Nowcasting LSM Growth in Pakistan

Abstract

This paper attempts to nowcast Large-scale Manufacturing (LSM) growth in Pakistan, which is generally used as proxy for economic activity in Pakistan. For this purpose, the dynamic factor and penalized regression models are used to extract the unique information set from a range of variables having close association with LSM. Given high seasonality induced volatility in LSM growth, we have also attempted to nowcast the trend and cycles separately. The estimation results show that the predicted LSM series fairly tracking the actual LSM series. Moreover, penalized regression models perform remarkably well in tracing cycles in LSM growth. However, dynamic factor model is quite successful in tracing the underlying trend growth but not the cycles.