A new method to assess the degree of information rigidity using fixed-event forecasts

• Published in: International journal of forecasting Vol. 37; no. 4; pp. 1576 - 1589
• Main Authors: Vereda, Luciano, Savignon, João, Gouveia da Silva, Tarciso
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01-10-2021
• Subjects: Forecast behavior; Inflation forecasts; Noisy information; Output growth forecasts; Sticky information; Noisy information; Sticky information; Forecast behavior; Inflation forecasts; Output growth forecasts
• ISSN: 0169-2070; 1872-8200
• DOI: 10.1016/j.ijforecast.2021.03.001

We propose a new method to explore the information content of fixed-event forecasts and estimate structural parameters that are keys to sticky and noisy information models. Estimation follows a regression-based framework in which estimated coefficients map one-to-one with parameters that measure the degree of information rigidity. The statistical characterization of regression errors explores the laws that govern expectation formation under sticky and noisy information, that is, they are coherent with the theory. This strategy is still unexplored in the literature and potentially enhances the reliability of inference results. The method also allows linking estimation results to the signal-to-noise ratio, an important parameter of noisy information models. This task cannot be accomplished if one adopts an “agnostic” characterization of regression errors. With regard to empirical results, they show a substantial degree of information rigidity in the countries studied. They also suggest that the theoretical characterization of regression errors yields a more conservative picture of the uncertainty surrounding parameter estimates.