Luc Anselin1, Patricio Aroca2 y Coro Chasco3.
Participación en sesión: Métodos y modelos de análisis regional.
In many areas of regional science, choices made by individuals are costly to collect or inaccessi- ble. However, analysts may have access to the choice data aggregated across groups of individuals in the form of counts or shares. Regression-based spatial interaction macro-models have frequently been used to estimate group-data choices. This is the case of many applications of gravity models to migration or other spatial interaction processes. In these models, the (aggregated) observations are treated as if they were single entities. In effect, they specify the dependent variable as mere (log- transformed) aggregations of individual data, which can produce –among others– severe problems of non-normality and heteroscedasticity, enhancing spatial autocorrelation in the error terms. Mo- reover, a simple aggregation of individual choices does not necessarily lead to grouped or “heard” behavior. Aggregation must result in models consistent with theory, which should be capable of iden- tifying overall regularities in collective population behavior. For this reason, we recommend following other strand of the literature based on a choice-theoretic perspective. Although typically concerned with the identification of individual behavior, choice models have also been specified for grouped da- ta when observations no longer consist of single individuals but sets of several persons who share similar characteristics (e.g. living in the same region). In these grouped-data choice models, the de- pendent variable consists of observed proportions or relative frequencies, which are estimated by nonlinear weighted least squares method. Grouped-data choice models can be easily generalized to spatial interaction models of migration or trade flows. In this context, the paper proposes another ex- tension which allows for spatial dependence in flow magnitudes, since the conventional assumption of independence between origin-destination flows can no longer be supposed. In order to deal with this spatial effect, the grouped-data discrete choice model can include spatially lagged variables, the spatial autoregressive (SAR) interaction model. Furthermore, other spatial discrete choice models for grouped-data flows could be specified like the spatial error model (SEM), spatial lag and error model (SARAR), spatial Durbin model (SDM) and the spatial cross-regressive model (SLX), among others. In the spatial discrete choice model for grouped-data flows, spatial dependence can take the form of spi- llovers to both regions neighboring the origin or destination in the dyadic relationships that characte- rize O-D flows. Specifically, this paper develops a spatial dependence probit models for grouped-data flows, revisiting the current literature and showing how the proposed method improve the current specification and estimation procedure.