A Forecasting Economic Growth of Iran: Univariate Versus Multivariate Time

This article aimed to forecast Iran's economic growth rate using annual time series data released in the period of 1338-1385, employing various econometric approaches. To prefer the most appropriate method based on a time-duration consideration, we applied a package of proper diagnostic tests on the output of different models and then singled out the most fitted model accordingly. We used these preferred models to forecast economic growth rate in a different duration period. Two classes of econometric approaches were considered; univariate and multivariate methods. Univariate methods incorporate Box- Jenkins method and State-Space method and the multivariate ones consist of VAR and VEC techniques. Results indicate that univariate approaches forecast Iran's economic growth more accurately. However, multivariate methods, which demonstrate very sensitive to the time-series properties of the variables and also to the specification of the models, emerge volatile performance. Nevertheless, the current methods display more powerful in forecasting just-one period ahead.