An empirical examination of the usefulness of value-added and cash flow from operations in a South African context

Thesis:

Young, Roland David
Abstract

ENGLISH ABSTRACT: Returns and the ability to statistically formulate the association between returns and other accounting variables have increased in significance to investors. Over the past three decades substantial effort in the subject-field of accounting research has been devoted to the definition of this association between earnings and firm values. The fundamental issue is the information content of earnings and the value thereof to investors. Usefulness, in the returns/prices context, can be defined as the R2 of the regression of returns/prices on earnings (Lev, 1989: 156). Empirical evidence consistently shows that earnings are a statistically significant explanatory variable. A natural extension aiming at improving the empirical results is to incorporate GAAP-related alternative information. One approach followed by numerous researchers is to decompose earnings into components, and show that the components provide more information than earnings alone. Another approach is to include additional variables such as working capital from operations and cash flow from operations (eFO) in the regression models. The empirical results generally showed that earnings components and other GAAP-related variables have explanatory powers. The improvements in the predictive capability (measured by R2) of these models were restricted. Lev (1989: 181 - 182) suggested an intrepid research direction that departs from GAAP by investigating the alternative definitions of earnings. Limited empirical studies on value-added to date have shown that it is a useful measure of performance. For example, Bao and Bao (1989) empirically demonstrated that value-added is positively associated with firm value in certain industries. The analysis employed, however, was the firm valuation analysis. Recently developed levels and changes analyses were not used. In 1998 Bao and Bao completed another study using earnings as a benchmark. Firm valuation analysis, levels analysis and changes analysis were all performed to evaluate the explanatory power of value-added. The results show that value-added is a significant explanatory variable and its explanatory power is higher than that of earnings. The results obtained in this research project indicate that in the South African context earnings is statistically the most explanatory variable. This is divergent from the findings obtained in other research performed in this area, where value-added was indicated as the statistically most significant variable. It does appear that the explanatory power of valueadded tends to increase over the review period. This can possibly be attributed to the relative novelty of value-added in South Africa as opposed to Europe and the USA.

AFRIKAANSE OPSOMMING: Opbrengste en die vermoë om die verhouding tussen aandeelpryse en rekeningkundige veranderlikes statisties te formuleer, wek toenemende belangstelling by beleggers. Oor die afgelope drie dekades het daar 'n noemenswaardige skuif in die fokus op die vakgebied van rekeningkundige navorsing plaasgevind. Aansienlik meer tyd word gespandeer om die verhouding tussen aandeelpryse en die waarde van firmas te kwantifiseer in statistiese modelle. Die onderliggende begrip vir die navorsing is die informasie-inhoud en die inherente waarde wat opbrengste vir beleggers lewer. Waarde, in die opbrengste/pryse konteks, kan as die R2 van die regressie van aandeelpryse teenoor opbrengste gedefinieer word (Lev, 1989: 156). Empiriese navorsing toon konstant aan dat verdienste 'n statistiese beduinde veranderlike is. Die logiese ontwikkeling gerig op die verbetering van die empiriese resultate is om alternatiewe GAAP veranderlikes te inkorporeer. Een benadering wat deur menige navorsers gevolg is, is om opbrengste in komponente te onderverdeel en sodoende te bewys dat die komponente meer informasie-inhoud verskaf as opbrengste in geheel. 'n Alternatiewe benadering is om addisionele veranderlikes soos bedryfskapitaal en kontantvloei in die regressie modelle in te sluit. In die algemeen toon die empiriese resultate dat die komponente van opbrengste en ander GAAP-verwante veranderlikes wel beskrywende moontlikhede het. Die verbetering in die statistiese vooruitskattings vermoë (gemeet deur R2) van hierdie modelle was wel beperk. Lev (1989: 181 - 182) het 'n radikale nuwe navorsings-benadering voorgestel wat afwyk van GAAP deur alternatiewe definisies van opbrengste te ondersoek. Tot op datum toon die beperkte hoeveelheid empiriese navorsing op toegevoegde waarde dat dit wel 'n noemenswaardige mate van opbrengste is. Byvoorbeeld, Bao en Bao (1989) het empiries bewys dat toegevoegde waarde positief geassosieer is met firma waarde in sekere industrieë. Slegs die firma valuasiemodel is gebruik om die analise uit te voer en nie die onlangs ontwikkelde vlakke en veranderingsanalises nie. In 1998 het Bao en Bao nog 'n studie onderneem met opbrengste as die maatstaf. Firma valuasie, vlakke analise en veranderingsanalise is alles toegepas om die beskrywende moontlikhede van toegevoegde waarde te ondersoek. Die resultate toon dat toegevoegde waarde 'n noemenswaardige beskrywende veranderlike is, meer so as opbrengste. Die resultate van hierdie studie projek toon dat in die Suid-Afrikaanse konteks, verdienste die mees beduidendste statistiese veranderlike is. Dit is teenstrydig met die bevindinge van ander navorsers, waar toegevoegde waarde statisties meer beduidend is. Dit wil wel voorkom asof toegevoegde waarde se statistiese informasie-inhoud toegeneem het oor die relevante periode. Dit kan moontlik toegeskryf word aan die relatiewe onlangse bekendstelling in toegevoegde waarde in Suid-Afrika in teenstelling met Europa en die VSA.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/70096