Abstract
We examine asymmetry in the loss function of corporate executives in their exchange rate forecasts and test for rationality of the forecast under the assumption of a possibly asymmetric loss function. We find evidence that the loss function of the forecast with a 1-year horizon is symmetric and support for rationality. These findings hold for manufacturing and nonmanufacturing sectors.
Notes
1 Since a weighting matrix depends on an estimate of α, our estimation is performed iteratively, assuming a weighting matrix in the first round to be an identity matrix. The continuously updating estimator of Hansen et al. (Citation1996) with the Bartlett kernel is used for possible gains to finite-sample efficiency.
2 A constant is included as an instrument in all models considered in this article. However, for notational convenience, it is dropped hereafter.
3 Because all empirical results are qualitatively equivalent, we do not explicitly mention it hereafter.