How Good are Dynamic Factor Models at Forecasting Output and Inflation? A Meta-Analytic Approach

This paper surveys existing factor forecast applications for output and inflation by means of a meta-analysis and assesses the determinants of the forecast performance of large factor models relative to other models. Our results suggest that factor models tend to outperform small models, whereas they perform slightly worse than alternative methods which are also able to exploit large datasets. The target variable (output/inflation and the country), the forecast horizon and the factor estimation technique matter. The size of the dataset from which factors are extracted positively affects the relative factor forecast performance, whereas pre-selecting the variables included in the dataset did not improve factor forecasts in the past. Quarterly data and rolling forecasts seem to lend themselves better to factor forecasts than monthly data and recursive forecasts, respectively. Whether forecasters rely on a balanced or an unbalanced panel and how the forecasting equation is specified are found to be irrelevant