Extant research on the gender pay gap suggests that men and women who do the same work for the same employer receive similar pay, so that processes sorting people into jobs are thought to account for the vast majority of the pay gap. Data that can identify women and men who do the same work for the same employer are rare, and research informing this crucial aspect of gender differences in pay is several decades old and from a limited number of countries. Here, using recent linked employer-employee data from 15 countries, we show that the processes sorting people into different jobs account for substantially less of the gender pay differences than was previously believed and that within-job pay differences remain consequential.