US institutions with US$100 million or more in assets under management must disclose part of their long positions via the Securities and Exchange Commission’s Form 13F Holdings Report on a quarterly basis. We consider the number of variations in institutions’ holdings between consecutive reporting periods, and we compute a normalized measure of the discrepancy (or “imbalance”) in the volume of shares bought or sold for each asset. Similarly, we also experiment with imbalances in terms of trade counts. Assets are then divided into quantiles according to the strength of the associated imbalance, whose sign is used to define the direction of investment. We test multiple hypothetical strategies, by varying the quantiles, time horizons and minimum numbers of institutions active on each asset. We compute Sharpe ratios and the related confidence levels, and we compare results to a benchmark that follows a basic price mean-reversion strategy. In this way we show that a significant opportunity for profit arises if an external investor is willing to trade contrary to the 13F filing imbalances. Indeed, imbalances capture the amount of information already consumed in the market, and the related trades tend to be inflated by crowding and herding behavior. Betting on a relatively short-term movement of prices against the sign of imbalances results in a profitable strategy, especially when using a time horizon of 21–42 trading days (ie, one to two calendar months) after the end of each financial quarter.