Building Communities or Building Empty Rooms? Social Education Spending and Municipal Outcomes in Japan

Abstract

This paper investigates whether Japanese municipalities that allocated a higher share of their budgets to social education expenditure experienced better demographic and economic outcomes between 2012 and 2021. Social education is defined as the government budget classification that includes operating costs for libraries, kominkan (community centers), museums, and culture activities. Applying a panel dataset of 1,718 municipalities and regressions with long difference specifications and pre-period controls, I find that municipalities with higher social education ratios in 2012 witnessed significantly lower population losses during the next 10 years, regardless of population size and income at baseline. Net migration and income per capita do not have a statistically significant long-term association, and using the lagged specification of the panel, there appears to be a negative short-term relationship between social education ratio and income. In panel fixed effects regressions, the lack of any yearly correlation between social education ratio and population growth is interpreted as being consistent with the notion that effects accumulate through time. An IV regression strategy proves ineffective because of the low power of first-stage regressions.