This study examines whether price momentum is driven by firms’ recent past performance or intermediate horizon past performance in the Korean stock market. Price momentum, a tendency of rising and falling stocks to keep rising and falling, is one of the most well-known stock market anomalies, and an extensive body of literature has shown that momentum is widely observed across a variety of countries and asset classes. However, previous studies provide the particularly interesting result that momentum is not strongly observed in the Korean stock market, and this is seen as exceptional. Several recent studies suggest that momentum strategies based on ranking or holding periods longer than six months were profitable in Korea in the period after the 1997 Asian financial crisis. Given these results and motivated by the recent empirical finding of Novy-Marx (2012, Is momentum really momentum? Journal of Financial Economics 103, 429-453) that momentum is primarily driven by intermediate past performance rather than by recent past performance, in this study, I examine whether the period over which past performance is measured plays an important role in determining the profitability of momentum strategies in the Korean stock market. I use monthly data for ordinary common stocks listed on the Korean Stock Exchange over the period February 1999 to December 2015. The main empirical findings are summarized as follows. First, momentum strategies based on performance during the period from 12 to seven months prior to portfolio formation earn positive and significant abnormal returns, while momentum strategies based on performance during the recent six months do not yield significant profits. The zero-cost portfolio buying the top 10% and selling the bottom 10% stocks based on intermediate past performance yields a Fama-French (1993) three-factor alpha of 1.51% per month, with a t-statistic of 2.06. The intermediate past performance can predict stock returns in the cross-section even after controlling for various firm characteristics, including firm size, the book-to-market ratio, idiosyncratic volatility, illiquidity, and turnover. Second, the abnormal profits of momentum strategies based on performance during the past 12 months are entirely driven by intermediate past performance, not by recent past performance. The past 12-month return loses its predictive power for the cross-section of stock returns after controlling for past returns during the period from prior 12 to seven months. At the same time, the predictive power of intermediate past returns is not affected by recent past returns. Stocks that are recent six-month winners but past 12-to-seven-month losers earn abnormally low returns on average. Third, the momentum anomaly based on intermediate past performance is more pronounced for stocks with a smaller size, higher liquidity, higher proportion of individual trading, and higher analyst coverage. The roles of liquidity and analyst coverage are not easily explained if momentum based on intermediate past performance was due to mispricing, because arbitrage is more likely to be limited for stocks with lower liquidity and lower analyst coverage. Finally, the abnormal profits of momentum strategies based on intermediate past performance are positive and significant only following periods of high aggregate liquidity and high stock market returns, but they are insignificant after periods of low aggregate liquidity and low stock market returns. This result suggests that momentum based on intermediate past performance could be at least partially due to a change in economic states. This study contributes to the literature on the cross-section of returns in the Korean stock market in the following ways. First, it is the first to investigate whether intermediate horizon past performance can predict the cross-section of returns in Korea. Previous studies on price momentum in the Korean stock market have focused on the predictive power of recent past performance and found only weak evidence that momentum profits are partially significant for stocks with particular characteristics or during particular periods. This study provides strong evidence that momentum strategies based on intermediate past performance yield significant and positive abnormal profits during the post-crisis period. Second, the empirical findings of this study help to reconcile conflicting results in the literature. It has been widely believed that momentum is relatively weak in Korea because momentum strategies based on past performance measured over no longer than six months are not profitable, which stands in contrast to the strong profitability of momentum strategies in the U.S. market. This study finds evidence that based on intermediate past performance, momentum strategies also become as profitable in the Korean stock market as in the U.S market. Although the profitability of momentum strategies based on recent past performance differs, intermediate past performance can significantly predict the cross-section of stock returns in both markets. These results suggest that price momentum cannot be simply interpreted as a short-run autocorrelation in returns and emphasize the importance of information in intermediate-horizon past performance.

