Portfolio Returns

Given the statistical evidence that autocorrelation coefficients are different during heavy institutional trading, we want to examine the profitability of a trading strategy that exploits the time-varying autocorrelation coefficients. The profitability provides a measure of economic significance.

We begin with a benchmark strategy that buys stocks with a positive return and sell them with a negative return. This strategy should generate a positive return given that the average autocorrelation coefficient is positive. During the sample period, we divide all stocks into two groups based on the sign of the return on day t: one group contains all stocks with positive returns and the second group contains stocks with negative returns. For each group, we first calculate the equal-weighted average return on day t+1 and then calculate its time-series average return over the sample period. To test the significance of the average return, we use the Newey-West standard errors of ten lags to account for possible autocorrelations of daily returns. The positive-return group

generates an average daily return of 0.21% and the negative-return group generates an average of -0.05%. The return of an arbitrage portfolio that longs the positive return and shorts the negative is 0.26% and is significant at a 0.01 level. Despite the statistical significance, its magnitude is small compared with a transaction cost of 0.87% (See Appendix for the discussion of the transaction cost).

To exploit the finding of autocorrelation on heavy trading days, we construct the following four portfolios. The first portfolio includes stocks that have heavy total trading volume (higher than its 200-day moving average), heavy institutional buy, and positive return on day t. The second portfolio includes stocks that have heavy total volume, heavy institutional selling, and negative return on day t. The third and fourth portfolios are similar to the first two portfolios except that they include stocks that have low, rather than high, institutional buy or sell. The criteria used to construct portfolios arise from the prediction of the LMSW model: the autocorrelation of returns is higher when total volume is high and the direction of the information trading is the same as the direction of returns.

Table 8 reports portfolio returns on day t+1. Panel A, B, and C report separately the portfolios based on trading from mutual funds, foreigners, and dealers. One thing common is that returns on portfolios based on a positive return on day t are all positive (columns 1 and 3) and that returns on portfolios based on a negative return on day t are all negative (columns 2 and 4). This reflects the fact that returns are positively autocorrelated.

When mutual funds and foreigners buy more on a positive-return day, the return on the next day is higher. For example, for the portfolio with large mutual fund buys, the average return on the next day is 0.51% (column 1, Panel A), whereas for the portfolio with light mutual fund buys and positive returns, the average return is 0.24% (column 3).

Table 8. Daily Portfolio Return Based on Lagged Returns, Aggregate Volume, and Institutional Trading

We construct three sets of portfolios from stocks which have large daily aggregate volume (turnover is larger than its 200-day moving average). Each set of portfolios includes four portfolios: large buy from institutional investor and positive return on day t, large sell and negative return, small buy and positive return, small sell and positive return; the portfolio composition changes daily. The first set of portfolios is based on foreign investors, the second set is based on mutual funds, and the third set is based on dealers. A stock is classified as “large buy (sell)” if the daily buy (sell) turnover is higher than its 200-day moving average, otherwise, it is “small buy (sell)”.

For each portfolio, we first calculate its daily equally weighted return for day t+1, and then calculate and report its time-series average; t-statistics in parentheses are based on Newey-West standard errors with 10 lagged autocorrelations. The sample includes 1,049 stocks for which there are at least 750 daily observations and that were listed on the Taiwan Stock Exchange and the Gre Tai Securities Market. The sample period is from 2000/12/12 to 2007/3/30.

Institutions have

Difference in returns between large and small trade portfolios

Panel A: Portfolios based on mutual funds’ trading

0.508 -0.149 0.244 -0.118 0.264*** -0.031 0.657*** 0.362***

Panel B: Portfolios based on foreigners’ trading

0.422 -0.161 0.275 -0.115 0.147*** -0.046** 0.583*** 0.390***

Panel C: Portfolios based on dealers’ trading

0.235 -0.072 0.329 -0.139 -0.094*** 0.067 0.308*** 0.468***

The difference in return is 0.26% (column 5) and is significant at a 0.01 level. The significant difference is consistent with the LMSW’s prediction that information trading increases the autocorrelation of returns.

Similarly, the portfolio return is more negative if mutual funds sell heavily on a negative-return day. The difference in return is -0.03%, which is not significantly different from zero at a 0.1 level. This nonsignificant difference is consistent with our hypothesis that sell volumes contain less information than bye volumes due to short sale constraints.

If we form an arbitrage portfolio that longs the portfolio with large mutual fund buy and shorts the portfolio with large mutual fund sell, the average return is 0.66%

(column 7) and is more than twice the return on a arbitrage portfolio that only exploits autocorrelation but ignores institutional trading.

