Does judgment improve macroeconomic density forecasts?

Abstract

This paper presents empirical evidence on how judgmental adjustments a¤ect the accuracy of macroeconomic density forecasts. Judgment is de…ned as the di¤erence between professional forecasters'densities and the forecast densities from statistical models. Using entropic tilting, we evaluate whether judgments about the mean, variance and skew improve the accuracy of density forecasts for UK output growth and in ‡ation. We …nd that not all judgmental adjustments help. Judgments about point forecasts tend to improve density forecast accuracy at short horizons and at times of heightened macroeconomic uncertainty. Judgments about the variance hinder at short horizons, but can improve tail risk forecasts at longer horizons. Judgments about skew in general take value away, with gains seen only for longer horizon output growth forecasts when statistical models took longer to learn that downside risks had reduced with the end of the Great Recession. Overall, density forecasts from statistical models prove hard to beat.

Combining Bayesian VARs with Survey Density Forecasts: Does it Pay Off?

“…When the individual models or the combination are tilted to both mean and variance of the SPF, there is a general worsening of the performance. Our results are similar to Galvao et al (2021) for U.K.'s GDP and inflation; they also found that judgement on the mean tends to improve model density forecasts at short horizons, whereas survey second moments hinder performance at short horizons. Combining individual BVARs both with and without judgement improves accuracy with respect to individual models according to the LPS metric, while in terms of CRPS the combination is worse than SPF for all variables and horizons, with the exception of the two-year-ahead GDP forecast.…”

“…Namely, we tilt either the individual models before combining them, or the model combination, to either the first moment or both first and second moments of the SPF. Therefore, we extend the literature that applies tilting to individual models (Krüger et al, 2017;Altavilla et al, 2017;Ganics and Odendahl, 2021) and model combinations (Galvao et al, 2021) or just combines macroeconomic models (Amisano and Geweke, 2017). To our knowledge it has not yet been considered to combine tilted forecasts or to include survey forecasts in macroeconomic model density combinations.…”

Do Inflation Expectations Improve Model-Based Inflation Forecasts?

“…Bańbura et al (2021) analyse for euro area data how to best combine subjective forecasts from the SPF and model forecasts from several BVARs and recommend tilting the model forecasts only to the first moments of the SPF (thus ignoring the information from the second) prior to performing forecast combination. Galvao et al (2021) also find improvements in forecast accuracy when tilting model forecasts to the mean of professional forecasts for output growth and inflation in the UK. Wright (2013) shows gains in forecasting performance from using long-term Blue Chip forecasts as priors for BVAR steady states.…”

“…We also do not consider a "glide path" model here as short-term (current quarter) inflation expectations are not available for our "main" measure of inflation expectations for the euro area (the SPF). Finally, we only use the first moment (mean) of the expectations given the findings of Bańbura et al (2021) and Galvao et al (2021) (see also Clements, 2014Clements, , 2018.…”

What is the predictive value of SPF point and density forecasts?

“…Several previous papers utilized information in survey histograms by fitting a parametric distribution and tilting a model's predictive density to the moments of the fitted distribution (see, e.g., Banbura et al (2021) and Galvão et al (2021)). As a robustness check of our baseline approach that applies tilting directly to the SPF bin probabilities and does not require additional distributional assumptions, we also consider tilting to higher-order moments of a fitted distribution.…”