No abstract

No abstract

We use a discrete choice model to study the impact of tax incentives on firm location choices in situations involving heterogeneous geographic characteristics, heterogeneous firm preferences and large choice sets. We apply our model to investigate the impact of movie production incentives on filming location choices for movies produced from 1999 to 2013. We gather the characteristics of filming locations and use a machine-learning technique to define choice sets. We find production incentives can attract movies to a state, but the impact depends on the type of incentive offered, studio characteristics and inherent location geographic characteristics. Mid-sized studios respond to all forms of incentives, major studios respond only to refundable and transferable tax credits, and independent studios are not sensitive to any incentives. We fail to find strong evidence that incentives create a more permanent movie industry in a state. A counterfactual identifies the states most impacted by these policies. We supplement our discrete choice model with a simple cost-benefit analysis, which indicates that movie incentive programs are revenue-negative for states.

In this paper, we examine how the decisions of churches are impacted by the decisions of rival churches. Using a novel data set of Christian churches in two suburban Nashville, TN counties, we estimate a model of strategic interaction based on empirical models of discrete games. We focus on a church's decision of whether or not to provide a week-day child care program. Empirical evidence indicates that churches are more likely to offer a service when a nearby same-denomination church offers it, but are less likely to offer the service if a nearby different-denomination church offers the service. This suggests that churches may feel pressure to match the services offered by same-denomination churches. Using our estimates, we conduct counterfactual simulations which indicate that a new church's entry may increase or decrease the number of churches offering child care depending on the size and denomination of the entrant. Additional counterfactual simulations indicate that in some cases, increases in population may decrease the probability of a church offering child care.