Ralph Lu - Academia.edu (original) (raw)

Papers by Ralph Lu

Social Science Research Network, 2011

ABSTRACT This study investigates whether the news release prior to the earnings announcement infl... more ABSTRACT This study investigates whether the news release prior to the earnings announcement influences different investors’ trading behavior during the earnings announcement period. We bridge the gap between this study and the previous analyses in that the technology of text mining is applied to extract the Chinese news, proxies of the news information are constructed, and comprehensive tick data are used to categorize trader types including dealers, mutual funds, foreigners, individuals and corporations. The empirical results show that individuals may have significant reactions to the news report prior to the earnings announcement. However the news report may have less of an effect on institutions’ trading behavior. The findings confirm that institutions are more rational than individuals.

Social Science Research Network, 2010

ABSTRACT In line with the stealth trading hypothesis advanced by Barclay and Warner (1993), Chakr... more ABSTRACT In line with the stealth trading hypothesis advanced by Barclay and Warner (1993), Chakravarty (2001) and Chakravarty, Van Ness and Van Ness (2005), this paper extends the test of fragmentation of trades for different types of investors in the Taiwan stock market – a pure order-driven emerging market. The trade-size clustering is investigated through a longitudinal analysis of the comprehensive tick data for Taiwan’s equity market from 1998/1/2 to 2006/12/31. We test the public information hypothesis that the stock price change associated with a release of public information that will occur on a trade of a given trade-size category is directly proportional to the percentage of trades in that category. Such a proposition goes against the stealth trading hypothesis which suggests that there is a disproportional association between the percentage cumulative price change and trade-size. To mitigate the research biases resulting from the deregulation of foreigners’ holdings during the study period, the sample is divided into three periods according to the investment caps for foreign investors in Taiwan for performing these tests (30%: 1998/1/2-1999/3/31, 50%: 1999/4/1-2000/12/31, no limit: 2001/1/2-2006/12/31). The empirical results indicate that, regardless of what type the investor is, most of the cumulative price changes are affected by small- and medium-size trades. By comparing the per percent volumes of cumulative absolute price changes for different institutional investors with individual investors, we further confirm that the extent of stealth trading inherent in institutions is much more intensive than it is for individuals. The clustering of the per percent trades of cumulative absolute price changes for large-size trades reveals the existence of trade-intention-induced price manipulation in Taiwan. The fact that the per percent trades of cumulative absolute price changes for large-size trades of institutions are significantly greater than those for individuals confirms the influence of institutional investors on price volatility. For the fully deregulated period (2001/1/2 – 2006/1/2), we find foreign institutional investors to be the most influential investors in Taiwan, even though they disguise their intention to trade. Our result is consistent with the prediction of Barclay and Warner, and Chakravarty’s stealth-trading hypothesis. While making the most of the information advantage, the institutional investors in Taiwan appear to engage in trade-intention-induced price manipulation in addition to stealth-trading behavior. The finding of the trade-intention-induced price manipulation of this paper is similar to the previous one, but different from the stark-naked phenomenon of the large block trade/ price manipulation in China’s stock market found in Cai, Cai, and Keasey (2006). Our findings appear to further confirm the market lore that institutions are informed traders.

Social Science Research Network, 2009

We document that, conditioned on a positive offer price revision from the midpoint of the initial... more We document that, conditioned on a positive offer price revision from the midpoint of the initial filing range, one extra piece of media coverage during the filing period for an IPO is associated with about two percentage points greater underpricing. Media coverage during the filing period doubles the adjusted R 2 in price revision regressions, with media coverage positively correlated with the absolute value of price revisions. One extra piece of media coverage generally leads to an additional 2.8% increase in the offer price when the price revision is positive, or to a 1.9% greater decrease if the price revision is negative. Combined with our results on underpricing, it appears that underwriters fully adjust for media coverage when revising the offer price downwards but only partially adjust when the offer price is revised upwards. We find that the positive relationship between media coverage and underpricing is stronger when ex ante uncertainty is greater, and fail to find any relationship between positive media coverage and IPO firms' long run underperformance. Overall, our findings are consistent with theories of underpricing being driven by the need to compensate investors for information acquisition, but are not consistent with investor sentiment or prospect theory explanations.

Social Science Research Network, 2013

ABSTRACT This study follows Sias (2004) in examining whether securities dealers’ herding behavior... more ABSTRACT This study follows Sias (2004) in examining whether securities dealers’ herding behavior exists and what is the main reason for their herding in the Taiwan stock market. By testing the cross-sectional dependence in dealers’ demand over two adjacent weeks and decomposing the data into their own cascades and other cascades, we demonstrate that dealers’ cascades mainly result from other cascades (herding), but their own trades are still significant for securities with at least low-to-medium trading frequency. We find little evidence that dealers’ herding is driven by habit investing in stocks which dealers trade with at least medium-to-high frequency. Momentum trading of dealers accounts for little of their herding behavior, and the obviously positive relationship between dealers’ demand and their lag demand changes little even with their momentum trading being taken into consideration. Most importantly, dealers are more likely to herd in large capitalization securities; thus, investigative herding rather than informational cascades are the main reason for dealers’ herding in the Taiwan stock market.

