Standard estimates of earnings profiles ignore the fact that, with unobserved heterogeneity, cross-section evidence need not reflect the `true' relationship between earnings and tenure. In this paper we argue that the observation of the position filled by an employee in the firm hierarchy is informative about both his/her quality and the quality of his/her match. Under certain assumptions, this information can be used to construct an unbiased estimator of the effects of tenure on earnings growth. We apply this simple idea to Japanese and UK data. We find that tenure effects on earnings are positive but smaller than the effects estimated with the traditional approach. In a comparative perspective, we also find that UK and Japanese earnings profiles are rather similar in banking and finance and remarkably different in manufacturing.