دانلود رایگان مقاله سوگیری بی ثباتی زمانی و وضعیت اقتصادی

abstract

This paper aims to investigate whether a forecast is optimal, given the information available when it is made. Going beyond the papers that study forecast errors based on the model of Nordhaus (1987), we use a time-varying procedure to forecast revisions and to account for the possibility that the duration of the state may also affect the bias. Three testable hypotheses are presented to help researchers test the optimality of forecasts, with the ultimate aim of determining whether these biases depend on the underlying economic state and whether they are persistent for the duration of the state. Corresponding biascorrected forecasts can then be made based on these results. The empirical study finds that the one-quarter-ahead official forecast of GDP growth in Taiwan does indeed suffer from state-dependent biases: a persistent under-estimation bias in the relatively good state, and an under-reaction bias that decays with duration in the relatively bad one. Eliminating these biases in the forecast can remove over 44.0% of the variation in forecast errors, and pseudo out-of-sample experiments further support the fact that the resulting biascorrected forecasts are markedly better than those made by Taiwan’s government or using other competing models. © 2016 International Institute of Forecasters. Published by Elsevier B.V. All rights reserved.

4. Concluding remarks

This paper aims to investigate whether or not forecasts are optimal, given the information that was available when the forecasts were made. Going beyond the papers that study forecast errors based on the model of Nordhaus (1987), we use a time-varying procedure to forecast revisions, and also account for the possibility that the duration of the state may affect the bias as well. Three testable hypotheses are presented to help researchers test the optimality of forecasts, with the ultimate aim of determining whether or not these biases depend on the underlying economic state and are persistent over the duration of the state. The corresponding bias-corrected forecasts can then be made based on these results. Briefly, this framework is novel and can be implemented using conventional estimation and hypothesis methods. In the empirical part, we apply the proposed framework to an investigation of Taiwan’s DGBAS forecasts for GDP growth rates. We find that the one-quarter-ahead forecast is not optimal, but actually suffers from state-dependent biases: a persistent under-estimation bias in the relatively good state and an under-reaction bias that decays with the duration in the relatively bad one. Eliminating these biases from the DGBAS forecast can remove over 44.0% of the variation in forecast errors, and pseudo out-of-sample experiments further support the fact that the resulting bias-corrected forecasts are markedly better than either those made by Taiwan’s government or those made using other competing models.