Paper Title
Fundamentals Selection on Nominal Exchange-Rate Predictability

Abstract
The existing empirical literature attempts to establish economic fundamentals under different theoretical frameworks and use them for short-term out-predictability of the nominal exchange rate. The results cannot berobustly consistent in fact. In short, the economic fundamental constructed by a single economic structure model does not beat the random walk with driftless on theshort horizon. The information attached to the economi`c fundamentals constructed by different economic structuremodels should be different, and the degree of influence onexchange rate changes also changes with time. This paperadopted the data-driven Bayesian Lasso regression toestimate the long-horizon predictive equation of thenominal exchange rate and compare it with the random walkmodel for short, medium and long-term predictioncapability. The empirical results show that the pooledestimation with Bayesian Lasso regression may help toimprove the out-predictability of exchange rate over theshort horizon. However, Bayesian lasso cannot providestatistical evidence to beat random walk model under longhorizonforecasts. Keywords - Fundamentals, Nominal Exchange Rate Forecast, Bayesian Lasso Regression.