. 143 
Figure 26: Schematic representation of metropolitan traffic.................................................. 144 
Figure 27: Schematic representation of corridor traffic ......................................................... 145 
Figure 28: Schematic representation of rural traffic .............................................................. 146 
Figure 29: Supply and demand of modes in the typologies .................................................. 147 
Figure 30: Rail competitiveness and sustainability................................................................ 150 
Figure 31: Constant rands (2000 figures) produced in the South African economy per ton. 158 
Figure 32: The performance per ton of the transportable economy in GDP terms ............... 159 
Figure 33:  Commodity flow modelling methodology............................................................. 161 
Figure 34: Cartographical indication of regions..................................................................... 169 
Figure 35: Growth rate of economic sectors by weight ......................................................... 181 
Figure 36: Expected future supply and demand in the South African economy ................... 182 
Figure 37: Flattening of number of times freight appears on typologies ............................... 183 
Figure 38: Verny's observation of freight transport statistics and GDP figures for Europe... 184 
Figure 39: Growth in demand by weight of the top 23 commodities ..................................... 185 
Figure 40: Growth in demand by weight for top 23 commodities (excluding coal)................ 186 
Figure 41: Percentage weight growth in commodities........................................................... 186 
Figure 42: Expected growth in total traffic on corridors......................................................... 189 
Figure 43: Dry bulk corridor demand..................................................................................... 190 
Figure 44: Liquid bulk demand on corridors (excluding pipelines)........................................ 190 
Figure 45: Break bulk corridor demand ................................................................................. 191 
Figure 46: Containerisation propensity.................................................................................. 192 
Figure 47: Western Cape vs. South African GDP growth ..................................................... 195 
Figure 48: US logistics costs as a percentage of GDP ......................................................... 201 
Figure 49: World logistics costs according to income groups ............................................... 202 
Figure 50: Trends in logistics cost components (2003 % contribution indexed, 2003=100). 203 
Figure 51: Break bulk corridor traffic in the context of total traffic (million tons 2004) .......... 205 
Figure 52: Regional GDP per capita as % of GDP per capita of the Western world ............ 219 
Figure 53: Demkes and Tavasszy's description of micro, meso and macro indicators ........ 223 
Figure 54: CO
2 
emissions - gram per ton-kilometer .............................................................. 225 
Figure 55: Obnoxious gas emissions - grams per ton-kilometer........................................... 226 
Figure 56: Future freight transport by mode.......................................................................... 227 
 vi
 vii
Figure 57: Future freight transport contribution to CO
2 
emissions ........................................ 228 
Figure 58: Verny's definition of sustainable development..................................................... 230 
Figure 59: Wealth and spatial organisation factors of sustainable development .................. 231 
Figure 60: Sustainability challenges in terms of spatial and community position ................. 232 
Figure 61: Measurement systems and objectives................................................................. 233 
 
 
 
 
 
 
 
 
 
 
 
Chapter 1 – Dissertation Summary 
South Africa currently experiences the double jeopardy problem of catching 
up to global economic competitiveness whilst at the same time feeling the 
pressures of sustainability management spearheaded by a global agenda. 
Global sustainability is defined as growth that is shared without depleting 
natural resources or damaging the environment. Academic disciplines are 
challenged to make a contribution and economics as such should contribute 
by providing the lead and lag indicators for the planning and measurement of  
scarce resources usuage. This integrative view includes economic sub-
disciplines, such as logistics. 
 
This integrative view is an acknowledged part of the economics discipline, 
except that the macro-economic context of some sub-disciplines, such as 
logistics, often receives less attention during the course of academic activities.  
The distribution of resources and outputs in the economy is a logistics 
controlled cross-cutting factor, but suffers from a lack of macro-economic 
perspective, and lead and lag orientated measurement. This state of the 
affairs is a historic backlog of logistics and its specific position within 
economics.  
 
During the primary economic era the world began to configure networks and 
markets, which became more pronounced and settled with the dawn and 
settling of the industrial era. Logistics then was a “given” and did not receive 
much thought even as industrial, market economies developed.  Transport 
was regarded as an administered cost, i.e. inefficiencies in logistics systems 
were evenly distributed between competitors, not giving any specific entity an 
advantage. With the advent of global competition and the diminishing returns 
on other cost saving measures, companies began to collaborate and integrate 
logistics functions within value chains, but the administered part of transport 
costs failed to receive the attention it required. In this way, global competitors 
did begin to experience disadvantages on a national level as whole 
 1
economies suffered from inefficiencies in logistics and specifically transport 
systems. 
  
Only a few theorists are beginning to emphasise the macro-logistics 
perspective, but not in depth, especially as it relates to infrastructure planning, 
relative sophistication between nations  and differences in “vision” between 
nations. Precious little is, however, offered in terms of performance and 
planning measurement systems. Well-respected national measurement 
systems have been developed in a few cases, but they are usually very 
broad, not integrative, do not provide analysis of specific macro-issues and 
are based on very old methodologies. 
 
In South Africa specifically, the problem of measurement absence was 
exacerbated over the past two decades when state assets were 
commercialised, the transport market deregulated and information on the 
performance of these sectors in the economy became even scarcer. South 
Africa’s rail system is old, has a narrow gauge and is uncompetitive in general 
freight, which places a massive burden on expensive road freight.  The 
specific structure of South Africa’s transport market is important, because of  
high concentrations of people, resources and manufacturing plants far from 
the coast. Requests for research have been raised, specifically on modal 
market share, freight flows, including commodity freight flows and forecasting, 
but the few research efforts in this regard over the past 50 years failed to 
provide any meaningful answers. 
 
South Africa’s transport development history followed five phases that can be 
described as foundations, industrial revolution, network, market connection 
and overcropping. During the foundation phase between 1830 and 1870 
spatial planning was more difficult and as population densities were low 
research based approaches were not always neccesary to the degree 
required in the industrial era. During the industrial revolution phase between 
1870 and 1910, the current narrow gauge was installed, more because of 
short-term profitability perspectives during road versus rail price wars rather 
than long-term based macro-research. During 1910 to 1950 in the network era 
 2
the final rail network was established, transport was regulated and network 
design errors made as the new constituents insisted on a rail network for 
commercial farming, which hampered investment in major densifying 
corridors. The network was funded by cross-subsidisation from high-value 
traffic on the corridors, where investment did not keep track with concomitant 
high tariff based returns. This caused initial surreptitious, but later open, often 
illegal road competition in the market connection era, between 1950 and 
1990, that was fraught with lawsuits and policing and general apprehension 
for the road mode from the railways. Rather than developing intermodal 
systems, as access roads improved and corridors densified even further,  the 
modes remained separate. All South Africa’s modal studies were done in this 
era, always through questionnaire-based approaches, but the results of the 
five known surveys (the fifth one was repeated a few times using the same 
methodology) were erratic and unreliable. No macro-economic perspectives 
were given and the stated intent to model total flows, commodity flows and to 
forecast never achieved. Only modal market share was reported on and only 
three of the studies (1958, 1971 and 1985) achieved some degree of 
acceptance as valid. The absence of research was a major cause of mistakes 
made during the first half of the overcropping era, 1990 to 2005 as transport 
was deregulated, the railways run down and transport densities (a major 
driver of efficiency) declined. Improved research was often requested, but to 
date precious little has been done. Information from Statistics South Africa is 
proven to be unreliable and even with the advent of a National Freight 
Logistics Strategy in 2005, little analysis is based on actual, repeatable 
research. 
 
To address the lack of research three models are proposed to provide a valid, 
sustained and repeatable measure of freight flows, i.e. the national freight flow 
model, the commodity flow model and the logistics cost model. 
 
The national freight flow model utilises vehicle counting technology at 
various permanent (398 stations) and secondary (430 stations) counting 
stations to model road data.  Actual rail data are used and the model was then 
run for 1993, 1997, 2003 and 2004. 
 3
 
The research confirmed that in the absence of intermodal traffic South Africa’s 
rail system (which is bigger than the next four in Africa combined) is in serious 
decline for the general freight mode. Total tonnage expanded by 50%, with 
about 60% growth on road over the last 11 years. The only rail growth was on 
ring-fenced export freight that grew by about a third. This means that South 
Africa’s economic growth will have to be carried by its road network and the 
country cannot afford it, because the dense corridors are becoming expensive 
(because it is not leveraged with intermodal technology).  This, in turn makes 
it difficult to enable the second economy to gain access to markets. The 
model was further applied to quantify transport in ton-kilometers (for exact 
scalability) to understand the distribution between various proposed 
typologies.  The primary typology describes world-class one-directional bulk 
transport with few origin to destination points of low value mining 
commodities.  The main challenge is systemic competitiveness with other 
deposits around the world that are mostly closer to the coast. The 
metropolitan typology describes many origin to destination points in cities with 
the main challenge to alleviate congestion. The rural typology describes many 
origin to destination points in rural areas with the main challenges to provide 
development as well as access to corridors and metropolitan markets. The 
corridor typology describes long haul, often high-value commodities, which 
originate at many points in cities, then converge, travel to a divergence point 
within other cities and are then distributed to consumption points. The 
challenge here is to alleviate South Africa’s spatial challenge by providing 
effective solutions, probably intermodal, to solve this sizeable portion of the 
country’s freight transport problems. 
 
South Africa’s population (under 1% of the world population) produces less 
than half a percent of the world GDP, but requires 2% of the world’s ton-
kilometers. This situation arises from the country’s economic development 
history that placed mining production and population development far from 
coastal areas, in a relatively open mineral export, and beneficiated product 
and energy import economy, thereby creating long export and import corridor 
requirements.   
 4
 
The measurement of performance for each mode over each typology was 
also necessary to understand whether objectives are being met. All growth 
over densified corridors occurred in the road mode which expanded by more 
than 60% over eleven years. This growth would be understandable if the 
corridors in question are short or the density per corridor low. In these 
instances the economy would have to absorb this growth in the road mode. 
Cheaper options are, however, available in intermodality if the density per 
corridor can be calculated as sufficiently high. In South Africa’s case the 
spatial efficiency objective of the corridor typology is not achieved. If this 
density is sufficient to sustain an intermodal solution, future investment should 
be channelled to such solutions. This could release funds for the development 
of the second economy in rural areas, which is a major economic objective of 
South Africa at present. Rural road traffic grew slower than corridor traffic 
(less than 60%), which correlates with the hypothesis that South Africa is not 
succeeding in the desired stimulation of rural economies. A major cause of 
this failure is deteriorating road infrastructure, which can now for the first time 
be measured. The development objective of this typology is therefore not 
achieved. Metropolitan growth is also slower than corridor growth. 
 
Performance measures could also be developed and it showed that South 
Africa’s GDP per ton generated fell from R1 000 about two decades ago to 
close to R900 in 1993 and lower than R800 in 2004. The transport system is 
therefor evaluated as not performing as it should, mostly because planning 
was not based on measurement and understanding of the realities of the 
situation. 
 
The commodity flow model’s objective was to translate total flows into 
commodity flows and develop a 20-year forecast. It was believed that 
forecasting on a commodity level will provide better insight into transport 
infrastructure demand planning. The model utilises the input-output model of 
the economy, disaggregated to 354 magisterial districts and translated from 
value to tons.  Gravity modelling principles were applied to determine flows for 
26 commodity groups. Forecasting was done on a commodity level, 
 5
 6
referenced independently and the top 20 commodities’ distribution, flow and 
forecasts were independently researched and verified. The model clearly 
illustrated that the dense corridors will grow by more than 100% over the next 
20 years, that the growth will occur faster in automotive, wood and chemical 
products and very fast in most commodities that can be containerised. It 
highlighted the specific nature of intermodal solutions and the size of 
container terminals that could be developed and where. 
 
As a final step the results of the models were compared to the cost of South 
Africa’s logistics in total and what the effect of these observations could be 
on logistics costs. The transport sector contributed more than 63% of logistics 
costs in South Africa (R141 billion of R223 billion) and grew faster as a cost 
factor than total logistics costs over the past two years. This compares poorly 
with the figure for the rest of the world which is below 40%. 
 
South Africa need real information to understand the logistics performance of 
the economy and this study was able to provide it and also indicate the 
direction for future planning in order to alleviate problem areas. Through this 
study a better understanding of an important part of South Africa’s economic 
system, that will inform and improve the management there-of was achieved. 
The study made it possible to measure on an annual and repeatable basis the 
volume, direction and modal market share of freight in South Africa, the 
commodities involved, its relation with logistics costs and forecasts for the 
future. For the first time the macro relationship between these major 
components are understood and the country can plan for future challenges. 
 
Chapter 2 - Problem statement  
and overall approach 
2.1 Introduction and background 
The first decade of the twenty-first century in South Africa will go down in 
history as the decade of debate regarding infrastructure, infrastructure 
planning and infrastructure requirements for the next half century. The context 
for these questions and the rationale behind them are South Africa’s national 
priorities of job creation, economic growth, poverty eradication, social 
cohesion and national identity
1
. Sustainable development is the core enabler 
of these priorities.  
 
Alfsen and Thorvald arrive at a definition of sustainable development by 
stating that: 
the point of departure should be economic development … to respect 
future environmental and social conditions that are acceptable to our 
grandchildren… (with) widespread agreement that a main threat to global 
sustainable development can be found in the uneven distribution between 
rich and poor.
2
 
 
Sustainable development is therefore defined as growth that is shared and 
that can be maintained within the context of the replenishability of the 
production factors consumed by the growth. This means that sustainable 
development requires the economy (i.e. wealth) to increase, requires society 
to ensure that everyone benefits from this increase in wealth, and requires the 
management of this process to ensure that the environment can continue to 
sustain the growth (Figure 1).  
 
 
 
                                                 
1
 Jordan, 2007, p. 1. 
2
 Alfsen and Thorvald, 2005, p. 5. 
 7
 
 
Economy Community Environment
 
 
 
 
Figure 1: The three pillars of sustainable development
3
 
 
In the ensuing debate around how these objectives will be met and what the 
infrastructure requirements are, another set of clear guidelines is emerging. 
The process requires planning, public/private partnerships and the practical 
application of theory to achieve these objectives.   In this regard, the various 
disciplines in the academic sciences are actively commandeered to apply 
scientific thought and practical contributions to South Africa’s infrastructure 
challenges. 
 
The one academic discipline that could be a major contributor in this area is 
economic sciences, for reasons both intrinsically described in the science 
itself and extrinsically described in the role of economics in human culture and 
scientific thought.  Intrinsically the discipline describes itself in well-known and 
accepted terms as an empirical social science
4
 studying those activities that 
involve the production and exchange of goods. It also analyses the 
movements in the total economy, commerce amongst economies and the 
choices involved in these processes.
5
 
 
Extrinsically economics could be described as one of the fundamental 
modalities of society
6
 that defines the human ability to save resources and 
distribute them over time.
7
   
 
                                                 
3
 Morris, 2005, p. 87. 
4
 Mohr and Fourie, 1996, p. 17-18. 
5
 Samuelson and Nordhaus, 1989, p. 4-5. 
6
 Strauss, 1978, p. 10. 
7
 Strauss, 1978, p. 20-21. 
 8
This means that economics could be regarded as a pure science with its own 
unique a priori reason for existence. It is defined by Samuelson and Nordhaus 
as: 
the study of how societies use scarce resources to produce valuable 
commodities and distribute them among different groups.
8
 
 
Within the context of this definition, economics has always acknowledged 
both a macro-economic and micro-economic field, where total economic 
functioning (macro) is distinguished from component behaviour (micro).
9
 
 
Its purpose in the macro field is not to make the direct technical contribution 
attributable to engineering, natural sciences and medicine.  However, 
economic sciences are positioned squarely in the ambit of the management of 
economic growth and therefore, by definition, planning. The planning of 
economic growth has always been driven by a solid set of researched 
parameters. The targets and performance on these parameters, as lead and 
lag indicators, have to be known in order to be managed. This is done in great 
detail for overarching economic performance as well as for many of the inputs 
into this performance, such as for banking, productivity, mineral resources 
and some other macro-economic factors. These inputs, often called 
production factors, are well known, but are by no means exhaustively 
described. In fact logistics, though often not specifically identified as a 
production factor, is a clear and unambiguous part of the Samuelson and 
Nordhaus definition quoted previously. Logistics is in fact the factor that 
enables regional specialisation (and thereby economic growth) through 
efficient and effective exchange systems.
10
 These relationships are depicted 
in Figure 2.  
 
 
 
 
                                                 
8
 Samuelson and Nordhaus, 1989, p. 5. 
9
 Samuelson and Nordhaus, 1989, p. 5. 
10
 Pienaar, 2005, p. 2. 
 9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Creation
of
wealth
Distribution
of
wealth
Sustainability
of
system
Req
u
ires
Sustainable development
Specialisation Logistics requirements
Requires
R
e
quir
e
s
Integration
Systemic
and
integrated
Systemic
and
integrated
Administration
Productive
regional
specialisation
Effective
exchange between
regions
Economy Community Environment
Specialisation Exchange Transport Storage Management
Physical 
assets
Flow 
management
Scientifically 
planned
Logistics system
Figure 2: From sustainable development to logistics requirements 
 
In the field of logistics a chronic and ever-widening gap emerged as far as the 
macro-economic context is concerned. Research is scant and over the past 
five decades very little in the development of a macro-economic context for 
logistics has been achieved, and definitely nothing comparable to the well 
researched field of pure macro-economics. As far as the macro-economic 
freight transport network of this country is concerned, research over the last 
fifty years has produced only five studies.  When they are analysed, it is found 
that none of the core objectives of these studies were achieved. The modal 
market share, flow of commodities, future demand for flow and costs are 
simply not known.  
 
The objective of the research presented in this dissertation is to introduce the 
fundamental model that can continuously provide these answers and to 
illuminate the process whereby these parameters can achieve the same 
macro-economic and national priority status as all other macro-economic 
parameters.  The overarching design of this approach requires a few specific 
 10
building blocks. In Chapter 2 the macro-economic point of departure is 
clarified; the history of logistics within this context described and a typology 
developed which can be used as a structuring tool for a literature survey of 
what was achieved in this field in South Africa in the past two centuries. Once 
this foundation is in place, the problem in more specific terms around the 
necessity for logistics measurement and the meaning of the concept of 
measurement are discussed within this context. This paves the way for a 
better understanding of the theoretical framework on which a macro concept 
is based and how its measurement support system should perform, informed 
by a comprehensive literature study (Chapter 3).  Chapter 4 provides a 
detailed overview from the available research literature of the measurement 
performance actually achieved. From this overview of actual performance, a 
specific set of guidelines emerges that informs the performance criteria for a 
macro-logistics measurement system.  These guidelines enable the proposed 
overarching design of such a measurement system, as set forth in Chapter 5. 
 
The proposed measurement system is designed around three core models 
that describe the macro freight transport performance of an economy. The 
first of these models describes total flows according to network types, leading 
to an understanding of freight supply in a modal and network sense (a modal 
network measurement of the South African economy has never been 
achieved before). The model design is described in Chapter 6 and applied in 
Chapter 7. The second model describes flows on a commodity demand side 
basis and, because it is demand driven, it enables forecasting of freight 
demand per commodity in South Africa. The model design is described in 
Chapter 8 and applied in Chapter 9.
11
 The third model calculates logistics 
costs for the macro-economy
12
 and is described and applied in Chapter 10. 
Chapter 11 illustrates the validity of the overarching results of this work and 
how it can be used in future to become a standard annualised system of lead 
and lag indicators for the managing of macro-logistics performance. 
                                                 
11
 The national freight flow model and commodity flow model have been researched and 
developed over the last 11 years in collaboration with Ilse Hobbs 
12
 This model was scoped together with this research and during the time period of this 
research (and in tandem with it) further developed by Wessel Pienaar, Francois Botes 
and Neil Jacobs. 
 11
2.2 Definitions of logistics and measurement 
This dissertation specifically proposes a fundamental measurement model for 
macro-logistics measurement in South Africa. Definitions of the concepts of 
measurement and logistics are therefore required. 
 
2.2.1 Definition of measurement 
Measurement as a concept needs to be seen against the canonical steps of 
a decision-making framework, since measurement does not add value unless 
it informs decision making.  Many decision-making models exist, but at its 
core the inputs are information gathering, information analysis or 
categorisation, and intelligence development as the final step in the decision-
making process.  This intelligence should adhere to certain principles, or at 
least endeavour to achieve certain performance goals.  
 
Hofstadter
13
 proposes certain general principles for intelligence which are 
valid in all contexts. He challenges the measurer to allow flexibility in data 
interpretation depending on the interpretive requirement, find new, novel and 
integrative areas to apply the data, deal with both paradoxes and similarities 
and highlight important areas for consideration. These principles call for 
significantly more than the reactive management information that is often 
available to decision makers and  encourage the measurer to integrate the 
measurement with the measurement subject itself. This approach is 
integrative and intrinsically macro-oriented in nature as the connectivity of 
each research subject to higher orders of being is taken into account. If the 
speed of inventory turnover for a business, for example, is being researched, 
Hofstadter’s definition would mean that the paradoxical result of higher speed 
making the administered cost of transport more expensive should also be 
considered.  (This is because lower inventory levels and smaller faster-
moving parcel sizes lead to reduced traffic densities, followed by modal shift). 
In this example the researcher might take the transport costs of any specific 
situation in a firm as a given administered cost, but the cumulative effect of all 
                                                 
13
 Hofstadter, 2000, p. 26. 
 12
the decisions will impact on the administered cost and thereby make the 
system as a whole either more or less competitive, in turn impacting on the 
cost of each shipment. Through this definition it is therefore shown that all 
measurement could, in fact, have a macro character. 
 
Measurement as a management activity is both a lead and lag indicator of 
efficient and effective management practices. As a lead indicator it supports 
the targeted deployment of operational and capital resources, and as a lag 
indicator it indicates performance and prepares the way for corrective action. 
The tool is usually deployed in micro-economics on a firm level to determine 
capital and operational investment requirements and measure logistics 
performance efficiency. On a micro scale in the logistics field, Kaydos 
describes measurement as assigning “a numerical scale to the size, value or 
other characteristic of a tangible or intangible object.”
14
  This requires 
scalability of the various performance measurements, an objective that is not 
often achieved with macro-logistics measurement. 
 
On a macro-economic scale measurement is common as a lead indicator to 
determine national policy, especially monetary and fiscal policy, and as a lag 
indicator to determine various elements of the performance of the national 
economy. It is less common as a tool in logistics infrastructure investment 
decisions and performance measurement, but sorely needed. As South Africa 
grapples with investment requirements for economic infrastructure 
development, guidance on where to spend limited resources is of critical 
importance. Over the next few years R400 billion
15
 (equivalent to one third of 
the country’s annual gross domestic product) must be spent by government 
on logistics infrastructure and it is self-evident that even relatively minor 
mistakes in this spending programme could have significant negative impacts. 
Specifically for rail, highways, ports and intermodal hubs, policy makers are 
experiencing an ever-increasing shortage of useful measurement tools to 
guide decision making and measure performance. This dissertation intends to 
fill this void specifically.  
                                                 
14
 Kaydos, 1999, p. 15. 
15
 “Billion” in this dissertation means one thousand million 
 13
Lakshmanan and Anderson emphasise the need for performance-based 
research to demonstrate the link between logistics infrastructure investment 
and economic growth.  This will enable an understanding of the “effects of 
logistical transformation, productivity enhancing location shifts, and value 
adding effects” and “ex post assessment of major infrastructure projects and 
programs”.
16
  Their request is clearly for both the input (lead) and output (lag) 
performance measurements of macro-logistics systems. 
 
2.2.2 A definition of logistics 
Logistics can be defined as an extended time and place utility for 
intermediate and finished goods. The fact that intermediate and final 
production usually does not occur at the time and place of consumption 
requires the intermediate storage and transportation of these goods up to the 
point of consumption. In micro-economics the costs of this displacement are 
usually carried by the freight owner, who can either be the seller of the goods 
(in which case it is sold delivered) or the buyer (in which case it is bought ex-
works). The cost of overcoming the time and place disparity, however, 
remains in the value chain, irrespective of who is responsible for its payment.  
The value of integrated logistics management is that it can reduce this cost. 
The phenomenon becomes even more striking where logistics is measured on 
a national basis and the specific logistics costs are carried by the national 
economy as the freight owner by proxy. This dissertation will address the 
issue on a national basis and make specific recommendations in this regard. 
 
Logistics costs can be broken down into three direct elements, namely 
transportation, warehousing and administration, and one indirect element, 
namely inventory carrying costs (the time-based working capital financing cost 
of inventory in the logistics chain). This dissertation will report on all four 
elements. Transportation costs will be highlighted in far more detail, as this 
cost is hypothesised to be not only the largest element of logistics costs, but 
                                                 
16
 Lakshmanan and Anderson, 2002, p. 17. 
 14
also to be growing faster than the other elements of logistics costs in most 
economies. 
 
2.3 Global trends 
2.3.1 Industrial development context 
The industrial revolution was enabled by automised production systems that 
had little to do, initially, with market balancing (i.e. the balancing of supply and 
demand) as initial demand outstripped supply. (Industrialisation meant many 
new products and also much more accessible products – due to cost 
rationalisation). The revolution was further fuelled by the two world wars and 
the production economy probably peaked (in terms of an ability to disregard 
demand-side issues) following Europe’s rebuilding after the end of the Second 
World War.  In 1946 Alford and Bangs described production control as: 
organization, planning, checking on materials, methods, tooling, and 
operation times, handling of routing, scheduling and dispatching, and 
coordination with inspection, so that the supply and movement of 
materials, operations of labour, utilization of machines, and related 
activities of factory departments – however subdivided – bring about the 
desired manufacturing results in terms of quantity, quality, time and 
place.
17
 
 
Demand-side economics is clearly not significant in Alford and Bangs’s 
description, which has a direct impact on how logistics systems were viewed 
(or not viewed) at the time when they described the theoretical principles of 
industrial management. At the same time logistics on a macro level is not 
taken into account, but rather treated as an administered cost, i.e. not 
providing a differentiation for companies, competing against each other, nor 
for countries competing globally. This points to the possible conclusion that 
the state of logistics and the acceptance, role and positioning of the function 
on a micro level relates to the macro-level structuring of the economic 
                                                 
17
 Alford and Bangs, 1946, p. 67. 
 15
platform from which it is launched. To test this conclusion a broad description 
or typology of economic platforms with an industrial development perspective 
could therefore better inform the current positioning of logistics and the global 
trends relating to this positioning. 
 
2.3.2 An underlying typology 
A discussion on global logistics trends should commence with a view of the 
different phases of economic platform development and the underlying 
physical support infrastructure of each phase, as depicted in abridged form in 
Figure 3.
18
  
 
 
 
 
 
 
 
 
 
 
 
 
 
Time
1
Industrial 
economy
2
Market 
economy
3
Electronic 
communities & 
trading platforms
4
Globally 
networked 
economy
5
Global bio-
engineering
6
Primary 
economy
Logi
st
i
c
s c
o
m
p
l
e
x
i
ty
High
Low
Figure 3: Abridged illustration of the growth in economic platforms 
 
The six phases depicted here refer to the development of local, national, 
regional and global economies over time. Primary economic development 
(characterised by land grab) is followed by industrial economic development 
(characterised by development of technology and registering of patents), 
which in turn develops into a more market-orientated economy (demand 
driven and service orientated).  With the development of electronic 
communities and electronic trading platforms, economies become more 
                                                 
18
 Havenga, 2005, p. 18. 
 16
globally networked and currently the world is entering a global bio-engineering 
phase on the back of these developments.  
 
A few important dimensions of this construct should be considered: 
� Movement along the axis differs for various parts of the world. In the last 
five decades the triad (USA, Germany, Japan) and their immediate 
geographical surrounds (North America, Europe and the Pacific Rim) 
developed faster along the curve, although catch-up by new entrants such 
as China and India is a specific feature of the last two decades. 
� Movement along the axis is fundamentally one-directional. It is difficult to 
catapult an economy into advanced phases without solid groundwork in 
earlier phases. 
� Entropy sets in, in earlier stages, as in all evolutionary systems. This 
requires continual revisiting of earlier building blocks. 
� Economies compete from different national and regional platforms, 
creating development disparities.  Attempts to address these disparities 
through world trade agreements, such as through GATT, are unfortunately 
often deficient. 
 
The underlying physical support architectures of the various economic 
platforms are depicted in Figure 4. 
 
 
 
 
 
 
 
 
 
 
 
 
 17
 
I
n
fr
a
s
tr
u
c
t
u
r
e
 
deve
l
opm
ent
P
r
oduc
t
i
on &
 
ope
rat
i
o
n
s 
m
anagem
ent
T
r
ans
po
r
t
 &
 
di
st
r
i
b
u
t
i
on 
(log
i
s
t
i
cs
)
Int
egr
ate
d
 
trad
i
ng 
pl
atf
o
r
m
s
G
l
o
ba
l
i
s
ed 
syst
em
s
Quest
 f
or
 
g
l
ob
a
l
 
“
ba
l
anc
i
n
g”
1
2
Market 
economy
3
Electronic 
communities & 
trading platforms
4
Globally 
networked 
economy
5
Global bio-
engineering
6
Industrial 
economy
Primary 
economy
Log
i
sti
cs
 compl
e
x
i
ty
Time
High
Low
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 4: Physical support architectures of economic platforms 
 
The physical support architectures describe the support for the underlying 
typology: 
� In the primary economic development era communities required 
infrastructure to reach new regions, often to exploit the mineral and 
agricultural wealth to be found in these new regions. Within defined 
boundaries, such as South Africa, the era became known as “hinterland 
development”. Infrastructure development to provide access to these 
areas, through roads, railways and harbours, played an important part in 
development and investment decisions. In some first world economies, 
such as Europe, this was preceded by maritime shipping technology and 
“hinterland development” was achieved much earlier than in South Africa. 
� In the industrial revolution era communities required manufacturing 
support systems usually enabled by efficiency of production lines. The 
growth in the manufacturing economy made more products available, but 
a prevailing push mindset often put the onus on the consumer to find 
suitable products. Power was in the hands of the industrialists. 
 18
� In the market economy era power shifted from production to consumption 
and most systems began a transformation process from push to pull. 
Transport and distribution became functions of the production process as 
consumers demanded final products closer to the point and time of 
consumption. 
� In the current electronic community and trading platform era, communities 
of consumers are more accessible through mega-portals (both virtual and 
bricks and mortar) and often enabled in both directions, i.e. virtual trading 
platforms required bricks and mortar logistics support infrastructure, but 
physical mega-portals required information systems enablement, through 
scheduling and demand forecasting. Suppliers in these areas realised the 
benefits of collaboration and even made use of third parties to engineer 
seamless delivery options. 
� In the globally networked economy these support systems become global 
and as a specific feature certain new global blocks are catching up to the 
initial triad. Remaining global imbalances restrain systems through 
renewed security concerns and the rerouting of basic commodities for food 
aid and primary health care. 
� In the global bio-engineering era the earth’s resources will have to be 
balanced for all communities. The anti-globalisation forum will gain support 
and the fringe demand for overall balancing of resources should become a 
mainstream phenomenon. 
 
2.3.3 Global trends 
A broad description or typology of economic platforms with an industrial 
development perspective has now been provided.  This enables an 
understanding of the current positioning of logistics and the global trends 
relating to this positioning.  Global trends have been, and will continue to be, 
one of the driving forces behind each of the economic platforms and is 
correlated to the underlying typology as is depicted in Figure 5. 
 
 19
In each of these eras the trends describe the basic mindset or point of 
departure of the logistics community. It also describes the key recurring theme 
that logistics professionals will or should revert to once entropy has set in. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Co
nf
i
g
ur
a
t
ion
 a
n
d 
r
e
co
nf
i
g
u
r
at
i
o
n
T
i
me
 &
 
mo
t
i
o
n
 
e
f
fi
c
i
enc
y
T
i
me
,
 mo
t
i
o
n
 
&
 
pl
ace
 u
t
i
l
i
t
y
Co
ll
ab
o
r
at
io
n &
 
in
t
e
g
r
a
t
i
o
n
I
n
fo
rm
at
i
o
n 
“
o
p
e
n acc
e
ss”
Quest for 
g
l
ob
a
l
 
susta
i
n
a
b
i
li
ty
In
f
r
a
s
tru
c
t
u
r
e
 
de
v
e
l
op
m
en
t
Pr
oduc
t
i
on
 &
 
oper
at
ions
 
m
anag
e
m
en
t
T
r
a
ns
por
t
 &
 
di
s
t
r
i
bu
t
i
on
 
(
l
o
g
i
sti
c
s)
In
t
e
g
r
a
te
d
 
tra
d
i
n
g
 
pl
a
t
f
or
m
s
G
l
ob
ali
s
ed 
sy
st
e
m
s
Qu
e
s
t fo
r
 
gl
ob
a
l
 
“
bal
an
c
i
ng”
1
2
Market 
economy
3
Electronic 
communities & 
trading platforms
4
Globally 
networked 
economy
5
Global bio-
engineering
6
Industrial 
economy
Primary 
economy
Logi
sti
c
s
 com
p
le
x
i
t
y
Time
High
Low
Figure 5: Global trends relating to the underlying typology 
 
� Configuration and reconfiguration. The issue of configuration and 
reconfiguration relates to the basics of logistics.  Yet, because of 
disparities in global development and entropy, this remains an issue and 
often requires new attention. The late Robert Delaney’s closing remarks in 
his final speech, when he presented the fourteenth state of logistics report, 
are almost prophetic in this regard: 
Let’s get our people right, our products right, and our production 
and distribution facilities network right and get our own house in 
order. Let’s RECONFIGURE our businesses operationally as a first 
step before we consider the benefits and risk of collaboration. In our 
experience, you should collaborate from a position of strength, not 
 20
weakness. We have to operate our way out of this challenge. We 
have to resist complaining about level playing fields. This is 
globalisation. There are no levelling playing fields. Recalling my 
favourite line from the award winning film Forrest Gump, ‘That’s all 
I’m going to say about that.’
19
 (Capitals in original.) 
 
When Delaney reverts to reconfiguration, he emphasises core structural 
problems that crept into systems within companies. This is, however, also 
true of national and regional economies, where the efficient development 
position that was evidently achieved is often lacking in some areas or has 
deteriorated over time. This could refer to people and processes on a 
micro level, but on a macro level it is often referred to as infrastructure 
development and investment and how to achieve a “perfect” balance for 
global competitiveness. 
 
The key recurring theme for this typology is therefore that core structural 
problems are increasingly identified in national logistics systems. Cases 
for reconfiguration on a global scale will emerge over the next ten years.   
 
� Time and motion efficiency. Time and motion efficiency was the initial 
stamping ground of experts such as Galbraith
20 
(the famous American with 
twelve children who transposed household issues of surviving in a 
household of 14 members to the field of time and motion study). The initial 
idea of time and motion study led to production line improvements, but 
rapidly spread to materials management and inventory control 
improvements. These improvements received a further boost in the last 
three decades with “just-in–time (JIT)” management, and companies such 
as FedEx and DHL built entire business models and value propositions 
around the concepts: 
One of the most tangible impacts of FedEx on business is the 
lowering of inventories and their associated carrying costs. In the 
                                                 
19
 Delaney and Wilson, 2003, p. 20-21. 
20
 Chase and Aquilano, 1981, pp. 16-19. 
 21
United States it is widely established that the inventory-to-sales 
ratio has experienced a steady decline in the past two decades as a 
result of advances in information technology and better logistics 
management. The overall U.S. inventory-sales ratio has fallen by 
over 20 percent over the past 18 years.
21
  
 
The double-edged sword of these improvements is the concomitant 
decrease in real interest rates, which accentuated improvements in 
materials handling, inventory management and the overall costs in the 
carrying of inventory. This, however, brought the rising cost of transport 
and transport inefficiencies in sharper focus. 
  
Little attention was given to distribution and very few logistics measures 
were developed. As Keebler and Durtsche points out, “supply chain power 
rested with manufacturers”.
22
 
 
The key recurring theme for this typology is therefore that “value chain 
efficiency” will in the next decade become a standard hygiene factor such 
as “production-line efficiency” has become over the past five decades. 
 
� Time and place utility. Time and place utility refer to the era when 
transport and distribution systems became more sophisticated, although 
the field was not referred to as “logistics”.  Gattorna and Walters refer to 
the start of this period “in the 1960s and early 1970s” when “physical 
distribution directors were very much an emerging breed.”
23
 The early 
1970s saw the emergence of cost trade-offs
24
 and the beginnings of what 
would later be known as “total cost of ownership”. Power slowly began 
moving down the value chain towards retailers
25
 and customer service 
became a competitive force. 
 
                                                 
21
 SRI International, 2004, p. 33. 
22
 Keebler and Durtsche, 1999, p. 47. 
23
 Gattorna and Walters, 1996, p. 2. 
24
 Gattorna and Walters, 1996, p. 3. 
25
 Keebler and Durtsche, 1999, p. 48. 
 22
This era’s sharp contrast in terms of distribution and customer focus is 
described in Wynne-Roberts’s short but well constructed treatise on 
industrial management, the end-to-end business process and its 
management just before the 1960s. Towards the end of his careful 
analysis of production management he simply refers to the “disposal of the 
finished product” and states that this should happen as fast as possible.
26
 
The arrival of the customer focus era began to change this as evidenced 
by Baily’s work in the 1960s where he talks about: 
The main task of any supply department is to supply the organization it 
serves with the goods and services required from outside the 
organization, in the right quality, at the right time, in the right quantity 
and at the right price. (Sometimes the right place must be added to the 
list…)
27
 
 
Baily initiated the contemplation of the place dimension and what its 
ramifications would be. As the understanding of the place dimension grew, 
manufacturers and producers began to understand the importance of 
distribution management. Customer demands became more sophisticated 
and undifferentiated long-distance solutions (often rail bound) were no 
longer adequate.  Lower utilisation of rail due to more tailored customer 
solutions led to a certain gradual demise in fixed, long-haul, heavy 
infrastructure, such as rail corridors, but at a cost that is still not 
understood in its widest context. 
…freight transport is likely to consume an increasing amount of energy 
and land, and it contributes to a wide range of problems such as air 
and noise emissions, congestion, traffic fatalities, etc. Social costs 
associated with road and air freight transport are reportedly much 
higher than those of rail and waterway freight modes.
28
  
 
Although more efficient in individual cases, the transport or place utility 
therefore tends to become more inefficient on a macro scale as 
                                                 
26
 Wynne-Roberts, 1956, p. 47. 
27
 Baily, 1969, p. 39. 
28
 Hesse and Rodrigue, 2003, p. 12. 
 23
collaboration and integration efforts are hampered by piecemeal 
optimisation.  The intrinsic costs could then be high, and rising, and in 
most cases extrinsic costs have not even been measured yet.  This is a 
clear case of where macro and micro approaches diverge and both should 
actually be considered. These challenges are usually addressed in a 
reactionary way, as reported by the Environment Directorate of the 
OECD
29
 Environment Policy Committee:
30
 
• Large-scale investments in highway systems; 
• Liberalisation of road freight over rail freight; 
• Vertical separation and open access on rail; 
• Lower real fuel prices; and 
• Increased purchasing power for private motor vehicles.  
 
Most of the configuration developments in basic transport infrastructure 
are therefore driven by factors outside the normal national cost 
consideration issues to the detriment of all users of the network. Reflecting 
on these developments over the past two decades, it is clear that the 
overall costs of logistics is declining, but that this decline is driven by 
factors other than basic transport infrastructure and the cost of transport. It 
seems, in fact, as if the costs of freight transport as a percentage of GDP 
might be rising and a case for national reconfiguration might be necessary. 
 
Some initial savings were realised in the US economy as deregulation 
improved the speed of inventory in transport, but most of the savings came 
from the reduction in inventory.
31
  Transport’s contribution to US GDP has 
declined over the past two decades (by 20%), but the decline in inventory 
carrying costs was triple that (60% decline).
32
 Even as long ago as in the 
early 1960s Heskett, Ivie and Glaskowsky reported that, in the decade 
between 1950 and 1960, transportation’s contribution declined by 6%, but 
                                                 
29
 Organisation for Economic Co-operation and Development. 
30
 Caid, 2003, p. 4-5. 
31
 Sutherland, 2003, p. 1. 
32
 Cooke, 2004, p. 2. 
 24
inventory carrying cost declined by 14%.
33
 In the last decade transport and 
storage’s contribution to the Australian GDP has risen from 4.7% to 5.1% 
(a rise of almost 10%).
34
  
 
The key recurring theme for this typology is therefore that both the intrinsic 
and the extrinsic costs of transport and distribution systems will become 
more transparent (as nations adjust to this problem) and will lead to new 
policy and regulatory constructs.  
 
� Collaboration and integration. As the power in value chains moved 
steadily down the chains towards final consumers, upstream players 
experienced margin squeezes. This phenomenon also holds true for 
supply chain practices. Dong et. al. report that buyers often transfer 
inventory costs to sellers and that relationships need to be formed to 
support both buyers and sellers: 
In order to build a long term supply relationship, it is likely that all 
members of the supply chain need to profit. Our exploratory research 
shows that JIT purchasing produces direct positive logistics results only 
for buyers, suggesting that long term JIT purchasing relationships may 
not necessarily be stable. In order to implement a successful JIT 
purchasing program from a supply chain perspective, managers in both 
the supplier and buyer organizations must act to produce the 
conditions conducive to JIT adoption and success for both buyers and 
sellers.
35
 
 
Collaboration and integration were pervasive themes during the past 
decade, as the industry progressed from logistics to supply-chain 
perspectives. Collaboration refers to the propensity to share common 
objectives, plan jointly and measure from the same platform. Integration 
refers to the propensity to integrate assets, processes and systems 
seamlessly. There is, however, a marked difference between what was 
                                                 
33
 Heskett et al. 1964, p. 15. 
34
 Australian Government, Transport Statistics, 2004, p. 1-2. 
35
 Dong et al., 2001, p. 479. 
 25
intended and what was achieved, i.e. the focus is often short term, and 
secrecy and cut-throat competition continue to play a role. As sustainable 
development as a core approach grows, this should change. 
 
Gattorna and Walters refer to the past decade as the time when 
information became a means to achieve competitive advantage, as 
opposed to the previous era (the 1960s to the 1980s), when information 
was purely a necessity.
36
 This role of information was both driven by and 
drove the rise of electronic information capacity and the emergence of 
electronic data interchange systems. 
 
The key recurring theme for this typology is therefore that collaboration 
and integration will move from JIT purchasing strategies to shared 
information and infrastructure, between buyers, sellers and logistics 
support service providers. 
 
� Information “open access”. The information “open access” era is driven 
by the arrival of globalisation in supply chains. Cohen and Roussel 
describe the two decades preceding the turn of the twentieth century as a 
time when “each site planned and scheduled its own operation with a 
focus on optimizing local results.”
37
 They emphasise the contribution of the 
growing complexity of business and added efficiency required by a global 
focus towards the end of the nineties to the arrival of this era. Christopher 
maintains that the management of global logistics “is in reality the 
management of information flows.”
38 
This is achieved through a balance of 
a local and a global focus where, for instance, local customer service is 
balanced by global network structuring and local market intelligence with 
global information systems.
39
 
 
Gollwitzer and Karl initiate their discussion on logistics on the basis of the 
role of the comptroller and how this role was changed and enriched by 
                                                 
36
 Gattorna and Walters, 1996, p. 4. 
37
 Cohen and Roussel, 2004, p. 3. 
38
 Christopher, 1998, p. 144. 
39
 Christopher, 1998, p. 145. 
 26
logistics. They demonstrate how logistics in the current era is changing 
into a management function with three basic cornerstones, i.e. co-
ordination, information and innovation. This required a process view of 
logistics and a process view is not possible without information which, 
according to them, changed to something much more than simply 
monetary information.
40
 
 
Ayers describes the trend of “open access” information as a cultural rather 
than a technical trend.
41
 He believes that supply-chain information sharing 
will be progressive as global relationships develop. In a sidebar on “A 
meeting of networked minds” he claims: 
In a networked economy, the company that has a distinct competitive 
advantage is the one whose supply chain has the largest span or 
presence – all the way from suppliers to the end consumers.
42
 
 
The competitive advantage is clearly not related to concealing information, 
but rather sharing from a platform of strength to achieve the greatest 
visibility through the chain. Handfield and Nichols describe how this 
visibility has become inter-organisational, powerful and strategic and 
“much more than a simple technology.”
43
 
 
The onset of the globally networked economy gives rise to new 
imperatives: 
While the fundamentals of logistics service are frequently addressed, 
understanding the application of logistics service strategies in a global 
context is still in the early stages of development. It is the role of future 
research to continue to investigate the influence of LSQ (Logistics 
Service Quality) in a global context.
44
  
 
                                                 
40
 Gollwitzer and Karl, 1998, p. 20-21. 
41
 Ayers, 2004, p. 2. 
42
 Ayers, 2004, p. 2. 
43
 Handfield and Nichols, 2002, pp. 296-334. 
44
 Mentzer, 2004, p. 10. 
 27
These imperatives will go beyond customer satisfaction to the 
understanding of value and costs relating to national and regional 
characteristics (Figure 6). Value and perceived value are always products 
of costs and service rating.
45
  It is therefore inevitable that a specific trend 
for the next decade will be the measurement of national and regional 
logistics costs as a percentage of GDP. The OECD has already estimated 
global logistics costs between 11-16% of world GDP.
46
  (This figure will 
probably become firmer over the next few years and country benchmarks 
against this figure should emerge.) 
 
National & 
regional 
characteristics
Organisational 
characteristics
Perception of 
order placement 
activities
Perception of 
order receipt
Satisfaction level 
response
 
 
 
 
 
 
 
 
 
 
 
Figure 6: Factors influencing customers' preference for global logistics services
47
 
 
The key recurring themes for this typology are therefore that: 
• logistics information “through the chain” will become more transparent; 
• the real winners will be supply-chain captains with more transparent 
and visible information systems from beginning to end; and 
• national logistics costs will be measured in more economies and 
managed as a lead and lag indicator. 
                                                 
45
 Bradley Gale (1994) developed this useful concept. which he calls “a plot of the prices and 
overall performance scores of competing offerings". The map contains reference lines for 
assessing customer-perceived fair value and the frontier prices that represent the best 
relative values available to the buyers in the category.  
46
 Australian Government, Transport Directions, 2004, p. 3. 
47
 Mentzer, 2003, p. 18. 
 28
� Quest for global sustainability. Global sustainability relates to the way in 
which global communities interact, and to the distinctions between these 
communities and the emerging new world order. Not the least of the 
derived issues from this trend are security threats and natural disasters, 
and the impact of these phenomena on logistics.
48 
Fawcett et al. highlight 
the various processes in which logistics are involved that have an effect on 
the environment, but single out three core concepts for consideration, i.e. 
ignorance of future damage by current actions, absence of direct “polluter” 
charges and the mindset of present gains for future costs.
49
 
 
A decrease in transport costs, globally, can also contribute to global 
upliftment. In a technical study to determine the quantitative role of 
transport in international business cycles, Ravin and Mazzenga found 
statistical proof for an interesting phenomenon: 
The welfare effects of changes in costs of transportation are more 
intriguing. We find large welfare effects: a drop in the costs of 
transportation from 20 percent to 15 percent is equivalent to a 
permanent increase in consumption of just above 1.5 percent.
50
  
 
The key recurring themes for this typology are therefore that global 
sustainability issues will drive logistics input cost decisions by the second 
decade of the 21st century, and that global security will play an important role 
in the logistics decisions over the next decade. 
 