Social Science Research Network, 2013

ABSTRACT In line with Wermers (1999), Choe et al. (1999) and Sias (2004), this paper extends Sias... more ABSTRACT In line with Wermers (1999), Choe et al. (1999) and Sias (2004), this paper extends Sias (2004)’s dynamic herding measure to identify the behavioural type of cascading or herding of UK fund managers in the stock market, their reasons for herding, whether their herding behaviours are different during bullish and bearish periods and whether or not their herding behaviours are informational. Our results demonstrate that UK fund managers’ cascades primarily result from their herding. Because they are more likely to herd in large capitalisation securities, their herding may result from investigative herding, contrary to the finding of Sias (2004). Moreover, our finding that fund managers’ cascades primarily result from their herding, which does not change in bullish and bearish stock markets. We also find that growth-type and international-type funds are more likely to herd with funds similar to their types while value-type funds are more likely to herd with funds different from their types. In addition, we confirm that positive correlation exists between the fraction of fund managers buying and subsequent stock returns, consistent with the results of Wermers (1999), Choe et al. (1999) and Sias (2004). We advance the herding results of Hung et al. (2010) to prove that their herding behaviours belong to informational herding within the subsequent year. The implication of our finding is that investors could follow UK fund managers to purchases stocks overbought by them with at least 15 traders quarterly in the following one year period, especially for growth, specific sector and international funds to improve the performances of their equity portfolios.

This working paper is produced for discussion purpose only. These working papers are expected to ... more This working paper is produced for discussion purpose only. These working papers are expected to be published in due course, in revised form, and should not be quoted or cited without the authors written permission.

This paper first extends Sias (2004) and Wylie (2005) to examine the existence of and the reasons... more This paper first extends Sias (2004) and Wylie (2005) to examine the existence of and the reasons for UK fund managers' herding behaviours in the stock market. Whether their herding behaviours are different during bullish and bearish periods and whether or not their herding behaviours are informational are explored. Our results demonstrate that UK fund managers' cascades primarily result from their herding, and habit investing does not primarily drive their herding behaviour. Different from the results of Sias (2004) and Wylie (2005), we find that momentum trading can be regarded as one of the main reasons for UK fund managers' herding. Because they are more likely to herd in large capitalisation securities, their herding may result from investigative herding, which is contrary to the finding of Sias (2004). Moreover, our results find that UK fund managers' cascades primarily result from their herding, which does not change in bullish and bearish stock markets. We al...

Emerging Markets Finance and Trade, 2014

ABSTRACT We use a dynamic herding measure to explore the causes of foreign institutional investor... more ABSTRACT We use a dynamic herding measure to explore the causes of foreign institutional investor (FII) herding in the Taiwan stock market and examine the effects of stock characteristics on the direction and extent of such herding. We find that FII herding primarily results from cascades rather than habit investing or momentum trading. The result of a panel smooth transition regression shows that FIIs' negative cascades focus on their largest net purchases of stocks, but FIIs' positive cascades focus on winner and small-sized stocks. To increase portfolio returns, investors can use FIIs' cascades to inform their stock purchases.

SSRN Electronic Journal, 2013

ABSTRACT This study follows Sias (2004) in examining whether securities dealers’ herding behavior... more ABSTRACT This study follows Sias (2004) in examining whether securities dealers’ herding behavior exists and what is the main reason for their herding in the Taiwan stock market. By testing the cross-sectional dependence in dealers’ demand over two adjacent weeks and decomposing the data into their own cascades and other cascades, we demonstrate that dealers’ cascades mainly result from other cascades (herding), but their own trades are still significant for securities with at least low-to-medium trading frequency. We find little evidence that dealers’ herding is driven by habit investing in stocks which dealers trade with at least medium-to-high frequency. Momentum trading of dealers accounts for little of their herding behavior, and the obviously positive relationship between dealers’ demand and their lag demand changes little even with their momentum trading being taken into consideration. Most importantly, dealers are more likely to herd in large capitalization securities; thus, investigative herding rather than informational cascades are the main reason for dealers’ herding in the Taiwan stock market.

