مقطع بازده سهام در بازارهای نوظهور

We are the first to investigate the cross-section of stock returns in the new emerging equity markets, the so-called frontier emerging markets. Our unique survivorship-bias free data set consists of more than 1400 stocks over the period 1997 to 2008 and covers 24 of the most liquid frontier emerging markets. The major benefit of using individual stock characteristics is that it allows us to investigate whether return factors that have been documented in developed countries also exist in these markets. We document the presence of economically and statistically significant value and momentum effects, and a local size effect. Our results indicate that the value and momentum effects still exist when incorporating conservative assumptions of transaction costs. Additionally, we show that value, momentum, and local size returns in frontier markets cannot be explained by global risk factors.

Traditional emerging markets have developed rapidly over the past decades, both economically and financially. A group of countries less developed than emerging markets with established stock exchanges has appeared on the radar screen of global investors. These new emerging markets as a group are also known as frontier emerging markets, or in short, frontier markets. These countries vary greatly in their economic development. The GDP per capita in 2008 of Bangladesh, for example is just $497 while that of Slovenia is $27,019.1 The market capitalization of stocks in frontier emerging markets in October 2008 is $113.6 billion.2 Although still smaller than traditional emerging and developed stock markets, these markets are becoming more important, as evidenced for example by recent listings of new mutual funds and exchange-traded funds on frontier markets.3 In addition, for academics, frontier emerging markets are an untapped data source that provides excellent out-of-sample research opportunities. Investors who are interested in improving the risk-return trade-off of their portfolios could expand their investment opportunity set by including frontier equity markets. Goetzmann et al. (2005) indicate that investors should be willing to keep expanding their investment horizon to new equity markets to get a better diversified portfolio. Speidell and Krohne (2007) also mention diversification benefits as a key motivation for investors to include frontier markets in their investment portfolios. Berger et al. (2011) investigate whether frontier equity markets are integrated with developed equity markets and conclude that this is not the case. These studies have in common that they consider frontier markets as a group or consider them at the country level. However, little is known about the risk, return, and diversification characteristics of return factors based on individual stock data in frontier markets.4 Our unique survivorship-bias free data set on individual stock characteristics in frontier markets allows us to construct portfolios based on other characteristics than the country of stock exchange listing. Hence, we are able to investigate the existence of value, momentum, size, and low-risk effects in these markets over the period 1997 to 2008 and gauge how much stronger these effects are when employed at the stock rather than the country level. Moreover, our data enable us to investigate whether investment strategies based on these cross-sectional stock attributes are correlated between developed, emerging, and frontier markets. Our paper aims to fill these gaps in the literature. This paper contributes to the literature on at least three dimensions. First, our results provide out-of-sample evidence for the existence of value, momentum, and local size effects. Sorting stocks in frontier markets on value characteristics, such as book-to-price ratios, momentum characteristics, such as past 6-month returns, or market capitalization per country yield statistically significant positive excess returns for the top quintile portfolios versus the index of 5% to 15% per annum. Our study extends the results by Fama and French (1998) and Rouwenhorst (1999) for international evidence on the value effect. Our results also reinforce the international evidence of the momentum effect reported by Griffin et al. (2003) and Rouwenhorst, 1998 and Rouwenhorst, 1999. Our results are further empirical evidence that value and momentum are present everywhere, as suggested by Asness et al. (2009). The presence of a local size effect confirms evidence in Europe by Heston et al. (1999) and emerging markets by Rouwenhorst (1999). Our results are important, as frontier markets are least integrated with developed and emerging equity markets, yet, the cross-section of stock returns seems to produce excess returns on exactly the same factors. Second, we are the first to investigate the profitability of value and momentum effects in frontier markets in detail when faced with real life market imperfections. We incorporate transaction costs estimates of 2.5% per single-trip transaction from Marshall et al. (2011) covering bid-ask spreads, market impact and commissions. We deem this to be a conservative estimate as we consider the largest half of our sample and apply a one-month lag between ranking and portfolio formation to account for possible opportunity costs. Our empirical findings indicate that transaction costs have a large impact on the profitability of value and momentum strategies. However, we still observe economically and statistically significant returns of approximately 6.6% to 7.7% per annum after incorporating transactions costs for value strategies and net returns of 4.6% to 7.2% for momentum strategies. These findings seem to be inconsistent with market efficiency. Third, we analyze whether exposure to global risk factors can explain the existence of the factor anomalies and whether the factors are prone to extreme downside risk. We document that the value, momentum, and local size effects in frontier markets cannot be explained by value, momentum, and local size effects in developed and emerging markets. This indicates that the excess returns are not caused by exposures to global risk factors and implies that our findings are independent of the existence of the effects in other markets. In addition we show that the downside risk of value, momentum and local size portfolios in frontier markets is lower than can be expected based on the assumptions of normality. Hence, we deem it unlikely that downside risk can explain the empirical results we document. Our paper is organized as follows. We start in Section 2 by describing the data and methodology used in our analyses. We investigate the value, momentum, size and low-risk effect in more detail in Section 3. In Section 4 we incorporate transactions costs in order to determine whether the cross-sectional return patterns still exist when faced with real life market imperfections. In Section 5 we investigate whether value, momentum, and local size effects in frontier markets can be explained by global risk factors. Finally, Section 6 concludes.

The new emerging equity markets, the so-called frontier emerging markets, are attracting increased attention from foreign investors. Research on these frontier markets is scarce and mostly conducted using the frontier market as a whole or at the country level. In this paper, we dig one step deeper and analyze the cross-section of individual stock returns. Our research on individual stocks in frontier emerging markets makes use of a unique high-quality and survivorship-bias free dataset. The use of individual stock characteristics data allows us to investigate the added value of investment strategies relative to strategies that only use aggregated data at the country level. We use data from more than 1,400 stocks from 24 frontier markets over a 12-year period from 1997 to 2008. This previously untapped data source provides excellent opportunities for out-of-sample research related to investment strategies that were previously analyzed in developed and emerging markets. Our empirical results indicate that portfolios based on value and momentum in frontier markets generate economically and statistically significant excess returns of about 5% to 15% per annum. The magnitude of these excess returns is at least as large as those found before in developed and emerging markets. We also find that there is a local size effect in frontier markets. These are striking empirical observations, as integration of frontier markets with developed and emerging markets is generally low. Our results are valuable out-of-sample evidence of the cross-section of stock returns previously documented in developed markets. These results are robust as they still hold after performing a battery of robustness analyses, such as an analysis by geographical region and financial liberalization. Investors who are interested to capture the value and momentum effect might be concerned with the transaction costs involved, as liquidity is typically lower than for more developed equity markets. We analyse the after transaction costs returns of value and momentum strategies using conservative estimates from Marshall et al. (2011) on a liquid sample of the largest 150 frontier market stocks including a one-month skip between ranking and implementation of the stocks in portfolio. Our results indicate that net excess returns are approximately 7% per annum for value and momentum strategies. These excess returns are both economically and statistically significant and therefore do not explain the existence of these factor returns. We additionally investigate whether the factor returns in frontier markets can be explained by risk. First, our results are not driven by frontier market, country- or region exposures, as our results still hold when correcting for these exposures. Second, our results cannot be explained by exposure to global risk factors, such as market, value, momentum and size. Third, it is unlikely that downside risk can explain the empirical results. Hence, we believe it is unlikely that transaction costs or risk can explain the strong factor returns. Although we cannot rule out that exposures to other global risk factors or local risk might explain the returns of the strategies, future research could investigate to which extent behavioural biases might explain the value, momentum and size effect in frontier markets.