2.4 The state of global macro-logistics measurement 
The case for understanding the cost and impact of global logistics is regarded 
as non-negotiable by benchmark researchers and authors of the discipline. It 
simply must be measured.  Bowersox and Closs, in fact, refer to a difference 
                                                 
48
 Sutherland, 2003, p. 2. 
49
 Fawcett et al., 1992, p. 286. 
50
 Ravin and Mazzenga, 2004, p. 657. 
 29
between “importance” in the micro field compared to “critical for global 
manufacturing and marketing”
51
 (Italics in original.) 
 
In terms of global measurement three aspects are receiving growing attention, 
i.e. typological differences, infrastructure requirements and the performance 
of logistics in terms of cost to the economy. 
 
2.4.1 Typological differences 
It is true that broad-based typological differences between logistics costs of 
various economies can be hypothesised: 
In more highly developed countries, such as Japan, Canada, the United 
States, and most countries in Western Europe, the supply chain and 
logistics systems are highly sophisticated. A firm entering those countries 
will find economies with good transportation systems, high-technology 
warehousing, skilled workers, and a variety of logistics support 
systems.
52
  
 
But is this broad-based understanding sufficient. Bowersox and Closs 
continue the argument by describing “vision” differences between North 
American, European and Pacific Rim perspectives. (The North American 
perspective is one of open geography with extensive land-based 
transportation needs, but relatively limited cross-border or international 
documentation requirements. Europe’s geography is much more compact, 
with greater diversity and regulatory barriers. In the Pacific Rim the focus is 
much more on water logistics management).
53
   
 
What is the overall typological perspective for South Africa? Conjectures in 
this regard are made, but are not based on any solid existing research results. 
The conjectures would revolve around the sprawling spatial nature of the 
country’s economy, just like the economy of the United States, but with 
                                                 
51
 Bowersox and Closs, 1996, p. 126. 
52
 Stock and Lambert, 2001, p. 516. 
53
 Bowersox and Closs, 1996, p. 127. 
 30
unfortunately a much lower gross domestic product in relation to these 
distances. A model solution would be to determine the flow volumes in 
general for the economy and relate these to GDP, in order to determine the 
size and depth of this typological difference. 
 
The intent of this dissertation is to develop a national flow model that can 
inform the development of such typologies. 
 
2.4.2 Infrastructure requirements 
In terms of infrastructure investments Lakshmanan and Anderson address 
this issue as follows: 
An important dimension of these economic impacts relates to the 
impacts of transport investments on the performance of the rapidly 
evolving freight services sector, whose improved productivity in turn 
enhances the productivity of the overall economy.
54
 
 
Productivity is therefore contextualised in the specific situation of the 
sprawling nature of the US economy (with long corridors) and the effects of 
this on infrastructure requirements and economic performance are 
understood. 
 
Demis and Tavasszy discuss logistics infrastructure in Europe, Asia-Pacific 
countries and North America, elaborating on “companies, industries and 
governments” and their different objectives in this regard. They summarise 
four objectives of measurement performance indicators for the macro 
performance of logistics infrastructure: 
• To establish a holistic view of the system; 
• To give feedback in order to initiate a new and better way to conduct 
and handle the measurement system; 
• To clarify the aim and goal for all participants in the system; 
                                                 
54
 Lakshmanan and Anderson, 2002, p. 3. 
 31
• To indicate overall important policy actions.
55
 
 
As far as South Africa’s logistics infrastructure is concerned, the best way to 
inform investment would be to determine the demand for the flow of all 
commodities now and in the future, so that infrastructure investments and the 
trade-offs between various investment options could be considered.  
 
2.4.3 Performance of logistics as a cost to the economy 
As far as the performance of logistics as a cost to the economy is concerned, 
the best known macro-measurement of logistics costs globally is the annual 
“State of Logistics Report” for the US economy, initiated by the late Robert 
Delaney and sponsored by Cass Information Systems and ProLogis. The 
report, according to Cooke,
56
 uses the Alford-Bangs formula to calculate the 
total value of logistics as the sum of three components: inventory carrying 
costs, transportation costs and administrative costs. The report actually uses 
a variety of methodologies, outlined by Robert Delaney in a 1987 policy 
position paper (three years before the publication of the first annual state of 
logistics report).
57
 Alford-Bangs’s approach is used to calculate inventory 
carrying charges
58
 and Frank A. Smith’s approach to calculate transportation 
costs. 
 
The 2003 report also quotes Heskett, Ivie and Glaskowsky as a methodology 
source, but the Alford-Bangs “standard” rate for inventory carrying charges is 
quite evident in the results of the reports over time. These standard rates are 
reflected in Table 1. 
 
 
 
 
 
                                                 
55
 Demkes and Tavasszy, 2000, p. 4-5. 
56
 Cooke, 2004. 
57
 Delaney, 1987. 
58
 Alford and Bangs, 1946, p. 396-397. 
 32
Item Percentage 
Storage facilities 0.25 
Insurance 0.25 
Taxes 0.50 
Transportation 0.50 
Handling & distribution 2.50 
Depreciation 5.00 
Interest 6.00 
Obsolescence 10.00 
Total 25.00 
 
Table 1: Alford-Bangs's standard inventory carrying charges table
59
 
 
What is mostly adjusted is the interest rate, according to Delaney himself: 
…but when comparing annual inventory carrying costs over time, it is 
necessary to replace the assumed 6 percent interest rate with the 
prevailing commercial paper rate, if it is higher. In 1981, for example 
when the average interest rate on commercial paper rose to 15.3 
percent, most companies’ carrying costs climbed to 34.3 percent of the 
market value of the inventories held in stockpiles and warehouses.
60
  
 
It is then also quite obvious that, when Delaney reports on the inventory 
carrying rate in 1981 for the US economy, he uses 34.7%. This figure is the 
original Alford-Bangs formula of 25% minus 6% plus 15.7%, which was close 
to the average interest rate for 1981. It therefore seems as if the other 
components of the Alford Bangs model are still used as it was developed 61 
years ago. 
 
The performance of logistics as a cost to the economy has by now been 
calculated for a number of countries and is depicted in Table 2. It is becoming 
standard practice in many of these economies to do the calculation annually 
and report on the findings. 
 
                                                 
59
 Alford and Bangs, 1946, p. 397. 
60
 Delaney, 1987, p. 3. 
 33
Region Country Logistics as % of GDP 
Australia 11.2 
Chin 14.5 
Hong Kong 15.9 
Indonesia 12.8 
Japan 10.1 
Korea 12.4 
Philippines 13.3 
Singapore 20.0 
Taiwan 13.5 
Asian region 
Asian total 11.0 
Austria 12.2 
Belgium 13.0 
Denmark 12.8 
Finland 11.5 
France 11.7 
Germany 11.8 
Greece 14.0 
Iceland 16.7 
Ireland 14.3 
Italy 12.6 
Netherlands 12.2 
Norway 13.2 
Portugal 13.6 
Spain 12.1 
Sweden 12.7 
Switzerland 13.2 
United Kingdom 12.2 
European region 
European total 12.2 
Canada 11.8 
Mexico 14.4 
United States 11.6 
North American 
region 
North American total 11.7 
Industrial total 11.7 
 
Table 2: Logistics cost comparisons for various economies
61
 
 
                                                 
61
 Bowersox and Closs, 1996, p. 128. 
 34
It would be possible for the South African economy, with a transportable GDP 
hypothesised to be less than one billion tons, to develop a much more robust 
model with fewer generalisations around inventory carrying cost (including 
interest charges and warehousing), administrative costs and transportation.   
 
A discussion on the global state of macro-logistics measurement leads to a 
more specific question of the position in South Africa. 
  
2.5 The South African situation 
Various events over the past two decades have highlighted core structural 
issues, problems and competitive disadvantages in South Africa’s logistics 
system. These include: 
• The results of the commercialisation, corporatisation and legal 
succession of the South African Transport Services in 1990 (which led 
to critical underinvestment in important areas causing an illogical 
fragmentation of assets, processes and systems); 
• The increasing scarcity of information; 
• The RDP policy statement in 1994, which indicated the social 
shortcomings of logistics assets; and  
• The Moving South Africa Study of 1998 (the most fundamental 
research project so far), which highlighted the possible future 
shortcomings of the system. 
 
2.5.1 Transnet events 
In 2004 the then CEO of Spoornet, Ms Mokgatle, referred to the situation as 
“severe”, adding that “Spoornet’s
62
 performance has deteriorated following 
years of neglect” and that “Spoornet’s long-term sustainability is in 
question”.
63
  Ms Ramos, the current CE of Transnet, talks about a business 
                                                 
62
 A division of Transnet. “Spoornet” refers to the business division of Transnet that is now 
known as “Transnet Freight Rail”, but was known as Spoornet when most of the research 
in the dissertation was conducted. 
63
 Spoornet Media Statement, 2004, p. 1. 
 35
that “has a long way to go in cleaning up its own house to become more 
efficient and accountable”.
64
  She also refers to the group’s assets as some of 
“SA’s most strategic assets”
65
 and the drive to restructure these. Singh
66
 says 
in the same article that Ms Ramos is in an unenviable position. She is dealing 
with a rail network, suffering from years of underinvestment, poor 
management systems and a skills shortage: “if Transnet does not get its 
house in order, it will not only have a continued negative impact on local 
business, but will also affect the New Partnership for Africa’s Development.” 
 
More recently, South Africa’s transport minister, Mr Radebe, referred to South 
Africa’s transport system’s performance as “poor” and a matter of great 
“concern”. He emphasised that “it was necessary to restructure the transport 
system generally to make sure that logistics, or the lack of it, did not act as a 
restraint on economic growth, employment and sustainable development”.
67
  
 
Yet none of these statements are based on real macro-economic research. 
No framework exists to measure the logistics performance of the country’s 
economy compared to internationally acceptable parameters. 
 
2.5.2 The increasing scarcity of information 
Another perspective on transportation cost performance and export 
competitiveness is provided by the World Trade Organisation, which 
determined that “the effective rate of protection provided by transport costs is 
in many cases higher than that provided by tariffs.”
68
 This statement is both 
ironic and painful in the South African situation, where tariffs and tariff 
protection are major topics of discussion and research, in spite of the fact that 
unnecessarily high transportation costs are probably a much more costly 
problem to the economy. Many databases, such as those of the South African 
Revenue Services and the National Ports Authority, have been established to 
                                                 
64
 Financial Mail, 2 July 2004, p. 30. 
65
 Financial Mail, 2 July 2004, p. 30. 
66
 Financial Mail, 2 July 2004, p. 30. 
67
 Star Business Report, 13 July 2004, p. 5. 
68
 World Trade Organisation, 2004, p. 114. 
 36
enable tariff research or adhere to some international conventions,
69
 often to 
the detriment of transportation research. The mere fact that tariffs are 
transparent leads to the self-fulfilling prophecy that this aspect is managed in 
detail, in spite of the fact that other issues might be more onerous.  
 
Information in general in South Africa is scarce and the railways and road 
hauliers do not disclose any useful information easily. It is true that the modes 
are locked in competition, but the actual competition is for an ever-decreasing 
competitive landscape as the railways sheds customers and concentrates on 
long-haul bulk “rail-locked” clients to the detriment of the national economy. 
Sharing of information could support the development of complementary 
systems, which is unfortunately not the case. 
 
South African society and business live in an economy where macro-
indicators such as the inflation rate, the balance of payments, the exchange 
rate and the interest rate are tracked closely, and in some cases even 
reported daily. Macro and micro decisions are made based on these 
indicators, but the country fails to make a concerted effort to understand 
logistics cost, which, by all indications, could account for more than 10% (and 
according to some broader assumptions more than 15%) of the country’s 
GDP. 
 
2.5.3 Social factors 
South Africa’s President refers to “a government of the left” that “includes a 
reduction in inequality, the provision of public services, the principle that 
workers should be treated as assets, rather than commodities, regulation of 
enterprise...”.
70
 Such ideals should, however, be measured and tracked, and 
the performance of various aspects of the economy, including the logistics 
system, monitored to see if these ideals are being achieved.  
                                                 
69
 Many of these databases are used by SARS and the Department of Trade and Industry to 
collect information on imports and exports to enable trade statistics that specifically 
enable tariff policy and reporting on National Accounts. It is not clear what other 
objectives are served by SAPO information, other than collecting Ad Valorem tax. 
70
 Cape Times, 1 July 2004, p. 6. 
 37
Moving South Africa, which embarked on a comprehensive data-driven 
approach,
71
 found during their study that this objective is often hampered by a 
chronic lack of information. The situation is, however, at its worst for rural 
communities (second economy),
72
 where the problem is described as 
“crippling.”
73
 Finally, the study notes with urgency that: 
The MSA strategy offers two responses to this situation.  One is to 
develop more data, and the second is to apply the general strategic 
principles developed in other parts of the strategy to the rural situation 
once the data are developed.
74
  
 
The development goals for South Africa around social upliftment are clear and 
unambiguous.  However, strategies, the lead data to inform the development 
of these strategies and the lag data to track its performance, are mostly 
absent. 
 
A recent statement of the Presidency of South Africa proposes: 
Monitoring and evaluation is the life-blood of sound and efficient 
planning and implementation. For it to add value to government work 
and to the broader process of social transformation, it should be based 
on objective measurements that reflect the ideals in our Constitution: to 
improve the quality of life of all South Africans and ensure that South 
Africa contributes to the creation of a better Africa and a better world. 
Cabinet has approved a set of key development indicators to provide 
evidence-based pointers to the evolution of our society. Based in part 
on Ten Year Review's human development indicators, they are 
markers that help define the milestones in the journey of social 
change.
75
 
 
Out of 72 indicators that the presidency lists, a very noble undertaking, not 
one refers to anything vaguely related to transport and logistics. 
                                                 
71
 Moving South Africa, 1998, p. 3, 5. 
72
 Moving South Africa, 1998, p. 11. 
73
 Moving South Africa, 1998, p. 178. 
74
 Moving South Africa, 1998, p. 179. 
75
 The Presidency of South Africa, 2007. 
 38
2.5.4 The quest for measurement 
The events described in paragraphs 2.5.1 to 2.5.3 above necessitate an 
understanding of what has been achieved in terms of logistics research and 
measurement in South Africa.  
 
The Transport White Paper (TWP) of 1997 and the fourteen-month Moving 
South Africa (MSA) research project that followed
76
 were designed to 
articulate strategies for the long-term development of South Africa’s transport 
and logistics infrastructure. The vision, in terms of the TWP, was for “efficient 
and fully integrated transport operations and infrastructure”.  MSA “was 
designed to produce a data-driven programme for strategic action” to enable 
this vision.
77
  The following specific challenges were highlighted during these 
studies: 
• Old priorities were still reflected in operational planning; 
• The overall system received systematic under-investment; 
• Broad prioritisation avoided focussing resources on specific 
customers; 
• Highly concentrated corridors have to carry freight from and to dense 
industrial locations, which requires intermodal systems that are absent 
to date; and 
• High prices, poor service levels and low reliability indicated a lack of 
support for export competitiveness and system sustainability on these 
corridors. 
MSA indicated that solutions to many of these problems could be facilitated by 
research-based regulatory involvement around infrastructure development. 
The study itself provides some measurement results and asks for it to be 
repeated at regular intervals. This never happened. 
 
Some policy guidelines were developed by the Department of Transport to 
address these issues, but now, almost a decade later, very few actual results 
have been realised. In the historical literature survey in Chapter 4 this element 
                                                 
76
 Moving South Africa, 1998. 
77
 Moving South Africa, 1998. 
 39
of the research will be developed further, but for now suffice it to say that the 
current situation was predicted, outlined and well formulated as a possible 
outcome about a decade ago. Furthermore, MSA was quite clear that the 
absence of research will seriously hamper future strategic management of the 
country’s transport and logistics system.  
 
Many government departments, such as the Department of Transport and the 
Department of Trade and Industry, are to some extent investigating strategies, 
policies and action plans to address this problem. Progress is slow, however, 
and erratic in nature. Transnet is also busy with a drive to inform national 
strategies and priorities, and to align Transnet’s transport and logistics 
strategy, market positioning, resource deployment and investment with these 
strategies. The dilemma is that very little discussion and alignment take place 
between the various stakeholders in this industry, e.g. government, parastatal 
organisations (such as Transnet and its rail, port and pipeline divisions), the 
national roads agency and road operators to align this research around a 
common platform. 
 
A contributing factor to this dilemma is that there is very little data and 
interpretation, and few frameworks, to provide a backdrop to these strategies, 
resulting in intermittent and disjointed research informing the discussions.  In 
fact, there is reason to believe that current initiatives will once again be 
hampered, delayed and even restricted, as so often in the past, because of 
the shortage of a strategic information system for logistics in South Africa. The 
country could consequently be in an even worse position a decade from now.  
The strategic research should therefore not only provide a framework for 
strategy and policy formulation, but also for shared and aligned points of 
departure for the disparate parties that have to contribute to this discussion. 
 
The conclusion is that, although the development of micro-logistics costs 
measurements is more pervasive in standard reference material, only a few 
references and measurements for macro-logistics costs are found in the 
 40
literature
78
 and some work in this regard is performed globally. In South 
Africa, however, no real attempt is being made to measure the cost, efficiency 
and effectiveness of South Africa’s logistics system. The capital investment in 
the county’s logistics infrastructure is also not guided by any measurement, 
research construct or model that could be monitored. Schneidewind maintains 
that: 
the established cost management system has direct effects on 
management decisions in supply chains.
79
 
 
Macro-management decisions will as a result be hampered by this state of 
affairs.  The absence of any cost management system on a macro level will 
also seriously impede national policy in this regard.  This is confirmed by 
Fredenhall and Hill: 
 As the logistics functions become more integrated, they are able to 
achieve many efficiencies. But, a barrier to fully implementing an 
integrated logistics function is the lack of accurate information about 
costs.
80
  
 
2.6 Proposed approach 
2.6.1 Methodology 
The overarching approach of this dissertation is quantitative and is based on 
economic modelling. Economic data are available in many forms and in many 
areas of the South African economy. The literature study of historical 
approaches will illustrate the failure of the survey methodology to solve the 
measurement problem. A modelling approach that works and could be tested 
over time for intrinsic and extrinsic congruity could provide a lasting solution to 
this problem.  
 
                                                 
78
 Voortman, 2004, p. 13. 
79
 Schneidewind, 2003, p. 92. 
80
 Fredendall and Hill, 2001, p. 213. 
 41
The array of data available to do this modelling is extensive, but success can 
only be achieved through the design process that will transform the applicable 
data sets, identify the underlying and often hidden relationships and integrate 
the data to answer the fundamental questions of size, flow, modal market 
share and future demand.     
 
The dissertation will describe the modelling of national surface freight 
transport flows, measuring the demand for freight infrastructure on a 
commodity or industry sub-sector level, and developing an understanding of 
logistics costs in South Africa. In order to achieve this, three models and the 
interpretation thereof will be proposed. Each of these models achieves a 
break-through in own right, but an integration case study of all three models 
will also be developed to illustrate the powerful strategic application of this 
work.  Strategic themes will be developed and recommendations made based 
on the results of the work as well. The models’ final test would be the 
repeatability of the construct on an annual basis, which will ensure application 
as both a lead and lag indicator of national logistics decision making and 
performance.  
 
2.6.2 The transport flow model 
The transport flow model is needed to answer the core question of modal 
market share in South Africa (an issue that has been irregularly and 
unsatisfactorily researched in the past), modal corridor flows (an issue that 
has been addressed in a few research works, but never resolved) and the 
core configuration of South Africa’s transport system. 
 
The transport flow model will utilise the South African Traffic Count Yearbook 
to develop a current and historical view of all freight traffic flows in South 
Africa. The model will account for the differences between corridor, rural and 
metropolitan freight, and the various net and average carrying capacities of 
the types of trucks that are used, and develop measurements for 1993, 1997, 
2003 and 2004. 
 42
 43
This information will then be collated and compared to actual Spoornet data 
for the same time periods to develop an insight into market shares, corridor 
densities and overall investment strategies for South Africa.  
 
2.6.3 The commodity flow model 
Different commodities have different logistics requirements.  It is therefore 
necessary to model the flow of all commodities in the economy – now and in 
the future – to enable planning for future requirements. 
 
The model will use a gravity modelling approach based on magisterial district 
supply and demand figures for South Africa that will be researched and 
developed from input-output models of the economy. Data will be developed 
for 2004. 
 
2.6.4 The logistics cost model 
The logistics cost model will be scoped and the measurement developed 
together with a research team from the Department of Logistics at the 
University of Stellenbosch.  The model will be used to identify transport, 
storage, inventory carrying, and management and administration costs for the 
South African economy from 2003 to 2005, and will be compared to the GDP. 
The measurement will use a “bottom-up” approach and will allow for detailed 
differences between industries for various cost drivers. 
 
2.7 Economic and historical base for continuance 
The problem statement refers to logistics’ macro-economic role and the lack 
of research in this regard in the history of the economy in South Africa. These 
statements require validation and explanation in this dissertation. It is 
therefore important to research the concept of logistics in a macro-economic 
context further.  This issue is addressed in Chapter 3. 
 
Chapter 3 - Macro-economic literature study 
3.1 Introduction and background 
Logistics is distinctive amongst oranisational functions as an integrator. It is 
process orientated internally, but also externally, in that it endeavours to 
integrate all organisational processes in a value chain view – even beyond the 
boundaries of a single organisational entity.  As such, it should be well suited 
to play a role in the integration of macro-economic production factors.  In 
order to fulfil this role, however, the discipline needs to come to grips with this 
macro-economic context. It also needs to develop, through introspection, an 
understanding for the reasons why the macro-economic context is so often 
neglected. 
 
3.2 The quest for a macro context 
3.2.1 The macro context of science 
One of the fundamental points of departure of this research is the distinction 
between micro and macro perspectives. This distinction is more than merely 
approaching a problem from two angles, in layman’s terms often referred to 
as “bottom-up” vs. “top-down” or in scientific terms “reductionism” in process, 
discovery or experimentation vs. integrative, synthesised and holistic thinking.  
It is in fact this incongruity that cuts through all scientific thought, since the 
overwhelming energy in scientific discovery tends to gravitate towards the 
micro or reductionist view. Jack Cohen and Ian Stewart plead for the macro or 
integrated view: 
Despite the success of conventional science, we think there should be 
more to the scientific endeavour than just the study of ever more refined 
internal bits and pieces.
81
 
 
                                                 
81
 Cohen and Stewart, 2000, p. 5. 
 44
This plea, from a well-known reproductive biologist (Jack Cohen) and one of 
the most active research mathematicians (Ian Stewart) of the current era, is 
made because they observe the increase and growth of reductionist thinking 
in all sciences.  This could in fact point towards a Zeitgeist of the sciences.  
 
Dodig-Cmkovic describes how the scientific process of antiquity has at its 
heart thought experimentation and holism rather than the practical 
experimentation of contemporary science and its reductionist nature. 
82
 
 
At the initiation of the experimental phase of science this holistic integration 
was well retained. John Brockman argues the point from the perspective of an 
“official culture”, which he terms the “humanities”, that abandoned science 
and technology after the fifteenth century.
83
 He refers to the thinkers of the 
fifteenth century, when the word “humanism” was tied in with the idea of one 
intellectual whole:
84
  
These men were intellectually holistic giants. To them, the idea of 
embracing humanism while remaining ignorant of the latest scientific and 
technological achievements would have been incomprehensible. The time 
has come to re-establish that holistic definition.
85
 
Brockman’s wish for holism is, however, still a problem and this is why Murray 
Gell-Mann relates the problem to management science and asks for a move 
away from “narrow discipline” to “integrative” thinking.”
86
 
 
Also within sciences this disconnect between the holistic and reductionist view 
becomes apparent. In physics a stand-off emerged between the practical 
experimentalists and theoretical modellers, with insufficient integration 
between the two. In economics the place of both macro-economics and micro 
economics is firmly established. But is this true of the sub-disciplines of 
economics?  
 
                                                 
82
 Dodig-Cmkovic, 2001, p. 11-13 
83
 Brockman, 2003, p. 3. 
84
 Brockman, 2003, p. 2. 
85
 Brockman, 2003, p. 2. 
86
 Gell-Mann, 1994, p. 346 
 45
Many of the operational applications of fields within economics seem to be by 
definition micro disciplines, i.e. relating to micro economics as such. 
Management science, for instance, might be seen as a study of an integrated 
or co-ordinated set of activities within economics. Within management science 
the more traditional fields of production, finance and marketing arose, more or 
less in that order. In this group business logistics is relatively new.
87
 The 
question arises, however, as to what extent business logistics should reach 
back into economics to establish the macro-economic parameters of the 
discipline, and whether this is the role of the business logistics researcher or 
economics researcher – the question specifically addressed by this 
dissertation. 
 
In terms of management process and development, logistics professionals are 
contributing to ever higher levels of decision making, as large corporations 
appoint senior executives to head the logistics function, which is often created 
by: 
pulling the sundry logistics functions out of marketing, operations, and 
elsewhere and placing them under a director or vice president of 
logistics.
88
 
 
The question that should be asked is “Who is responsible for this 
development and growth of the logistics function and all its applications (both 
micro and macro)”? In many respects the discipline, through research and 
development as well as informal and formal activities (such as the work of the 
Supply Chain Council), took charge of its own development and positioning. 
The assumption that the logistics discipline should at least be partially 
responsible for its own positioning in macro-economic thought could therefore 
hold true and should be considered. In this line of reasoning a more specific 
treatment of macro-economics and the role of logistics becomes necessary. 
                                                 
87
 Ballou, 2004, p. 3. 
88
 Russel, 1994, p. 2. 
 46
3.2.2 Macro-economics 
Macro-economics accounts for the economy of a country as a whole and 
informs the overall strategic management of the country’s economy in terms 
of the contribution of holistic components as well as the interaction between 
these components. This is achieved through measuring, modelling, reporting 
and strategising around these components as well as the whole, normally 
aggregated as holistic benchmarks such as gross national product (GNP) and 
gross domestic product (GDP). In short: 
[it] is the study of the behaviour of the economy as a whole.
89
  
 
Mohr and Fourie
90
 continue by setting specific goals for the macro-economy, 
i.e. growth, full employment, price stability, external (balance of payments) 
stability and fair income distribution.  In order to achieve these objectives the 
components or “production factors” of the economy need to achieve optimal 
interaction and output, very much in the same way as in which the production 
factors of a firm (on the micro level) is expected to achieve optimal interaction 
and output. 
 
In the South African economy “cluster studies” were initiated by the 
Department of Trade and Industry to evaluate and consider the contribution, 
effectiveness and prospects of various economic clusters in South Africa. 
During the course of these studies, and in some sense almost by chance, the 
researchers realised that certain cross-cutting issues emerged, of which 
logistics emerged as an important factor. These cross-cutting issues do in fact 
relate to the production factors that economists define in macro-economy, but 
in a different way as the defined standard, and it seems that macro-economic 
theory neglected to follow these emerging issues as in micro-economy, where 
the understanding of production management has expanded greatly over the 
last few decades. The macro-economy is therefore not necessarily equipped 
to use micro-economic principles in economic management and it is therefore 
no surprise that in spite of the pervasive logistics problems identified, no effort 
                                                 
89
 Samuelson and Nordhaus, 1989, p. 76. 
90
 Mohr and Fourie, 1996, p. 95. 
 47
was made to develop a macro-economic view of logistics or at least the role of 
logistics in the macro-economy. 
 
The fact is that the distribution of resources in the economy, and therefore 
macro logistics, is a macro-economic production factor and requires 
measurement together with issues such as employment, crime, the production 
price index (PPI) and investment.  The lack of macro-economic thought in 
logistics leads to a concomitant gap in macro-measurement thinking. 
 
3.2.3 Macro-measurement 
Textbooks on transportation, distribution, logistics management and supply 
chain management are usually completely or partially silent on macro-logistics 
cost issues.  Cursory references to transport policy, deregulation and the 
effects of modal shifts do occur, but these topics are seldom analysed in 
depth and the macro-economic results of these structural changes are hardly 
ever discussed. Schary and Skjott-Larsen
91
 provide a reasonable analysis for 
Europe and discuss the conflicts between national interests and EU policy, 
infrastructure investment decisions in this context and the significance of 
supply and demand against this background. The fact that a similar conflict of 
interest could arise between firm, industry and national interests, is not 
addressed.  These relationships are depicted in Figure 7.   
 
 
 
 
 
 
 
 
 
 
                                                 
91
 Schary and Skjott-Larsen, 2001, pp. 220-223. 
 48
 
Infrastructure, Investment 
Development and Usage
>
European
Interests
Germ
a
n
y
Fra
n
ce
Et
c
…
A. Economic community
B
Countries of the 
European community
Infrastructure, Investment 
Development and Usage
>
VW BMW Et
c…
D. Industry
Companies in 
Auto Manuf. Industry
Firm-level 
interests
Infrastructure, Investment 
Development and Usage
>
National
Interests
Me
t
a
l
s
Ener
gy
Et
c…
B. Country
C
Clusters in Germany
Infrastructure, Investment 
Development and Usage >
Cluster
Interests
Au
t
o
C. Cluster
St
ee
l
Et
c…
D
Industries in 
Metals Cluster
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 7: From firm-level interests to macro trans-national interests 
 
The areas marked A and D are commonly discussed, but B and C are often 
not considered. It is important to note that it is always the same infrastructure 
that is at stake and that the field of logistics seeks integration in a systemic 
context. This integration could therefore not be limited only to areas A and D.  
 
Baudin, in his work on how to make logistics “leaner”,
92
 addresses logistics 
performance at the firm level, but unfortunately not the performance of all 
systems within industries nor industries in a national or societal framework. 
He does, however, provide examples of “dysfunctional outcomes of free 
interactions”, which is essentially the problem associated with allowing the 
optimisation of inter-firm-level systems (as opposed to intra-firm-level 
systems), rather than regional cross-cutting systems. Logisticians understand 
collaboration, and the collaborative aspect of inter-firm-level strategies, design 
and flow management is a well developed field. Collaboration is, however, 
                                                 
92
 Baudin, 2004, p. 14. 
 49
always seen as groups of firms utilising and optimising pieces of infrastructure 
and processes.  
 
The hidden and often invisible effects of many firms (often unrelated to each 
other) using the same infrastructure, often overcropping it to the detriment of 
the whole, are never discussed. This is in contrast to other factor conditions 
such as banking, labour and resources that are discussed, developed and 
measured. (If bankers extend too much credit too easily, inflation rises and 
monetary policy responds with interest rates. Yet if modal shifts because of 
firm-level interest in speed or any other reason reduces the overall efficiency 
of a country’s transport system, who responds or should respond?) 
 
Logistics textbooks, such as Grant et al., make a distinction between the role 
of logistics in the economy
93
 (described as important because of its cost to 
the economy and its role in the facilitation of transactions) and the role of 
logistics in the organisation
94
 (described as important because of its link to the 
marketing concept and because it provides one of the last remaining areas of 
major improvements in profitability and competitive performance). Pienaar 
makes a similar distinction between the macro-economic perspective
95
 
(where he discusses the actual costs for the South African economy) and the 
emergence of logistics in the business context
96
 (where he describes the 
function’s organisational context in history, its tools and position in the firm). 
 
Murphy and Wood specifically define a “macro perspective” for logistics and 
also refer to the economic and growth impetus that logistics and especially 
transport provide to an economy. They highlight the Irish case study, where 
the country decided to upgrade transport systems to sustain recent growth in 
the economy.
97
 
 
                                                 
93
 Grant et al., 2006, p. 7. 
94
 Grant et al., 2006, p. 8. 
95
 Pienaar, 2005, p. 3. 
96
 Pienaar, 2005, p. 4. 
97
 Murphy and Wood, 2004, p. 3-4. 
 50
Healey and Ilbery describe the logistics function in terms of linkages, and 
distinguish material and information flow as the core linkage requirement with 
associated monetary and sometimes personnel flow.
98
 They continue to 
argue that: 
                                                
strong material and information linkages are often associated with 
agglomerations of economic activity, but to what extent linkages 
encourage agglomeration…is a more debatable question. To begin to 
unravel the relationship between linkages and agglomeration it is first 
necessary to discuss the nature of agglomeration economics. 
Agglomeration effects which arise from the clustering of activities and 
which are external to the firm are called external economies of scale.
99
 
 
They go on to describe how all firms that share these external economies of 
scale can benefit from resultant cost savings
100
 – an important and very much 
overlooked point in the economic development of South Africa, especially in 
how it pertains to the management of long-haul corridors where purely micro-
economic considerations drive modal split, with no policy to drive long-term 
economic agglomeration advantages. 
 
The element of macro-economics that does receive attention relates to 
investment in logistics infrastructure. In this case, infrastructure is viewed as a 
“direct injection to the economy”
101
 and an additional production factor. 
Lakshmanam and Anderson describe this better-known field of study in terms 
of the accepted positive correlation and the regulatory impacts that are 
therefore required, but they are careful to note the deficiencies in the body of 
knowledge around the networks created by different modes and the effect of 
this on production factors.
102
 In this way they indicate the way towards a more 
robust understanding of the effect of different modes on network design to 
benefit an economy as a whole. 
 
 
98
 Healey and Ilbery, 1990, p. 84. 
99
 Healey and Ilbery, 1990, p. 88. 
100
 Healey and Ilbery, 1990, p. 88. 
101
 Lakshmanam and Anderson, 2002, p. 5. 
102
 Lakshmanam and Anderson, 2002, p. 6. 
 51
Similarly, in South Africa, the infrastructure component is understood in 
general, but without an understanding of the concomitant network and modal 
view. Logistics News reports on the Barloworld Supply Chain Foresight Study 
and the fact that “many SA businesses have cited the imbalances in the 
country’s Supply Chain and logistics infrastructure as the cause of their 
increasing lack of competitiveness.”
103
 In a follow-up study the same target 
research audience (chief executives and senior supply chain managers) 
reports concern for “pressures placed on SA’s supply chains by increasing 
demand locally, and by the increases in supply chain complexity brought 
about by rapid globalisation and a new diversity in the customer base of many 
companies.”
104
  
 
These concerns are often raised and also topical right now in South Africa for 
large parts of the country’s infrastructure, but a move forward would be to: 
1. Describe the calculation, cost drivers and supporting processes of the 
macro logistics costs as an important field of study (i.e. open up the 
process of the macro-economic study of logistics); and 
2. Describe the inter-relationships between the two areas (macro and 
micro) and to contemplate whether these areas offer a complete 
description of logistics (a question often best answered by taking a 
more in-depth approach to the problem). 
 
The answer to the first question is an outcome of this study, but it is the 
second question that requires an answer now. 
 
3.3 The systemic relationship between the various roles of logistics 
The systemic relationship between logistics roles is often discussed on the 
level context, such as the strategic and operational level issues within the 
firm, giving rise to discussions on centralisation versus decentralisation, the 
positioning of logistics management within the firm, and the appointment of 
                                                 
103
 Logistics News, March 2006, p. 13. 
104
 Logistics News, April 2007, p. 2. 
 52
the logistic management function on the executive and board. The 
hierarchical relationships between firms, organised industry, clusters and 
national economies are not so frequently analysed. South Africa’s cluster 
study group of the mid-1990s came across the issue of logistics as a cross-
cutting issue literally by chance, when the analysis of various economic 
clusters gave rise to what was called “cross-cutting” issues and the state of 
logistics came under serious scrutiny for the first time. 
 
The cluster studies originally set out to understand various important 
industries in South Africa (such as the steel, automotive and petrochemical 
industries) in order to facilitate the global competitiveness of these industries. 
The Logistics “society” was silent during this process, but the researchers 
soon found that certain issues impact on all industries, and logistics problems 
in South Africa seemed to top the list. 
 
The studies soon arrived at Porter’s competitive forces constructs and the 
related topics of how the production factors of nations should be organised.
105
 
Yet, incredibly, the logistics community once again missed the opportunity 
and remained largely silent on the type of contributions that the discipline 
could make to further advance integration on a national and regional level. 
Even in international texts these issues are largely ignored or at best receive 
cursory reference. Coyle et al. discuss the importance of a competitive 
environment in global logistics and do, in fact, make a connection between 
Porter’s model elements and logistics. The model elements are summarised 
as factor and demand conditions, on the one hand, and company strategy and 
related and supporting industries (i.e. cluster conditions), on the other hand. 
They also maintain that the United States specifically is “slipping in 
international trade”, because of poor performance in transportation and 
technology.
106
  
 
Louw, in his work on advanced supply chain planning processes and decision 
support systems in the petrochemical industry, arrives at industry views for 
                                                 
105
 Porter, 1990, pp. 74-76. 
106
 Coyle et al., 2003, p. 152. 
 53
inter-supply chain focus.
107
 He defines clusters as “groups of businesses that 
need to cooperate closely due to interdependencies and synergies that exist 
between their supply chains or segments thereof.”
108
 The work proceeds to 
identify some interesting distinctions between what is called “advancement 
dimensions”, i.e. moving from firm level to value clusters through industry 
clusters. The final and highest dimension is called “logistics network 
integration” and is typified as a move from transactional relationships to a 
deeper trust level across industries.
109
 
 
Critically though, and given that the identification of macro perspectives is not 
the focus of Louw’s work, he suggests a company-driven approach to move 
through the dimensions. He describes each stage or dimension as “the 
building block required for the next level of advancement”.
110
 This is 
acceptable for the context of the work, but could also be compared to a 
situation where micro-economists argue that firms, industry and commerce 
should organise themselves along “deep trust” levels to achieve the macro-
economic objectives of a national economic system. This description is 
entirely correct in the micro-economic context, as no company has a 
responsibility to act in the “common good” of a country (i.e. to act in such a 
way that its actions are detrimental to the company, purely to satisfy some 
national objectives). The situation from a national strategic and policy setting 
perspective is, however, clearly not tenable, as macro-economic objectives 
are usually set independently, measured according to defined parameters and 
then managed according to performance within these parameters. Is there 
any reason why this should be different for logistics than for economic 
systems in general? Louw’s “advancement dimensions” are therefore exactly 
correct as companies move through the dimensions to achieve own 
objectives, but in the same way the macro-economic policy maker (usually the 
regulator) should initiate the move through the dimensions from a macro-
economic starting point. 
 
                                                 
107
 Louw, 2006, p. 384. 
108
 Louw, 2006, p. 382. 
109
 Louw, 2006, p. 382-383. 
110
 Louw, 2006, p. 383. 
 54
In the next section, a basic definition of an economic system and the role of 
logistics in that system is provided. 
 
3.4 Economic systems and the role of logistics 
3.4.1 Macro-economics, clusters and cross-cutting factors 
Terreblanche describes economic systems as the various methods that 
countries (or communities) can use to organise their economic life in the 
broadest sense.
111
 What is important is the totality of economic life that is 
included in the definition and Terreblanche’s contention that an economic 
system at least requires economic activities that are collated according to a 
certain principle or principles into a coherent structure and that this structuring 
activity can be enforced.
112
 Hayes describes the alternatives inherent in this 
enforcement as two “sharply contrasting methods of bringing about 
cooperation in an economic society.”
113
 These are strict regimentation 
(socialism or collectivism) and spontaneous and voluntary individualism 
(capitalism). Notably Hayes describes the extreme forms of these systems as 
either communism or anarchism, and then argues that even the more 
individual nature of the United States system (compared to other nations) 
requires restrictions and that “these restrictions upon private property are so 
extensive today that our system is considerably removed from a state of 
Individualism”.
114
  
 
The process of economic organisation within a community is therefore total (it 
includes all aspects of economic life), structured (collated according to 
principles) and enforceable (not only spontaneous, but by a specified power). 
The resulting macro-organisation’s existence is qualified by certain objectives 
(increased output, high employment, stable prices and proper foreign 
economic policy – i.e. the relationship between the macro-organisation and 
                                                 
111
 Terreblanche, 1994, p. 3. 
112
 Terreblanche, 1994, pp. 5-14. 
113
 Hayes, 1928, p. 21. 
114
 Hayes, 1928, p. 22-23. 
 55
others),
115
 and this qualification is achieved by certain instruments that involve 
trade-offs or choices.
116
  
 
A further perspective on macro-economics is provided by McCarthy, who 
refers to the aggregation principle in economics.
117
 Micro-economy 
aggregates common goods such as the demand and supply of oranges, whilst 
macro-economy aggregates uncommon goods, such as the total production in 
the economy. He provides the standard definition of the production factors of 
macro-economics, i.e. land, labour, natural resources and enterprise.
118
  Is 
this list complete or could it at least be expanded horizontally? Are transport 
facilities, for instance, a micro-economic cluster or a macro-economic 
production factor or both? Unmined coal (resources in the earth) could be 
seen as a natural resource (macro-economic production factor to some 
extent), but becomes a micro-economic cluster when mined. As soon as it is 
mined, it can be used by other clusters (as energy in boilers, to manufacture 
steel or to generate electricity), but is, as per the definition of a cluster, not a 
cross-cutting factor. It can be replaced by other alternatives, it can be 
imported and certain sectors of the economy can do without it. This is not the 
case for functions such as banking or transport, which cannot be substituted 
nor imported, and the economy basically cannot exist without these factors. It 
seems as though cross-cutting factors in the economy, in addition to the 
standard production factors, also require aggregation and understanding at a 
macro-economic level.  
 
The fact that businesses require, for instance, a transport or treasury function 
and that other businesses such as hauliers and investment banks provide 
these services on an outsourced basis leads to a misconception in that 
logistics is seen as a cluster (and therefore micro-economically orientated) 
function only. All the micro-economic principles necessary to run a transport 
company or bank would be present, as well as the micro-economic principles 
to run the transport function and treasury function in each non-transport or 
                                                 
115
 Samuelson and Nordhaus, 1989, p. 83. 
116
 Samuelson and Nordhaus, 1989, p. 86. 
117
 McCarthy, 1998, p. 104. 
118
 McCarthy, 1989, p. 105. 
 56
non-banking business.  These activities, however, also have a macro-
economic dimension in that when they are added together, a cross-cutting 
factor arises where, for instance, if all orange production functions or coal 
mining functions in a country are added together, at most a cluster arises. 
Clusters are important, but not critical; clusters need all cross-cutting factors, 
but cross-cutting factors do not need all clusters. This principle is appreciated 
fairly well in non-logistics functions such as banking and labour, but not really 
appreciated as far as the logistics function is concerned. It is this macro-
economic dimension of cross-cutting factors (that also has micro-economic 
counterparts such as employment agencies and human resources functions 
for labour and transport companies, and transport functions for logistics) that 
provides an interesting descriptive challenge to the macro-economy in that the 
macro-economic context are often ignored. This leads to the final deduction in 
this regard, i.e. that if logistics’ objectives are the creation of maximum value 
for the least cost, then the macro-economic objective of this cross-cutting 
factor would be in the net value added to a country’s Gross Domestic Product, 
and in order to achieve this it should be managed as such. 
 
To proceed from this point forward requires an understanding of clusters 
compared to cross-cutting factors, a definition of cross-cutting factors and the 
general role of these factors in the economy. 
 
3.4.2 Economic clusters 
The cluster concept probably originated when Porter expanded his five-forces 
concept to national competitiveness and developed his “diamond of national 
advantage”.
119
 Porter calls the formation of clusters a “pervasive” 
characteristic of all “advanced national economies” and it relates to the fact 
that successful industries “help to create others in a mutually reinforcing 
process”. The process can be either vertical (buyer/supplier) or horizontal 
(common customers, technology, channels, etc.).
120
 
 
                                                 
119
 Porter, 1991, p. 73. 
120
 Porter, 1990, p.149. 
 57
South Africa has well-known examples of such clusters, for instance the 
vertical relationship between the country’s extensive coal and iron ore mining 
industries, iron and steel manufacturing, automotive and steel distribution 
commerce. Other vertical examples include chemicals (from conversion of 
raw materials such as coal and other minerals to chemical manufacturing and 
downstream processes), food production (from agriculture to downstream 
food production and branding) and paper (from forestry to downstream 
printing, ink production and usage). These vertical chains are often described 
as value chains, but a cluster would form where these value chain activities 
are grouped together in close geographic, process, ownership or other group 
configurations. 
 
Clusters are often seen as only a geographic concept, such as in the 
conclusion of Wisner et al. that “the concept of business clusters represents a 
new way of thinking about facility decisions”. This could be construed as 
including various typologies, but they continue to refer to “geographic 
concentrations.”
121
 
 
Clusters are, however, a multidimensional concept, incorporating both 
arguments mentioned above. Two of the most important dimensions are 
geography and industry sectors (including the “agglomeration” of enterprises 
in these sectors).
122
 One could argue that these are pre-determinants for 
cluster development, i.e. economic clusters cannot develop if both of these 
dimensions are not in place. If, however, these dimensions are in place, other 
“external economic factors”
123
 such as market access, labour market pooling, 
intermediate input effects and technological spill-overs come into play as 
“unintended or incidental by-products”.
124
 In order to identify a fully developed 
cluster there should also be joint action
125
 or what Zeng
126
 describes as 
taking “deliberate efforts to cooperate and collaborate.”  
                                                
 
 
121
 Wisner et al., 2005, p. 355 . 
122
 Schmitz, 1992, pp. 64-69. 
123
 McCommick, 2004, p. 1. 
124
 McCommick, 2004, p. 1. 
125
 Nadvi, 1999, p. 1. 
126
 Zeng, 2006, p. 3. 
 58
The concepts of cooperation and collaboration are not foreign to logistics, but 
what is interesting here is that the act of collaboration (in any economic 
context) is not possible without common platforms or points of departure. 
Movement through a value chain always implies that physical product (via 
beneficiation) and transactions flow downstream, with supported upstream 
and downstream flow of information – core to the logistics concept. This 
definition therefore implies, as illustrated in Figure 8, important flows in the 
economy, i.e. of goods, transactions and information – without these 
important linkages neither unintentional nor intentional co-operation and 
collaboration could be possible. 
Clusters
Land
Resources
Labour
Growth
Enterprise
Employment
International
Stability
P
r
i
m
a
r
y 
= B
a
si
c
 p
r
oduc
t
i
on
S
e
conda
r
y
 = 
M
an
uf
ac
t
u
r
i
n
g (
I
n
dus
t
r
y
)
T
e
rti
a
ry
 =
 S
e
rv
i
c
e
s
Micro-economic
M
ac
r
o-
ec
ono
mi
c
 I
n
p
u
t
s
M
a
c
r
o-
ec
ono
m
i
c
 O
u
t
p
u
t
s
 
Figure 8: The development of clusters 
 
A further requirement of this study would be to define the list of underlying 
economic sectors that can be used to construct cluster identities and provide 
a stable platform from which these could be studied. No clear framework 
exists in the economy to do this, which in turn gives rise to the unfortunate 
situation where research databases differ. This is usually acceptable for 
micro-level work, but hampers exactly the macro framework that the economy 
requires. Data produced by the South African Revenue Services, the National 
Ports Authority, Spoornet, Statistics South Africa and other sources all differ in 
 59
frameworks. This is clearly an unfortunate state of affairs and should be 
addressed. 
 
