«Equity Style Investing RONG, WU How to cite: Equity Style Investing, Durham theses, Durham University. RONG, WU (2013) Available at Durham E-Theses ...»
Second, over the past decades and in the long-run, value investing tends to generate higher returns than growth strategy on most equity markets around the world. The reward to value investing is more pronounced for small-cap stocks, but it is also present in large-cap companies. Chan et al. (1991) first document that the return spread between the Japanese value and growth stocks defined by BM ratios is 1.1% per month. In the U.S. markets, Fama and French (1992) show value portfolios generate average monthly returns of 1.83% as compared to 0.30% of the growth portfolios. They also find that the size of value stocks with higher BM ratios on average tend to be smaller than growth stocks. Capaul et al. (1993) argue that the value premiums are pervasive in the international market, and Fama and French (1998) provide similar findings that a global value investing outperform the global growth investing for 7.6% annually from 1975Lakonishok et al. (1994) also document the outperformance of value investing on NYSE and AMEX stocks sorted by different valuation descriptors. They report that value portfolios sorted by BM ratios outperform the growth counterparts by 10.5% annually over the five years after formation, and such superior returns persisted if using different valuation criteria like PE ratios or PC ratios. Besides, the average size-adjusted value investing return is 3.5%, indicating a 7.8% spread relative to the growth strategy.
(2001) report the outperformance of value strategies using U.K. stock data for the period 1975 to 1998.
2.4 Explanations for size and value premiums Although the existence of size and value premium is relatively uncontroversial, there is much debate about the underlying reason behind it. The explanations regarding the size and value premiums split in the academic community.
For the size premium, some papers argue that small stocks tend to have high liquidity risk. For example, Amihud and Mendelson (1986) find that the size effect is linked to liquidity risk (measured as bidask spread) and therefore conclude that the size effect is largely a liquidity effect. Similarly, Liu (2006) argues that small-cap stocks perform better because they have low liquidity and hence investing in such smaller firms require higher returns for the compensation of bearing liquidity risk. Vassalou and Xing (2004), on the other hand, link the default risk to the size effect. They argue that small firms with highest default risk can earn high returns hence size premium can be viewed as a default risk effect. More recently, Zhang (2006) links the size premium to ‘information uncertainty’ provided to investors about small stock’s volatile fundamentals. Overall, these explanations are based on the classical financial theory that smaller firms are riskier than larger firms in general and hence conclude the outperformance of small-cap investing is driven by underlying sources of risk.
There are some competing theories to explain why value investing outperforms growth investing in general. Fama and French (1992, 1993, 1995) argue that value premium is the compensation for the higher risk of value stocks that is not explained by Capital Asset Pricing Model (CAPM). With the use of the multifactor asset pricing model in the context of Merton (1973), they link the higher returns of value stocks to exposure to the financial distress. The risk-based explanation is supported by authors like Liew and Vassalou (2000), Cooper et al. (2001), Lettau and Ludvigson (2001) and Petkova and Zhang (2003) and Vassalou and Xing (2004). Lakonishok et al.
(1994), however, do not support that value stocks are fundamentally risky. Lakonishok et al. (1994) compare value and growth stock performance under different economic conditions. They find that value stocks still outperform growth stocks in bad economic states and when the marginal utility of wealth is high. Hence it is concluded that value stocks actually have lower downside risk than growth stocks. Lakonishok et al. (1994) therefore suggest mispricing is the cause for the outperformance of value stocks. La Porta (1996) also argue that value investing works because expectations about future growth in earnings are too optimistic. Investors undervalue the value stocks and overvalue the growth stocks and the reward of value investing results from the correction of such mispricing. The mispricing story about value premium is also supported by Haugen
and Baker (1996) and Daniel and Titman (1997) in a behaviouralfinance framework.
There is another interpretation for the value premium which rests on the data-snooping hypothesis and poses a tough challenge to style investing. Lo and MacKinglay (1990) argue that the findings of value premium is due to data mining. Thus the methodological issue of sample selection bias causes the relative returns between value and growth strategies (c.f. Kothari et al. (1995), Conrad et al. (2003)).
Banz and Breen (1986) and Kothari et al. (1995) also suggest that ‘survivorship bias’ may contribute to the observed value premium.
Since some authors exclude delisted/dead companies in the year-toyear test and therefore fail to take into consideration the risk of financial distress for value stocks. Hence the cross-sectional return differences across stocks might be a statistical fluke.
2.5 Contrarian and Momentum investing
Parallel to style investing based on the classification of firm-specific characteristics, the implementation of investing strategies based on the correlations of asset returns is very popular. The properties of the short-term positive autocorrelation and long-term negative serial correlation of stock returns are well documented in the literature.
This academic finding forms the theoretical basis for contrarian and momentum investing widely recognised in the market. Contrarian investing of De Bondt and Thaler (1985, 1987) is to buy stocks that have performed poorly and sell stocks that have performed well in the past period. This strategy ignores the market trend and only focuses on the stocks which are considered to be mispriced. De Bondt and Thaler (1985, 1987) document that stocks experienced poor performance over a 3-5 years period subsequently outperform those that have previously performed well, and vice versa. The contrarian strategies of buying past losers and selling past winners can earn average profit of 25% over 3-year period. While this strategy is a relative long-run investing, Jegadesh (1990) and Lehman (1990) also find that it works in the short-term. Although studies on contrarian investing are initially based on the U.S. markets, it has also been widely investigated across continents both in developed markets and emerging markets. For example, in the U.K. market, Lonie and Lonie (1991), MacDonald and Power (1991) and Dissanaike (1997) document the abnormal returns from contrarian strategies based on monthly returns of UK stocks. Rouwenhorst (1998), Bildik and Gulay (2007), Galariotis (2004) and Antoniou et al.
(2005) find similar results. These studies all suggest that contrarian investing can generate economically significant profits.
However, similar to value premium, there is no general consensus regarding the cause of this profitability. De Bondt and Thaler (1985) interpret the contrarian profit being driven by investors’ overreaction to good and bad news, while Chan (1988) argues it is caused by the instability of risks for winner and loser stocks. Apart from the above explanations, there are schools of other thoughts such as the size effect (Clare and Thomas (1995)), January effect (Zarowin (1990) and the stock market microstructure bias (Conrad and Kaul (1993).
In contrast to contrarian strategy, momentum investing comes in various guises. Price momentum and earnings momentum are two of the most common types. Unlike contrarian strategy that exploits the long-run reversals of stock returns, the price momentum is based on the continuation of short-term and intermediate of cross-sectional stock returns. Such strategy follows the ‘trends’ to buy the past ‘winners’ and sell the past ‘losers’. The usual justification for this investing strategy is that the performance of both overall market and individual stocks is largely driven by investors’ sentiment which itself follows trends. Jegadeesh and Titman (1993) use stocks on the NYSE and AMEX markets to form self-financed portfolios and find that buying stocks with high returns over the previous 3-12 months and selling stocks with low returns over the same time period perform well in the following 12 months. When dealing with data, ten equally weighted deciles portfolios are constructed according to the ranking of returns in the past 3 to 12 months. The ‘winner’ is defined as the top deciles portfolios and ‘loser’ is identified as the bottom deciles. In their later study, Jegadeesh and Titman (2001) extend the dataset to 1998 and show that the initial results still held, suggesting that their initial findings are robust to the criticism of data-snooping.
Momentum profit is not only found in the individual stock level, but is also observed in the industry and country level. Moskowitz and Grinblatt (1999) document the large abnormal returns for industry momentum of buying past winner industries and selling past losing industries. Asness et al. (1997) also test the momentum strategies in industry portfolios and country portfolios. Furthermore, Lewellen (2002) finds that momentum strategy based on size and book-tomarket portfolios are at least as profitable as individual stock momentum. The profitability of momentum strategy is not only identified in the U.S. markets, but in international markets as well.
For example, in the U.K. market, Liu et al. (1999) document the profitability of momentum strategies over the period 1977-96. They argue that UK momentum effects are robust across two sub-samples in their dataset. Based on a different data sample source, Hon and Tonks (2003) also find that UK momentum effects exist in the subsample 1977-96, but not in the earlier 1955-76 period. Other studies such as Rouwenhourst (1998) and Bird and Whitaker (2003) all document the momentum effect in the European markets during periods of 1980-1995 and 1990-2002, respectively. Furthermore Richards (1997) find the monthly momentum profit in international markets from 16 countries during the period of 1970-1995. Overall, these studies would suggest that price momentum is a worldwide phenomenon in the investment marketplace.
The earnings momentum investing is based on the assumption that the reported earnings of a firm is a major source of information to which its underlying stock prices react. Ball and Brown (1968) suggest that the change in a company’s earnings from one reporting period to the next would cause a consistent movement in stock prices, and the post announcement earnings drift is also found to be relevant. This suggests that investment strategies based on earnings momentum are likely to be rewarded. Earnings momentum strategy forms the investing portfolios based on the direction and the magnitude of analysts’ earnings forecasts. Bird and Whitaker (2003) implement such strategy in major European markets for the periods of 1990-2002. They show that across the markets the performance of the quintile portfolios formed using the direction of ‘agreement’ as the criterion is significant for a period of up to 12 months, and the performance differentials between the low and high momentum portfolios is 7.5% annually. However, the performance of the portfolios based on the magnitude of the earnings forecast revisions is much weaker and inconsistent.
The profitability of momentum strategies seem to be at odds with the efficient market hypothesis since asset pricing models such as CAPM and Fama and French (1993) three-factor model all fail to explain it.
The academic view for the source of momentum profits is divided. A number of influential theoretical papers have sought to explain momentum effects based on cognitive biases in the behavioural finance framework. For example, De Bondt and Thaler (1985), Jegadeesh and Titman (1995), Daniel et al. (1998) propose the overreaction hypothesis. They argue that investors tend to overreact to news (both bad and good) and such overreaction could lead past losers to be underpriced and past winners to be overpriced, therefore resulting in better returns and worse returns for the losers and winners in the future, respectively. On the other hand, papers like Hong and Stein (1999) favour the underreaction hypothesis. They contend that momentum effect is related to underreaction since the positive autocorrelations of stock returns over short periods may reflect the slow transition of firm-specific news into its underlying stock prices. Specifically, stock prices may underreact to firm-related news like earnings announcements. If the underlying news is good in nature, stock prices may keep going up after the initial positive reaction. Conversely, stock prices will continue to fall down following the initial negative reaction when receiving the bad news. In addition to the overreaction and underreaction propositions, Barberis et al.
(1998) argue that momentum is caused by irrational investors’ underreaction to corporation news because investors suffer from representativeness bias and conservatism.
Recently, in addition to these behavioural explanations, Chordia and Shivakumar (2002) link the momentum effect to business cycles.
They find some evidence that momentum profits can be attributed to business cycle conditions and be predicted by lagged macroeconomic variables. However, this risk-based explanation is challenged by Griffin et al. (2003). Cooper et al. (2004) also argue that profits to momentum strategies depend critically on the state of the market, thus market state is the sort of conditioning information that is relevant for predicting the profitability of the momentum investing.
While the studies for properties of long-term and short-term stock return reversals have been well undertaken, previous researches focus primarily on the price and earnings side, rather than style side.