Manifestations of inequality in three developing countries : an investigation of differential education and labour market outcomes

Abstract

ENGLISH SUMMARY : Economic inequality is a pervasive phenomenon that has long been a feature of both developed and developing countries. Such inequality is particularly problematic in developing countries because of its potential to exacerbate already established and yet detrimental features that characterise most developing countries, such as high unemployment, a large informal sector, and high population growth rates. This thesis investigates how inequality manifests itself through differential student outcomes and also attempts to examine whether or not technological growth destroys jobs and worsens wage inequality. Chapter 2 investigates the relationship between student performance and socio-economic status (SES) among grade six students in Malawi and Namibia. Malawi was specially included in this cross-country comparison because it is likely that the underlying mechanisms that govern the typical/expected SES-performance relationship do not hold in countries as poor as Malawi as they would in more advanced developing countries like Namibia. Using OLS regressions and hierarchical (multilevel) models, the results show an approximately flat socio-economic gradient for student education performance for Malawi, for both the full sample and the reduced samples (urban and rural). In Namibia, in contrast, SES appears to be correlated with student performance. However, this is primarily driven by students who live in urban areas, whereas, like Malawi, rural Namibia also has an approximately flat socio-economic gradient. Chapter 3 builds on this by taking special interest in research in low-income countries like Malawi and the challenge that arises when, in the absence of income/expenditure information, one has to rely on an asset index to distinguish among individuals of comparable SES levels. This follows on discussions in the literature that have well articulated the difficulty asset-based measures have of doing so especially among various shades of poor individuals, that is, differentiating the poor from the very poor: a feature which a measure used in very poor countries like Malawi should have. Chapter 3 explores the use of finite mixture modelling as an alternative approach to achieving this goal. The findings suggest that using this approach makes it possible to distinguish between individuals’ relative SES level in a meaningful way. Lastly, Chapter 4 is primarily interested in examining if technological growth in South Africa contributed to exacerbation of wage inequality and job loss during the period from 1997 to 2015. This analysis is done through the lens of a routine-biased technological change framework whose main hypothesis is that recent technological advancements are biased towards replacing labour in routine tasks. This chapter presents findings from descriptive analysis, OLS regressions, as well as a non-linear systems estimator applied to a normalised CES production function. The results show both descriptive and empirical evidence of a hollowing out of middle-skilled work (done by workers whose occupations typically involve a high share of routine tasks). Further, these findings are differentiated by gender and race.

AFRIKAANSE OPSOMMING : Ekonomiese ongelykheid is ‘n algemene verskynsel in beide ontwikkelde en ontwikkelende lande. Sulke ongelykheid is veral problematies in ontwikkelende weens sy potensiaal om reeds bestaande negatiewe kenmerke van die meeste sulke lande te vererger, soos hoe werkloosheid, ‘n groot informele sektor en snel bevolkingsaanwas. Hierdie proefskrif ondersoek hoe ongelykheid in verskille in leerling-uitkomste manifesteer, en bestudeer ook in watter mate tegnologiese groei werkgeleenthede uitwis en loonongelykheid vererger. Hoofstuk 2 ondersoek die verhouding tussen leerling-uitkomste en sosio-ekonomiese status (SES) onder graad 6 leerlinge in Malawi en Namibia. Malawi is veral by hierdie vergelyking tussen twee ontwikkelende lande ingesluit, omdat die meganismes wat gewoonlik die verband tussen SES en leerling-prestasie onderlˆe moontlik nie so goed in ‘n land so arm soos Malawi sal geld as in ‘n land wat verder ontwikkel is soos Namibia nie. Met gebruik van gewone kleinste-kwadrate regressies en hierargiese (multivlak) modelle word bevind dat Malawi se sosio-ekonomiese gradient byna heeltemal plat is, beide vir die volle steekproef en vir aparte landelike en stedelike groeperings. In Namibia, daarenteen, styg leerling-uitkomste o¨enskynlik met SES, maar dit word veral deur stedelike leerlinge verklaar, terwyl die sosio-ekonomiese gradi¨ent in landelike gebiede plat is, netsoos in Malawi. Hoofstuk 3 bou hierop voort deur spesiale aandag te gee aan navorsing in lae-inkomste lande soos Malawi en die uitdaging wat ontstaan wanneer navorsers weens die afwesigheid van inkomste of bestedingsdata verplig is om ‘n bate-indeks te gebruik om tussen leerlinge met soortgelyke SES te onderskei. Dit volg op besprekings in die vakliteratuur wat uiteensit hoe moeilik bate-indekse veral tussen arm en baie arm individue kan onderskei – ‘n kenmerk wat erg nodig is in baie arm lande soos Malawi. Hoofstuk 3 gebruik sogenaamde “finite mixture”-modelle as alternatiewe metode om hierdie doel te bereik. Die resultate dui daarop dat dit ‘n sinvolle benadering bied om tussen verskillende individue se relatiewe SES te onderskei. Ten slotte is die vraag in Hoofstuk 4 in watter mate tegnologiese verandering in Suid-Afrika van 1997 tot 2015 tot loonongelykheid en verlies aan werkgeleenthede bygedra het. Hierdie analise gebruik ‘n roetine-gebaseerde tegnologiese veranderingsraamwerk. Die belangrikste hipotese van hierdie raamwerk is dat onlangse tegnologiese verandering geneig is om arbeid te vervang vir roetine take. Hierdie hoofstuk bied resultate van beskrywende analise, OLS-regressies, sowel as ‘n nie-linieere stelsel-beramer toegepas op ‘n genormaliseerde CES produksiefunksie. Die resultate toon beide beskrywende en empiriese getuienis van ‘n uithol-proses van beroepe met ‘n groot mate van roetine-take, meestal beroepe in die middel van die geskoolheidsverdeling. Die resultate word ook volgens geslag en ras gedifferensieer.