We analyse the heterogeneity of exchange rate forecasts by a panel of professional forecasters. Adopting the view that forecasters’ economic behaviour is such that they constantly collect, process and analyse relevant information when producing forecasts, we apply a Mixed-Data Sampling (MIDAS) regression approach. This enables us to explore the roles played by key drivers for which available data are at different frequencies from forecast disagreement. Examining the Colombian peso/U.S. dollar exchange rate, we find that central bank intervention is most effective in reducing heterogeneity in the very short-run, and when conducted against a background of high exchange rate volatility.