The pattern is different for portfolios based on dealers’ trading. When dealers buy more on a positive-return day, the return on the next day is lower rather than higher. For the portfolio with large buys from dealers, the average return is 0.24% (column 1, Panel C), whereas for the portfolio with light dealer-buys, the average return is 0.33%

(column 3). The difference in return -0.09% (column 5) and is significant at a 0.01 level.

The significant difference is consistent with the earlier finding that dealer's trading reduces the autocorrelation of returns.

Andrade, Chang, and Seasholes (2008) find that the change in shares held in margin accounts in Taiwan is a measure of liquidity demand and is related to price reversals. If we take their measure of liquidity demand into account, does it increase or reduce the return of our arbitrage portfolios? If foreigners or mutual funds happen to trade against margin traders, then our results may recede and the return of our arbitrage portfolios will drop significantly. On the other extreme, if the direction of margin

trading is the same as the trading by foreigners or mutual funds, we can improve the return of our arbitrage portfolios by taking the margin trading into account.

Following Andrade, Chang, and Seasholes (2008), we calculate the daily change in shares for each stock held in margin accounts normalized by the number of shares outstanding. We first calculate the correlation coefficients between the imbalance of margin trading and institutional buy or sell volume for each stock and then take the cross-sectional average. The average correlation coefficients are not high; they range from -0.05 to 0.13. Therefore, the imbalance of margin trading is only weakly related to institutional buys or sells.

Next, we examine the profitability of portfolios taking into account both the imbalance of margin trading and institutional trading. Each day, we sort all stocks with heavy trading volume into one of six portfolios based on institutional trading and margin trading. There are three groups (low, medium, and high) of margin trading using the 20^th and 80^th percentile of the imbalance of margin trading as the cutoff points.

There are two groups of institutional trading: large institutional buys on a positive-return day and large institutional sells on a negative-return day. We calculate the time-series average return for each portfolio and report them in Table 9.

As found in Andrade, Chang, and Seasholes (2008), the higher the imbalance of margin trading, the lower the return on the following day. Holding the institutional trading constant, the differences in the average return between the low and high imbalance of margin trading are all positive and range from 0.29% to 0.52%.

Taking into account the imbalance of margin trading, however, does not change our conclusion that the arbitrage portfolios based on institutional trading are profitable.

Holding the imbalance of margin trading constant, the average return of arbitrage portfolios remains positive with a range from 0.45% to 0.88%. Therefore, our results are

Table 9: Daily Portfolio Return Based on Lagged Returns, Aggregate Volume, Institutional Trading, and Margin Trading

We construct two sets of portfolios from stocks which have large daily total volume (turnover is larger than its 200-day moving average). Each set of portfolios includes six portfolios based on two criteria:

the first criteria is based on institutional trade and return (large buy from institutional investor and positive return on day t, large sell and negative return), the second criteria is based on the net margin trading (Z_i,t), which is the daily change in shares held in margin accounts normalized by the number of shares outstanding (Andrade, Chang, and Seasholes, forthcoming). The portfolio composition changes daily. In Panel A, each day we sort the stocks with large daily total volume into five quintiles based on Zi,t. In Panel B, each day we sort the stocks in the largest size quartile and with large daily total volume into five quintiles based on Z_i,t.The first quintile, the second to fourth quintiles, and the fifth quintile are denoted by Low Z_i,t, Medium Z_i,t, and High Z_i,t, respectively. The first set of portfolios is based on foreign investors, and the second set is based on mutual funds. A stock is classified as “large buy (sell)”

if the daily buy (sell) turnover is higher than its 200-day moving average, otherwise, it is “small buy (sell)”. For each portfolio, we first calculate its daily equally weighted return for day t+1, and then calculate and report its time-series average; t-statistics in parentheses are based on Newey-West standard errors with 10 lagged autocorrelations. The sample includes 1,049 stocks for which there are at least 750 daily observations and that were listed on the Taiwan Stock Exchange and the Gre Tai Securities Market. The sample period is from 2000/12/12 to 2007/3/30.

Based on mutual funds trading Based on foreigners’ trading Net

not driven by the liquidity demand of margin traders.

We can even improve the performance of our arbitrage portfolios if we combine information trading with margin trading. For example, we can long the portfolio that includes stocks with a low imbalance of margin trading and a strong buy from mutual funds, and short the portfolio that includes a high imbalance of margin trading and a strong sell from mutual funds. The average daily return of the improved arbitrage portfolio would be 1.17%, which is higher than the round trip transaction cost of 0.87%.