Horizon Bias in Expectations Formation

Abstract

We provide empirical evidence that the optimism bias in expectations of economic quantities increases with the forecasting horizon. We label this empirical regularity the horizon bias. We document significant horizon bias in the macroeconomic expectations of professional forecasters, both in the U.S. and abroad. We show that horizon bias is unlikely to be the result of career concerns, asymmetric loss, information rigidities, or cognitive mistakes. We then provide theoretical support to the notion that horizon bias can arise within theories of motivated beliefs. Finally, following the conceptual framework of Benabou(2015), we test whether theory-based drivers of motivated beliefs can explain time-series variation in the horizon bias. We find strong support for many of these drivers. Overall, we offer strong evidence that motivated beliefs lead to biases in expectations formation.

Keywords: Motivated beliefs, Optimal expectations, Wishful thinking, Economic forecasting, Economic models, Expectations, Optimism Bias, Survey Forecasts

JEL Classification: C53, D03, D83, D84, D90, D91, E37