APPLICATION OF A DYNAMIC PANEL DATA ESTIMATOR TO CROSS-COUNTRY COFFEE DEMAND: A TALE OF TWO ERAS

We estimate price and income elasticities of demand for green coffee beans in panels of up to 40 countries, both during and after the operation of export quotas under International Coffee Agreements. The dynamic panel estimator proposed in Han and Phillips (2007) is used because it is a consistent estimator, for any length of panel, regardless of the presence of unit roots. Dynamic panel data models, of any type, do not seem to have been previously applied to coffee demand. We find evidence of a concave relationship between income and coffee consumption for countries which are members of the International Coffee Organization, but no evidence of such a relationship for other countries. A further conclusion is that measures which increase the price of coffee beans can be expected to have little effect on coffee sales.