Yoshinori KUROKAWA

Which is better for a firm, job rotation or specialization, can be considered as an endogenously formed worker-indivisible job matching problem. We model this problem as a firm's profit maximization problem under uncertainty, with and without overlapping generations. In both models, we show that among all possible job allocations, the rotation and specialization schemes are the only variations that can be optimal in terms of profits. Moreover, the rotation scheme is better when the productivity difference between post- and under-training workers is smaller, the uncertainty about job continuation in the future is more significant, or the slope of seniority wages is larger. The results indicate that firms in different environments prefer different worker-job matchings.

Recent research has begun to imply intermediate goods-skill complementarity; however, this possible complementarity has been hypothesized but not statistically tested, despite the increasing importance of intermediate goods in production. This study provides statistical evidence regarding whether intermediate goods are more complementary with skilled labor than with unskilled labor. Using panel data from 40 countries over the period 1995-2009, we estimate a two-level constant elasticity of substitution (CES) production function. Our major findings are fivefold. First, at the aggregated one-sector level, the elasticity of substitution between intermediate goods and unskilled labor is 1.22, which is significantly greater than that between intermediate goods and skilled labor of 1.05, indicating intermediate goods-skill complementarity. Second, at the disaggregated level, such complementarity is primarily observed in heavy manufacturing industries and the service sector, whereas complementarity is observed between intermediate goods and unskilled labor in the primary sector and light manufacturing industries. Third, the normalization of the data and the cumulant estimators exhibit stronger results. Fourth, our baseline results are confirmed applying several robustness checks, such as switching skilled and unskilled labor or considering capital-skill complementarity. Finally, intermediate goods-skill complementarity tends to be higher for industries that use more imported intermediate goods.

This study examines whether a country's trade significantly increases industry-level output and whether this impact is greater for industries with higher levels of differentiation. To empirically answer these questions, the Frankel and Romer [(1999). "Does Trade Cause Growth?" American Economic Review 89 (3): 379-399] method, which uses the geographical component of trade as an instrument, is extended in two ways. First, the study examines the impacts on industry-level output instead of country-level output. Second, an index of industry differentiation based on Rauch [(1999). Networks Versus Markets in International Trade." Journal of International Economics 48 (1): 7-35] is introduced. Using a two-stage least squares estimation and data for 20 manufacturing industries and 99 countries in 2015, it is found that the impact of a country's trade on industry-level output and output per worker is heterogeneous across industries. A one-percentage-point increase in the trade/GDP ratio results in a larger effect for industries with higher degrees of differentiation, ranging from 0.340% to 1.736% for industrial output and from 0.754% to 1.566% for industrial output per worker. This heterogeneity with respect to industry differentiation is robust even when controlling for latitude and institutional variables. These results suggest that a country's trade affects industry-level output differently depending on the industry's level of differentiation.

Two- or more-factor general equilibrium models commonly assume no factor intensity reversals (FIRs): a good that is relatively capital intensive compared with other goods within a country/region is also relatively capital intensive within another country/region. This assumption is so important that its breakdown results in the collapse of several well-known theorems, such as the Heckscher-Ohlin theorem and the Stolper-Samuelson theorem. Seeing a recent observation, however, Feenstra (2015, Advanced international trade: Theory and evidence (2nd ed.), Princeton University Press) argues that FIRs are quite realistic. Our nonparametric test finds that recent regional-level data support his view. At the two-digit industry level, the degree of FIRs among regions is higher than those found in previous studies, which is accompanied by wide differences in relative factor prices among regions. The degree has also increased over the last two decades. FIRs are even stronger at the disaggregated four-digit industry level and, though not many, exist at the two-aggregated-industry level as well. At all the three levels, FIRs do not disappear even when we take into account other factors such as human capital and land. Thus, considering a model without restrictions on FIRs might be a possible direction of research.