Testing for Employee Discrimination in Britain using Matched Employer-Employee Data.

We use recent matched employer-employee data to directly test if white workers have a taste for racial discrimination in Britain. We formally introduce individual and firm heterogeneity into the discrimination model used by Becker (1957, 1971) which we extend to generate predictions consistent with an employee taste for discrimination. We argue firstly that white employees with a taste for discrimination should report lower levels of job satisfaction the larger the proportion of ethnic minorities at their workplace. Secondly, white employees would have to be compensated by higher wages if required to work alongside ethnic minority co-workers. Both hypotheses are clearly supported for white males in our data, after comprehensively controlling for individual, job, and workplace characteristics. The white male wage premium for working amongst only ethnic minority co-workers, as compared to working only with whites, is about 12%. Importantly, it appears that neither of these effects operates via realised racial prejudice at the workplace or white employees' feelings concerning their job security.