Maximum Likelihood

Abstract

As mentioned in the introduction to the book, the key use for the gravity model is forecasting and for forecasting, it is important to know which aspects of the base period remain unchanged into the forecast period. Therefore, in the context of the gravity model, we would usually need to know which aspects of the model are configuration-free. In Part I of this book, we saw the conditions under which one or more of the functions A _c (i), B_c(j) and F _c (c _ij ) are configuration-free and therefore can be assumed to remain invariant from base to forecast period. However, we still need to get numerical estimates of the parameters in these functions. This chapter and the next are devoted to this topic.