If we do not observe herding in the above setting, then the conclusion that institutional investors herd and destabilize prices is questionable.

Lack of herding at the aggregate 4 level, however, does not eliminate the possibility that there were cross-sectional differences in the level of herding across different types of institutional investors (e.g., mutual funds versus banks). On the other hand, if we do find herding, then in addition to highlighting the role of institutional investors in one of the largest price run-ups in US stock markets, we would be in a better position to comment on the circumstances in which herding manifests itself.

It would also be interesting to determine whether institutional investors herded more on the buy-side or sell-side during the run-up than during the subsequent decline. If institutional investors have better information than individuals then, in aggregate, institutions should have herded more on sell-side when the market was going up (contrarian strategy) and individuals should have been herding more on buy side.

We would expect to see exactly the opposite behavior during the price decline. Furthermore, the claim that institutional herding, if present, is stabilizing in nature could be examined. In addition, the observed cross-sectional differences in herding between different types of institutions could help us in better understanding the different roles of institutions, which has not been documented so far.

Our first major finding is that in contrast to prior research, institutional investors, regardless of the type, herded into internet stocks with substantial intensity. Second, institutional investors exerted no correcting force on new economy stocks during the period of 1998 to March 2000. On the contrary, they participated in it by buying in herds. We find no cross-sectional differences in herding by banks, insurance firms, mutual funds, etc. Thus, herding by institutions was much more than what could be explained by inflow of funds to internet funds by retail investors.

Third, amongst internet stocks, herding was spread across all size quintiles, and was largely unrelated to past performance. Lastly, institutional investors’ buying exerted upward price pressure, and the reversal of excess returns in the subsequent quarter indicates that institutional trading was destabilizing and not based on information. However, the results on institutional selling are consistent with the idea that institutional herding enhances market efficiency by incorporating information into stock prices.

Related Pages

Tags: , , ,