Monetary exchange rate models in a small open economy w/ alternative inflation expectation

In this paper, we empirically test the validity of four popular monetary exchange rate models under alternative inflation expectation measurements using the NOK/USD exchange rate. The selection of Norway seems ideal as it is a small open economy that does not participate in any significant economic or political organization, nor has pegged its currency with another one. In consumption with the creation of a dedicated fund for concentrating all oil revenues its economy exhibits sustainable robustness to exogenous shocks and can thus be modeled accurately for measuring the effect of monetary and fiscal policy on exchange rate determination. We approach the problem of testing exchange rate models under a machine learning framework. Starting from a general form monetary exchange rate model, the main innovation of the paper is the extraction of the model in a two-step procedure; after training a Support Vector Regression (SVR) model for fitting the data we infer upon its structure with the use of the Dynamic Evolving Neural Fuzzy Inference System (DENFIS), something that has never been attempted before. The aforementioned procedure tackles the evaluation problem without limiting itself to the usual problems of its econometric counterparts proposed in the existing literature (i.e. non-stationarity, endogeneity, heteroskedasticity etc.). Moreover, we approximate the unobserved inflation expectation with five alternative methods including and excluding oil prices effects on CPI in order to better isolate the role of oil price fluctuations in the determination of the NOK/USD rate. The SVR methodology appears superior to the alternative econometrics one (VECM) in exchange rate modelling. The best overall model in terms of accuracy is an SVR model with autoregressive inflation expectations excluding energy prices from inflation differential. The estimated coefficients show that the sign of the coefficient on the interest rate differential corroborates only with the model proposed by Bilson (1978), while we detect a significant inflation rate differential. We conclude that for a small open petroleum producing country such as Norway, fundamentals possess enough forecasting ability when used in exchange rate forecasting but no monetary exchange rate model can describe explicitly the evolution path of the exchange rate.