Do the size, value, and momentum factors drive stock returns in emerging markets?

Accepted Manuscript Title: Do the size, value, and momentum factors drive stock returns in emerging markets? Author: Nusret Cakici, Yi Tang, An Yan PII: DOI: Reference:

S0261-5606(16)30047-X http://dx.doi.org/doi: 10.1016/j.jimonfin.2016.06.001 JIMF 1676

To appear in:

Journal of International Money and Finance

Please cite this article as: Nusret Cakici, Yi Tang, An Yan, Do the size, value, and momentum factors drive stock returns in emerging markets?, Journal of International Money and Finance (2016), http://dx.doi.org/doi: 10.1016/j.jimonfin.2016.06.001. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Do the Size, Value, and Momentum Factors Drive Stock Returns in Emerging Markets? Nusret Cakici *, Yi Tangξ, and An Yan† Current version: June 2016

Abstract This paper investigates the size, value and momentum effects in 18 emerging stock markets during the period 1990 − 2013. We find that size and momentum strategies generally fail to generate superior returns in emerging markets. The value effect exists in all markets except Brazil, and it is robust to different periods and market conditions. Value premiums tend to move positively together across different markets, and such inter-market comovements increase overtime and during the global financial crisis.

JEL Classification code: F21, F65, G12, G15 Keywords: Emerging markets, cross-sectional stock returns, market comovements

*

Corresponding author, Gabelli School of Business, Fordham University, 45 Columbus Avenue, New York, NY 10023. Email: [email protected]. Phone: (212) 636 6776. Fax: (212) 586-0575. ξ Gabelli School of Business, Fordham University, 45 Columbus Avenue, New York, NY 10023. Email: [email protected]. Phone: (646) 312-8292. Fax: (646) 312-8295. † Gabelli School of Business, Fordham University, 113 West 60th Street, New York, NY 10023. Email: [email protected]. Phone: (212) 636-7401. Fax: (212) 586-0575.

We thank the special issue editor, Professor Iftekhar Hasan, and three referees for their extremely helpful comments and suggestions. We also benefited from discussions with participants at the 2015 Conference honoring Professor James R. Lothian. 1 Page 1 of 47

1.

Introduction Emerging markets have become an important part of a global equity portfolio allocation.

According to MSCI Inc., in 1988 when Morgan Stanley Capital International (MSCI) launched the first comprehensive emerging markets index there were just 10 countries in the MSCI Emerging Markets Index, representing less than 1% of world market capitalization. Since then, the index covers more than 800 securities across 23 emerging markets and represents approximately 11% of world market capitalization. There is considerable empirical research identifying value and momentum effects in the U.S. and other develop markets.1 Despite the fact that emerging markets constitute an increasing share of the world's equity markets, there are far less empirical studies that investigate the size, value and momentum effects for individual emerging markets. Some pioneering empirical studies have documented the presence of value and momentum effects in emerging markets (e.g., Fama and French (1998, 2012), Griffin, Ji, and Martin (2003), and Rouwenhorst (1998, 1992)). They show that value stocks with higher book-to-market equity ratios (BM) generate higher average returns than growth stocks with low BM, and that stocks with higher returns over the past year outperform those with lower returns. Cakici, Fabozzi, and Tan (2014), a more recent paper, study the value and momentum effects in three emerging regions (Asia, Eastern Europe, and Latin America), and find strong evidence for the value effect in all regions and the momentum effect in all regions but Eastern Europe.

1

Among many others, see Stattman (1980), Debondt and Thaler (1987), Fama and French (1992, 1996), and Lakonishok, shleifer and Vishny (1994) for evidence on the value effect, and Jegadeesh and Titman (1993, 2001) on the momentum effect in the U.S; See Fama and French (1998, 2012), Rouwenhorst (1998, 1999), Griffin (2002), Griffin, Ji and Martin (2003), Chui, Titman, and Wei (2010), and Hou, Karolyi, and Kho (2011), and Asness, Moskowitz, and Pedersen (2013), among many others, for evidence on the value and momentum effects in developed markets.

1 Page 2 of 47

Different from previous studies on emerging markets, which often group different markets into several regions, and explore the value and momentum effects within a region, this paper analyzes the effects for individual emerging markets, including China, India, Indonesia, South Korea, Malaysia, Philippines, Taiwan, and Thailand in Asia, Argentina, Brazil, Chile, Colombia, and Mexico in Latin America, and Czech Republic, Hungary, Russia, Poland, and Turkey in Eastern Europe. It is crucial for investors who diversify internationally using a bottomup portfolio approach to understand whether these well-documented patterns in stock returns exist at the market level, how they are correlated within the local market and across different markets, and how they move together with macroeconomic and financial risks. Our work is similar to Cakici and Tan (2014) who investigate value and momentum factors in 23 developed stock markets. However, there are ample reasons to suspect that the value and momentum effects may behave differently in emerging markets. On one hand, although it is well documented by numerous studies that buying stocks with low prices relative to their fundamentals (i.e., value stocks) as well as stocks with higher past medium-term returns (i.e, past winners) generates superior future returns, the interpretation of these return patterns has been very controversial. For example, Fama and French (1992, 1993) and Chen Zhang (1998) attribute the value premium to distress risk, while Lakonishok, Shleifer, and Vishny (1994), La Porta (1996), and La Porta, Lakonishok, Shleifer, and Vishny (1997) show evidence that the value premium is associated with mispricing due to naive investor expectations of future growth. Similarly, there are two competing explanations of the momentum premium − the risk-based (e.g., Jegadeesh and Titman (1993), Fama and French (1996), Grundy and Martin (2001), Chordia and Shivakumar (2002), Griffin, Ji, and Martin (2003) , Liu and Zhang (2008)) and the behavioral-based (e.g., Barberis, Shleifer, and Vishny (1998), Daniel, Hirshleifer and

2 Page 3 of 47

Subrahmanyam (1998), Hong and Stein (1999), Hong, Lim and Stein (2000), Grinblatt and Moskowitz (2004), and Israel and Moskowitz (2013)). On the other hand, emerging markets usually have slower information diffusion, higher transaction costs, lower institutional investor participation than developed markets. Retail investors in emerging markets tend to passively hold suboptimal portfolios due to underdeveloped financial markets. These market frictions intertwined with the mechanisms underlying the value and momentum premiums can either exacerbate or dampen the value and momentum effects on cross-sectional variation in expected stock returns. Following Fama and French (1993), we construct factor mimicking portfolios based on size, book-to-market equity ratio, and momentum, and calculate monthly factors with respect to size (SMB), value (HML), and momentum (UMD) for individual markets. 2 We find that during the period 1990 − 2013, the size effect does not exist in all emerging markets except China. In contrast, the average SMB is negative in 14 of 18 markets, ranging from –0.10% per month in Czech Republic to –1.02% per month in Hungary. The average HML is always positive in a range from 0.41% per month in Brazil to 2.34% per month in India, and it is statistically significant in all markets except Brazil. On the other hand, the momentum effect, which tends to be stronger than the size and value effects in the U.S. and other developed markets, is surprisingly weak in most emerging markets. Although the average UMD is positive in 14 markets with a minimum of 0.05% per month in Malaysia and a maximum of 1.99% per month

2

We focus on the SMB, HML, and UMD factors because they along with the market factor are commonly used in empirical asset pricing studies (e.g., Fama and French (1993, 2010), Carhart (1997), among numerous others), and because they have been carefully analyzed in developed markets including U.S. and emerging regions. We closely follow the empirical methodology of previous studies to ensure that our findings are comparable to theirs. However, we agree that there are more return patterns worth investigating. For example, Francis, Hasan, and Hunter (2008), by applying a conditional asset pricing model to 36 US industries, document strong evidence on the importance of currency movements to industry competitiveness. We are planning to explore these ideas in our future work.

3 Page 4 of 47

in India, it is statistically significant only in Chile (with a mean HML of 0.74%) and India (with a mean HML of 2.34% per month). Next, we correlate value and momentum factors. We show that they often move inversely together both within the same market and across different markets. Moreover, the negative correlations between the value and momentum factors are stronger for big stocks than for small stocks and in the periods when the U.S. stock market posts positive returns than in the periods of negative U.S. stock market returns. On the other hand, the value factor tends to move positively together across different markets, so does the momentum factor. We then regress value and momentum factors for individual markets against stock market returns calculated for the local, regional, global and U.S. markets. We show that the value and momentum factors, the former in particular, remain intact after controlling for the stock market returns, suggesting that the value and momentum factors are not driven by market returns. We trace the value and momentum factors to macroeconomic fundamentals, funding liquidity, stock market liquidity, and credit risks measured for local, regional, global and U.S. markets. We find that value factors on average have positive loadings on local, regional, global, and U.S. future GDP growth, and especially so to local GDP growth. Specifically, exposure of the value factor to local future GDP growth is positive in five (India, South Korea, Malaysia, Philippines, and Thailand) out of the eight markets in Asia and in all Latin America and Eastern Europe markets, with a mean coefficient (averaged across all markets) of 1.21, implying that 1% higher GDP growth over the future three-year period on average is associated with 1.21% higher value return. The momentum factor has relatively weak correlation with future GDP growth. On the other hand, both the value and momentum factors are insensitive to funding liquidity, market

4 Page 5 of 47

liquidity risk, and credit risk, suggesting that the 18 emerging markets may have not been adequately integrated into world markets. We further explore a risk-based explanation of the value premium in individual markets by examining whether the systematic risk of value stocks is on average greater than that of growth stocks. We find that it is so only in half of the emerging markets we study, of which the difference is statistically significant only in Tailand and Mexico. Hence, we do not find convincing evidence that support a risk-based explanation of the value premium in emerging markets. Finally, we perform subperiod analyses. We find that the value strategy in emerging markets is robust to different sample periods and market conditions, while the momentum effect is generally weak. Moreover, the value premium has become more positively correlated across emerging markets overtime, and are even more so during the global financial crisis period. The remainder of the paper proceeds as follows. Section 2 describes the data, the variables, and the methodology that we use to form the factors. Section 3 discusses the results. Section 4 concludes the paper.

2.

Data In this section, we describe the data that we use in the paper. We measure all variables in

U.S. dollars. The sample period is January 1990 to December 2013.

2.1.

Equity data We retrieve stock-level data on monthly returns, market capitalization, and book equity

for 18 emerging markets from Datastream. We obtain the one-month U.S. Treasury bill rate from

5 Page 6 of 47

Kenneth French's online data library to proxy for the monthly risk-free rate. Following Fama and French (1992, 1993, 2000), we calculate stock size (SIZE) as the product of price per share and number of shares outstanding (in millions of dollars) and the ratio of book value of equity to market value of equity (BM). Following prior research, we lag BM by six months to ensure that the accounting information is available in the portfolio formation month. Following Jegadeesh and Titman (1993), we calculate stock-level momentum (MOM) as the cumulative stock return over a period of 11 months ending one month prior to month t.

2.2.

Data on macroeconomic, liquidity, and credit risks To study macroeconomic condition in individual markets, we download data on annual

nominal GDP growth from the World Bank website. We measure macroeconomic condition by change in annual nominal GDP growth rate. To study funding liquidity, we consider three variables: the VIX, the LIBOR rate, and the U.S. TED rate. The daily implied volatility of S&P 500 index options (VIX) is from the Chicago Board of Options Exchange (CBOE). We average daily VIX to obtain the annual measure. The LIBOR rate is the 3-month London Interbank Offered Rate (LIBOR) based on the U.S. Dollar. The U.S. TED spread is the difference between the LIBOR rate and the 3-month U.S. T-bill rate. The LIBOR and Treasury constant maturity rates are from the FRED database of the Federal Reserve Bank of St. Louis. To study market liquidity risk, we obtain two sets of variables from the Center for Research in Security Prices (CRSP): 1) the level, the innovation, and the traded factors of Pastor and Stambaugh (2003), and 2) the fixed and the variable factors of Sadka (2006).

6 Page 7 of 47

Finally, to study credit risk, we extract data on the 10-year constant maturity U.S. Treasury rate (T-note) and Moody's seasoned Aaa (Aaa) and Baa (Baa) corporate bond rates are from the FRED database of the Federal Reserve Bank of St. Louis. We measure credit risk by Aaa minus T-note, and Baa minus Aaa. We annualize the aforementioned liquidity and credit risk variables by averaging their monthly values.

2.3.

Constructing factors We calculate monthly market, size, value, and momentum factors for individual markets.

For each market, the market factor (MKT) is the monthly excess return on the value-weighted portfolio that consists of all stocks in the market covered by Datastream. We closely follow Fama and French (2012) when calculating the size and value factors. At the beginning of each month over the period January 1990−December 2013, we first group the largest market capitalization stocks into the big portfolio (B), which accounts for 90% of the total market capitalization. All remaining stocks are grouped into the small portfolio (S). Within the big portfolio, we determine the bottom 30% (L), middle 40% (M), and top 30% (H) breakpoints for the BM ratio and apply these BM breakpoints to both the big and small stocks. These classifications allow us to form six value-weighted portfolios: S/L, S/M, S/H, B/L, B/M, and B/H. The size factor (SMB) is the average of the returns on the three small stock portfolios (S/L, S/M, S/H) minus the average of the returns on the three big stock portfolios (B/L, B/M, and B/H). We measure the value factor (HML) as the average of the monthly returns on S/H and B/H portfolios minus the average of the monthly returns on S/L and B/L portfolios.