SSRN Electronic Journal, 2013

SSRN Electronic Journal, 2013

ABSTRACT In line with Wermers (1999), Choe et al. (1999) and Sias (2004), this paper extends Sias... more ABSTRACT In line with Wermers (1999), Choe et al. (1999) and Sias (2004), this paper extends Sias (2004)’s dynamic herding measure to identify the behavioural type of cascading or herding of UK fund managers in the stock market, their reasons for herding, whether their herding behaviours are different during bullish and bearish periods and whether or not their herding behaviours are informational. Our results demonstrate that UK fund managers’ cascades primarily result from their herding. Because they are more likely to herd in large capitalisation securities, their herding may result from investigative herding, contrary to the finding of Sias (2004). Moreover, our finding that fund managers’ cascades primarily result from their herding, which does not change in bullish and bearish stock markets. We also find that growth-type and international-type funds are more likely to herd with funds similar to their types while value-type funds are more likely to herd with funds different from their types. In addition, we confirm that positive correlation exists between the fraction of fund managers buying and subsequent stock returns, consistent with the results of Wermers (1999), Choe et al. (1999) and Sias (2004). We advance the herding results of Hung et al. (2010) to prove that their herding behaviours belong to informational herding within the subsequent year. The implication of our finding is that investors could follow UK fund managers to purchases stocks overbought by them with at least 15 traders quarterly in the following one year period, especially for growth, specific sector and international funds to improve the performances of their equity portfolios.

SSRN Electronic Journal, 2013

ABSTRACT The study extends Sias’s (2004) method of analysing institutional investors’ inter-tempo... more ABSTRACT The study extends Sias’s (2004) method of analysing institutional investors’ inter-temporal herding in securities to examine whether the herding effect among banks in industrial lending exists in Taiwan, determine the reasons for herding among these banks, and examine whether their herding behaviour differs across bullish and bearish periods and bank types. This study demonstrate that the cascades among banks in Taiwan primarily result from their herding and that habit loans are not the primary factor in banks’ herding behaviours. Because banks are more likely to engage in herding behaviour in the case of lending to large-capitalisation industries, their behaviours may stem from investigative herding. Moreover, our results show that banks’ cascades result primarily from their herding, which does not differ across bullish and bearish periods. Meanwhile, their herding behaviours are more significant for middle- and large-capitalisation industries in bullish periods. This study also finds that private banks are more likely to follow the same type of bank classification, whereas public banks are more likely to follow a different type. Commercial banks are more likely to follow the same type of bank class, whereas SME (small and medium-sized enterprise) banks are more likely to follow a different type. Our finding confirms the reputational herding behaviour by commercial banks in Taiwan. This study also finds that banks outside the system of financial holding companies are more likely to engage in a different type of herding, which may be evidence of characteristic herding among banks outside the financial holding system that follow those within the financial holding system. Finally, large banks are more likely to follow banks of the same type, whereas small banks are more likely to follow different types of banks, a finding consistent with the results of Barron and Valve (2000). To our best knowledge, this study is the first research which systematically explore banks’ herding behaviour on corporate loans in Taiwan.

Social Science Research Network, 2013

We apply text mining (TM) techniques to extract and quantify relevant Chinese financial news, in ... more We apply text mining (TM) techniques to extract and quantify relevant Chinese financial news, in an attempt to further develop the classical early warning models of financial distress. We extend the work of Demers and Vega (2011) by proposing a measure of the degree of credit default, referred to as the ‘distress intensity of default-corpus ’ (DIDC), and investigate the predictive power of DIDC on default probability by incorporating it into the signaling model, along with the classical financial performance variables (liquidity, debt, activity and profitability ratios). We construct a logistic regression (LR) model to better integrate the DIDC and financial performance variables into a more effective early warning signal model, with the incorporation of DIDC into the LR model revealing a significant reduction in Type I errors and an apparent increase in classification accuracy, thereby proving the effectiveness of the additional information from TM on financial corpus, and also con...

Journal of Financial Studies, 2012

This study examines the determinants of New Taiwan Dollar interest rate swap (NTD IRS) spreads. P... more This study examines the determinants of New Taiwan Dollar interest rate swap (NTD IRS) spreads. Prior literature provides evidence that the term structure of interest rates, liquidity, and credit risk comprise the swap spreads. The term structure of interest rates includes the interest rate level, the slope of the yield curve, and the interest rate volatility. The empirical results for the full sample period show that the interest rate level, the slope of the yield curve, interest rate volatility, liquidity risk, and credit risk are all important factors affecting the swap spreads. Results of the variance decomposition for the full sample period indicate that default risk is the most important factor among the five components. Furthermore, the results show that the explanatory power of the regression model for the term structure of interest rates, liquidity, and credit risk is higher in a bear market than in a bull market. Finally, the results of variance decomposition indicate that...

Romanian journal of economic forecasting

This study explores the dynamic relationship between the sentiment of institutional investors and... more This study explores the dynamic relationship between the sentiment of institutional investors and market returns in the futures market. Using data from the Taiwan futures market, the empirical results show that the dynamic relationship between the sentiment of foreign institutional investors and the futures returns is much stronger than that of the sentiment of domestic institutional investors and the futures returns. Our empirical results also display that the sentiment of foreign institutional investors Granger-causes the sentiment of domestic institutional investors, but not vice versa. Finally, the sentiment of foreign institutional investors has a larger effect on subsequent market returns and market states than that of the sentiment of domestic institutional investors. Overall, our empirical results suggest that the relationship among the institutional investor sentiment, market returns, as well as market conditions in the Taiwan futures market is dominated by the sentiment of...

