Rational Expectations and Importance of the Loss Function Assumption: Evidence from Farmers’ Harvest Forecasts
Most tests of the rational expectations hypothesis (REH) in agricultural settings implicitly assume that farmers have a quadratic loss function. This raises the possibility of false rejection of REH if the loss function is not quadratic, even when expectations are being formed rationally under the true loss function. We use survey data on harvest expectations and actual harvests to test REH under alternate loss function assumptions. We reject REH under a quadratic loss function, but cannot reject it under a linear loss function. Our findings suggest caution when implicitly assuming quadratic loss functions to test rationality in agricultural settings.
|Journal||The Empirical Economics Letters|