A spatial choice model based on random utility

Decreasing number of students in Germany lead to the closures of schools due to lack of utilization. Hence, competition between schools will increase. We use a mixed multinomial logit model in order to identify influencing factors of school choice and to regard realistic substitution patterns within the context of forecasting school choice probabilities with changing choice sets over time. Results yield commuting distance as the most important factor, but school characteristics like profiles (math or languages, e.g.) and spatial attributes like centrality have remarkable impact, too. Moreover, the analysis shows heterogeneity in population. For our sample we identify authority responsible (private/public school) as the most important variable to specify substitution patterns. Using the results, we simulate the effects of school closings and openings in realistic scenarios.