In terms of logistics two important drivers should be considered when 
developing a framework for economic data gathering. Globally the standard 
industrial classification system is used for economic data reporting, with 
various derivatives that differ between countries, but are almost always 
comparable through various concordances like the one that makes 
comparison between NAICS (the North America system) and NACE (the 
grouping of European systems)
127
 possible. The harmonised system is often 
used for international trade and is unique in that services are excluded and 
only physical goods are classified, but once again comparisons within the 
larger groupings of standard industrial classification systems can be 
reached.
128
 It should be apparent that any system of reporting on economic 
production should refer back to these related frameworks. Statistics South 
Africa uses this classification scheme and the overarching framework is 
depicted in Table 3.
129
 (Statistics South Africa does unfortunately not always 
publish data at this level of detail.) Other comparative problems exist even 
within the organisation such as the fact that much of the data are aggregated, 
some publications have been discontinued, some are in rand value and some 
are in weight. 
 
 
 
 
 
 
 
 
 
 
                                                 
127
 Communication and Resource Centre Administrator, European Commission, 2007, p. 5. 
128
 Communication and Resource Centre Administrator, European Commission, 2007, p. 7-8. 
129
 Central Statistical Services, 1993, p. 1. 
 60
First Digit 
SIC Code 
Overarching 
Economic 
Sector 
Economic Division 
1 Primary Agriculture, hunting, forestry and fishing 
2 Primary Mining and quarrying 
3 Secondary Manufacturing 
4 Tertiary Electricity, gas and water supply 
5 Tertiary Construction 
6 Tertiary 
Wholesale and retail trade; repair of motor vehicles, motor 
cycles, personal and household goods; hotels and restaurants 
7 Tertiary Transport, storage and communication 
8 Tertiary 
Financial intermediation, insurance, real estate and business 
services 
9 Tertiary Community, social and personal services 
0 Consumption 
Private households, exterritorial organisations, representatives 
of foreign governments and activities not adequately defined 
 
Table 3: Overarching economic classification 
 
From a demand side, the overarching system makes a clear distinction 
between the primary, secondary, tertiary and consumption sectors of the 
economy, or from a supply side view only of primary, secondary and tertiary 
sectors. According to a value chain perspective, one could refer to the 
extraction, beneficiation, services and consumption sectors of the economy 
(Figure 9). 
 
 
 
 
 
 
 
Extraction
(Mining & Agriculture)
(Primary)
Beneficiation
(Manufacturing)
(Secondary)
Consumption
(Private 
households)
Services
(Energy, construction, trade, transport, professional, community)
(Tertiary)
 
Figure 9: The economic value chain according to the standard industrial classification 
 
 61
The problem with this classification arises at the next level SIC code, i.e. the 
further sub-classification of the system. (For the purposes of this discussion 
the study concentrates on the extraction and beneficiation sectors, as the 
other sectors’ weight is negligible compared to the first two). The global 
system subdivides agriculture into a further three subdivisions, mining in six 
and manufacturing in 10. On the third level, as depicted in Table 4, 10 
agricultural, nine mining and 60 manufacturing subdivisions are included in 
the classification. 
 
  2nd div. 3rd div. 4th div. 5th div. % of total 
Agriculture 3 10 14 14 6 
Mining 6 9 18 21 9 
Manufacturing 10 60 131 190 85 
 
Table 4: Detailed sub-sectors of the extraction and beneficiation sectors of the economy  
 
The problem that arises from this classification scheme is that it is 
hypothesised that the mining sector, which has access to only 21 5
th
 divisional 
sub-sectors (6% of the total), comprises about half of our economic 
production by weight. On the other hand, the manufacturing sector, which has 
many more divisional sub-sectors, also has many more internal complexities 
irrespective of its lower weight. The fact is, however, that neither weight nor 
logistics complexity plays a role in the determination of these groupings, as 
logistics is not considered nor forms part of the decisions in this regard. This 
is a clear indication of how logistics considerations (which strive to provide 
systemic integration on a firm level) have failed to provide the same guidance 
on a macro-economic level. The outcome of this state of affairs is that various 
entities with various considerations use systems that serve their own 
objectives (e.g. Spoornet – which has more bulk commodity classifications 
than average; National Ports Authority – ad valorem considerations; SARS – 
tariffing) and therefore design distinct and incomparable systems to serve 
these objectives. This is understandable and acceptable, but the simple 
missing link, i.e. relating all these considerations back to the standard system, 
 62
is not executed nor required by any entity in the process and especially not 
the regulatory authority that should be concerned with the recording of 
information for macro-economic guidance around logistics issues, i.e. the 
Department of Transport. 
 
The process of further refining this system, as well as informing the weight 
versus complexity debate, is a natural further by-product of the study and will 
be considered in the outputs. 
 
3.4.3 Cross-cutting factors defined 
Cross-cutting factors identified in South Africa and often referred to in the 
literature include transport and logistics, the banking system, AIDS, sectoral 
business structuring and clustering, international trade and tariffs and tariff 
strategy, multi-factor productivity (both management and workers), import and 
export parity management and the development of downstream industries. 
 
The process of identifying cross-cutting factors was driven by the foreign 
investment that followed the attainment of democratic freedom in South Africa 
in the early nineties and the wish to establish strategies that will accelerate 
growth. One of the first initiatives was studies funded by Japan, called the 
Japanese Grant Fund (JGF) studies. The JGF's initial focus was on the 
following cross-cutting studies:  
• Pre-shipment export finance;  
• Strengthening business services for South Africa's SMMEs;  
• Promotion of foreign direct investment; 
• Labour skills upgrading: financing of training; and 
• Regional industrial development policy: an evaluation of the RIDP.
130
 
 
As the process of specific cluster studies continued, this list grew rapidly to 
incorporate previously unidentified factors, such as transport which, though 
                                                 
130
 Nedlac, 1996. 
 63
initially overlooked, soon became the one issue that seemed the most difficult 
to deal with and resolve. 
 
A clear and unequivocal distinction between economic clusters and cross-
cutting issues arises from this list. These cross-cutting factors all relate back 
to macro-economic management and the achievement of the national 
objectives of growth, employment, stability and international trade growth, as 
well as the protection and advancement of the macro-economic production 
factors of land, resources, labour and enterprise. 
 
3.5 From macro-economic imperative to measurement 
Logistics and its management is clearly shown as a macro issue that requires 
management and then, by definition, measurement. The case for 
measurement is articulated in Chapter 3 of the study, but to what extent has 
the measurement ideal been achieved in South Africa? The historical context 
and past achievements in this regard are discussed in Chapter 4.  
 
Distinct time periods can be identified that match closely with the phases of 
logistics development as identified in Chapter 2.  These time periods are 
summarized in Table 5. The framework can now be used to consider the 
history of transport logistics measurement in South Africa. 
 
 
 
 
 
 
 
 
 64
Description of 
Economic Platform 
Time 
Logistics Support and 
Architecture 
Global Trend 
Primary economy < 1870 Infrastructure development 
Configuration & 
reconfiguration to 
achieve efficient 
access 
Industrial economy 1870-1960 
Production & Operations 
Management 
Time & motion 
efficiency 
Market economy 1960-1990 
Transport & distribution 
(logistics) 
Time, form & place 
utility 
Electronic 
communities & 
trading platforms 
1990-2000 Integrated trading platforms 
Collaboration & 
integration 
Globally networked 
economy 
2000-2020 Globalised systems 
Information “open 
access” 
Global bio-
engineering 
2020 + Quest for global balancing 
Quest for global 
sustainability 
 
Table 5: Phases of logistics development  
 
 
 
 
 
 
 
 
 
 
 
 
 65
Chapter 4 - Literature study and historical context of 
macro transport logistics measurement history in 
South Africa 
4.1 Introduction 
In the context of the importance of the macro-economic dimension of 
transport logistics and the resultant management and measurement 
imperatives, this chapter discusses the history of measurement, milestones 
and problems. This discussion uses the framework typology developed in 
Chapter 2 as background. 
 
The performance of macro-economic measurement in South Africa 
specifically (against the backdrop of a local and historical context) was 
researched and the available literature studied to form a picture of what was 
achieved and where improvements were possible. Only five surveys (Figure 
10) conducted over the past five decades (of which the last one was repeated 
a few times) could be found and the relative successes and shortcomings of 
these will be discussed in this chapter. In the absence of surveys, the role and 
position of measurement in general are also analysed. 
 
Primary 
economy
Industrial 
economy
Market 
economy
Electronic 
communities
Global 
networking
Global bio-
engineering
1870 1960 1990 2000 2020
Verburgh 
1957
Smith 1971
Hamilton 
1981, 1985
Pretorius 
1985-1993
 
Figure 10: Logistics phases and surveys 
 
4.2 Infrastructure development during the primary economy phase 
In the years between the invention of steam transport and the second 
industrial revolution, first world rail transport began to blossom. Europe saw 
 66
the development of a major rail network and the transcontinental rail link was 
completed in the United States of America in 1869.
131
 
 
Transport in South Africa in the pre-industrial era was slow and painful, and 
people often did not move much from place to place during their lifetimes. 
Incidents like the sinking of the Birkenhead close to Gansbaai led to a 
situation where Afrikaans-speaking people with English surnames could trace 
their lineage back to English soldier survivors of the shipwreck that simply 
remained in Gansbaai after the incident, because it was so difficult to get back 
to Cape Town. The possibility of building a railway was therefore seen as 
“immensely beneficial”
132
 and “enterprising local merchants and speculators 
were pleading for a link to the interior”.
133
  
 
What about planning and measurement? This has to be seen against the 
context of the cost of investment, the planning horizon and the long-term 
effect of planning decisions. In many respects this context places the planners 
of the mid-19th century in very similar position to the one South Africa faces 
now. A lack of foresight and a critical lack of planning around the required 
links between the coast and fast-developing hinterland were evident. In 
1845
134
 illustrious figures such as the Attorney General of the Cape Colony, 
William Porter and the Surveyor General of the first Railway company to be 
formed in South Africa, i.e. The Cape Town Railway and Dock Company,
135
 
refused to have their names associated with the first railway construction in 
South Africa. Conradie states that political agendas rather than economic 
considerations are “a fundamental principle on which transport development 
in South Africa was built” and that this state of affairs still persists today.
136
 
 
The political hue that affected transport issues was also evident in the first 
road-rail debates in South Africa. A number of people had a vested interest in 
road transport on ox wagons. As long ago as the early years of railway 
                                                 
131
 Coyle et al. 1986, p. 6. 
132
 South African Railways, 1960, p. 1. 
133
 South African Railways, 1960, p. 3. 
134
 South African Railways, 1960, p. 2. 
135
 South African Railways, 1960, p. 3. 
136
 Conradie, 2007. 
 67
development the short-term advantage was enjoyed by the road transporters, 
because of low capital and maintenance costs (i.e. not paying for road 
infrastructure).
137
 The emotional debate, however, revolved around vested 
interests, and the measurement and demand for transport in the future were 
not considered. It was reported in 1885 that the rail system saved the Cape a 
million pounds per annum,
138
 but the argument did not contribute to future 
investment scenarios. Burman draws an interesting parallel with the 
development of commercial farming when he talks about transport in the early 
19
th
 century.  This should serve as a warning regarding South Africa’s national 
objective for the second economy and the development push from 
subsistence to commercial farming: 
Why should a farmer in the Overberg risk his wagon, his precious oxen 
and even his life, to carry produce over the mountains to Table Bay? 
So the farmer grew just enough crops for subsistence, and produce at 
Table Bay remained scarce.
139
 
 
One of the fundamental problems of the change from subsistence farming to 
commercial farming is access, which was a problem nearly 200 years ago and 
is still a problem today. 
 
At the start of the second industrial revolution in Europe, South Africa had 104 
km of railway (between Cape Town, Wynberg and Wellington; and in Durban 
between the Point and the city).
140
 
 
4.3 Production and operations management during the industrial 
economy phase 
4.3.1 Introduction 
In Europe the industrial age truly burgeoned during the late 1800s. The 19
th
 
century was typecast as the century of the “transport revolution” and 1870 as 
                                                 
137
 Heydenrich, 1965, p. 101. 
138
 Heydenrich, 1965, p. 109. 
139
 Burman, 1984, p. 11. 
140
 Van Rensburg,  1996, p. 2. 
 68
the year of the start of the “second industrial revolution”.
141
 The establishment 
of industrial complexes had two immediate effects on logistics, i.e. the 
decision criteria for long haul and short haul, which mostly implied rail 
transport for long haul and local distribution by truck,
142
 and the decision that 
plants should be located in the “locality where the aggregated transportation 
costs are the least”.
143
 
 
These approaches are clearly devoid of a total logistics view and the more 
advanced cost and competitive advantage trade-offs that the logistics 
profession is capable of in the modern era. Most theory concentrates on 
production and operations management and sees the logistics function more 
as the work-in-process function on the factory floor. The inbound 
transportation of raw materials and the outbound delivery of finished goods 
receive little strategic attention. Cost measurement on a micro-scale is 
considered, but the long-haul imperative for rail is hardly questioned, nor 
measured as an economic activity. Wynne-Roberts discusses three causes 
for increased product cost, i.e. the cost of raw materials, sub-optimisation of 
assets used for production and labour unproductivity. He lists eight reasons 
for raw material cost increases, but fails to mention inbound transportation or 
logistics costs. Nowhere in the text does outbound distribution play any 
role.
144
 Evidently in this era transportation was still seen as an administered 
cost, with a belief that management could not have an impact on this cost, nor 
use it to provide a competitive advantage for the firm. According to Keebler 
and Durtche, “transportation services such as rail and trucking were heavily 
regulated and comprised 70 percent of a manufacturers’ logistics expenses, 
compared with 57 percent today.”
145
 The concept of an administered cost is, 
however, a macro-economic concept per se, and this therefore places 
logistics squarely in the macro-economic arena.  
 
                                                 
141
 Terreblanche, 1988, p. 106. 
142
 Spriegel and Lansburgh, 1947, p. 114 
143
 Locklin, 1935, p. 114. 
144
 Wynne-Roberts, 1956, p. 52-53. 
145
 Keebler and Durtsche, 1999, p. 48. 
 69
In the Heckscher-Ohlin model that has been used since the 1930s as an 
orthodox approach to the ultimate causes of international trade, it is stated 
that commodities differ in their factor requirements and that countries differ in 
their factor endowments. Ten assumptions are used in the running of the 
model, of which the tenth is that transportation costs are zero.
146
 The model 
concluded that factor endowments of one nation competing with another 
drove competitive advantage (just as Ricardo’s model stated that labour 
productivity is the key)
147
 and that transportation costs do not play a role in 
the determination or acquisition of competitive advantage. McDougall tested 
these models in 1951 and determined that there is a direct positive correlation 
between relative trade advantages and improved labour productivity, but that 
even in the case of vast factor and labour productivity advantages, markets 
are not captured by competing nations completely, as the models suggested. 
He lists the exclusion of transport costs from the model and the relative effect 
of negative transportation abilities in spite of positive factor endowments as 
one of the reasons for this phenomenon.
148
 
 
On a firm level in this era operations management came of age as a 
quantitative discipline. Most contributions during the 1920s and 1930s were 
quantitatively driven with statistical quality controls as a benchmark,
149
 but 
inbound and outbound logistics played little or no role.  
 
In South Africa in this era the industrial revolution hadn’t taken off yet to the 
same degree as in Europe, but the discovery of gold and diamonds in the 
latter half of the 19
th
 century led to increased requirements to reach the 
hinterland effectively. Opposition to a railway still prevailed and especially in 
the East Free State rallies opposing this development were held.
150
  Details 
on the “traditional” transport mode (on road and by ox wagon) were known, 
                                                 
146
 Chacholiades, 1990, pp.63-66. 
147
 Chacholiades, 1990, p. 88. 
148
 Chacholiades, 1990, p. 90. 
149
 Cook and Russel, 1984, p. 4. 
150
 Van Rensburg, 1996, p. 3. 
 70
i.e. it cost about 40 shillings per 100 lb. to move around 1 000 miles. The 
wagon and mules required 65 days to do the trip.
151
 
 
Disagreements over the rail gauge, which led to the replacement of the first 
track of 54 inches with the current 42 inches, also broke out. The winning 
arguments were not based on proper research, but short-term interests and 
powerbases.
152
 The decision of 1871 to revert to the narrow 3’6” gauge was 
based more on speed of construction than actual cost savings (which were in 
dispute) and not thought through on all dimensions.
153
  
 
Railway development was, however, rapid and the two colonies and two 
republics that would later form the Union of South Africa independently 
established railways, connecting the gold and diamond fields of Johannesburg 
and Kimberley with Cape Town, Port Elizabeth, East London and Lourenço 
Marques (then Delagoa Bay, today Maputo).
154
 These developments often 
entailed acrimonious deliberations;
155
 they were one of the main causes of the 
first Anglo-Boer war in the early 1880s, and although the first railway 
companies were in private hands, as in the United States, governments 
became more and more involved as the territorial scramble continued.
156
  
 
The establishment of the Union of South Africa in 1910 led to these lines 
being officially connected and the establishment of one national railway 
organisation. This organisation controlled more than 12,000 route kilometers 
(60% of today’s total), about double the 6,500 km at the turn of the twentieth 
century. This grew to approximately 22,000 kilometers by the late 1920s.
157
 
The peak of development was reached at this stage – compared to the United 
States, where track mileage peaked in 1916.
158
 
 
                                                 
151
 South African Railways, 1960, p. 15. 
152
 Burman, 1984, p. 50. 
153
 Heydenrich, 1965, p. 62. 
154
 Van Rensburg, 1996, p. 4-5. 
155
 Coetzee, 1940, p. 215. 
156
 Coetzee, 1940, p. 218-219. 
157
 South African Railways, 1960, p. 14. 
158
 Coyle et al., p. 6. 
 71
Transport policy of the Union between 1910 and 1948 was straightforward. 
The basic objective was to develop and run a railway on business principles 
(i.e. it should pay for itself), but that agricultural and industrial development 
should be taken into account.
159
 Somehow this consideration led to differential 
tariffing: low tariffs for mining and agriculture and high tariffs for industrial 
goods. This was the first form of cross-subsidisation and the first threat to rail 
density as higher-value goods immediately sought to move to road transport, 
because of the inordinately high tariffs that were required for rail 
transportation.  This reduced rail density and applied cost pressures on lower-
value commodities. The total cost concept for the economy was, however, not 
considered and possible solutions ignored this basic approach. The solution 
to “control” road freight (as proposed by the Le Roux commission in 1929 and 
confirmed by the Page commission in 1947) didn’t really address the problem 
structurally and in fact, even today, the problem has not been solved. Simple 
calculations of macro rail economics, current and future demand, and costs to 
the economy could have suggested better solutions, but these were, however, 
never addressed. 
 
The drivers of regulation were much simpler, as indicated by the 1927 report 
of the General Manager of the South African Railways and Harbours. He 
specifically argued that road hauliers are not “common” carriers and can 
charge any tariff and ignore any traffic. They could in effect “reverse” railway’s 
strategy of charging more for higher-value traffic and the only reason for 
prohibiting this practice would be to protect the railway investment.
160
 The 
options available in terms of protecting the self-interest of railways by 
considering density and other transport economic arguments were clearly 
ignored. This issue surfaces again in an interesting way when the same 
General Manager quotes Mr R.N. Harvey (Vice-Chairman of the Executive 
Committee of the Association of Chambers of Commerce) from an address 
delivered at the annual congress of Assocom held at East London during 
October, 1959: 
                                                 
159
 Van Rensburg, 1996, pp. 8-11. 
160
 South African Railways and Harbours, 1927. 
 72
I leave aside the aspect that some reduction in high-rated railway traffic 
would not necessarily have the dire effects upon Railway finances 
which the Railways apparently fear, because high-rated traffic is not 
always highly remunerative traffic, high costs are associated with a 
good deal of high-rated traffic, notably where the shorter hauls are 
involved and proportionately high terminal costs are experienced.
161
 
 
The General Manager interestingly continues to complain that the Railways 
needs to “be administered on business principles” and should “balance its 
own annual budget”.
162
 The issues of density, the opportunity costs of lost 
traffic and fixed versus variable costs, are ignored. This problem persists even 
today and Spoornet continues to shed customers without considering 
opportunity costs and density. 
 
4.3.2 Verburgh’s 1958 contribution 
Verburgh, in the years directly after the new government came into power in 
1948, began detailed studies on transportation volumes and demand for 
transport. He did pioneering work and should be appreciated for many 
reasons, including the ground-breaking nature of the surveys. His first 
analysis defines the “safe” domains for rail and road in two categories. These 
are fragile, urgent and perishable for road, and bulk long-haul for rail.
163
 
Verburgh conducted his research long before the advent of intermodal 
solutions and containerisation and his transport classes therefore do not 
include other more sophisticated categories. 
 
Verburgh begins his work by defining capacity in terms of equipment for road 
and rail. He includes in his datasets these capacities as well as rail tonnages. 
He also makes a graphic comparison of these data against net national 
income at factor prices (Figure 11). 
                                                 
161
 South African Railways and Harbours, 1959. 
162
 Ibid. 
163
 Verburgh, 1958, p. 13 
 73
 
Figure 11: Verburgh's 1958 illustration (indexed 1946 = 100) 
 
In his analysis Verburgh expresses satisfaction with rail’s growth of 78% over 
the period, attributes this to the growth in inter-city transport in “splendid 
isolation” and does not express concern for what the data might show initially, 
i.e. that rail volumes are growing slower than the commercial road vehicle 
population.
164
 It is possible to analyse Verburgh’s data further, both in terms 
of what the data intrinsically express and in terms of what extrinsically could 
be added to the data. 
                                                
 
4.3.2.1 Intrinsic problems highlighted with Verburgh’s data 
Verburgh’s satisfaction with rail market share and the protection of intercity 
rail freight is disconcerting. He ignores, at first, the real relationship between 
the capacities of road and rail. He tabulates in a clear and unambiguous way 
 
164
 Verburgh, 1958, p. 15-18 
 74
the growth in road operational capacity, compared to rail operational capacity, 
i.e. wagon and truck populations
165
 (an important element of operational 
capacity). Even if no data were available, he could have hypothesised other 
elements of capacity, i.e. the growth in truck capacity (which improved 
markedly over this time period),
166
 and more importantly the relationship 
between road and rail fixed infrastructure. Rail’s fixed infrastructure reached 
its current levels in 1956 and was therefore able to carry 180 million tons 
easily (if not more) compared to the 64 million tons that Verburgh reported for 
1958. This must mean, in an “administered” tariff environment, that additional 
freight on the rail infrastructure could be carried more cheaply (even if the 
ratio of fixed to variable costs is low – and some authors today postulate that 
as much as 70% of rail costs are fixed). This then also suggests that one of 
four options exists around the user-pay principle: 
• Either trucks are paying the full costs of fixed road development, in 
which case the country as a whole is paying too much for transport; or 
• Trucks are subsidised by motor vehicles in terms of fixed road 
development, in which case a distortion is occurring, and the economic 
cost will rise; or 
• The economy is subsidising road development, in which case a 
distortion is occurring, and the economic cost will rise; or 
• There is limited road transport next to railway lines, i.e. all transport is 
on feeder routes, in which case no specific problem exists. 
 
The only sensible option would be option four, but Verburgh reports 
exemptions granted to road operators that allow long-haul operations, which 
contradict this statement
167
 and therefore at least the possible effect of the 
other three options should have been considered. In reality, a combination of 
                                                 
165
 The capacity of a transport system can be explained in three classes, i.e. installed, 
maintained and operational. Installed capacity is the initial fixed investment and can only 
be improved in leaps; maintained capacity is a view on the annual level of maintenance of 
this infrastructure and can also only be improved in leaps if the maintenance schedule 
deteriorated, whilst operational capacity is the real capacity available on the fixed 
infrastructure, given current levels of employees, systems, motive power and carrying 
capacity.  These definitions have become more poignant in today’s world of infrastructure 
capacity issues and were developed further by the candidate for current policy makers. 
166
 Australian Bureau of Statistics, 2001, p. 3. 
167
 Verburgh, 1958, p. 36-37, p. 44. 
 75
the fourth and other options occurred in South Africa, which means that 
railway tariffs remained unnaturally high, making the system more vulnerable 
than it should have been when deregulation finally occurred.  The high tariffs 
also created a false sense of security in the narrow gauge, which have 
created undue investment challenges today.
168
  
 
4.3.2.2 Extrinsic problems highlighted with Verburgh’s data 
A comparison of Verburgh’s rail data with national accounts information which 
was available in 1958 leads to more interesting extrinsic questions. The data 
can be correlated with production in agriculture, mining and manufacturing, 
often called the “transportable” portion of the economy. These statistics are 
available and point to structural problems in transport. The comparison is for 
obvious reasons not a precise measurement of market share, as structural 
changes in the locations of production and consumption are not included, but 
it remains a good approximation of indicative rail market share in the absence 
of other data. The correlation is depicted in Table 6. 
                                                 
168
 The carrying capacity of railway trucks – which remains the biggest driver of marginal 
costs and means that larger carrying capacities decrease costs – increases with wider 
gauge. In most cases this capacity can at least be doubled with a wider gauge. Compared 
to Europe, which has a wider gauge and half of South Africa’s road truck carrying capacity 
limits, an important marginal cost component of South Africa’s rail system today has an 
installed backlog by a factor of four – a result of the lack of this reasoning by many 
researchers of the day. 
 76
0
20
40
60
80
100
120
140
160
180
200
1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956
I
nde
x
 
Mining Agriculture Manufacturing
 
Figure 12: Rail transport as percentage of transportable GDP in tons (Indexed 1937=100) 
 
The time series points to interesting developments: 
• Rail’s share of agriculture traffic is rising. This sector is the most 
heavily subsidised, difficult to transport (due to many collection points) 
and one of the most difficult to invest in as far as transport equipment is 
concerned (as the traffic is seasonal and the equipment is therefore not 
utilised throughout the year – hence even more fixed costs compared 
to road). It is also cyclical, thereby exacerbating all these challenges. In 
addition, the sector reaches equilibrium points, where neither imports 
nor exports of commodities are necessary, directly influencing the 
share of transport versus production. 
• Rail’s share of mining remained static. As most mining commodities 
are exported from fixed points of production through fixed harbours, 
this is by nature rail’s bread and butter transport. 
• Rail’s share of manufacturing declined by a fifth. This could be due to 
structural problems, but it is seriously questioned whether the total loss 
could be for this reason. By now the railways are complaining about 
competition and are clearly observing the trend. The fairly ridiculous 
 77
situation in which manufactured transport continues to subsidise low-
value transport and then declines because of service issues (return is 
invested in agricultural transport through cross-subsidisation) 
continues, without serious management of the problem. 
 
The issue of agricultural market share and the development of branch lines 
was always a political one.  It only changed in the 1980s as South Africa and 
Transnet were preparing for change. As far back as 1928, and before formal 
road transport regulation, Frankel reported: 
Instead of the decisions on new railway construction being left to an 
expert body acting on commercial principles and co-ordinating the 
expenditure on it with the capital expenditure necessary in other 
directions, not only is the amount to be spent every year decided 
mainly by the Minister of Railways and Harbours and the Government, 
but it appears that even the decision as to which of the proposed lines 
of railway are to be constructed depends largely on the wishes of the 
minister…..the building of branch lines has become a matter of annual 
competition in Parliament for the favours of a Minister cajoled and 
threatened by a large number of politicians, each clamouring for a line 
in his own constituency.
169
 
 
These issues require management and should have been considered with the 
other data relating to Verburgh’s initial summary of the situation.  The data are 
also confirmed by the reaction of freight owners, with approximately 20% of all 
manufacturers, retailers and wholesalers complaining about delivery area 
exemption certificates at this time.
170
 
 
The other reasonably obvious extrinsic consideration is the unnatural spatial 
problems that South Africa faces. The country is now preparing for inter-
metropolitan traffic and faces spatial problems that no other comparatively 
sized developed country on earth faces, i.e. long-haul corridors between six 
                                                 
169
 Frankel, 1928, p. 113-114. 
170
 Verburgh, 1958, p. 196-197. 
 78
metropolitan centres, i.e. Cape Town, Durban, Johannesburg, Port Elizabeth, 
East London and Bloemfontein. This spatial organisation and the data that 
could support it actually point to a sensible network design. This design 
should concentrate on heavily engineered rail corridor connections between 
metropoles and between mining deposits and manufacturing centres and 
ports. This network should be supported by a secondary feeder road system 
supported by branch lines, where densities allow for it. Whilst the export of 
mining commodities is important from a development perspective (and the 
feeder option would today be labelled part of  the second economy), the 
corridor element will be important for the first economy of the future – the data 
therefore point to a future structural problem, when unsustainable rail 
business receives investment (branch lines to remote agricultural areas over 
routes of light density) and future, sustainable rail business (heavy haul 
corridors) receives less investment. 
 
4.3.2.3 Verburgh’s conclusions of 1958 
Verburgh’s survey methods of 1957 (the year in which the survey was done) 
were based on a special study of the Transportation and Communications 
Commission of the United Nations Organisation.  This study suggested that in 
most countries the technique of sampling and limited period questioning 
provided the best method to obtain road transport performance figures.
171
 
Verburgh and his team at the Bureau for Economic Research at the University 
of Stellenbosch drew a sample of private and ancillary road transport service 
providers to develop an understanding of the industry. 
 
Pertinent statistical conclusions on market share from Verburgh’s study 
undertaken in 1957 are as follows:
172
 
• About 8 000 firms were involved in haulage for reward in 1957. They 
operated 12 000 lorries, 1 500 mechanical horses and 4 500 trailers 
with an average carrying capacity of 5 tons for lorries and 10 tons for 
                                                 
171
 Verburgh, 1958, p. 210. 
172
 Some of these statistics were calculated from Verburgh’s data by the candidate, as not all 
relevant calculations necessary for future reference were done by him. 
 79
trailers. Only 20% (2 400) of the lorries had a carrying capacity of more 
than 2.5 tons, but 55% of trailers had a carrying capacity of more than 
8 tons. Only 8% of the private hauliers (about 64 firms) owned more 
than three vehicles; 
• These firms, together with road transport done by the railways (about 
10% of total haulage for reward), transported 67 million tons in 1957 
compared to railway traffic of about 75 million tons. 
 
These statistics enable the calculation of market share as depicted in Table 6. 
 
Tons Ton-kilometers 
  
  Rail 
Road 
for 
reward 
Road 
ancillary Total Rail 
Road 
for 
reward
Road 
ancillary Total 
Millions 75 67 113 255 32 000 1 955 1 960 35 915 
% 29 26 44 100 89 5 5 100 
 
Table 6: Modal market share in 1957
173
 
 
Transport supply in 1957 was therefore 255 million tons and 36 billion ton-
kilometers
174
. R659 of gross domestic product at constant 1995 prices were 
therefore produced for every ton that was shipped and R4.68 of gross 
domestic product for every ton-kilometer performed.
175
 The dominance of rail 
in 1957 is clearly evident. Infrastructure decisions as far as long-haul freight 
transport goes are still essentially rail decisions. If this is transposed to 
average transport distance, the figure for rail is 427, for road for reward 29 
and for road ancillary 17. 
 
Average transport distances are in balance with a “pipe and feeder” system, 
where railways are the pipe, fed over shorter distances by road transport for 
                                                 
173
 Calculated from Verburgh's (1958) distribution of vehicles per firm type (p. 21), average 
distance loaded per annum (p. 217) and ton-miles per vehicle (p. 168). Verburgh 
excluded fuel, farming, service and farming vehicles. 
174
 This figure is about 20% of supply today for tons and 5% for ton-kilometers. For a spatially 
challenged economy the problem became exponentially worse over 50 years – stressing 
the issue of long-term planning. This will be further discussed later in this study. 
175
 Comparisons between transportable volumes and the transportable portion of the GDP will 
be developed in later parts of the study 
 80
reward and very short distances by freight owners’ own ancillary transport 
systems.   
 
Verburgh does not specifically calculate and compare the percentage 
distribution between cargo types within 50 km and beyond 50 km for transport 
for reward, but it can be done from his data
176
 (Table 7). 
 
Distribution 
% Within 50 
km 
% Beyond 50 
km 
Household removal 0 1 
New furniture/household goods 1 0 
Metals/machinery 3 4 
Aggregate 53 27 
Building material 23 11 
Coal 4 24 
Oil 0 2 
Ores 1 0 
Agricultural products 4 8 
General Cargo 13 23 
 
Table 7: Distribution of road transport for reward within and beyond 50 km zones in 1958 
 
Aggregate has low value and transport drops off, but so does coal which is 
transported over longer distances by road. 
 
The primary goal of Verburgh’s work was to balance supply and demand.  
This was, however, done from the perspective that insufficient supply should 
be provided by the extension of road transport, not from the perspective of 
balancing the transport modes. This is clearly the Archimedean Point
177
 of 
Verburgh’s study. Verburgh complains about Local Road Transportation 
                                                 
176
 Calculated from Verburgh, (1958), cargo tons within 50 miles (p. 101) and cargo tons 
further than 50 miles (p. 103). 
177
 The Archimedean point is described by Dooyeweerd as the “standpoint which transcends 
modal (in this case thought structure and not transport modes) aspects”, that “may not be 
divorced from our own subjective self, but is still subject to the law (in this case scientific 
structure)”  Dooyeweerd, 1969, Volume I, pp. 8-12. 
 81
Boards’ lack of experience and inconsistent application of regulations. He 
refers to price collusion, a small amount of price regulation and cut-throat 
competitive practices in the road industry. This leads to many haulier business 
failures, in general due to over-cropping. He campaigns for quality standards 
and more freedom for larger firms to develop longer-distance routes, exactly 
for the reason that standards can be developed and over-cropping minimised. 
He explains freight owners’ disadvantages because of these circumstances. 
Through this study, Verburgh’s pioneering work has set a clear and 
unambiguous benchmark for road transport evaluation.  However, by not 
following through on a wider modal balancing approach, his study does not 
succeed in providing long-term infrastructure planning frameworks. At the 
same time, though, Verburgh’s work, which is 50 years old at the time of the 
candidate’s research, is the first and last attempt in the history of macro-
transport logistics measurement in South Africa that tries to develop a detailed 
strategic perspective. 
 
4.4 Transport and distribution logistics during the market economy 
phase 
4.4.1 Introduction 
A physical infrastructure landmark that inaugurated the market economy age 
in the first world is the “American interstate system of highways” bill that was 
passed by congress in 1956.
178
 Another landmark of this era is the 
introduction and, by the end of the 1960s, widespread use of container 
technology.
179
 It is ironic that, although the combination of freeway calibre 
roads, containers and installed technologies, i.e. railways and shipping, 
clearly pointed the way towards a new era of intermodalism, South Africa 
never really caught on to this concept. Intermodal systems benefit all 
stakeholders (business, the public, freight owners, distributors and freight 
service providers), but as Alan Jorgenson maintained in 1999: 
                                                 
178
 Coyle et al. p. 6. 
179
 Baluch, 2005, p. 126. 
 82
This potential, already successfully implemented in the America’s, 
Europe and Australasia, has unfortunately, not been realised in 
Southern Africa.
180
  
 
The initial reason for this is provided by Mitchell, who says: 
During the first two decades of this new regime
181
 a programme was 
put into place to restore and upgrade infrastructure which had become 
run down during the war years. However….no coordinated, integrated 
or even rational approach to the provision of infrastructure was 
apparent – both between modes and between authorities. The general 
approach was mainly ad hoc responses to specific issues, and most 
definitely there was no overall transport infrastructure funding strategy 
in place.
182
 
 
Two reasons could be postulated for this state of affairs, i.e. that strategy 
follows vision and research
183
 (which were both absent to some degree) and 
that intermodalism of this nature in South Africa would have required macro-
economic intervention, which was absent. 
 
The issue of the relative share of “high-rated” traffic on railways was finally 
understood to some extent in this era as rail’s tonnage share of said traffic 
rose from around 18% in the beginning of the 1970s to about 20% in the 
middle of that decade and then fell to 12% at the end of that decade. Revenue 
share for this traffic declined from around 50% to about 40% during the same 
time period.
184
 
 
As far as road transport and intermodalism are concerned, Jones reports that 
during this era: 
                                                 
180
 Jorgensen, 1999, p. 1. 
181
 Mitchell refers to the regime of the Nationalist government that came into power in 1948, 
soon after the Second World War, and therefore refers to the 1950s and 1960s. 
182
 Mitchell, 2004, p. 2. 
183
 Havenga and Hobbs, 2004, p. 58. 
184
 Jones, 1999, p. 181. 
 83
A detailed picture of modal shares in land transport has been 
notoriously difficult to construct in South Africa, principally because of 
the paucity of robust road transport data.
185
 
 
This apparently leads to a similar “paucity” in his own work on intermodal 
discussion (irrespective of the title of the relevant chapter).  Unfortunately, 
both the absence of intermodal discussions and the lack of robust road 
transport data remain a problem to this day, despite the efforts of some 
researchers and Statistics South Africa.  It is one of the objectives of this 
study to suggest solutions to these issues, as will be discussed in later 
sections of this work. 
 
This era, when the market economy took off, was a time when the United 
Nations was thinking about containerisation and its role in intermodality, and 
especially its role in developing countries, where infrastructure development is 
expensive: 
Therefore, any programme for the improvement of transport should be 
considered in the context of the entire economic and social system of 
the area concerned and drafted within the general development 
programme of the area. Transport studies and surveys of any kind 
should not be regarded as isolated actions, but guided by the policy and 
targets of a long-range transport plan based on the future transport 
needs and available resources.
186
 
 
4.4.2 Verburgh’s 1968 contribution 
In 1968 Verburgh, 10 years after his initial work, is again one of the only 
researchers in South Africa who considers transport data and attempts a 
research-driven approach to current and future problems. The opening 
statement of his work is a landmark in itself: 
                                                 
185
 Jones, 1999, p. 186. 
186
 Traut, 1967, p. 8. 
 84
The pre-requisite to decision-making on matters of transport policy, 
particularly where the infrastructure is involved, is a well-conceived 
anticipation of not only the long-term development of the traffic, but 
also the distribution of this traffic among the various forms of 
transport.
187
 
 
He also says that: 
One of the major problems encountered in the process of planning the 
nation’s transportation facilities was the lack of basic data of 
tendencies in respect of the volume, composition and ‘desire lines’ of 
the demand for transport and goods.
188
 
Evidently this is a problem with a 40-year history and no solution yet. 
 
Verburgh suggests through this statement a possible change in his 
Archimedean Point. He believes research data should not only be developed 
to inform road liberalisation, but that infrastructure planning, based on future 
demand determination, should also play a role. The forecast nature of this 
work confirms his change in approach. The work, however, is based on 
another critical assumption, namely that it is limited to middle- and long-
distance transport and that this transport is provided by rail as rail transport is 
sufficiently representative of the total demand.
189
 
 
The main thrust of Verburgh’s study was to determine the adequacy of rail 
infrastructure over the 10 years following his study. It is therefore possible to 
test Verburgh’s forecast for rail transport for this 10-year period,
190
 compared 
to the actual rail data.  The deviations from his forecasts are represented in 
Table 8. 
 
 
 
                                                 
187
 Verburgh, 1968, p. 1. 
188
 Verburgh, 1968, Foreword. 
189
 Verburgh, 1968, p. 1. 
190
 Verburgh, 1968, p. 61. 
 85
Rail forecast Rail actual % Deviation 
 
Mining Ag
ricu
ltu
r
e 
Ma
nufa
c
t
uring 
Mining Ag
ricu
ltu
r
e 
Ma
nufa
c
t
uring 
Mining Ag
ricu
ltu
r
e 
Ma
nufa
c
t
uring 
1965 50 491 15 854 24 521 50 491 15 854 24 521 0 0 0 
1966 51 709 16 171 25 222 52 666 14 478 25 071 2 -12 -1 
1967 52 925 16 488 25 922 54 347 16 804 25 300 3 2 -2 
1968 54 142 16 805 26 623 54 159 20 010 26 381 0 16 -1 
1969 55 360 17 122 27 323 57 082 19 568 27 396 3 13 0 
1970 56 576 17 412 28 025 59 383 16 716 29 321 5 -4 4 
1971 57 793 17 756 28 726 57 813 14 419 27 724 0 -23 -4 
1972 59 011 18 073 29 426 58 244 15 846 28 282 -1 -14 -4 
1973 60 227 18 390 30 127 60 053 18 378 28 134 0 0 -7 
1974 61 444 18 707 30 827 61 466 14 903 29 323 0 -26 -5 
1975 62 662 19 024 31 528 64 078 17 125 29 535 2 -11 -7 
 
Table 8: Verburgh's forecasts for rail transport in tons (‘000) vs. actual data
191
 
 
The cyclical nature of agricultural demand and supply, and its effect on 
forecasting, is clearly evident from the data (exacerbating rail infrastructure 
planning and cost coverage problems). Mining forecasts seem to be generally 
in line, but manufacturing transport by rail shows a steady decline when 
compared to forecasted volumes. This could be an indication of what was 
previously discussed regarding Verburgh’s first study and the incorrectness of 
the core assumptions around modal roles, the poor infrastructure decisions by 
government and the lack of multi- and intermodal planning.
192
 It could, 
however, be that manufacturing (and other sectoral) production lagged 
forecasts. It is therefore necessary to compare Verburgh’s forecasts of total 
production in the economy with actual data.  
                                                 
191
 Verburgh’s data had to be normalised to fit actual rail tonnages for 1965. The 
normalisation is applied to the complete subsequent analysis of the data and is therefore 
comparable. 
192
 Both multimodal (the provision of the applicable modes to supply transport effectively) and 
intermodal (the supply of effective linkages) are important in this regard. 
 86
Table 9 represents Verburgh’s forecast for physical production in the 
economy for the ten years following his study,
193
 the actual data and deviation 
from forecasts. 
 
Economic production: forecast Economic production: actual % Deviation 
 
Mining 
Agriculture 
Manufac
turi
ng 
Mining 
Agriculture 
Manufac
turi
ng 
Mining 
Agriculture 
Manufac
turi
ng 
1965 76 763 17 835 26 081 76 763 17 835 26 081 0 0 0 
1966 80 309 19 143 26 976 79 025 22 612 28 126 -2 15 4 
1967 84 017 20 547 27 901 83 239 19 427 29 521 -1 -6 5 
1968 87 898 22 054 28 859 87 273 20 701 32 543 -1 -7 11 
1969 91 958 23 672 29 848 90 910 20 701 34 774 -1 -14 14 
1970 96 204 25 408 30 874 98 109 23 568 37 052 2 -8 17 
1971 100 647 27 272 31 933 81 839 24 842 38 493 -23 -10 17 
1972 105 295 29 272 33 028 103 682 20 383 42 026 -2 -44 21 
1973 110 158 31 420 34 160 110 474 20 064 44 723 0 -57 24 
1974 115 250 33 725 35 334 123 154 26 115 46 489 6 -29 24 
1975 120 568 36 198 36 546 133 902 26 434 47 605 10 -37 23 
 
Table 9: Verburgh's forecasts for economic production in tons (‘000) vs. actual data  
 
The evidence is quite clear: by the end of the 10-year period rail transported 
7% less manufacturing commodities than forecasted by Verburgh, but 
production in the manufacturing sector outstripped Verburgh’s projections by 
23%. The market economy had truly arrived in South Africa.  The 
manufacturing economy was doing everything it could to move away from 
subsidising farmers by creating its own transport and logistics solutions and 
highlighting the poor state of intermodal planning by government. 
 
Verburgh’s final conclusions are interesting. He does not consider modal 
balancing, but he does talk about the adequacy of rolling stock, fixed 
infrastructure and marshalling yards (although he suggests mechanisation)
194
 
                                                 
193
 Verburgh, 1968, p. 61. 
194
 Verburgh, 1968, p. 82-83. 
 87
and sounds a word of warning that none of these can work effectively without 
adequate manpower, management and planning.
195
 
 
4.4.3 Smith’s contribution 
In the early 1970s Verburgh’s initial work of 1958 was repeated in detail by 
Smith. Smith cites difficulties at the Department of Statistics, high percentage 
of business mortalities and the “stringent limitations” of the effects of the 1930 
Act which regulated road freight transport as the core issues that led to his 
study.
196
  Similar to Verburgh’s work of 1957, Smith’s work was based on 329 
returned questionnaires (Verburgh 403) from a sample of 4,000 out of the 
deemed 12,000 road hauliers, which resulted in a response rate of 8% 
(Verburgh 16%). (Smith’s study is silent on many of the ancillary or freight-
owner transport statistics that Verburgh did report on, but it was possible to 
reconstruct some trends and data from his base data.) The study was 
therefore essentially a supply-side study, as was Verburgh’s. 
 
Recreating core statistics from Verburgh’s data (which Smith – although 
referring to Verburgh extensively – interestingly doesn’t do) made it possible 
to do the comparisons between the two studies for: 
1. Ton and ton-kilometer (Table 10) 
2. Market share (Table 11) 
3. Average transport distance (Table 12) 
 
 
 
 
 
 
 
 
                                                 
195
 Verburgh, 1968, p. 83-84. 
196
 Smith, 1973, p. 1-2. 
 88
Million Tons Million Ton-kilometers 
 
 
Rail 
Road for 
reward 
Road 
ancillary 
Total 
Rail 
Road for 
reward 
Road 
ancillary 
Total 
Verburgh – 1957 75 67 113 255 32 000 1 955 1 960 35 915 
Smith – 1971 100 245 279 624 53 000 7 850 6 387 67 237 
 
Table 10: Modal market share
197
 
 
In fourteen years tons shipped more than doubled and ton-kilometers nearly 
doubled. Road activity grew exponentially. 
 
Tons - % market share Ton-kilometers - % market share 
  
  
Rail 
Road for 
reward 
Road 
ancillary 
Total 
Rail 
Road for 
reward 
Road 
ancillary 
Total 
Verburgh – 1957 29 26 44 100 89 5 5 100 
Smith – 1971 16 39 45 100 79 12 9 100 
 
Table 11: Modal market share in percentage 
 
Rail market share declined markedly, even considering that it came from a 
low base; both road for reward and ancillary road market share doubled.  The 
growth in average transport distances compared to the faster growth in tons 
shipped versus ton-kilometers performed is counter-intuitive. This is, however, 
a direct result of the changing structure of the market. 
 
