MODELING THE DEMAND FOR DURABLE INPUTS: DISTRIBUTED LAGS AND CAUSALITY

Vector-autoregressive-moving-average (VARMA) modeling was used to identify distributed lag relationships among farm tractor derived demand variables and to provide a basis for formally testing the hypothesis that the price of new tractor horsepower is exogeneous to its quantity demanded. Similar causality tests were used for a number of other explanatory variables, including the interest rate, price of diesel fuel, and price of used tractors. Results indicate that several lagged variables are significant causal factors and that the dynamic nature of the demand structure cannot be ignored when explaining tractor demand.