Empirical Analysis of the Impact of School Closure on Population Change in Japan

Abstract In Japan, 4,907 public elementary schools and 1,201 public junior high schools were closed between 1988 and 2018. The government has been promoting consolidation and closure projects to optimize school size in response to the rapidly declining birth rate. Consensus building among stakeholders is essential for school consolidation, which requires an extended discussion period. It is believed that a population decreases if schools are closed in a community; however, no studies in Japan have objectively demonstrated this. This study examines the impact of the closure of elementary and junior high schools on population change in Japan by comparing affected areas with those with existing elementary and junior high schools by considering regional characteristics using digital national land information. We evaluate the impact of school closures on population changes using propensity score matching (PSM). A difference-in-differences (DID) model was developed to verify the trends before and after school closures. The selected covariates for good PSM models included the number of railway stations, the presence of urban facilities, urban areas, metropolitan areas, total area, and the average age of inhabitants. The good-accuracy models in PSM showed that the change in population before and after school closures (closed schools minus existing schools) ranged from −37.7 to 44.4. However, these results showed no significant differences, as indicated by the t -test values ranging from −0.26 to 1.44. Also, the coefficients of the good-accuracy models in DID ranged from −45.9 to 44.43. The t -test values of these models ranged from −0.2 to 1.31, indicating no significant differences in population trends before and after school closures. Our results can contribute to stakeholder consensus-building when policymakers include population issues resulting from school closures on their agenda.