Portfolio asset selection through the use of modified moving averages and steepest gradient techniques

Petzer, Greydon E. (2001-03)

Thesis (MBA)--Stellenbosch University, 2001.

Thesis

ENGLISH ABSTRACT: Many tracker funds exist in the South African market in which investors can invest their money. Growing in:popularity is the index funds that, instead of investing in individual shares invest into funds that track and guarantee returns related to specific indices. One such fund is the All Share Index 40 (ALSI40) Tracker Fund. The index is equity based to reflect the performance of the ordinary South African share market. Companies selected for inclusion in the ALSI40 Index are generally larger companies of sound financial standing having widely traded and marketable securities. As the ALSI40 is therefore a reflection of the total market as a whole, investing into such a Tracker Fund would only gain average returns over the long run. A modeled developed through using a twenty-day moving average to signal buy and sell periods, coupled with individual share gradients likened to future share growth potential is evaluated to determine if such a model would gain returns above the Tracker Fund and therefore gain returns above the ALSSI40 Index. The study project is wholly based on technical analysis, specific to the shares that constitute the ALSI40 Index. Through the selection of these shares fundamental analysis is taken care of and emphasis is placed on a technical trading technique developed. The technique is based on a two-stage model. Firstly, switch points are determined that indicate buy and sell signals through applying a 20-day moving average to the daily closing price of the selected shares. When the moving average is tracking below the closing price a buy signal is generated, and when the gradient of the moving average turns negative a sell signal is generated. Using these 'switch points', the second stage of the model is entered into through the allocation of weights to individual shares that conform to the buy selection criteria. The weights are determined using the gradient of the linear regression analysis equation. The gradient is synonymous with the growth potential of the share at the time a switching takes place. Through the use of the above model it was found that returns well above the holding period return of the ALSI40 Index are achievable. Evaluation of the model on a calendar year basis yielded a 113% over and above the return yielded by the ALSI40 Index for 1999. Similarly, positive returns were yielded for the 2000 calendar year and thus entrenching the trading technique as being successful. The major downside to the model is the number of switches dictated through strict adherence to the developed model. For example, 110 switches were necessary during 1999 to achieve the 113% over and above yield. Assuming switching fees of 1% per switch, margins that beat the holding period return of the ALSI40 would rapidly be eroded away. Although successful in achieving the aim of beating returns of the ALSI40, the model and computer code developed is robust and primitive in form. Numerous options exist to optimise the model, and thereby the potential to generate even greater returns. Optimisation would include a better 'gradient' function and procedure to reduce switching costs. The switching technique used is the most efficient obtainable from a single moving average.

AFRIKAANSE OPSOMMING: Daar bestaan vele 'tracker" fondse in Suid-Afrika waarin beleggers hulle geld kan belê. Die mees gewilde van hierdie fondse is die indeksfondse wat groei verwant aan spesifieke indekse waarborg, en dus direkte belegging in individuele aandele vervang. Een so 'n indeksfonds is die "All Share Index 40 (ALSI40) Indeks "Tracker" Fonds. Hierdie indeks is gebaseer op aandeelhouersbelang (gewone aandele) om die opbrengs van die gewone Suid-Afrikaanse aandelemark te reflekteer. Die gekose maatskappye vir insluiting in die ALSI40 indeks is gewoonlik die groter maatskappye met 'n gesonde finansiële status, met verhandelbare en verkoopbare sekuriteite. Aangesien die ALSI40 dus 'n weerspieëling is van die totale Suid-Afrikaanse aandelemark as 'n geheel, sal belegging in 'n 'tracker" fonds slegs gemiddelde groei oor die langtermyn lewer. 'n Model ontwikkel deur middel van die gebruik van 'n 20 dag bewegende gemiddelde om koop en verkoop tydstippe aan te dui, gekoppel aan individuele aandeelhellings te einde toekomstige aandeelgroei-potensiaalaan te toon, word beoordeel om te bepaal of so 'n model groei bo the 'tracker" fonds sal lewer, en dus ook groei bo die ALSI40 Indeks. The studieprojek is in geheel gebaseer op tegniese analise, met spesifieke verwysing na die aandele wat in die ALSI40 Indeks bestaan. Deur die seleksie van hierdie aandele is die fundamentele analise by implikasie reeds aangespreek en word die ontwikkelde tegniese verhandelingstegniek beklemtoon. Die tegniek is gebaseer op 'n twee-fase model. Eerstens, herbelegginspunte word bepaal, welke punte koop- en verkoopseine aandui deur middel van die toepassing van 'n 20 dag bewegende gemiddelde op die daaglikse sluitingsprys van die aandele. Wanneer die bewegende gemiddelde benede die sluitingsprys beweeg, word 'n koopsein genereer, en wanneer die helling van die bewegende gemiddelde negatief draai, word 'n verkoopsein genereer. Deur hierdie punte te gebruik, word die tweede fase van die model binnegetree deur die allokasie van gewigte aan die individuele aandele wat voldoen aan die koop-seleksie kriteria. Die gewigte word bepaal deur die helling van die lineêre regressie vergelyking. Die helling is sinoniem met die groeipotensiaal van die aandeel soos op die tydstip wat die herbelegging plaasvind. Deur gebreuik te maak van die bogenoemde model, is dit bevind dat die groei bo die hou periode van die ALS140 bereik kan word. Beoordeling van die model gebaseer op 'n kalenderjaar, het groei van 113% bo die 1999 groei van die ALS140 indeks gelewer. Soortgelyk is positiewe groei vir die 2000 kalenderjaar gelewer, wat derhalwe die verhandelingstegniek as suksesvol bevestig. Die grootste teenkant van die model is die aantal herbeleggings wat genereer word deur streng toepassing van die model. Byvoorbeeld, 110 herbeleggings was nodig gedurende 1999 om die 113% groei bo die indeks te bereik. Met die aanname dat herbeleggingskostes van 1% per herbelegging gehef word, word marges wat deur die houperiode-opbrengs van die ALS140 bereik kan word, vinnig uitgeroei. Alhoewel die model suksesvol is in die bereiking van die doel om die ALS140 opbrengs te verbeter, is die model en die rekenaarprogram wat ontwikkel is kragtig en primitief. 'n Groot aantal opsies is beskikbaar om die model te verbeter, en derhalwe die potensiaal om selfs hoër groei te genereer. Sulke opsies sal 'n verbeterde helling-funksie insluit asook 'n proses om herbeleggingskostes te verminder. Die herbeleggingstegniek wat gebruik is, is die effektiefste tegniek wat beskikbaar is vir 'n enkele bewegende gemiddelde.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/52406
This item appears in the following collections: