Marshall Digital Scholar

Real-Time Nowcasting of Short-Run of the Euro-Dollar Exchange Rate with Economic Fundamentals: Does the measure of Money Supply Matter?

Marshall University

This paper proposes a fundamental-monetary-based econometric model with mixed frequency data to now caste the Euro-Dollar short run exchange rate. We use the exact amount of information that are available to researchers or police makers at the time of prediction of exchange rate. The spot exchange rate information are available at week basis while other macroeconomic data like money supply, in action, industrial production, and interest rate are available at monthly basis. Since the monetary model consists of stable money demand functions (Bianco, 2012), not all the measure of money supply guaranties a stable money demand. Barnett (1978, 1980) has proven that the measure of money supply (simple sum) that central banks published monthly does not provide a stable money demand. Only, the Divisia monetary aggregates provide a stable money demand at any level of aggregation. This paper proposes a multivariate state space model that takes into account not only the asynchronous information inflow, but also the instability of the money demand to predict four weeks spot exchange rate. We obtain great improvement of the model proposed by Bianco (2012) with Divisia monetary aggregates as measure of money supply. The latter performs much better at all forecasting horizons (one, two, three, and four weeks) than all of others models that predict the spot exchange rate with Significant improvements. We estimate our model (dynamic factor model) with the procedure proposed by Barnett et al (2016).