Abstract
Taking the aftermath of Chi-chi earthquake in 1999 as an example, the article investigates the social stratification of disaster hazard and its consequences on social inequality. Following the literature on “social vulnerability,” we argue that class, ethnicity and uneven rural-urban development shaped the unequal distribution of disaster hazards, which led to social inequality in the affected area. We use the first wave of the Taiwan Education Panel Survey (TEPS), a national dataset of high school students collected in 2001, and identified the affected households in the statistics. It is found that the hazard is not neutral but is robustly associated with class, ethnicity and the rural areas. The households from the working class, peasants and self-employed origins were more likely to be affected, while those from the capitalist and manager origins suffered less from the earthquake. Minorities from aboriginal and Hakka were more likely to suffer, while urban residents were less likely to be affected the disaster. To estimate the redistributive effects in the process of reconstruction, we introduce the treatment regression model for the control of the risk inequality. Modeling the household income, economic crisis, self-reported happiness and depression, we find that the victims were worse off two years after the earthquake. However, once the treatment model randomized the hazards, the differences between the affected families and the non-affected families disappeared. The outcomes suggest that the deterioration of social inequality mostly results from the unequal distribution of hazard rather than the redistribution in the reconstruction. The conclusion also implies that for reducing the unequal hazard, the reconstruction strategy for the affected area should focus not only on the compensation of the victims, but also on the reforms of primarily socialinequalities.