ATD – km 
  Rail 
Road 
for 
reward 
Road 
ancillary
Verburgh – 1957 427 29 17 
Smith – 1971 530 32 23 
 
Table 12: Average transport distance 
                                                 
197
 Smith, 1973, p. 139-140. Ancillary transport statistics calculated from average distance 
and tons per vehicle per annum (p. 144-145), per industry group and vehicle fleet 
composition from growth rates in commercial vehicles, NAAMSA, and certificates issued 
(p. 5). 
 89
Economic performance in terms of gross domestic product per ton shipped 
has now deteriorated to R542 (from R659 in 1957) at constant 1995 prices, 
but in terms of ton-kilometer delivered the performance improved to R5.03 
(from R4.68). This is an interesting phenomenon, because even when the 
average transport distance for all modes increased, the average transport 
distance in the economy decreased from 141 kilometers to 108 kilometers. 
This is a phenomenon of urbanisation, but could also be a lead indicator of 
specialisation, following urbanisation, in a multi-metropole economy that 
should predict corridor formation. 
 
Smith does not attempt the same level or depth of economic analysis as 
Verburgh and no further clarification of his interpretations is possible as he is 
almost completely silent on many of the salient issues that his research could 
uncover. The data do point to an increase in road activity over rail; especially 
as “rail captured” traffic picks up and manufacturing transport does everything 
possible to no longer make use of the rail mode. He does, however, make an 
interesting observation: 
As long ago as 1957, the Minister of Transport indicated that he had 
given instructions to the National Transport Commission to conduct a 
survey of all road traffic. As far as can be ascertained, such a survey 
has never been successfully completed by any Government 
department.
198
 
 
The issues that were overlooked in the previously mentioned research and 
planning were beginning to emerge in the 1970s. Jones refers to a view of the 
“separation of modes as an artificial one”
199
 that emerges in the 1970s, as 
“rail grapples with an increasingly uneasy competitive relationship with road 
transport” and being “increasingly at odds with its changing tariff base.” This 
happens at a time of coal and iron export line development and 
containerisation projects, which drive these changes even further.
200
 
 
                                                 
198
 Smith, 1973, p. 4. 
199
 Jones, 1999, p. 176. 
200
 Jones, 1999, p. 177. 
 90
Jones quotes interesting performance figures for rail in this period, namely 
that tonnages shipped increased by 50% over the decade and ton-kilometers 
by 63%. Average transport distance grew faster, because of the export lines 
and more cross-border traffic. What Jones picks up, however, is that high-
rated traffic’s contribution to tonnages and rail revenue peaked during the 
middle of the decade and then declined. Jones’ illustrative figures are 
depicted in Figure 13. 
 
 
Figure 13: Jones's analysis of traffic (note uneven distribution in ton and income).
201
 
 
The question arises as to whether this is because of market structure or 
market share changes? For the only available comparison, Verburgh’s original 
data are extended to the end of the 1970s and are depicted in Figure 14 (note 
that Verburgh created the first base data sets, but did not analyse the data in 
this fashion). 
                                                 
201
 Jones, 1999, p. 181. 
 91
0
20
40
60
80
100
120
140
160
180
200
1
937
1
938
1
939
1
940
1
941
1
942
1
943
1
944
1
945
1
946
1
947
1
948
1
949
1
950
1
951
1
952
1
953
1
954
1
955
1
956
1
957
1
958
1
959
1
960
1
961
1
962
1
963
1
964
1
965
1
966
1
967
1
968
1
969
1
970
1
971
1
972
1
973
1
974
1
975
1
976
1
977
1
978
1
979
I
n
d
ex 
Mining Agriculture Manufacturing
 
Figure 14:  Rail transport as percentage of transportable GDP (indexed 1937=100) 
 
The flattening of manufacturing “market share” is clearly evident, as is the 
peak that Jones refers to, i.e. the relationship and correlation between 
performance data and market share by proxy, seems to be in place. Jones 
discusses the rate problem in detail:  (i) in terms of price elasticity, the 
railways’ assumption that high-value cargo is price-inelastic and the 
subsequent cross-subsidisation phenomenon, which was by then clearly 
installed; and (ii)  in terms of the difficulty of the railways to continue defending 
this position during the 1970s.
202
  
 
Jones makes no specific attempt to measure market share, but quotes Van 
der Veer’s
203
 market share figures and states that it is a “crude indicator” and 
that “market share figures are notoriously difficult to obtain”. At least by now 
the principle of integrated and inter-modal planning as a necessity has been 
confirmed and has the fact that information is an important prerequisite to 
make it work. 
                                                 
202
 Jones, 1999, p. 183-184. 
203
 Van der Veer, 1982, pp. 19-26. 
 92
Van der Veer compares 1972 to 1981 and reports the following market share 
figures (Table 13): 
 
 Rail Road for reward Road ancillary 
1972 61% 4% 35% 
1981 51% 12% 37% 
 
Table 13: Van der Veer's market share figures of 1981 
 
Van der Veer’s figures include pipelines in road for reward, but because this 
figure is known, as is rail tonnages, a complete picture (which he didn’t draw), 
comparing his figures with those of Verburgh can be calculated (Table 14). 
 
Million Tons 
 Rail Road for reward Road ancillary Total 
1957 – Verburgh  75 67 113 255 
1972 – Van der 
Veer 102  7  59 167 
1981 – Van der 
Veer 176 41 128 345 
 
Table 14: Comparison of Van der Veer's 1981 figures with those of Verburgh 
 
Jones warns that Van der Veer’s data should be interpreted with great 
caution, but they are the best that he could find. Van der Veer’s calculations 
are unknown, but the emerging problem according to Jones is quite clear, i.e. 
that intermodal planning is important (to solve the tariff issue) and that reliable 
data do not exist.  
 
By the beginning of the 1980s the cross-subsidisation issue and the tariffing 
process became rather ridiculous and a self-fulfilling disaster, as the age of 
logistics and total cost of ownership approached. Kennedy quotes a road 
transport board official who maintains that: 
 93
...if a company needs road transport to stay in business, it is a marginal 
firm and probably should not be in business anyway.
204
 
 
He continues to cite another event where: 
...an applicant at another board who justified his application with 170 
pages of railway documentation showing erratic transit times and 
substantiated damage claims paid by the railway. It was also shown 
that the actual loss incurred as a result of the frequent damage to the 
commodity was twice the amount paid by the railways in claims. Using 
this evidence, the applicant showed that by using the railway, his profit 
margin on his product fell from 56 cents to 6 cents per unit.
205
 
He got a permit. 
 
By now cross-subsidisation became untenable, as service standards of the 
railways fell behind at the same time that the road mode increasingly became 
a viable option. This clearly caused a road vs. rail debate rather than a road 
and rail debate, especially as a result of the absence of research and 
strategy. 
 
4.4.4 Hamilton’s first study 
In 1983 Hamilton repeats Verburgh’s methodology in response to a request 
from the South African Transport Services (who were then the operator of 
South Africa’s railway) organisation. He states that the survey is needed for 
planning and that a geographical distribution of the results will also be 
attempted, as well as commodity and route descriptions.
206
 Two other 
important aspects about Hamilton’s survey approach and mandate need to be 
mentioned: 
• The Road Freight Association, which initially promised co-operation, 
withdrew their support; 
                                                 
204
 Kennedy, 1982, p. 19. 
205
 Kennedy, 1982, p. 19-20. 
206
 Hamilton, 1983, p. 1-2. 
 94
• Hamilton questions the validity of the Central Statistical Services’ list of 
professional hauliers – an important consideration for later 
developments. 
Both these incidents led to a low response rate for his survey.
207
 
 
Hamilton’s survey is once again a supply-side survey of road hauliers and 
ancillary hauliers in South Africa. He considers South Africa’s commercial 
vehicle population and professional haulier addresses as retrieved from the 
Central Statistical Service, and the addresses of ancillary transport users as 
received from various sources. It is important to note that the addresses were 
stratified and that no stratification based on certificates, i.e. the more detailed 
stratification data available based on exemption certificates issued to road 
hauliers, (which proved to be differentiated and therefore important), was 
done.  
 
By the time that Hamilton completed his survey he realised that his sample 
was too small and erratic to distinguish between road for reward and ancillary 
transport. He therefore assumes his figure as a total figure and determines 
ancillary transport by subtracting the Central Statistical Services’ figure 
(supposed to be the transport for reward figure) from his total figure. The net 
result of this was that Hamilton never reported on commodity or route 
distribution, and that his usage of the Central Statistical Services’ information 
to determine ancillary transport figures is highly suspect. The summarised 
results were startling (when compared to Verburgh’s 1957 survey and Smith’s 
1971 survey – refer Table 15. 
 
 
 
 
 
 
 
                                                 
207
 Hamilton, 1983, p. 4-5. 
 95
Million Tons Million Ton-kilometers 
  
  
Rail Road for reward Road ancillary Total Rail Road for reward Road ancillary Total 
Verburgh - 
1957 75 67 113 255 32 000 1 955 1 960 35 915 
Smith – 1971 100 245 279 624 53 000 7 850 6 387 67 237 
Hamilton - 1981 185 239 2 204 2 628 98 406 27 515 136 007 261 928 
 
Table 15: Hamilton's modal market share figures compared to Verburgh and Smith 
 
Since Hamilton’s data are suspect, it was decided to analyse his findings in 
more detail, in order to learn from possible problematic issues in his 
approach. 
 
Hamilton provides a breakdown of fleet sizes, number of vehicles per fleet 
size per category, tonnage for November 1981 (the month in which the 
sample was surveyed), carrying capacities per vehicle group in each fleet size 
and average transport distance per vehicle group in each fleet size (Table 
16).
208
  
 
Fleet size Trips per day Average trip length 
1 7 25 
2-10 16 67 
11-20 5 110 
21+ 12 73 
 
Table 16: Fleet performance calculations based on Hamilton's 1981 data 
 
From these data vehicle performance data per fleet size could be calculated. 
It is inconceivable in 1981 that this level of productivity could be achieved, i.e. 
that a vehicle can be loaded and off-loaded 16 times per day (i.e. 32 load 
handlings) and travel 67 km between each loading. As in Smith’s study of 
1971, Hamilton provides little interpretation of his results, especially in the 
                                                 
208
 Hamilton, 1983, fleet sizes (p. 32), number of vehicles per fleet size per category (p. 26-
31), tonnage for November 1981 (p. 37), carrying capacities per vehicle group in each 
fleet size (pp. 26-32). 
 96
macro-economic sense, but urges policy makers to improve information 
gathering.
209
 
 
In the last decade of what could be described as the development of a market 
economy, between 1980 and 1990, quite a few changes occurred in South 
Africa.  
 
The new General Manager of the South African Railways and Harbours, Bart 
Grové, a transport economist, took over from Kobus Laubscher (a mechanical 
engineer). Grové promised to change the organisation into a market-driven 
business.  One of the key initial steps was to change its name to South 
African Transport Services to ensure the inclusion of all modes in this new 
market-driven vision.  The key objective for the railway specifically was to 
prepare it for competition. 
 
Kuhn’s overview provides an overview of the work of the National Institute for 
Transport and Road Research at the CSIR, and discusses new government 
priorities that require the transport system to be better utilised.
210
 Kennedy did 
a detailed analysis of road and rail costs in 1984 and highlights the two 
reasons for road transport regulation, i.e. to protect the railway’s traffic base 
and delay the deterioration of the road network.
211
 He highlights the fact that 
the railways had not been balancing their books during preceding years (due 
to cross-subsidisation within rail and from other modes within Transport 
Services).  According to him, the transport system was therefore not working.  
He suggested that road levies charged at true usage costs will protect the 
road system.
212
 
 
Kennedy’s work is extensive, but does not consider externalities and makes 
an interesting comment on the resultant traffic loss of the railways (if full 
deregulation is implemented and full costs are charged to both the railways 
                                                 
209
 Hamilton, 1983, p. 22. 
210
 Kuhn, 1983, p. 201. 
211
 Kennedy, 1984, p. 57. 
212
 Kennedy, 1984, p. 57-58. 
 97
and road). The comment refers to the loss of income of the railways and the 
resultant effect on the railways’ cost base: 
A very short time after such traffic losses, several days or weeks, there 
will be likely to be very little cost reduction, apart from reduced energy 
usage and damage inflicted upon the track structure. After several 
months, if the traffic reductions are significant on certain routes, the 
scheduling of trains, assignment of motive power and the organisation 
of yard and station forces can be adjusted to meet the new traffic 
levels.
213
 
 
Kennedy ignores the fact that most rail costs are fixed over the long term 
(they are sunken to some extent) and that all railways require high densities 
for them to be successful. He does not consider intermodality and future 
freight levels that could possibly change the income structure of a railway over 
dense long-haul corridors. 
 
4.4.5 Hamilton’s second study 
In 1985 Hamilton repeated his study of 1981, once again for the sole purpose 
of defining the extent of the road haulage industry. He admits shortcomings in 
the 1981 study, which he attempts to overcome. His objective is again to 
determine origin/destination data by commodity, as well as the modal 
structure of all surface freight in South Africa (except finally consumed 
product).
214
 Hamilton mentions in passing why some of the conclusions of the 
1981 survey were suspect, i.e.: 
• That fleet sizes were used as a sampling unit rather than vehicles, and 
that the number of operators per fleet size was difficult to establish; 
• That vehicles with a gross mass of less than 5 ton are often registered 
as utility vehicles or in the name of private persons or other non-
                                                 
213
 Kennedy, 1984, p. 100. 
214
 Hamilton, 1986, p. 1. (Hamilton defines public road transport as a transport operator who 
possesses a motor transport certificate and whose main activity is transport for reward, 
and ancillary road transport as a business institution trading for profit but who transports 
his own goods not for reward.) 
 98
transport bodies such as municipalities, statutory bodies, vehicle 
dealers, etc. (i.e. part of transport of finally consumed produce or for 
other purposes).
215
 
 
Hamilton’s approach is once again to survey road for reward and ancillary 
operators with a questionnaire, which means that the basic supply-side 
approach did not change from Verburgh, Smith and his own first study. He 
achieves a high response rate of 13.5% (due to a shorter questionnaire and 
follow-up interviews), but this response rate represents only 113 satisfactorily 
completed questionnaires.
216
 
 
Hamilton’s survey indicates the following modal market share results 
(compared to Verburgh, Smith and his own first study - Table 17). 
 
Million Tons Million Ton-kilometers 
 
 
Rail Road for reward Road ancillary Total Rail Road for reward Road ancillary Total 
Verburgh – 1957 75 67 113 255 32 000 1 955 1 960 35 915 
Smith – 1971 100 245 279 624 53 000 7 850 6 387 67 237 
Hamilton – 1981 185 239 2 204 2 628 98 406 27 515 136 007 261 928 
Hamilton – 1985 170 263 241 674 91 861 31 297 14 219 137 377 
 
Table 17: Modal market share in 1985 (Hamilton)
217
 
 
The modal market share in tons can also be expressed in percentage terms 
(refer to Table 18). 
 
 
 
 
 
                                                 
215
 Hamilton, 1986, p. 3-4. 
216
 Hamilton, 1986, p. 7. 
217
 Hamilton splits tons between road for reward and road ancillary, but not ton-kilometers. 
Based on previous studies’ split between the ATD of these two categories, a rough 
estimate of the ton-kilometer split was possible. 
 99
Tons – (% market share) Ton-kilometers – (% market share) 
  
  
Rail Road for reward Road ancillary Total Rail Road for reward Road ancillary Total 
Verburgh – 1957 29 26 44 100 89 5 5 100 
Smith – 1971 16 39 45 100 79 12 9 100 
Hamilton – 1981 7 9 84 100 38 11 52 100 
Hamilton – 1985 25 39 36 100 67 23 10 100 
 
Table 18: Hamilton's modal market share % of 1985 
 
From the available data, average transport distances can also be calculated 
(Table 19). 
 
ATD – km 
  Rail 
Road for 
reward 
Road 
ancillary
Verburgh – 1957 427 29 17 
Smith – 1971 530 32 23 
Hamilton – 1981 532 115 62 
Hamilton – 1985 540 119 59 
 
Table 19: Hamilton's average transport distance per mode 
 
The results of the 1985 study seem to indicate an arrest of the deterioration of 
the railways’ performance, apparently confirmed by an improvement in market 
share. However, if the approximately 40 to 50 million tons transported on the 
export lines at this stage (the export lines had begun functioning since the 
1971 survey) is taken into account, as well as the natural growth in other 
mining commodities, this is not the case and rail market share remained flat at 
best. In fact, at this stage the original Verburgh data sets of 1958 (Figure 14) 
could be considered again and extrapolated up to 1985 to get a better picture 
of market share by proxy for the three transportable economic sectors (Figure 
15). 
 100
0
20
40
60
80
100
120
140
160
180
200
19
37
19
38
19
39
19
40
19
41
19
42
19
43
19
44
19
45
19
46
19
47
19
48
19
49
19
50
19
51
19
52
19
53
19
54
19
55
19
56
19
57
19
58
19
59
19
60
19
61
19
62
19
63
19
64
19
65
19
66
19
67
19
68
19
69
19
70
19
71
19
72
19
73
19
74
19
75
19
76
19
77
19
78
19
79
19
80
19
81
19
82
19
83
19
84
19
85
In
d
e
x 
Mining Agriculture Manufacturing
 
Figure 15: Extrapolation of Verburgh's original 1957 data sets (indexed 1937=100) 
 
These figures do point towards a flattening of the downward trend of rail 
performance since 1957 during the 1970s and early 1980s, and therefore 
seem to correspond well with Verburgh’s, Smith’s and Hamilton’s second data 
sets. 
 
If Hamilton’s first study is ignored and the second study accepted, initial 
performance calculations are possible (Table 20). 
 
Year GDP per ton GDP per ton-km ATD 
1957 R659 R4.68 141 
1971 R542 R5.03 108 
1985 R717 R3.52 204 
 
Table 20: Economic performance related to transport input 
 
The trend from 1957 to 1971 is now exactly reversed from 1971 to 1985. GDP 
per ton improved, but deteriorated per ton-kilometer. This is attributable to the 
 101
building and commissioning of mass low-value export systems, which are ton-
kilometer “hungry”, but produce low relative income. 
 
Once again Hamilton makes no strategic deductions from his survey or offers 
any strategic infrastructure suggestions. He does ask for the co-operation of 
associations such as the Public Carriers Association and Association for 
Private Transport Owners, which was once again lacking during the 
research.
218
 In spite of this he believes that the study will now be repeated 
often.
219
 
 
As the world was changing at the end of the 1980s from the bi-polarity of two 
superpowers and the Cold War between them into a more “open village”, 
systems changed into collaborative and integrated electronic communities. 
South Africa experienced multiple changes in society and the market 
economy slowly changed into a market, channel and cluster economy. Many 
economic clusters (such as grain, fruit, wine and transport) were preparing for 
deregulation, government businesses (such as Sasol, Iscor and the S. A. 
Transport Services) for privatisation, and the economy as a whole for 
globalisation.  
 
At this time the transport sector had little understanding of where it fitted into 
this picture. Research was scant, unformalised, not trusted and poorly 
managed from a central point (different, for instance, from that of the clusters 
such as mining, fruit, chemicals and fuel). Researchers provided little strategic 
context for their work and hardly ever showed how it should be applied and 
what role it should play in planning. It was accepted practice to some extent 
that the researcher should be distant from application and context, but this 
only works if central planning is effective. Transport was poorly managed in 
South Africa from a central point of view and this affected all aspects of the 
management cycle. Research, claimed by all researchers to be important, 
therefore had to show for itself why it mattered - and it was not able to. 
 
                                                 
218
 Hamilton, 1986, p. 28. 
219
 Hamilton, 1986, p. 28. 
 102
Smith talks about freight transport in 1980: 
The conveyance of goods is not a subject that is receiving much public 
attention in the Republic…..In the Republic research has revealed that 
heavy vehicle movement problems are largely in urban areas.
220
 
 
He doesn’t specify the research that he refers to, but the absence of real 
guidance for macro-planning from proper research is clearly evident. The 
densification of corridors, problems with high-rated traffic and a huge outcry 
for exemption certificates from road users, if not total deregulation, was taking 
place and it is hardly conceivable why it did not lead to a more critical view of 
the total freight transport system, except for the fact that the research society 
itself did not sound any alarm bells. 
 
4.5 Integrated trading platforms during the establishment of 
electronic communities 
4.5.1 Introduction 
Banfield “announces” this era of electronic integration with a statement about 
transformation in strategic supply management: 
Fundamental change – nothing short of transformation – has occurred. 
The result of this transformation is a new way of doing business. 
Collaboration is the key behaviour change.
221
 
 
The customer-centric behaviour of the market era never changed, though. 
More than ever customers were important, but the benefit of collaboration was 
cost cutting, often achieved through supply-chain streamlining as internal 
process views took root, redundant work was eliminated and different 
organisations within the supply chain gained clarity about their roles.
222
 
 
                                                 
220
 Smith, 1980, p. 10. 
221
 Banfield, 1999, p. 4. 
222
 Banfield, 1999, p. 13. 
 103
A core driver is the information revolution. New communication technologies 
impact on logistical improvements to the same degree as the microprocessor 
did a few decades ago.
223
 Bar-coding, electronic data interchanges and 
tracking and tracing technologies provide opportunities for process integration 
which remains the “one aspect of logistics that continues to improve.”
224
  
 
In this era most industries rely on information technology as a core driver of 
change, but these industries also have other core drivers unique to each 
industry to drive intra-industry (cluster) behavioural changes. In transport 
logistics the global era of containerisation and intermodality has arrived as the 
tool for collaboration in higher value and seamless logistics. Containerisation 
revolutionised freight during the preceding decades,
225
 but it came into full 
global play in this era and was shown to have cost advantages in all aspects 
of intermodalism.
226
 
 
This improvement from secrecy and cut-throat competition to collaboration 
presented opportunities for an improved transport interface not only between 
marine and overland modes, but also between overland modes.
227
 
 
In the 1980s Coyle et al. described how “intermodal or multimodal 
transportation has been long hailed as offering significant potential to improve 
transportation services with lower cost.”
228
 They do, however, refer to 
“potential that wasn’t attained though”; they attribute the problem to regulation 
and emphasise how this failure to achieve efficiencies through intermodal 
transport was beginning to “change vastly” with deregulation. They predicted 
that the “time was ripe for significant growth in intermodal traffic” and that the 
“environment was right for intermodal advances.”
229
 
 
                                                 
223
 Bowersox and Closs, 1996, p. 16. 
224
 Bowersox and Closs, 1996, p. 16. 
225
 Baluch, 2005, p. 127. 
226
 Kuo, 1987, pp. 86-89. 
227
 Jorgensen, 1999, p. 1. 
228
 Coyle et al., 1986, p. 493. 
229
 Coyle et al., 1986, p. 494. 
 104
Globally freight developed around intermodality, regional logistics hubs and 
integrated road, rail and sea movements. Intermodality in domestic transport, 
such as in Europe, meant accompanied and unaccompanied
230
 combined 
transport movements
231
 that contributed to seamless collaboration, 
decongestion and cost efficiency in the networks. 
 
Globally the information era also revolutionised research in all aspects of 
transport logistics. In North America a special agency was created that 
reported to the President and Congress annually on modal market share, 
intermodal developments and infrastructural requirements. Research in 
Europe reached levels where the European Commission was able to 
understand the exact trade-off between the levels of congestion and rail 
subsidies in Europe. Nelldal et al. propose that Europe should consider even 
shorter distance intermodality: 
In the US the concept of large-scale systems is well-developed – long 
trains carrying Double-Stack Containers, i.e. two container layers on 
each wagon. This is working well with large traffic volumes over long 
distances, in fact very much like a container ship ashore! As to Sweden 
and other parts of the European continent, it should therefore be of 
interest to have alternative concepts also, working jointly with trucking 
over shorter distances and offering an increased number of 
destinations, where the major markets are.  
 
By means of “liner train traffic”, meaning that the train makes a number 
of intermediate stops under way, a larger market share may be 
achieved. This requires simple terminals situated at train sidings with 
both ends connected to the main line, so that the train will not have to 
be shunted. An example of this is the “Light Combi” system employed 
in Sweden. A fork lift truck is carried in each train, and this is operated 
                                                 
230
 Intermodal transport is mostly referred to in Europe as combined transport and categorised 
as either accompanied (when drivers accompany their trucks that were loaded on a 
rolling highway train – “Rollende Landstrasse”) or unaccompanied (when containers, 
trailers or swap bodies are lifted on trains and therefore disconnected from truck 
tractors). 
231
 Baluch, 2005, p. 130. 
 105
by the loco driver himself, who loads and unloads the swap bodies.  
“Light Combi” is in fact used for sending consumer goods – a market 
that was lost by most railways long ago.  
 
The “Light Combi” concept may be competitive within the 300 to 400 
km range, while traditional intermodal concepts often require twice that 
distance. Note that “Light Combi” may be the answer to two different 
needs – satisfying smaller traffic flows over short and medium-length 
hauls, or being a feeder to heavy-haul intermodal traffic. By adopting 
“Light Combi”, traditional intermodal services may be concentrated to 
large terminals only, improving their profitability.
232
 
 
In South Africa progress is painfully slow in terms of intermodality. During the 
1990s no direction was given and no research developed on which to base a 
change in policy. 
The current policies are not effective for a variety of reasons including 
previous distortions in the policy formulation process, in that modes 
were considered in isolation. Ineffective implementation of policy, and 
political and sectional pressure, all led to a lack of a holistic, 
sustainable, multimodal transportation policy.
233
 
 
The railway is considering change and many business cases for intermodal 
options are considered, but all eventually rejected. Only the maritime 
container option remains, and domestic container movements are all but non-
existent. By the end of the 1990s Parker interviewed Leo Petkoon, then the 
container terminal and operations General Manager of the railways: 
Petkoon said that by leaving road transport to its own devices, 
Spoornet had put its greatest emphasis on its own ability to provide 
mass transport of bulk loads and sensible economies of scale over 
long and even medium distances. ‘The ideal solution for many 
customers is an intermodal one, combining the benefits of road and 
                                                 
232
 Nelldal et al. 2000, p. 1. 
233
 TMT Projects, 1995, p. 1. 
 106
rail’. A harmonious relationship between the two modes is not only 
possible, but is in fact a ‘match made in heaven’. It will be a positive 
step if national transport policy encourages this kind of relationship.
234
 
 
In other words, the railways were waiting for national policy to indicate a way 
forward and support this development. Moving South Africa did address this 
issue, but a lack of research and government inertia meant that nothing was 
done. The Minister of Transport’s intermodal cargo transfer policy discussion 
document of 2006 does little to provide the direction required and once again 
focuses on mode balancing rather than real intermodalism. The policy 
proposal merely requests that transporters “migrate, where possible, certain 
categories of cargo from road haulage back to rail haulage.”
235
 
 
In 1990 Pretorius referred to the Demand Modelling Work Group’s (a sub-
committee appointed by the National Transport Commission to study rural 
road needs) proposal, in 1982, to the effect that a national flow model be 
developed but that a shortage of funds meant that it was never done.
236
 
 
4.5.2 Freight Transport Data Bank 
Towards the late 1980s the Research Unit for Transport Economic and 
Physical Distribution Studies of the Rand Afrikaans University developed a 
database to fill the gap for “reliable information relating to the macro as well 
as the micro economic aspects of freight transport activities.”
237
 The report 
provides the first reference to “macro” issues; it is extensive, provides some 
forecasts and develops indicators. The samples on which the surveys were 
based were initially extensive – close to 40 000 in 1985, 1986 and 1987 – but 
were reduced (after “more knowledge about the composition of the population 
was obtained”) to below 10 000 in 1988, 1989 and 1990.
238
 The stratified 
                                                 
234
 Parker, 1999, p. 1. 
235
 Transport Analysis Unit, 2006, p. 21. 
236
 Pretorius, 1990, p. 20. 
237
 Pretorius, 1991, p. 1. 
238
 Pretorius, 1991, pp. A.1-A.7. 
 107
random sample considered population, gross geographical product and 
commercial vehicle population.
239
 
 
The results of the work were extensive and added for the first time “Rental” 
transport (i.e. ancillary transport, but not with its own equipment) to the road 
modes. Two important transport performance parameters were unfortunately 
missing, i.e. ton-kilometers and origin-destination pairs. The results for tons 
on a macro level are shown in Table 21 (rental is excluded from this table as it 
was small and not comparable to other studies). 
 
Million Tons 
  Rail 
Road for 
reward 
Road 
ancillary Total 
1985 176 239 363 778 
1986 131 220 412 763 
1987 157 199 437 793 
1988 187 168 414 769 
1989 155 134 382 671 
1990 164 131 353 648 
 
Table 21: Pretorius' modal market share 1985-1990
240
 
 
The only known data for this time period are rail tonnages and Pretorius 
depicts the sample’s response and then compares this with actual rail 
tonnages,
241
 but never discusses the dissimilarities between the two data sets 
(which were quite extensive). It seems as if the deterioration in the sample 
size did not influence the reliability of the data too much as rail’s performance 
as reported by the sample undercounts rail’s actual performance by 12% on 
average. (Pretorius does not provide this analysis.) The difference is depicted 
in Table 22. 
 
 
                                                 
239
 Pretorius, 1991, p. A.2. 
240
 Pretorius, 1991, Appendix 2, Table 2.1-2.8. 
241
 Pretorius, 1991, p. 31. 
 108
 
Rail – 
Pretorius 
(mt) 
Rail – 
Actual 
(mt) 
Difference 
(mt) 
% 
Difference 
1985 176 182 6 3 
1986 131 178 47 26 
1987 157 188 31 16 
1988 187 187 0 0 
1989 155 179 24 13 
1990 164 184 20 11 
 
Table 22: Undercounting of rail performance in Pretorius' sample 
 
A comparison of Pretorius’s work with all the previous market share work 
reported is provided in Table 23. 
 
Million Tons 
  Rail 
Road for 
reward 
Road 
ancillary Total 
Verburgh - 1957 75 67 113 255 
Smith – 1971 100 245 279 624 
Hamilton - 1981 185 239 2 204 2 628 
Hamilton - 1985 170 263 241 674 
Pretorius - 1985 176 239 363 778 
Pretorius - 1986 131 220 412 763 
Pretorius - 1987 157 199 437 793 
Pretorius - 1988 187 168 414 769 
Pretorius - 1989 155 134 382 671 
Pretorius - 1990 164 131 353 648 
 
Table 23: Modal market share from 1957 - 1990 based on all available sources 
 
The data represent an interesting picture. If Hamilton’s first survey is ignored 
(because of intrinsic problems identified by analysis and admitted by him), 
Pretorius’s rail data are replaced with actual data, and if road transport data 
are “normalised” with the 12% undercounting observation for rail, the following 
picture emerges (Table 24). 
 
 
 109
Million Tons 
Road for 
reward 
Road 
ancillary   Rail Total 
75 67 113 255 Verburgh - 1957 
100 245 279 624 Smith - 1971 
170 263 241 674 Hamilton - 1985 
182 272 414 868 Pretorius - 1985 
178 251 470 899 Pretorius - 1986 
188 227 498 913 Pretorius - 1987 
Pretorius - 1988 187 192 472 851 
Pretorius - 1989 179 153 435 767 
Pretorius - 1990 184 149 402 735 
 
Table 24: Modal share - Hamilton's first study ignored and Pretorius’ data normalised 
 
The “split” between road for reward and road ancillary provides a good 
approximation of two thirds/one third which, together with propensity to 
outsource, could provide some indications for future studies. At best 
Pretorius’s data could be quoted to say that on the eve of deregulation the 
economy required around 700 to 800 million tons of freight transport, with 
ancillary road accounting for about 50% of the share and road and rail each 
approximately 25%. 
 
Pretorius’s surveys continued for a few years and were then discontinued. 
Once again it is assumed that the lack of strategic context, lack of funding and 
lack of interest from the Department of Transport led to the survey’s demise. 
The data sets and analysis that compare the history of surveying (as provided 
above) were never constructed. These data are also the last semi-reliable 
data that existed before the study in this dissertation was initiated. Another 
application of the data are to correlate the indicated transport volumes with 
physical production of the GDP, i.e. to try and determine an indicator for future 
reference. This indicator will then estimate the number of times that each ton 
of output of the economy is shipped (Table 25). 
 
 
 
 110
Year 
Shipment 
frequency
1957 2.48 
1971 2.45 
1985 1.45 
1985 1.86 
1986 1.93 
1987 1.96 
1988 1.76 
1989 1.57 
1990 1.53 
 
Table 25: Shipment frequency of the transportable GDP 
 
4.5.3 Land freight transport statistical review 
Statistics South Africa also produced a survey of road transport for reward, 
but this survey was discontinued in 2003. The purpose of the survey was to 
sample “private sector enterprises (firms) primarily engaged in the 
conveyance of goods for remuneration in South Africa.”
242
 The survey 
reported figures that were much higher than other available data and it was 
always in dispute. Closer analysis revealed that the sample was dated and 
reverted back to a time when a lot of transport for reward was carried out by 
bus companies and small operators, i.e. before the emergence of mega-
carriers. The results were often based on returns from mega-carriers and then 
extrapolated to account for a number of smaller enterprises who couldn’t or 
wouldn’t respond, or no longer existed. This could account for higher than 
expected figures. 
 
The surveys quote tonnages and gross income for road hauliers for various 
commodities. This makes it possible to calculate rand per ton for the last 
bulletin published in 2003 (Table 26). 
 
 
 
                                                 
242
 Statistics South Africa, 2003, p. 4. 
 111
  
Tonnage 
'000 
Income 
R000 
Rand per 
ton 
Livestock, fresh produce and crops 3 952 101 055 26 
Mining and quarrying products 10 409 96 255 9 
Food, beverages and tobacco products 2 619 144 821 55 
Textiles, clothing and leather goods 1 206 15 692 13 
Wood, products of wood and cork 1 367 40 735 30 
Paper and paper products 810 27 834 34 
Coke, petroleum, chemicals, rubber and plastic products 8 772 227 761 26 
Non-metallic products 4 663 103 263 22 
Basic metal products 932 28 917 31 
Electrical machinery and electronics 2 734 52 768 19 
Furniture 1 581 68 720 43 
Parcels 4 567 326 574 72 
Other goods 2 317 78 296 34 
Total 45 929 1 312 691 29 
 
Table 26: Commodity transport income for reward - Statistics South Africa 2003 data
243
 
 
These figures are completely unlikely for any operator over any distance (the 
cheapest transport in South Africa, i.e. iron ore transport over the Sishen-
Saldanha export machine, was already at this stage achieving more than R30 
per ton). The review also reports the number of vehicles in operation (31 830) 
and kilometers travelled by all these vehicles, laden (128 262 000) and 
unladen (64 748). Ton-kilometers cannot be calculated, but to achieve the 
ton-kilometers reported by Hamilton in 1985, each vehicle had to travel 4 030 
km per month laden and 2 034 km unladen, which is possible, but when laden 
each of the vehicles would have to carry an average of 21 tons all the time, 
which seems highly unlikely. This data was therefore obviously seen as 
flawed and Statistics South Africa discontinued the survey.  
                                                 
243
 Statistics South Africa, 2003, p. 13. 
 112
4.6 Globalised systems during the globally networked economy 
(2000-2020) and the expectations around global bio-engineering 
As South Africa enters the globally networked economy, questions are being 
asked about the competitiveness of the country’s economic system, the 
maintenance of its physical infrastructure and whether South Africa’s transport 
network will support the country in these global challenges. Change is 
necessary not only to fix design inefficiencies and a backlog of economic 
infrastructure supply, but also to prepare for some switches to more 
sustainable forms of transport, but no data exist to inform a new strategy. This 
study aims to fill this gap. 
 
In many countries in the first world the debate around global competitiveness 
has moved on, away from national priorities to global sustainability issues. 
This does not mean that measurement on a macro scale will become less 
important, but rather that the word “macro” will get a new global significance. 
As Kyoto, or at least the principles of Kyoto, become accepted (and 
everybody expects that they eventually will be accepted), nations will have to, 
rather than develop national competitive stances, adhere to national limiting 
norms. Emissions will be tracked and monitored, emissions trading will 
become the norm and the sustainability issues around wealth distribution and 
environmental damage will become more onerous and transparent. 
 
The World Business Council for Sustainable Development admits that 
transport’s functions are essential, but that: 
“...in performing these essential functions, mobility must not sacrifice 
other essential human or ecological values, today or in the future.”
244
 
 
The Council continues to report on seven goals for 2030, i.e.: 
1. Reduced emissions; 
2. Limited greenhouse gasses; 
3. Reduced transport-related deaths and injuries; 
                                                 
244
 World Business Council for Sustainable Development, 2006. 
 113
4. Reduced transport-related noise; 
5. Mitigated traffic congestion; 
6. Narrowed mobility divides between rich and poor; 
7. Improved mobility opportunities in developing societies.
245
 
 
The Council also addresses freight mobility specifically in its extensive 2001 
report and admits the necessity of freight mobility, although it consumes 43% 
of all transportation fuel.
246
 It does, however, raise the same concerns about 
congestion, infrastructure, security, stability and sustainability that exist for 
transport in general. It also proposes three overarching strategies, i.e.: 
1. Cleaner and more fuel-efficient trucks for urban areas; 
2. Intermodal systems; 
3. Modern heavy-haul rail corridors.
247
 
 
Stigson highlights the core requirements for managing these goals and 
strategies, i.e. “informed and well researched descriptions of mobility” and 
“modelling challenges to measuring the gap between where we are, and 
where we want to be.”
248
 
 
South Africa’s measurement backlog will become a real albatross around the 
country’s neck as externalities that the country has absolutely no sense of at 
the moment will also be a consideration in all measurement. The worst 
problem of all is that the network configuration that the country develops today 
(not even based on proper intrinsic data) will be subjected to the sharp edge 
of both intrinsic and extrinsic factors in the next two to three decades. The 
next generation will therefore inherit the worst possible position, if something 
is not attempted right now. 
                                                 
245
 World Business Council for Sustainable Development, 2006. 
246
 Sustainable Mobility Working Group, 2001, p. 6-1. 
247
 Sustainable Mobility Working Group, 2001, p. 6-16. 
248
 Stigson, 2004, p. 1. 
 114
 115
                                                
4.7 Lessons learned 
The history of freight transport measurement on a macro scale in South Africa 
may be chequered, but lessons can be learned to prepare for the future and 
define achievable research objectives. Looking at the history one of the 
greatest lessons is summarised well by Darwin, when he analyses the 
development of intellect and reasoning: 
Little is known about the functions of the brain, but we can perceive 
that as the intellectual powers become highly developed, the various 
parts of the brain must be connected by intricate channels of the freest 
intercommunication; and as a consequence each separate part would 
perhaps tend to be less well fitted to answer to particular sensations or 
associations in a definite and inherited – that is instinctive – manner. 
There seems even to exist some relation between a low degree of 
intelligence and a strong tendency to the formation of fixed, though not 
inherited habits, for as a sagacious physician remarked to me, persons 
who are slightly imbecile tend to act in everything by routine or habit; 
and they are rendered much happier if this is encouraged.
249
 
 
The biggest mistake is therefore not to learn from history, but to keep 
repeating the same mistakes over and over again, and therefore regressing. 
To remain in the same place is not being static; it is falling backwards into a 
deep abyss of incompetence. The lessons learned should therefore be 
collected and understood to arrive at a set of guidelines, as discussed in the 
next chapter. 
 
 
 
 
249
 Darwin, 1901, p. 103. 
Chapter 5 - Research objectives and methodology 
5.1 Introduction 
An understanding of the history of modal market share research, the macro 
issues at hand and the role of the logistics discipline in resolving these issues 
indicate the way towards possible solutions for the continuous measurement 
of macro-logistics performance. Past research in this field, even if considering 
a narrowly defined field in logistics, has been scant and consistently non-
performing.  In addition, even if the field of research within logistics might 
seem small – and to some, even unimportant – significant questions are 
confronting the discipline at this specific point in South Africa’s history. These 
questions include an understanding of current network design and its 
performance, future demand and where investment would be required, how 
performance in this regard will be measured and, finally, how the transition to 
sustainable technologies will be managed.  
 
The fact that these answers are not readily available is an obvious 
shortcoming of the management and strategic planning around South Africa’s 
transport and logistics infrastructure, but the academic management of the 
discipline should surely also take responsibility.  All indications are there that 
these questions will continue to be asked in the future as global challenges of 
disparity between development and scarce resources, between rich and poor 
nations, and between global order and national objectives peak in the next 
two to three decades. 
 
Any attempt to answer these questions should carefully consider the task and 
its challenges, and what might have caused past failures in this regard. 
Lessons from history usually contribute towards better solutions. 
 
 116
5.2 Problems identified in past research 
Macro-economic transport and logistics measurement could at the very least 
be described as having a chequered history. It is important to summarise 
lessons from this history and, based on these observations, identify some 
objectives for future research in this regard. 
 
5.2.1 Methodology 
All past research (including the work by Statistics South Africa that was 
discontinued) utilised a questionnaire-based methodology. The approach is 
valid, but expensive, cumbersome and fraught with the dangers and 
difficulties caused by secrecy in the industry and unwillingness to co-operate. 
These difficulties can be mitigated by large sample sizes, personal interviews 
or even personal follow-ups of non-responses, but these interventions are 
time consuming. All research suffers from the trade-off between speed, cost 
and quality (Figure 16), and the methodology that could provide the most 
effective trade-off should be identified. The position (identified as “poor” in the 
diagram), where research is expensive, slow and of low quality, is probably an 
exception (especially if the researcher is effective and efficient). At the same 
time achieving the cost-effective, fast and high-quality position could be 
difficult. It is important, however, for the researcher to understand this 
position, especially in areas such as macro-logistics measurement, where 
funding is difficult to obtain. 
 
 117
Cost
ExpensiveCheap
Quality / 
accuracy
High Q
Low Q
Speed
Slow
Fast
Cheap, 
slow, low Q
Expensive, 
slow, low Q
Expensive, 
fast, low Q
Expensive, 
fast, high Q
Cheap, fast, 
high Q
Cheap, slow, 
high Q
Below averagePoor
Above average Winning zone
9
9
 
Figure 16: Research trade-offs
250
 
 
5.2.2 Repeatability 
Measurement should be repeated annually and with the same methodology. 
In the absence of repeat data, the data itself falls into disuse. The Verburgh, 
Smith and Hamilton studies provided a continuum to some extent, but over 
too many years. It is also virtually impossible to maintain methodology in the 
exact same state over so many years (1957–1985), which made comparisons 
between the surveys difficult. In this regard the Pretorius surveys between 
1985 and 1993 did achieve repeatability to some extent and were used widely 
at that time, but also failed eventually because of other problems. 
 
5.2.3 Scope 
The scope of the surveys changed between surveys and addressed macro-
economic issues in different fashions.  All the surveys included questions on 
many variables which probably contributed to the fact that many were not 
completed. It would be advisable, under these circumstances, to concentrate 
on what is important and endeavour to resolve these important issues 
                                                 
250
 Havenga and Hobbs, 2004, p. 55. 
 118
effectively, rather than to include questions on areas that simply can’t be 
answered or are not so important.  
 
Outsourcing, for instance, is a difficult phenomenon to measure and 
understand. The outsourcing of an activity does not remove the specific 
activity, but rather just the management of, or responsibility for, the activity. If 
Sasol, as happened recently, outsources X percentage of their workforce, 
does it mean that X percent less people are needed to deliver the output that 
the business usually produces? If total output were Y, does it mean that 
productivity improved by X/Y? In the same sense the outsourcing of 
transportation services does not impact on the number of ton-kilometers 
required in the economy directly, but it could improve their performance, 
which is an entirely different issue. 
 
5.2.4 Context 
The Verburgh survey was the only survey that provided a strong context and 
raison d’etre for its existence. The state of transport planning in South Africa, 
the relative weakness of the Department of Transport (compared, for 
instance, to the vigour with which fiscal policy and its effect are measured and 
delivered by the treasury, and how monetary policy and its effect are 
measured and delivered by the Reserve Bank) and the absence of research 
make it impossible for researchers to adopt an “ivory tower” approach. 
Research is worthless if not used and should be undertaken only to either 
improve the discipline or lead to better decision making. Research as such is 
a processing function, which is informed by vision or “sense” and in turn 
should inform vision in order to make decisions. These decisions should in 
turn always lead to the resourcing and communication of the decision (Figure 
17). 
 119
Decide
Sense
Research
Resource
Communicate
 
 
Figure 17: The role of decision making in the strategic process
251
 
 
Without understanding the context and application of research, this loop is 
broken and the research will fall into disuse. 
 
In addition, by failing to contextualise research within other economic data, a 
research “island” is created. Users find it difficult to relate the use and 
interpretation of the data with other constructs in the economy and this causes 
disuse and also poor funding of research.  
 
5.2.5 Micro-economic detail 
All research requires accuracy and quality effort from the researchers, and 
modellers strive to develop models that depict and predict reality as closely as 
possible. It is a fact, however, that many macro-economic models are 
approximations and cannot achieve the same level of accuracy as is possible 
in the micro-economic field. It is important to make a distinction between 
absolute accuracy (which is unfortunately always impossible in the macro-
economy) and consistency of assumptions. Some research work (not 
necessarily the four studies discussed in more detail in this dissertation
252
) 
which attempts industry-level understanding of the state of logistics often 
                                                 
251
 Havenga and Hobbs, 2004, p. 58. 
252
 Verburgh, 1957; Smith, 1971; Hamilton, 1981 and 1985; Pretorius, 1985 to 1990. 
 120
suffer from the researchers’ unwillingness to engage in the macro-economic 
side of the research for the reasons mentioned above.  Industry-level studies 
are therefore often incomplete and lack a holistic view of the industry. 
 
Even in the micro-economy on business level the “industry architects” (firms 
that distinguish themselves as leaders) will know the size, drivers, constructs, 
volumes and forecasts for their industries, and will manage their businesses in 
the light of this knowledge. The logistics industry within a nation should, at 
least, together with its distinct capability to model macro-logistics provide the 
macro answers, even if these are not 100% correct.   
 
5.2.6 Systemic relationships 
Logistics forms a systemic relationship with other parts of a firm and value 
chain in the business and industry context. It is a process-orientated discipline 
that targets least total cost through trade-offs between all the functions 
involved in solving time and place disparity, contingent on specific customer 
service levels for maximum competitive advantage.  
 
In order to do this, intra-systemic relationships between functions such as 
transportation, distribution and materials management exist, but also inter-
systemic relationships between functions such as logistics, marketing, 
manufacturing and finance. These relationships are always considered in 
micro research and should always be considered in macro research, albeit on 
a higher level. This means that macro research should inform questions such 
as the link between a nation’s logistics facilities (e.g. ports, roads, railways 
and transhipment hubs), production facilities, spatial organisation and 
manufacturing practices, as well as the cost trade-offs in decision making 
around these constructs. 
 
