Estimation and Testing of Forecast Rationality under Flexible Loss
提出一种方法，在未知损失函数形状时，从预测序列中反推出与理性预测一致的损失函数参数，并构建允许非对称损失的新理性检验，应用于IMF和OECD对G7国家预算赤字预测。
In situations where a sequence of forecasts is observed, a common strategy is to examine \n“rationality” conditional on a given loss function. We examine this from a different perspective— \nsupposing that we have a family of loss functions indexed by unknown shape parameters, then given \nthe forecasts can we back out the loss function parameters consistent with the forecasts being rational \neven when we do not observe the underlying forecasting model? We establish identification of the \nparameters of a general class of loss functions that nest popular loss functions as special cases and \nprovide estimation methods and asymptotic distributional results for these parameters. This allows us \nto construct new tests of forecast rationality that allow for asymmetric loss. The methods are applied \nin an empirical analysis of IMF and OECD forecasts of budget deficits for the G7 countries. We find \nthat allowing for asymmetric loss can significantly change the outcome of empirical tests of forecast \nrationality.