7 Page 8 of 47

We construct the monthly momentum factor (UMD) in the same fashion as the HML factor, except that the sample starts from January 1991. We lose the monthly data from January 1990 to December 1990 to compute momentum of individual stocks at the end of December of 1990.

3.

Results

3.1.

Market, size, value, and momentum In this section, we provide sample statistics on the size, value and momentum factors in

individual markets. We start with calculating the cross-sectional medians of stock size (SIZE) measured in millions of dollars and book-to-market equity ratio (BM). We divide stocks into small (S) and big (B) portfolios using the size median. Table 1 reports the time-series averages of the cross-sectional SIZE and BM medians, the number of stocks each month (N), the monthly excess returns on the value-weighted market portfolio (MKT), the one-month-ahead returns on the value-weighted small and big portfolios, and the monthly size-related factor (SMB), measured as the return difference between portfolios S and B. [Insert Table 1 about here] Table 1 shows that the average MKT over the sample period is statistically indistinguishable from zero in all markets except Indonesia. The average excess market return in Indonesia is 1.84% per month with a t-statistic of 2.14.3 Table 1 also shows that the average monthly SMB is negative in 14 of the 18 emerging markets, ranging from –0.10% with a tstatistic of –0.27 in Czech Republic to –1.02% with a t-statistic of –2.69 in Hungary. SMB is significantly positive only in China with an average premium of 1.06% per month and t-statistic

3

Throughout the paper, we calculate t-statistics (t-stat) using the Newey-West (1987) procedure with six lags. We use 5% as our statistical significance level and present significant values in bold.

8 Page 9 of 47

of 3.13. These observations are largely consistent with Fama and French (2012) and Cakici and Tan (2014) who report insignificant size premium in developed markets. [Insert Table 2 about here] Table 2 presents the time-series means and t-statistics of the value factor (HML) for individual markets. It shows that the average HML factor is positive in all 18 markets ranging from 0.41% per month in Brazil to 2.34% per month in Indonesia. It is statistically significant in 17 markets with the exception of Brazil. The corresponding annualized Sharpe ratio is economically significant as well, in a range from 0.42 in Taiwan to 1.22 in Thailand. To investigate whether the value effect varies systematically with stock size, we also measure value premium for stocks in the small portfolio (HML_S) and for stocks in the big portfolio (HML_B), as well as the difference between HML_S and HML_B (HML_S–B). Similar to Cakici and Tan (2014), we find that the value premium for small stocks is somewhat stronger, indicating that the results at the regional level documented in Fama and French (2012) are robust at the market level. For example, Table 2 shows that the average HML_S is between 0.53% per month in Brazil to 2.30% per month in Indonesia, and is statistically significant in 14 markets. On the other hand, the average HML_B is between 0.34% in Chile and 2.38% in Indonesia, and is statistically significant only in seven markets. Moreover, the average difference between HML_S and HML_B is positive in 12 markets, and is statistically significant only in South Korea and Thailand. Table 3 presents the time-series means and t-statistics of the momentum factor (UMD) for individual markets. The momentum factor is on average much weaker than the value factor. Specifically, although the UMD factor is positive in 14 markets, it is statistically significant in two markets: India with an average UMD of 1.99% per month and a t-statistic of 2.89, and Chile with an average UMD of 0.74% per month and a t-statistic of 2.75. The annualized Sharpe ratio, 9 Page 10 of 47

ranging from 0.02 in Malaysia and Philippines to 0.76 in India, is on average less economically significant than that of HML. [Insert Table 3 about here] Table 3 further shows that the momentum factor for small stocks (UMD_S) is slightly stronger than for big stocks (UMD_B). The average UMD_S is positive and statistically significant in three markets: India with an average UMD_S of 1.30% and a t-statistic of 2.22, Chile with an average UMD_S of 1.21% and a t-statistic of 3.91, and Mexico with an average UMD_S of 1.17% and a t-statistic of 2.35. On the other hand, the average UMD_B is significantly positive only in India with an average UMD_B of 2.68% and a t-statistic of 2.86. Moreover, the average difference between UMD_S and UMD_B (UMD_S–B) is positive in 11 markets, though it is statistically insignificant in all these markets except Chile with an average UMD_S–B of 0.94% per month and a t-statistic of 2.86. Next, we investigate the source of value premiums: Is it due to underperformance of the growth stocks in the lowest BM group, outperformance of the stocks in the highest BM group, or both? Specifically, for each market we regress returns on value-weighted portfolios sorted based on the book-to-market equity ratio against returns of the local market. We find that the CAPM alpha for value stocks (i.e., stocks in the highest BM group) is positive in all 18 markets, and statistically significant in all but Brazil, Argentina, and Columbia. On the other hand, the CAPM alpha for growth stocks is negative and statistically significant only in China, Indonesia, Argentina Columbia, and Turkey. These results indicate that the value premium in China, Indonesia, and Turkey is due to both outperformance of value stocks and underperformance of

10 Page 11 of 47

growth stocks, that in Argentina and Columbia is due to underperformance of growth stocks, and that in the remaining markets is primarily driven by outperformance of value stocks. 4

3.2.

Factor correlations In this section, we analyze correlations between value and momentum factors within each

market as well as across different markets.

3.2.1

Intra-market correlation between the value and momentum factors We first examine the correlations between the factors within the same market. The intra-

market correlations are particularly important for investors pursuing a menu of value and momentum strategies within a local market. For each market, we correlate the value and momentum factors calculated for all stocks (HML and UMD), stocks in the small portfolio (HML_S and UMD_S), and stocks in the big portfolio (HML_B and UMD_B). Table 4 reports the correlation coefficients and the NeweyWest adjusted t-statistics. [Insert Table 4 about here] Panel A of Table 4 shows that the correlation coefficient between HML and UMD is negative in 17 markets, ranging from –0.03 in Hungary to –0.44 in Malaysia, and is statistically significant in six of the 17 markets. Panel A further shows that the correlation coefficient between HML _S and UMD_S is negative in 14 markets ranging from –0.04 in China to –0.26 in Taiwan, and is statistically significant in five of the 14 markets. The correlation coefficient

4

We also investigate the source of the size effect in China and the source of the momentum effect in Chile and India. We find that the size effect in China is due to outperformance of small stocks, but not to underperformance of big stocks; the momentum effect in India is driven by outperformance of winner stocks as well as underperformance of loser stocks, whereas that in Chile is mainly driven by outperformance of winner stocks.

11 Page 12 of 47

between HML_B and UMD_B is negative in 15 markets ranging from –0.05 in Indonesia to – 0.42 in Malaysia, and is statistically significant in seven of the 15 markets. The negative intra-market correlations between HML and UMD suggest that investors may be able to improve risk-return efficiency by combining value and momentum strategies. To test this conjecture, we use a naive portfolio strategy that gives equal weight to the HML and UMD factors. The last three columns in Panel A of Table 4 report the mean, t-statistic, and annualized Sharpe ratio of the naive strategy for each market. Not surprising, the Sharpe ratio of the combination strategy is not significantly improved relative to the simple strategy that invests 100% in the HML factor since the UMD factor is insignificant in most emerging markets. However, investors in India and Chile, the only markets in which the average UMD factor is significantly positive, are indeed better off by taking the combination strategy. Panel A of Table 4 also shows that the Sharpe ratio increases from 0.53 (0.76) of the HML (UMD) factor to 1.11 in India, and from 0.61 (0.68) of the HML (UMD) factor to 0.97 in Chile. Panel B reports the averages of the intra-market correlation coefficients between HML and UMD averaged across all markets within the same region as well as the average correlation coefficients in the up markets (when the U.S. equity market return is positive) and the down markets (when the U.S. equity market return is negative). It shows that the average correlation coefficients between HML and UMD are, respectively, –0.34 (–0.20), –0.18 (–0.06), and –0.20 (0.00) in Asia, Latin America, and Eastern Europe markets when the U.S. stock market moves up (down). These coefficients indicate that the negative correlation between HML and UMD within the same market is stronger in the up markets.

12 Page 13 of 47

Overall, our results suggest that the value and momentum factors tend to move inversely together within the same market, and the negative intra-market correlations are stronger for big stocks than for small stocks and in the up markets than in the down markets.

3.2.2

Inter-market correlations Next, we examine the correlations between factors across different markets. The inter-

market correlations are particularly useful for investors investing in foreign stock markets. [Insert Table 5 about here] We first correlate the value factor in two markets. Table 5 reports the pairwise correlation coefficients between the value and momentum factors. The diagonal values in Table 5 are the correlations within the same market, are they are the same as those reported in Panel A of Table 4. Table 5 shows that 225 of 324 pairwise correlations are negative, with an average correlation of –0.11. 14% of these negative correlations are statistically significant. [Insert Table 6 about here] Panel A of Table 6 reports the averages of the inter-market correlation coefficients between the value and momentum factors calculated based on all stocks (HML and UMD), small stocks (HML_S and UMD_S), and big stocks (HML_B and UMD_B) averaged across all markets within the same region. Panels B and C Table 6 present the similar inter-market correlation coefficients for the up and down markets, respectively. Panel A shows that the average inter-market correlations between HML and UMD are, respectively, –0.06, –0.07, and – 0.04 in Asia, Latin America, and Eastern Europe markets. It also shows that the negative correlations between the value and momentum factors are stronger for big stocks than for small stocks in every region. Panels B and C show that the negative intra-market correlations between the value and momentum factors are stronger in the up markets than in the down markets. 13 Page 14 of 47

[Insert Table 7 about here] Table 7 further shows that 108 out of the 153 pairwise correlations between the value factors are positive, of which more than 17% are statistically significant. Table 8 shows that 144 out of the 153 pairwise correlations between the momentum factors are positive and about 40% of these positive correlations are statistically significant. Thus, the value factors as well as the momentum factors tend to move together across different markets. [Insert Table 8 about here] In sum, consistent with our findings based on intra-market correlations between the value and momentum factors, our inter-market results confirm that the value and momentum factors tend to move inversely together across different markets. The negative inter-market correlations are stronger for big stocks than for small stocks and in the up markets than in the down markets. On the other hand, the value factors often move positive together across different markets, so do the momentum factors.

3.3.

CAPM alphas and betas In this section, we test whether value and momentum factors can be absorbed by the

stock market factors. Specifically, for each market we estimate the following CAPM models: (1) where

and

are, respectively, market i’s value and momentum factors;

denotes excess returns of a market portfolio, including the market itself (Market CAPM), the region in which the market resides (Regional CAPM), the global market (Global CAPM) that consists of the developed markets studied in Fama and French (2012), and the U.S. stock market. [Insert Table 9 about here]

14 Page 15 of 47

Table 9 presents the alphas and slope coefficients from regression (1) by including separately the returns of the five market portfolios as independent variables. It shows that the alphas of the value factors remain positive and highly significant. For example, the alphas from the Market CAPM are in a range from 0.34% per month in Brazil to 2.06% per month in Indonesia, and they are statistically significant in all markets but Brazil. On the other hand, the value factors have positive loadings on returns of the local market (15 out of 18), and about half of the loadings are statistically significant. The results from the other CAPM specifications are qualitatively similar, though the value factors are less sensitive to returns of the regional, global, and U.S. stock markets. Finally, we run regression (1) by including the returns of all five market portfolios as independent variables. The regression results are reported in the last two columns of Table 9. Consistent with the results from the univariate regressions, the alphas from the multivariate regressions remain positive and statistically significant in all markets except Brazil. [Insert Table 10 about here] Table 10 presents the alphas and slope coefficients of the CAPM models similar to model (1) but with UMD as the dependent variable. Consistent with the results reported in Table 3, Table 10 shows that the alphas of the UMD factors are positive and statistically significant only in India, Chile, and Mexico. On the other hand, the UMD factors generally have negative loadings on returns of the five market portfolios. Overall, our results above suggest that the value and momentum factors cannot be explained by the local, regional, global, and U.S. factors. Furthermore, the value factors on average have positive loadings on the stock market factors, whereas the momentum factors have negative loadings. These loadings indicate that the value factors move positively with the stock markets, whereas the momentum factors move inversely.

15 Page 16 of 47

3.4.

Exposures to macroeconomic condition, funding liquidity, market liquidity, and

credit risk Previous studies show that asset exposures to macroeconomic risks have profound implications for return dynamics in the cross-section. For example, Bali and Engle (2010) find a positive and significant link between time-varying conditional default beta and future portfolio returns. Bali, Brown, and Caglayan (2011) find hedge funds with higher/lower exposure to default/inflation risk generate higher expected returns in the cross-section. Motivated by these findings, in this section, we trace the value and momentum factors to macroeconomic and financial risks by estimating exposures of value and momentum factors to macroeconomic conditions, funding liquidity, stock market liquidity, and credit risk.5

3.4.1. Exposures of the value factors We first investigate exposures of value factors. We run regressions of annual value factors against future GDP growth and contemporaneous annual stock market returns. For each market, we estimate three sets of regressions: The first regression uses the GDP growth and the stock return of the local market as the independent variables; the second uses the U.S. future GDP growth and the U.S. market return; and the last uses the global future GDP growth and the global market return. Annual GDP data are from the World Bank website. Future GDP growth is the three-year GDP growth and is defined as (GDPt+3–GDPt)/GDPt, where t indexes years. [Insert Table 11 about here]

5

Caution is needed in interpreting the results from the tests that trace the value and momentum factors to macroeconomic and financial because the sample period 1990−2013 includes a number of structural changes possibly induced by the 1991-1992 credit crunch, the 1998 Russian default and LTCM debacle, the 2000-2001 bursting of the tech bubble, and the 2007-2009 global financial crisis, and because the emerging markets studied in the paper are at different stages of economic development.

