Wage Offer and Vacancy - Source of Friction in the Japanese Labor Market -(in Japanese)

1 Motivation The main theme of this research is to examine empirically where the source of search friction exists in Japanese labor market. There has been a new influential theory called "endogenous matching function", which regards matching efficiency to be determined through interdependency among individual agents. In this research, we aim to verify empirically the importance of such interdependency among agents in markets in case of Japan. The main reason why we particularly focus on the relations between offered wage and vacancy size is because they are essential factors for vacancies, so that we could naturally regard these two factors affect individual agent's behavior to form the above-mentioned interdependency. By verifying whether negative relations between offered wages and vacancy sizes exist as theory predicts, we expect to examine if the source of search friction in labor market actually lies in such interdependency among agents. 2 Estimation Method We employed establishment panel dataset of the year 1993-1995, which is constructed from "Employment Trend Survey" by Ministry of Health, Labor, and Welfare, and examined if there actually exists negative relationship between the offered wage and vacancy size. In particular, we constructed a model which explains offered wage levels by vacancy sizes, controlling both individual characters and establishment attributes. Since we could only use wage change rates instead of actual offered wages, we decomposed wage change rates into previous wages and current wages, and assumed we could estimate wages approximately for both wages from Mincer-type wage equations. If we could derive a negative coefficient for vacancy size as an explanatory variable, then this indicates interdependency between job-seekers and establishments actually holds as premises for agents with their search activities in labor market, and such relations have been acting as frictional factors in search activities. 3 Results The estimation result of vacancy coefficients is not significant with the entire sample of fulltime workers. This implies that the interdependency between job-seekers and establishments are not so important at the entire labor market level. At the next step, we tried the same estimation by dividing the sample by job category, assuming job-seekers look for a job within specific job categories. The result shows coefficients of vacancy size have negative sign with clerks or with sales staffs, while they are not significant with technical staffs or with managers. From this result, we could reckon that in case of vacancies for unskilled jobs like former ones interdependency would be quite relevant, while in case of vacancies for skilled jobs like latter ones it does not count very much. Rather, random matching could be more relevant with them, or they even might be perfectly coordinated. There might exist endogeneity issues with regards to vacancy size with the above estimation, as well as selection bias issues since we only employed samples of job changers for our estimation. However, after we have tried several estimation patterns which could be effective for the above-mentioned issues, we still have not derived negative coefficient for the variable in concern, which confirmed our conclusion that theoretical expectation of negative relations between vacancy sizes and offered wage levels does not necessarily hold in Japanese labor market in general. 4 Policy Implications The above conclusion suggests that we should not overvalue the interdependency among agents in labor market. However, with jobs such as unskilled ones, there might be interdependent relations between job seekers and firms, so that policies should focus more on controlling such relations to enhance matching efficiency. In particular, public job centers are mainly dealing with unskilled vacancies, so it might be useful for them to try to focus more on strengthening coordination skills on such interdependent relations. For example, it might be effective if job centers arrange not to allocate too many applicants on specific (i.e. popular) vacancies when they make referrals.