Effects of reconstruction work on private employment after a natural disaster.

“

Although this study focuses on labor markets after a disaster, its implications are useful for multiple facets of economics. We briefly summarize this point here. First, for disaster economics, our results can contribute to the discussion of long-run economic growth such as the creative destruction hypothesis. As noted, the direction of long-run growth after a disaster is still controversial, especially for geologic disasters. Moreover, it seems that there is no theoretical proposal on the background mechanism without the Schumpeterian one, and no analysis on indirect effects of reconstruction policy. In contrast, our analysis shows the possibility that reconstruction policy plays an important role for growth through the labor market channel of fiscal policy in Alesina et al. [1]. Future study should estimate whether private investment is decreased by the resulting reduction of employment in non-construction sectors after reconstruction. Second, our study can be regarded as confirmation of the existence of the labor market channel of fiscal policy in Alesina et al. [1] since most reconstruction consists of public works. Because our study employs daily microdata, our identification is more concrete than in Alesina et al. [1]; which employs annual multi-county data. We should realize the possibility that reconstruction projects for natural disasters may result in a reduction in private investment and thereby long-term stagnation. Third, for labor economics, this study has two contributions. The first is to confirm the crowding-out effects of private employment by fiscal expansion. It is quite rare to find this effect empirically. One case we know of is Algan et al. [33]; which finds the effects of public employment using OECD panel data over 40 years. The present study contributes to this existing literature on the crowding-out effects using microdata of job postings. The second contribution is to provide the first analysis of vacancy duration in Japan using a duration modeling method. Although the analysis on vacancy duration has been done in some countries, it is far from satisfactory. Our estimation results for covariates would be useful for understanding employer’s hiring behavior.

”