Measuring predictability: theory and macroeconomic applications

The authors propose a measure of predictability based on the ratio of the expected loss of a short-run forecast to the expected loss of a long-run forecast. This predictability measure can be tailored to the forecast horizons of interest, and it allows for general loss functions, univariate or multivariate information sets, and stationary or nonstationary data. The authors propose a simple estimator and suggest resampling methods for inference. They then provide several macroeconomic applications. First, on the basis of fitted parametric models, the authors assess the predictability of a variety of macroeconomic series. Second, they analyze the internal propagation mechanism of a standard dynamic macroeconomic model by comparing predictability of model inputs and model outputs. Third, they use predictability as a metric for assessing the similarity of data simulated from the model and actual data. Finally, the authors sketch several promising directions for future research.