SSRN Electronic Journal, 2008

This paper performs an empirical investigation on the relationship among the possibility of finan... more This paper performs an empirical investigation on the relationship among the possibility of financial distress, firm performance, corporation governance and macroeconomic environment using the companies listed on the Taiwan Stock Exchange (TSEC) and the financial distress events took place between 1997 and 2003. It is observed that the firms on the brush with bankruptcy defined as follows: insolvent companies; firms under 100%-cash margin requirement; trading-halt companies; firms de-listed from TSEC; net worth is negative; serious loss. The sample data set are formed according to the industry-matched and size-matched principle. We build up a quarters-ahead-predicting model of financial distress allowing for measuring treatment effects of governance characteristics. Meanwhile, we present a "moving-window with augmenting information set" research design to evaluate the capacity of our forecasting model. It is found that the capacity of poor company to come out of financial distress is determined by how transparent the governance is, especially for one-quarter ahead of prediction. All of the poor companies falling into the observed group of CEO with high stock-holding and those pledging ratio almost went into bankruptcy. Through the comparison of moving-window research design, we found that the early warning model with treatment effects performs better in prediction power than those competitive models.

SSRN Electronic Journal, 2013

ABSTRACT The study extends Sias’s (2004) method of analysing institutional investors’ inter-tempo... more ABSTRACT The study extends Sias’s (2004) method of analysing institutional investors’ inter-temporal herding in securities to examine whether the herding effect among banks in industrial lending exists in Taiwan, determine the reasons for herding among these banks, and examine whether their herding behavior differs across bullish and bearish periods and bank types. The study results demonstrate that the cascades among banks in Taiwan primarily result from their herding and that habit loans are not the primary factor in banks’ herding behaviors. Because banks are more likely to engage in herding behavior in lending to large-capitalization industries, their behaviors may stem from investigative herding. Moreover, our results show that banks’ cascades result primarily from their herding, which does not differ across bullish and bearish periods. Meanwhile, their herding behaviors are more significant for middle- and large-capitalization industries in bullish periods. The study also finds that private banks are more likely to follow the same type of bank classification, whereas public banks are more likely to follow a different type. Commercial banks are more likely to follow the same type of bank class, whereas SME (small and medium-sized enterprise) banks are more likely to follow a different type. This finding is consistent with reputation herding by commercial banks in Taiwan. Next, the study finds that banks outside the Taiwanese financial holding system are more likely to engage in a different type of herding, which may be evidence of characteristic herding among banks outside the financial holding system that follow those within the financial holding system. Finally, larger banks are more likely to follow banks of the same type, whereas smaller banks are more likely to follow different types of banks.

SSRN Electronic Journal, 2013

This study extends Sias (2004) to examine whether herding exists among foreign institutional inve... more This study extends Sias (2004) to examine whether herding exists among foreign institutional investors (FIIs), what is the cause of their herding, and whether their herding is stable across both bullish or bearish periods or institutional types in the Taiwan stock market. By testing the cross-sectional dependence in the FIIs' demand for stocks in two adjacent months and decomposing them into their own cascades and other cascades, we demonstrate that the FIIs' cascades mainly result from their herding for securities traded at medium to high frequency even though their own cascades still exist. We find little evidence that FIIs' herding behavior is driven by habit investing in stocks which FIIs trade with at least medium to high frequency. The momentum trading of FIIs is found to account for little of their herding, and the obviously positive relationship between the FIIs' demand and their lag demand changes little, even with their momentum trading being taken into consideration. Moreover, FIIs are more likely to herd in large capitalization securities, and thus investigative herding, rather than informational cascades, is the main reason for herding among FIIs in the Taiwan stock market. One of our contributions may be to find that the phenomenon that FIIs' cascades mainly result from their herding does not change in the bullish and bearish Taiwan stock market. However, since FIIs are more likely to herd in large-capitalization securities, their herding will obviously increase (decline) when they herd in large-capitalization (small-capitalization) stocks in the bullish (bearish) stock market period. This study further finds that FIIs and dealers are more likely to follow similar-type institutions than different-type institutions, respectively, which is consistent with the view proposed by Del Guercio (1996) and Bennett, Sias and Starks (2003) that institutional herding is driven by reputational or characteristic herding. Security dealers in Taiwan tend more often to follow same-type herding while FIIs more often negatively follow different-type herding, which implies that security dealers exhibit the strongest tendency to herd especially in same-type institutions possibly due to reputation concerns, which is consistent with the view put forward by Sias (2004).

SSRN Electronic Journal, 2011

This study investigates whether the news release prior to the earnings announcement would influen... more This study investigates whether the news release prior to the earnings announcement would influence different investors" trading behavior during the earnings announcement periods. We bridge the gaps between this study and the previous analyses that the technology of text mining is applied to extract the Chinese news, the proxies of the news information are constructed and the comprehensive tick data is used to categorize trader types including dealers, mutual funds, foreigners, individuals and corporations. The empirical results show that individuals may have significant reactions to the news report prior to the earnings announcement. However the news report may have less effect to institutions" trading behavior. The findings confirm that institutions are more rational than individuals.

