An unobserved component panel data model to study the effect of earnings surprises on stock prices, trading volumes, and spreads

Previous empirical work on the effect of earnings announcements on trading volume and bid-ask spreads relied on OLS estimation of equations using price changes or errors in analystsʹ earnings forecasts as proxies for the unobserved earnings surprises. We show that this leads to substantial errors in variables biases (which account for the often insignificant relationships found earlier). We estimate an unobserved components model by instrumental variable method using both these variables as proxies. The results show significant effects of earnings surprises on price, volume, and bid-ask spreads.