This paper begins with a description of simple select models of inflation a
nd their ability to fit the data. The paper in turn measures the stability
of particular parsimonious models. One-step forecast tests are applied, whi
ch establish the instability of the money demand model compel-ed with other
models, despite its ability to fit the data nearly as well as the P-star m
odel. Further, the forecasting performances of the selected parsimonious mo
dels are analyzed, though only the performance of one-step forecasts are fu
lly treated given the insufficient descriptive ability of the AR processes
of explanatory variables. As a measure of forecast ability, the author empl
oys U-statistic, RMSE, and MAE tests. The above statistics assign the best-
forecast performance to the modified version of the P-star model of inflati
on for a small, open economy.