Previous research seemed to have concentrated on only one isolated aspect 
of macro-logistics (Verburgh’s study is a possible exception) and failed to 
deliver on these systemic relationships. 
 121
5.2.7 Global links and benchmarking 
Some macro-economic intelligence on logistics systems is emerging globally. 
This creates opportunities for benchmarking as well as incorporating the 
theoretical part of South Africa’s macro-logistics systems into a more global 
relationship. The size of the global logistics marketplace, its future and its 
structure are becoming more and more known as some researchers reveal 
total tonnage, ton-kilometers, GDP and global growth.  It therefore becomes 
possible to measure a nation’s performance and productivity in this regard.  
 
5.2.8 Memories of the future 
Research must lead to a view of the future. Forecasting is a core and critical 
component of economics and it is expected of macro-economists to have a 
view of the future in order to make it possible for monetary authorities, 
industry and large business planners to prepare for future events. In-depth 
forecasting should also be scenario-based (Figure 18) and describe the 
underlying assumptions of each scenario. 
Scenario 1
Scenario 2
Scenario 3
Scenario 4
Tomorrow
Tomorrow
Tomorrow
Tomorrow
Tomorrow
Tomorrow
Tomorrow
TodayYesterday
Intermodality?
Re-regulation?
Technology?
Current Logistics Trends
 
 
Figure 18: Scenarios of the future
253
 
                                                 
253
 Havenga and Hobbs, 2004, p. 86. 
 122
 123
5.2.9 Scalability 
Qualitative discussions are important, but in the final instance measurement 
by numbers is required.  Scalability is the ability to access large datasets, in 
terms of either the number or the dimension of individual data elements, and 
the ability to grow the data in various directions as the field of research grows, 
without losing the core relationships within the data. This means, for instance, 
that if the ton-kilometers on a specific corridor is understood, it should also 
relate to ton-kilometers on other corridors, to the metropolitan areas where it 
originates and ends, to the total ton-kilometers in a region and even to the 
ton-kilometers in the world.  It is an expansive rather than reductionist 
approach. 
 
5.3 Overarching methodology 
The overarching methodological approach of this dissertation is therefore to 
describe the constructs of logistics, and especially transport, in South Africa 
and scale it in order to understand current and future demand. This requires a 
view of what is produced and consumed (by weight), how it is moved (modal 
market share), where it is moved (typologies) and what is moved 
(commodities). This raises the question that, if economists can value the 
economy, why can macro-logisticians not weigh it.  If economists can report 
on national, regional and district economic performance, why can logisticians 
not report on spatial performance such as corridors and hubs? Ultimately 
these aspects should be measured in terms of relative performance in, and 
contribution to, the economy as well as forecasted into the future to enable 
planning and investment. 
 
Economists often use models to develop these answers and the approach in 
this dissertation will also be modelling, enhanced with desktop research in 
certain cases. The models must be repeatable annually, based on the same 
assumptions, relate to the other aspects of the macro-economy, create 
forecasts and enable global comparisons.  With these objectives in mind the 
actual model development can be discussed. 
Chapter 6 - National freight flow model 
6.1 Introduction and background 
The first objective of the overarching methodology is to achieve a valid and 
repeatable picture of modal market share in South Africa. The Verburgh, 
Smith, Hamilton and Pretorius research projects all attempted to arrive at 
some approximation of modal market share and size, but achieved this 
objective with varying degrees of success. The work was, however, not 
repeatable on an annual basis.  It is hypothesised that this is because of the 
huge effort required to do the research but that the results never produced 
what the researchers tried to achieve. The fact that the work was never 
repeated therefore also meant that it couldn’t be used for planning purposes 
nor for measuring economic performance on a consistent basis. Their 
attempts to measure traffic flows with some approximation of origin and 
destination points all failed. 
 
The objective of the national freight flow model is therefore to fill this gap and 
provide the first repeatable observations on modal market share and freight 
flows in South Africa. It is believed that this work could easily be repeated 
annually and therefore used for infrastructure policy decisions, planning, and 
as a lead and lag indicator of the performance of the network. 
 
6.2 The structure of the national freight flow model 
The national freight flow model utilises the South African National Roads 
Agency’s (SANRAL) Traffic Count Yearbooks compiled by Mikros Traffic 
Monitoring, as well as actual freight flow data obtained from Spoornet, as 
basis to develop a current and historical view of freight traffic flows in South 
Africa.   
 124
6.2.1 Rail freight flows 
The rail freight data obtained by previous researchers were one-dimensional, 
i.e. they reported only volumes shipped and not rail traffic flows between 
origins and destinations. As traffic flows can be developed from Spoornet’s 
data, this will be the first step in the process. Once road freight flows have 
been determined, a matching between the two data sets will have to be 
achieved. 
 
6.2.2 Road freight flows 
Comprehensive traffic observations started in South Africa in 1984, when a 
pilot study was conducted on the 600 km N3 route between Johannesburg 
and Durban. As a result of the success of this study the National Transport 
Commission (now SANRAL) decided in June 1985 to expand the Traffic 
Counting Network to traffic counting stations.
254
 
 
SANRAL now repeats this work annually and a reasonable degree of stability 
in the work process has been achieved since the 1990s. The main objective 
of SANRAL’s efforts is not to develop freight flows, but to understand 
congestion points and enable planning in terms of all vehicular flows in South 
Africa (of which heavy vehicles are only a subset). No attempt has ever been 
made to use the information for freight flow purposes and a reasonable 
amount of modelling of the available data are therefore necessary. 
 
The SANRAL Traffic Counting Yearbook is a compendium of traffic 
information obtained at traffic counting stations on primary roads, 255 
highlighting the latest available traffic characteristics. A station is an 
installation on a road which enables the collection of data; both permanent 
and secondary stations are utilised.  A permanent station is one where 
continuous traffic observations are made.  A secondary station is one where 
traffic observations are made on a sampling basis for at least 168 consecutive 
                                                 
254
 Mikros, 2006, p. iv. 
255
 Primary roads refer to all declared national roads and to other roads of strategic or 
economic importance (Mikros, 2006, p. i). 
 125
hours per annum. The Yearbook contains information on 398 permanent 
stations and 430 secondary stations. The stations are placed on selected links 
of the national and primary road network and yield information such as 
Average Daily Traffic and Average Daily Truck Traffic.
256
  
 
The sampling base for the model therefore compares well with previous work, 
where questions were asked from a small sample of actual hauliers, but often 
for just a short period of time. The traffic counts for permanent stations are 
continuous, while for a secondary station the counts are made on a sample 
basis for at least 168 hours, which compares well with the previous sample 
studies, where data for only one week were collected. The counting station 
data are also automated and therefore far less prone to human error as would 
be the case in a sampling framework where a human response is required. 
The data are also available annually, which means that a validated model 
would be easily repeatable and would be more permanent and stable as a 
base, provided that a link between the traffic counts and freight volumes could 
be found. 
 
6.3 Calculation of road tonnages 
The number of counting stations per year analysed in this dissertation is 
shown in Table 27. 
 
Year Number of stations 
1990 346 
1993 368 
1997 239 
2003 626 
2004 583 
2005 521 
 
Table 27: Number of CTO stations per year analysed 
                                                 
256
 Mikros, 2006, pp. i-ii. 
 126
The data from all these counting stations were obtained from SANRAL and 
captured in an Excel database. SANRAL unfortunately does not provide 
computerised data sets in a usable Excel format.  It is a secondary objective 
of this work, if successful, to negotiate the electronic provision of data from 
SANRAL to facilitate the repetition of this work through limiting opportunities 
for human error and also fast-tracking data capturing. 
 
The modelling approach followed to calculate road tonnage from traffic 
observations is described below. 
� SANRAL allocates counting stations to specific routes across the country 
e.g. N1, R30, etc. in a geographical order, e.g. on the N1 stations start in 
Cape Town and end in Beitbridge. 
� The average daily truck traffic (ADTT) (i.e. number of trucks) and the 
percentage split of these trucks between short, medium and long trucks 
(SMLT) are captured per counting station per route from SANRAL data. 
The SANRAL counting methodology enabled the counting monitors at the 
stations to distinguish between trucks and passenger vehicles by the 
length of the vehicle. The trucks can then also be categorised further in 
terms of axles. This provides a good approximation of vehicle type, 
especially compared to the uncertainty in Hamilton’s work in this regard on 
what should be included as a true freight vehicle and what not. Hamilton 
decided in his second survey to exclude all trucks of five tons and smaller, 
as these were, according to his qualitative research, mostly in the hands of 
final consumers and mostly not used for freight and ancillary freight 
transport. SANRAL’s split excludes only the very small vehicles in this 
base, but it is assumed that only the very small vehicles actually conform 
to Hamilton’s assumption. Vehicles from four tons and higher could 
therefore be included in the research. 
o ADTT is the total number of trucks observed in each direction during 
the actual period monitored divided by the total number of hours 
monitored multiplied by twenty-four. 
o SMLT truck split percentage is the percentage of trucks in each 
direction which fall into each of the following categories: 
 127
• A short truck is typically a rigid-chassis two-axle vehicle designed 
for transport of goods, or a bus with at least one axle with four 
wheels; 
• A medium truck is typically a truck-tractor, plus semi-trailer 
combination; 
• A long truck is typically a combination of a truck-tractor plus a semi-
trailer and a full trailer. 
o The indicated split is established from the combination of 
measurements of vehicle length and chassis height as follows: 
• A vehicle shorter than 4.6m is always regarded as a light vehicle, 
not a truck; 
• A vehicle between 4.6 and 11.0 m long is classified as a short truck 
if the signal indicating the chassis height is “high”. (If the signal 
indicated a “medium” or “low” chassis height the vehicle is 
considered to be a long light vehicle e.g. a car towing a caravan.); 
• A vehicle between 11.0 and 16.8 m long is classified as a medium 
truck, irrespective of the chassis height; 
• A vehicle longer than 16.8 m is classified as a long truck, 
irrespective of the chassis height.
257
 
� The number of trucks had to be translated into the actual weight of the 
freight.  For this purpose two figures were calculated – the total weight 
(truck + freight) and the truck tare (only truck weight) – the difference 
between the two will then be the actual weight of the freight. 
• The average total ton per SMLT was calculated from (Road Freight 
Association) RFA data. (Table 28).  The total weight per SMLT was 
calculated by multiplying the truck mass with the number of trucks.  
 
 
 
 
 
 
 
                                                 
257
 Mikros, 2005, p. 10. 
 128
Truck type Average ton 
Short  18.8 
Medium 42.4 
Long 51.6 
 
Table 28: Average gross ton per truck type as calculated from RFA data 
 
� The tare for SMLT was then calculated based on the average tare per 
vehicle type as published by the Road Freight Association (Table 29). 
 
Truck type Tare 
Short  5.1 
Medium 14.1 
Long 18.2 
 
Table 29: Tare based on vehicle concepts published by the Road Freight Association 
 
� Useful analysis can, however, only be conducted if the data per counting 
point are translated into corridor, metropolitan and rural totals. 
� For the national routes, counting stations were depicted graphically to 
determine the split between metropolitan peaks, rural traffic and long 
distance (corridor) traffic.   
� The assumption is that corridor traffic is indicated when a “flattening” of 
traffic counts occurs, while other stations are either metropolitan or rural, 
depending on their count size and location: 
• Corridor traffic:  The stations where flattening occurred were allocated to 
national routes.  The average of the annual weight for all the counting 
stations per corridor was calculated to reflect the tonnage per corridor; 
• Metropolitan traffic:  The key metropolitan areas were identified (through 
sharp peaks in traffic counts).  Different routes lead into these 
metropolitan areas.  For each route the annual average was calculated.  
The annual totals per route were then added together to obtain total 
metropolitan traffic; 
 129
• The remaining stations were allocated to rural traffic.  Similar to 
metropolitan traffic (where a group of different routes combine to form a 
metropolitan infrastructure typology), each province has a number of 
rural routes that also combine to form a rural infrastructure typology.  
For each route, the annual average was calculated.  The annual totals 
per route were added together to obtain total rural traffic; 
• Discussions with SANRAL indicated problems with some counting 
stations.  Extreme outliers were therefore excluded from the analysis 
(i.e. counting stations significantly higher or lower than other stations on 
the same route, and often significantly different from a series of adjacent 
stations, also clearly visible in the graphs). 
 
Figure 19 to Figure 22 below show the graphical depiction of Cape Town-
Johannesburg and Durban-Johannesburg for 2004 and 2005. (All routes are 
depicted graphically in this fashion to determine corridor flows, but only the 
two main corridors are illustrated in this research.) 
Metropolitan areasCorridors
Not taken into accountRural areas
Rural 
Free 
State
0
1000
2000
3000
4000
5000
6000
Ol
d
 Oa
k
S
t
el
l
enb
er
g I
/
C
B
r
ack
e
nf
e
l
l
C
ap
e R
en
dezvous
E
i
kend
al
V
a
n R
i
e
bee
c
k I
C
K
r
aa
i
f
on
t
ei
n
O
os
t
enb
er
g
Wi
ne
l
a
n
d
s W
I
M
Joo
s
t
e
nb
er
gvl
a
k
t
e
 I
C
K
l
a
p
m
u
ts
 IC
Pa
a
r
l
 I
C
 
W
B & EB
H
ug
uen
ot
 P
l
aza 
1
H
ug
uen
ot
 P
l
aza 
2
H
u
gu
en
o
t
 N
o
r
t
h P
i
ez
o
Wo
rce
s
t
e
r
De
 D
o
o
r
n
s
T
o
u
w
s
r
iv
ie
r
 No
r
t
h
Le
t
j
i
e
s
b
os
Th
r
e
e
 S
i
s
t
er
 W
I
M
H
ano
v
e
r
 R
oa
d
Ga
r
i
e
p
S
p
r
i
ng
f
on
t
ei
n
E
d
enb
ur
g S
out
h
U
ncl
e
 T
om
'
s P
l
ac
e
F
e
rre
i
ra
J
a
ge
r
s
f
o
nt
e
i
n
Fi
c
h
ar
d
t
Ha
ld
o
n
 Ro
a
d
L
an
gen
hoven
 P
ar
k
N
e
l
son M
a
nde
l
a B
oul
ev
a
r
d
E
nge
n O
n
e S
t
o
p
P
i
ts
to
p
B
l
oe
m
f
o
nt
e
i
n
 by
pa
s
s
V
er
keer
d
evl
ei
 P
l
aza 
1
V
er
keer
d
evl
ei
 P
l
aza 
2
K
r
oo
ns
t
ad bypass
K
r
oo
nst
ad
 N
B
 W
I
M
V
a
al
 pl
a
z
a
 
1
V
a
al
 pl
a
z
a
 
2
O
r
a
n
j
e
v
ille
V
aal
r
i
vi
er
Va
n
d
e
r
b
i
j
l
 SB+
N
B
S
e
boken
g N
ew
G
r
ass
m
er
e
 m
ai
n
 p1
G
r
ass
m
er
e
 m
ai
n
 p2
G
r
ass
m
er
e r
am
ps 
p1
G
r
ass
m
er
e r
am
ps 
p2
M
i
s
g
und
 I
C
Die
p
k
l
o
o
f
 I
C
Intra Cape Town
Rural 
Western 
Cape
Cape Town 
– Gauteng 
via Bloem
Intra 
Bloem
Cape Town 
– Gauteng 
via Bloem
Intra 
Gauteng
ignore -North 
bound only
 
Figure 19: Average daily truck traffic N1 Cape Town-Johannesburg (bi-directional) – 2004 
 
 
 
 130
0
500
1,000
1,500
2,000
2,500
3,000
3,500
4,000
4,500
5,000
Ol
d
 O
a
k
S
t
el
l
en
ber
g I
/
C
B
r
ackenf
el
l
W
i
nel
a
nds W
I
M
PAW
C
Fr
ans
chhoek
W
e
l
l
i
ngt
on
H
uguenot P
l
aza 1
H
uguenot P
l
aza 2
RAW
C
RAW
C
H
ugue
not
 N
or
t
h
Touw
sr
i
v
i
er
 N
or
t
h
P
r
i
nce A
l
ber
t
 R
d
Thr
ee S
i
s
t
er
s W
I
M
H
anove
r
 R
oad
Ga
r
i
e
p
F
i
char
dt
Gl
e
n
 A
/
C
 I
/
C
B
l
oem
f
ont
ei
n
V
e
rk
e
e
rd
e
v
l
e
i
V
e
rk
e
e
rd
e
v
l
e
i
W
i
nburg N
or
t
h
K
r
oonst
a
d S
out
h
K
r
oonkom
K
r
oonr
us
K
r
oonst
a
d
 B
y
pass
K
r
oonl
ey
K
r
oon U
l
t
r
aci
t
y
Ko
pp
i
e
s
 I
/
C
Va
al
 Pl
a
z
a
 1
Va
al
 Pl
a
z
a
 2
Gr
a
s
me
r
e
 Ma
i
n
 P
Gr
a
s
me
r
e
 Ma
i
n
 P
Gr
a
s
me
r
e
 R
a
mp
s
Gr
a
s
me
r
e
 R
a
mp
s
M
i
sgund
G
ol
den H
i
ghw
ay
D
i
epk
l
oof
 I
/
C
Intra Cape Town
Rural Western 
Cape
Cape Town 
– Gauteng 
via Bloem
Intra Bloem
Cape Town 
–Gauteng 
via Bloem
Intra Gauteng
Rural areas
Corridors
Not taken into account
Metropolitan areas
Rural Free 
State
 
Figure 20: Average daily truck traffic N1 Cape Town-Johannesburg (bi-directional) – 2005 
 
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
11000
12000
13000
C
l
o
e
te
 Ea
s
t +
 W
e
s
t
St J
a
m
e
s
 
IC
P
a
r
a
di
s
e
 V
al
l
ey
W
e
s
t
me
a
d
  
W
I
M E
B
+
W
B
W
e
s
t
m
e
ad I
C
M
a
ho
ga
ny
M
ar
i
an
hi
l
l
 P
l
a
z
a
 1
M
ar
i
an
hi
l
l
 P
l
a
z
a
 2
S
ho
ngw
en
i
 I
C
Ke
y
 R
i
d
g
e
C
at
o R
i
dg
e W
B
C
am
p
er
do
w
n I
C
Da
r
d
en
el
l
e
s
Li
on
 P
ar
k
 
I
C
A
s
hbu
r
t
o
n
A
s
hb
ur
t
on I
C
A
s
hb
ur
t
on
 W
I
M
M
a
rk
e
t
 S
t
re
e
t
Ne
w
 E
n
g
l
an
d WI
M
N
ew
 E
n
gl
an
d I
C
G
r
e
y
to
w
n
 R
o
a
d
 IC
S
anc
t
uar
y
 R
oad
 
S
B
+
N
B
Cha
t
t
e
rt
on
Du
nc
an
 M
c
K
e
n
z
i
e
Hi
l
t
o
n
 NB
+
S
B
Ced
a
r
a
 NB
+
S
B
M
er
r
i
v
al
e I
C
Mo
o
i
 Ma
i
n
 1
Mo
o
i
 Ma
i
n
 2
Mo
o
i
 r
a
mp
 
1
 S
B
+
N
B
Mo
o
i
 r
a
mp
 
2
 S
B
+
N
B
T
r
ev
e
r
t
o
n 1
T
r
ev
e
r
t
o
n 2
M
i
dw
ay
 W
I
M
T
u
ge
l
a
 P
l
az
a
 
1
T
u
ge
l
a
 P
l
az
a
 
2
V
a
n Re
en
en
W
i
l
g
e P
l
az
a  1
W
i
l
g
e P
l
az
a  2
D
e
 
H
oek
 P
l
az
a 1
D
e
 
H
oek
 P
l
az
a 2
H
e
i
d
el
be
r
g
 S
c
r
e
e
n
 
NB
+
S
B
Ba
l
f
o
u
r
Hei
d
e
l
be
rg
Ja
co
b
 S
t
r
e
e
t
D
a
s
v
ill
e
B
e
no
ni
 I
C
Gl
e
n
r
o
y
S
ui
k
er
bo
s
r
an
d
B
a
rr
y
 M
a
ra
i
s
 I
C
Leo
nd
al
e I
C
Le
on
d
al
e
H
e
i
d
el
be
r
g R
oa
d I
C
Al
b
e
r
t
o
n
G
r
ey
 A
v
en
ue
 I
C
R
a
n
d
 Ai
r
p
o
r
t 
IC
 n
o
r
t
h
e
r
n
 s
i
d
e
Ra
nd
 A
i
r
p
ort
 I
C
 
s
o
ut
he
r
n
 s
i
de
E
l
an
ds
 I
C
 
S
+
N
Ros
h
er
v
i
l
l
e N
B
+
S
B
G
e
l
d
en
hu
y
s
 I
C
 S
W
+
N
E
C
r
es
t
o
n Hi
l
l
V
a
n 
B
u
uren
 I
C
 NB
+
S
B
G
i
llo
o
l
y
'
s
 S
W
+
N
E
G
i
llo
o
l
y
'
s
 S
E
+
N
W
Li
n
k
s
f
i
el
d
 S
o
ut
h
Li
nk
s
f
i
el
d I
C
E
den
v
a
l
e
M
o
dd
erf
o
nt
e
i
n
 I
C
L
o
m
b
ardy
Da
 C
o
s
t
a
S
a
pp
er
s
 H
a
v
e
n
A
l
ex
an
dr
a
M
a
rl
b
o
ro
 I
C
F
r
an
k
e
nw
a
l
d
Bu
c
c
l
e
u
c
h
Metropolitan areasCorridors
Not taken into accountRural areas
Intra 
Durban
Rural 
KZN
Gauteng
-Durban
Intra 
Gauteng
 
Figure 21: Average daily truck traffic N3 Durban-Johannesburg (bi-directional) – 2004 
 
 131
Rural areas
Corridors
Not taken into account
Metropolitan areas
0
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
9,000
10,000
11,000
12,000
13,000
EB C
l
o
e
te
 
E+
W
S
t
 Ja
m
e
s I
C
Pa
r
a
d
i
s
e
 Va
l
l
e
y
Que
e
ns
b
ur
gh
 N
+
S
Wes
t
m
e
ad I
/
C
Wes
t
m
e
ad WI
M
M
ari
anh
i
l
l
 P
l
a
z
a 
1
M
ari
anh
i
l
l
 P
l
a
z
a 
2
C
a
m
perdo
w
n I
/
C
D
a
r
d
e
n
e
lle
s
Li
on
 P
ark
 
I
C
A
s
h
burt
on
A
s
h
burt
on I
/
C
Ma
r
k
e
t
 s
t
r
e
e
t
N
e
w
 E
ngl
a
nd WI
M
N
e
w
 E
ngl
a
nd I
/
C
G
r
e
y
to
w
n
 R
o
a
d
 I
C
Sa
n
c
t
u
a
r
y
 R
d
 S
B+
N
B
Ch
a
t
t
e
r
t
o
n
Hi
l
t
o
n
 NB
+
S
B
 
I
/
C
Ceda
r
a
 N
B
+
S
B
 
I
/
C
D
u
nc
a
n M
c
K
e
nz
i
e
M
e
r
r
i
v
al
e I
C
Mo
o
i
 R
a
mp
 
N
B
+
S
B
 1
Mo
o
i
 R
a
mp
 
N
B
+
S
B
 2
Mo
o
i
 Ma
in
 1
Mo
o
i
 Ma
in
 2
T
rev
ert
o
n 1
T
rev
ert
o
n 2
T
uge
l
a
 P
l
a
z
a 
1
T
uge
l
a
 P
l
a
z
a 
2
V
an
 Reen
en
W
i
l
ge
 P
l
az
a 
1
W
i
l
ge
 P
l
az
a 
2
De H
oek
 
P
l
a
z
a 1
De H
oek
 
P
l
a
z
a 2
H
e
i
d
el
berg
Ba
l
f
o
u
r
H
e
i
d
el
berg
Ja
c
o
 st
r
e
e
t
D
a
s
v
i
lle
B
e
no
ni
 I
/
C
Gl
en
r
o
y
S
u
i
k
erbos
r
a
nd
B
a
r
r
y m
a
r
i
a
s
 I/
C
Le
ond
al
e I
/
C
L
eon
dal
e
H
e
i
d
el
be
r
g
 R
d
 I
/
C
A
l
bert
on
G
rey
 A
v
e 
I
/
C
R
a
nd A
i
r
p
ort
 I
/
C 
S
Ra
n
d
 A
i
rp
o
r
t
 I
/
C n
El
a
n
d
s
 I
C
 
S+
N
R
o
sh
e
r
vi
l
l
e
 N
B
+
S
B
G
e
l
den
huy
s
 I
/
C
 S
W
+
N
E
Cre
s
t
o
n
 Hi
l
l
V
an B
u
uren I
C
 
NB
+
S
B
G
ill
o
o
ly
'
s
 S
W
+
N
E
Gi
l
l
oo
l
y
'
s
 N
+
E
F
ran
k
en
w
al
d
B
u
ccl
e
u
ch
Intra 
Durban
Rural 
KZN
Gauteng
-Durban
Intra 
Gauteng
 
Figure 22: Average daily truck traffic N3 Durban-Johannesburg (bi-directional) – 2005 
 
6.4 Calculation of rail tonnages 
Actual rail data are available per commodity and per origin-destination station.  
Rail data therefore do not have to be modelled.  The Spoornet origin-
destination station pairs, however, had to be allocated manually to the same 
corridor, metropolitan and rural classification as the SANRAL counting 
stations to enable comparison. 
   
6.5 List of corridors, metropolitan and rural areas 
The above analysis resulted in the definition of the following corridors, rural 
and metropolitan areas for analysis. 
 
� Corridors (bi-directional): 
The corridors were grouped with a specific typology in mind that would 
enable reporting on more than one of the overarching groupings, i.e. 
Western, Eastern, Northern, etc. The two main corridors (i.e. Gauteng-
 132
Durban and Gauteng-Cape Town) deemed to carry the heaviest traffic, 
however, were kept separate.  The following typology was therefore used 
(Table 30): 
 
ROUTE SUMMARY ROUTE DETAIL 
Durban-Gauteng  
Cape Town-Gauteng   
Gauteng-Richards Bay   
Cape Town-Port Elizabeth 
East London-Port Elizabeth 
Durban-East London 
Cape Town-Durban 
Coastal 
Durban-Richards Bay 
Gauteng-Witbank 
Gauteng-Nelspruit Eastern Corridors 
Gauteng-Swaziland 
Gauteng-Upington 
Cape Town-Upington Western Corridors 
Gauteng-Lobatse 
from/to Port Elizabeth 
from/to East London Eastern/Southern Cape-Gauteng 
from/to Mossel Bay 
Beitbridge-Gauteng 
Northern Corridors 
Gauteng-Polokwane 
Cape Town-Namibia   
 
Table 30: Corridors 
 
� For metropolitan and rural areas the following typology was used (Table 
31): 
 
 
 
 
 
 133
Intra Gauteng 
Intra Cape Town 
Intra Durban 
Intra Bloemfontein 
Intra Port Elizabeth 
Intra Witbank/Middelburg 
Metropolitan 
Intra East London 
Rural Eastern Cape 
Rural Free State 
Rural Kwazulu Natal 
Rural Limpopo 
Rural Mpumalanga 
Rural North West 
Rural Northern Cape 
Rural 
Rural Western Cape 
 
Table 31: Rural and metropolitan areas 
 
6.6 The national freight flow model 
The results from the national freight flow model were tabulated to provide 
modal market share per corridor, rural and metropolitan area in South Africa. 
The developed model was applied to counting data from counting stations and 
actual Spoornet data for 1993, 1997, 2003 and 2004. This application would 
therefore indicate whether (i) a comparable link with the previous sporadic 
surveys that ended in 1990 could be established; (ii) whether the measures 
from the late 1990s up to 2004 have a reasonable correlation with GDP; and 
(iii) whether the 2003 and 2004 data (when the model was applied annually) 
seem stable and useful enough to serve as the basis for an annual model.  
The results from the modelling exercise are reported in the next chapter. 
 
 
 
 134
Chapter 7 - National freight flow model results and 
strategic interpretation 
7.1 Introduction 
The national freight flow model (NFFM) was applied to data from Spoornet 
and SANRAL for 1993, 1997, 2003 and 2004.
258
 The results produced the 
most complete picture of surface freight yet produced in South Africa, 
including modal market share, and was the first successful attempt to model 
and report on flow data in the country. The data were tested for internal 
consistency, compared with other economic data and also with the Verburgh, 
Smith, Hamilton and Pretorius studies for relative performance comparisons.  
The results are also applied to initiate the process of generating economic 
performance data for transport in the economy. 
 
7.2 Total supply of, and demand for, surface freight transport 
The most definitive measures of surface freight transport supply are ton and 
ton-kilometer. Two of the initial objectives that the candidate attempted to 
determine were the tons shipped in the economy and the ton-kilometers of 
transport that were delivered. The data collection methodology of the data 
used in this model, however, means that it was not possible to determine 
where freight counted specifically originated per load and what the destination 
was going to be. Certain trucks counted could therefore have been traversing 
more than one of the three typologies, i.e. corridor, rural or metropolitan, and 
even more than one sub-typology within the main typologies without the 
specific distinctions being noticeable.
259
 In terms of all three of the typologies 
that the overarching measurement is trying to achieve, this is not a problem. It 
merely means that freight, in situations like these, impacts on more than one 
                                                 
258
 The candidate was assisted in the application by his collaborator, Ilse Hobbs. 
259
 This means, for example, that if a specific load originated in Caledon in the Western Cape 
and was sent to Johannesburg, it will be observed by the model as a rural Western Cape 
load and a load that used the Cape Town to Gauteng corridor. In another example, if a 
load originated in Durban and was sent to Beitbridge, it will be observed as using both 
the Durban to Gauteng and Gauteng to Beitbridge corridor. 
 135
classification, which is actually technically correct. The specifics of this 
observation should, however, always be remembered and, when the model 
application is repeated, the same set of assumptions should be applied. The 
correct measure description therefore is tons shipped and ton-kilometers 
performed over every classification of the freight network of South Africa, i.e. 
corridors, rural areas and metropolitan areas. 
 
In terms of this description the tons and ton-kilometers obtained from the 
modes (supplied) in South Africa for the years of measurement are depicted 
in Table 32. 
 
 
Road 
total 
Rail (excluding 
export 
machines)
260
 
Rail 
(export 
machines)
Total 
1993 692 107 72 870 
1997 848 120 85 1 053 
2003 1 026 105 94 1 225 
2004 1 105 106 96 1 307 
 
Table 32: Surface freight supply in South Africa between 1993 and 2004 (million tons) 
 
The above data produce the following market share figures (Table 33). 
 
 
Road 
total 
Rail total 
(excluding export 
machines) 
Rail (export 
machines) 
1993 79% 12% 8% 
1997 81% 11% 8% 
2003 84% 9% 8% 
2004 85% 8% 7% 
 
                                                 
260
 “Export machines” refers to the railway’s dedicated coal and iron ore export lines 
 136
Table 33: Surface freight market share between 1993 and 2004 
The surface freight transport supply from this model can be compared to 
previous surveys, as depicted in Table 34. 
 
Million ton Rail Road Total 
1957 – Verburgh 75 180 255 
1971 – Smith 100 524 624 
1985 – Hamilton 170 504 674 
1985 – Pretorius 182 686 868 
1986 – Pretorius 178 720 898 
1987 – Pretorius 188 725 913 
1988 – Pretorius 187 663 850 
1989 – Pretorius 179 588 767 
1990 – Pretorius 184 552 736 
1993 – NFFM 178 692 870 
1997 – NFFM 204 848 1053 
2003 – NFFM 199 1026 1225 
2004 – NFFM 202 1105 1307 
 
Table 34: Survey comparison: All surveys
261
 
 
A market share comparison of all surveys is depicted in Table 35. 
 
 
 
 
 
 
 
 
 
 
                                                 
261
 Hamilton’s flawed survey of 1981 excluded and Pretorius’s data normalised with actual rail 
data. 
 137
 
 Rail Road 
1957 – Verburgh 29% 71% 
1971 – Smith 16% 84% 
1985 – Hamilton 25% 75% 
1985 – Pretorius 21% 79% 
1986 – Pretorius 20% 80% 
1987 – Pretorius 21% 79% 
1988 – Pretorius 22% 78% 
1989 – Pretorius 23% 77% 
1990 – Pretorius 25% 75% 
1993 – NFFM 20% 80% 
1997 – NFFM 19% 81% 
2003 – NFFM 16% 84% 
2004 – NFFM 15% 85% 
 
Table 35: Market share comparison: All surveys (based on tons) 
 
Irrespective of the flaws of previous surveys, the definite downward trend in 
rail market share is clear. Unfortunately, because of the flaws of previous 
surveys and the absence of flow data, the specifics of this decline were never 
really understood. If the effect of the rail transport of ring-fenced commodities 
(i.e. bulk commodities that came on stream in the 1970s) is added, the  
decline of rail is much more pronounced (Table 33). 
 
Growth rates for constant GDP at 2000 prices, physical production in the 
economy and tons transported can now be compared between the national 
freight flow model and the previous studies analysed. This is depicted in Table 
36.  
 138
 
 
Constant 
GDP at 2000 
prices 
Physical 
volume of 
production in 
the economy 
Million tons 
transported 
Million tons 
transported 
(known 
errors 
removed) 
CAGR - 
Verburgh 1957 
to Hamilton 1985 
3.85% 5.54% 3.53% 3.53% 
CAGR - 
Pretorius 1985 to 
1990  
1.67% 0.61% -3.59% -3.27% 
CAGR - 1993 to 
2004 NFFM 
application 
3.15% 1.71% 3.77% 3.77% 
 
Table 36: GDP growth rate comparison
262
 
 
The survey methods followed by Verburgh, Hamilton, Smith and Pretorius are 
essentially the same and could therefore serve as a good base for 
comparison.  
 
The Verburgh to Hamilton survey results perform satisfactorily in this 
comparison as the growth in tons transported correlates well with the growth 
in GDP. The faster growth in the physical volume of production in the 
economy (compared to GDP and transport growth), however, could not be 
explained by the commissioning of South Africa’s two export machines in the 
1970s, i.e. the Richards Bay coal line and Sishen to Saldanha iron ore line,
263
 
because in both these cases it should also have translated into higher 
transport volumes. It is also hypothesised that more double handling of goods 
in a more mature economy (caused by specialisation) should occur, which 
means that tons transported should probably grow somewhat faster than the 
physical volume of production in the economy.  
 
The Pretorius surveys did not do well in these comparisons. Constant GDP 
growth in South Africa (1.67%) was slow in the years just before the 
                                                 
262
 Hamilton’s flawed 1981 survey excluded. 
263
 It is indicative of the impact of these “export machines”that more than 10% of the 
transportable economy is shipped on these two lines today. 
 139
unbanning of the ANC and the release of Nelson Mandela, and the physical 
volume of production slowed down to a compound annual growth rate of a 
mere 0.61%, but a negative correlation of more than 3% with tons transported 
is highly unlikely.  
 
The national freight flow model correlation performs well in all comparisons. 
Constant GDP grows faster than the physical volume of production as the 
economy matures and the tertiary sector expands.  The expected increase in 
double handling results in a faster transport growth rate than that of physical 
production, but more in line with GDP growth. 
 
A more precise measure of correlation would be to measure the correlation 
coefficient between the various data sets. This calculation is depicted in Table 
37. 
 
 Studies 
 
Verburgh, 
Smith and 
Hamilton 
Pretorius NFFM 
GDP to physical production 0.992 0.935 0.986 
GDP to tons transported 0.627 -0.811 0.996 
Physical production to tons 
transported 
0.653 -0.720 0.985 
GDP to tons transported: Known 
errors removed 
0.933 -0.868 0.996 
Physical production to tons 
transported: Known errors removed 
0.874 -0.794 0.985 
 
Table 37: Correlation coefficient comparison for all surveys 
 
Only limited data points are currently available, but initial observations for 
2005 and 2006 indicate that the same high correlation for the national freight 
flow model results persists. The data have been workshopped amongst peers 
at the CSIR, Transnet, the Department of Trade and Industry, the Department 
 140
of Agriculture and Spoornet, and current indications are that the high 
correlation will continue in future applications of the model. As the work 
continues, the academic soundness of the research should be continuously 
tested and reported on in due course to determine if the positive correlation is 
valid for longer time series. 
 
The total supply of transport in 2004 for all typologies has been established to 
be 1,307 million tons and 336 billion ton-kilometers.  The next section 
describes supply within these typologies and the salient features of each one. 
 
7.3 Typologies and the current distribution of freight 
In South Africa 336 billion ton-kilometers of freight movement is provided to 
the economy, compared to the approximately 15,000 billion ton-kilometers 
supplied in the world.  South Africa has just less than 1% of the world’s 
population, produces less that 0.4% of the world’s GDP, but requires more 
than 2% of global transport in terms of ton-kilometers (Figure 23) – the 
country’s transport demand is therefore excessive in the light of these 
indicators.
264
 
 
Surface freight
GDP 
Population
0.0
0.5
1.0
1.5
2.0
2.5
3.0
 
 
 
 
 
 
 
 
Figure 23: South African indicators as a percentage of world indicators 
                                                 
264
 Calculated from Rodrigue, et al, 2005, Gielen, D. 2004, and State of Logistics Survey 
(CSIR, 2005) modelling. 
 141
 
This situation arises from the country’s economic development history that 
resulted in production and population development far from coastal areas, in a 
relatively open mineral export economy, and a beneficiated product and 
energy import economy (because of minerals found far from the coast), 
thereby creating long export and import corridor requirements. 
 
This situation places, ipso facto, a massive burden on transport infrastructure 
and (because of the poor productivity
265
) specifically a need for excessively 
cheap transport. The South African economy is still relatively primary, 
especially when compared to developed economies, and a better 
understanding of the specifics of the typologies is required in order to 
formulate an appropriate freight transport strategy. 
 
The 336 billion ton-kilometers of freight can be divided into four distinct 
typologies i.e. primary, metropolitan, rural and corridor (Figure 24). 
 
Primary, 73, 
22%
Metropolitan, 
47, 14%
Rural, 63, 
19%
Corridor, 146, 
45%
 
                                                 
265
 The global economy produces $2.76 per required ton-kilometer, compared to South 
Africa’s $0.65 for each ton-kilometer provided (calculated from global and domestic GDP 
data). 
 142
Figure 24: Freight typologies (ton-kilometers and percentage share) 
 
The primary typology describes the world-class one-directional bulk transport 
corridors to move South Africa’s low-value mining commodities efficiently to 
the coast for export, and to internal beneficiation facilities.  Approximately 
85% of this typology is supplied for iron ore transport from Sishen to Saldanha 
and coal from Mpumalanga to Richards Bay. Efficiency is crucial in order to 
compete with other deposits around the world that are mostly closer to the 
coast (in the country of origin – therefore requiring less land transport) and 
often closer to global demand points (therefore requiring less maritime 
transport).  Origin-destination pairs, which facilitate bulk heavy-haul, are 
usually few in number.
266
  This typology is typically rail-orientated and 
migration to the road mode is practically impossible. Schematically it can be 
described as a pipe with two fixed ends, as depicted in Figure 25. 
 
 
 
Figure 25: Schematic of primary traffic 
 
The metropolitan typology describes transport in metropolitan areas over 
shorter distances. It often relates to local deliveries from intermediate 
production points and distribution centres to final points of consumption as 
well as transport between intermediate demand and supply points. This freight 
traffic causes congestion because of the many origin-destination pairs 
distributed over small areas; it is expensive (both from a capital supply and 
usage point of view) and is typified by rapid growth in South Africa.  This 
typology is predominantly road-based, but rail-based solutions where traffic 
behaves like primary traffic are possible to alleviate congestion.
267
 
                                                 
266
 Fewer origin-destination pairs for bulk movements mean that unit trains with a ring-fenced 
wagon fleet, requiring little or no shunting, can be used, drastically reducing the cost of 
rail transport. 
267
 Examples include the bale-by-rail solution that was developed for Vissershok in Cape 
Town, where a ring-fenced fleet delivers baled waste over defined and limited origin-
destination pairs to a landfill site outside Cape Town. 
 143
Schematically this can described as a circle with many origin-destination pairs 
inside, as depicted in Figure 26. 
 
 
 
Figure 26: Schematic representation of metropolitan traffic 
 
The corridor typology describes transport over long distances between 
metropolitan production and consumption centres. More than 40% of current 
corridor demand is for freight flows between Gauteng and the Durban and 
Cape Town metropoles. Origin-destination pairs for the longer part of the 
journey are few and distances long, but with distributed local collection and 
delivery points. The long-distance portion of the freight could theoretically be 
transported more cost effectively by rail, but the many origin and destination 
points make this difficult. Rail transport between the convergence points only 
would mean the double handling of break bulk. This situation makes corridor 
demand an ideal candidate for intermodal solutions to avoid double handling 
or excessive long-haul road demand. Schematically this can be described as 
a pipe which is exactly the same as the schematic for primary traffic, but 
connected to a metropolitan circle with convergence points at the end points 
as depicted in Figure 27. 
 144
Corridor
Convergence point
 
 
Figure 27: Schematic representation of corridor traffic 
 
The rural typology is a longer-distance typology than metropolitan traffic, and 
could be shorter or longer than corridor traffic.   It describes transport within 
rural areas, between rural areas and from rural areas to connection points on 
corridors. This typology is currently significantly under-supplied in South 
Africa, with the consequence that commercial farming feels the strain of a 
deteriorating rural transport infrastructure network. The provision of rural 
freight transport is a core requirement for enabling the “second economy” as 
South Africa tries to elevate subsistence farming to commercial farming. Many 
origin-destination pairs are distributed over large, sparsely populated areas. 
This typology is primarily road-based as the development of new rail networks 
is not justified by low densities, but better rail solutions are possible where 
infrastructure is in place. Schematically this can be described as the pipe and 
circles of corridor traffic, but without convergence points. Rural traffic is rather 
independent from the pipe and circles in terms of origin-destination pairs, 
although it could sometimes be linked to the corridor traffic, at many different 
and often independent connection points (Figure 28). 
 
 145
 
Figure 28: Schematic representation of rural traffic 
 
The characteristics of the various typologies are summarised in Table 38. 
 
 Primary Corridor Rural Metropolitan 
Traffic 
type 
Bulk (rail only, 
mainly export 
coal & iron ore) 
Mostly 
manufacturing, 
some agriculture 
Mostly 
agriculture 
Mostly final delivery 
Distance Long Long & short 
Medium & 
short 
Short 
OD’s 
Few, one-
directional 
Long distance, few 
ODs; short 
distance: many 
OD's 
Many Many 
Major 
challenge 
Global 
competitiveness 
Spatial 
organisation, 
efficiency 
Development 
Congestion 
alleviation 
Possible 
logistics 
approach 
Ring-fenced, 
bulk rail 
systems 
Intermodal 
solutions 
Effective road 
feeder 
system 
World-class 
commuter systems 
amidst effective 
freight delivery 
 
Table 38: Summary of typology characteristics 
 
Supply and demand for the typologies per mode can be calculated as supply 
equal to observed behaviour and demand equal to what primary customers 
report (i.e. what freight owners indicated they have available for transport), 
what intermodal best standards would indicate for corridor traffic and what the 
distribution between rail and road should be, given certain densities on 
corridors and rural areas.  This is depicted in Figure 29 
 
The data from Table 32, Table 33 and Table 35 compared to the current 
structure of the transport market in South Africa as depicted in Figure 29, 
point directly to the decline of the rail transport mode in South Africa with no 
concomitant rise in intermodalism as in the rest of the world. The issues of 
high-rated traffic initially cross-subsidising agriculture (i.e. causing investment 
in an unprofitable and unnecessary rural rail network), and the absence of 
research and planning based on the research, led to this distortion in South 
 146
Africa’s economy, where transport efficiencies, because of the spatial 
challenges, are of critical importance.  
Primary
Primary
Metro
Rural
Rural
Corridor
Corridor
Corridor
Metro
0
20
40
60
80
100
120
140
160
180
Pe
rc
e
n
ta
g
e
Primary 84 0 100
Metro 29 0 113
Rural 73 0 125
Corridor 90 10 167
Rail Intermodal Road
Tonkm supply expressed as % of tonkm demand for each mode
Tonkm supply = 
tonkm demand
Tonkm supply > 
tonkm demand
Tonkm supply < 
tonkm demand
 
Figure 29: Supply and demand of modes in the typologies
268
 
7.4 The current position of rail 
The economic causes of rail supply are best described by a theoretical case, 
which provides a good average of the situation in South Africa as well as an 
understanding of the drivers of rail sustainability. 
 
In the average theoretical case for South Africa, prior to the deregulation of 
the transport sector, a theoretical assumption of the transport of 10 million 
tons over a 1 000 km corridor per annum, i.e. 10 billion ton-kilometers, can be 
made. Income and costs are more or less in balance before deregulation (with 
some inefficiencies, but a lower profit drive), which means that capital 
replacement cost, maintenance and running cost covers the income – in this 
theoretical case this would be (at an assumed 20c per ton-kilometer) around 
R2 billion (at today’s prices).  
 
                                                 
268
 Supply in the typology is calculated from known observations. The intermodal supply 
typology can be calculated from known rail volumes. Demand is calculated 
conservatively assuming 25% of traffic being high-value with road as a balancing 
element. The primary rail demand is calculated from reports of under-supply by large-
volume exporters of mining commodities. 
 147
In normal rail economics about 75% of costs would be fixed over the short 
term and 50% over the medium term. What would happen during deregulation 
is that traffic would begin to shift from rail to road (especially time- and value-
sensitive cargo) – and this movement will be accelerated and pronounced in 
the absence of intermodal solutions.  
 
Corridor traffic in South Africa could have shifted far beyond 50% from rail to 
road in a relatively short period, but assuming a 50% shift, this would mean a 
loss of R1 billion income on the theoretical corridor and only a reduction of 
R0.25 billion of costs in the short term and R0.5 billion in the medium term. 
This means that R0.75 billion  would have to be cross-subsidised from rail 
“captured” freight in the short term and R0.5 billion in the medium term. In the 
absence of cross-subsidisation the remaining traffic would have to carry a 
tariff of 35c per ton-kilometer in the short term and 30c in the medium term 
(an immediate tariff increase of between 50 and 75%). 
 