Social Science Research Network, 2011

ABSTRACT This study investigates whether the news release prior to the earnings announcement infl... more ABSTRACT This study investigates whether the news release prior to the earnings announcement influences different investors’ trading behavior during the earnings announcement period. We bridge the gap between this study and the previous analyses in that the technology of text mining is applied to extract the Chinese news, proxies of the news information are constructed, and comprehensive tick data are used to categorize trader types including dealers, mutual funds, foreigners, individuals and corporations. The empirical results show that individuals may have significant reactions to the news report prior to the earnings announcement. However the news report may have less of an effect on institutions’ trading behavior. The findings confirm that institutions are more rational than individuals.

Social Science Research Network, 2010

ABSTRACT In line with the stealth trading hypothesis advanced by Barclay and Warner (1993), Chakr... more ABSTRACT In line with the stealth trading hypothesis advanced by Barclay and Warner (1993), Chakravarty (2001) and Chakravarty, Van Ness and Van Ness (2005), this paper extends the test of fragmentation of trades for different types of investors in the Taiwan stock market – a pure order-driven emerging market. The trade-size clustering is investigated through a longitudinal analysis of the comprehensive tick data for Taiwan’s equity market from 1998/1/2 to 2006/12/31. We test the public information hypothesis that the stock price change associated with a release of public information that will occur on a trade of a given trade-size category is directly proportional to the percentage of trades in that category. Such a proposition goes against the stealth trading hypothesis which suggests that there is a disproportional association between the percentage cumulative price change and trade-size. To mitigate the research biases resulting from the deregulation of foreigners’ holdings during the study period, the sample is divided into three periods according to the investment caps for foreign investors in Taiwan for performing these tests (30%: 1998/1/2-1999/3/31, 50%: 1999/4/1-2000/12/31, no limit: 2001/1/2-2006/12/31). The empirical results indicate that, regardless of what type the investor is, most of the cumulative price changes are affected by small- and medium-size trades. By comparing the per percent volumes of cumulative absolute price changes for different institutional investors with individual investors, we further confirm that the extent of stealth trading inherent in institutions is much more intensive than it is for individuals. The clustering of the per percent trades of cumulative absolute price changes for large-size trades reveals the existence of trade-intention-induced price manipulation in Taiwan. The fact that the per percent trades of cumulative absolute price changes for large-size trades of institutions are significantly greater than those for individuals confirms the influence of institutional investors on price volatility. For the fully deregulated period (2001/1/2 – 2006/1/2), we find foreign institutional investors to be the most influential investors in Taiwan, even though they disguise their intention to trade. Our result is consistent with the prediction of Barclay and Warner, and Chakravarty’s stealth-trading hypothesis. While making the most of the information advantage, the institutional investors in Taiwan appear to engage in trade-intention-induced price manipulation in addition to stealth-trading behavior. The finding of the trade-intention-induced price manipulation of this paper is similar to the previous one, but different from the stark-naked phenomenon of the large block trade/ price manipulation in China’s stock market found in Cai, Cai, and Keasey (2006). Our findings appear to further confirm the market lore that institutions are informed traders.

Social Science Research Network, 2009

We document that, conditioned on a positive offer price revision from the midpoint of the initial... more We document that, conditioned on a positive offer price revision from the midpoint of the initial filing range, one extra piece of media coverage during the filing period for an IPO is associated with about two percentage points greater underpricing. Media coverage during the filing period doubles the adjusted R 2 in price revision regressions, with media coverage positively correlated with the absolute value of price revisions. One extra piece of media coverage generally leads to an additional 2.8% increase in the offer price when the price revision is positive, or to a 1.9% greater decrease if the price revision is negative. Combined with our results on underpricing, it appears that underwriters fully adjust for media coverage when revising the offer price downwards but only partially adjust when the offer price is revised upwards. We find that the positive relationship between media coverage and underpricing is stronger when ex ante uncertainty is greater, and fail to find any relationship between positive media coverage and IPO firms' long run underperformance. Overall, our findings are consistent with theories of underpricing being driven by the need to compensate investors for information acquisition, but are not consistent with investor sentiment or prospect theory explanations.

Social Science Research Network, 2013

ABSTRACT This study follows Sias (2004) in examining whether securities dealers’ herding behavior... more ABSTRACT This study follows Sias (2004) in examining whether securities dealers’ herding behavior exists and what is the main reason for their herding in the Taiwan stock market. By testing the cross-sectional dependence in dealers’ demand over two adjacent weeks and decomposing the data into their own cascades and other cascades, we demonstrate that dealers’ cascades mainly result from other cascades (herding), but their own trades are still significant for securities with at least low-to-medium trading frequency. We find little evidence that dealers’ herding is driven by habit investing in stocks which dealers trade with at least medium-to-high frequency. Momentum trading of dealers accounts for little of their herding behavior, and the obviously positive relationship between dealers’ demand and their lag demand changes little even with their momentum trading being taken into consideration. Most importantly, dealers are more likely to herd in large capitalization securities; thus, investigative herding rather than informational cascades are the main reason for dealers’ herding in the Taiwan stock market.

