Publications

Summary: Using data from over 80 million mobile devices, we construct a city-by-day mobility index during the COVID-19 pandemic to examine how social capital shapes voluntary prevention and policy compliance. In the second year, both behaviors declined sharply in low–social capital areas but remained stable in high–social capital ones. The effects are heterogeneous: norms and culture matter less in developed, urbanized regions. In Japan, variation in party support was linked not to ideology or ruling status but to whether a party held majority or minority standing. These findings suggest that conformity to community norms, rather than political orientation, is a key driver of collective behaviors that protect public health.

Summary: This study proposes mixed spatial ANOVA (MS-ANOVA), a spatial extension of mixed MANOVA models, to evaluate correlations among random effects in hierarchical spatial data. Formulated as a Bayesian hierarchical model, MS-ANOVA is estimated through an empirical Bayesian procedure: hyperparameters are obtained via quasi-maximum likelihood, and posterior distributions are then derived using these estimates. We also demonstrate the asymptotic validity of the first-step estimator. Applying the model to Japanese happiness survey data, we find that social capital—understood as community beliefs and norms supporting collective action—significantly enhances individual happiness.

Summary: This study examines how scattered greenery (street trees and yard bushes), rather than cohesive greenery (parks and forests), affects housing prices. Using high-resolution satellite images combined with condominium sales and rental data, we estimate hedonic pricing models. Results show that scattered greenery within 100 m significantly raises housing prices, especially near highways, while more distant greenery has no effect. Inexpensive and small properties, whether for sale or rent, are less responsive. These findings reveal substantial heterogeneity in preferences for greenery by property type and location, suggesting that increased green amenities may contribute to environmental gentrification through concentrations of expensive housing.

Summary: Airborne pollen from seasonal allergies imposes overlooked economic costs by raising healthcare spending, lowering productivity, and reducing mobility. This study uses daily scanner data linked with pollen counts from 120 observation stations in Japan to estimate its impact on consumption. Expenditures decline by about 2% on days with unusually high pollen, and lagged results show the effect persists for more than a week. Weekly and monthly analyses also confirm a reduction in spending, suggesting that pollen exposure decreases total expenditure rather than simply delaying purchases. These findings underscore the hidden economic burden of pollen and the importance of urban and health policies to address seasonal allergies.

Summary: This study investigates how disclosure of school quality information influences the behavior of students and parents. Using the event in which school-level test scores were first disclosed in October 2014 in Matsue City, Japan, I estimate the change in the effect of test scores on housing rent and population in a school district between pre- and post-disclosure. Following the disclosure of scores, the school catchment areas with higher scores witnessed a significant increase in the rents of apartments for families and the population of school-age children. These results provide novel evidence that disclosure of school quality information significantly impacts people's school and residential choice behavior.

Summary: This study investigates the effect of public school quality on housing rents within a school district. I estimate the causal effect of school quality as measured by average test scores on housing rents, by using a regression discontinuity design to control for unobserved characteristics of neighborhoods. Specifically, I focus on apartment buildings located within school attendance district boundaries. I find that school quality has a significantly positive effect on housing rents of apartments for families, but school quality does not have a significant effect on housing rents of dwellings for single people. These results show that parents are willing to pay more to send their children to a better school.