Using personnel data from a Japanese machine-producing company, we examine the nature of employment adjustments using buyout formulas. Performance ratings and other characteristics for employees who accepted buyouts in 1993 and 1996 are compared with those of employees who remained or left voluntarily. The empirical results reveal the following points. In the case of 1993 buyout, while the skill grade level and associated level of earnings of buyout takers are relatively high, their performance ratings are the same as those of incumbents and lower than those of voluntary leavers. Therefore, the buyout takers were those whom the company wanted to dismiss. For the 1996 buyout, although performance ratings of buyout takers were the same as those of incumbents in the case of managers, the ratings of non-managerial, lower-graded (i.e. lower ranked position) buyout takers were higher than those of incumbents and the same as those of voluntary leavers. Therefore, the non-managerial buyout takers were those whom the company would have preferred to retain. To conclude, the company was able to specify who was dismissed but was unable to specify who took buyouts despite efforts to ensure retention of value employees, including careful setting of incentive schemes. The case presented in this paper demonstrates that employment adjustment using a buyout framework invites "adverse selection" despite careful employer precautions.