Invariably in South Africa the following hybrid, unsystematic strategy was 
followed to alleviate this situation: 
1. Cross-subsidisation did occur to some extent, making South Africa’s 
“captured” rail traffic slightly less competitive – this was partly offset by 
state of the art world-class engineering, especially on the export 
lines;
269
 
2. Relative tariff growth on some higher-value, time-sensitive freight did 
occur at faster than PPI growth rates, but this led to a worsening of the 
problem as more traffic left the railway;
270
 
3. The complete difference could never be earned, resulting in declining 
investment and expenditure on maintenance, inducing further freight 
losses and a never-ending downward spiral. 
                                                 
269
 It is common knowledge that Spoornet at some stage in the late 1990s made a profit of R1 
billion on the export coal and iron ore lines and a loss of R1 billion on its general freight 
business. Strategies exist to eradicate this cross-subsidisation, but it still persists to some 
extent. 
270
 Total road rand per ton costs increased by 3.7% from R108.36 to R112.39 from 2003 to 
2004. Total rail costs increased by 12.2% over the corresponding period from R67.99 to 
R76.26. The cents per ton-kilometer figure for road showed a 0.7% increase from 57.3c 
to 57.7c, whilst rail increased by 10.4% from 11.5c to 12.7c. Note the still comparatively 
low figure for rail, if all traffic is included in the equation. 
 148
 
Globally this problem usually has three possible outcomes: 
1. Railway decline on shorter haul is allowed to continue and the railways 
concentrate on bulk heavy haul over long distances (the American 
model); 
2. Investment in intermodal solutions to get the best of both possible 
worlds (part of the European and American models); 
3. Re-regulation (part of the European model). 
 
The first option could be expensive for the economy and could be identified 
with the typical “conspicuous consumption” culture of North America. 
Congestion and social factors in South Africa necessitate the adoption of 
Europe’s option. The problem is that South African railway management in the 
mid-1990s planned for the American option, even though this would have 
been impossible to implement.  It does not fit South Africa’s objectives, would 
cause job losses and place unnecessary strain on the economy. This means 
that the “normal” problems of transport deregulation were compounded by 
poor strategy and that assets were allowed to deteriorate even further to the 
brink of collapse. No investment in intermodal solutions occurred and these 
problems were even further compounded by an absence of guidance from the 
policy makers, lack of the development of a revised National Transport Policy 
following the De Villiers deregulation report and lack of implementation of 
MSA initiatives. 
 
One of the major causes of this situation was a poor understanding of the 
network typologies, the real trends in transport for these typologies and the 
limited or erroneous assumptions that followed this shortage of information. 
The belief in the economy was that rail decline was “normal” (erroneous - it 
shouldn’t have been so pronounced for corridors and especially long-haul 
corridors) caused by poor service (erroneous – poor service was definitely not 
the only cause of decline as the theoretical example shows) and not rapid 
(erroneous – masked by the rise of “rail captured” export machines and the 
concomitant absence of a network view).  
 
 149
The drivers of rail sustainability are best described by comparing or offsetting 
the positive contribution of a railway to an economy with its disadvantages. A 
railway will, under the right circumstances, save an economy money, provide 
systemic access for freight and passenger movements, and provide 
environmentally sustainable transport solutions. These advantages should be 
offset or compared to its major disadvantage, i.e. that it provides only one 
degree of freedom of movement. The only way in which the advantages of rail 
can be monetised in spite of the disadvantage is by not competing with other 
modes directly, but by exploiting the intrinsic technologies of rail, i.e. its 
bearing, guiding and coupling technologies compared to the capacity that it 
can leverage. Bearing, which indicates the weight of axle load that can be 
maintained (and therefore volumes) and guiding, which indicates the wheel on 
track differentials (and therefore speed of movement) are added to coupling, 
which means long trains with massive volumes (therefore combining high 
volume time and long-distance solutions).
271
  These intrinsic technologies 
consequently describe two drivers of rail technologies, i.e. the speed of 
guiding and the axle load of bearing (Figure 30). 
 
Be
ar
i
n
g
Guiding
Hea
v
y 
axle load
Lig
h
t
 ax
le
 
loa
d
Low speed High speed
AB
CD
 
Figure 30: Rail competitiveness and sustainability
272
 
 
                                                 
271
 Van der Meulen, 2007.  
272
 Van der Meulen, 2007. 
 150
The four areas of competitiveness in this depiction of drivers indicate that 
position: 
• A is suitable for heavy haul traffic; 
• B is suitable for heavy intermodal traffic; 
• C is suitable for fast intercity high-value traffic or passengers; 
• D is suitable for general freight solutions in a regulated environment. 
 
Van der Meulen maintains that all railways in the D group (where South 
Africa’s rail system is located) will gradually become redundant and that the 
problem can only be solved at the state level, where it was created (i.e. by 
redesign).
273
 
 
South Africa’s rail system was designed without A, B and C in mind, was 
highly regulated for a long time (and could therefore survive for a long time), 
but was destined to fail with deregulation. This failure is therefore a 
combination of the incorrect application of rail economics caused by de-
densification of loads (as a result of deregulation), which in turn was caused 
by the absence of intermodal solutions, and gauge and network design errors. 
All of these errors could have been avoided by macro-economic research on 
actual demand, coupled to design options to meet that demand in a 
sustainable package of solutions and design. 
 
Macro-economic research of this kind requires time-series information, and to 
understand this it is necessary to see the detailed picture that underlies the 
typologies and the trend in traffic movement over the typologies within the 
national freight flow model’s measurement history.   
 
7.5 Trend in freight movements over the typologies 
The growth in tonnages over the various typologies is depicted in Table 39. 
 
 
                                                 
273
 Van der Meulen, 2007. 
 151
 1993 1997 2003 2004 
Road 96 122 142 166
Corridor tra
ffic 
Rail (excluding 
dedicated export 
lines) 
54 61 57 56
Rail (dedicated export 
lines) 
72 85 94 86
Road 208 243 306 328
Rural traffi
c 
Rail 33 39 34 37
Road 387 483 578 612
Metropolita
n
 
traffic 
Rail 20 20 13 13
 
Table 39: Tonnage growth over various typologies (million tons) 
 
The change over time can also be depicted as market share within each 
typology (Table 40). 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 152
 1993 1997 2003 2004 
Road 64% 67% 71% 75% 
Corridor tra
ffic 
Rail (excluding 
dedicated export 
lines) 
36% 33% 29% 25% 
Rail (dedicated export 
lines) 
100% 100% 100% 100% 
Road 86% 86% 90% 90% 
Rural traffi
c 
Rail 14% 14% 10% 10% 
Road 95% 96% 98% 98% 
Metropolita
n
 
traffic 
Rail 5% 4% 2% 2% 
 
Table 40: Percentage market share within typology 
 
The issue around real rail decline is now clear, indisputable and entirely 
scalable. The results of these observations can now be compared to the 
typology challenges as summarised in Table 38 and the specific and known 
challenges of the South African economy in terms of transport. 
 
All growth over dense corridors occurred in the road mode, which expanded 
by more than 70% over eleven years. This growth would be understandable if 
the corridors in question were short or the density per corridor low. In this 
instance the economy will have to absorb this growth in the road mode. 
Cheaper options are, however, available in intermodality, if the density per 
corridor can be calculated as sufficiently high. In South Africa’s case, as 
observed by the national freight flow model, the spatial efficiency objective of 
the corridor typology is not achieved. If this density is sufficient to entertain an 
intermodal solution, future investment should consider investing in such 
solutions. This could release funds for the development of the second 
economy in rural areas, which is a major objective of the country at the 
moment. 
 153
 
Rural road traffic grew more slowly than corridor traffic (less than 60% growth) 
which correlates with the hypothesis that South Africa is not succeeding in the 
stimulation of rural economies as desired. A major cause of this failure is 
declining road infrastructure, the impact of which can now for the first time be 
measured. The development objective of this typology is therefore not 
achieved.  Metropolitan growth is also slower than corridor growth.   
 
The detailed results per typology will point towards specific issues and 
possible solutions.  These results are discussed in the following sections. 
 
7.6 Trends in movement over the various corridors 
The various corridors performed differently in terms of growth over the past 11 
years. The growth is depicted in Table 41. 
 
Gauteng-
Durban 
Eastern 
corridors
Gauteng-
Cape 
Town 
Northern 
corridors
South 
Eastern 
Western 
corridors
Road 21 18 8 11 6 5
1993 
Rail 14 9 2 5 3 1
Road 20 24 12 17 9 7
1997 
Rail 16 12 2 5 3 1
Road 25 27 15 23 10 6
2003 
Rail 13 13 3 4 3 1
Road 32 28 18 25 11 8
2004 
Rail 12 11 3 3 4 1
 
Table 41: Corridor growth from 1993 to 2004 (million tons) 
 
 154
Growth on the most dense corridor, the route between Gauteng and Durban, 
was relatively slow (26%), but this is also the corridor that is probably the 
most overstretched in the country because of current density. Alternative 
routes (even as far afield as the Energy-Demoina route - the alternative route 
between Gauteng and Durban over Piet Retief that is much longer) are often 
used because of inter alia lack of policing, lower density and fewer toll roads. 
From this high base, if the road/rail market share position should continue, 
major long-haul congestion problems will arise in the future. 
 
The Eastern corridors describe the roads to Witbank, Nelspruit and Maputo 
from Gauteng. There has been a major initiative over the past few years to 
develop this corridor, but freight has grown by only 46% (the 2nd lowest and 
slower than economic growth) over the time period. The expected growth in 
rail traffic has also not been realised yet. 
 
The Gauteng to Cape Town corridor achieved 135% growth over the time 
period, almost all on road. This is perhaps the greatest error in South Africa’s 
infrastructure planning framework, as this is also the longest corridor and 
should, by any standard, be more rail-bound than the rest. 
    
7.7 Trends in movement in the various metropolitan areas 
Almost all known areas (or areas usually classified as metropolitan) 
experienced growth of between 50% and 60% over the period, as depicted in 
Table 42. 
 
 
 
 
 
 
 155
 Gauteng Durban 
Cape 
Town 
PE 
Bloem-
fontein 
Witbank
274
 
EL 
Road 216 61 61 20 17 11 2
1993 
Rail 11 2 2 1 Neg 3 1
Road 267 76 76 23 24 12 3
1997 
Rail 12 2 2 1 Neg 4 Neg
Road 317 89 97 29 28 12 6
2003 
Rail 6 1 1 1 Neg 4 Neg
Road 331 103 99 32 27 13 6
2004 
Rail 6 1 1 1 Neg 4 Neg
 
Table 42: Metropolitan freight transport growth between 1993 and 2004 (million tons) 
 
This downward pattern of rail freight decline is probably quite normal for the 
shorter distances involved and might seem acceptable to planners, but it 
should be remembered that dense metropolitan areas are growing from an 
already high base (given the current available infrastructure that is installed 
and planned).  In addition, 86% of all metropolitan freight in South Africa 
moves within three metropolitan areas and 41% of all freight shipped in South 
Africa originates and terminates within one of these same three metropolitan 
areas. 
 
7.8 Trends in movement in the various rural areas 
The biggest rural growth was experienced in KZN, where rural freight 
movements grew three times faster than in the Eastern Cape, as indicated in 
Table 43. 
 
                                                 
274
 Witbank includes the Witbank and Middelburg areas. 
 156
 
Gauteng 
KZN 
Mpuma-
langa 
Free 
State 
Nort
h 
Wes
t
 
Limpopo 
Wes
t
ern 
Cape 
Easter
n 
Cape 
Northern 
Cape 
Road 54 29 44 23 16 19 13 9 1 
1993 
Rail  9 1 8 11 Neg 4 Neg 1 
Road 61 41 46 23 21 23 18 9 1 
1997 
Rail  10 1 10 12 Neg 5 Neg Neg 
Road 72 60 54 31 29 27 19 12 2 
2003 
Rail  6 1 8 11 Neg 8 Neg Neg 
Road 75 67 58 33 30 28 22 14 2 
2004 
Rail  7 1 8 12 Neg 9 Neg Neg 
 
Table 43: Rural freight transport growth between 1993 and 2004 
 
Government’s failure to deliver on objectives for the rural Eastern Cape region 
is clearly evident. 
 
 
7.9 Developing transport performance measures 
The application of the national freight flow model as lead indicator has now 
been adequately defined, given the specific planning information for the 
typologies that was generated. Possibilities also exist to use the data to 
develop lag indicators that can measure the performance of the economy in 
terms of transport consumed. GDP data, physical production data and 
sectoral GDP are known and a calculation is therefore possible (Figure 31). 
 157
700
800
900
1000
1100
1200
1300
195
7
197
1
198
5
198
6
198
7
198
8
198
9
199
0
199
1
199
3
199
7
200
3
200
4
200
5
R
a
nd pe
r
 t
o
n
 
Figure 31: Constant rands (2000 figures) produced in the South African economy per ton
275
  
 
The time-frame over which the performance of the national freight flow model 
is measured is short (with only five data points), and the previous 
measurements erratic and unreliable.  The increasingly poor performance of 
the economy in terms of transport productivity is, however, clearly visible 
where the economy generated less than R800 for each ton transported in 
2005. This measure over time could inform the spatial performance of the 
economy (about which only hypotheses have existed up to now) and 
contribute towards a better understanding of the spatial dilemma. As the 
tertiary sector in a mature economy grows, the measure should actually 
improve and, furthermore, also improve with increased productivity in 
transport. This is, however, not the case in South Africa. 
 
The measure can be extended to indicate measures for the transportable 
economy (Figure 32). This measure also indicates a declining trend and the 
poor performance of spatial reorganisation, spatial requirements and transport 
productivity in South Africa. 
                                                 
275
 Per ton moved for distinct typologies and for the sub-typologies within each typology.  
(note unequal intervals between years). 
 158
150
200
250
300
350
1957 1971 1985 1991 1993 1997 2003 2004 2005
R
a
nd
 pe
r
 t
on
 
Figure 32: The performance per ton of the transportable economy in GDP terms
276
  
 
7.10 Infrastructure planning 
The lead indicators (modal market share in total, modal market share per 
typology and correlation of freight transport with GDP per typology) which 
could inform transport and logistics infrastructure planning, and the lag 
indicators (performance of GDP and transportable GDP per ton) that measure 
past performance of transport in the economy have now been defined. 
 
The lead and lag indicators that were informed by the application of the 
national freight flow model not only prove to be a highly reliable measure of 
transport, but they also lead to the development of data that can begin to 
inform infrastructure decision making based on the structure of transport in 
the economy, trends and performance indicators. In order to look into the 
future of South Africa’s economic infrastructure requirements, a more detailed 
view of future transport requirements will be necessary, specifically around 
the commodities that are transported now and those that will require transport 
over the next two decades.  The model developed for this purpose is 
discussed in the next chapter.  
                                                 
276
 Per ton moved for distinct typologies and for the sub-typologies within each typology (note 
unequal intervals between years). 
 159
Chapter 8 - Commodity flow model methodology 
8.1 Introduction and background 
The development of freight flow data in South Africa was an important first 
step towards producing intelligence for future infrastructure planning and 
performance measurement of transport in the economy. For the first time 
transport supply and modal market share can be measured reliably and flows 
of freight determined, as discussed in Chapters 6 and 7. These flows are, 
however, for total freight without an indication of the composition of this freight 
(which is not possible given the methodology). An important missing link in the 
picture is therefore the determination of commodity flows over the typologies. 
Just as freight flow determination was attempted by previous surveys and all 
failed in this regard, so determining commodity flows was also attempted, but 
with no success.  
 
A commodity flow model was therefore developed to quantify commodity 
flows over the typologies that were developed in the national freight flow 
model. The commodity flow model should also provide forecasts that can be 
used to model various future infrastructure solution scenarios as well as 
developing low road, likely road and high road scenarios of commodity 
transport demand in the future. 
 
The core modelling process is based on 354 magisterial districts, resulting in 
finely granulated data sets for flow modelling. The process followed is 
depicted in Figure 33. 
 
 
 
 
 
 
 
 160
National I-O model
Apportionment -
Supply and demand 
per commodity on a 
geographical basis
Allocation - Flows 
per commodity
Actual data – based 
on publications and 
personal interviews
Macro-economic 
data
Macro-economic 
forecast
Commodity 
forecasts
Verify with: 
•Known flows
•Rail data
• National freight flow model
Consolidation of 
data into corridor 
and rural flows
Step 1 Step 2
Strategic interpretation and 
presentation of results
Step 3
 
Figure 33:  Commodity flow modelling methodology 
 
8.2 Supply and demand per commodity on a geographical basis 
8.2.1 Inter-sectoral tons moved 
The modelling of supply and demand on a geographic basis per commodity 
uses, as a first step, the input-output table (I-O table) of the economy as a 
platform.  
 
A modern I-O table is a tool in economics by which a system of national 
accounts is extended, classified and depicted in tabular format. The basic 
structure of an I-O table is based on the same framework as Leontief’s 
original statistical I-O table, which quantifies all transactions that took place 
between the main economic stakeholders or industries in a particular year.
277
 
The key feature of the table is that it divides these economic transactions into 
the main sectors of the economy, starting with Agriculture, Forestry and 
Fishing right through to Community Services. The I-O table is therefore an 
extension of the National Accounts of a country, i.e. disaggregating National 
Accounts into the various sectors of the economy.  Thus by its nature the I-O 
table gives detailed information on the intermediate and final demand 
components of each commodity in the economy. 
 
                                                 
277
 McDonald and Punt, 2005, p. 65. 
 161
De Jong et al. discuss the various options for supply and demand modelling 
(in their case called production and attraction); they state that advantages of 
input-output models are the link to the economy, the possibility of getting land-
use interactions and the policy effects of elastic coefficients. The listed 
disadvantages are the need for an input-output table (which could be 
constructed in this case), the demands for multi-regionality (it is often not 
possible to distribute inputs and outputs over various sub-regions of a country, 
but it was possible for South Africa to be divided into 19 sub-regions and 
magisterial districts), restrictive assumptions (which were solved by enhancing 
the models with desktop research) and the need to identify imports and 
exports (which is possible for the South African economy).
278
 
 
For the analysis two models needed to be developed. Firstly, a customised 
model, based on the I-O table of the economy to estimate the supply and 
demand,
279
 in volumetric terms, according to product type on a national level. 
Secondly, a model was developed to disaggregate these national figures on a 
regional and magisterial district basis.
280
  
 
Step 1 – Setting up the base year (2006) 
A 2006
281
 I-O table was used as the basis for the model.  The 2006 I-O table 
from Quantec shows imports as an aggregated single figure in the production 
process for each sector.  For the purposes of this model, the I-O table had to 
be transformed to accommodate the flow of imports into the economy in each 
of the transactions that occur in the economy.  Import coefficients were 
developed using the proportion of imports to production. These import 
coefficients were multiplied across the matrix to incorporate imports. The 
following equation was used to develop the import coefficients:  
                                                 
278
   De Jong et al. 2004 p. 108. 
279
 Demand and supply are disaggregated into the following components. Supply 
components: imports and production. Demand components: intermediate demand, 
exports, investments and consumption expenditure. 
280
 I-O modelling was done with the assistance of Conningarth Economists. The objective of 
the dissertation is to develop a modelling system, and not to contribute to the specific I-O 
modelling system 
281
  The 2006 I-O Table was developed by Quantec, a South African firm that specialises in 
data analyses. 
 162
 
 
(P+IM)/P 
where  
P=Production 
IM=Imports.  
 
Step 2 - Forecasting 
This step involved creating I-O tables for each year from 2004 to 2025. This 
was achieved by using forecasts of local demand, exports, imports and output 
for each year. 
 
Forecasting is designed to help decision making and planning in the present.  
Forecasting is by nature rooted in uncertainty.  In order to arrive at the best 
possible answer, a combination of forecasting techniques was used in this 
model. Three different methodologies around forecasting were followed for 
the three different transportable sectors of the economy, because of the 
inherently different structures of these sectors. Agricultural forecasting is the 
most difficult and requires specific treatment with respect to the unpredictable 
nature of the sector. Mining output is highly predictable and planned as 
continuous production runs years in advance. Manufacturing output is a 
combination of the two, but it is not dependent on weather and meteorological 
factors.  The manufacturing sector is also well forecasted in the economy, 
with reliable models that can be used. It should be noted that the primary 
objective of this study is not forecasting output per se, but developing a model 
that translates forecasts of output into flow forecasts per typology for the 
various commodities in the economy. 
 
This dissertation has argued that the researcher, especially in this specific 
field where strategic guidance from the regulator is observed to be poor, 
should become involved in the strategic positioning of the research. The down 
side (and especially if the involvement is too intense and close to 
management of infrastructure planning) is also a possibility, i.e. scientists 
 163
involved in future studies have indicated that the future that is forecast in such 
a case could be expressed as a desire to create that specific future. Scientists 
often refer to “creating the future” – in the words of physicist Dennis Gabor  
“the future is invented, not predicted”.
282
   Stakeholders’ current and medium-
term actions will therefore influence the future that is created. 
 
8.2.1.1 Forecasting agriculture sector  
In this sector available forecasts differ widely and it was therefore decided to 
use the consensus method amongst experts to obtain valid forecasts. 
 
The consensus method involves seeking expert opinions from more than one 
person. Each is an expert in his own discipline, and it is through the synthesis 
of these opinions that a final forecast is obtained.  This was attained in the 
model by the involvement of experts from various industries and disciplines 
within the agricultural sector during workshop sessions and individual 
interactions. This included developing a narrative of the future, which could 
later be translated into an actual figure.  The results of these sessions were 
compared to historical trends (trend extrapolation) and debated by agricultural 
experts to ensure that all known current and future events are discounted in 
the three forecasting scenarios. 
 
8.2.1.2 Forecasting mining sector 
Various published industry sources that are publicly available were used to 
determine mining output. Where uncertainty existed, specialists within the 
industry were consulted. Thereafter a consensus forecast was generated 
between a group of economists and the Department of Mineral and Energy’s 
mineral economists in charge of various mineral groups. Forecasts were 
compared to historic trends as well as information published by the 
Department. 
                                                 
282
 Source: StatPac Inc, http://www.statpac.com. 
 164
8.2.1.3 Forecasting manufacturing sector 
Standard manufacturing output forecasting models like those of ABSA are 
available and were used. These forecasts are the results of an elaborate 
system of quantitative analyses coupled with, and to some extent controlled 
by, qualitative evaluation of each sector’s unique characteristics. 
 
8.2.1.4 Intermediate transactions/demand in the economy 
Sections 8.2.1.1 to 8.2.1.3 describe the methodology to forecast the three 
main sectors individually.  However, it is important to take note that these 
sectors have an interrelationship.  An example of this is the agriculture sector 
that buys materials from the manufacturing sector (for instance, chemicals, 
fertilizers and fuel) to produce agricultural products.  If the agriculture sector 
expands in terms of certain products, this will have a positive effect on certain 
manufacturing industries. 
 
The input-output model was used to calculate these various interrelationships 
by making use of the standard input-output formula to calculate the output per 
sector by taking into account the interrelationships in the economy. The 
equation is as follows: 
 
q
t
 = [I-A
t
]
-1
*f
t
 
 
where  
• q =  the vector of total sectoral production;  
• I =  the identity matrix;  
• A = the matrix of technical coefficients; 
• f = vector of final demand by sector; 
•  [I-A]
-1
 = Inverse of technical coefficient matrix measuring direct and 
indirect effects. 
 
The final demand vector f consists of individual demand elements, which are 
indicated by the symbols c, g, i, s, e and m, where: 
 165
 
c = final sales/outputs from the different sectors to households;
283
 
g = current purchases by the government sector from the different  
  sectors; 
iI = gross domestic fixed investments (sales/outputs by the different  
sectors) which are regarded as capital goods; 
s = changes in inventories.  The relevant figure can have either a  
  positive or negative value, depending on whether inventories  
 increased or decreased in the various sectors.  The changes in  
 inventories in the Input-Output Table denote the total change of 
a specific product irrespective of the sector of the economy 
where the changes occurred; 
e = exports by the different sectors; 
m = imports by the different sectors. 
 
If intermediate demand that takes place in the economy is taken into account 
it could be expected that production of the individual sectors that was 
originally forecast for the economy, would differ from the results of the input-
output system. An iterative process was followed to reconcile the two 
approaches in the end.  The advantage of the Input-Output Model approach is 
that in the end demand and supply must totally balance, which is the 
prerequisite of the freight modelling system which follows the forecasting 
process. 
 
8.2.1.5 Verifying forecasting 
Forecasting for all commodities in the model was verified by the Bureau of 
Economic Research (BER) at the University of Stellenbosch.  
 
It is emphasized again that the objective of this study was not forecasting 
commodity output, a well researched field in many cases, though detailed 
commodity forecasts on some commodities are not readily available, but 
                                                 
283
 The term “household” implies private consumption expenditure. 
 166
rather the modelling methodology required to translate commodity forecasts 
into commodity flow data. The BER model is discussed in more detail in  
section 4. 
 
This exercise was undertaken for three scenarios – low road, likely road and 
high road. 
 
Step 3 – Changing monetary terms to weight 
The objective of step 3 is to convert the I-O tables for each year from 
monetary terms into volumetric terms.  The estimation and projection of 
commodity markets are normally expressed in monetary terms. For this 
reason most of the modelling is done in monetary terms.  However, to 
facilitate analysis in the transport sector, it is more practical to express 
production magnitudes in volumetric terms. It was at this point that the 
production and demand components of the I-O table were converted from 
monetary to volumetric terms using a rand per ton ratio. This allows for the 
generation of the total volume (tonnage) that is produced in the economy, 
which ultimately needs to be transported by South Africa’s transport network. 
 
8.2.2 Estimation of freight on a regional basis 
The next task was to distribute this national supply and demand on a 
magisterial district basis. 
 
Step 1 – Regionalising volumes: Identifying regions 
Step 1 involved identifying and defining practical and workable regions that 
could be used to illustrate freight in South Africa transported on a regional 
basis.  Nineteen regions were identified for this purpose. The nineteen regions 
are based on the 1998 Development Profile, compiled by the Development 
Bank of South Africa (DBSA), where the provinces of South Africa were 
divided into a number of sub-regions. The DBSA, in collaboration with 
development organizations in the various provinces, subdivided the provinces 
 167
into functional sub-regions for analytical and planning purposes. Some of 
these sub-regions were then aggregated together to make them even more 
applicable for transport and planning purposes. They were aggregated on a 
homogeneous structure from an economic point of view. Areas of similar 
agricultural economic activity were grouped together. For example, areas in 
the Eastern Cape comprised mainly of Karoo-type vegetation, where sheep 
farming is the prominent economic activity, were grouped together. Areas of 
predominantly grasslands-type vegetation, consisting mainly of cattle 
ranching, were also grouped together. However, the boundaries of the original 
sub-regions set out by the DBSA were not broken; the sub-regions were 
merely added together, creating 19 larger sub-regions (listed in Table 44 and 
depicted on a map in Figure 34).  
 
Code Region description Province 
1 Sub-regions 1 and 3 – PE Eastern Cape 
2 Sub-regions 2, 4 and 5 – EL Eastern Cape 
3 Gauteng Gauteng 
4 West Coast – Saldanha Western Cape
5 Cape metropole & winelands Western Cape
6 Southern Cape Western Cape
7 Free State Free State 
8 Eastern, Far Eastern and Central North West 
9 Western and Southern North We
10 Bo-Karoo, Diamond Fields & Kalahari Northern Cape
11 Namaqualand, North Northern Cape
12 West, Lower Orange Northern Cape
13 Bushveld & Western Limpopo 
14 Northern, Lowveld, Central, Southern & Central Limpopo 
15 Zululand KwaZulu Natal
16 Thukela, Port Natal, Southern Natal, East Griqualand, Natal Midlands KwaZulu Natal
17 Lowveld Mpumalanga 
18 Highveld Mpumalanga 
19 Eastveld Mpumalanga 
 
Table 44 : Regional descriptions 
 
 168
 
 
Figure 34: Cartographical indication of regions 
 
Step 2: Regionalising volumes: Apportioning volumes 
In this step supply and demand needed to be apportioned to the different 
regions. The lack of data on a regional basis necessitated the use of 
secondary keys in certain cases to apportion supply and demand.  An 
approximation method was then used to apportion the demand and supply of 
a product in these 19-sub regions. The analysis was done on an individual 
demand-and-supply basis.  
 
An example of this approximation method can be shown by looking at private 
consumption expenditure. There are no figures available for private 
consumption expenditure on a product basis in a specific region. Thus various 
secondary keys were used to develop a structure whereby private 
consumption of a product can be apportioned to various regions. In the case 
of motor vehicles for private use, keys such as the following can be used: the 
distribution of higher-income people in the country; household income, 
consumption of electricity; car ownership; and the GGP of transport. These 
 169
secondary keys are then integrated by using weights assigned to each key 
according to its importance relative to other keys. The total consumption or 
production of a product was then apportioned on a regional basis using this 
calculated structure.  However, in certain cases, such as for instance the 
production of maize, there was no need to use this method, as it is possible to 
identify where maize was produced. 
 
Another example is the case of the intermediate use of coal, whereby 
intermediate use is estimated from the production of, for instance, electricity 
(specific coal-fired power-stations and other industries such as Sasol 1, 2 and 
3).  The location where such industries are situated is known. 
 
Step 3 – From regions to magisterial districts 
In this step the apportionments to the 19 sub-regions were advanced to 354 
magisterial-district levels. The same essential methodology was used. 
Magisterial districts are the smallest available geographical unit on which 
some forms of data are available from time to time. To achieve the scalability 
objective this is therefore the natural smallest unit that could be used.  
 
8.3 Estimation of flows per commodity 
8.3.1 Modelled flows 
Estimation of flows per commodity is possible through a gravity modelling 
approach using supply and demand data. The candidate tested this approach 
with visualisation tools depicting areas and the application of gravity decay or 
transport resistance rules per commodity manually on a large-scale map of 
the country commodity by commodity. (This process involves taking each 
commodity and writing each supply and demand weight of a commodity on 
each region in different colours and then doing the gravity-deduced flows 
according to specified rules manually. The derived flows are then tested and 
compared with known flows. The process worked, but was far too time-
consuming as a tool for the magnitude of the flows in question - theoretically 
 170
125 316 flows per commodity).  A software program specifically designed for 
gravity modelling was therefore used with the help of Stephan Krygsman of 
the Department of Logistics at the University of Stellenbosch. 
 
Gravity-based approaches (according to Krygsman
284
) are based on the 
premise that trade flows between origins and destinations are determined by 
measures of supply and demand, and a measure of transport resistance. In 
such models, according to Rand Europe, “the flow between zone i and zone j 
is a function of the product of production and attraction measures of zone i 
and zone j respectively divided by a some measure of the (gereralised) 
transport cost.”
285
 Production and attraction models are derived from zonal 
trip rate models or Input-Output tables and related models. Input-Output 
models have been discussed and zonal trip rates for production and attraction 
“are usually derived from classifying cross-sectional data on transport 
volumes to/from each zone in the area under investigation into a number of 
homogeneous zone types.”
286
 
 
These measures can be given in monetary values (values of interaction flows 
between supply and demand) or in trade flow tonnage, disaggregated per 
commodity type. 
 
Krygsman
287
 further explains that the amount of interaction – freight flows – 
between two areas is therefore presumed to be directly related to the 
attraction of the areas and inversely to the distance (or other resistance 
measure) between the two.  The fact that the amount of interaction is related 
to the distance between origins and destinations means that the attraction 
value of a destination is inversely proportional to the distance between origins 
and destinations. And inversely, the interaction between an origin and its 
surrounding destinations decreases as the destinations are further away from 
the origin. The function describing the attraction value between origins and 
destinations within a certain distance is called a distance decay function. The 
                                                 
284
 Krygsman, 2006. 
285
 Rand Europe, 2001, p. 7. 
286
 Rand Europe, 2001, pp. 4-5. 
287
 Krygsman, 2006. 
 171
principle of this function is derived from physics, specifically from the law of 
gravity formulated by Newton. 
 
The measure of transport resistance refers to a transport cost variable of 
overcoming the spatial distribution between supply and demand locations.  
Travel time (measured in minutes) or travel distance (measured in kilometers) 
or a more complex generalised transport cost function combining travel time 
and cost is often used as transport resistance. The modelling rules for decay 
in freight transport specifically that arose from the manual test performed by 
the candidate are: 
• Granularity of commodity – Where commodities are defined 
homogeneously, decay functions are high. This means that commodities 
that are exactly the same will always be consumed where the demand is 
closest to production. In cases where the commodity description is broad, 
decay functions are lower. This means that commodities are often 
exchanged between regions, because supply and demand that might be 
in geographical proximity are not necessarily for the same sub-type of the 
commodity; 
• Brand considerations – Brand considerations work in the same way as 
granularity, except that certain high-value commodities, such as cars, 
might have a fine granularity of description, but very high brand affiliation; 
• Value – Value combines granularity and brand in that low-value 
commodities will tend to increase the decay function irrespective of brand; 
• International borders and geographical and other spatial characteristics 
can also advance the decay function.  
 
In the application of these rules on the computer model, three types of 
distance decay functions are usually used (and tested for appropriateness): 
• A neutral function. This is used only to demonstrate that there actually is 
distance decay.  It was only used to test various trade flows and whether 
or not any distance decay does exist; 
• An exponential function. Compared to a power function (see paragraph 
below), an exponential function represents a quickly declining distance 
 172
decay.  Based on Spoornet data, various commodities were modelled 
using this function;  
• A power function. Compared to an exponential function (see paragraph 
above), a power function represents a gradually declining distance decay.  
This is often the case with commodities such as manufactured foods. 
The functions used in this specific model are summarised in Table 45. The 
generation of decay functions is a specific study in itself and the research in 
this dissertation did not include further analysis of the appropriateness of 
specific techniques to develop these functions. This a suggested topic for 
further research and for that reason the decay functions actually used are 
given. A follow up study in the use of gravity modelling for freight flow 
quantification per commodity could research this specific aspect further. 
 
Commodity 
Distance Decay 
Function 
Distance Decay 
Parameter (Beta) 
Manganese Exponential 0.0009675 
Cement Exponential 0.0064419 
Bricks Exponential 0.035 
Other chemicals Exponential 0.035 
Coal mining Exponential 0.0417226 
Stone quarrying: Other Exponential 0.0417226 
Stone quarrying: Limestone Exponential 0.0417226 
Other basic iron and steel products Exponential 0.0417226 
Mining of chemical and fertilizer minerals Exponential 0.0417226 
Industrial chemicals Power 0.01 
Automotive Powe 0.01 
Processed foods Power 0.15 
Beverages Powe 0.15 
Break bulk Power 0.15 
Dry bulk Power 0.15 
Perishables Powe 0.15 
Wood and wood products Power 0.5 
Other non-metallic minerals Power 0.5 
Crude petroleum and natural gas Power 0.5 
Non-ferrous metal basic products Power 2 
Fertilizer Power 
Fuel & petroleum products Power 2 
 173
Chrome Power 2.041654 
Iron ore Power 2.776749 
Maize Powe 4 
Sugar cane Power 5.6 
 
Table 45: Decay functions for various commodities 
In general, one of the generic approaches available consists of three 
formulas: 
Tij = Ai.Bj.Oi.Dj.f(Cij,b)…………………………………………………(1) 
Ai = 1/(Sj.Bj.Dj.f(Cij,b))……………………………….……………….(2) 
Bj = 1/(Si.Ai.Oi./f(Cij,b))……………………………………………….(3) 
 
f(Cij,b) = exp(-b.Cij) in case of an exponential function 
f(Cij,ß) = Cij
-ß
) in case of a power function 
 
where: 
Tij = the estimated number of trips between origin i and destination j 
Ai = the balancing factor for origin i 
Bj = the balancing factor for destination j 
Oi = the constraint value for origin i 
Dj = the constraint value for destination j 
b = the distance decay parameter 
Cij = The distance between origin i and destination j 
 
The balancing factors ensure that the sum of the estimated outflows per origin 
equals the known origin total, and the sum of the estimated inflows per 
destination equals the known destination total. 
 
Formula 1 calculates the actual trips in the origin/destination matrix. Formula 
2 equates the total number of trips from origins in the matrix to the set number 
(the “origin constraint”). Formula 3 equates the total number of trips to the 
destinations in the matrix to the set number (the “destination constraint”). 
 
 174
The value of the distance decay parameter should be known before the 
equations above can be solved.  
 
In order to derive a distance decay parameter, observed freight flow 
information and distance or travel time between the various supply/demand 
locations are required.  The latter requires, for each origin-destination pair, a 
distance/travel time measure of separation.  The travel time/distances can be 
based on airline (straight line) distances or on a transport (rail/road or 
intermodal) network. As no suitable (that is connected or cleaned network) rail 
network existed, a national rail network had to be constructed for South Africa.  
Travel distances between all origins and destinations were derived using this 
network.  Transport costs (or a generalised cost) could not be derived as 
costs (freight rates per commodity per distance) were not available. 
 
In addition to the rail network, a detailed national road network was also 
constructed.  This allowed road travel times (and costs) to be estimated 
between the various origins and destinations.  While this is not directly 
relevant for this study, it does allow researchers to assess the travel time 
differential between road and rail for the origin-destination pairs.  The travel 
time differentials (and function derivatives such as the travel time ratio = travel 
time by road divided by travel time by rail) has been known to play a key role 
in road/rail mode choice decisions.  This information will be incorporated in 
future versions of a commodity flow model. 
 
Observed (i.e. rail) freight flows provided information on the travel distance 
that freight was moved.  Existing freight flow information is used to calculate 
the mean trip length (MTL) of the observed trips. With this information the 
distance (decay) freight will be transported (for the different commodities) can 
be calculated. The calculated mean trip length of observed freight flows and 
the origin/destination distance matrix is subsequently used as input to 
calculate the distance decay function.  
 
Unfortunately, due to the fact that only actual rail freight flow information is 
available (which in most instances accounted for only a small market share), 
 175
the distance decay could not be derived (for all commodities) from these 
observed data.  Based on hypothesized distance decay supported by 
theoretical knowledge gained from the visualisation exercise, various distance 
decay parameters were tested (and compared where possible) with rail data.  
The ‘best fitting’ distance decay parameters were subsequently selected.   
 
Gravity models, as discussed above, are the most widely used approaches 
internationally to distribute freight flows between origins and destinations and 
have been operationalised in various international freight flow models, 
including the Dutch SMILE (Strategic Model for Integrated Logistics 
Evaluations) and TEM (Transport Economic Model),
288
 the Finnish study on 
different distribution models and the Great Belt traffic model.
289
   
 
8.3.2 Enhancing modelled flows based on the Pareto process 
Once commodity flows were estimated, the top 23 commodities by weight 
(approximately 90% of total weight) were investigated further by desktop 
research and interviews to compare modelled data with what could be 
established in real life. All discrepancies were investigated to: 
• Improve the modelled data in the I-O approach, but also to improve the 
actual workings of the model – this means that the data were improved for 
the commodities investigated, but the improvement of the model also 
caused improvement of data for commodities that were not investigated, 
through application of the rules that could be generated by the research. 
• Improve the modelled data of the flow model, but also improve the actual 
workings of the model – with the same data improvements for 
uninvestigated data as for the I-O model relating to decay functions. 
 
The approach therefore allowed for both highly reliable models and flow data. 
 
                                                 
288
 Raha and Williams, 2002, pp. 38 and 40. 
289
 Rand Europe, 2001, p. 7. 
 176
8.4 Long-term forecast of the South Africa economy – the validation 
process 
The BER’s (Bureau of Economic Research) long-term forecast of the South 
Africa economy was used to validate the forecasts used in the I-O model. 
8.4.1 Broad methodological processes 
The BER’s methodology broadly consists of three processes. 
• Generating a standard macro-econometric forecast of the SA economy 
over the forecast horizon, 2006 to 2025.  Assumptions are made for the 
exogenous variables in the model; simulations were run and then  
subjectively adjusted by BER researchers.  The output of the model 
consists of the main expenditure components of GDP (household 
expenditure disaggregated into 22 sub-components; fixed investment into 
5 asset categories), imports, exports, balance of payments data, price 
and financial variables (see a description of the model below).  A long-
term forecast is essentially a trend forecast.  It is not possible to project 
the business cycle over a 20-year period, something which can be done 
over a 5-6 year forecast period.  It is therefore acceptable to provide 
projected growth rates in 5- or 10-year averages.  It is also important to 
conduct a proper supply-side analysis from a growth accounting point of 
view.  This entails an analysis of the economy’s production potential, 
being a function of three variables, i.e. the growth in the fixed capital 
stock (determined by the fixed investment rate), the growth in the labour 
force (or employment) and, thirdly, the growth in Total Factor Productivity.  
The BER’s macro-econometric model contains such a (Cobb-Douglas) 
production function. 
• The expenditure component forecasts were then fed into Quantec 
Research’s RSA Inter-Industry input-output model to estimate real value-
added, output, employment, import and export growth across 46 industry 
groups based on the BER demand-side forecast. 
• These projections were then evaluated and adjusted where necessary by 
BER researchers. 
 
 177
For the purpose of this study, where the BER had been tasked to evaluate the 
macro-economic assumptions and projections underlying the  freight flow 
projections, a demand-side macro-econometric long-term forecast was 
compiled (2006-2025) and used to recalibrate an existing supply-side 
projection compiled in January 2006.  Given the nature of the macro-
econometric modelling, the BER could not evaluate the projections for the 
individual commodities.  The BER/Quantec projections were used as a control 
for the aggregated sector projections.   
 
8.4.2 The BER’s macro-economic forecasting model 
The broad structure of BER's model can be described as that of a demand-
orientated macro-econometric model. Although the model essentially 
determines South Africa's gross domestic product (GDP) from the demand 
side (i.e. GDP is determined as the sum of final consumption expenditure by 
households, final consumption expenditure by general government, gross 
fixed capital formation, inventory investment and exports of goods and 
services, less imports of goods and services), specific supply elements in the 
form of a measure of potential output and economy-wide capacity utilisation 
have been included in an attempt to capture the production side of the 
economy. Capacity utilisation, which is measured as the inverse of the gap 
between actual and potential output, enters the equations for imports and 
prices as a variable supply constraint. 
 
The BER's model of the South African economy contains 135 equations, of 
which 30 are econometrically estimated equations and 105 are identities and 
transformations. The latest version of the E-views econometric package, E-
Views 5, was employed to estimate the equations and to compile the model. 
Co-integration techniques were used to estimate the majority of the 
behavioural equations in the BER's model. These techniques have several 
advantages compared to the standard techniques such as Ordinary Least 
Squares (OLS). The most important advantages are that they provide an 
answer to the so-called spurious correlation problem and provide for 
 178
 179
                                                
specification of both the long-run theory-based relationships between the 
variables as well as the short-run dynamic relationships.
290
  
 
 
The BER also verified the sectoral forecasts (i.e. on a sector level rather than 
commodity level). 
 
8.5 The final model output 
The final model output was a 2006 view as well as a forecast view of 2006-
2020 of all land freight transported in South Africa, per commodity, between 
354 possible locations (origins or destinations). This specific output provided, 
for the first time in South Africa’s history, an answer to a question posed by 
Verburgh in 1957 and for which Verburgh, Smith, Hamilton and Pretorius 
sought to find a solution, but couldn’t with the survey approach that they used. 
This made it possible to at least begin to make infrastructure investment 
decisions based on what the economy would require rather than what 
politicians dictate. 
 
 
 
 
 
 
290
 Laubscher, 2006. 
Chapter 9 - Commodity flow model results and 
strategic interpretation 
9.1 Introduction 
The application of the commodity flow model to the South African economy 
requires a few initial modelling rules around scenario delineation. Commodity-
level forecasts were produced for three growth scenarios, i.e. low growth, high 
growth and likely growth. The underlying macro-economic assumptions for the 
scenarios assume ongoing political stability and that the government 
continues to provide market-friendly economic policies.  Furthermore, it is 
assumed that the new economic growth strategy (ASGISA) and its 
successors are reasonably well implemented and supported by the private 
sector and other stakeholders.  Monetary policy will continue to be conducted 
within an inflation-targeting framework and the target remains at the current 3-
6% range.  Fiscal policy also continues to be characterised by disciplined 
expansion as has been the case since 2001, resulting in budget deficits 
ranging between 0 and 3% of GDP. These broad assumptions were therefore 
accepted as the likely scenario for the forecasting framework in which the 
model was developed whilst more onerous assumptions as a low growth 
scenario were used. 
 
The assumptions and scenario delineation were discussed with economists in 
detail, but it is once again emphasised that the objective of the modelling 
exercise was not forecasting per se, but the creation of the tool. (It is true that 
the tool will be worthless without decent forecasting, but the process of 
forecasting is not the focus here.)  
 
9.2 Analysis of the economy by weight 
The first consideration of this process is to model the weight of the GDP of 
South Africa, i.e. to determine the total tonnage produced and imported into 
South Africa or consumed and exported from South Africa. Although total 
 180
production and imports were 575 million tons in 2004, with mining supply at 
379 million tons, agricultural supply at 55 million tons and manufacturing 
supply at 142 million tons, the fastest growth rate over the past 70 years was 
achieved for manufacturing (Figure 35). 
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
193
6
194
1
194
6
195
1
195
6
196
1
196
6
197
1
197
6
198
1
198
6
199
1
199
6
200
1
200
6
I
n
dex
 19
36=
100
Mining Agriculture Manufacturing
 
Figure 35: Growth rate of economic sectors by weight 
 
Transport strategy in South Africa up to 1990 was to concentrate on 
agricultural transport in rail, to develop mining rail solutions and ignore 
densification of the manufacturing economy on corridors. An analysis of 
growth rates in the economy by weight alone and coupled with the now known 
growth rates and modal shifts on corridors, indicates the inappropriateness of 
this strategy, to some extent because it was informed by little or no  research 
data. 
 
The results of the forecasting generate the following view of the future 
economy in terms of supply in tons (Figure 36). 
 
 
 
 
 
 181
600
800
1,000
1,200
1,400
1,600
1,800
2,000
2004 2009 2014 2019 2025
Year
M
i
ll
io
n
 to
n
s
Likely Low Growth High Growth
 
Figure 36: Expected future supply and demand in the South African economy 
 
Supply and demand does not equal freight movements over the typologies as 
the same freight would sometimes utilise more than one typology or sub-
typology. A link is therefore required between the future view of the economy 
by weight in tons and the number of times, on average, each ton would be 
observed on the typologies. The four measurements taken between the 1993 
and 2004 (Table 46) suggest that this figure is growing inversely exponential  
and it looks like a flattening could occur at approximately 2.3 observations, but 
further research is necessary (Figure 37). 
 
 
Economy by 
weight (millions) 
Observed on 
typologies 
(millions) 
Number of times 
observed on 
typologies 
1993 477 870 1.82 
1997 506 1 053 2.08 
2003 542 1 225 2.26 
2004 575 1 307 2.27 
 
Table 46: Number of times each ton produced is observed on typologies 
 
 182
1.5
1.6
1.7
1.8
1.9
2.0
2.1
2.2
2.3
2.4
1993 1997 2003 2004
N
u
mb
e
r
 o
f
 
t
i
me
s
 
Figure 37: Flattening of number of times freight appears on typologies 
 
It would seem that the number of observations over all typologies of each ton 
is increasing, but the growing granularity of the truck counts based on an 
increasing number of counting stations should be taken into account. The 
figure (observations of each ton over all typologies) grew by about 3.4% per 
annum between 1993 and 1997, 1.4% between 1997 and 2003 per annum 
and 0.5% between 2003 and 2004. A flattening of this increase is therefore 
observed that could flatten out completely at about 2.3 observations per ton, 
given the current trend.  This would suggest that the current 1.3 billion tons 
observed on the typologies could grow to more than 4.1 billion tons in the next 
20 years. 
 
