SUMMARY
The purpose of this work is to empirically study the importance of agglomeration economies in the selection of geographical zones in which foreign investors are willing to develop their investment projects in Chile. For this purpose a discrete choice model, te conditional logit model (CLM), first proposed by McFadden (1974), was used in this study and the application of the IIA test prposed by Hausman and McFadden (1984) is reported. Furthermore, the use of a less restrictive model is proposed. The main advantage of this latter model is that it allows the variances of the random components to be different across alternatives, that is to say it is a Heteroscedastic Ectreme Value Model (HEVM). The results suggest that agglomeration economies significantly influence the selection of the geographical zone in which the investment is to be located. The results are robust to different specifications of the model. Furthermore, the IIA assumption seems to be appropriate and the result of the HEVM is not significantly different of the CLM.