2. METODO
2.5 Variables y medidas
Since Fama's formulation of weak, semi-strong, and strong-form market efficiency in the 70s, hundreds of papers have empirically tested market efficiency and as Fama (1991:1575) notes himself, "The literature is now so large that a full review is impossible...". In this spirit, the subsequent chapter will briefly review the most relevant articles that have appeared on each of the three forms of market efficiency. Further, I will aim to highlight the position of my thesis with respect to existing empirical tests of market efficiency.
2.4.1 Weak Form Tests
Weak form tests of market efficiency were performed to test the price or return independence assumption of the Random Walk Model (or Random Walk Theory) (Fama, 1965). In other words, markets are said to be weak form efficient when technical traders cannot use information incorporated in historical prices to predict future prices. Meaning that prices or returns should be independent of each other. In order to test this assumption, Fama (1965) computed serial correlation coefficients, runs tests15, and Alexander's filter techniques on daily, four-day, nine-day, and sixteen-day stock market price changes, to conclude that there is no evidence of dependence, and that the independence assumption of the Random Walk Model can be confirmed. In the light of subsequent research (see e.g. Lo and MacKinlay, 1988), Fama (1991:1580) corrected his initial conclusion that returns could not be predicted. He argues that "...recent research is able to show confidently that daily and weekly returns are predictable from past returns." Not only did scholars find that past returns help predict future returns, other variables such as inflation, interest rates, dividend yields, earnings/price ratios, helped predict stock returns (Bodie, 1976).
2.4.2 Semi-Strong Form Tests, i.e. Event Studies
The goal of short-term semi-strong form market efficiency tests is to examine whether stock markets16 use available information correctly in setting stock prices (Fama, 1976), and the speed at which stock prices adjust to the release of new information (Fama, 1970). The dominant empirical tool to test for short-term semi-strong market efficiency is the event study methodology, using the CAPM or the market model (See e.g. Equation 4) (Dimson and Mussavian, 1998). Essentially, event studies examine the short-term price behaviour of securities around specific events such as changes in legislation, earnings announcements, or
stock splits (Binder, 1998). Fama et al. (1969) was the first study to perform an event study. To empirically test how quickly prices adjust to new information, Fama et al. (1969) investigated the effect of stock splits on securities' stock prices.17 The authors concluded that the stock market "almost immediately" incorporates stock split information after the announcement date (Fama et al., 1969:20). In relation to stock market efficiency, their findings implied that stock markets are indeed efficient given the rapid adjustment of stock prices to that particular event. Subsequent event studies have used various events to investigate how stock prices react to the arrival of new information in the form of mergers and acquisitions (Halpern, 1983), seasoned equity offerings (Akhigbe and Whyte, 2015), sovereign debt rating announcements (Michaelides et al., 2015), corporate social responsibility (Krüger, 2015), environmental regulation (Ramiah et al., 2013), and environmental programmes (Fisher-Vanden & Thorburn, 2011) have been studied extensively.
In the context of socially responsible investing, many event studies have been performed to investigate the stock price behaviour of firms to the arrival of new information in relation to environmental, social, and governance information and its implications for market efficiency (Frooman, 1997). Frooman's meta study of 27 published works on this topic revealed that shareholder wealth decreases if firms act in an irresponsible or illegal manner (Frooman, 1997). In particular, he observed significantly negative abnormal returns for firms with many controversial and illegal incidents (or events) such as criminal misconduct, fraud, pollution, product recalls, fines for safety violations, and price fixing. These findings are in line with recently published works (Krüger, 2015). Similarly, Hamilton (1995) finds that firms with higher pollution events such as air emissions or toxic spills experience significantly negative abnormal returns after the announcement. Klassen and McLaughlin (1996) even observe significantly positive abnormal returns for firms with strong environmental management initiatives.
With respect to the implications for stock market efficiency, the predominant view held among scholars using the event study methodology in this field is that usually new events related to environmental, social, and governance issues tend to be relatively quickly incorporated into the stock prices of firms (See e.g. Capelle-Blancard and Laguna, 2010; Endrikat, 2015; Jacobs et al., 2010; Little et al., 1995). As such, stock markets tend to be efficient relative to new publicly available information in the short-term. The emphasis on the
short-term is important, as the event study methodology is not capable of assessing whether the market is efficient over longer periods of time, and whether market efficiency persists or not. In the following chapter I will review studies testing market efficiency over longer time periods.
