An application of montier’s c-score to the johannesburg securities exchange: a tool for short selling

Abstract

One of the assumptions upon which modern portfolio theory is based is the efficient market hypothesis which postulates that market prices fully reflect all available information, which implies that an abnormal return cannot be made. Evidence has amassed in contradiction to the efficient market hypothesis as demonstrated by Jegadeesh and Titman (1993); Mohanram (2005); Montier (2009) and Piotroski, (2000). However these studies demonstrated earning an abnormal return by buying an asset as opposed to selling an asset. Evidence by Altman (2000) and Beneish, Lee and Nichols (2013) affirmed that abnormal returns may be earned by selling a declining asset. There has been no published work conducted on the South African market pertaining to an instrument that may be used to detect a decline in share price due to prior earnings manipulation, thereby providing the scope of this research. In recent years the focus of the discipline of asset pricing has shifted away from theoretical modelling towards empirical analysis. The C-score by Montier (2008) is a binary earnings manipulation detection model, designed to identify stocks that may be shorted for an abnormal return. An exploratory study of stocks on the Johannesburg Stock Exchange (JSE) from 2002 to 2010 was conducted. Vital focus areas included the resources and industrials sector. Results of this research prove that C-score is insufficient as a stand-alone tool for detecting shortable stocks on the JSE. Whilst negative relative returns were earned for certain holding periods of certain sectors, a consistent trend could not be isolated.