Momentum and contrarian strategies in international stock markets: Further evidence

Abstract

Previous studies have shown that market participants underestimate earnings growth for past winner stocks, and that growth stocks are more sensitive to earnings surprises. These findings suggest implementing momentum strategies with growth stocks. This study investigates linkages between value versus growth investment styles and momentum strategies in international markets. In addition, we extend Jegadeesh and Titman (2001)-type tests, which attempt to distinguish between competing explanations of the momentum phenomenon, to international market indices. Our full sample results show that momentum profits are concentrated in the growth indices, and that there is evidence of short-term overreaction in these and other indices that is subsequently corrected. Our subsample results are mixed; there is some evidence that the profitability of momentum (but not contrarian) strategies persists in the post-December 1987 period. However, unlike the earlier period, there is no evidence that markets overreact and that these overreactions are subsequently corrected.

Introduction

In recent years, practitioners and academic scholars have found that some relatively simple trading strategies based on past cross-sectional stock returns yield significant abnormal profits. Researchers divide these strategies into two major categories: the contrarian strategy that relies on price reversals and the momentum strategy based on price continuations. For the U.S. Stock Market, DeBondt and Thaler (1985) investigate return patterns over extended periods of time and find that contrarian strategies perform exceptionally well over 3–5-year horizons. In contrast, Jegadeesh and Titman (1993) document that strategies that buy winner stocks and sell loser stocks (i.e. momentum strategies) generate significant positive returns (about 1% per month) over 3–12-month holding periods. In a comprehensive investigation, Conrad and Kaul (1998) find both momentum and contrarian profits in the U.S. market, depending on the time horizon investigated. Specifically, the contrarian strategy is profitable for short-term (weekly, monthly) and long-term (2–5 years, or longer) intervals, while the momentum strategy is profitable for medium-term (3–12-month) holding periods.³

Many studies have found that momentum strategies work in international stock markets, although findings have not been uniform.⁴ Rouwenhorst (1998) examines twelve European markets’ stock returns between 1980 and 1995. He finds that an internationally diversified portfolio of past medium-term winners outperforms a portfolio of medium-term losers by more than 1% per month after taking risk factors into consideration. He also presents evidence that European and U.S. momentum strategies have a common component suggesting that a common factor may drive the profitability of momentum strategies in both the U.S. and international markets. Echoing these findings, Schiereck et al. (1999) report substantial intermediate horizon profits to momentum strategies in the German market, and find that the German and U.S. markets behave very similarly. Van Dijk and Huibers (2002) link momentum profits in European markets to analyst behavior. Specifically, they find that analysts systematically underestimate earnings for strong price-momentum stocks, underestimate autocorrelation in earnings growth between consecutive years, and are in general too slow to adjust their earnings forecasts.⁵ They suggest that further research should focus on the relative importance of the underreaction explanation for momentum. Hon and Tonks (2003) examine momentum strategies in the U.K. stock market. Interestingly, they report strong evidence for momentum profits out to 24-month horizons, but only for 1977–1996. They find no evidence of momentum profits in the earlier 1955–1976 period in the U.K.

In general, evidence for momentum profits is weaker in Asian markets. Chui et al. (2000) report evidence for momentum profits in these markets, but with the notable exceptions of those in Japan and Korea. Hameed and Kusnadi (2002) find virtually no evidence of momentum profits in the stock markets of six Pacific Basin countries, and although Kang et al. (2002) do find significant momentum profits at 3–6-month horizons in the Chinese market, their study was confined to the “A” shares traded exclusively by local investors.

Because of the wide availability of country index data, many studies of momentum and contrarian strategies in international equity markets have examined these strategies at the country (as opposed to individual firm) level.⁶ These studies attempt to assess if investors under and/or overreact to information not only at the level of the individual stock, but more systematically on a countrywide basis. Evidence for mean reversion in country stock indices, which is related (albeit imperfectly) to the success of DeBondt and Thaler (1985)-style contrarian strategies, is reported by Richards (1995) and Balvers et al. (2000). Richards (1997) finds that the DeBondt and Thaler contrarian strategies applied to country indices earn annualized excess returns of approximately 6% at three and 4-year horizons, and that these excess returns are statistically significant. Richards also reports annualized momentum profits of around 3% per annum at 3–12-month horizons, but these are not significant. However, Asness et al. (1997) and Balvers and Wu (2002) do report significant momentum profits in country indices at 6–12-month horizons. Chan et al. (2000) also use international equity indices and examine three different phenomena in their study, i.e. country-specific momentum profitability; the relation among momentum profits, interest rate and exchange rate movements; and, the relation between momentum profits and trading volume. From these tests, they conclude that the momentum profits arise mainly from time-series predictability in stock market indices, rather than from interest rate or exchange rate movements. In contrast to the U.S. market, where momentum profits generally require ranking and holding periods in excess of 3 months, they also find significant positive profits for short holding periods (less than 4 weeks) in the international equity indices.

The causes of momentum and/or contrarian profits have been the subject of considerable debate. One debate revolves around whether momentum profits can be reconciled with market efficiency. Those who argue that such reconciliation is possible stress that observed momentum profits may be a product of data mining, and/or that momentum profits are simply fair compensation for risk. While data mining is always a possibility, it is unlikely to be the major explanation for the momentum anomaly, given that roughly similar strategies have been shown to work in a variety of global settings, and for different subperiods in the U.S. as documented by Conrad and Kaul (1998). Furthermore, Jegadeesh and Titman (2001) find that in U.S. markets, over the 1990–1998 sample period (which was not included in their original 1993 paper) momentum strategies continue to be profitable, and past winners outperform past losers by about the same magnitude as in the earlier period.⁷

Similarly, as Korajczyk and Sadka (2004) note, the consensus in the literature is that risk factors do not completely explain intermediate horizon momentum profits. Fama and French (1996) concede that their three-factor asset-pricing model does not explain returns to momentum portfolios. Nevertheless, some studies, e.g. Conrad and Kaul (1998), have indicated momentum profits could be a by-product of certain stocks being riskier than others in some unknown way, and thus having higher expected returns. This is because momentum strategies take long (short) positions in stocks with high (low) past returns; if these past returns are high (low) because of unknown systematic risk factors, then the same stocks should continue to earn relatively high (low) returns in future periods. If this interpretation is correct, then momentum profits can be consistent with market efficiency. Conrad and Kaul (1998) argue that the cross-sectional variation in the mean returns of individual securities plays an important role in their profitability, and could potentially account for the profitability of momentum strategies. However, recent findings reported by Jegadeesh and Titman, 2001, Jegadeesh and Titman, 2002 cast substantial doubt on the Conrad and Kaul Hypothesis. If momentum profits were due primarily to cross-sectional differences in mean returns, then past winners (losers) should continue to be superior (inferior) performers indefinitely into the future. Instead, Jegadeesh and Titman (2001) find that momentum portfolio (winners minus losers) returns are positive only during the first 12 months of portfolio formation; if anything, momentum portfolio returns beyond this 12-month horizon are negative. As another risk-based explanation, Chordia and Shivakumar (2002) allege that momentum profits are explained by exposure to macroeconomic risk factors; however, Griffin et al. (2003) demonstrate that this linkage does not exist in an international context.

In addition to Jegadeesh and Titman (2001), other studies have reported results consistent with behavioral models (e.g. Barberis et al., 1998, Daniel et al., 1998, Hong and Stein, 1999) that suggest that investors initially underreact and, ultimately, overreact to new information. Chan et al. (1996) show that stock prices underreact to earnings news and momentum profits concentrate on subsequent earnings announcements. Also, they suggest that investors tend to engage in herding behavior. Lakonishok et al. (1992) find evidence of underreaction for pension fund managers, while Grinblatt et al. (1995) report similar behavior by mutual funds. However, in a comprehensive study of institutional investors, Badrinath and Wahal (2002) find significant differences in trading practices across different classes of institutions (e.g. banks, pension funds, investment advisors and mutual funds), and caution against making broad generalizations concerning institutional trading behavior.

Another anomaly that has received a great deal of attention in the finance literature is the value/glamour anomaly. Many studies have found that a strategy of investing in value stocks (i.e. those with relatively low market-to-book ratios, P/E ratios or past earnings growth) produces higher returns than investing in growth stocks in the U.S. as well as in international markets. Researchers have offered a variety of reasons for this performance difference. Fama and French (1993) argue that higher average returns on value stocks merely compensate for the higher risk they bear; value stocks have positive loadings on a factor related to relative distress. On the other hand, Lakonishok et al. (1994) and other studies argue that value strategies earn higher returns because these strategies exploit the suboptimal behavior of the typical investor. Investors irrationally extrapolate past earnings growth, thereby overvaluing companies that have performed well in the past, and undervaluing those that have performed poorly. Finally, still another branch of the asset pricing literature (see, for example, Kothari et al., 1995) holds that the superior past performance of value stocks likely arises from data snooping, which finds accidental patterns in historical data that are unlikely to be repeated in the future.

The main purposes of this study are to investigate linkages between value/glamour and momentum strategies in international markets, and to extend Jegadeesh and Titman (2001)-type tests, which attempt to distinguish between competing explanations of the momentum phenomenon, to international market indices. There are two reasons why momentum strategies may work better if implemented with growth stocks. First, as Chan et al. (1996) show for the U.S. and Van Dijk and Huibers (2002) show for European markets, analysts underestimate future earnings growth for past winner stocks. Skinner and Sloan (2002) demonstrate that growth stock prices are considerably more sensitive to earnings surprises; consequently, if past winner stocks experience positive earnings surprises on average, these surprises will have a greater impact on those winner stocks that are also growth stocks. A second issue regarding growth stocks is that there is greater uncertainty in their valuation. As Miller (1977) argues, in an environment in which short selling is at least somewhat restricted, the most optimistic investors have a disproportionate impact on stock prices. The greater the dispersion of opinion as to what a stock is worth, the greater the likelihood that a stock will be overvalued. Thus, we conjecture that growth stocks are more likely to be mispriced. If our conjecture is correct, we would expect much of the momentum profits generated by holding past winner growth stocks to reverse if these stocks are held beyond the 1-year horizon past studies have found optimal for implementing momentum strategies.

Linkages between the momentum and the value/growth phenomena have been examined in U.S. markets, but not in a global context. Asness (1997) directly studies this interaction in the U.S. market during the period of 1963–1994 and concludes that both value and momentum strategies are effective, although value and momentum measures are negatively correlated. Grinblatt and Moskowitz (2003) similarly report that, controlling for market capitalization, momentum strategies work better if implemented with low book-to-market (i.e. growth) stocks.

In this study, we examine momentum and contrarian strategies at the country index level using the Morgan Stanley Capital International (MSCI) growth, value and blended indices. If growth stocks are more sensitive to earnings surprises than value stocks, we would intuitively expect that momentum strategies would work better within the growth indices, and this is exactly what our results show. An examination of performance by months after portfolio formation reveals that somewhere between 7 and 10 months after formation, the momentum portfolios, which initially earn large positive returns, begin earning negative returns. This underperformance continues out to 36 months after formation. However, this underperformance, while highly significant, is qualitatively similar for the growth and value indices. Thus, our second conjecture that growth stocks are more likely to be mispriced, is not strongly supported.

In addition to testing for momentum and contrarian profits in our full sample (December 1974–November 2000), we also conduct tests for two approximately equal-length subperiods, December 1974–November 1987, and December 1987–November 2000. During the later period, overall equity indices for 13 emerging markets are also available from MSCI, enabling us to ascertain the profitability of momentum/contrarian strategies in emerging market indices, and to compare these with developed markets. Our subperiod results show that, in developed markets, evidence for contrarian profits is much weaker in the post-December 1987 period than in the previous period. The momentum profits may persist to some degree (depending on the methodology employed); however, contrary to the earlier period, there is no evidence of investor overreaction post-1987. Similarly, in emerging markets, momentum profits appear to be present post-1987, but at most time horizons the variance of returns is too large to draw meaningful statistical inferences at conventional levels.

The balance of this paper is organized as follows: the basic methodology and data are described in Section 2. Section 3 contains more detailed descriptions of the specific empirical tests and their results. Section 4 concludes the paper.