Selective Attention in Exchange Rate Forecasting

We analyze the exchange rate forecasting performance under assumption of selective attention. Although currency markets react to a variety different information, we hypothesize that market participants process only limited amount information. Our analysis includes more than 100,000 news articles relevant six most-traded foreign pairs for period 1979-2016. employ dynamic model averaging approach reduce selection uncertainty and identify time-varying probability include regressors in our models. results show smaller sizes models accounting presence attention offer improved fitting results. Specifically, document growing impact trade monetary policy on euro/dollar following global financial crisis. Overall, point existence case most pairs.