Nonlinear Smooth Transition Auto-regressive Method in Investigation of Influence of Saving on Economic Growth: A Case Study of Iran

This study aims at the effects of saving variation on Iran's economic growth. The study is conducted in the framework of the nonlinear smooth transition auto-regressive method covering the period 1976-2012. To this end, we have tested a linear model against a nonlinear model and shown that the nonlinear model has a better approximation than the linear model. Then, a logistic nonlinear model is estimated by using Trasvirta test. Variations in the independent variables on the basis of the logistic nonlinear model are divided into three segments include down, middle and up thresholds. The result indicate that the effect of differential variables such as GDP, saving and ex-period error is reduced by moving from down to the middle threshold. This decreased effect, that is included in the model exogenously, shows itself in a dummy variable named ‘war impact’. Then, by moving from the middle toward the up threshold, this process is inversed, and it increases. In all the three thresholds, saving and the GDP variation of the ex-period have the least and the most influence on the next period GDP respectively.