Meeting, Beating, Streaks and Bubbles

Abstract

The literature in economics and finance document that asset bubbles can emerge and remain sustained for a variety of reasons. In this paper, we develop an analytical model to characterize two types of rational bubbles linked to accounting disclosures, drift bubbles and sensitivity bubbles. We conjecture that both types of bubbles are likely to do so when firms experience streaks of good news. We use our analytical model to generate a number of hypotheses that are expected to hold if drift and/or sensitivity bubbles arise when firms experience streaks of good news, which we proxy for using lengths of meet or beat streaks. In price level regressions, we provide evidence consistent with meet or beat streaks associated with both non-linear drift and non-linear sensitivity price bubbles, consistent with the predictions drawn from the model. We offer less compelling evidence consistent with the model predictions that the drift and sensitivity bubbles should accelerate as a firm continues a streak. We provide reasonable evidence consistent with the drift and sensitivity bubbles unwinding when a firm fails to continue a meet or beat streak. Finally, we find evidence that the industry-wide streak related bubble-like patterns are incremental to the firm-specific patterns we document.

Keywords: Earnings response; Earnings association; Meet or beat analyst forecasts; Stock market bubbles; earnings string

JEL Classification: M4