Social Science Research Network, 2013

ABSTRACT In line with Wermers (1999), Choe et al. (1999) and Sias (2004), this paper extends Sias... more ABSTRACT In line with Wermers (1999), Choe et al. (1999) and Sias (2004), this paper extends Sias (2004)’s dynamic herding measure to identify the behavioural type of cascading or herding of UK fund managers in the stock market, their reasons for herding, whether their herding behaviours are different during bullish and bearish periods and whether or not their herding behaviours are informational. Our results demonstrate that UK fund managers’ cascades primarily result from their herding. Because they are more likely to herd in large capitalisation securities, their herding may result from investigative herding, contrary to the finding of Sias (2004). Moreover, our finding that fund managers’ cascades primarily result from their herding, which does not change in bullish and bearish stock markets. We also find that growth-type and international-type funds are more likely to herd with funds similar to their types while value-type funds are more likely to herd with funds different from their types. In addition, we confirm that positive correlation exists between the fraction of fund managers buying and subsequent stock returns, consistent with the results of Wermers (1999), Choe et al. (1999) and Sias (2004). We advance the herding results of Hung et al. (2010) to prove that their herding behaviours belong to informational herding within the subsequent year. The implication of our finding is that investors could follow UK fund managers to purchases stocks overbought by them with at least 15 traders quarterly in the following one year period, especially for growth, specific sector and international funds to improve the performances of their equity portfolios.

This working paper is produced for discussion purpose only. These working papers are expected to ... more This working paper is produced for discussion purpose only. These working papers are expected to be published in due course, in revised form, and should not be quoted or cited without the authors written permission.

This paper first extends Sias (2004) and Wylie (2005) to examine the existence of and the reasons... more This paper first extends Sias (2004) and Wylie (2005) to examine the existence of and the reasons for UK fund managers' herding behaviours in the stock market. Whether their herding behaviours are different during bullish and bearish periods and whether or not their herding behaviours are informational are explored. Our results demonstrate that UK fund managers' cascades primarily result from their herding, and habit investing does not primarily drive their herding behaviour. Different from the results of Sias (2004) and Wylie (2005), we find that momentum trading can be regarded as one of the main reasons for UK fund managers' herding. Because they are more likely to herd in large capitalisation securities, their herding may result from investigative herding, which is contrary to the finding of Sias (2004). Moreover, our results find that UK fund managers' cascades primarily result from their herding, which does not change in bullish and bearish stock markets. We al...

Emerging Markets Finance and Trade, 2014

ABSTRACT We use a dynamic herding measure to explore the causes of foreign institutional investor... more ABSTRACT We use a dynamic herding measure to explore the causes of foreign institutional investor (FII) herding in the Taiwan stock market and examine the effects of stock characteristics on the direction and extent of such herding. We find that FII herding primarily results from cascades rather than habit investing or momentum trading. The result of a panel smooth transition regression shows that FIIs' negative cascades focus on their largest net purchases of stocks, but FIIs' positive cascades focus on winner and small-sized stocks. To increase portfolio returns, investors can use FIIs' cascades to inform their stock purchases.

SSRN Electronic Journal, 2013

ABSTRACT This study follows Sias (2004) in examining whether securities dealers’ herding behavior... more ABSTRACT This study follows Sias (2004) in examining whether securities dealers’ herding behavior exists and what is the main reason for their herding in the Taiwan stock market. By testing the cross-sectional dependence in dealers’ demand over two adjacent weeks and decomposing the data into their own cascades and other cascades, we demonstrate that dealers’ cascades mainly result from other cascades (herding), but their own trades are still significant for securities with at least low-to-medium trading frequency. We find little evidence that dealers’ herding is driven by habit investing in stocks which dealers trade with at least medium-to-high frequency. Momentum trading of dealers accounts for little of their herding behavior, and the obviously positive relationship between dealers’ demand and their lag demand changes little even with their momentum trading being taken into consideration. Most importantly, dealers are more likely to herd in large capitalization securities; thus, investigative herding rather than informational cascades are the main reason for dealers’ herding in the Taiwan stock market.

SSRN Electronic Journal, 2013

SSRN Electronic Journal, 2013

ABSTRACT In line with Wermers (1999), Choe et al. (1999) and Sias (2004), this paper extends Sias... more ABSTRACT In line with Wermers (1999), Choe et al. (1999) and Sias (2004), this paper extends Sias (2004)’s dynamic herding measure to identify the behavioural type of cascading or herding of UK fund managers in the stock market, their reasons for herding, whether their herding behaviours are different during bullish and bearish periods and whether or not their herding behaviours are informational. Our results demonstrate that UK fund managers’ cascades primarily result from their herding. Because they are more likely to herd in large capitalisation securities, their herding may result from investigative herding, contrary to the finding of Sias (2004). Moreover, our finding that fund managers’ cascades primarily result from their herding, which does not change in bullish and bearish stock markets. We also find that growth-type and international-type funds are more likely to herd with funds similar to their types while value-type funds are more likely to herd with funds different from their types. In addition, we confirm that positive correlation exists between the fraction of fund managers buying and subsequent stock returns, consistent with the results of Wermers (1999), Choe et al. (1999) and Sias (2004). We advance the herding results of Hung et al. (2010) to prove that their herding behaviours belong to informational herding within the subsequent year. The implication of our finding is that investors could follow UK fund managers to purchases stocks overbought by them with at least 15 traders quarterly in the following one year period, especially for growth, specific sector and international funds to improve the performances of their equity portfolios.

