This paper utilizes a rich data set on workers and their employers in the US and Japan to test several predictions of efficiency wage theories. The data set incorporates numerous objective and subjective performance measures including turnover, effort, absences, satisfaction, and commitment. It also contains extremely good measures of establishment, worker, and job characteristics. For almost all of the performance measures in both countries efficiency wage theories are supported; that is, workers receiving particularly high wages given their observable characteristics report that they are less likely to quit, more satisfied with their pay, and so forth. The between-establishment component of wages is a more reliable predictor of performance than the within-establishment component.