• EndNote
• RefWorks
• Scholar's Aid
• BibTeX

Export Citation Cancel

Following Fama and French’s (1992, 1993) findings that firm characteristics or factors such as size and value have forecasting power for excess returns, academics and the financial industry are investigating and investing in styles of value and growth. The literature examining the value premium (the excess performance of value portfolio over growth portfolio) typically focuses on long-term performances in international stock markets, including the Korean stock market. In addition, according to the literature, participants such as institutional investors avoid investing in growth stocks. However, many empirical studies indicate that the value premium is related to the current state of the economy. They argue that value premiums are time-variant. Therefore, we analyze the existence of value premiums and the performance of style allocation strategies across different regimes in Korean stock market. We use style indices data for the Korean stock market developed by three indices companies, FnGuide, WISEfn, and MSCI, comprising three value indices and three growth indices from 2001 to 2014. Although value indices outperform growth indices for three indices companies during the total sample period, the outperforming of styles is not stable over time. That is, the sign and degree of the value premium varies each year. This time-varying value premium suggests that style allocation strategies to spread investment across value and growth stocks outperform strategies to invest only in value stocks. Thus, to construct a dynamic style allocation strategy based on stock market regimes, we identify two regimes using the regime-switching model introduced by Hamilton (1989, 1990) using the KOSPI index. The first regime based on the KOSPI index has high average returns and low volatility—this is the expansion regime or low-volatility regime. The second regime has low average returns and far high volatility—this is the recession regime or high- volatility regime. According to the result from the regression analysis regarding value and growth indices, growth stocks strongly outperform value stocks in low-volatility regimes. Our strategy is to adjust the weight between the value portfolio and the growth portfolio constructed on different regimes to show better performance over value-only investment strategy. In particular, the performance of dynamic style allocation strategy is best when we use the value index and growth index of MSCI. This result comes from the better performance of growth indices over value indices during the expansion regime (or low-volatility regime). The style allocation strategy that adjust the weights of growth stocks and value stocks of MSCI in different regimes or according to the probabilities of regimes shows an annual average return of 13.86% (in-sample) and 14.09% (out-of-sample), while the strategy to invest only in value stocks of MSCI earns an annual average return of 12.86%. That is, the dynamic style allocation strategy using MSCI indices produces 1.0% (in sample) and 1.23% (out-of- sample) more per annum than the value-only investment using the MSCI value index. Furthermore, for all three companies, the style allocation strategy outperforms the value-only investment over a considerable period. In particular, it is important to properly adjust the portfolio weights of styles in the large-value premium period. The style allocation that profits promoted in 2004 and 2012 as styles are rotated appropriately according to the regimes. In 2008 and 2014, when the returns on growth stocks were relatively high, the style allocation profits developed due to the increasing investment weight of the growth index. Note that the style allocation performance using MSCI indices is best and has the strongest correlation across the three indices companies. We expect that the style allocation can produce higher returns when we use data that have clearer characteristics or style, such as pure-value- and pure-growth- oriented funds. In conclusion, we identify the time-varying value premium in the Korean stock market. A proper style allocation strategy based on the economic situation or market conditions can generate better performance than investments that exclude growth stocks.

• EndNote
• RefWorks
• Scholar's Aid
• BibTeX

Export Citation Cancel

“A volatility interruption (VI) is a sophisticated microstructure mechanism providing cooling-off periods and effective price discovery functionalities in brief periods of abnormal volatility for an individual stock” (Eom, Ra, Park, and Ahn, 2015). VIs consist of two types: dynamic and static. The dynamic VI is invoked when a price fluctuation due to a single order exceeds a predetermined range, e.g., ±2% in the Korea Exchange (KRX). The static VI is activated when a cumulative price fluctuation due to multiple orders and transactions exceeds a predetermined range, e.g., ±10%. On September 1, 2014, KRX preemptively adopted the dynamic VI. Then, on June 15, 2015, the exchange additionally introduced the static VI while simultaneously expanding the price limit to ±30% to alleviate a wider range of price changes. In this paper, focusing on the adoption that occurred on June 15, 2015, we examine whether the newly introduced static VI, in addition to the existing dynamic VI, produces economic benefits consistent with the purpose of introducing VIs, such as price stabilization and discovery. In this process, we also identify the characteristics involved in dynamic and static VI invocations and discuss what type of VIs or price stabilization mechanisms are best suited to the Korean stock market. The related domestic and foreign papers are limited in number. The unique feature of this paper is the direct comparative analysis of the economic effectiveness of both dynamic and static VIs. We were able to perform this analysis because KRX adopted the dynamic and static VIs consecutively. In addition, the price-limit system, which is unique to the Korean stock market, allows us to discuss the economic relationship between the static VI and price limit. We analyze 1,937 stocks listed on the KOSPI and KOSDAQ markets in KRX over 2 periods of 48 trading days each before and after June 15, 2015, when the static VI was introduced. The previous period denotes the period when only the dynamic VI was implemented, and the latter period refers to the period when both the dynamic and static VIs were implemented in tandem with an increase of the price limit to ±30%. The results of our analyses using trade and quote data are as follows. First, the stocks for which the static VI is invoked are mainly small and medium-sized, low-priced, and highly volatile, as are those for which the dynamic VI is invoked. Unlike the dynamic VI, however, the static VI tends to be invoked more frequently as trading volume increases, suggesting that the static VI invocation is a phenomenon that occurs when investors’ opinions vary. The static VI is invoked more often than the dynamic VI in terms of the number of stocks or frequencies. Second, there is little relation between dynamic VI invocation and the hit occurrence of upper and lower limits. Meanwhile, the static VI has a significant influence on the hit occurrence of upper and lower limits in direction and magnitude. This suggests that the static VI plays a role in enforcing narrow upper and lower limits. However, assuming that the intrinsic volatility of the KOSPI and KOSDAQ markets remained constant, the expansion of the price limit to ±30% appears to have increased the realized volatility of these markets by about 14~15%. Third, the price stabilization and price discovery effects of the dynamic VI are fairly solid, and the effects themselves are consistent regardless of the introduction of the static VI or the expansion of the price limit. These effects are also almost the same quantitatively and qualitatively as those of the dynamic VI that was adopted in September 2014. The static VI, in contrast, appears to contribute to some degree to price discovery, but greatly impairs price stability. This is attributable to the static VI and price limit, which function fairly similarly, being implemented at the same time. The VI system in Korea was adopted to eventually replace the existing price-limit system, which has been evaluated as ineffective in preventing temporarily abnormal fluctuations in prices that may occur in high-frequency trading, manipulative trading, etc. Our results show that the dynamic VI is serving its intended purposes relatively well, while the static VI has a significant negative impact on price stability. This seems to be due to a conflict between the designed functionality of the static VI and its practical role as a narrow upper and lower price limit within the existing price-limit system. In addition, the dynamic VI has a limited effect on the establishment of an equilibrium price immediately after the VI invocation period in cases where the stock price overshoots to a remarkable level just before the dynamic VI invocation. Therefore, in-depth verification of the various parameters (e.g., ±10% variation in the static VI) that constitute the specific activating requirements of the dynamic and static VIs adopted by KRX is advised.

• EndNote
• RefWorks
• Scholar's Aid
• BibTeX

Export Citation Cancel

By empirically testing short selling activities that occur before changes in analysts’ recommendations, this paper examines the “informed front-running hypothesis” that predicts an abnormal decrease (increase) of short selling before analyst upgrades (downgrades). According to Blau and Wade’s (2012) conceptual framework, this study divides analyst recommendation changes into analyst upgrades and downgrades and uses four measures of abnormal short turnover used in Christophe et al. (2010) and Blau and Wade (2012). This study shows that abnormal shorting increases significantly before both downgrades and upgrades. After dividing the sample into KOSPI and KOSDAQ categories based on market type, domestic and foreign analysts’ recommendation changes, and investor type (institutions, foreigners, and individuals), the study finds no evidence that short sellers have the ability to acquire information about upcoming recommendation changes before the information becomes publicly available. The study also finds that pre-recommendation short selling does not predict upcoming events. These findings disprove the informed front-running hypothesis and are generally consistent with Blau and Wade’s (2012) argument that short selling before recommendation changes is more likely to be speculative than informed.

• EndNote
• RefWorks
• Scholar's Aid
• BibTeX

Export Citation Cancel

This study investigates the effect of external shocks on stock market volatility, focusing on improving the explanatory power and prediction ability of the heterogeneous autoregressive model of realized volatility (HAR-RV) using intraday high-frequency 5-minute return data from the KOSPI market index over the period from January 2004 to June 2016. Based on previous studies, we use improved methods in our empirical design to enhance the reliability of the empirical results: a method extracting jump components from measurements of realized volatility based on statistical significance evaluation; a method incorporating nighttime market information (without trading) into the measurements of realized volatility; a method assessing whether to improve the prediction ability of a proposed model through statistical significance evaluation; and a robustness test comparing the proposed model with models containing well-known explanatory variables of the volatility leverage effect and realized skewness and kurtosis. This study creates time series data on the external shock variable (ES) using principal components analysis by combining the selected 10-type variables that have the property of external shocks in the international financial markets, raw material markets, and commodity markets. It then uses the ES variable in empirical tests. The main results are as follows. First, the ES variable created from the principal components analysis does well at reflecting large changes in international markets. Second, the ES variable has a significant Granger causality relationship to the realized volatility over the whole period, while each of the selected 10-type variables has a significant causality relationship only for the specific periods of interest. Third, from the perspective of in-sample analysis, the HAR-RV model with the ES variable significantly improves the explanatory power of changes in the future realized volatility. In the out-of-sample analysis, the ES variable has the significant information value of enhancing the predictive power of the model in future periods. Finally, these are robust results regardless of whether the volatility leverage effect and realized skewness/kurtosis variables are included in the HAR-RV model. Our results show robust evidence that the external shocks have meaningful information value for explaining and predicting changes in future volatility in the HAR-RV model, and imply that the methodology of constructing the ES variable may provide new directions for future research using the HAR-RV model.

• EndNote
• RefWorks
• Scholar's Aid
• BibTeX

Export Citation Cancel

Aït-Sahalia, Parker, and Yogo (2004) suggest using luxury goods retail sales data as an alternative measure of consumption to obtain more reasonable estimates for the coefficient of relative risk aversion that better match the observed equity premium. We apply their novel idea of using data that reflect discretionary consumption by the wealthy to the Korean context by using sales revenue of three largest sellers of high-end whisky and two major airlines as proxies for discretionary consumption that are more likely to respond to movements in the stock market. When theses proxies for discretionary consumption are used in place of standard consumption, the estimates for relative risk aversion are an order of magnitude smaller and economically plausible, similar to the findings reported by Aït-Sahalia et al. (2004) for the United States.