SSRN Electronic Journal, 2013

ABSTRACT The study extends Sias’s (2004) method of analysing institutional investors’ inter-tempo... more ABSTRACT The study extends Sias’s (2004) method of analysing institutional investors’ inter-temporal herding in securities to examine whether the herding effect among banks in industrial lending exists in Taiwan, determine the reasons for herding among these banks, and examine whether their herding behaviour differs across bullish and bearish periods and bank types. This study demonstrate that the cascades among banks in Taiwan primarily result from their herding and that habit loans are not the primary factor in banks’ herding behaviours. Because banks are more likely to engage in herding behaviour in the case of lending to large-capitalisation industries, their behaviours may stem from investigative herding. Moreover, our results show that banks’ cascades result primarily from their herding, which does not differ across bullish and bearish periods. Meanwhile, their herding behaviours are more significant for middle- and large-capitalisation industries in bullish periods. This study also finds that private banks are more likely to follow the same type of bank classification, whereas public banks are more likely to follow a different type. Commercial banks are more likely to follow the same type of bank class, whereas SME (small and medium-sized enterprise) banks are more likely to follow a different type. Our finding confirms the reputational herding behaviour by commercial banks in Taiwan. This study also finds that banks outside the system of financial holding companies are more likely to engage in a different type of herding, which may be evidence of characteristic herding among banks outside the financial holding system that follow those within the financial holding system. Finally, large banks are more likely to follow banks of the same type, whereas small banks are more likely to follow different types of banks, a finding consistent with the results of Barron and Valve (2000). To our best knowledge, this study is the first research which systematically explore banks’ herding behaviour on corporate loans in Taiwan.

Social Science Research Network, 2013

We apply text mining (TM) techniques to extract and quantify relevant Chinese financial news, in ... more We apply text mining (TM) techniques to extract and quantify relevant Chinese financial news, in an attempt to further develop the classical early warning models of financial distress. We extend the work of Demers and Vega (2011) by proposing a measure of the degree of credit default, referred to as the ‘distress intensity of default-corpus ’ (DIDC), and investigate the predictive power of DIDC on default probability by incorporating it into the signaling model, along with the classical financial performance variables (liquidity, debt, activity and profitability ratios). We construct a logistic regression (LR) model to better integrate the DIDC and financial performance variables into a more effective early warning signal model, with the incorporation of DIDC into the LR model revealing a significant reduction in Type I errors and an apparent increase in classification accuracy, thereby proving the effectiveness of the additional information from TM on financial corpus, and also con...

Journal of Financial Studies, 2012

This study examines the determinants of New Taiwan Dollar interest rate swap (NTD IRS) spreads. P... more This study examines the determinants of New Taiwan Dollar interest rate swap (NTD IRS) spreads. Prior literature provides evidence that the term structure of interest rates, liquidity, and credit risk comprise the swap spreads. The term structure of interest rates includes the interest rate level, the slope of the yield curve, and the interest rate volatility. The empirical results for the full sample period show that the interest rate level, the slope of the yield curve, interest rate volatility, liquidity risk, and credit risk are all important factors affecting the swap spreads. Results of the variance decomposition for the full sample period indicate that default risk is the most important factor among the five components. Furthermore, the results show that the explanatory power of the regression model for the term structure of interest rates, liquidity, and credit risk is higher in a bear market than in a bull market. Finally, the results of variance decomposition indicate that...

Romanian journal of economic forecasting

This study explores the dynamic relationship between the sentiment of institutional investors and... more This study explores the dynamic relationship between the sentiment of institutional investors and market returns in the futures market. Using data from the Taiwan futures market, the empirical results show that the dynamic relationship between the sentiment of foreign institutional investors and the futures returns is much stronger than that of the sentiment of domestic institutional investors and the futures returns. Our empirical results also display that the sentiment of foreign institutional investors Granger-causes the sentiment of domestic institutional investors, but not vice versa. Finally, the sentiment of foreign institutional investors has a larger effect on subsequent market returns and market states than that of the sentiment of domestic institutional investors. Overall, our empirical results suggest that the relationship among the institutional investor sentiment, market returns, as well as market conditions in the Taiwan futures market is dominated by the sentiment of...