16 Page 17 of 47

Table 11 shows that the value factor on average has a positive loading on the GDP growth of the local market. It is positive in five (India, South Korea, Malaysia, Philippines, and Thailand) out of the eight markets in Asia and in all Latin America and Eastern Europe markets, with a mean coefficient (averaged across all markets) of 1.21. This mean coefficient implies that an increase of 1% in GDP growth over the future three-year period on average is associated with an increase of 1.21% in value return. The results are qualitatively similar but slightly weaker when the global and U.S. data are used in the regressions. Next, we run regressions of annual value factors against a funding liquidity variable or a credit risk variable after controlling for future U.S. GDP growth and contemporaneous annual return of the U.S. stock market. The three funding liquidity variables are the VIX, the LIBOR rate, and the U.S. TED rate, and two credit risk variables are Moody's seasoned Aaa corporate bond rate minus the 10-year constant maturity U.S. Treasury rate (Aaa-T-note) and the difference between Moody's seasoned Baa and Aaa corporate bond rates (Baa-Aaa). [Insert Table 12 about here] Table 12 shows that the value factors on average have weak loadings on funding liquidity and credit risks. The slope coefficients on the funding liquidity variables are generally insignificant and swing between positive and negative values. The slope coefficients on the credit risk variables are always insignificant. Finally, we run regressions of annual value factors against a variable that captures the stock market liquidity risk after controlling for future U.S. GDP growth and contemporaneous annual return of the U.S. stock market. We use two sets of stock market liquidity variables: 1) the level, the innovation and the traded factor variables from Pastor and Stambaugh (2003), and 2) the fixed and the variable components from Sadaka (2006). Table 13 shows that the value factors are on average insensitive to liquidity risk of the U.S. stock market. 17 Page 18 of 47

[Insert Table 13 about here] Our results indicate that the value factors have positive exposures to future GDP growth. The value factors are more sensitive to local GDP growth than regional, global, and U.S. GDP growth. On the other hand, the value factors are generally insensitive to funding liquidity, market liquidity, and credit risks of the U.S. market.

3.4.2. Exposures of the momentum factors In this section, we investigate the exposures of the momentum factors to the macroeconomic conditions, funding liquidity, and stock market liquidity, and credit risk. As in Section 3.4.1, we first run regressions of annual momentum factors against future GDP growth and contemporaneous annual stock market returns. We then regress annual momentum factors against a funding liquidity variable or a credit risk variable after controlling for future U.S. GDP growth and contemporaneous annual return of the U.S. stock market. Finally, we run regressions of annual momentum factors against a variable that captures the stock market liquidity risk after controlling for future U.S. GDP growth and contemporaneous annual return of the U.S. stock market. [Insert Tables 14, 15, and 16 about here] Different from the results based on value factors, Table 14 shows that the slope coefficients of future GDP growth swing from negative to positive values, and they are insignificant in most regressions. These results indicate that the momentum factors are insensitive to future GDP growth. On the other hand, similar to the results based on the value factors, Tables 15 and 16 show that the momentum factors on average have insignificant loadings on funding liquidity, credit, and market liquidity risk. These results are not surprising given the fact that the momentum effect is weak in most markets. 18 Page 19 of 47

In sum, value and momentum factors of individual markets are weakly exposed to macroeconomic conditions, liquidity, and credit risk of regional, global and U.S. markets.

3.5.

Explore a risk-based explanation of the value premium While there is some agreement that value strategies that call for buying stocks with low

prices relative to book equity, earnings, dividends, or other measures of fundamental value produce superior returns, the interpretation of why they do so is controversial. Fama and French (1992, 1993) and Chen Zhang (1998) attribute the value premium to distress risk, while Lakonishok, Shleifer, and Vishny (1994), La Porta (1996), and La Porta, Lakonishok, Shleifer, and Vishny (1997) show evidence that the value premium is associated with mispricing due to naive investor expectations of future growth.. In this section, we explore a risk-based explanation. Specifically, we estimate the panel regression for each market: (2) where

is return of the local market i in month t;

denotes returns for value-weighted

portfolio of value stocks (i.e., stocks in the highest BM group) and returns for value-weighted portfolio of growth stocks (i.e., stocks in the lowest BM group);

is a dummy variable

taking the value of 1 for returns of value stocks and 0 otherwise. Hence, the interaction term (

) captures difference in average systematic risk between value stocks and growth

stocks. We find that the slope coefficient on the interaction term is significantly positive in Tailand and Mexico, it is significantly negative in Russia, and it swings between positive and

19 Page 20 of 47

negative signs in the other markets and is statistically insignificant. Collectively, these results do not strongly support a risk-based explanation of the value premium in the emerging markets.

3.6.

Subperiod analyses Emerging stock markets have become less segmented from world stock markets in the

last two decades characterized by increased integration in trade, capital flows and movement of labor. These changes have sophisticated implications for stock returns. For example, Jong and Roon (2005), after accounting for the time-varying nature of market integration, find that decrease in segmentation significantly reduces cost of capital and stock returns in emerging markets. Bae and Zhang (2015) find that stock markets more integrated towards world markets experience larger price drops during the 2008 financial crisis. Motivated by these important findings about emerging markets, we perform subperiod analyses of the value and momentum premiums in this section. We split the sample into four subperiods: 1) the period 1990−2001 (first half), 2) the period 2002−2013 (second half), 3) the global financial-crisis period December 2007−June 2009 (crisis), and 4) the non-crisis period 1990−2013 after excluding the global financial-crisis period (non-crisis). First, we examine the monthly value and momentum premiums for the four subperiods. Panels A of Table 17 in the online appendix reports the results for the value premium. In the panel, entries under "First versus second half" present average monthly premiums for the first and second subperiods and their difference; entries under "Non-crisis versus crisis period" present average monthly premiums for the non-crisis and crisis periods and their difference. Panel A shows that the average value premium remains positive in all markets during the first and second subperiods. The difference in the value premium between the two subperiods is

20 Page 21 of 47

statistically insignificant in all markets except for India. Panel A further shows that the value premium during the global financial crisis period remains positive in all emerging markets except for Columbia and Czech. The difference in the value premium between the non-crisis and crisis periods is statistically insignificant in all markets except for Brazil. These results indicate that the value strategy in emerging markets is generally robust to different sample periods and states of the economy. [Insert Table 17 about here] Panels B of Table 17 presents the subperiod results for the momentum premium. The momentum effect does not exist in all markets during the period 1990−2001. It becomes significantly positive in India, Malaysia and Chile during the period 2002−2013. The difference in the momentum premium between the first and the second half of the sample period is statistically insignificant in all markets except for Malaysia. Panel B further shows that the momentum premium during the global financial crisis period is on average negative in 15 emerging markets, though it is statistically insignificant in all markets potentially due to small number of observations. Overall, the subperiod results indicate that the momentum effect is surprisingly weak in emerging markets. [Insert Table 18 about here] Next, we perform subperiod analysis of correlations between value premiums across different markets (i.e., inter-market correlations). 6 Panels A and B of Table 18, respectively, report differences in inter-market correlations between the first and the second half, and differences between the non-crisis and the crisis periods. Panel A shows that the average of the 153 differences in pairwise correlations between the first and the second subperiods is −0.07, and

6

Given the weak cross-sectional return predictability of the momentum effect in the emerging markets, we concentrate on the value premium in the remainder of Section 3.6.

21 Page 22 of 47

99 out of the 153 differences are negative, of which more than 22% are statistically significant. Panel B shows that the average of the 153 differences in pairwise correlations between the noncrisis and crisis subperiods is −0.13, and 102 out of the 153 differences are negative, of which more than 20% are statistically significant. These results indicate that the value premium has become more positively correlated across emerging markets overtime, and are even more so during the global financial crisis period. The increased comovements in value premiums across emerging markets are consistent with the notion that emerging markets are more integrated into world markets, and may exert significant impacts on efficiency of international portfolio diversification.

4. Conclusions This paper contributes to the empirical asset pricing literature by exploring the stocklevel data on 18 emerging markets and examining the size, value, and momentum effects in individual markets. We find that during our sample period from 1990 to 2013, the size effect does not exist in all markets except China. We identify the value effect in all markets but Brazil. On the other hand, the momentum effect is surprisingly weak in most emerging markets. We further show that stock market returns cannot subsume value and momentum factors. Correlation analyses indicate that value and momentum factors tend to move inversely together both within the same market and across different markets. The negative comovements are generally stronger among big stocks and in periods during which the U.S. stock market posts positive returns. On the other hand, the value factors tend to move positively together across different markets, so do the momentum factors.

22 Page 23 of 47

We trace the value and momentum factors to macroeconomic fundamentals as well as funding liquidity, stock market liquidity risk, and credit risk measured for local, regional, global, and U.S. markets. We find that the value factors on average have positive exposures to local, regional, global, and U.S. future GDP growth, and especially so to local GDP growth. On the other hand, the value and momentum factors are generally not sensitive to funding liquidity, market liquidity risk, and credit risk. These results suggest that the 18 emerging markets may have not been adequately integrated into world markets. We futher explore a risk-based explanation of the value premium in individual markets by comparing the systematic risk of value and growth stocks. We find that the systematic risk of value stocks is significantly greater than that of growth stocks only in Tailand and Mexico. Hence, the value premium in the emerging markets may not be driven by a risk-based mechanism. Finally, we perform subperiod analyses. We find that the value strategy in emerging markets is robust to different sample periods and states of the economy, while the momentum effect is generally weak. Furthermore, the value premium has become more positively correlated across emerging markets overtime, and are even more so during the global financial crisis period. The increased comovements in value premiums across emerging markets are consistent with the notion that emerging markets are more integrated into world markets, and may significantly impact efficiency of international portfolio diversification. For future research, it would interesting to exploit the information embedded in the crosssectional variation in the momentum factors across different markets, and understand why the momentum effect is on average more pronounced in developed markets than in emerging markets.

23 Page 24 of 47

References Asness, Clifford S., Tobias J. Moskowitz, and Lasse H. Pedersen, 2013, "Value and Momentum Everywhere,” Journal of Finance, 929–985. Bae, Kee-Hong and Xin Zhang, 2015, “The Cost of Stock Market Integration in Emerging Markets,” Asia-Pacific Journal of Financial Studies, 44, 1-23. DeBondt, Werner F. M. and Richard H. Thaler, 1987, “Further evidence on investor overreaction and stock market seasonality,” Journal of Finance 42, 557–581. Bali, Turan G. and Robert F. Engle, 2010, "The Intertemporal Capital Asset Pricing Model with Dynamic Conditional Correlations," Journal of Monetary Economics, 57, 377–390. Bali, Turan G., Stephen J. Brown, and Mustafa O. Caglayan, 2011, "Do Hedge Funds’ Exposures to Risk Factors Predict Their Future Returns?" Journal of Financial Economics, 101, 36–68. Barberis, Nicholas, Andrei Shleifer, Robert Vishny, 1998, "A model of investor sentiment," Journal of Financial Economics, 49, 307–343 Cakici, Nusret, Frank J. Fabozzi, and Sinan Tan, 2013, "Size, Value, and Momentum in Emerging Market Stock Returns," Emerging Markets Review, 16, 46–65. Cakici, Nusret and Sinan Tan, 2014, "Size, Value, and Momentum in Developed Country Equity Returns: Macroeconomic and Liquidity Exposures," Journal of International Money and Finance, 44, 179–209. Carhart, Mark M., 1997, "On persistence in mutual fund performance," Journal of Finance, 52, 57–82. Chen, Nai-fu and Feng Zhang, 1998, "Risk and Return of Value Stocks," Journal of Business, 71, 501– 535. Chordia, Tarun and Lakshmanan Shivakumar, 2002, "Momentum, Business Cycle, and Time-varying Expected Returns," Journal of Finance, 57, 985–1019. Chui, Andy, Sheridan Titman, and K.C. Wei, 2010, “Individualism and Momentum around the World,” Journal of Finance 65, 361–392. Griffin, M. John, 2002, "Are the Fama and French Factors Global or Country Specific," Review of Financial Studies, 15, 783–803. Griffin, M. John, Xiuqing Ji, and Spencer Martin, 2003, " Momentum Investing and Business Cycle Risk: Evidence from Pole to Pole," Journal of Finance, 58, 2515–2547.

24 Page 25 of 47

Grinblatt, Mark and Tobias J. Moskowitz, 2004, "Predicting Stock Price Movements from Past Returns: The Role of Consistency and Tax-Loss Selling," Journal of Financial Economics, 71, 541–79. Grundy, Bruce D. and J. Spencer Martin, 2001, "Understanding the Nature of the Risks and the Source of the Rewards to Momentum Investing," Review of Financial Studies, 14, 29–78. Daniel, Kent,David Hirshleifer, Avanidhar Subrahmanyam, 1998, "Investor Psychology and Security Market Under- and Overreactions," Journal of Finance, 53, 1839–1885. Fama, Eugene F., and Kenneth R. French, 1992, "The cross-section of expected stock returns," Journal of Finance, 46, 427–466. Fama, Eugene F., and Kenneth R. French, 1993, "Common risk factors in the returns of stocks and bonds," Journal of Financial Economics, 33, 3–56. Fama, Eugene F., and Kenneth R. French, 1996, "Multifactor Explanations of Asset Pricing Anomalies," Journal of Finance, 51, 55–84. Fama, Eugene F., and Kenneth R. French, 1998, "Value versus Growth: International Evidence," Journal of Finance, 53, 1975–1999. Fama, Eugene F., and Kenneth R. French, 2000, "Characteristics, covariances, and average returns: 1929– 1997," Journal of Finance, 55, 389–406. Fama, Eugene F., and Kenneth R. French, 2010, " Luck versus skill in the cross-section of mutual fund returns," Journal of Finance, 65, 1915–1947. Fama, Eugene F., and Kenneth R. French, 2012, "Size, Value, and Momentum in International Stock Returns," Journal of Financial Economics, 105, 457–472. Francis, Bill B., Iftekhar Hasana, and Delroy M. Hunter, 2008, "Can hedging tell the full story? Reconciling differences in United States aggregate- and industry-level exchange rate risk premium," Journal of Financial Economics, 90, 169–196. Hong, Harrison and Jeremy C. Stein, 1999, "A Unified Theory of Underreaction, Momentum Trading, and Overreaction in Asset Markets," Journal of Finance, 54, 2143–2184. Hong, Harrison, Terence Lim, and Jeremy C. Stein, 2000, "Bad News Travels Slowly: Size, Analyst Coverage, and the Profitability of Momentum Strategies," 55, 265–295. Hou, Kewei, Bong-Chan Kho, and Andrew Karolyi, 2011, “What Factors Drive Global Stock Returns,” Review of Financial Studies, 24, 2527–2574.

25 Page 26 of 47

Israel, Ronen and Tobias J. Moskowitz, 2013, "The Role of Shorting, Firm Size, and Time on Market Anomalies," Journal of Financial Economics, 2013, 108, 275–301. Jegadeesh, Narasimhan, and Sheridan Titman, 1993, "Returns to buying winners and selling losers: Implications for stock market efficiency," Journal of Finance, 48, 65–91. Jegadeesh, Narasimhan, and Sheridan Titman, 2001, "Profitability of Momentum Strategies: An Evidence of Alternative Explanations," Journal of Finance, 56, 699–720. Jong, Frank de and Frans A. de Roon, 2005, "Time-Varying Market Integration and Expected Returns in Emerging Markets," Journal of Financial Economics, 78, 583–613. Lakonishok, Josef, Andrei Shleifer and Robert W. Vishny, 1994, “Contrarian investment, extrapolation, and risk,” Journal of Finance 49, 1541–1578. LaPorta, Rafael, 1996, "Expectations and the Cross-Section of Stock Returns," Journal of Finance, 51, 1715–1742. LaPorta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer, and Robert W Vishny, 1998, “Law and Finance,” Journal of Political Economy, 106, 1113–1155. Liu, Laura Xiaolei and Lu Zhang, 2008, "Momentum Profits, Factor Pricing, and Macroeconomic Risk," Review of Financial Studies, 21, 2417–2448. Newey, Whitney K., and Kenneth D. West, 1987, "A simple, positive-definite, heteroscedasticity and autocorrelation consistent covariance matrix," Econometrica, 55, 703–708. Rouwenhorst, 1998, "International Momentum Strategies," Journal of Finance, 53, 267–284. Rouwenhorst, 1999, " Local return factors and turnover in emerging stock markets," Journal of Finance, 54, 1439–1464. Stattman, Dennis, 1980, “Book values and stock returns”, The Chicago MBA: A Journal of Selected Papers. 4, 25–45.

26 Page 27 of 47

Table 1. Average size, book-to-market equity, number of stocks, and the market and size factors We calculate the cross-sectional medians of stock size (SIZE) measured in millions of dollars and book-to-market equity ratio (BM) at the beginning of each month over the period of January 1990 to December 2013. We divide stocks into small (S) and big (B) portfolios. This table reports the time-series averages of the cross-sectional medians, the number of stocks each month (N), the monthly excess returns on the valueweighted market portfolio (MKT), the monthly returns on the value-weighted small and big portfolios, and the monthly size-related factor (SMB), defined as the return difference between S and B. The t-statistics (t-stat) are calculated using the Newey-West (1987) procedure with six lags. All returns are in U.S. dollars. SIZE Asia China India Indonesia South Korea Malaysia Philippines Taiwan Thailand Latin America Argentina Brazil Chile Columbia Mexico Europe Hungary Poland Turkey Czech Russia

BM

N Mean

MKT t-stat

Mean

S t-stat

Mean

B t-stat

Mean

SMB t-stat

345.30 5.00 89.50 75.90 78.30 44.70 211.90 50.80

0.34 0.83 0.72 1.37 1.08 1.06 0.78 0.98

1,077 1,143 157 900 655 187 585 388

0.86 0.68 1.84 0.81 0.51 0.37 0.36 0.43

1.24 1.02 2.14 1.06 0.75 0.56 0.60 0.58

2.04 0.54 1.59 0.88 0.46 0.51 0.64 0.58

2.58 0.72 2.13 1.02 0.55 0.71 0.94 0.81

0.98 1.56 2.40 1.18 0.92 0.74 0.63 0.77

1.39 1.82 2.69 1.50 1.35 1.04 1.07 1.00

1.06 –1.01 –0.82 –0.30 –0.46 –0.23 0.01 –0.19

3.13 –1.86 –1.69 –0.86 –1.41 –0.83 0.03 –0.56

113.40 167.10 171.70 318.70 387.90

1.15 0.43 0.81 1.19 0.82

63 405 168 37 138

0.61 0.41 0.58 0.90 0.81

0.91 0.58 1.22 1.55 1.49

0.17 0.90 0.84 0.92 0.73

0.21 1.27 1.78 1.88 1.28

0.87 0.64 0.82 1.17 1.12

1.22 0.94 1.73 1.97 2.05

–0.71 0.26 0.02 –0.25 –0.39

–1.83 0.99 0.13 –0.68 –1.74

45.00 57.60 84.00 109.50 203.20

1.02 0.81 0.76 1.18 2.18

34 179 204 37 136

0.90 0.73 1.72 0.80 1.65

1.28 1.15 1.89 1.44 1.30

0.25 0.88 1.95 0.89 1.26

0.44 1.12 2.04 1.80 1.35

1.27 1.23 2.10 0.99 1.56

1.87 1.93 2.32 1.89 1.61

–1.02 –0.34 –0.15 –0.10 –0.30

–2.69 –0.75 –0.41 –0.27 –0.68

27 Page 28 of 47

Table 2. Value factors At the beginning of each month over the period January 1990–December 2013, we first group the largest market capitalization stocks into the big portfolio (B), which accounts for 90% of the total market capitalization. All remaining stocks are grouped into the small portfolio (S). Within the big portfolio, we determine the bottom 30% (L), middle 40% (M), and top 30% (H) breakpoints for the BM ratio and apply these BM breakpoints to both the big and small stocks. These classifications allow us to form six value-weighted portfolios: S/L, S/M, S/H, B/L, B/M, and B/H. We measure the value factor (HML) as the average of the monthly returns on S/H and B/H portfolios minus the average of the monthly returns on S/L and B/L portfolios. HML_S, HML_B, and HML_S–B, respectively, denote the monthly return difference between the high and low BM portfolios within the small and big portfolios, as well as the difference between HML_S and HML_B. This table reports the time-series averages of HML_S, HML_B, HML_S–B, and HML, and the corresponding Newey-West adjusted t-statistics. The last column reports the annualized Sharpe ratio of the HML factor in each market. All returns are in U.S. dollars. HML_S Mean t-stat Asia China India Indonesia South Korea Malaysia Philippines Taiwan Thailand Latin America Argentina Brazil Chile Columbia Mexico Europe Hungary Poland Turkey Czech Russia

HML_B Mean t-stat

HML_S–B Mean t-stat

Mean

t-stat

HML Sharpe ratio

1.108 1.223 2.299 2.013 1.347 1.462 1.028 2.282

3.04 3.41 4.65 4.19 4.50 3.25 2.70 5.99

0.916 1.235 2.380 0.666 0.725 0.848 0.347 0.892

2.67 1.39 3.28 1.17 2.62 1.60 0.93 1.82

0.192 –0.012 –0.082 1.346 0.622 0.614 0.681 1.390

0.51 –0.01 –0.10 2.53 1.67 0.87 1.95 2.40

1.012 1.229 2.339 1.339 1.036 1.155 0.688 1.587

3.38 2.53 4.91 2.96 4.70 3.38 2.06 4.80

0.73 0.53 0.92 0.76 1.05 0.76 0.42 1.22

1.453 0.553 0.744 1.365 1.137

1.95 1.50 2.40 1.77 2.84

1.526 0.268 0.337 1.256 0.522

2.60 0.41 1.34 2.46 1.37

–0.073 0.284 0.408 0.109 0.614

–0.08 0.44 1.24 0.14 1.20

1.489 0.411 0.540 1.311 0.829

2.93 0.97 2.36 2.48 2.79

0.73 0.22 0.61 0.67 0.60

1.700 0.803 1.313 1.727 0.702

2.85 2.09 2.79 2.88 1.09

1.172 2.352 1.333 0.728 2.270

1.64 3.19 2.34 1.26 1.78

0.528 –1.549 –0.020 0.999 –1.567

0.53 –2.01 –0.04 1.18 –1.09

1.436 1.577 1.323 1.227 1.486

3.31 3.55 3.00 2.99 2.10

0.72 0.76 0.81 0.81 0.58 28 Page 29 of 47

Table 3. Momentum factors At the beginning of each month over the period January 1991–December 2013, we first group the largest market capitalization stocks into the big portfolio (B), which accounts for 90% of the total market capitalization. All remaining stocks are grouped into the small portfolio (S). Within the big portfolio, we determine the bottom 30% (L), middle 40% (M), and top 30% (H) breakpoints for firm momentum (MOM) and apply these MOM breakpoints to both the big and small stocks. These classifications allow us to form six value-weighted portfolios: S/L, S/M, S/H, B/L, B/M, and B/H. We measure the momentum factor (UMD) as the average of the monthly returns on S/H and B/H portfolios minus the average of the monthly returns on S/L and B/L portfolios. UMD_S, UMD_B, and UMD_S–B, respectively, denote the monthly return difference between the high and low MOM portfolios within the small and big portfolios, as well as the difference between UMD_S and UMD_B. This table reports the time-series averages of UMD_S, UMD_B, UMD_S–B, and UMD, and the corresponding Newey-West adjusted t-statistics. The last column reports the annualized Sharpe ratio of the UMD factor in each market. All returns are in U.S. dollars.

Asia China India Indonesia South Korea Malaysia Philippines Taiwan Thailand Latin America Argentina Brazil Chile Columbia Mexico Europe Hungary Poland Turkey Czech Russia

UMD_S Mean t-stat

UMD_B Mean t-stat

UMD_S–B Mean t-stat

Mean

UMD t-stat

0.136 1.300 –1.125 0.274 0.049 0.130 0.237 0.431

0.39 2.22 –1.87 0.62 0.10 0.21 0.53 0.73

0.061 2.682 –1.237 0.653 0.044 –0.037 0.367 0.327

0.16 2.86 –1.71 1.14 0.10 –0.06 0.80 0.49

0.075 –1.382 0.112 –0.378 0.004 0.167 –0.131 0.104

0.20 –1.87 0.15 –0.72 0.01 0.28 –0.42 0.20

0.099 1.991 –1.181 0.464 0.046 0.047 0.302 0.379

0.31 2.89 –2.15 1.06 0.11 0.09 0.71 0.66

0.07 0.76 –0.49 0.24 0.02 0.02 0.18 0.15

1.042 0.679 1.212 –0.213 1.165

1.72 1.41 3.91 –0.37 2.35

0.143 0.785 0.273 –0.038 0.303

0.22 1.18 0.85 –0.06 0.77

0.899 –0.106 0.939 –0.175 0.861

1.26 –0.18 2.86 –0.23 1.98

0.592 0.732 0.743 –0.125 0.734

1.14 1.47 2.75 –0.27 1.88

0.26 0.39 0.68 –0.06 0.48

1.009 0.781 –0.503 0.418 –1.412

1.80 1.41 –0.93 0.66 –1.16

0.690 0.607 –0.419 0.412 0.484

0.89 0.80 –0.77 0.59 0.43

0.319 0.174 –0.084 0.006 –1.896

0.41 0.21 –0.13 0.01 –1.74

0.850 0.694 –0.461 0.415 –0.464

1.54 1.33 –1.08 0.83 –0.45

0.40 0.31 –0.25 0.22 –0.14

Sharpe ratio

29 Page 30 of 47

Table 4. Intra-market correlations between the value and momentum factors Panel A reports the intra-market correlation coefficients between the value and momentum factors calculated for all stocks (HML and UMD), small-cap stocks (HML_S and UMD_S), as well as big-cap stocks (HML_B and UMD_B). The last three columns in Panel A, respectively, report the mean, tstatistic, and annualized Sharpe ratio of a combination strategy that gives equal weight to the HML and UMD factors. Panel B reports the averages of the intra-market correlation coefficients between HML and UMD averaged across all markets within the same region as well as the average correlation coefficients during up (U.S. equity market return positive) and down (U.S. equity market return negative) markets. The t-statistics are calculated using the Newey-West (1987) procedure with six lags. The sample period is from January 1991 to December 2013. Panel A. Intra-market correlations HML_S and UMD_S Corr. coef. t-stat Asia China –0.04 –0.22 India –0.06 –0.43 Indonesia –3.48 –0.22 South Korea 0.01 0.02 Malaysia –0.30 –1.40 Philippines –0.26 –1.69 Taiwan –2.29 –0.26 Thailand 0.11 0.99 Latin America Argentina –3.06 –0.17 Brazil –2.27 –0.20 Chile –0.08 –1.02 Columbia –0.12 –1.66 Mexico 0.02 0.19 Europe Hungary 0.05 0.65 Poland –2.24 –0.19 Turkey –0.08 –0.66 Czech 2.70 0.25 Russia –0.18 –1.54

HML_B and UMD_B Corr. coef. t-stat

HML and UMD Corr. coef. t-stat

Mean

0.5×(HML+UMD) t-stat Sharpe ratio

–0.28 –0.32 –0.05 –0.41 –0.42 –0.27 –0.30 –0.45

–2.13 –1.60 –0.44 –2.58 –4.66 –1.80 –2.35 –3.07

–0.25 –0.32 –0.16 –0.30 –0.44 –0.32 –0.33 –0.29

–1.68 –1.87 –1.67 –1.82 –4.64 –1.91 –2.78 –2.51

0.56 1.61 0.58 0.90 0.54 0.60 0.49 0.98

2.67 3.91 1.66 3.99 2.52 2.15 2.26 3.62

0.63 1.11 0.36 0.82 0.58 0.49 0.52 0.80

–0.18 0.05 –0.16 –0.25 –0.33

–1.42 0.27 –1.63 –2.30 –3.64

–0.26 –0.05 –0.12 –0.18 –0.18

–3.83 –0.25 –1.60 –1.79 –1.74

1.04 0.57 0.64 0.59 0.78

3.01 1.89 3.68 2.03 3.83

0.79 0.44 0.97 0.46 0.84

0.04 –0.24 –0.15 0.01 –0.17

0.34 –0.76 –2.24 0.05 –0.90

–0.03 –0.28 –0.24 0.18 –0.36

–0.33 –1.14 –2.56 1.78 –3.64

1.14 1.14 0.43 0.82 0.51

3.09 4.17 1.93 2.20 1.19

0.79 0.88 0.40 0.63 0.31

Panel B. Average Intra-market correlations across individual regions All months Asia –0.13 –0.31 Latin America –0.11 –0.17 Europe –0.03 –0.10

–0.30 –0.16 –0.15

Up market Asia Latin America Europe

–0.13 –0.17 0.01

–0.37 –0.18 –0.19

–0.34 –0.18 –0.20

Down market Asia Latin America Europe

–0.14 –0.03 –0.07

–0.19 –0.13 0.06

–0.20 –0.06 0.00

30 Page 31 of 47

Table 5. Inter-market correlations between the value and momentum factors This table reports the inter-market correlation coefficients between the value (HML) and momentum (UMD) factors. Numbers in bold indicate significance at the 5% level or better. The sample period is from January 1991 to December 2013.

Momentum factor (1) China (2) India (3) Indonesia (4) SouthKorea (5) Malaysia (6) Philippines (7) Taiwan (8) Thailand (9) Argentina (10) Brazil (11) Chile (12) Columbia (13) Mexico (14) Hungary (15) Poland (16) Turkey (17) Czech (18) Russia

Asia (1) –0.25 –0.01 –0.05 0.09 –0.08 0.10 0.03 0.09 0.12 0.09 0.08 –0.04 –0.08 –0.03 0.00 –0.03 0.08 –0.04

(2)

(3)

(4)

(5)

(6)

(7)

(8)

0.09 –0.32 0.05 –0.08 0.08 0.00 –0.14 0.02 0.09 –0.04 0.00 –0.01 –0.15

0.10 –0.01 –0.16 –0.03 –0.03 0.11 –0.10 0.05 0.05 –0.10 –0.12 –0.10 –0.04 –0.04 0.02 –0.09 –0.08 –0.01

–0.06 –0.15 –0.15 –0.30 –0.01 –0.16 –0.19 0.03 0.03 –0.17 –0.08 –0.10 –0.09 –0.03 0.03 –0.02

–0.11 –0.13 –0.14 –0.07

0.01 –0.11 –0.17

–0.03

–0.20 –0.17

–0.11 –0.04 –0.03 0.15 –0.06

0.18 –0.07

–0.44 –0.14 0.12 –0.15 –0.09 –0.04 –0.03 –0.24 –0.07 –0.10 –0.12 0.07 –0.09 –0.09

–0.12 –0.29 –0.32 –0.13 –0.32 –0.02 –0.16 –0.18 –0.13 –0.08 0.05 –0.07 –0.13 –0.07 0.05

–0.19 –0.01 –0.06 0.12 0.01 –0.33 0.13 0.00 –0.06 0.04 –0.02 –0.04 –0.01 0.01 0.07 0.12 –0.07

–0.25 –0.22 –0.24 –0.25 –0.13 –0.29 –0.04 –0.18 –0.19 –0.24 –0.02 –0.07 –0.06 –0.07 –0.08 –0.40

Value factor Latin America (9) (10) –0.01 0.03 –0.02 0.01 –0.03 –0.10 –0.07 –0.05

0.06 –0.07 –0.17 –0.17 –0.08 –0.08 –0.01 –0.11

–0.26 –0.03 –0.01 –0.02 –0.19 0.00 –0.15 –0.09 –0.08 –0.06

–0.13 –0.05 –0.22 –0.39 –0.01 –0.17 0.01 0.02 –0.01 –0.21

(11)

(12)

(13)

0.08 0.05 –0.11 –0.04 –0.04 –0.14 –0.05 0.01 –0.02

–0.06 0.09 –0.27 0.03 –0.11 –0.03 0.00 0.03 –0.11 –0.09 –0.15 –0.18 –0.03 0.02 0.03 –0.12

–0.15 –0.23 0.15 –0.02 0.03 0.06 –0.11 0.00 0.14 0.10 0.05 0.11 –0.18 0.07 0.10 –0.03

0.14 –0.09

0.22 0.01

–0.19 –0.12 0.00 –0.01 0.05 0.17 –0.14 –0.09 –0.15

Europe (14) (15) –0.02 –0.05 0.00 0.06 –0.01 –0.10 0.04 0.01 –0.02 –0.04 –0.02 0.00 –0.07 –0.03 –0.06 –0.14 –0.04 –0.04

–0.03 –0.07 –0.02 0.00 –0.14 –0.12 –0.06 0.05 –0.06 –0.10 –0.01 –0.01 –0.23 0.04 –0.28 –0.11 0.10 0.00

(16)

(17)

(18)

0.01 0.03 –0.06 –0.03 –0.13 –0.16 0.00 –0.13

0.10 0.05 0.07 –0.07 –0.07 0.02 0.09 0.03 –0.05 0.08 0.01 –0.06 0.07 0.00 –0.10 0.09

–0.02 –0.02 –0.03 –0.04 –0.11 –0.06 –0.05 –0.19 –0.03 0.05 0.04 –0.08 –0.02 0.01 –0.04 –0.05 0.14

–0.16 –0.33 –0.11 –0.08 –0.17 –0.09 –0.24 –0.24 –0.11 –0.19

0.18 –0.01

–0.36

31 Page 32 of 47

Table 6. Inter-market correlations between the value and momentum factors: Small versus big stocks and up versus down markets Panel A reports the averages of the inter-market correlation coefficients between the value and momentum factors calculated for all stocks (HML and UMD), small-cap stocks (HML_S and UMD_S), and big-cap stocks (HML_B and UMD_B), averaged across all markets within the same region. Panel B and C report the same set of statistics calculated for up (U.S. equity market return positive) and down (U.S. equity market return negative) markets, respectively. Panel A. All months All stocks Small stocks Big stocks

Asia –0.06 –0.02 –0.07

Latin America –0.07 –0.03 –0.05

Europe –0.04 –0.01 –0.02

Asia –0.08 –0.03 –0.08

Latin America –0.09 –0.04 –0.07

Europe –0.05 –0.01 –0.03

Asia –0.02 –0.01 –0.02

Latin America 0.01 0.02 0.00

Europe –0.01 –0.01 0.00

Panel B. Up market All stocks Small stocks Big stocks Panel C. Down market All stocks Small stocks Big stocks

32 Page 33 of 47

Table 7. Correlations between the value factors across markets This table reports the inter-market correlation coefficients between the value factors. Numbers in bold indicate significance at the 5% level or better. The sample period is from January 1990 to December 2013. Asia (1) China (2) India (3) Indonesia (4) SouthKorea (5) Malaysia (6) Philippines (7) Taiwan (8) Thailand (9) Argentina (10) Brazil (11) Chile (12) Columbia (13) Mexico (14) Hungary (15) Poland (16) Turkey (17) Czech (18) Russia

Latin America

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

1.00

–0.04

–0.01

–0.06

–0.02

–0.01

0.17

0.19 0.00

–0.03

1.00

0.11

0.24

1.00

–0.04

–0.05

0.03

1.00

–0.08

0.04

1.00

Europe

(9)

(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17)

(18)

0.14

0.05

–0.04

–0.04

0.07

0.14

0.05

0.05

0.06

–0.13

0.14

0.05

0.00

0.32

0.05

0.18 0.07

–0.03

0.01

–0.13

0.08

–0.02

0.06

–0.07

–0.06

0.29

0.15

0.12

0.11

0.10

0.00

0.13

–0.04

–0.11

0.30

0.16

0.01

0.08

0.11

0.16

0.04

0.09

0.06

0.03

0.12

0.19

–0.23

0.32

–0.04

0.05

–0.08

–0.03

0.19 –0.09

0.08

0.00

0.09

0.12

0.14

1.00

–0.04

0.20

0.03

0.01

0.09

0.05

0.07

0.07

0.06

0.03

0.07

–0.17

1.00

0.01

0.08

0.08

0.08

0.09

0.20

0.01

0.02

–0.14

–0.07

0.02

1.00

–0.02

0.11

–0.04

0.05

0.00

0.03

0.03

0.16

–0.05

0.36

1.00

–0.03

–0.02

0.01

–0.11

0.03

–0.04

–0.04

1.00

0.13

0.22

0.05

0.17

0.20 0.00

0.13 0.04

0.04

0.07

1.00

0.15 1.00

0.02

0.11

0.00

0.09

0.01

0.01

–0.03

0.07

0.10

0.07

0.15

–0.05

1.00

0.02

–0.02

–0.06

0.04

–0.06

1.00

0.16

0.09

–0.05

0.10

1.00

0.23

0.02

–0.17

1.00

–0.07

0.09

1.00

–0.10 1.00

33 Page 34 of 47

Table 8. Correlations between the momentum factors across countries This table reports the inter-market correlation coefficients between the momentum factors. Numbers in bold indicate significance at the 5% level or better. The sample period is from January 1991 to December 2013. Asia (1) (1) China (2) India (3) Indonesia (4) SouthKorea (5) Malaysia (6) Philippines (7) Taiwan (8) Thailand (9) Argentina (10) Brazil (11) Chile (12) Columbia (13) Mexico (14) Hungary (15) Poland (16) Turkey (17) Czech (18) Russia

1.00

Latin America (9) (10)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

0.18 1.00

0.08 0.06 1.00

0.02 0.17 0.09 1.00

0.11 0.16 0.25 0.17 1.00

0.06 0.09 0.18

0.02 0.21 0.02

0.15 0.19 0.15

0.04 0.04 0.03

0.13 0.20 0.13

0.22 0.39 1.00

0.35 0.01 0.10 1.00

0.30 0.37 0.56 0.07 1.00

0.26 0.07

0.23 0.09

0.26 0.09 0.27 1.00

0.22 0.19 0.18 0.23 1.00

(11)

(12)

(13)

Europe (14)

0.08 0.12

0.03 0.03

0.21 0.26 0.15

0.31 0.22 0.18

0.10 0.32 0.08

–0.01 0.10 0.09

0.20 0.20 0.24 0.21 0.40 1.00

0.18 0.00 0.21 0.18 0.15 0.21 1.00

0.23 0.16 0.21

0.12 0.08 0.01 0.01 0.11 –0.02 0.00 0.03 0.10

0.20 0.18 0.16 0.25 0.11 –0.01 1.00

0.15 1.00

(15)

(16)

(17)

(18)

0.13 0.22

–0.03 0.13 0.13

0.07 0.03 0.06 0.07 0.08 0.02 –0.05 0.12 0.13

–0.03 0.12 0.11

0.15 0.05 0.32 0.11 0.02 0.02 0.10 0.12 0.02 –0.01 0.26 0.17 1.00

0.13 0.07 0.11 0.16 0.24 0.12 0.20 0.15 –0.03 0.24 –0.06 0.17 1.00

0.15 0.15 0.08 0.05 0.26 0.19 0.10 1.00

0.31 0.19 0.24 0.17 0.39 0.25 0.27 0.22 0.19 0.32 0.20 0.06 0.25 0.09 1.00

34 Page 35 of 47

Table 9. Alphas and betas of value factors relative to the market, regional, global, and U.S. CAPMs For each market, we run monthly time-series regressions of the value factors (HML) separately on returns of the market, returns of the region in which the market resides, returns of a value-weighted global portfolio that consists of the developed markets studied in Fama and French (2012), and returns of the U.S. equity market. This table reports the intercepts (α), the slope coefficients (β), and the corresponding t-statistics (t-stat) from the regressions. The last two columns report the intercept (α) and the t-statistic (t-stat) from regressing HML on returns of the market, the region, the globe, and the U.S. markets. The t-statistics are calculated using the Newey-West (1987) procedure with six lags. The sample period is from January 1991 to December 2013. Market CAPM

