Quantifying seasonal affective disorder in the South African capital market

Wagner, Anton Herman (2009-12)

Thesis (MBA (Business Management))--University of Stellenbosch, 2009.

Thesis

ENGLISH ABSTRACT: Experimental research in psychology and economics indicates that depression causes heightened risk aversion. Previous research has documented robust links between seasonal variation in length of day, seasonal depression (known as seasonal affective disorder, or SAD), risk aversion and stock market returns. One such study provides international evidence that stock market returns vary seasonally with the length of the day, a result called the SAD effect. Stock returns are shown to be significantly related to the amount of daylight throughout the autumn and winter. Another study examines the SAD effect in the context of an equilibrium asset pricing model to determine whether the seasonality can be explained using a conditional version of the capital asset pricing model (CAPM) that allows the price of risk to vary over time. Given the above as the base departure point, this report analyses the SAD effect in the context of the South African capital market, where, firstly, the variation in length of day during the year is not so severe compared with other countries like Sweden and the UK and, secondly, a more recent dataset includes the effects of integrated markets and globalisation, that possibly resulted in a shift of seasonal behaviour in the market. It quantifies the SAD effect in general, across industry sectors, over time periods and confirms that a conditional CAPM holds in explaining the seasonality due to SAD. The results differ substantially from those of prior studies. The expected signs of the SAD and Aut coefficients are reversed. Closer analysis shows that seasonality in stock returns has undergone a shift compared to seasonality in an older dataset. A prior finding that effects, such as the SAD, are better explained using excess returns than using raw total returns of the market, is reinforced. The analysis of SAD over time sheds some light on the unexpected outcome of the SAD and Aut coefficients by providing evidence that the validity of regression models deteriorated over time and, more conclusively, that in two consecutive periods, the SAD and Aut coefficients decreased in absolute value. It also found that the coefficients are linearly related to excess returns during the latter period only. The conditional CAPM provides evidence that the effect of SAD is captured in the time variation in the price of risk. The factor of reducing the remainder of SAD in error terms is, however, remarkably smaller. The implication is that market risk already accounts for the SAD effect, but only to a degree, and that the remaining contribution of the SAD effect contained in varying the price of risk is substantially less significant. This finding coupled with the contradictory results in the signs of SAD and Aut coefficients renders evidence of SAD in the South African market rather inconclusive.

AFRIKAANSE OPSOMMING: Navorsing in die gebied waar sielkunde en ekonomie oorvleuel, toon dat depressie ‘n verlaagde risiko-aptyt meebring. Meer spesifiek, vorige navorsing dokumenteer dat robuuste skakels tussen seisoenale lengtes van dae, seisoenale depressie (beter bekend as winter depressie), risiko-aptyt en opbrengste op die andelebeurs bestaan. Een van hierdie studies, wat uitgevooer was op ’n aantal hoof internasionale markte, bevind dat opbrengste op die aandelebeurse wel seisoenaal varieer in ooreenstemming met die lengtes van dae. In dié studie word daar getoon dat opbrengste noemenswaardig ooreenstem met die hoeveelheid sonlig gedurende herfs en winter. ’n Ander studie bestudeer weer seisoenale depressie in konteks met ’n ekwilibrium kapitaal-bate prysmodel om vas te stel of seisoenaliteit verklaar kan word deur ’n kondisionele weergawe van hierdie model waar die prys van risiko varieer oor tyd. Met bogenoemde as vertrekpunt, analiseer die verslag wat volg winterdepressie in die Suid-Afrikaanse aandelemark waar, eerstens, die variasie in lengtes van dae gedurende die jaar minder is as vir ander lande soos Swede en Engeland, en tweedens, waar ’n meer onlangse datastel die effekte van geintegreerde markte en globalisasie insluit. Die verslag kwantifiseer seisoenale depressie in die algemeen op die effektebeurs, pas dan regressie-modelle toe op verskeie industrie-sektore en oor verskillende periodes. Ten einde, word bevestig dat ’n kondisionele kapitaal-bate prysmodel seisoenaliteit as gevolg van winterdepressie kan verklaar. Vergeleke met vorige studies, word daar teenstrydighede in die resultate opgemerk ten opsigte van die verwagte tekens van die koeffisiënte. Met nadere ondersoek word bevind dat die seisoenaliteit in opbrengste ’n verskuiwing ondergaan het vergeleke met ’n ouer datastel. ’n Vorige bevinding dat faktore, tipies soos seisoenale depressie, beter verklaar kan word deur alpha opbrengste (risiko vrye obrengste afgetrek) as rou opbrengste, word bevestig. Die analise oor tyd verklaar gedeeltelik die onverwagte koeëfisiënte. Daar word waargeneem dat die geldigheid van regressie modelle oor twee opeenvolgende tydperke verswak het. Verder word daar ook bevind dat die absolute waardes van die koeëfisiënte verklein het en dat koeëfisiënte ’n liniêre verwantskap met opbrengste toon, slegs in die latere tydperk. Die toepassing van ‘n kondisionele kapitaal-bate prysmodel bevestig dat seisoenale depressie teenwoordig is in die prys van risiko. Daar word verder bevestig dat mark risiko gedeeltelik seisoenale depressie verklaar en dat die oorblywende teenwoordigheid daarvan in die prys van risiko heelwat minder statistiese geldigheid het. Gegewe dit, tesame met die teenstrydigheid in die tekens van die koeëfisiënte, word enige duidelike konklusies ten opsigte van seisoenale depressie se teenwoordigheid in die Suid-Afrikaanse mark verhoed.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/954
This item appears in the following collections: