A note on "tractors in Spain: a logistic analysis"

Abstract

In his well-known article Mar-Molinero (1980) used the logistic curve in order to forecast the stock of agricultural tractors in Spain and introduced an autoregressive error term to model the effect of "economic influences and government policies". Other models and methods proposed by Oliver, Meade, McGowan and Harvey are referred in the bibliography. In this paper, we apply an alternative method instead in order to cope with the autoregressive errors, that is the addition of two periodic (sine) functions in the model, which represent the effects of the economic cycles. The model is estimated by non-linear regression techniques and its fitting and forecasting accuracy is compared with those of other models, when possible. It turns out that this technique is a worthwhile alternative to situations like the present one.