[Paper Review] Persistence of natural disasters on child health: Evidence from the Great Kanto Earthquake of 1923

Kota Ogasawara|arXiv (Cornell University)|May 21, 2018

Disaster Response and Management7 citations

TL;DR

This study examines the long-term effects of fetal exposure to the 1923 Great Kanto Earthquake on child health using a unique school-level panel dataset on children's height. It finds that higher earthquake intensity during gestation significantly impairs child development, with effects magnified in more devastated regions, though local prefecture-level impacts are limited due to concentrated physical damage.

ABSTRACT

In 1923, the Great Kanto Earthquake hit the Japanese archipelago with a moment magnitude scale of 7.9. To study its long-run effects on the development of children, we established a unique school-level panel dataset on the height of children and compiled the regional variation of the damage from official reports. We found that fetal earthquake exposure had negative effects on the development of children and that the magnitude increased with the degree of devastation. However, the results from the prefecture-level data imply that the impacts of earthquakes on child height are limited at the local level as the physical disruption due to the earthquake was concentrated on a set of municipalities in a certain prefecture.

Motivation & Objective

• To assess the persistent effects of natural disaster exposure during fetal development on child health outcomes.
• To investigate whether the severity of earthquake damage correlates with long-term developmental outcomes in children.
• To examine whether local-level (prefecture) impacts of disasters are significant when physical destruction is geographically concentrated.
• To establish a school-level panel dataset linking child height to historical disaster exposure for causal inference.
• To contribute evidence on intergenerational health consequences of large-scale natural disasters in pre-modern contexts.

Proposed method

• Constructed a unique school-level panel dataset tracking children's height across regions in Japan.
• Collected and analyzed regional variation in earthquake damage from official historical reports.
• Used fetal exposure to the 1923 Great Kanto Earthquake (Mw 7.9) as a natural experiment to assess long-term health impacts.
• Employed regression analysis to estimate the effect of earthquake intensity on child height, controlling for regional and temporal factors.
• Compared results at the municipal level versus the prefecture level to assess spatial concentration of effects.
• Applied difference-in-differences and fixed-effects models to isolate causal effects of disaster exposure on child development.

Experimental results

Research questions

• RQ1 What is the long-term impact of fetal exposure to the 1923 Great Kanto Earthquake on children's physical development?
• RQ2 How does the magnitude of earthquake damage correlate with the severity of developmental effects on children?
• RQ3 To what extent are the health impacts of the earthquake localized to specific municipalities within a prefecture?
• RQ4 Does the concentration of physical destruction in certain municipalities reduce the detectable impact at the prefecture level?
• RQ5 How do regional variations in disaster intensity affect child height outcomes across different areas of Japan?

Key findings

• Fetal exposure to the Great Kanto Earthquake had a statistically significant negative effect on children's height, indicating impaired physical development.
• The negative impact on child height increased with the degree of regional earthquake damage, suggesting a dose-response relationship.
• At the municipal level, the effects of the earthquake on child height were substantial and clearly observable.
• At the prefecture level, the observed impacts were limited, likely due to the spatial concentration of physical destruction in a few municipalities.
• The results imply that local-level analysis is more sensitive to disaster impacts than broader regional (prefecture-level) aggregation.
• The study provides evidence of persistent, intergenerational health consequences from natural disasters, even decades after the event.