Cointegration test of the monetary theory of inflation and forecasting accuracy of the univariate and vector ARMA models of inflation

Abstract

Conducts both the cointegration test of the monetary theory of inflation and the Granger‐causality test between the variables in the system, and also develops univariate and multivariate time series models to forecast inflation rates. Quarterly time series data for Pakistan, from 1972‐2 to 1993‐4 is used for empirical investigation. Results suggest no cointegrating or long‐run relationship between the variables in the monetary model. Observes that there is some evidence of Granger‐causality running from inflation to output growth. Comparison of out‐of‐sample quarterly forecasts for the 1988‐1 to 1993‐4 period are made for univariate and vector ARMA models of inflation. States that the forecasting accuracy of the multivariate ARMA model is not statistically different from that of the univariate ARMA model.