Cointegration, causality and the transmission of shocks across wheat market in Pakistan

Abstract

This paper uses quarterly price data and examines the transmission of shocks across different spatially separated locations besides identifying causality among these locations. Johansen and Juselius's (Econ. Stat., 52, 160-210, 1990) multivariate cointegration procedure identified two cointegrating vectors among these locations. Following Toda and Yamamoto (J. Econom., 66, 225-250, 1995), causality tests showed only one bi-directional causality and it was between Peshawar and Hyderabad locations. Faisalabad and Sargodha appeared independent (i.e. exogenous) market locations in price discovery process. Peshawar market showed maximum (i.e. 5) number of significant links. The generalized impulse response functions, though, suggested similar (cyclical) pattern of responses across the markets, but their time profile, which provides insight into the system's speed of convergence to long run equilibrium path, varied with different level of extent and persistency. Responses to shock originating in consumption markets (i.e. Karachi, Peshawar and Lahore) remained short lived; whereas the shocks stemming from surplus wheat producing locations (i.e Multan, Sargodha and Faisalabad) produced long and more persistent responses.