We find analysts convey information about a firm’s earnings without fully revising their earnings forecast by increasing bundling intensity, which is the extent to which an analyst report that has an earnings forecast revision includes also price target and/or recommendation revisions with the same sign as the earnings forecast revision. We develop a firm-level measure of bundling intensity, BF_Score, and find it is an economically meaningful predictor of analyst-based earnings surprises. The surprises reflect bias in consensus earnings forecasts related to information analysts convey through bundling intensity. Analysts’ use of bundling and the predictive power of BF_Score are higher when macroeconomic uncertainty is higher, which is when analysts’ incentives to avoid bold earnings forecast revisions are greater. Additionally, firms with higher BF_Score are more likely to report earnings that barely meet or beat the consensus forecast. This finding suggests analysts make more beatable earnings forecasts to curry favor with management by bundling rather than reflecting all the positive news in higher earnings forecasts. Adjusting analyst-based earnings surprises for the implications of BF_Score results in a distribution of earnings surprises that more closely resembles a normal distribution. Notably, the adjustments reduce the well-known kink asymmetry around zero for consensus analyst forecast-based earnings surprises by 66%, and markedly reduce skewness and kurtosis. Prior research attributes the kink primarily to earnings management. Instead, our findings suggest the kink reflects predictable analyst-based earnings surprises, and highlight the need for research utilizing consensus analyst earnings forecasts and analyst-based earnings surprises to account for these biases.
Note: Research papers posted on SSRN, including any findings, may differ from the final version chosen for publication in academic journals.