Fama and French (1996) derived another multi-factor model known as the Fama-French Three- Factor model which at present is the principal approach to identifying factors for sources of systematic risk (Bodie et al., 2014). This model applies firm traits that appear on empirical grounds to be representation of exposure to market or systematic risk. The factors selected are variables have in the past predicted average returns accurately and may therefore be depicting risk premiums. The Fama and French three-factor model is given as:
π ππ‘ = ππ + π½ππ(πΎππ‘β π ππ‘) + π½ππππ΅πππ΅π‘+ π½ππ»ππΏπ»ππΏπ‘+ πππ‘ Where:
π ππ‘ = The expected returns of the portfolio
Page | 30 πΎππ‘ = The return of the market portfolio
π ππ‘ = Risk-free rate
SMB = Small Minus Big, the difference between the return of a portfolio of large stocks and a portfolio of small stocks.
HML = High Minus Low, the difference between the return of a portfolio high book-to-market- value ratio stocks and the return of a low book-to-market ratio stocks portfolio.
The market index in this model captures systematic risk emanating from macroeconomic factors. Although HML and SMB are in themselves not obvious candidates for pertinent risk factors, it is argued that they may be proxies for more fundamental yet-unknown variables.
Fama and French for example, posit that firms in distress financially are more likely to have high book-to-market value ratios and small stocks may be more sensitive to variations in business conditions hence, these variables may portray sensitivity to macroeconomic risk factors.
Black (1993) however argues that whenever researchers continuously scan the database of asset returns when searching for explanatory factors, they can eventually unearth past βpatternsβ that purely are due to chance, a practice commonly referred to as data-snooping. Black (1993) further points out that since discovery, the return premiums to factors like firm size have turned out to be inconsistent. Fama and French however have proved that the book-to-market and size ratios have in various time periods and markets over the world predicted average returns therefore diminishing possible effects of data-snooping (French, 2015).
With regards to factors, there are at least 3166 factors that have been tested by financial market researchers that explain the cross-section of expected returns (Harvey, Liu and Zhu, 2015). In fact, Cochrane (2011), describes this as βa zoo of new factorsβ. Figure 2.6 shows the growth in the publications of factors over the years.
6 Appendix D presents a detailed list of all the factors
Page | 31 Figure 2.6: Factors and Publications
Source: Harvey, Lui and Zhu (2015: 19)
Harvey, Lui and Zhu (2015) argue that many of the factors discovered are only significant by chance and therefore, it is a dangerous mistake in asset pricing tests to apply the typical statistical significance cutoffs such as a t-statistic exceeding 2.0.
Furthermore, McLean and Pontiff (2016) studied the return predictability of 97 factors that academic studies have shown to forecast the cross-section of stock returns using out-of-sample and post-publication and found that factors lose 26% of their power after discovery. This inter alia, may be attributed to the effects of data mining. Factors further lose 32% of their predictability power after they appear in academic papers suggesting that investors only learn about this mispricing only after they have been published in academic papers. Financial markets can therefore not be construed to incorporate all relevant information since factor models purely reflect risk-return trade-offs and should not be affected by the publications done by academics.
Ranguelova, Feeney and Lu (2015) discovered that there has been a significant decline in alpha as shown in Figure 2.7. Hedge funds included in the HFRI Composite index in 2001 achieved a rolling 3-year alpha of 25%, peaking at around 35% in 2002, before declining and finally
Page | 32 plateauing between 5 and 10% after 2008. Strategies that required lower entry barriers, for example long/short equities, suffered the steepest alpha generation decay as a result of more players entering the industry. In the long/short equities strategy group however, Figure 2.8 shows managers focusing on small and mid-cap equities generated more alpha and experienced less alpha decay relative to their peers concentrating on large-caps since mid-2004.
Figure 2.7: Alpha Generation Decay: HFRI Composite from January 2001 β January 2015
Source: Adopted from Investcorp, Bloomberg, cited in Ranguelova, Feeney and Lu (2015: 1) It is well known that the small-cap equities offer a more attractive opportunity for generating alpha through investing in βunder the radarβ equities discovered by talented stock pickers.
Managers who concentrate on small and mid-cap equities benefit from the βstructural alphaβ
existing in this segment by exploiting lingering inefficiencies such as the quality of information for example, fewer analysts and less frequent publication of research reports on these stocks, and the limited volume and also, from a playing field that is less crowded relative to that of large cap equities (Ranguelova et al., 2015). Figure 2.9 shows that the number of analysts assigned to small cap equities as at March 2015 were far fewer than the number of analysts assigned to large cap equities.
Page | 33 Figure 2.8: Less Alpha Decay in Small-Cap Hedge Fund Strategies from January 2001 β January 2015
Source: Adopted from Investcorp, Bloomberg, cited in Ranguelova, Feeney and Lu (2015: 1)
Figure 2.9: Number of Analysts per Market Cap Size Decile (March 2015)
Source: Adopted from Investcorp, cited in Ranguelova, Feeney and Lu (2015: 4)
Ranguelova et al. (2015) compiled data from Capital IQ data service to tally the number of publications for the largest 10 and smallest 10 smallest companies of the Russell 2000 and the S&P 500 indices over a period of 30 days, three months, six months and one year. Defining publications as earnings estimates, research reports, research notes, fixed income reports, and articles on industry overview as they relate to the selected companies, financial models, initiation of coverage memos, notices of rating change and reporting results summaries. The study reports that there is a significant difference in publication frequency for constituents of
Page | 34 the two indices (Figure 2.10). There were on average, 36 publications for a large S&P 500 stock compared to just 10 for a large Russell 2000 stock.
Figure 2.10: Dispersion of Analyst Forecasts per Market Cap Size Decile (March 2015)
Source: Adopted from Investcorp, cited in Ranguelova, Feeney and Lu (2015: 6)
The dispersion of analystsβ forecasts for companies with market capitalizations of $7 billion or more diverge in a tight range of less than 10% increasing monotonically as the size of the market capitalization drops. This implies smaller companies attract less analyst attention, leading to limited coverage by few analysts. Therefore, there are structural inefficiencies in the U.S. small-cap equity market that can be exploited to generate alpha (Ranguelova et al., 2015).