Earnings moves and pre-earnings implied volatility

Arjun K. M., Mike Lipkin and Leon Tatevossian

Abstract

We investigate price moves by listed American stocks upon the release of quarterly earnings data to show the relationship between these (frequently outsized) return realizations and pre-earnings implied volatility. Specifically, we examine short-expiry and intermediate-expiry implied volatility on the business day immediately preceding the earnings date. In conjunction with standard modeling assumptions, we extract from this implied volatility data a derived variance for the instantaneous return in the upcoming earnings window. We examine, for three discrete years, both the raw realized return and, based on this derived variance, the predicted magnitude of that return. Two naive models are used for the predicted move, one of which appears to work very well. Across the three calendar years of our study there is a degree of consistency. The raw price moves are highly fat-tailed and, by our choice of measure, strongly peaked about the origin, with 2.5% in each of the tails. In most cases the options market does a good job of predicting the (magnitude of the) price impact of the earnings event, but again there are significant numbers of outliers. It is accepted wisdom that, in aggregate, the distribution of returns over earnings windows is symmetric, with equal incidence of up and down moves. We find this to be true. Practitioners are generally aware of this behavior, yet we believe such results have not previously been published. Our observation of the apparent stability of this behavior across a decade and a half, as well our remarks on the profile of the aggregate return distribution, may make these findings useful to the trading community.