4,420千円 (直接経費: 3,400千円、間接経費: 1,020千円)
2018年度: 1,300千円 (直接経費: 1,000千円、間接経費: 300千円)
2017年度: 1,300千円 (直接経費: 1,000千円、間接経費: 300千円)
2016年度: 1,820千円 (直接経費: 1,400千円、間接経費: 420千円)
I investigated if the subjective performance evaluation contains useful information above and beyond the objective performance evaluation. Using the personnel data of a major car sales company in Japan. I found that the current evaluation predicts one year ahead evaluation even after controlling for current sales, validating the above claim. More interesting results came from our analysis of unexpectedly negative evaluation. I found that when a worker receives an unexpectedly negative evaluation conditional on the sales, the next period sales tend to fall, but only when the evaluation is made by inexperienced supervisor. I also found that the worker is likely to quit after receiving an unexpectedly negative evaluation, but only when the evaluation is made by inexperienced supervisor. This pattern permits two interpretations: (1) evaluations are more accurate among experienced supervisors than among inexperience ones, and thus evaluations made by inexperienced supervisors might be more biased (passive evaluator hypothesis); and (2) evaluations are accurate for both experienced and inexperienced supervisors, but the former communicates them better and induces more desirable worker behavior by providing feedbacks and perhaps coaching (active coach hypothesis). I find little systematic differences in the distributions of evaluations between experienced and inexperienced supervisors, however, which seems to suggest that evidence is more consistent with the latter.