Implied Biases in Forecasts of Earnings for Different Horizons

Security analysts generally provide forecasts of earnings for the current period as well as oneyear ahead earnings at fiscal year end. In this study, we derive an estimation procedure, which infers forecast bias from equivalent price expressions that utilize different horizon earnings forecasts. It is well documented that analyst long-horizon earnings forecasts tend to be more optimistic (ex post) than short-horizon forecasts. We develop a system of equations that simultaneously estimates the expected rates of return and growth rates implied by analysts’ forecast of current period earnings and one-year ahead earnings. We attribute the difference between the two expected rates of returns as the relative extent to which investors adjust analyst earnings forecasts (downward) farther out when calculating the present values of future payoffs, knowing that earnings forecasts for more distant period are, on average, more optimistically biased. We refer the adjustment to forecast of one-year ahead earnings as our measure of ex ante forecast bias. We find that investors, on average, adjust one-year ahead earnings forecasts downwards by approximately 10%. To further validate our bias measure, we partition our sample using determinants, which have been shown, in prior literature, to explain ex post forecast bias. Our results show that risk-return proxies and information content of forecasts identified in prior literature as explaining ex post forecast bias also explain ex ante forecast bias. In addition, investors do not heavily weight accounting attributes in their adjustment of the more distant earnings forecasts. In sum, we develop an estimation procedure, which estimates the implied expected rate of return and growth rate from analyst forecasts, while adjusting for the forecast bias in more distant earnings. Furthermore, our ex ante forecast measure is the first to quantify investors’ adjustment of earnings forecasts at different horizon at consensus forecast formation.