Abstract
The study investigates import demand for Pakistan considering four different models for the period 1978 to 2016. The study estimates standard model, revised traditional, dynamic structural and dynamic financial import demand models. Although import demand is found to be cointegrated with relative prices of imports and real activity variables in all four models yet, activity variable plays an important role in classifying imports as necessity or luxury goods. Moreover, the results show that estimated short-run elasticities are smaller than long-run elasticities. It reveals the ineffectiveness of exchange rate policy in influencing import demand in the short-run but not in the long-run. Besides, the study finds that the estimates of import demand elasticities and the slope of demand curve are subject to the choice of the model. Therefore, the study conducts within sample as well as out of sample forecasting analysis of the estimated models to evaluate forecasting performance. The study finds that the standard import demand model performs the best for Pakistan. The study concludes that gains from trade can be maximized when a suitable model is focused for the formulation of trade policy.