Industrial location and interpretation of zero counts.

This paper intends to shed light on the treatment of zero counts in count data (CD) industrial location models. An excess of zeros and overdispersion has been dealt with by estimating mixture models such as negative binomial, zero-inflated and hurdle models. However, the existence of a considerable share of zero counts could also have an economic interpretation that is worth investigating. To do so, several CD models focussing specifically on zero-inflated and hurdle models have been fitted and compared to examine empirical data for Catalonia. [ABSTRACT FROM AUTHOR]