SSRN Electronic Journal, 2008

This paper performs an empirical investigation on the relationship among the possibility of finan... more This paper performs an empirical investigation on the relationship among the possibility of financial distress, firm performance, corporation governance and macroeconomic environment using the companies listed on the Taiwan Stock Exchange (TSEC) and the financial distress events took place between 1997 and 2003. It is observed that the firms on the brush with bankruptcy defined as follows: insolvent companies; firms under 100%-cash margin requirement; trading-halt companies; firms de-listed from TSEC; net worth is negative; serious loss. The sample data set are formed according to the industry-matched and size-matched principle. We build up a quarters-ahead-predicting model of financial distress allowing for measuring treatment effects of governance characteristics. Meanwhile, we present a "moving-window with augmenting information set" research design to evaluate the capacity of our forecasting model. It is found that the capacity of poor company to come out of financial distress is determined by how transparent the governance is, especially for one-quarter ahead of prediction. All of the poor companies falling into the observed group of CEO with high stock-holding and those pledging ratio almost went into bankruptcy. Through the comparison of moving-window research design, we found that the early warning model with treatment effects performs better in prediction power than those competitive models.

SSRN Electronic Journal, 2013

ABSTRACT The study extends Sias’s (2004) method of analysing institutional investors’ inter-tempo... more ABSTRACT The study extends Sias’s (2004) method of analysing institutional investors’ inter-temporal herding in securities to examine whether the herding effect among banks in industrial lending exists in Taiwan, determine the reasons for herding among these banks, and examine whether their herding behavior differs across bullish and bearish periods and bank types. The study results demonstrate that the cascades among banks in Taiwan primarily result from their herding and that habit loans are not the primary factor in banks’ herding behaviors. Because banks are more likely to engage in herding behavior in lending to large-capitalization industries, their behaviors may stem from investigative herding. Moreover, our results show that banks’ cascades result primarily from their herding, which does not differ across bullish and bearish periods. Meanwhile, their herding behaviors are more significant for middle- and large-capitalization industries in bullish periods. The study also finds that private banks are more likely to follow the same type of bank classification, whereas public banks are more likely to follow a different type. Commercial banks are more likely to follow the same type of bank class, whereas SME (small and medium-sized enterprise) banks are more likely to follow a different type. This finding is consistent with reputation herding by commercial banks in Taiwan. Next, the study finds that banks outside the Taiwanese financial holding system are more likely to engage in a different type of herding, which may be evidence of characteristic herding among banks outside the financial holding system that follow those within the financial holding system. Finally, larger banks are more likely to follow banks of the same type, whereas smaller banks are more likely to follow different types of banks.

SSRN Electronic Journal, 2013

This study extends Sias (2004) to examine whether herding exists among foreign institutional inve... more This study extends Sias (2004) to examine whether herding exists among foreign institutional investors (FIIs), what is the cause of their herding, and whether their herding is stable across both bullish or bearish periods or institutional types in the Taiwan stock market. By testing the cross-sectional dependence in the FIIs' demand for stocks in two adjacent months and decomposing them into their own cascades and other cascades, we demonstrate that the FIIs' cascades mainly result from their herding for securities traded at medium to high frequency even though their own cascades still exist. We find little evidence that FIIs' herding behavior is driven by habit investing in stocks which FIIs trade with at least medium to high frequency. The momentum trading of FIIs is found to account for little of their herding, and the obviously positive relationship between the FIIs' demand and their lag demand changes little, even with their momentum trading being taken into consideration. Moreover, FIIs are more likely to herd in large capitalization securities, and thus investigative herding, rather than informational cascades, is the main reason for herding among FIIs in the Taiwan stock market. One of our contributions may be to find that the phenomenon that FIIs' cascades mainly result from their herding does not change in the bullish and bearish Taiwan stock market. However, since FIIs are more likely to herd in large-capitalization securities, their herding will obviously increase (decline) when they herd in large-capitalization (small-capitalization) stocks in the bullish (bearish) stock market period. This study further finds that FIIs and dealers are more likely to follow similar-type institutions than different-type institutions, respectively, which is consistent with the view proposed by Del Guercio (1996) and Bennett, Sias and Starks (2003) that institutional herding is driven by reputational or characteristic herding. Security dealers in Taiwan tend more often to follow same-type herding while FIIs more often negatively follow different-type herding, which implies that security dealers exhibit the strongest tendency to herd especially in same-type institutions possibly due to reputation concerns, which is consistent with the view put forward by Sias (2004).

SSRN Electronic Journal, 2011

This study investigates whether the news release prior to the earnings announcement would influen... more This study investigates whether the news release prior to the earnings announcement would influence different investors" trading behavior during the earnings announcement periods. We bridge the gaps between this study and the previous analyses that the technology of text mining is applied to extract the Chinese news, the proxies of the news information are constructed and the comprehensive tick data is used to categorize trader types including dealers, mutual funds, foreigners, individuals and corporations. The empirical results show that individuals may have significant reactions to the news report prior to the earnings announcement. However the news report may have less effect to institutions" trading behavior. The findings confirm that institutions are more rational than individuals.