When reputation evaluation indicators become targets, existing indicators will lose the role of indicating the true quality; At present, the evaluation of discipline reputation mostly focuses on subjective evaluation based on objective data, and there is a dispute about reliability and validity; Due to different indicators and weight settings, it is difficult to make horizontal comparison among disciplines; The evaluation also has a certain time lag. In order to solve the above four problems, this study explores a new method of discipline reputation evaluation. Taking the business administration discipline as an example, it collects data of 5848 doctoral graduates who first entered teaching posts, establishes a directed adjacency matrix from the employment unit to the doctoral degree awarding unit, and uses the theory and method of social network analysis to conduct quantitative analysis on the doctoral mutual employment network. The results show that: (1) PhD exchange network can explain discipline reputation and is a new indicator to measure discipline reputation; (2) From the perspective of employment behavior among colleges and universities, there is horizontal flow and downward flow between the head colleges and universities, and downward flow is mainly among the middle and lower colleges. There is a time lag between college talent recruitment and academic achievement output. Therefore, the mining of the structural characteristics and network evolution trend of the PhD exchange network based on the "foot voting" of doctoral graduates is faster than the discipline ranking based on the follow-up achievement indicators to reflect the changes in the discipline quality, which can be used to warn the changes in the discipline quality.