Once Lucky, Always Lucky? Institutional Trading in a Connected World

We employ transaction-level trading data to examine institutional trading before the public revelation of option backdating and find that some institutions made lucky trades, i.e., they trade ahead of other investors and gain significantly. Based on these trades, we develop a proxy for the luckiness of a given institution on a specific firm and examine whether institutions continue to be lucky in their trades on the same set of firms outside the backdating setting. We find that institutions are more likely to make correct trades on their lucky firms before earnings announcements, and that in general they trade more actively and perform better on these firms. We provide evidence that these results cannot be explained by institutions extracting information from publically available data. Instead, our results are more consistent with the conjecture that institutions repeatedly exploiting preferential information through private institution-firm connections.