Search

This paper analyzes the impacts of economic shocks on Japanese regional employment using causal forest algorithms, and our data-drive findings are compared with those estimated under the standard OLS estimation. Our results are as follows. The largest declines in employment growth during the recessionary period associated with the crisis are in industries and regions where employment grew before the crisis, and in industries that have a large share of employment in the region. These results differ from the standard OLS results, which show that the negative impact of shocks is greater in regions with higher unemployment rates. In addition, testing the robustness with respect to outliers, which are samples that could be arbitrarily removed, we show that the results obtained by causal forests are robust, but those of OLS are not.

This paper examines the effects of concluding a free trade agreement (FTA) in the presence of international trade costs between countries. In the traditional arguments, the optimal external tariffs set by the FTA members are always lower than the pre-FTA optimal tariffs, which implies that there are the tariff complementarity effects as the FTA forming. To reexamine this argument, we construct a simple three-country model of imperfect competition with endogenously determined (external) tariffs, and demonstrate that in the presence of trade costs, the member countries may employ the higher external tariff as they form the FTA. That is the tariff complementarity effects disappear. We also find that in contrast to traditional argument, the non-member country's welfare may worsen even if there are tariff complementarity effects. Furthermore, the findings show that the FTA is likely to result in the deterioration of the member countries' welfare, depending on the trade costs.