• EndNote
• RefWorks
• Scholar's Aid
• BibTeX

Export Citation Cancel

We examine the effects of idiosyncratic and systematic volatility and stock return synchronicity on stock liquidity using a sample of firms listed on the Korea Exchange (KRX) from 2002 to 2013. The association between volatility and liquidity is extensively studied in the literature, with a typical focus on how “total” volatility affects liquidity. Few studies, however, divide volatility into idiosyncratic and systematic components to see how these individually influence liquidity. Distinguishing between the two components is important to studies examining the relationship between volatility and liquidity because they may influence liquidity in significantly different ways. Market makers face two sources of risk in providing liquidity to the market: inventory risk and adverse selection risk. Greater inventory risk, greater adverse selection risk or both lead to greater spreads posted by market makers to cover their potential losses, resulting in reduced liquidity. To understand how this mechanism works, we need to understand how idiosyncratic and systemic volatility each pose inventory concerns and adverse selection risk to market makers. Let us first consider how market makers’ inventory risk is affected by the two volatility components. Predictions are possible in either direction. Since systematic volatility can be hedged away by market makers, it may not pose much inventory concern. Idiosyncratic volatility, however, may have a direct effect on inventory costs because it cannot be removed by hedging. Portfolio diversification has a completely opposite implication. Although idiosyncratic volatility can be diversified away easily, systematic volatility cannot. Hence, the former may not have as significant an effect on inventory costs as the latter. The two volatility components may also have different influences on adverse selection risk. The costs related to information asymmetry tend to be greater with idiosyncratic volatility than with systematic volatility because it is easier for a market maker to access and interpret common signals to the market. As a result, firms with greater idiosyncratic volatility are more likely to have greater adverse selection risk and thus lower liquidity. Studies investigating the association between the individual volatility components and liquidity fall into two groups: those that examine the link between systematic volatility and liquidity (e.g., Baruch et al., 2007; Baruch and Saar, 2009) and those that explore the effect of idiosyncratic volatility on liquidity (Bali et al., 2005; Spiegel and Wang, 2005). Chan et al. (2013) merge these two strands by examining how both idiosyncratic and systematic volatility affect liquidity at the same time. They propose that collecting market-wide information costs uninformed market makers less than collecting firm-specific information, and thus firms with greater systematic risk have smaller adverse selection costs and greater liquidity. Based on their empirical analysis of U.S. stocks, Chan et al. show that stocks with greater systematic volatility and greater return synchronicity with the market have greater liquidity. We investigate whether the same relationship holds in the Korean stock market, where there are no market makers and liquidity is provided collectively by investors placing limit orders. We measure idiosyncratic volatility using residuals from the market model. Systematic volatility is then obtained by subtracting the idiosyncratic variance from the total variance. Return synchronicity, meanwhile, is measured as R2 from the market model. For liquidity measures, we use the Amihud liquidity measure and Roll’s spread, the two most popular liquidity proxies available with daily data. We regress the two liquidity proxies on the volatility components and return synchronicity measures together with a group of control variables that are known to affect liquidity in the literature. The latter includes stock price, equity market capitalization, turnover, and institutional ownership. We find that idiosyncratic volatility decreases liquidity, whereas systematic volatility and stock return synchronicity increase liquidity. Our results are consistent with the explanation that the collective liquidity providers in the Korean stock market also find it cheaper to gather market-wide information than firm-specific information, which leads to lower adverse selection costs and greater liquidity for stocks with greater systematic volatility. Our empirical findings are robust to alternative variable definitions and model specifications. Market makers play a central role in providing liquidity on the New York Stock Exchange (NYSE). However, on the KRX, there are no market makers. Limit orders submitted by the investing public serve as the primary repository of liquidity in stock trading. Despite this apparent difference in market structure, our empirical results generally support the findings of Chan et al. (2013) on the NYSE, which have an interesting implication for the efficacy of the limit-order trading system. Although individual limit-order traders do not possess the same set of advantages that market makers have in terms of their monopolistic access to order flow information, it appears that “collectively” the investing public substitutes market makers and plays the same role in providing liquidity to the market.

• EndNote
• RefWorks
• Scholar's Aid
• BibTeX

Export Citation Cancel

The literature on tax-expense momentum is unclear as to whether its effect is due to anomaly or risk. We apply the approach developed by Ohlson and Bilinski (2015) to assess whether the positive relationship between tax expense momentum (surprise) and future stock returns is explained by anomaly or risk. We find that tax expense momentum increases the probability of a high return and decreases that of a low return. This supports an anomaly-based explanation for the tax expense momentum.