Regional CAPM

α%

t-stat

β

t-stat

α%

t-stat

China

0.99

3.60

0.03

0.46

0.96

3.17

India

1.16

2.56

0.10

1.64

1.30

2.69

Indonesia

2.06

4.22

0.15

2.01

2.30

4.88

South Korea

1.37

3.21

–0.04

–0.41

Malaysia

0.98

4.71

0.11

Philippines

1.08

3.83

0.20

Taiwan

0.69

2.13

–0.02

Thailand

1.55

4.99

β

Global CAPM

U.S. CAPM

t-stat

α%

t-stat

β

t-stat

α%

t-stat

0.07

1.19

1.02

3.37

–0.02

–0.27

1.00

3.32

–0.10

–0.75

1.26

2.74

–0.05

–0.29

1.29

2.82

0.05

0.46

2.20

4.74

0.27

2.23

2.17

4.71

1.31

3.02

0.05

0.66

1.37

3.20

–0.06

–0.47

1.83

0.92

4.55

0.16

3.66

0.95

4.53

0.17

2.71

1.02

3.59

0.19

2.56

1.07

3.57

0.16

–0.22

0.71

2.16

–0.03

–0.42

0.77

2.35

–0.16

0.08

2.16

1.50

4.87

0.12

1.86

1.55

4.87

β

Altogether t-stat

α%

t-stat

0.01

0.27

0.90

3.38

–0.10

–0.54

1.32

2.84

0.26

2.46

2.03

4.51

1.39

3.29

–0.08

–0.64

1.36

3.36

3.09

0.94

4.40

0.15

2.57

0.93

4.45

1.77

1.06

3.46

0.14

1.80

1.08

3.64

–1.36

0.83

2.52

–0.21

–1.74

0.83

2.60

0.07

0.84

1.55

4.84

0.05

0.71

1.54

5.12

Asia

Latin America Argentina

1.35

2.81

0.24

3.75

1.36

2.82

0.24

4.09

1.35

2.65

0.26

2.97

1.33

2.57

0.23

2.61

1.35

2.67

Brazil

0.34

0.89

0.16

3.20

0.33

0.84

0.16

2.59

0.37

0.91

0.07

0.89

0.43

1.07

–0.03

–0.33

0.58

1.64

Chile

0.53

2.40

0.03

0.49

0.54

2.39

0.01

0.16

0.56

2.53

–0.04

–0.54

0.57

2.60

–0.05

–0.75

0.57

2.65

Columbia

1.03

2.12

0.31

3.23

1.23

2.36

0.16

2.38

1.30

2.45

0.03

0.31

1.29

2.43

0.04

0.47

1.04

2.26

Mexico

0.81

2.86

0.03

0.51

0.83

2.86

0.00

–0.09

0.87

2.94

–0.08

–0.93

0.89

2.97

–0.09

–1.19

0.88

2.97

Hungary

1.37

3.18

0.08

1.27

1.32

2.93

0.19

1.33

1.32

2.94

0.22

1.33

1.31

2.88

0.19

1.28

1.31

2.84

Poland

1.58

3.74

0.00

–0.02

1.59

3.68

–0.02

–0.32

1.61

3.75

–0.06

–0.76

1.65

3.86

–0.11

–1.24

1.68

3.86

Turkey

1.03

2.79

0.17

7.64

1.14

2.76

0.29

3.76

1.12

2.75

0.39

4.48

1.07

2.69

0.38

4.71

1.01

2.87

Czech

1.19

3.03

0.05

0.76

1.26

3.19

–0.05

–0.78

1.27

3.19

–0.08

–0.99

1.28

3.23

–0.10

–1.09

1.25

3.19

Russia

1.54

2.02

–0.03

–0.50

1.39

1.87

0.16

1.18

1.38

1.87

0.21

1.17

1.37

1.86

0.21

1.15

1.47

1.96

Europe

35 Page 36 of 47

Table 10. Alphas and betas of momentum factors with respect to local, regional, global, and U.S. CAPMs For each market, we run monthly time-series regressions of the momentum factors (UMD) separately on returns of the market, returns of the region in which the market resides, returns of a value-weighted global portfolio that consists of the developed markets studied in Fama and French (2012), and returns of the U.S. equity market. This table reports the intercepts (α), the slope coefficients (β), and the corresponding t-statistics (tstat) from the regressions. The last two columns report the intercept (α) and the t-statistic (t-stat) from regressing UMD on returns of the market, the region, the globe, and the U.S. markets. The t-statistics are calculated using the Newey-West (1987) procedure with six lags. The sample period is from January 1991 to December 2013. Market CAPM

Regional CAPM

Global CAPM

U.S. CAPM

Altogether

α%

t-stat

β

t-stat

α%

t-stat

β

t-stat

α%

t-stat

β

t-stat

α%

t-stat

β

t-stat

α%

t-stat

China

0.11

0.36

–0.01

–0.20

0.13

0.42

–0.05

–1.06

0.11

0.34

–0.01

–0.19

0.12

0.37

–0.03

–0.42

0.16

0.53

India

2.06

3.20

–0.11

–0.71

2.04

3.09

–0.07

–0.33

2.01

3.06

–0.03

–0.13

1.96

2.92

0.04

0.18

1.95

2.80

Indonesia

–1.05

–2.01

–0.07

–1.05

–1.16

–2.23

–0.03

–0.24

–1.20

–2.25

0.03

0.22

–1.20

–2.24

0.02

0.18

–1.05

–2.02

South Korea

0.48

1.23

–0.02

–0.26

0.53

1.29

–0.10

–1.12

0.51

1.24

–0.09

–0.70

0.52

1.26

–0.08

–0.64

0.53

1.31

Malaysia

0.26

0.56

–0.41

–2.49

0.25

0.64

–0.29

–2.35

0.17

0.42

–0.23

–1.93

0.22

0.59

–0.26

–2.18

0.31

0.68

Philippines

0.23

0.47

–0.50

–2.82

0.34

0.81

–0.43

–2.60

0.28

0.63

–0.45

–2.40

0.31

0.69

–0.40

–2.45

0.22

0.43

Taiwan

0.31

0.76

–0.03

–0.33

0.31

0.78

–0.02

–0.18

0.28

0.71

0.04

0.38

0.27

0.69

0.04

0.45

0.28

0.70

Thailand

0.54

1.12

–0.38

–2.94

0.63

1.31

–0.37

–2.30

0.60

1.18

–0.43

–2.37

0.64

1.26

–0.39

–2.31

0.51

1.00

Argentina

0.76

1.54

–0.27

–4.34

0.75

1.60

–0.31

–4.26

0.79

1.60

–0.39

–3.35

0.81

1.57

–0.32

–2.84

0.75

1.53

Brazil

0.79

1.60

–0.14

–2.12

0.81

1.66

–0.16

–1.68

0.91

1.89

–0.34

–2.85

0.92

1.99

–0.29

–2.65

0.82

1.73

Chile

0.84

3.27

–0.17

–2.79

0.78

3.04

–0.08

–1.80

0.79

3.08

–0.10

–1.39

0.80

3.06

–0.08

–1.28

0.84

3.09

Columbia

–0.01

–0.02

–0.13

–1.22

–0.09

–0.19

–0.07

–1.11

–0.09

–0.20

–0.06

–0.66

–0.08

–0.16

–0.08

–0.83

0.03

0.07

Mexico

0.88

2.54

–0.18

–2.15

0.83

2.37

–0.19

–2.64

0.85

2.49

–0.23

–1.94

0.85

2.44

–0.17

–1.53

0.80

2.29

Hungary

0.96

1.76

–0.12

–1.62

0.95

1.70

–0.16

–1.49

0.96

1.70

–0.20

–1.57

0.96

1.72

–0.17

–1.28

0.90

1.72

Poland

0.84

1.77

–0.20

–2.33

0.82

1.75

–0.20

–1.89

0.83

1.74

–0.26

–2.05

0.87

1.83

–0.26

–2.11

0.85

1.71 –0.78

Asia

Latin America

Europe

Turkey

–0.27

–0.69

–0.11

–2.43

–0.36

–0.87

–0.15

–1.90

–0.35

–0.84

–0.21

–2.16

–0.33

–0.80

–0.20

–2.06

–0.30

Czech

0.51

1.08

–0.12

–1.40

0.50

1.00

–0.13

–1.17

0.49

1.00

–0.16

–1.33

0.51

1.03

–0.17

–1.50

0.56

1.15

Russia

–0.23

–0.22

–0.14

–2.19

–0.27

–0.27

–0.34

–2.04

–0.21

–0.22

–0.52

–2.25

–0.18

–0.19

–0.53

–2.26

–0.20

–0.21

36 Page 37 of 47

Table 11. Value factor loadings on future GDP growth and the market return For each market, we run annual time-series regressions of the value factor (HML) separately on 1) the future GDP growth rate and the contemporaneous return of the market, 2) the future GDP growth rate and the contemporaneous return of the U.S. market, 3) the future GDP growth rate and the contemporaneous return of the global market. This table reports the slopes on GDP growth rate ( ), the market return 2 (MKT–RF), the corresponding t-statistics (t-stat), and R-squared (R ) from the regressions. The t-statistics are calculated using the Newey-West (1987) procedure with six lags. The sample period is from January 1990 to December 2013. Local market

U.S. market

GDP

t-stat

MKT-RF

t-stat

R

–0.11

–0.10

0.13

1.42

2

Global market GDP

t-stat

MKT-RF

t-stat

R2

0.09

0.67

1.15

0.11

1.34

0.02

GDP

t-stat

MKT-RF

t-stat

R

0.16

1.58

1.26

0.09

0.78

2

Asia China India

1.51

0.76

0.33

3.12

0.42

–0.79

–0.51

0.59

2.14

0.17

3.76

1.65

0.82

2.88

0.35

Indonesia

–0.80

–1.40

0.03

0.26

0.06

3.59

2.07

–0.11

–0.40

0.25

2.88

1.25

–0.09

–0.29

0.09

South Korea

0.05

0.04

0.03

0.22

0.01

–2.09

–2.93

–0.23

–1.00

0.22

–0.85

–0.58

–0.08

–0.37

0.02

Malaysia

1.07

1.70

0.17

3.14

0.35

–2.10

–2.83

0.20

2.61

0.45

–1.95

–3.69

0.20

4.08

0.33

Philippines

0.26

0.28

0.38

5.70

0.64

–1.38

–1.42

0.52

3.55

0.31

–1.11

–0.84

0.55

4.91

0.37

Taiwan

–0.06

–0.10

–0.07

–0.37

0.01

1.26

1.02

–0.38

–1.21

0.09

1.86

0.86

–0.21

–0.86

0.05

Thailand

1.54

2.33

0.12

1.27

0.28

1.41

0.85

–0.09

–0.23

0.04

2.62

1.91

0.07

0.16

0.06

Argentina

1.22

1.20

0.07

0.41

0.08

2.67

1.02

0.07

0.13

0.04

1.11

0.44

0.10

0.17

0.00

Brazil

2.19

1.23

0.29

3.37

0.38

–2.77

–1.20

–0.09

–0.39

0.14

–0.22

–0.12

0.17

1.00

0.02

Chile

0.75

1.21

0.20

1.37

0.27

–0.55

–0.63

0.15

0.79

0.05

0.94

0.64

0.26

1.15

0.13

Columbia

0.57

0.39

0.17

1.28

0.03

1.71

0.82

–0.70

–1.29

0.09

2.81

1.42

–0.57

–1.03

0.09

Mexico

1.90

2.25

0.27

3.64

0.26

0.68

0.49

0.10

0.57

0.03

3.79

2.14

0.32

2.81

0.27

Hungary

1.20

1.14

0.20

2.31

0.19

1.85

1.40

0.35

1.54

0.18

2.96

1.44

0.38

1.69

0.15

Poland

4.89

2.12

–0.02

–0.07

0.16

5.12

2.70

0.53

2.73

0.34

4.92

2.44

0.52

1.83

0.15

Latin America

Europe

Turkey

1.83

3.98

0.18

3.21

0.33

0.51

0.29

0.20

0.88

0.03

2.16

1.10

0.34

1.68

0.08

Czech

1.85

2.42

0.48

3.58

0.48

–1.72

–1.82

0.29

0.70

0.07

3.01

1.07

0.50

1.05

0.16

Russia

1.89

1.47

–0.55

–2.09

0.45

0.41

0.14

–0.05

–0.14

0.00

1.89

0.65

–0.01

–0.03

0.01

37 Page 38 of 47

Table 12. Funding liquidity and credit risk exposure of value factors For each market, we run annual time-series regressions of the value factor (HML) on a funding liquidity variable or a credit risk variable after controlling for the future GDP growth rate and contemporaneous return of the U.S. market. The funding liquidity variables include the VIX, the TED spread for the U.S., and the U.S. LIBOR rate. The credit risk variables are the Aaa – T-note and Baa – Aaa spreads for the U.S. This table reports the regression slopes on the funding liquidity and credit risk variables (β) and the corresponding Newey-West adjusted t-statistics (t-stat). The sample period is from January 1990 to December 2013. VIX β Asia China India Indonesia South Korea Malaysia Philippines Taiwan Thailand Latin America Argentina Brazil Chile Columbia Mexico Europe Hungary Poland Turkey Czech Russia