2.4.3 Strong-form Tests
The strong-form of the Efficient Market Theory implies that all available information (public and private information) is entirely reflected in stock prices (Fama, 1970). As the theory distinguishes between public and private information, it is important to explain what is meant by private information. The literature distinguishes between two types of private information, a) inside information held by corporate insiders, and b) private information held by professional investment managers, which could be in form of private assessments based on public information (e.g. an analyst's assessment report) (Fama, 1970, 1991). Thus, the goal of strong-form market efficiency tests is to investigate the long-term profitability of mutual fund managers and investment trading strategies with specialist information (Dimson and Mussavian; Fama, 1970). These tests are concerned with the ability of investors or trading strategies to consistently outperform the market, which should not be possible if the market is truly efficient (Fama, 1970). Jensen (1968) was probably one of the first and most comprehensive tests of this kind of strong form market efficiency tests, at the time. He analysed the investment performance18 of 115 mutual funds over the period from 1945 to 1964, using an "absolute" measure of investment performance, i.e. the multiperiod Capital Asset Pricing Model (CAPM) based on Lintner (1965) and Sharpe's (1964) single-period model (See e.g. Equation 4), and found very little evidence that mutual funds outperformed the market significantly, even before deducting mutual fund fees of a fund manager (Jensen, 1968). Based on his evidence, he concluded that markets are efficient and investors are better off with a passive buy-and-hold strategy (Jensen, 1968).
However, since the 70s "There is evidence that some security analysts (e.g., Value Line) have information not reflected in stock prices", Fama concluded in his second extensive review on the strong-form market efficiency (Fama, 1991:1603). He argues that some scholars (see e.g. Copeland and Mayers, 1982; Stickel, 1985) have found evidence that Value Line and other security analysts produce private information and when that information is revealed to the market it leads to stock price adjustments. These adjustments occur because producing private information is costly and investors are compensated for incurring these
costs (Fama, 1991). This implies that the stock market is less efficient because there will be some private information that is not entirely reflect in stock prices (Fama, 1991).
In the context of socially responsible and ESG investing, and similar to Jensen's work on mutual funds, several studies have investigated the financial performance of socially responsible mutual funds relative to conventional funds over the long term (Barnett and Salomon, 2006; Bauer et al., 2005; Geczy, et al., 2005; Hamilton et al., 1993; Kreander et al., 2005; Renneboog et al., 2008a). The main conclusion of these studies tends to be that socially responsible funds neither significantly outperform nor underperform relative to conventional funds19. In terms of the implications for market efficiency, this research would indicate that markets seem to be efficient in the long term.
However, in addition to studies on mutual funds, much research has assessed whether certain investment trading strategies based on ESG information generate abnormal returns over the long term (See e.g. Edmans, 2011; Gompers et al., 2003; Kempf and Osthoff, 2007), which would imply that markets are not efficient with respect to certain environmental, social, and governance information. As these studies find that investment trading strategies based on high social responsibility, good corporate governance, and positive employee relations generate significant abnormal returns, one could question whether markets are truly efficient in the long run with respect to certain information sub-sets.
In my thesis, the efficiency of capital markets is understood along the lines of Jensen's somewhat more economically relevant definition of an efficient market (See Chapter 2.2. 'Definition of the Efficient Market Theory'. This means that stock markets are efficient based on a certain information set as long as it is impossible to make economic profits based on that information set (Jensen, 1978). I will address the implications of each of my empirical findings for the efficiency of capital markets based on Jensen's definition. Having reviewed several studies that test Fama's weak, semi-strong, and strong-form tests of market efficiency, my thesis is most similar to the kind of market efficiency tests conducted by Jensen (1968), that aim to investigate the long-term profitability of investment trading strategies (portfolio trading strategies), equity indexes, and companies, rather than market efficiency tests in the strictest sense such as studies on insider trading20 (See e.g. Tavakoli et al., 2012). In this spirit, the studies closest to mine are those of Edmans (2011), Gompers et al. (2003), Kempf and Osthoff (2007), and Schröder (2007).
19 While some studies tend to find a significant outperformance of socially responsible mutual funds relative to conventional mutual funds (See e.g. Mallin et al., 1995).