CHANGE LANGUAGE | Home > Doc > Continuous overreaction, insiders trading activities and momentum strategies > Introduction

Continuous overreaction, insiders trading activities and momentum strategies


Empirical findings and discussions

Insider trading refined momentum strategies

Adjust the size and BM factors

Long-term reversal



Continuous overreaction, insiders trading activities and momentum strategies


A number of papers have provided evidence that the trading strategies based on past securities returns can beat the market. Jegadeesh and Titman (1993) show that the strategies by buying stocks with high returns over the previous 3–12 months and selling stocks with poor returns over the same time period can generate significant abnormal returns over a medium holding period of 3–12 months. A lot of explanations have been provided.

Jegadeesh and Titman (1993) suggest that a more sophisticated model based on investor behaviour is needed to explain the anomaly, while some others as Conrad and Kaul (1998) attribute the source of the profits as the cross-sectional difference of mean stock returns. Grinblatt and Moskowitz (1999) proposed that the industry momentum is the key element in explaining the return persistence anomalies. However, we find that the industry factor does not sound that good in explaining our insiders trading activities refined momentum strategies.

Grundy and Martin (2001) also document that industry momentum alone does not explain the profitability of momentum strategies. Nowadays, more and more people don’t believe that the investors act rationally as assumed in the classical asset pricing theory. Some researchers turn to look for the irrational explaining of the momentum effects. Most of these behavioural studies are based on the experimental cognitive psychology findings.

Daniel, Hirshleifer and Subramanyam (Daniel et al., 1998) propose a continuous overreaction model based on two psychological findings: people are overconfident on the private signals and attribution bias. They attribute the momentum effects to the overreaction to the private signals and underreaction to the public signals and the eventual correction by public signals is the cause of long-term reversals.

On the other hand, Barberis, Shleifer and Vishny (Barberis et al., 1998) propose that the short-term momentum effects and long-term contrarian effects are caused by the investors’ falsely perceiving that there are two earning regimes. Hong and Stein (1999) propose another irrational model that does not rely on psychology findings. In their model, there are ‘two boundedly rational agents’: newswatchers and momentum traders.

The underreaction of newswatchers causes short-term momentum effects. The early momentum traders will take the chance and profit by trend-chasing and push up/down the winner/loser’s price further and attract more momentum traders enter in. These later entered momentum traders will push up/down the winner/loser’s price too high and cause overreaction.

The motivation of the study is that if the momentum effects are the results of investors irrational reaction, how will they react on the insiders trading activities? The paper investigates the influence of insiders trading and explanatory power of these trading activities on the momentum effects. As described in some literatures, the insiders, such as executives and managers, should know much more about their own company than any outsiders. Many previous literatures have provided evidence that insiders trading activities can predict the cross-sectional returns.

Seyhun (1992) shows that the aggregate insiders trading activity predicts cross-sectional future stock returns during the period 1975–1989. He attributes his findings as ‘both the changes in business conditions as well as movements away from fundamentals contribute to the information content of aggregate insider trading’. Lakonishok and Lee (1998) also provide evidence that the aggregate insiders trading can predict the market movements over the period from 1975 to 1995.

Consistency, we also find that the aggregate insiders trading activities contain valuable information in predicting cross-sectional stock returns during the period from 1985 to 1997. Hence, the insiders trading should be a good indication of how prospective the company is. In the study, we introduce the insiders trading activities in the naı¨ve momentum strategy portfolio selection process to refine the naı¨ve momentum strategies.

All stocks are sorted by two standards: past return performance and insiders trading activities in the previous period. We then form portfolios on the intersections of the sorted groups on the two standards. The performance of these intersectional portfolios is studied. The stocks performed good (bad) and bought (sold) by insiders in the past falls into the trend regime of Barberis et al. (1998) and stocks performed good (bad) and sold (bought) by insiders in the past falls into mean-reverting regime of Barberis et al. (1998).

As predicted by Barberis et al. (1998), we may observe that the momentum effects are negative among stocks in the trend regime because of overreaction and positive among stocks in the mean-reverting regime because of underreaction. As stated above, Daniel et al. (1998) tells a different story. The story of Daniel et al. (1998) is that the short-term momentum effects and long-run reversals are caused by the selfattribution bias. Specifically, Daniel et al. (1998) predicts that stocks performed good (bad) and bought (sold) by insiders in the past will show a very strong momentum effect and long-run reversal.

The contribution of the paper lays on the follows: First, The insiders trading activities have the ability to predict cross-sectional stock returns in US market during the period January 1985–November 1996. We find that it can earn positive profit as long as 36 months after the portfolio formation. Second, the momentum effects exist among stocks in US markets during our study period. We repeated the methods of Jegadeesh and Titman (1993).

The results are similar with their findings. Third, the risk factors cannot explain the insiders trading activities refined momentum effects. The momentum effects are still obvious after we control the size and BM factors. Forth, the industry factor has weak influence on our refined momentum effects. Fifth, we attribute the continuous overreaction as the source of momentum effects.

Our results supports the prediction made by the model of Daniel et al. (1998). The remainder of the paper is organized as follows. Section 2 describes the databases and the insiders trading activity measurement methods used in the study. Section 3 reports the empirical findings are discussions. Section 4 concludes the paper.

Databases and measurement methods of insiders trading activities

Our sample includes all non-financial common shares listed in CRSP and COMPUSTAT files. Because of data availability of insiders trading data record, our research only covers the period from January 1985 to November 1996. Section 16 of Securities and Exchange Act of 1934 (SEA) requires all insiders report any transactions with the SEC by 10th of the month following their transaction. SEA defines an insider as an officer, director, an individual in policymaking position or a beneficial owner (holder of 10% or more).

We drew our data from the First Call Insider Research (First Call). In our research, we only consider open market transaction of common shares.In the study, we choose the aggregate insiders trading ratio (AITR) as the measurement method. Seyhun (1992) used the net trading numbers as the measurement methods of aggregate insiders trading, while Lakonishok and Lee (1998) use the net trading ratio as the measurement of insiders trading activities. The reason we choose the method is that the method makes the balance between the quality and quantity of insiders trading information.

When the Net Trading Number method is used, an insider-buy is recognized when we observe more buy transactions executed by insiders during the forming period, i.e. we pay more attention to the quality of these transactions, and less to the quantity of these transactions. On the other hand, when we use the Net Trading Shares method, we pay more attention to the transaction quantity and less to the transaction quality. By using the AITR, we can quantify the strength of these trading signals. The aggregate insiders trading ratio is defined as:


where B is dollar value of shares bought by insiders during the forming period, S is the dollar value of shares sold by insiders in the forming period. If the ratio is greater than 0, we treat this as an insiders-buy, and if the ratio is less than 0, we treat it as an insiders-sell.

Jihong Xiang, Jia He , Min Cao

Performance Trading

Home | Mappa | Staff | Disclaimer | Privacy | Supportaci | Contact

Copyright © ed il suo contenuto sono di esclusiva propriet� di DHDwise. E' vietata la riproduzione anche parziale di qualsiasi parte del sito senza autorizzazione compresa la grafica e il layout. Prima della consultazione del sito leggere il disclaimer. Per informazioni consultare la sezione info.