TED t-stat

β

t-stat

LIBOR β t-stat

Aaa–T-note β t-stat

Baa–Aaa β t-stat

–0.02 –0.04 0.00 0.01 –0.56 –0.16 –1.39 3.23

–0.53 –0.78 0.08 0.12 –0.42 –0.21 –2.23 1.97

–0.05 0.02 0.01 –0.06 –0.43 –1.44 2.12 0.19

–0.70 0.34 0.16 –1.04 –0.36 –1.19 2.03 1.42

–0.06 0.10 0.00 0.19 –0.26 1.11 –0.30 –0.66

–0.78 2.36 0.13 2.34 –0.71 1.47 –0.45 –3.24

–0.05 0.03 0.05 0.05 –1.27 0.66 0.15 0.15

–0.95 0.49 0.63 0.10 –1.88 1.36 0.17 0.83

0.01 0.09 –0.07 –2.08 –0.72 –0.34 –1.12 –0.05

0.54 1.92 –1.12 –1.80 –0.93 –0.40 –1.13 –0.24

0.03 –0.13 0.16 0.64 0.38

0.27 –0.65 0.74 2.19 1.51

–0.17 0.00 0.02 –0.02 0.37

–1.16 –0.02 0.10 –0.06 1.85

–0.22 0.22 0.01 –0.10 0.40

–1.71 1.21 0.04 –0.37 1.11

–0.04 –0.25 0.46 –0.12 0.32

–0.27 –1.95 1.39 –0.85 1.22

–0.16 0.27 –0.20 0.11 0.16

–0.85 1.51 –1.22 0.45 1.06

–0.29 –0.25 0.43 0.12 1.45

–0.94 –0.94 0.50 0.14 1.60

0.43 –0.22 0.33 1.59 –0.13

2.21 –0.45 0.78 1.79 –0.13

0.10 –0.20 1.06 0.38 1.06

0.31 –0.35 1.10 0.89 1.13

0.03 0.81 1.14 –0.54 –1.23

0.12 0.56 1.29 –0.77 –1.29

0.33 0.51 1.51 1.10 1.58

1.33 0.46 1.64 1.61 1.31

38 Page 39 of 47

Table 13. Exposure of value factors to the liquidity factors of Pastor and Stambaugh 2003 and Sadka 2006 For each market, we run annual time-series regressions of the value factor (HML) on a liquidity variable after controlling for the future GDP growth rate and the contemporaneous return of the U.S. market. The three liquidity variables from Pastor and Stambaugh (2003) are the level, the innovation, and the traded factor variables. The two liquidity variables from Sadaka (2006) are the fixed and the variable components. This table reports the regression slopes on the liquidity variables (β) and the corresponding Newey-West adjusted t-statistics (t-stat). The sample period is from January 1990 to December 2013.

β Asia China India Indonesia South Korea Malaysia Philippines Taiwan Thailand Latin America Argentina Brazil Chile Columbia Mexico Europe Hungary Poland Turkey Czech Russia

PS level t-stat

PS innovation β t-stat

PS traded factor β t-stat

Sadka fixed factor β t-stat

Sadka variable factor β t-stat

0.05 0.05 0.01 0.04 –0.04 0.09 0.04 –0.02

1.10 0.85 0.20 0.74 –0.52 1.45 0.53 –0.26

0.19 –0.04 0.09 –0.09 0.09 0.04 0.06 0.04

1.98 –0.58 1.14 –1.89 1.37 0.59 0.85 0.78

–0.03 0.00 0.04 0.00 0.05 0.07 0.03 0.14

–0.43 0.10 0.60 –0.06 1.22 0.80 0.46 0.79

0.02 –0.03 0.04 0.04 0.05 0.01 0.09 0.11

0.33 –0.50 0.56 0.78 0.77 0.33 1.28 0.74

0.02 0.02 0.01 0.23 0.02 0.09 –0.10 –0.02

0.46 0.22 0.24 2.20 0.26 0.96 –1.90 –0.25

0.10 0.04 0.12 7.46 1.12

1.66 0.55 1.46 1.57 0.52

0.10 –0.01 0.04 4.04 5.39

1.22 –0.13 0.48 0.97 1.87

–0.05 0.13 –0.05 2.07 6.37

–0.38 1.17 –0.50 0.65 1.79

0.00 –0.01 0.18 0.39 4.85

0.01 –0.10 1.33 0.20 1.35

0.23 0.01 2.47 2.68 0.42

2.07 0.14 1.63 1.09 0.31

1.64 –2.42 0.77 0.75 –0.44

0.25 –0.69 1.06 1.03 –0.50

2.50 4.53 –0.13 0.82 1.06

1.03 1.21 –0.44 1.45 1.54

7.50 –0.82 –0.08 –0.68 0.21

1.76 –2.05 –0.18 –1.79 0.27

2.92 0.75 0.22 –1.53 –0.06

0.97 1.04 0.42 –2.41 –0.12

–0.48 –0.63 0.00 0.84 –0.67

–0.12 –0.67 –0.01 1.79 –0.75

39 Page 40 of 47

Table 14. Momentum factor loadings on future GDP growth and the market return For each market, we run annual time-series regressions of the momentum factor (UMD) separately on 1) the future GDP growth rate and the contemporaneous return of the market, 2) the future GDP growth rate and the contemporaneous return of the U.S. market, 3) the future GDP growth rate and the contemporaneous return of the global market. This table reports the slopes on GDP growth rate ( ), the market return 2 (MKT–RF), the corresponding t-statistics (t-stat), and R-squared (R ) from the regressions. The t-statistics are calculated using the Newey-West (1987) procedure with six lags. The sample period is from January 1991 to December 2013. Local market

U.S. market

GDP

t-stat

MKT–RF

t-stat

R

–0.45

–0.42

–0.04

–0.38

2

Global market GDP

t-stat

MKT–RF

t-stat

R2

0.11

1.48

1.24

0.18

1.44

0.04

GDP

t-stat

MKT–RF

t-stat

R

0.01

1.88

1.57

0.12

0.91

2

Asia China India

0.07

0.03

–0.05

–0.22

0.00

–2.12

–0.85

0.09

0.17

0.02

–4.03

–1.81

–0.02

–0.03

0.04

Indonesia

–0.93

–2.59

–0.02

–0.33

0.07

1.42

1.13

0.38

1.12

0.13

2.43

1.15

0.34

0.95

0.09

South Korea

–0.26

–0.71

–0.01

–0.08

0.01

–0.03

–0.03

0.00

0.03

0.00

–0.08

–0.09

0.01

0.09

0.00

Malaysia

–0.71

–0.97

0.01

0.07

0.02

–1.17

–0.68

–0.25

–1.43

0.06

–2.84

–1.94

–0.33

–1.95

0.10

Philippines

–1.35

–0.67

–0.38

–2.42

0.22

2.29

0.71

–0.49

–1.65

0.11

–1.25

–0.46

–0.79

–4.24

0.20

Taiwan

–1.21

–3.52

–0.06

–0.43

0.15

–2.19

–2.35

0.20

0.92

0.15

–2.38

–1.86

0.14

0.59

0.11

Thailand

–0.17

–0.25

–0.29

–1.39

0.16

2.26

0.70

–0.23

–1.07

0.05

0.14

0.05

–0.35

–1.71

0.04

Argentina

0.73

1.12

–0.33

–2.12

0.18

4.77

1.89

–0.82

–2.58

0.37

1.06

0.49

–0.92

–3.00

0.30

Brazil

–0.93

–0.39

–0.17

–1.60

0.09

3.44

1.46

–0.18

–0.80

0.14

2.82

1.22

–0.18

–0.98

0.07

Chile

0.16

0.25

–0.07

–0.58

0.02

0.34

0.48

–0.34

–2.54

0.15

–0.24

–0.26

–0.35

–2.20

0.17

Columbia

–2.37

–1.49

0.01

0.09

0.21

3.59

2.21

–0.14

–0.69

0.27

1.74

1.38

–0.25

–1.55

0.09

Mexico

0.54

0.38

–0.22

–0.94

0.14

1.58

1.32

–0.42

–1.68

0.19

1.34

0.74

–0.43

–1.60

0.20

2.57

1.63

–0.44

–2.09

0.26

4.52

1.68

0.54

1.19

0.18

4.19

1.81

0.30

0.97

0.05

Latin America

Europe Hungary Poland

0.12

0.06

–0.07

–0.31

0.00

0.19

0.15

–0.57

–2.22

0.14

0.43

0.18

–0.37

–1.16

0.06

Turkey

–0.25

–0.46

0.06

0.66

0.04

0.56

0.25

0.00

0.02

0.00

0.82

0.34

–0.06

–0.38

0.01

Czech

1.68

2.17

0.23

1.76

0.19

1.44

0.75

0.18

1.09

0.04

3.75

1.05

0.21

1.44

0.10

Russia

0.29

0.37

–0.04

–0.27

0.01

1.83

0.72

–1.13

–3.64

0.38

1.02

0.40

–1.07

–3.45

0.37

40 Page 41 of 47

Table 15. Funding liquidity and credit risk exposure of momentum factors For each market, we run annual time-series regressions of the momentum factors (UMD) on a funding liquidity variable or a credit risk variable after controlling for the future GDP growth rate and contemporaneous return of the U.S. market. The funding liquidity variables include the VIX, the TED spread for the U.S., and the U.S. LIBOR rate. The credit risk variables are the Aaa – T-note and Baa – Aaa spreads for the U.S. This table reports the regression slopes on the funding liquidity and credit risk variables (β) and the corresponding Newey-West adjusted t-statistics (t-stat). The sample period is from January 1991 to December 2013. VIX β Asia China India Indonesia South Korea Malaysia Philippines Taiwan Thailand Latin America Argentina Brazil Chile Columbia Mexico Europe Hungary Poland Turkey Czech Russia

TED t-stat

β

t-stat

LIBOR β t-stat

Aaa–T-note β t-stat

Baa–Aaa β t-stat

–0.08 –0.13 –0.07 –0.10 –2.42 –1.68 0.52 –5.62

–1.94 –1.19 –1.86 –2.13 –1.85 –0.84 0.50 –2.15

–0.16 –0.07 –0.05 –0.09 –2.02 –0.74 0.09 –0.09

–1.41 –1.23 –1.00 –1.65 –1.42 –0.54 0.10 –0.67

–0.13 –0.17 –0.12 –0.41 –1.53 –1.13 –1.32 0.22

–1.75 –1.59 –1.88 –3.26 –1.63 –1.14 –1.37 0.71

–0.11 –0.08 –0.13 –1.27 –1.80 –0.87 –1.43 –0.37

–1.85 –1.24 –1.74 –1.89 –0.85 –1.32 –1.60 –1.70

–0.18 –0.07 –0.10 0.34 –0.23 –0.25 –1.18 0.03

–2.68 –1.03 –1.22 0.21 –0.25 –0.27 –1.12 0.13

–0.18 0.06 –0.15 –0.64 –0.35

–0.99 0.28 –0.69 –1.36 –1.29

0.06 –0.13 –0.06 0.21 –0.04

0.29 –1.20 –0.33 0.54 –0.12

0.16 0.01 –0.05 –0.22 –0.02

0.90 0.04 –0.23 –0.72 –0.06

–0.07 0.20 –0.36 –0.07 –0.28

–0.28 1.24 –0.76 –0.21 –0.97

0.15 0.41 0.11 –0.22 0.09

0.71 1.61 0.48 –0.64 0.51

0.03 –0.20 –2.57 –1.32 –1.54

0.08 –0.72 –3.68 –1.22 –1.12

–0.29 –0.15 –1.33 –2.88 –1.70

–1.06 –0.23 –1.21 –3.38 –0.99

–0.67 –1.74 –2.23 –0.61 –1.92

–1.82 –2.83 –1.45 –0.77 –2.28

0.08 –3.82 –2.11 –0.30 –1.45

0.21 –1.41 –2.07 –0.43 –1.53

–0.16 –1.58 –1.76 –2.30 –3.95

–0.53 –1.05 –1.40 –1.97 –2.42

41 Page 42 of 47

Table 16. Exposure of momentum factors to the liquidity factors of Pastor and Stambaugh 2003 and Sadka 2006 For each market, we run annual time-series regressions of the momentum factor (UMD) on a liquidity variable after controlling for the future GDP growth rate and contemporaneous return of the U.S. market. The three liquidity variables from Pastor and Stambaugh (2003) are the level, the innovation, and the traded factor variables. The two liquidity variables from Sadaka (2006) are the fixed and the variable components. This table reports the regression slopes on the liquidity variables (β) and the corresponding Newey-West adjusted t-statistics (t-stat). The sample period is from January 1991 to December 2013.

