Investigating the Co-movement of Cocoa and Coffee Prices in International Markets: Evidence from Non-linear Autoregressive Distributed Lag Model and Rolling Window Methods

Abstract

The focus of this study is to empirically investigate the co-movement of Cocoa, Arabica, and 
Robusta coffee prices for 1960-2018. An essential contribution of the present study is using the 
non-linear autoregressive distributed lagged co-integration technique and rolling window
based methods for empirical analysis and to check the co-integration among variables. At the 
same time, the Auto Regressive Distributed Lagged (ARDL) bound test and Gregory-Hansen 
co-integration techniques are employed for robust analysis. The empirical estimations indicate 
that the positive shocks in Cocoa prices have affected Arabica and Robusta coffee prices in the 
short and long run. For Cocoa and Arabica coffee estimations, the one unit of positive change 
of cocoa price tends to increase the Arabica price with 0.1445 and 0.806 for the short run and 
long run, respectively. These findings suggest that the increase in cocoa price causes an 
increase in the demand for Arabica coffee from importing countries. Similar results are 
confirmed in the case of Cocoa and Arabica. At the same time, the adverse shocks in Robusta 
coffee prices have affected Cocoa prices with greater magnitude in the long run. Thus, our 
outcomes also confirm the co-movement of the coffee and cocoa prices.