The previous Verburgh, Smith, Hamilton and Pretorius surveys indicated a 
reverse trend. A decoupling of the relationship between shipment volumes, 
average length of haul and GDP was reported by Raha and Williams for the 
United Kingdom since the early 1990s,
291
 but in a recent study by Verny
292
 
the number of times that the transportable economy is transported for Europe 
shows an increasing trend (though still correlated), which agrees with the 
observations of this study (Figure 38). 
                                                 
291
 Raha and Williams, 2002, p. 47. 
292
 Verny, 2007, p. 120. 
 183
 
 
Figure 38: Verny's observation of freight transport statistics and GDP figures for Europe 
 
It could be that this trend is related to economic maturity, but further 
observations would be necessary to clarify this point. It does seems as if 
Raha and Williams’s further observation, i.e. that structural changes, caused 
by modal shift, caused the decoupling (they talk about modal shift causing 
improved road productivity and length of haul)
293
 and if this is true it can be 
hypothesised that the figure for the number of times that the transportable 
GDP is transported will rise initially, but then stabilise as the structural effects 
of modal shift stabilise. This observation, if proved correct, will also correlate 
with the findings of this study. 
 
It is at this point of the analysis that the commodity flow model begins to 
inform strategy in a real way, in that this growth cannot inform long-term 
infrastructure investment decisions without understanding where infrastructure 
would be needed, and the size and type of infrastructure that should be 
developed. These questions require a consideration of the types of 
commodities that will need transportation and the typologies on which the 
transportation needs per commodity will be required. 
                                                 
293
 Raha and Williams, 2002, p. 47. 
 184
9.3 Commodity growth 
At a commodity level the growth in economic demand for the top 23 
commodities by weight is shown in Figure 39 below. 
0 50 100 150 200 250 300 350 400 450 500
COAL MINING
IRON ORE
FUEL & PETROLEUM PRODUCTS
ST ONE QUARRYING: OT HER
CRUDE PETROLEUM & NATURAL GAS
PROCESSED FOODS
WOOD & -PRODUCTS
OTHER BASIC IRON & STEEL PRODUCTS
CEMENT
STONE QUARRYING: LIMESTONE
CHROME
MA IZE
AUTOMOTIVE
NON-FERROUS METAL BASIC PRODUCTS
BEVERAGES
SUGAR CANE
INDUSTRIAL CHEMICALS
OTHER CHEMICAL PRODUCTS
BRICKS
MANGANESE
OTHER NON-METALLIC MINERALS
FERT ILIZER
MINING OF CHEMICALS & FERTILIZER MINERALS
Tons('000 000)
2004 2025
 
Figure 39: Growth in demand by weight of the top 23 commodities
294
 
 
The demand for commodities can also be depicted by showing the growth in 
tons excluding coal (coal is the only commodity in South Africa that is often 
used by mine-mouthed industries, mostly electricity and conversion to fuel 
and other chemicals, therefore excluding 60% of the commodity from road 
and rail transport demand – a further 30% is transported on a ring-fenced 
export machine with defined and known needs).  The results show that the 
economy will grow the fastest by weight in basic energy-related commodities, 
basic earthworks, and building-related and food commodities (Figure 40). 
 
 
 
 
                                                 
294
 The 2025 forecast that was originally used, were changed to 2019 for further analysis due 
to specific data constraints. 
 185
-10.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0
IRON ORE
FUEL & PETROLEUM PRODUCTS
CRUDE PETROLEUM & NA TURA L GA S
STONE QUARRYING: OTHER
PROCESSED FOODS
WOOD & -PRODUCTS
OTHER BA SIC IRON & STEEL PRODUCTS
CEMENT
AUTOMOTIVE
CHROME
NON-FERROUS META L BA SIC PRODUCTS
OTHER CHEMICA L PRODUCTS
BEVERAGES
INDUSTRIA L CHEMICA LS
BRICKS
MA IZE
STONE QUARRYING: LIMESTONE
MA NGA NESE
OTHER NON-META LLIC MINERA LS
FERTILIZER
MINING OF CHEMICA LS & FERTILIZER MINERA LS
SUGA R CA NE
Tons('000 000)
 
Figure 40: Growth in demand by weight for top 23 commodities (excluding coal)  
 
When depicted as a percentage growth, however, it is apparent that growth in 
beneficiated products will be far greater than in unbeneficiated products 
(Figure 41).  
 
-30 0 30 60 90 120 150 180 210 240 270 300
AUTOMOTIVE
WOOD & -PRODUCTS
OTHER CHEMICAL PRODUCTS
FUEL & PETROLEUM PRODUCTS
IRON ORE
NON-FERROUS METAL BASIC
CRUDE PETROLEUM & NATURAL GAS
CHROME
OTHER BASIC IRON & STEEL PRODUCTS
INDUSTRIAL CHEMICALS
BRICKS
CEMENT
MANGANESE
BEVERAGES
PROCESSED FOODS
OTHER NON-METALLIC MINERALS
STONE QUARRYING: OTHER
COAL MINING
MAIZE
FERTILIZER
STONE QUARRYING: LIMESTONE
MINING OF CHEMICALS & FERTILIZER
SUGAR CANE
Growth %
 
 
Figure 41: Percentage weight growth in commodities 
 
 186
This analysis indicates that it is not only apparent (as shown in the previous 
historical analysis) that the manufacturing sector is growing faster by weight, 
but also that some of the high-value elements of the manufacturing sector are 
growing faster than the lower-valued elements. The observation about the 
folly of previous investment decisions is therefore highlighted and it should be 
remembered that this view is not necessarily commonly known, as the sheer 
size and voice of the mining community by weight tended to overshadow the 
wisdom of other points of view. 
 
The next step in the process is to answer the question of where these 
commodities will flow by applying the commodity flow model to the data. The 
data could potentially generate 354 by 354 by 23 distinct observations (i.e. 
close to three million observations, if 23 commodities are used) and it was 
therefore decided to concentrate on corridor flows for this analysis. 
 
9.4 Corridor demand 
In the calculation of corridor demand, a different approach was used to define 
corridors than the one used in the NFFM. This is because the model is an 
origin to destination-type model and not a model based on physical counts of 
truck movements. Furthermore, certain choices are involved in deciding to 
what extent certain origin to destination pairs are linked to a corridor or not. 
The final refinement of this approach would be to link the model to a 
geographic information system tool to inform this mapping more scientifically 
– a line of further research that is proposed in this dissertation and in which 
the candidate is already involved. 
 
The definitions used were quite flexible in defining specific origin to 
destination pairs on corridors. This means that if an origin to destination pair 
was deemed to include the corridor infrastructure for more than 50% of the 
possible route, the traffic was included on that corridor. Clearly a geographic 
information system tool will inform this part of the model design better and 
should be applied. 
 187
On two corridors a close correlation was found between the models. The 
national freight flow model estimated the South Eastern corridors to carry 15.4 
million tons of freight compared to the 14.7 million tons of the commodity flow 
model. The Gauteng-Durban corridor was estimated to carry 43.4 million tons 
by the national freight flow model and 42.2 million tons by the commodity flow 
model. It is believed that the same level of correlation could be possible if: 
• The model run is repeated with less  flexible corridor descriptions; 
• A geographic information tool is used as a platform for both models.  
 
The analysis shows that freight demand will continue to concentrate around 
the main freight corridors, if change in the current spatial structure of the 
economy is not engineered over the medium to long term. This is fortunate, 
since it allows for the construction of a high-density core network which will 
reduce unit transport costs and allow for the focusing of investment. 
 
Figure 42 below shows the total growth in corridor freight on the eight biggest 
freight corridors in the country. The forecasted demand suggests that it will be 
practically impossible to provide the infrastructure that the economy will need 
in the same rail/road configuration than is currently the case. Even the 
doubling of current rail supply will mean that road volumes will have to 
increase by 60 percent on average to meet demand in 2019 and almost 
double by 2025. By 2025 this will make transport extremely expensive, even if 
the extrinsic costs of road transport supply are excluded. Finding a 
sustainable way to address the challenge of the growth in corridor freight is a 
strategic priority for the country. 
 
 188
Gauteng-Lobatse
Gauteng-Beitbridge
Gauteng-Maputo
Gauteng-Durban
Gauteng-East London
Gauteng-PE
Cape Town-Durban/RBay
Gauteng-Cape Town
Road 2004
6.1
Spoornet 2004
2.2 2009 – 9.5
2019 – 13.0
Road 2004
9.1 
Spoornet 2004
5.7 2009 – 19.3
2019 – 29.8
Road 2004
28.5
Spoornet 2004
3.6 2009 – 37.4
2019 – 51.3
Road 2004
3.5
Spoornet 2004
2.8 2009 – 7.3
2019 – 10.1
Road 2004
31.9
Spoornet 2004
10.3 2009 – 50.1
2019 – 69.4
Road 2004
5.0
Spoornet 2004
0.4 2009 – 6.7
2019 – 9.2
Road 2004
8.4
Spoornet 2004
0.9 2009 – 11.4
2019 – 16.0
Road 2004
33.3
Spoornet 2004
1.6 2009 – 42.0
2019 – 61.7
Million tons 
2004, 2009, & 
2019
 
 
Figure 42: Expected growth in total traffic on corridors 
 
Corridor freight can typically be categorised as dry bulk, break bulk and liquid 
bulk.  Figure 43, Figure 44 and Figure 45 below show the projected corridor 
demand for these categories. 
 
Dry bulk demand will grow, but not very rapidly. In percentage terms, the 
Gauteng-Maputo corridor is expected to display the largest growth in dry bulk 
freight as the Maputo corridor is developed, because Maputo is closer to 
Gauteng than the other coastal harbours. Steady growth is expected on the 
Gauteng-Durban and Gauteng-Cape Town corridors as well. Rail’s under-
supply of dry-bulk requirements (as evidenced by its low market share) is 
clear and unambiguous proof of under-investment in infrastructure. It also 
represents most of the traffic that is well understood by rail marketers and 
some of the first flows that should be targeted for a possible modal switch. 
 
 189
Gauteng-Lobatse
Gauteng-Beitbridge
Gauteng-Maputo (incl. Witbank, Nelspruit)
Gauteng-Durban
Gauteng-East London
Gauteng-PE
Cape Town-Durban.RBay
Gauteng-Cape Town
Road 2004
0.9
Spoornet 2004
1.2 2009 – 2.2
2019 – 2.6
Road 2004
3.7
Spoornet 2004
4.3 2009 – 11.0
2019 – 17.5
Road 2004
9.9
Spoornet 2004
2.4 2009 – 14.2
2019 – 18.8
Road 2004
1.4 
Spoornet 2004
0.8 2009 – 2.3
2019 – 3.1
Road 2004
11.2
Spoornet 2004
4.0 2009 – 17.3
2019 – 21.4
Road 2004
2.3 
Spoornet 2004
0.2 2009 – 2.9
2019 – 3.9
Road 2004
4.3
Spoornet 2004
0.2 2009 – 5.1
2019 – 6.8
Road 2004
6.2
Spoornet 2004
0.2 2009 – 7.0
2019 – 8.9
Million tons 
2004, 2009, & 
2019
 
 
Figure 43: Dry bulk corridor demand 
 
 
Million tons 
2004, 2009, & 
2019
Excluding pipelines
Gauteng-Lobatse
Gauteng-Beitbridge
Gauteng-Durban
Gauteng-East London
Gauteng-PE
Cape Town-Durban/RBay
Gauteng-Cape Town
Road 2004
0.9
Spoornet 2004
0.6 2009 – 1.2
2019 – 1.5
Road 2004
0.4
Spoornet 2004
0.1 2009 – 0.6 
2019 – 0.8
Road 2004
3.9
Spoornet 2004
0.2 2009 – 4.6
2019 – 6.8
Road 2004
0.3 
Spoornet 2004
0.1 2009 – 0.4
2019 – 0.5
Road 2004
2.2 
Spoornet 2004
0.2 2009 – 2.7
2019 – 4.3
Road 2004
0.4
Spoornet 2004
0.2 2009 – 0.5
2019 – 0.7
Road 2004
0.4
Spoornet 2004
0.1 2009 – 0.6
2019 – 0.8
Road 2004
8.2
Spoornet 2004
0.2 2009 – 10.0 
2019 – 16.2
 
 
Figure 44: Liquid bulk demand on corridors (excluding pipelines) 
 
 190
Liquid bulk demand (excluding pipeline demand), on the other hand, is 
expected to show very little growth. It is anticipated that the bulk of new 
corridor demand will be accommodated by the pipeline network. 
 
Gauteng-Lobatse
Gauteng-Maputo (incl. Witbank, Nelspruit)
Gauteng-Durban
Gauteng-East London
Gauteng-PE
Cape Town-Durban
Gauteng-Cape Town
Road 2004
3.2 
Spoornet 2004
0.3 2009 – 4.3
2019 – 6.7
Road 2004
3.7
Spoornet 2004
1.2 2009 – 6.0
2019 – 9.4
Road 2004
11.0
Spoornet 2004
1.0 2009 – 14.1
2019 – 19.9
Road 2004
1.7
Spoornet 2004
2.0 2009 – 4.4
2019 – 6.3
Road 2004
15.6
Spoornet 2004
6.0 2009 – 25.9
2019 – 37.9
Road 2004
1.7 
Spoornet 2004
0.2 2009 – 2.2
2019 – 3.3
Road 2004
2.6
Spoornet 2004
0.6 2009 – 3.9
2019 – 5.8
Road 2004
14.7
Spoornet 2004
1.2 2009 – 19.3
2019 – 28.4
Million tons 
2004, 2009, & 
2019
 
 
Figure 45: Break bulk corridor demand 
 
It is estimated that about 80% of non-bulk dry and liquid cargo worldwide 
moves in containers (Figure 46).  
 
 191
 
Figure 46: Containerisation propensity
295
 
 
As international trade volumes continue to grow, developed countries in 
particular are driving the further containerisation of break bulk freight. The 
propensity to containerise will increase over the next 20 years, which will lead 
to an even faster growth in traffic that could easily use intermodal facilities.  
 
This situation clearly points the way towards an intermodal strategy and the 
availability of intermodal facilities where break bulk reaches critical densities. 
This level is deemed to be four million tons of break bulk on any corridor 
longer than 500 kilometers, based on Harris’s work, which will be explained 
further in the next chapter.
296
 
 
In order to understand this situation an analysis was done to determine how 
much freight will be connected by intermodal facilities through a network of 
connections as each new node in the network is commissioned.  (The 
analysis was based on terminal density in Gauteng – i.e. it was done in 
                                                 
295
 Container penetration refers to the degree to which all cargo that could be containerised 
actually is containerised. 
296
 Harris, 1977, p. 563. 
 192
declining order of the weight of freight arriving and departing from Gauteng for 
each node by weight). (Table 47) 
 
Tons added to Gauteng as node as each connection in terms of total and 
break bulk is added 
Tons 2004 Tons 2009 Tons 2019 
 
 Total 
Break 
bulk 
Total 
Break 
bulk 
Total 
Break 
bulk 
Durban 42 22 50 26 69 38 (c) 
+ Cape Town 74 38 92 45 130 66 
+ Maputo 89 43 111 51 160 75 
+ Port Elizabeth 98 46 122 55 176 81 
+ Lobatse 104 50 129 59 186 87 
+ Beitbridge 113 54 139 63 199 94 
+ East London 118 54 146 65 208 (a) 99 (b) 
 
Table 47: Impact of a cumulative set of intermodal facilities 
 
This means that, if all freight is containerised by 2019, total freight between 
Gauteng and the other seven nodes will amount to 208 million tons, with 99 
million tons of break bulk translating into about nine million loaded containers 
(TEUs) in Gauteng (see (a) and (b) in Table 47).  It also means that 38% [(c) 
as a percentage of (b)] of the potential domestic break bulk intermodal 
solutions could be achieved in 2019 by a solution connecting Durban and 
Gauteng. This figure will increase to 67% by adding Cape Town, 76% by 
adding Maputo, 82% by adding Port Elizabeth, and so forth. For the first time 
in South Africa’s history a planning hierarchy, based on actual commodity flow 
data, can be developed. 
 
9.5 The indirect advantages of commodity flow views 
Some indirect advantages of a commodity flow view were also identified. Of 
these the most important was the settling of the harbour-hinterland debate. In 
the absence of actual data, major corridors are often seen as feeder routes to 
and from ports. This approach means that the corridors are often expected to 
be “geared” to function as such and that major interventions, like the 
movement of terminals from one port to another or the development of new 
 193
infrastructure at one port rather than another, are expected to have a major 
effect on corridor density. 
 
This situation is especially true for South Africa’s major general cargo port, i.e. 
Durban, and its major hinterland feeding route, i.e. the N3 from Durban to 
Gauteng. Of the 22 million tons of imports and exports handled by the harbour 
in 2004, it would have been difficult to estimate what tonnage would be used 
on the corridor, but all documents, plans and approaches for the corridor 
usually assume that this number is high. A simple population density 
assumption (which says that 50% of total tonnage is for local and direct 
hinterland consumption and 50% for consumption around the corridor 
hinterland) would mean that 11 million tons of the 42 million tons would be 
classified as port traffic on the corridor. With commodity matching (i.e. now 
having data with a finer granularity) and even with the most negative 
assumptions around the calculation – i.e. allowing for a two-thirds 
consumption of harbour imports at the end of the corridor per commodity and 
allowing for two-thirds exports for commodities coming from the Gauteng end 
of the corridor, it is difficult to identify more than 10% of traffic on the corridor 
that relates to the port – the figure is probably lower. (This calculation means, 
for example, that if the quantity of rice imported through Durban harbour is 
known – and it has usually been known up to now, though not always – and if 
the amount of rice that flows north on the corridor is known, a limiting factor 
for rice emerges. If more rice is on the corridor than the quantity imported, the 
rest must be domestic traffic, plus at least some of the rice that is imported 
must be for consumption in Durban. By repeating this sum per commodity and 
also doing the sum the other way around for exports, clarity can be achieved 
on the maximum of freight on the corridor that relates to harbour activities – 
and this volume is much lower than what is often expected). The case for 
information clearly emerges from this case study. 
• In many reports today the corridor is still classified as a major contributory 
feeder to the harbour and its raison d’etre described as such. 
• An analysis of traffic flows on corridors indicates that at least 75% of the 
traffic on the corridor is unrelated to the harbour. 
 194
• An analysis of commodities on the corridor shows that at least 90% of the 
traffic on the corridor is unrelated to the harbour. 
 
It is true that the corridor is important for the port, but not necessarily the other 
way round. If this insight  can be confirmed, the case for domestic intermodal 
traffic could have been made earlier and the problem of spatial organisation 
(or rather disorganisation) identified earlier. 
 
The spatial challenges are not only an issue for South Africa’s densest 
corridor, but also for the second biggest corridor, i.e. Gauteng to Cape Town. 
In the 1970s Gauteng industries consolidated and mining experienced 
exponential growth. This situation changed around as many large companies 
moved their headquarters to Cape Town, professionals relocated there and 
the deregulation of agricultural industries encouraged market entry. These 
aspects stimulated growth in the Western Cape (Figure 47) and increased the 
bidirectional food traffic between Gauteng and Cape Town. This situation 
went by “unnoticed” by freight and logistics planners, as no flow model existed 
to inform planning. The corridor grew faster over the 11-year period between 
1993 and 2004 than the Gauteng to Durban corridor (135% vs. 26%) and the 
expected growth in bidirectional processed foods creates specific 
opportunities for national efficiencies to serve spatially disabled markets, as 
discussed below. 
-1
0
1
2
3
4
5
1
973
1
975
1
977
1
979
1
981
1
983
1
985
1
987
1
989
1
991
1
993
1
995
1
997
1
999
2
001
2
003
2
005
P
er
c
e
nt
age
South Africa Western Cape
 
Figure 47: Western Cape vs. South African GDP growth 
 195
 
9.6 Understanding break bulk commodities – the case of processed 
foods 
Processed foods is an important commodity that is transported between the 
three major nodes of Gauteng, Durban and Cape Town, amounting to 12.3 
million tons in 2004 and forecasted to be 19.6 million tons by 2019. 
 
About 50% of the break bulk commodities between Gauteng and Cape Town 
is processed foods and this position will not change by 2019. This figure is 
about 30% for the corridor between Gauteng and Durban and will not change 
relative to other commodities over the medium term. The model indicates that 
2.8 million tons of processed foods flowed from Gauteng to Cape Town, and 
3.2 million tons in the opposite direction from Cape Town to Gauteng. It 
further predicts that this distribution should be close to perfect balance in 
2019, with 4.6 million tons of processed foods flowing in both directions. 
Between Durban and Gauteng 4.2. million tons of processed foods flow from 
Durban and 2.1. million tons in the opposite direction. This distribution should 
remain more or less the same by 2019, when 7.0 million tons is predicted 
from Durban and 3.4 million tons towards Durban. 
 
This densification of a single commodity group (all perishables and liquids 
were excluded from this comparison) on two corridors with a combined length 
of more than 2 000 kilometers and a near perfect bidirectional flow, connected 
at a single congested node, i.e. Gauteng, is a clear and unambiguous 
indication of intermodal solutions. Never before has this case been stated in 
such stark and incontestable relief. 
 
This case, as a specific opportunity within the break bulk flows, will be 
analysed further in the next chapter, once costs can be added to the analysis. 
 
 196
 197
9.7 Conclusion 
The commodity flow model described for the first time specific demand 
conditions for freight and logistics infrastructure now and in the future. It also 
highlighted errors in certain assumptions about freight types in South Africa 
and specific opportunities that had never before been identified.  
 
The commodity flow model is now ready for an update and will be updated on 
an annual basis (should funding be available) in the quest for finding solutions 
to South Africa’s spatial challenges. The final element in macro-economic 
understanding of transport and logistics in South Africa is an understanding of 
costs. 
 
 
 
 
 
 
 
Chapter 10 - Logistics cost perspectives 
10.1 Introduction 
During the course of the research to determine modal market share the need 
arose also to develop an understanding of logistics costs in South Africa. It 
has been hypothesised already that surface freight transport is a big 
challenge in South Africa, probably more than in most other countries 
because of this country’s spatial challenges. The rest of the world has, 
however, moved on to a holistic logistics view, of which transport is seen as 
but a subcomponent. 
 
It should be remembered that transport is: 
• The largest cost factor in South Africa; 
• The most noticeable administered cost; 
• The most visible in terms of modal competition or coopetition. 
Transport also requires the most investment for combined domestic as well as 
international freight movements. 
 
A view of total logistics cost in South Africa would therefore: 
• Make the picture more complete and enable planners to understand 
transport in its total context, understand the various relationships of 
logistics stack elements and provide a complete view of logistics 
infrastructure requirements in South Africa; 
• Close the loop to determine performance of macro-logistics, also for 
transport. 
 
During the period in which the models in this dissertation were applied 
(between 2005 and 2007), the initial aim was to determine logistics costs and 
therefore the state of logistics in South Africa, but the scope of the dissertation 
was changed to concentrate on the finer details of the transport dimension of 
logistics, such as modal market share, freight flows and commodity flows. 
Whilst the candidate concentrated on this research and the spatial challenges 
 198
presented by transport, he initiated and obtained funding for a  task team of 
researchers at the Logistics Department at the University of Stellenbosch to 
develop a logistics cost model of the South African economy. The challenge 
was to incorporate the findings of such a model into the findings of the 
candidate’s own research. 
 
10.2 The development of the model 
Botes et al.
297
 refer to current methodologies on logistics costs measurements 
and Delaney’s popularisation of the concept.  Delaney applies a transportation 
cost component using Smith’s methodology and an inventory carrying cost 
component using the Alford-Bangs formula.
298
 The accuracy, difficulty of 
compilation and unsuitability of the methodology for logistics cost calculation 
are well described by the researchers (Botes et al.), but the historical context 
is of interest and is described below.  
 
As mentioned in Chapter 2, it is clear that the Delaney study still applies the 
Alford-Bangs formula,
299
 which is usually quoted from Alford and Bangs’s 
work of 1955 (probably because when Delaney himself sets out his approach, 
he quotes the 1955 version
300
). The original work, published in 1944, almost 
certainly contained the formula, but the candidate could locate only the 1946 
edition.  In the 1946 edition the formula is described
301
 and the authors initially 
estimated an inventory carrying charge of between 10% and 20%.
302
 Alford 
and Bangs then continue to question this wide range and eventually put 
forward a figure of 25%, based on a previous discussion of Alford in a book on 
Industrial Management, of which the oldest reference that can be found is a 
3rd edition, published in 1941.
303
 Furthermore, according to Alford and Bangs, 
the previous work of Alford in deciding on the inventory charge was based on 
the work of Parrish, which obviously predated Alford’s 1941 work by at least 
                                                 
297
 Botes et al., 2006, p. 3. 
298
 Botes et al., 2006, p. 5. 
299
 MacroSys Research and Technology, 2005, p. 6. 
300
 Delaney, 1987, p. 6. 
301
 Alford and Bangs, 1946, pp. 396-397. 
302
 Alford and Bangs, 1946, p. 396. 
303
 Alford, 1941. 
 199
one year. (No further information on Parrish could be found.) It is therefore 
safe to assume that the Alford-Bangs formula dates back to at least 1940 (if 
not earlier) and is still applied, to some extent, in its original form by Wilson 
today. (Delaney passed away after the publication of the 14th annual state of 
logistics report and Rosalyn Wilson, his earlier collaborator, is continuing the 
work, with the latest report, the 18th, published on 6 June 2007
304
.) 
 
The historic context of the Alford Bangs formula is important, because it 
highlights the fact that researchers in the world’s largest economy will not 
hesitate to use reasonably broad estimates to determine trends, as long as 
assumptions are applied uniformly and work is repeated on a regular basis. 
Researchers who insist on the finest detail of absolute correctness will 
struggle in the macro-economic and macro-logistics field.  An initial point of 
departure would be to at least improve on the very coarse granularity of 
Delaney’s and Wilson’s work. 
 
10.3 The global state of logistics 
The most publicised global research that has been conducted on logistics cost 
is that of Delaney, continued by Rosalyn Wilson after Delaney passed away.  
The 18th report indicates that in the United States a steady upswing of 
logistics costs as a percentage of GDP has been noticed in the last three 
years (Figure 48).  Wilson reported that transportation costs rose slower than 
expected, but costs in total were driven up by higher fuel prices, inventory and 
interest rates and warehousing costs.  Transportation costs were still at record 
highs.
305
 
 
                                                 
304
 Wilson, 2007. 
305
 Wilson, 2007, p. 12-14. 
 200
8
8.5
9
9.5
10
10.5
199
6
199
7
199
8
199
9
200
0
200
1
200
2
200
3
200
4
200
5
200
6
Pe
r
c
e
n
t
a
g
e
 
 
Figure 48: US logistics costs as a percentage of GDP
306
 
 
Rodrigues et al.
307
 developed an artificial neural network approach
308
 to 
compare global logistics costs by country.  They use 24 countries 
representing 75% of the global GDP in 2002 (37 859 billion USD – countries 
included in the study -  of 48 771 USD – total world GDP for 2002) for their 
approach
309
 and a combination of Wilson and Bowersox, Calantone, Closs 
and Stank’s approaches and estimations.
310
 
 
The findings of Rodrigues et al. are worrying for developing nations as they 
report that logistics costs of the countries used in their study rose from 13.4% 
in 1997 to 13.8% in 2002.  They also summarise their results based on the 
income of countries in the study, as depicted in Figure 49. 
                                                 
306
 Wilson, 2007, p. 4. 
307
 Rodrigues, 2005, p. 1. 
308
 Rodrigues, 2005, p. 3. 
309
 Rodrigues, 2005, p. 5. 
310
 Rodrigues, 2005, pp. 4-5. 
 201
10
11
12
13
14
15
16
17
18
1997 2000 2002
Pe
r
c
e
n
ta
g
e
Low Lower middle Upper middle
High (non-OECD) High (OECD)
 
 
Figure 49: World logistics costs according to income groups
311
 
 
For the period of their research logistics costs decreased for all nations, 
except for the nations with lower incomes, and it is also higher for nations with 
lower incomes. Rodrigues et al. hypothesise that this could be caused by 
operational pressures in the first world and density considerations.
312
 Both of 
these observations are of importance for the situation in South Africa today 
where, as was illustrated, the development potential of the second economy 
has not yet been achieved and the full potential cost savings on densified first 
economy corridors have not yet been exploited. 
 
10.4 Results for South Africa 
The article by Botes et al. includes figures for 2003 indicating the results of 
their model’s first application to South African data. The model, together with 
the national freight flow model, is now repeated annually and it is possible to 
discuss their results for 2005. The results are published in the CSIR’s state of 
logistics survey, which the candidate co-authors on an annual basis. It 
estimates logistics costs for 2005 at R223 billion (R141 billion for transport), or 
                                                 
311
 Rodrigues, 2005, p. 9. 
312
 Rodrigues, 2005, p. 13. 
 202
14.5% of GDP, slightly down from 14.8% in 2003 and 14.6% in 2004. If the 
data provided are analysed in a different fashion from what the researchers 
provided in the state of logistics report, however, an interesting picture 
appears.  
 
Transport and profits (including management and administration) account for 
the major contribution in cost increases. The timeframes are short, but it could 
be worrying as South Africa’s transport costs as a percentage of logistics 
costs are 63.1%, compared to the lower than 40% in the rest of the world 
(Figure 50). 
 
80
90
100
110
120
130
140
2003 2004 2005
I
nde
x
Transport Storage and ports
Management, admin & profit Inventory carrying cost
 
Figure 50: Trends in logistics cost components (2003 % contribution indexed, 2003=100)  
 
Another interesting perspective is to measure the conversion factor of logistics 
spending (Table 48).  
 
 
Conversion of logistics 
rand into GDP rand 
Conversion of tons 
handled into GDP rand 
Primary 3.3 140 
Secondary 1.4 500 
 
Table 48: Conversion factors of logistics 
 203
The conversion factor indicates that the primary sector consumes logistics in 
an efficient manner. It only generates R140 of GDP for every ton handled 
(and of this, mining generates around R90 per ton handled and agriculture 
R250), but it converts every R1 spent on logistics into R3.30 of GDP. The 
conversion rate for the manufacturing sector is very low. It is obvious that the 
manufacturing sector would require more logistics complexity in solutions, but 
what is not so obvious is why few stakeholders acknowledge that South Africa 
is spatially challenged and that the manufacturing sector’s logistics should 
behave more like that of the primary sector, and especially mining, to be more 
efficient.   
 
It is here that the typologies point towards a solution, i.e. part of the 
manufacturing sector’s freight on corridors does behave like mining 
commodities, being the freight between convergence points as depicted in 
Chapter 7.  In fact, even more efficient solutions can be developed as this 
freight is bidirectional, which mining commodities are not. The case study that 
the candidate aims to propose in the next paragraph therefore poses the 
question that, if 44% of the freight on corridors is break bulk (as shown in 
Chapter 7) and if this tonnage is enough for efficient mining type solutions 
when transported between intermodal hubs or convergence points by rail,
313
 
should a large saving not be possible for about 11% (break bulk freight on 
corridors forms around 11% of all freight in South Africa - Figure 51) of all 
freight in South Africa, i.e. of R141  billion of logistics costs. Surely if 11% of 
total costs can be handled more efficiently, the logistics system deals with 
nearly R15 billion of costs systemically? This is once again an example where 
macro-economic data point the way towards specific areas of inquiry. 
 
 
 
 
 
 
                                                 
313
 As a crude rule of thumb it could be assumed that local delivery is paid for by what is 
saved on bidirectional traffic – refer Chapter 6. 
 204
Metropolitan, 66, 
9%
Breakbulk, 82, 
11%
Perishables, 13, 
2%
Liquid bulk, 21, 3%
Dry bulk, 68, 9%
Automotive, 4, 1%
Primary, 111, 15%
Rural, 376, 50%
Processed foods, 
27, 4%
Other, 55, 7%
Corridor traffic
 
 
Figure 51: Break bulk corridor traffic in the context of total traffic (million tons 2004) 
 
10.5 A cost case study for South Africa – taking macro-logistics cost 
measurement to the next level 
The following case study was developed to show how the argument, as set 
out in the previous section, could be developed further with the data now 
available. It is also used to determine whether at least a pre-feasibility study 
on certain solutions should have been developed earlier.  
 
Erero and Van Heerden (2005) analyse the role of freight rates in South Africa 
in depth and come to some interesting conclusions: 
• More study is necessary to determine the real costs, especially in terms of 
road vehicles and in terms of externalities; 
• Road vehicles are often over-cropped (leading to lower prices); 
• High traffic densities on rail are only possible on a few routes; 
• Future trends should be towards containerised rail transport between 
dense metropolitan hubs; 
 205
• Road tariffs are hard to find.
314
 
Further research in this regard is necessary but, now that densities are 
known, a pre-feasibility case study is developed based on 2004 data. 
 
10.5.1 Finding tariffs 
A major impediment in comparative research is usually road tariffs (and 
recently often also rail), which are deemed to be secret and unavailable. In 
terms of the macro-economic nature of the case study, a rough rule of thumb 
indication of costs would probably be sufficient, but it was decided to do a 
reasonably accurate calculation based on the state of logistics survey, and 
also to poll hauliers and/or freight owners for actual tariffs on the two main 
corridors. 
 
The logistics cost model pegs total surface transport costs in South Africa at 
R140.5 billion, of which rail earns R14.1 billion, leaving R126.4 billion for road, 
for which it delivers 223 billion ton-kilometers, i.e. at 57c per ton-kilometer. 
This figure is obviously different for freight movements over the various 
typologies, i.e. metropolitan, rural and corridor freight. Using rail distribution of 
tariffs, which are known, as a training case, the road cost per ton-kilometer 
could be distributed between the typologies with the same weights (Table 49): 
 
 
Rail cent per 
ton-km 
Road cent 
per ton-km 
Corridor 23 38 
Metro 70 114 
Rural 24 39 
 
Table 49: Comparable typology costs in cent per ton-kilometer
315
 
 
The rail training case suggested that metropolitan tariffs would be three times 
higher than corridor tariffs. Polling hauliers and freight owners indicated a tariff 
                                                 
314
 Erero and Van Heerden, 2005. 
315
 Based on logistics costs survey and rail cost distribution as training data. 
 206
of between 34c and 40c per ton-kilometer for corridor distances between 500 
kilometers and 1 000 kilometers (mostly for Gauteng to Durban traffic), and 
between 22c and 28c for longer distances (mostly for Gauteng to Cape Town 
traffic). It would therefore seem that an average bidirectional rate of 37c for 
the Gauteng to Durban corridor and 26c for the Gauteng to Cape Town 
corridor could be accepted. It should be remembered that this rate is 
bidirectional and, if an assumed 15% empty haul is included, a very close 
match between the logistics cost model figure and the polled figure from 
hauliers is achieved. 
 
Rail tariffs also differ for the different corridors and for the ring-fenced heavy-
haul export machines (which are based on one-directional traffic). These 
tariffs are 23.5c per ton-kilometer for the Gauteng to Cape Town corridor and 
32.6c for the Gauteng to Durban corridor. Once again, if the rail tariffs were 
taken as a training set and applied to the road tariff in order to determine the 
corridor split, the answer would have been exactly the same (i.e. the Gauteng 
to Durban corridor’s tariff is more or less 40% times higher than the Gauteng 
to Cape Town corridor tariffs). 
 
The 577 kilometer average transport distance coal export machine conveys 
around 65 million tons of coal at 8.3c per ton-kilometer and the iron ore export 
machine conveys 25 million tons at 3.9c per ton-kilometer. In the rail proof of 
concept case
316
 the 8.3c per ton-kilometer will be used for the Gauteng to 
Durban corridor and the 3.9c per ton-kilometer for the Gauteng to Cape Town 
corridor.  
 
Rail pick-up and delivery tariffs are well known. Even if Goosen maintains that 
all terminal costs are comparatively low in terms of total freight movement,
317
 
it should still be included. Erero and Van Heerden
318
 quote R328 for a 12-
meter container (with a maximum carrying capacity of 27 ton) over a 
                                                 
316
 “Proof of concept” refers to a system’s proven ability to deliver on a specific solution within 
specified parameters. The term is used here to describe rail’s proven ability to perform at 
specified low cost levels on dedicated export lines – given adequate densities 
317
 Goosen, 1979, p. 7. 
318
 Erero and Van Heerden, 2005. 
 207
maximum distance of 40 kilometers for Spoornet cartage rates. That will be 
R12.15 per ton for a maximum of 40 kilometers. 
 
10.5.2 Developing a model for a possible macro saving 
1. N
s
 = W
s
 - R
s  
a. Ns = Nett result for scenario 
b. Ws = Road cost for scenario 
c. Rs = Rail cost for scenario 
2. W
s
 =
 
T
tc 
x
 
V
s 
x D
 
a. T
tc
 = Road tariff for case 
b. V
s
 = Volume for scenario 
c. D = Distance 
3. Rs = T
rc
 x V
s
 x D + (2P x V
s
) 
a. T
rc
 = Rail tariff for case 
b. V
s
 = Volume for scenario 
c. P = Pick-up and delivery for rail 
4. Scenarios, 
s(pf,25%,50% or 75%)
:  
a. Primary commodity (processed foods) on corridor total switch = 
s(pf)
 
b. 25% of road break bulk to rail switch = 
s(25%)
 
c. 50% of road break bulk to rail switch = 
s(50%)
 
d. 75% of road break bulk to rail switch = 
s(75%)
 
5. Cases, 
c(tc,rbc,rpoc,rs)
  
a. 
c
 = Base case 
b. 
poc
 = Prove of concept case (Rail) 
c. 
s
 = Stretch case (Rail)  
 
Volumes and tariff cases for the two corridors are depicted in Table 50. 
 
 
 
 
 208
 Gauteng – Cape Town Gauteng – Durban 
Processed Foods 6.0 6.3Volume (in 
million ton) Break bulk 11.9 21.6
Road 37c 26c
Rail base case 32.6c 23.5c
Rail proof of 
concept 
8.3c 3.9c
Tariffs (in 
cent per 
ton-km) 
Rail stretch
319
4.2c 1.9c
 
Table 50: Volume and tariff cases for case study 
 
The tariff cases are based on the following assumptions: 
• The base cases are the current tariffs (for which rail is not profitable and 
road may be over-cropped); 
• The proof of concept case is the current rail export machine tariff. The 
export machines are engineered for few origin-destination pairs, heavy 
haul and are profitable. The exact same heavy haul design is therefore 
necessary between corridor convergence points, and pickup charges will 
be added; 
• The stretch case is the proof of concept case, but for bidirectional traffic. 
 
Some of the tariff cases might seem extreme, but since the volumes are now 
known, the question should rather be asked under what conditions these 
tariffs could be achieved? The export machines’ engineering provides some 
indication: 
• No shunting (break up) of trains; 
• Longer unit trains; 
• Even tempo; 
• Large volumes. 
 
The export machines were also designed for heavy haul in terms of gradients, 
movable frogs and loops. The extent to which this engineering would be 
required will depend on density. In the following scenarios densities such as 
                                                 
319
 The rail “stretch” case refers to what could be achieved if export line efficiency were 
possible 
 209
those achieved on the coal export machine will probably still not be 
necessary. 
 
It should be remembered that both export machines were built from scratch 
30 years ago, both are profitable and became profitable at volumes around 10 
million tons per annum. The case study therefore proposes a change from 
thinking that these conditions could only exist for the export machines to 
asking why specifically these conditions couldn’t exist for high-volume corridor 
transport between convergence points. 
 
One of the reasons why the export machines are successful is also the 
singularity of commodity. For this reason processed foods (excluding liquids 
and perishables), which are nearly always transported palletised in South 
Africa, were specifically included as a ring-fenced option in the model. Food 
production power is also highly concentrated in South Africa (mostly in the 
hands of only three companies), which could facilitate design discussions. 
 
10.5.3 Applying the model 
Applying the model results in the following answers for the two corridors 
(Table 51 and Table 52). 
 
 
 
 
 
 
 
 
 
 
 
 
 210
 
 
 
 Switching scenario 
 s=25% s=50% s=75% 
Processed foods 
(million tons) 
6.3 5.4 10.8 16.2 
Current 
volume & cost 
for road 
Road base        1 399 1 199     2 398     3 596 
Rail base        1 390      1 191    2 382     3 574 
Rail proof of 
concept 
          471         404       808     1 212 
Cost to economy 
Rail stretch           316         271       542        813 
Rail base               9             8         15          23 
Rail proof of 
concept 
          927         795    1 590     2 385 
Savings 
Rail stretch        1 082         928    1 855     2 783 
 
Table 51: Gauteng to Durban model result (rand millions, except first data row)  
 
 
 
 
 
 
 
 
 
 
 
 
 
 211
 
 
 
 s=25% s=50% s=75% 
Processed foods 
(million tons) 
6.0 3.0 6.0 8.9 
Current 
volume & cost 
for road 
Road base  2 184  1 083  2 166   3 249  
Rail base  2 124  1 053  2 106   3 159  
Rail proof of 
concept 
 478  237  474   710  
Cost to economy 
Rail stretch  310  154  307   461  
Rail base  60  30  60   89  
Rail proof of 
concept 
 1 706  846  1 692   2 538  
Savings 
Rail stretch  1 874  929  1 859   2 788  
 
Table 52: Gauteng to Cape Town model result (rand millions, except first data row)) 
 
The cost saving for the two corridors combined at the 75% switch scenario 
would therefore be R4 921 million for the proof of concept scenario and R5 
571 million for the stretch scenario. 
 
The question is sometimes asked whether the export machines are really 
profitable and if this is feasible. The model was therefore repeated to 
generate, for an indicative summarised figure, just these 75% switch scenario 
totals for the two corridors for the poc and stretch case. The results are 
summarised in Table 53. 
 
 
 
 
 212
 
 
Percentage 
price increase 
Rail poc cost 
saving (Rm) 
Rail stretch cost 
saving (Rm) 
None 4 923 5 571 
10 4 794 5 507 
20 4 664 5 442 
30 4 535 5 378 
40 4 405 5 313 
50 4 276 5 249 
60 4 146 5 184 
70 4 017 5 119 
80 3 888 5 055 
90 3 758 4 990 
100 3629 4 926 
 
Table 53: Cost saving for 75% switch scenario with price increase options 
 
Even if double the current export machine tariffs could be charged by 
Spoornet and all other efficiencies can be equalled, Spoornet will become 
highly profitable and the country (in the poc case) could save 4% of its total 
freight bill. 
 
The model delivers a few interesting results: 
• It proves that rail, in its current form, could not survive (a view shared by 
Van der Meulen); 
• It shows that, even with a one-directional proof of concept case, the 
switching solution would be worthwhile; 
• It shows that if the same conditions as with the export machines could be 
achieved: 
o The switching solution would pay for itself; 
o The solution would save the country (for just these two corridors) 6.6% 
of the nation’s freight bill; 
 213
o The saving would be equal in return to the country as if Spoornet in its 
current form lowered its rates by 60% for all traffic and turned 
profitable. 
It is true that some of the assumptions might be contestable, but the indicated 
return is so vast that further research is imperative. The case study also 
confirms the value of macro data to indicate possible avenues for exploration. 
 
A few salient points should be remembered in this regard: 
• The tariffs are not strictly tariffs in the normal sense of the word. They are, 
however, in the case of rail the exact average that was applied in each 
typology, and in the case of road the exact average that was calculated on 
a macro level and quoted on a micro level (no large difference was 
noticed). 
• Tariffs are known more generally than is usually assumed. 
• The export machines were developed and paid for through own funding – 
no government subsidy was required. 
• If a systemic development could be achieved, as in this case study, further 
advantages could develop: 
o Logistics hubs around the convergence points could develop with 
distribution centres – such as for major food manufacturers and 
retailers in the processed foods case that was specifically highlighted; 
o Placing the hubs at convergence points will alleviate congestion by 
taking one of the two trips (being from long haul to distribution centre 
and distribution centre to retailer) out of the congested metropole; 
o As carbon emission issues grow, only a single power source (the 
source that drive the trains) will have to change, but even with coal-
based electricity generation for rail, it would still emit less carbon per 
ton. 
The most important and salient feature of the argument is a density argument. 
A benchmark study by Harris researched the economies of density carefully 
for various rail operators in the United States.  Harris never comes to a 
specific conclusion, except that he maintains that it is important and that 
railways should be allowed to increase low-density rural rates and decrease 
 214
high-density corridor rates.
320
 This observation was sorely needed in South 
Africa between 1950 and 1990. Harris does, however, plot the negative 
exponential nature of his results and, interestingly, an inflection point of the 
curves can be read from his plot, i.e. at around 4 million tons for a 500-
kilometer length of railway. Anselm studied the German railway and came to 
the same conclusion; he indicated that wherever the slots are available to 
double volume, the costs and tariffs should decrease by 40%.
321
 This is a 
major, and often forgotten, factor in cases where capacity utilisation is below 
50% and traffic is available. This case is often not considered in rail business 
planning in South Africa, especially since it requires a step-function thinking 
approach rather than a gradual ramp-up, something which rail planners are 
often unwilling to do. 
 
As Kennedy indicated before deregulation and when the deregulation debate 
became widespread, a specific role for rail was expected with deregulation. 
This role was bulk long haul of low-value commodities, and “movement of 
containers in special trains…to provide excellent door to door service for 
many commodities between major centres.”
322
  The question should be asked 
why Kennedy’s prediction did not materialise? It is true that Spoornet probably 
failed in service delivery and the required competitive posture, but other 
reasons could also be proposed. 
 
10.6 Conclusion  
The last step in the process of a macro-economic view of logistics, and 
specifically transport, in South Africa was the understanding of costs. Through 
this, the loop from modal flows over typologies, by specific commodities, 
forecasted into the future and with a view on costs to the economy and 
possibilities for improvement, was closed. 
 
                                                 
320
 Harris, 1977, p. 563. 
321
 Anselm, 2002, p. 11. 
322
 Kennedy, 1982, p. 26. 
 215
 216
Through these data, and the annual repeat of the process, infrastructure 
planning and opportunities for improvement could be better informed and 
performance could be measured. 
 
Chapter 11- The contribution of this research and 
future challenges 
11.1 Introduction 
History has shown that man’s transport ability plays a core, even fundamental, 
role in development. The development of the wheel was a foundation on 
which more or less all further growth in society was founded. Without the 
ability to move people and produce, no specialisation would be possible with 
no complex economies, no complex manufacturing, no international trade and 
no globalisation. Yet, man would measure all other things on a macro scale 
and assume that transport would merely be “there”, automatically available 
and invisible. Like the sun that rises in the morning, the weather or the tides, 
transport would be available when needed. 
 