β Asia China India Indonesia South Korea Malaysia Philippines Taiwan Thailand Latin America Argentina Brazil Chile Columbia Mexico Europe Hungary Poland Turkey Czech Russia

PS level t-stat

PS innovation β t-stat

PS traded factor β t-stat

Sadka fixed factor β t-stat

Sadka variable factor β t-stat

–0.02 0.10 0.01 0.11 –0.02 0.06 –0.02 0.18

–0.52 1.14 0.20 1.89 –0.10 0.66 –0.46 1.89

–0.02 0.06 –0.14 0.07 0.04 –0.10 –0.04 0.05

–0.18 0.94 –1.27 1.03 0.47 –1.18 –0.56 0.53

0.05 0.14 0.00 0.25 0.13 –0.14 0.05 0.19

0.35 1.62 0.07 2.79 1.24 –1.28 0.94 0.86

0.11 –0.10 0.04 –0.04 0.03 –0.03 0.08 0.15

1.64 –1.16 0.59 –0.87 0.26 –0.67 1.29 0.77

0.22 –0.03 0.09 –0.09 –0.02 –0.21 0.09 0.04

2.06 –0.37 1.48 –1.01 –0.38 –1.41 1.06 0.43

–0.11 0.07 0.18 –8.47 –0.17

–0.70 0.72 1.52 –1.20 –0.03

0.02 0.09 0.09 –11.36 –8.28

0.13 1.10 0.95 –2.46 –1.51

–0.09 –0.22 0.27 3.52 6.57

–0.75 –1.85 2.90 1.34 2.42

–0.08 –0.05 0.24 8.97 4.69

–0.61 –0.46 1.29 1.54 1.11

0.02 –0.01 2.00 –3.90 0.22

0.19 –0.08 0.83 –1.17 0.08

–8.31 –1.39 0.11 –1.11 –0.65

–3.12 –0.31 0.14 –1.02 –0.98

–1.64 1.53 1.44 –0.23 –0.23

–0.58 0.22 1.58 –0.24 –0.40

0.97 0.16 0.46 1.10 0.70

0.22 0.45 0.64 2.42 1.16

–1.48 –1.14 0.73 –0.04 0.00

–0.59 –1.26 1.01 –0.06 0.00

–1.95 1.83 0.86 –0.83 0.30

–0.50 1.86 0.74 –1.28 0.23

42 Page 43 of 47

Table 17. Value and momentum factors - subperiod analyses We calculate the average monthly value and momentum factors for four subperiods: 1) the period 1990−2001 (first half), 2) the period 2002−2013 (second subperiod), 3) the global financial-crisis period December 2007−June 2009 (crisis), and 4) the non-crisis period 1990−2013 after excluding the global financial-crisis period (non-crisis). Panels A and B of this table, respectively, report the results for value and momentum factors. In each panel, entries under "First versus second half" present average monthly premiums for the first and second subperiods and their difference; entries under "Non-crisis versus crisis period" present average monthly premiums for the non-crisis and crisis periods and their difference. The t-statistics are calculated using the Newey-West (1987) procedure with six lags. Panel A. Value factors First half Mean t-stat Asia China India Indonesia South Korea Malaysia Philippines Taiwan Thailand Latin America Argentina Brazil Chile Columbia Mexico Eastern Europe Hungary Poland Turkey Czech Russia

First versus second half Second half Mean t-stat

First minus second Mean t-stat

Non-crisis Mean t-stat

Non-crisis versus crisis period Crisis Non-crisis minus crisis Mean t-stat Mean t-stat

1.450 0.289 3.010 1.094 0.929 0.754 0.781 2.109

2.80 0.45 3.69 1.32 2.51 1.80 1.34 4.39

0.574 2.170 1.669 1.585 1.142 1.556 0.594 1.065

2.20 3.20 3.79 4.36 4.79 2.90 1.82 2.50

0.876 –1.881 1.341 –0.491 –0.213 –0.802 0.187 1.044

1.51 –2.02 1.44 –0.55 –0.48 –1.17 0.28 1.63

0.993 1.237 2.365 1.298 1.033 1.108 0.636 1.514

3.06 2.45 4.84 2.67 4.46 3.38 1.80 4.81

1.217 1.144 2.058 1.796 1.062 1.677 1.261 2.392

2.22 0.67 1.21 2.37 1.51 1.30 1.31 1.56

–0.224 0.093 0.308 –0.498 –0.029 –0.569 –0.625 –0.878

–0.34 0.05 0.18 –0.56 –0.04 –0.44 –0.61 –0.57

1.910 0.298 0.677 1.979 1.093

2.37 0.40 1.81 2.30 2.24

1.068 0.523 0.404 0.643 0.566

1.78 1.33 1.55 1.14 1.75

0.842 –0.226 0.274 1.337 0.528

0.84 –0.27 0.60 1.30 0.90

1.598 0.289 0.503 1.433 0.763

3.00 0.64 2.11 2.52 2.48

0.292 1.744 0.948 –0.029 1.561

0.19 2.92 1.14 –0.05 2.10

1.306 –1.455 –0.444 1.462 –0.798

0.81 –2.03 –0.51 1.82 –1.02

2.001 2.133 1.790 1.114 2.225

3.17 2.75 2.35 1.80 1.53

0.871 1.021 0.856 1.329 0.902

1.54 2.59 2.02 2.43 1.84

1.130 1.112 0.934 –0.215 1.323

1.33 1.28 1.07 –0.26 0.86

1.255 1.555 1.284 1.429 1.307

2.85 3.25 2.89 3.37 1.71

3.427 1.821 1.750 –0.863 3.231

2.55 1.91 1.14 –0.66 3.15

–2.173 –0.266 –0.466 2.292 –1.925

–1.54 –0.25 –0.30 1.66 –1.53

43 Page 44 of 47

Panel B. Momentum factors

First half Mean t-stat

First versus second half Second half Mean t-stat

First minus second Mean t-stat

Non-crisis Mean t-stat

Non-crisis versus crisis period Crisis Non-crisis minus crisis Mean t-stat Mean t-stat

Asia China India Indonesia South Korea Malaysia Philippines Taiwan Thailand

0.357 1.574 –1.908 0.511 –0.840 –0.130 0.173 0.070

0.75 1.62 –2.44 0.70 –1.27 –0.15 0.24 0.07

–0.160 2.408 –0.454 0.416 0.933 0.224 0.431 0.688

–0.41 2.50 –0.62 0.85 2.00 0.36 0.95 1.26

0.517 –0.835 –1.455 0.095 –1.772 –0.355 –0.258 –0.618

0.83 –0.61 –1.36 0.11 –2.19 –0.33 –0.30 –0.54

0.288 2.355 –0.973 0.717 0.214 0.185 0.534 0.537

0.94 3.72 –1.77 1.62 0.50 0.36 1.27 0.93

–1.983 –2.020 –3.472 –2.329 –1.795 –1.476 –2.252 –1.361

–1.74 –0.53 –1.63 –1.74 –0.97 –0.53 –1.42 –0.62

2.271 4.376 2.499 3.046 2.008 1.661 2.786 1.898

1.91 1.14 1.15 2.17 1.06 0.59 1.69 0.84

Latin America Argentina Brazil Chile Columbia Mexico

1.237 1.237 0.426 –0.069 0.912

1.57 1.49 1.12 –0.08 1.70

–0.053 0.227 1.059 –0.182 0.556

–0.08 0.44 2.87 –0.42 0.98

1.290 1.010 –0.632 0.112 0.356

1.27 1.03 –1.19 0.12 0.46

0.678 0.953 0.796 –0.188 0.860

1.32 1.95 3.09 –0.37 2.59

–0.356 –1.706 0.157 0.560 –0.649

–0.15 –0.98 0.10 0.73 –0.26

1.034 2.660 0.639 –0.748 1.509

0.43 1.50 0.42 –0.80 0.61

Eastern Europe Hungary Poland Turkey Czech Russia

1.651 0.922 –0.900 0.577 –1.234

1.87 1.13 –1.31 0.61 –0.59

0.048 0.465 –0.022 0.270 0.144

0.08 0.72 –0.04 0.63 0.17

1.603 0.457 –0.878 0.307 –1.377

1.48 0.44 –1.04 0.29 –0.61

0.914 0.883 –0.284 0.580 –0.214

1.58 1.82 –0.66 1.10 –0.20

0.142 –1.391 –2.405 –1.294 –2.896

0.09 –0.52 –1.43 –1.05 –0.79

0.772 2.275 2.121 1.874 2.682

0.46 0.85 1.23 1.40 0.70

44 Page 45 of 47

Table 18. Inter-market correlations of value premiums - subperiod analyses Panels A and B, respectively, report the difference in inter-market correlations of value premiums between the period 1990−2001 (first half) and the period 2002−2013 (second half), and the difference in inter-market correlations between non-crisis period (i.e., the period 1990-2013 after excluding the global financial crisis period December 2007−June 2009) and the global financial-crisis period. Numbers in bold indicate significance at the 5% level or better. Panel A. Difference in correlation between the first half and the second half Asia (1) (1) China (2) India (3) Indonesia (4) SouthKorea (5) Malaysia (6) Philippines (7) Taiwan (8) Thailand (9) Argentina (10) Brazil (11) Chile (12) Columbia (13) Mexico (14) Hungary (15) Poland (16) Turkey (17) Czech (18) Russia

(2) 0.07

(3) −0.03 −0.20

(4) −0.09 0.06 −0.44

(5) 0.09 −0.16 −0.40 −0.32

(6) 0.06 −0.34 −0.33 −0.17 −0.12

(7) −0.12 0.22 −0.09 0.21 −0.50 −0.07

(8) 0.13 −0.23 −0.57 −0.05 −0.01 −0.24 −0.22

Latin America (9) (10) 0.01 −0.17 −0.24 −0.05 −0.18 0.24 −0.10 −0.23 −0.07 −0.09 0.20 −0.12 −0.25 −0.11 0.09 −0.09 −0.31

(11) −0.08 −0.07 −0.11 0.01 0.03 −0.11 0.07 −0.32 0.05 −0.06

(12) 0.01 0.16 0.12 0.26 −0.05 0.19 0.38 0.20 0.35 0.28 0.29

(13) −0.02 0.22 0.08 0.02 −0.24 −0.27 0.09 −0.24 −0.25 −0.15 −0.13 −0.01

Europe (14) −0.08 −0.19 −0.10 −0.17 −0.04 −0.25 −0.01 −0.37 0.16 0.05 −0.11 0.13 −0.25

(15) 0.17 −0.01 −0.24 −0.09 0.17 −0.06 0.04 −0.30 0.09 −0.27 −0.37 0.21 −0.16 −0.11

(16) −0.17 −0.35 −0.24 −0.31 −0.35 −0.35 −0.18 −0.30 0.10 −0.20 0.18 0.10 −0.42 0.07 0.12

(17) 0.11 0.16 −0.05 −0.22 0.06 0.21 0.17 0.10 0.05 −0.12 0.00 −0.08 0.03 0.09 −0.07 −0.20

(18) −0.16 0.00 −0.24 −0.07 −0.02 −0.23 −0.15 0.11 0.17 −0.20 −0.18 −0.17 −0.24 −0.28 −0.39 −0.22 −0.16

45 Page 46 of 47

Panel B. Difference in correlation between the non-crisis and crisis periods Asia (1) (1) China (2) India (3) Indonesia (4) SouthKorea (5) Malaysia (6) Philippines (7) Taiwan (8) Thailand (9) Argentina (10) Brazil (11) Chile (12) Columbia (13) Mexico (14) Hungary (15) Poland (16) Turkey (17) Czech (18) Russia

(2) −0.25

(3) −0.06 −0.56

(4) −0.03 −0.02 −0.50

(5) 0.02 −0.47 −0.73 −0.44

(6) −0.20 −0.55 −0.71 −0.17 −0.29

(7) −0.12 −0.08 −0.23 0.14 −0.26 −0.30

(8) −0.08 −0.58 −0.84 −0.14 −0.45 −0.60 −0.32

Latin America (9) (10) 0.15 −0.15 −0.21 −0.50 0.10 0.08 −0.26 −0.25 0.04 −0.41 0.08 −0.24 0.07 −0.19 0.01 −0.39 −0.08

(11) −0.20 −0.49 −0.49 −0.06 −0.06 −0.38 −0.06 −0.38 0.13 −0.06

(12) −0.03 0.19 0.36 0.41 −0.18 0.35 0.38 0.20 0.34 0.41 0.42

(13) 0.09 −0.11 0.04 0.24 −0.17 −0.50 0.04 −0.30 −0.22 −0.19 −0.31 0.02

Europe (14) −0.43 −0.45 −0.16 −0.01 −0.02 −0.42 −0.46 −0.53 0.28 −0.20 −0.24 0.22 −0.21

(15) −0.15 −0.40 −0.18 0.20 0.08 −0.20 −0.03 −0.32 −0.16 −0.10 −0.53 0.22 −0.30 −0.20

(16) −0.20 −0.58 −0.57 −0.49 −0.90 −0.55 −0.18 −0.63 0.14 −0.41 0.03 −0.03 −0.31 0.06 0.28

(17) 0.48 0.61 0.13 −0.03 0.34 0.65 0.34 0.26 0.27 0.25 0.03 0.22 0.34 0.34 0.33 0.10

(18) −0.13 −0.39 −0.25 0.04 −0.11 −0.23 −0.25 0.05 0.35 −0.61 −0.33 −0.24 −0.43 −0.44 −0.25 −0.13 −0.19

46 Page 47 of 47