In this study, we separate institutional investors into the categories of transient, dedicated, and quasi-indexer and examine whether the ownership percentages of these different types of institutional investors are associated with the accuracy of analysts’ earnings forecasts. We hypothesize a positive association between ownership by transient investors and earnings forecast accuracy, as forecast accuracy may be a signal of decreased price impact of trades and increased profit trading opportunities. We predict a negative association between ownership by dedicated investors and earnings forecast accuracy because more accurate forecasts eliminate dedicated investors’ informational advantage. Finally, we predict that forecast accuracy is not relevant for the investment decisions of quasi-indexers.

Our empirical evidence suggests that transient investors are indeed drawn to companies with lower forecast errors and increase (decrease) their holdings when forecast errors decrease (increase). However, the evidence on the behavior of dedicated investors and quasi-indexers reveals more complex decision pattern, consistent with investment decisions not driven by analysts’ information. Dedicated investors’ ownership decreases with forecast accuracy, especially earnings are lower than forecasts, and are more likely to increase ownership levels when forecast accuracy decreases. Quasi-indexers appear to reduce ownership as forecast accuracy decreases, but only when earnings exceed forecasts.

These findings highlight the importance to adjust for investors’ heterogeneity in research models rather than to cast institutional investors as a homogeneous group. Our findings also indicate that attempts to achieve "predictable earnings" appeal primarily to one type of institutional investors: transient.