The unwillingness to measure the transport function on a macro scale has an 
interesting explanation and even more interesting side effects. It has to be 
assumed that transport is such a fundamental pre-condition to everything else 
that it has to be in place before anything else can happen. If it takes a team of 
people with oxen sixty-five days to move a few tons between Cape Town and 
Gauteng at a cost of R15 000 per ton in today’s terms, nothing much would 
move. On the other hand, if a railway is built and the journey can be 
completed in two days and at a fraction of the cost, it can be assumed that it 
is not really an issue any longer. 
 
The interesting implication of this explanation is that the same could be said of 
all basic resources in the world. Once the mere flick of a switch can provide 
power and lights, the turning of a valve can provide water, a single movement 
of the hand can dispose of sewerage and refuges, and the pressing of a 
button can provide communication, one could assume that these issues do 
not require much further thinking. Yet most of these normal daily functions are  
usually carefully managed, measured and planned. And all of these functions 
are suddenly heading for a critical point of depletion in their current form for 
 217
the first time in the world’s history.
323
  In addition, the damage caused by 
these functions will for the first time, within the next three decades, become 
untenable for man and will have to change. Where does this leave transport in 
South Africa? This is the case of double jeopardy in its worst form. 
 
The management of the South Africa’s transport infrastructure is behind 
developed nations (industrialised countries) in terms of depth and 
performance measurement. 
• Transport as such is often managed more poorly than other 
infrastructure functions. 
• All infrastructure functions are currently challenged and often believed 
to be inadequate for current development challenges. 
• Very little work has been done around considering sustainability issues 
in infrastructure development, especially in South Africa. 
• Of all the measures that would be required to manage sustainability, 
those that relate to transport (for instance CO
2  
emission tons per ton-
kilometer) are the furthest behind for transport – especially in South 
Africa. 
 
Smits
324
 analyses the reasons for poor performance in Sub-Saharan 
economies (Figure 52) and proposes three, i.e. geographical and resultant 
spatial factors, colonial legacies and institutional environment, which he 
explains as a lack of “equal access”.  
 
 
 
 
 
 
 
 
                                                 
323
 Martin, 2007, pp. 7-9. 
324
 Smits, 2006, pp. 2-3. 
 218
0
5
10
15
20
25
30
35
40
45
50
1913 1950 1973 2001
P
erc
e
n
t
age
Asia Latin America Sub-Saharan Africa
 
 
Figure 52: Regional GDP per capita as % of GDP per capita of the Western world
325
 
 
All three of his core fundamentals relate to transport, i.e. spatial factors could 
lead to high transport costs, the colonial legacy was to pump mining 
commodities out of the country (rather than beneficiate these at the point of 
origin) and thereby create centres far from the sea and access challenges, 
especially for the second economy, including transport deficiencies. 
 
11.2 The contribution of this research 
This research proposed a macro-economic dimension and in its outcomes 
improved the understanding of the South African economy in total. It was 
illustrated how limited this understanding used to be and how little attention 
transport infrastructure development planning and economic management on 
a macro-level used to be (and in many respects still is until the communication 
cycle that follows from this research is fully developed.) By quantifying flow 
and size as well as market share and future demand, sensible planning and 
management is possible and the current bottlenecks could be eradicated and 
its future manifestations minimised. 
 
                                                 
325
 Compiled from Smits, 2006, p. 8. 
 219
The research uncovered the volume and modal market share of all freight 
flows in South Africa. The methodology made historical comparisons possible 
and therefore highlighted problem areas for improvement. It provided the first 
comprehensive data set to allow for network segmentation and the 
understanding and quantification of the freight transport typologies in South 
Africa. The management of these typologies as both separate, but also 
integrated constructs is now possible. Corridors’ efficiency requirements, 
primary traffic’s global competitiveness requirements, rural traffic’s 
developmental requirements and metropolitan traffic’s congestion alleviating 
requirements can now receive the necessary individual and administrative 
focus as well as integrated funding and policy-making focus. Before this 
research many hypotheses about segmentation and network typologies were 
proposed, but the absence of quantification led to confusion as far as 
propositions about what was needed are concerned. 
 
The characterisation of these freight flows, within all these typologies on a 
quantitative commodity level, now and forecasted into the future, not only 
allows for better infrastructure planning requirements identification, but also 
for an improvement in modal balancing. By understanding the nature of freight 
on a quantitative level, specific modal balancing case studies can be 
researched. If the economic regulator and modal competitor understand the 
exact nature of what is required, the most efficient solutions can be sought, 
the underlying policies developed and management structures installed to 
improve freight transport in South Africa. 
 
The contextualisation of these freight flows in a logistics costs framework 
allows for the utilisation of the systemic trade-off characteristics of logistics 
within the total national context. It is now possible to measure the overall 
performance of the current system, and propose and measure possible 
solutions. 
 
Through this work the overall management of macro freight transport can be 
improved and the national economy benefited.  
  
 220
National economies in the world have unique challenges, but one of South 
Africa’s own unique problems is of a spatial nature. No other comparable 
economy exhibits the quantum of South Africa’s spatial problem (as the 63% 
of logistics spend on transport compared to the less than 40% for the world 
illustrates), which means that the transport sector should be especially well-
managed, which is often not the case. By proposing, developing and proving 
a unique and working modelling system a ground-breaking management tool 
becomes available. 
 
This is necessary, because not only is the comparable problem smaller in the 
first world, the first world is also moving forward to address new future 
challenges. These challenges should be defined and the usefulness of this 
research to contribute to these challenges illustrated. This should be done to 
also understand what future directions follow-up research should take. 
 
11.3 Development and growth objective of the economy – how to 
implement change 
One of the major impediments to growth is often the focus on individual 
advantage versus the common good construct. This focus on individual 
advantage could be direct, such as the migration of each extra ton from rail to 
road (to gain an individual speed or service advantage) up to the point that a 
large migration carries huge costs for the economy that has to be carried by 
the freight owners. This huge cost that is then occurred can also be indirect 
where the resultant damage of this action must be borne by a total economy. 
 
Hardin maintains that “the population problem has no technical solution; it 
requires a fundamental extension in morality”.
326
 Hardin’s excellent exposition 
of problems associated with micro-management results in an interesting 
proposal which he calls “mutual coercion mutually agreed upon” based on 
rational thought.
327
 He asks for careful measurement, then comparisons and 
then agreement on the necessity for change. He specifically warns that 
                                                 
326
 Hardin, 1968, p. 1243. 
327
 Hardin, 1968, pp. 1247-1248. 
 221
systems will not change if a “perfect” alternative cannot be found, even though 
the less than perfect alternative might be better than the status quo. The only 
way for advancement in this situation (which is more or less the position in all 
cases when change is necessary) is to rationally measure the two alternatives 
(status quo and alternative for change), decide on a course of action rationally 
and then implement it. Hardin calls this a “moral” approach, not technical, and 
through this description elevates rational planning and the implementation of 
this planning to the highest social order possible. 
 
This work solves the measurement problem, and therefore enables society to 
engineer alternatives that move away from the status quo. On the other hand, 
some of the underlying issues that Hardin’s work uncover refers back to 
human nature itself. Future work should therefore look at the macro change 
management requirements for the implementation of the planning which this 
work enables. 
 
11.4 Distribution of wealth objective of the community – the national 
development application of this work 
Demkes and Tavasszy discuss the difference between macro- and micro-level 
indicators in depth and find the problem in the meso area (Figure 53), 
because “the macro indicators focusing on welfare maximisation are mostly 
decomposed into meso level indicators focusing on welfare optimisation, 
under the condition of subsidiarity, for sectors or regions, and not on supply 
chains”.
328
 They therefore blame both macro-economics and the regulatory 
side for not having a supply chain view and the supply chain side for not 
connecting with macro objectives. 
 
 
 
 
 
                                                 
328
 Demkes and Tavasszy, 2000, pp. 15-16. 
 222
Macro 
level
Meso 
level
Micro 
level
Traditional macro 
level goals
Welfare maximisation:
Long term focus on sustainability and growth
Decomposition
Geographical or 
industry goals
Supply chain 
goals
usually no linkage
Welfare optimisation:
Sector focus on shorter term
(under condition of subsidiarity)
Supply chain optimisation:
Sector focus on shorter term
(under condition of win-win)
Combination of company goals
Traditional micro 
level goals
Profit maximisation:
Short term focus on shareholder maximisation
 
 
Figure 53: Demkes and Tavasszy's description of micro, meso and macro indicators 
 
They continue to say that: 
It appears that traditional micro indicators are useful for internal and 
external measurements of single companies’ logistic performance, but 
inadequate for the whole, interrelated, supply chain. A similar 
statement can be made for traditional logistics performance indicators 
on a macro level. They are useful for comparative analyses among 
different nations, but they are inadequate for comparing different 
supply chains. Moreover, there is very little information on regional 
differences i.e. intra-European differences and the effectiveness of 
policy measures in the area of logistics and telematics facilities. Supply 
chain indicators could be seen as a combination of aggregated micro 
indicators, disintegrated macro indicators and meso indicators. Yet, the 
macro indicators focusing on welfare maximisation are mostly 
decomposed into meso level indicators focusing on welfare 
optimisation, under the condition of subsidiarity, for sectors or regions, 
and not on supply chains.
329
 
                                                 
329
 Demkes and Tavasszy, 2000, p. 16. 
 223
  
As given in Figure 53, policy-makers should seek the linkage between on the 
one hand the macro and meso level indicators and on the other the supply 
chain indicators, aimed to create a win-win situation.
330
 
 
Demkes and Tavasszy add a new and interesting perspective to the issue. 
They claim, by definition, that macro-level measurement of logistics 
performance is not only important as lead and lag indicators for macro-
management of logistics and infrastructure development, but for all societal 
development (the welfare imperative in their definition) and that meso welfare 
objectives and measurement should be decomposed into supply chain goals.   
 
People in South Africa might think that the issues of the second economy are 
important, but hidden to the rest of the world. It seems that this might not be 
the case and that the differences between rural and corridor transport and its 
effects are highly visible and as such could become a global issue. The World 
Trade Report of 2004 gives various transport costs, of which some of the 
salient examples are summarised in Table 54. 
 
Route US Dollar per kilometer 
Johannesburg to Maputo 1.40 
Durban to Lusaka 1.56 
Durban to Maseru 2.16 
 
Table 54: Transport costs for a ton of containerised cargo
331
 
 
The more rural the transport becomes, the more expensive it often is – with 
the obvious negative community effects associated with it. One of the specific 
outcomes of this work was the enablement, of a much more specific 
quantification of the wealth distribution challenges as it pertains to access of 
the second economy and tailor-made solutions are now possible. 
                                                 
330
 Demkes and Tavasszy, 2000, p. 16. 
331
 Panitchpakdi, 2004, p. 118. 
 224
11.5 Sustainability of the environment 
Whitelegg
332
 measures various environmental impacts, such as CO
2 
emissions (Figure 54) per ton-kilometer, compares modes and illustrates 
improvements over time. 
 
0
50
100
150
200
250
1987 2005 1987 2005 1987 2005
Rail Inland Waterways Road
 
Figure 54: CO
2 
emissions - gram per ton-kilometer 
 
His comprehensive study includes other obnoxious emissions as well (Figure 
55) and illustrates the poor position of road freight in terms of future 
sustainability issues. 
 
                                                 
332
 Whitelegg, 1994, p. 6. 
 225
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
1987 2005 1987 2005 1987 2005
Rail Inland Waterways Road
Methane TVOC Nitrogen oxide Carbon monoxide
 
 
Figure 55: Obnoxious gas emissions - grams per ton-kilometer 
 
It is, however, not the specifics and various arguments and counter 
arguments of the road vs. rail debate and the need for intermodal solutions 
that are important here, but rather the fact that measuring is taking place and 
certain shifts will be managed. From the time that Whitelegg has written his 
article the debate around these issues has heated up considerably. Many 
conventions in the first world are arising and it is becoming increasingly 
certain that Kyoto will finally be implemented and even strengthened with 
more stringent world protocols. It is difficult to imagine how this situation will 
be managed by a society that can’t even perform the simplest of measures, 
i.e. count ton-kilometer per mode. 
 
The World Business Council for Sustainable Development is taking this issue 
further by predicting the contribution of various modes to transport supply 
(Figure 56). 
 
 
 226
 
 
Figure 56: Future freight transport by mode
333
 
 
The growth prediction from the 15,000 billion ton-kilometers at the turn of the 
century to close to 50 000 billion ton-kilometers by the middle of the century is 
staggering in terms of the potential damage to society. Freight transport 
consumes 43% of all transport fuel and causes the most damage in terms of 
environmental impact.
334
 The specific impacts of various transport vehicles 
are also predicted and depicted in Figure 57. 
 
 
 
 
 
 
 
 
 
 
 
 
 
                                                 
333
 Stigson, 2004, p. 32 
334
 Massachusetts Institute of Technology, 2001, p. 6-2 
 227
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 57: Future freight transport contribution to CO
2 
emissions
335
 
 
Once again the specifics of the debate are not what is important here, but the 
foundation on which the debate and resultant decisions will be based, 
especially in the absence of measurement from an economy that contributes 
2% to this problem (but has far less than 1% of the world’s population and 
relative economic size).  
 
Through this work this challenge can be addressed in the future and South 
Africa can prepare itself for a more environmental friendly transport system. 
The country would be able to report the relationships between underlying 
transport options for freight and how these constructs impact on the 
environment. It will eventually be able to join the first world in a more pro-
active stance towards environmental control. 
 
                                                 
335
 Stigson, 2004, p. 37 
 228
11.6 Measurement for future direction – a new model for sustainable 
measurement 
McKinnon illustrates measurement in terms of sustainability quite clearly by 
outlining proposed policy measures and the reasons for these measures. 
These include: 
• Provision of additional infrastructure capacity (because of data that 
suggested that congestion-related delays cost the United Kingdom 
economy £1.2 billion); 
• Restraint on freight movement growth (because data suggested that 
ton-kilometers rose 146% in 39 years); 
• Modal shift (because of a rail modal market share drop from 30% to 
7%); 
• Further support for improving back-haul (because empty-haul has 
dropped from 33% to 27% and could go further); 
• Improved energy efficiency (because freight transport accounted for 5-
6% of all CO
2
 emissions); 
• And the reduction of externalities (also all dimensions measurement 
based).
336
 
The unavoidable fact is that measurement will be required to identify and 
measure these matters. 
 
Demkes and Tavasszy
337
 discuss the role of performance indicators for 
companies, industries and governments, and identify four goals: 
• To establish a holistic system view; 
• Feedback to improve the measured system; 
• Clarification of aim and goal (essentially alignment); 
• Development measurement for policy direction. 
 
All of the issues summarised above could be modelled back to the United 
Nations’ pillars for sustainable development, summarised in Chapter 2 (Figure 
1) and relating it back to South Africa’s future.  
                                                 
336
 McKinnon, 2003, pp. 133-149. 
337
 Demkes and Tavasszy, 2000, pp. 4-5. 
 229
The three pillars are ubiquitous for all nations and have recently been 
enhanced by Verny’s definition of sustainable development Figure 58. 
 
Economic 
growth
Spatial equity
Social progress
Environmental 
protection
SUSTAINABLE 
DEVELOPMENT
 
Figure 58: Verny's definition of sustainable development
338
 
 
Verny adds a spatial dimension to the mix and this is an essential 
improvement on the definition. He argues that: 
Distance is the heart of the problem. Stakes consist of preserving the 
quality of transport services and economic growth. The aim is to carry 
shorter distances or to carry differently. So our hypothesis supposes a 
link between productive logics, spatial organization and supply 
chain.
339
 
 
Logisticians should consider a spatial dimension and therefore a new model 
for sustainable development is proposed that depends on spatial physicality 
and community welfare (Figure 59). 
 
 
                                                 
338
 Verny, 2007, p. 115. 
339
 Verny, 2007, p. 115 
 230
R
e
lati
ve 
wealt
h
Spatial 
organisation
Hi
gh
 weal
th
 
eve
n
l
y
 
di
stri
bu
ted
Pover
t
y
 
wi
th
 a f
e
w 
ri
c
h
Confined Sprawling
B – Wealthy 
enclaves
D – Specialised 
industry
C - Siberia
A – Forgotten 
islands
 
Figure 59: Wealth and spatial organisation factors of sustainable development 
 
The groups are described as follows: 
• The A group could be described in terms of the forgotten Christmas 
Islanders on the Polynesian islands who lost touch with the global 
community and starved once this connection was lost; 
• The B group, such as rich enclaves in the West that live behind high walls 
or on islands, such as Japan, which need precious connections to the 
world, but have their own successful individualised cultures; 
• The C group, such as South Africa, with some wealth, but in the hands of 
a few, with widespread poverty over large geographical areas, but not on 
an island, in fact with porous borders, where more poverty is imported, but 
trade is also possible; 
• The D group, such as Europe with specialised and smooth flowing 
interchanges between nations. 
 
From this model the challenges in terms of the United Nations definition can 
be summarised (Figure 60). 
 
 
 231
Type Growth of the economy Distribution of growth 
Environmental 
protection 
A Aid Population control External agencies 
B Manage security of supply Fine tune social systems 
Manage natural 
resources carefully 
C 
Competitiveness of first 
economy 
Education, wealth 
distribution through job 
creation, stimulation of 
second economy 
Create environmental 
management systems 
and prepare for future 
management 
D 
Steady monetary and fiscal 
management and exchange 
agreements 
Social fine tuning, 
growing aged population 
burden 
Global protection 
agreement frameworks 
  
Figure 60: Sustainability challenges in terms of spatial and community position 
 
Transport, logistics and performance-management priorities can now be 
deduced from this. In A types food and aid must reach impoverished 
communities over poor infrastructure. In B types communities must develop 
their own effective transport systems to ensure security, often requiring a 
strong maritime culture with effective ports. In C types communities must 
stimulate growth nodes to make them globally competitive, but provide access 
to the “unconnected”. In the D types communities must manage the 
environmental damage of transport systems carefully. All the types require 
measurement and all of these will require management, from different 
perspectives and with different emphases, but eventually performance and 
planning measures will improve all systems and (as depicted in Figure 61) 
these measurements and measurement objectives will become more complex 
as new perspectives are added.       
 
 
 
 
 
 232
Type Measurement Objective 
A 
Understand basic measures of modal market 
share, transport efficiency and hinterland 
corridor blockages. Measure infrastructure 
development 
To measure Lead and lag for 
infrastructure development 
B 
Add port efficiencies and cargo transfer 
specifics 
To ensure effective linkages for 
security of supply 
C Add network typologies, intermodality 
To minimise transport costs in 
spatially challenged environments 
D Add sustainable measures 
To manage environmental 
charges 
 
Figure 61: Measurement systems and objectives 
 
This research makes it possible to manage these priorities and quantify the 
objectives, irrespective of where a country is on the spatial and community 
position matrix. 
 
11.7  Final thoughts on the macro perspective 
Davies contemplates reductionism in science and he refers to the “rampant 
reductionism” that was “the main thrust of Western scientific thinking over the 
last three centuries”.  He expands on this viewpoint by adding: 
But of course some problems (such as jigsaws) are only solved by putting 
them together – they are synthetic or ‘holistic’ in nature. The picture on a 
jigsaw, like the speckled newspaper image of a face, can only be 
perceived at a higher level of structure than the individual pieces – the 
whole is greater than the sum of its parts.
340
 
 
Davies continues to discuss the apparent paradox in the creation of order by 
comparing it to the second law of thermodynamics.  It appears as if the 
creation of order reverses entropy, but it does not. The overall “entropy 
balance sheet”, as he calls it, remains balanced and the order is paid for in 
                                                 
340
 Davies, 1990, p. 61. 
 233
the external environment of the open system.
341
 (Entropy in a closed system 
cannot be avoided, according to the second law). This is a ubiquitous fact and 
in all sciences it means that in the race for order, in the final reckoning, there 
will be winners and losers. In the specific field of this study a country like 
South Africa can either win or lose in the field of global competitiveness (with 
global competitiveness supplied inter alia by the transport system), it could 
renew natural resources or deplete these resources (through incorrect energy 
allocations – also inter alia through transport and spatial design) and it could 
uplift society or destroy the societal fabric of the country (through denying 
access to the second economy). These issues are issues of order that must, 
by definition, be measured and managed. 
 
The “order” decisions that South Africa make will be the culmination of the 
national will translated into policy and implementation. This research can 
provide inputs into the understanding of the freight transport dimension’s 
capability to improve this order. 
 
11.8 Conclusion 
In this research the macro-economic imperatives of logistics and the resultant 
burden of measurement for macro-logistics were highlighted. An in-depth 
analysis of the history of measurement confirmed that the problems 
associated with poor information was often understood, but never really 
addressed. A few attempts were made, with varying degrees of success, but 
with no permanent solution. Three interactive models were proposed and the 
success of these models for measurement on a continuous basis and 
examples of contribution to lead and lag policy and infrastructure 
management systems were illustrated. 
 
It is possible to use these models to begin to answer the questions of not only 
the past, but also the future, i.e. what would South Africa need in terms of 
infrastructure, how will the performance of this infrastructure be measured for 
                                                 
341
 Davies, 1990, p. 65. 
 234
 235
renewed planning and how will policy makers ensure that future growth, and 
societal and sustainability objectives are met. 
 
The interaction amongst communities, as described in Chapter 2, in the 
emerging new world order will place immense new pressures on the 
management of South Africa’s transport system, which is not even on the 
agenda yet. Eventually countries will be graded on the performance of their 
domestic transport system (that will have to be measured) and an inability to 
demonstrate performance against specific sustainability criteria might 
jeopardise the country’s foreign trade system in many respects. This study 
proposes immediate action to rectify this situation. 
 
 
 
 
 
 
 
List of sources 
Alford, L.P. 1941. Principles of Industrial Management, The Ronald Press 
Company, New York. 
Alford, L.P. and Bangs, J.R. 1946. Production Handbook, The Ronald Press 
Company, New York. 
Alfsen, K.H. and Thorvald, M. 2005. An international framework for 
constructing national indicators for policies to enhance sustainable 
development, Background paper prepared for the UN Expert Groups Meeting 
on Indicators of Sustainable Development in New York, 13-15 December, 
2005. 
Anselm, O. 2002. Comparable cost calculation for infrastructure of road and 
rail, Institut fur Wirtschaftpolitik und Wirtschaftsforschung (IWW), Universitat 
Karlsruhe, Karlsruhe. 
Australian Bureau of Statistics. 2001. Year Book Australia: Australia’s motor 
vehicle fleet since the 1920s, Available on the web, www.abs.gov.au. 
Australian Government, Department of Transport and Regional Services. 
2004. The Commonwealth’s Transport Directions, [online], Available from 
www.dotars.gov.au. 
Australian Government, Department of Transport and Regional Services, 
Bureau of Transport and Regional Economics. 2004. Transport Statistics: 
Transport and storage contribution to GDP, [online], Available from 
www.dotars.gov.au, in Year Book, Australia, Canberra. 
Ayers, J. 2004. Rooting out Supply-Chain Costs, [online], Available from 
www.optimizemag.com.  
Baily, P.J.H. 1969. Purchasing and Supply Management, second edition, 
Chapman and Hall, London. 
Ballou, R. H. 2004. Business Logistics/Supply Chain Management: Planning, 
Organizing, and Controlling the Supply Chain, 5
th
 edition, Pearson Education 
International, Upper Saddle River, New Jersey. 
Baluch, I. 2005. Transport Logistics, Past, Present and Predictions, Winning 
Books, Dubai. 
Banfield, E. 1999. Harnessing Value in the Supply Chain, John Wiley & Sons, 
Inc., New York. 
Baudin, M. 2004. Lean Logistics, the Nuts and Bolts of Delivering Materials 
and Goods, Productivity Press, New York. 
Botes, F.J., Jacobs, C.G. and Pienaar, W.J. 2006. A model to calculate the 
cost of logistics at a macro level: a case study of South Africa, in Southern 
African Business Review, Volume 10, Number 3, December 2006, pp 1-18. 
Bowersox, D.J. and Closs, D.J. 1996. Logistical Management, The Integrated 
Supply Chain Process, The McGraw-Hill Companies, Inc. New York. 
 236
Brockman, J. 2003. Introduction: The new humanists, in Brockman, J. ed. 
Science at the edge, Weidenfield and Nicolson, London. 
Burman, J. 1984. Early railways at the Cape, Human & Rosseau, Cape Town. 
Caid, N. 2003. Analysis of the links between transport and economic growth, 
OECD, Paris. 
Cape Times, 1 July, 2004. Quintal, A. Mbeki commits his 2
nd
 term to delivery, 
saying “we are a government of the left”. Cape Town. 
Central Statistical Services. 1993. Standard Industrial Classification of all 
Economic Activities, Central Statistical Services, Pretoria. 
Chacholiades, M. 1990. International Economics, International ed., McGraw-
Hill Publishing Company, New York 
Chase, R.B. and Aquilano, N.J. 1981. Production and operations 
management: A Life Cycle Approach, Richard D. Irwin, Inc., Humewood, 
Illinois. 
Christopher, M. 1998. Logistics and Supply Management: Strategies for 
Reducing Cost and Improving Service, Prentice-Hall, London. 
Coetzee, D.J. 1940. Spoorwegontwikkeling in die Suid-Afrikaanse Republiek 
(1872-1899), Nasionale Pers, Beperk, Kaapstad. 
Cohen, S. and Roussel, J. 2004. Strategic Supply Chain Management: The 5 
Disciplines for Top Performance, McGraw-Hill, New York. 
Cohen, J. and Stewart, I. 2000,  The collapse of chaos: Discovering Simplicity 
in a Complex World, Penguin Books, London. 
Communication and Resource Centre Administrator, European Commission, 
http://circa.europa.eu/irc/dsis/bmethods/info/data/new/embs/EconomicClassifi
cations.pdf, accessed 24 June 2007. 
Conradie, Eric. 2007. Interview with transport historian conducted by 
candidate,  South African Railway Museum, Johannesburg. 
Cook, T.M. and Russel, R.A. 1984. Contemporary Operations Management, 
text and cases, 2
nd
 ed. Prentice-Hall, Inc. Englewood Cliffs, New Jersey. 
Cooke, James, A. 2004. The two faces of Globalization, [online], Available 
from http://www.manufacturing.net.  
Coyle, J.J., Bardi, E.J. and Cavinato, J.L. 1986. Transportation, 2
nd
 ed., West 
Publishing Company, St. Paul. 
Darwin, C. 1901. The Descent of man and selection in relation to sex, John 
Murray, London. 
Davies, P. 1990. God & the new physics, Penguin Books, London. 
De Jong, G., Gunn, H. and Walker, W. 2004 National and International Freight 
Transport Models: An Overview and Ideas for Future Development, in 
Transport Reviews, Vol. 24, No. 1, January 2004. 
Delaney, R.V. 1987. The Disunited States: A Country in Search of an Efficient 
Transportation Policy, Cato Policy Analysis No. 84.  
 237
Delaney, R.V. and Wilson, R. 2003. 14th Annual “State of Logistics Report, 
ProLogis and Cass Information Systems, Inc. Washington. 
Demkes, R. and Tavasszy, L.A. 2000. Benchmarking infrastructure and 
logistics services across Europe, Asia-Pacific and North America, Paper 
presented at IMRL 2000 Third International Meeting for research in Logistics, 
Trois-Rivieres, 9-11 May 2000. 
Dodig-Cmkovic, G. 2001. Theory of Science, MRTC report, Malardalen Real-
Time Research Centre, Malardalen University, Vasteras. 
Dong, Y., Carter, C.R. and Dresner, M.E. 2001. JIT purchasing and 
performance: an exploratory analysis of buyer and supplier perspectives, in 
Journal of Operations Management, Elsevier. 
Dooyeweerd, H. 1969. A new Critique of Theoretical Thought, The 
Presbyterian and Reformed Publishing Company, Amsterdam. 
Erero, J.L. and Van Heerden, J. 2005. The role of freight rates in the race 
between road and rail in South Africa: A computable general equilibrium 
analysis, Proceedings of Biennial conference of the Economic Society of 
South Africa: Development perspectives: Is Africa different?, 7-9 September, 
2005 , Durban. 
Fawcett, P., McLeish, R. and Ogden, I. 1992. Logistics Management, 
Prentice-Hall, Harlow. 
Financial Mail, 2 July 2004. Transnet: Changing Culture Is Painful. 
Johannesburg. 
Frankel. S.H. 1928. The Railway Policy of South Africa: An analysis of the 
effects of railway rates, finance and management on the economic 
development of the Union, Hortors Limited, Johannesburg.  
Fredendall, L.D. and Hill, E. 2001. Basics of Supply Chain Management, The 
St. Lucie Press, Bocka Raton, Florida, USA. 
Gale, B.T. 1994. Managing customer value: Creating quality and service that 
customers can see, The Free Press, A Division of Macmillan, Inc., New York. 
Gattorna, J.L. and Walters, D.W. 1996. Managing the Supply Chain: A 
strategic Perspective, Macmillan Business, Houndmills, Basingstoke, 
Hampshire. 
Gell-Mann, M. 1994. The quark and the jaguar: Adventures in the simple and 
the complex, W.H. Freeman and Company, New York.  
Gielen, D. 2004. Transportation in the world and the EU: A bird’s eye view, 
Proceedings 7
th
 EU Hitachi science and technology forum. 
Gollwitzer, M. and Karl, R. 1998. Logistik-Controlling: 
Wirkungszusammenhange – Leistung, Kosten, Durchlaufzeiten und 
Bestande, Wirtscaftsverlag Langen Muller/Herbig, Munchen. 
Goosen, J. 1979. Line-haul road transport terminal costs, South African 
Council for Scientific and Industrial Research, National Institute for Transport 
and Road Research, Pretoria. 
 238
Grant, D.B., Lambert. D.M., Stock, J.R. and Ellram, L.M. 2006. Fundamentals 
of Logistics Management: European edition, McGraw-Hill, London. 
Gribbin, J. 2002. Science: A History: 1543-2001, Penguin Books, London. 
Hamilton, C.C. 1983. Die Suid-Afrikaanse goederevervoermark, Nasionale 
Instituut vir Vervoer- en Padnavorsing, WNNR, Pretoria. 
Hamilton, C.C. 1986. The estimated market share of public and ancillary road 
carriers in a specifically defined South African goods transport market, 
National Institute for Transport and Road Research, CSIR, Pretoria. 
Handfield, R.B. and Nichols, E.L. Jr. 2002. Supply Chain Redesign: 
Transforming Supply Chains into Integrated Value Systems, Prentice-Hall, 
Upper Saddle River, New Jersey. 
Hardin, G. The Tragedy of the Commons, in Science, New Series, American 
Association for the Advancement of Science, Vol. 162, No. 3859, pp..1243-
1248. 
Harris, R.G. 1977. Economies of Traffic Densities in the Rail Freight Industry, 
in The Bell Journal of Economics, Vol. 8, No. 2. (Autumn, 1977) pp. 556-564. 
Havenga, J.H. 2005. Global logistics trends, in CSIR centre for Logistics and 
Decision Support, The first Annual State of Logistics Survey for South Africa, 
CSIR, Pretoria 
Havenga, J.H. and Hobbs, I.E. 2004. A Practical Guide to Strategy: Making 
strategic thinking, development and implementation accessible, Sun Press, 
Stellenbosch. 
Hayes, H.G. 1928. Our Economic System, Vol. 1, Henry Holt and Company, 
New York. 
Healey, M. and Ilbery, B.W. 1990. Location and Change: perspectives on 
economic geography, Oxford University Press, Oxford. 
Heskett. J.L., Ivie, R.M. and Glaskowsky, N.A. Jr. 1964. Business Logistics: 
Management of Physical Supply and Distribution, The Roanald Press 
Company, New York. 
Heydenrich, D.H. 1965. Die geskiedenis van die spoorweë in die Kaapkolonie 
tot 1885, M.A. Verhandeling, Universiteit van Stellenbosch, Stellenbosch. 
Hesse, M. and Rodrigue, J. 2003. The transport geography of logistics and 
freight distribution, in Journal of Transport Geography, Elsevier, Berlin. 
Hofstadter, D.R. 2000. Godel, Escher, Bach: an Eternal Golden Braid, 
Penguin Books, London. 
Jones, T. 1999. The South African transport sector in the 1970s, in South 
African Journal of Economic History, Volume 14: September 1999, pp. 162-
194. 
Jordan, P.Z. 2007.  Key note address at the National Film and Video 
Foundation Indaba, Midrand (Edited Précis), available from 
http://www.nfvf.co.za/A%20REPORT%20ON%20THE%20NFVF%20FILM%2
0INDABA%202005.pdf.  
 
 239
Jorgenson, A.A. 1999. Intermodal transport options for international and 
domestic freight requirements in southern Africa, Proceedings of Council of 
Logistics Management Conference, 14-15 April, Johannesburg. 
Kaydos, W. 1999. Operational Performance Measurement: Increasing Total 
Productivity, St. Lucie Press, Boca Raton. 
Keebler, J.S. and Durtche, D.A. 1999. Keeping Score: Measuring the 
Business Value of Logistics in the Supply Chain, Council of Logistics 
Management, Oak Brook, Illinois.  
Kennedy, T.L. 1982. Costs of road and rail transport, National Institute for 
Transport and Road Research, CSIR, Pretoria. 
Kennedy, T.L. 1984. Road and rail transport and distribution costs – A modal 
choice model, National Institute for Transport and Road Research, CSIR, 
Pretoria. 
Krygsman, S. 2006. Project notes for Transnet’s commodity flow project, 
Centre for Supply Chain Management, Department of Logistics, University of 
Stellenbosch, Stellenbosch. 
Kuhn, S.H. 1983. Overview and trends of research and development in roads 
and transport at the NITRR, in The Civil Engineer in South Africa, Vol. 25, No. 
4, April 1983, pp. 193-204. 
Kuo, K. 1987. The effect of freight transportation on foreign trade 
development, in Van der Walt, N.A. and Van Pletzen, L. Modern trends in 
industrilization and the third world, Rand Afrikaanse University and the 
Industrial Development Corporation of South Africa Limited, Sandton.  
Lakshmanan, T.R. and Anderson, W.P. 2002. Transportation infrastructure, 
Freight Services and Economic Growth: A Synopsis of a White Paper, The 
U.S. Department of transportation Federal Highway Administration, 
Washington. 
Laubscher, P. 2006. An assessment of the macro-economic projections 
underlying the Transnet freight flow model – 2006 to 2025. A confidential 
study prepared for the Department Logistics,  University of Stellenbosch. 
Locklin, D.P. 1935. Economics of Transportation, Business Publications, 
Chicago. 
Logistics News. March 2006. Supply chain foresight shows SA supply chain is 
a time bomb.  
Louw, J.J. 2006. Advanced supply chain processes and decision support 
systems for large-scale petrochemical companies, PhD dissertation, 
University of Stellenbosch, Stellenbosch. 
MacroSys Research and Technology. 2005. Logistics Costs and U.S. Gross 
Domestic Product, MacroSys Research and Technology, Washington. 
Martin, J. 2007. The Meaning of the 21
st
 Century: A vital blueprint for ensuring 
our future, Eden Project Books, Transworld Publishers, London.  
Massachusetts Institute of Technology. 2001. Mobility 2001: Dedicated to 
making a difference, Sustainable Mobility Working Group of the World 
Business Council for Sustainable Development, Geneva. 
 240
McCarthy, C.L. 1988. Monetêre Ekonomie, ‘n Suid-Afrikaanse Perspektief, 
HAUM Opvoedkundige Uitgewery, Pretoria. 
McCormack, D. 2004. Upgrading MSE clusters – Theoretical frameworks and 
practical approaches in Innovation systems and innovative clusters in Africa, 
paper presented at the Regional Conference on Innovation Systems and 
Innovative Clusters in Africa, Bagamoyo, 18-20 February.  
McDonald, S. and Punt, C. 2005. General equilibrium modelling in South 
Africa: What the future holds, in Agrekon, Vol. 44, No. 1 (March 2005), pp. 60-
98. 
McKinnon, A. 2003. Sustainable Freight Distribution, in Hine, J. and Preston, 
J. ed. Integrated Futures and Transport Choices: UK Transport Policy Beyond 
1998 White Paper and Transport Acts, Ashgate, Aldershot, Hants. 
Mentzer, J.T., Myers, M.B. and Cheung, M.S. 2004. Global market 
segmentation for logistics services In: Industrial Marketing Management, pp. 
1-18. 
Murphy, P.R. Jr. and Wood, D.F. 2004. Contemporary Logistics, 8
th
 edition, 
Pearson Prentice-Hall, Upper Saddle River, New Jersey. 
Mikros Traffic Monitoring (Pty) Ltd. 2005. SANRAL traffic counting yearbook 
2004, South African National Roads Agency Limited, Pretoria. 
Mikros Traffic Monitoring (Pty) Ltd. 2006. SANRAL traffic counting yearbook 
2005, South African National Roads Agency Limited, Pretoria. 
Mitchell, M.  2004. The history of transport infrastructure in South Africa 
during the past 50 years, Unpublished report, Development Bank of South 
Africa, Pretoria. 
Mohr, P. and Fourie, L. 1996. Ekonomie vir Suid-Afrikaanse studente, J.L. van 
Schaik Akademies, Pretoria. 
Morris, R. 2005. Rapid review of golf course and polo field developments, 
Department of Environmental Affairs & Development Planning, Cape Town. 
Moving South Africa, 1998. A Transport Strategy for 2020, Report and 
Strategy Recommendations, Department of Transport, Pretoria. 
Nadvi, K. 1999. Collective efficiency and collective failure: the response of the 
Sialkot Surgical Instruments Cluster to Global Quality pressures, World 
Development, Vol. 27/9. 
National economic development and labour council (Nedlac) 1996, Report on 
activities, 
http://www.nedlac.org.za/top.asp?inc=docs/reports/annual/1996/sum_activi.ht
ml 
Nelldal, B.; Troche, G. and  Wajsman, J 2000. Railway development for future 
transportation markets, Swedish National Rail Administration (“Banverket”), 
Swedish State Railways (SJ),  KTH, Adtranz, the Swedish Transport and 
Communication Board (KFB), Stocholm. 
Panitchpakdi, S. 2004. World Trade Report 2004: Exploring the linkage 
between the domestic policy environment and international trade, World 
Trade Organisation, Geneva. 
 241
Parker, A. 1999. Managing Spoornet’s transition ‘will remain difficult for some 
time, in S.A. Transport, December, 1999. 
Pienaar, W.J. 2005. Introduction to business logistics, Chapter 1, in Vogt, J.J., 
Pienaar, W.J. and De Wit, P.W.C. 2005. Business Logistics Management, 
Oxford University Press, Cape Town. 
Porter, M.E. 1991. On competition and Strategy, Harvard Business Review 
Paperback, Boston. 
Porter, M.E. 1990. The competitive advantage of nations, The Macmillan 
Press Ltd., London. 
Pretorius, A.M.J. 1990.  Inter-regional transport and energy modelling in 
South Africa: A brief review of past research and future directions, Division of 
Roads and Transport Technology, CSIR, Pretoria. 
Pretorius, W.P. 1991, Freight transport data bank report, 2nd Edition, 
Research Unit for Transport Economic and Physical Distribution Studies, 
Rand Afrikaans University, Johannesburg. 
Raha, N. and Williams, I. 2002. DfT Integrated Transport and Economic 
Appraisal: Review of Freight Modelling, University of Westminster, London. 
RAND Europe, 2001. Study on Ideas on a new National Freight Model 
System for Sweden: Report to the SAMGODS Group, RAND Europe and 
Transek AB, Leiden. 
Ravin, M.O. and Mazzenga, E. 2004. International business cycles: The 
qualitative role of transportation costs, in Journal of International Money and 
Finance 23 (2004), Elsevier. 
Rodrigue, J.P. 2005. Transport geography on the web, Hofstra University, 
Department of Economics and Geography, Hempstead, New York. 
Rodrigues, A.M., Bowersox, D.J. and Calantone, R.J. 2005. Estimation of 
Global and National Logistics Expenditures: 2002 Data Update, in Journal of 
Business Logistics, Vol. 26, No. 2, 2005, pp. 1-16. 
Russel, S.H. 1994. The coming of age of Business Logistics, in Journal of 
Education for Business, Nov/Dec 1994, Vol. 70, Issue 2.   
Samuelson, P.A. and Nordhaus, W.D. 1989. Economics, McGraw-Hill Book 
Company, New York. 
Scharry, P.B. and Skott-Larsen, T. 2001. Managing the Global Supply Chain, 
2
nd
 ed., Copenhagen Business School Press, Copenhagen. 
Scmitz, H. 1992. On the clustering of small firms, In: IDS Bulletin 23(3), pp. 
64-69. 
Schneidewind, U. 2003. “Symbols” and “Substances” – An Interpretive Supply 
Chain Management Perspective. In: Seuring, S., Muller, M., Goldbach, M. et. 
al. 2003. Strategy and Organization in Supply Chains, Physica-Verlag, 
Heidelberg.  Pages. 
Smith, R.A.F. 1980. A Challenge to the Civil Engineer, in the proceedings of 
Transportation options for the future symposium, 29 October 1980, CSIR, 
Pretoria. 
 242
Smith, W.J.J.  1973. A Quantitative Study of Road Hauliers and Ancillary 
Road Transport Users in the Republic of South Africa, Rand Afrikaans 
University, Johannesburg. 
Smits, J.P. 2006. Economic Growth and Structural Change in Sub Saharan 
Africa during the Twentieth Century: New Empirical Evidence, paper 
presented at Session 41: “International differences in Economic Welfare: A 
Long-Run perspective” at the International Economic History Association 
Conference, August 2006, Helsinki. 
South African Railways and Harbours. 1927. Report of the General Manager 
of Railways and Harbours, South African Railways and Harbours, 
Johannesburg. 
South African Railways and Harbours. 1959. Report of the General Manager 
of Railways and Harbours, South African Railways and Harbours, 
Johannesburg. 
South African Railways and Harbours. 1960. A Century of Transport: A 
Record of the Achievement of the Ministry of Transport of the Union of South 
Africa, Da Gama Publications, Johannesburg. 
Spoornet Media Statement, 13 May 2004. Mokgatle, D. Spoornet is being 
reborn, Spoornet, Johannesburg. 
Spriegel, W.R. and Lansburgh, R.H. 1947. Industrial Management, 5
th
 ed., 
John Wiley & Sons, inc., New York. 
SRI International, 2004. Global impacts of Fedex in the New Economy, 
[online], Available from www.sri.com, Menlo Park. 
Star Business Report, 13 July 2004. Cokayne, R. Poor transport costs 
business, says Radebe, Johannesburg. 
Statistics South Africa, 2003. Land freight transport, Statistical release P7142, 
September 2003, Pretoria. 
Stigson, B. 2004. Mobility 2030: Meeting the challenges of sustainability, 
World Business Council for Sustainable Development, Geneva. 
Stock, J.R. and Lambert, D.M. 2001.  Strategic Logistics Management, 4
th
 ed.  
McGraw-Hill, Singapore. 
Strauss, D.F.M. 1978. Inleiding tot die Kosmologie, Sacum Beperk, 
Bloemfontein. 
Sustainable Development Working Group. 2001. Mobility 2001, World Council 
for Sustainable Development, Geneva. 
Sutherland, J. 2003. Logistics and Supply Chain Predictions for the 
Transportation Industry in 2003, [online], Available from 
www.eyefortransport.com, London.  
Terreblanche, S.J. 1988.  Geskiedenis van die Westerse Ekonomie, Die 
verhaal vanaf die Griekse statstate tot by Demokratiese Kapitalisme, 
Academia, Pretoria.  
Terreblanche, S.J. 1994. ‘n Geskiedenis van opeenvolgende ekonomiese 
stelsels, Southern Boekuitgewers, Halfweghuis. 
 243
The Precidency of South Africa. 2007. Development indicators: Mid term 
review, Government Communications, 
http://www.info.gov.za/speeches/2007/07062208151001.htm, 21 June 2007. 
TMT Projects, 1995. Heavy Vehicle Overloading in South Africa, the Public 
Affairs division of the Automobile Association of South Africa, Johannesburg. 
Transport Analysis Unit, Department of Transport. 2006. Intermodal cargo 
transfer policy: Discussion document, Transport Policy Analysis, Department 
of Transport, Pretoria. 
Traut, R.E. 1967. Some United Nations Activities in Transport, in proceedings 
of Seventh Annual Conference – Containerization And Packaging 24 January, 
1967, , Bulk Packaging and Containerization Institute, Inc., New York. 
Van der Meulen, D. 2007. Leveraging the competitive strengths of railways in 
South Africa, Proceedings of transport infrastructure and development 
conference, 30-31 May, 2007, Sandton. 
Van der Veer, G.D. 1982. The role of rail transport in the total transport scene: 
Freight transport, Proceedings of the Annual Transportation Conference, 
Pretoria. 
Van Rensburg, R.A.J. 1996. The history of the  rail transport regulatory 
environment in South Africa, in Spoornet project: The transport function in 
South Africa, The Department of Transport Economics of the University of 
Stellenbosch, Stellenbosch. 
Verburgh, C. 1958.  Road transport of goods in South Africa, Bureau for 
Economic Research, Faculty of Commerce, University of Stellenbosch, 
Stellenbosch. 
Verburgh, C. 1968, The Relationship between the Transport of Goods by Rail 
and Economic Development in South Africa, Transport Research Centre, 
University of Stellenbosch, Stellenbosch. 
Verny, J. 2007. The importance of decoupling between freight transport and 
economic growth, in EJTIR, 7, No. 2 (2007). Pages. 
Voortman, C. 2004. Global Logistics Management, Juta Academic, Cape 
Town. 
Whitelegg, J. 1994. Freight Transport, Logistics and Sustainable 
Development, A report for WWF, Eco-Logica Lmtd., White Cross, Lancaster. 
Wilson, R. 2007. 18
th
 Annual State of Logistics Report: The new face of 
logistics, Council for Supply Chain Management Professionals, Washington. 
Wisner, J.D., Keong Leong, G. and Tan, K. 2005. Principles of Supply Chain 
Management: A Balanced Approach, Thomson South-western, Mason, Ohio. 
World Business Council for Sustainable Development. 2006. Geneva. 
www.wbcsd.org/Plugins/DocSearch/details.asp?ObjectId=MTc4NTE - 66k - 
World Trade Organization, 2004. World Trade Report: 2004, World Trade 
Organization. 
 244
 245
Wynne-Roberts, C.R. 1956. Industrial Management, Pakistan Institute of 
Personnel Administration, Karachi. 
Zeng, D.Z. 2006. Knowledge, Technology and Cluster-based Growth in Africa 
– findings from Eleven Case Studies of Enterprize Clusters in Africa, 
Knowledge for Development (K4D) Program, World Bank Institute, World 
Bank.