• EndNote
• RefWorks
• Scholar's Aid
• BibTeX

Export Citation Cancel

This study documents that predictability of insider trading on future stock price crash varies according to the types of insiders and the timing of the sale. Using insider trading data from Korea between 2005 and 2014, we find that largest shareholders tend to sell far before a stock price crash, while other types of insiders, including other large shareholders and executives, are more likely to sell immediately prior to a crash. Such pattern is more pronounced in firms with low CSR scores and low R-squares, but not observed among firms with high CSR scores or high R-squares. We also find that our results are stronger amongst firms with higher litigation risk. These findings suggest that largest shareholders may be well aware of the potential legal or reputational risk associated with insider trading while the remaining insiders may be less concerned.

• EndNote
• RefWorks
• Scholar's Aid
• BibTeX

Export Citation Cancel

Asset allocation is one of the most important decisions made in the asset management process to achieve a portfolio’s target return within the constraints of its risk tolerance level. Brinson et al. (1986, 1991) conclude that asset allocation policy explains 93.6% or 91.5% of performance. This greatly helps practitioners understand the importance of asset allocation in asset management. However, at the same time, it triggers the publication of other studies with different views on its empirical testing and different interpretations of its results. For example, Ibbotson et al. (2000) and Vardharaj et al. (2007) conduct empirical tests using regression models to measure the explanatory powers of asset allocation in fund performance and find that the dispersion of R2 in their models is wider than those of Brinson et al. Alternatively, Xiong et al. (2010) decompose the total return into three components: market movement, asset allocation policy return in excess of the market return, and active return. They then run time-series and cross-sectional regressions on total returns and excess market returns for both equity and blend funds, respectively. Based on their regression analyses, Xiong et al. (2010) conclude that market movement dominates the other two components. Additionally, they find that the explanatory powers of passive management (asset allocation policy) and active management are at similar levels. In sum, although multiple studies propose different conclusions from Brinson et al. (1986), they all agree that asset allocation is significant in the asset management process. Applying the methodologies of Brinson et al. (1986) and Xiong et al. (2010), this study performs time-series and cross-sectional analyses of equity and blend fund returns in Korea, the U.S., the U.K. and Australia, based on each country’s style benchmark and leading market index returns. A performance attribution analysis is conducted to determine the contribution of passive and active management to fund performance by comparing the coefficients of determination (R2) for each country’s regression models. The results indicate that in accordance with the time-series regression analysis of Brinson et al. (1986), the explanatory power of asset allocation reaches its highest level when the total fund return is explained by benchmark returns reflecting market movement. However, in all four countries, when market movement is separated, it’s the greatest contribution to the performance of both equity and blend funds. When a market-movement-excluded excess return is used, in three countries, excluding the U.K., active management has higher explanatory power for equity funds and passive management has higher explanatory power for blend funds. When a cross-sectional regression analysis is conducted using the methodology of Xiong et al. (2010) to measure the explanatory power of asset allocation performance without market movement, the explanatory power of active management is higher for equity funds in all four countries. In the case of blend funds, however, the explanatory power of active management is higher in the U.S. and U.K., where equities hold a larger portion of funds than bonds. Finally, passive management shows higher explanatory power in Korea, where its funds are composed of similar levels of equities and bonds. Comprehensively reviewing the results of this study from a mid- and long-term investment policy perspective, we conclude that the selected variables in strategic asset allocation consist of investors’ risk preferences and asset class risk premiums, not short-term market movements. Therefore, it is difficult to agree with Xiong et al. (2010) that it is reasonable to exclude market movements from returns because the contribution to investment performance for mid- and long-term market movements are already reflected in risk premiums. We add, however, that if market movements are under control, which can be explained as constant asset class risk premiums, then we support the finding of preceding studies that active management is as significant as passive management in terms of its contribution to performance.

• EndNote
• RefWorks
• Scholar's Aid
• BibTeX

Export Citation Cancel