Since 1971 the number of doctors providing ambulatory care in France has been restricted by the limited number of places for students in the second year of medical studies. Initially set at 8588 places, the number only started to fall in 1978, until it reached 3500 places in 1993. The arrival on the scene of the large baby-boom cohorts and inadequate initial restrictions meant that many generations of young doctors started practicing. It was only in 1987 (i.e. nine years after 1978, due to the length of medical studies) that we can start to see the impact of the limited number of medical school places on the number of young doctors. Taking a representative sample of ‘sector 1’ GPs (public doctors who charge the officially negotiated rates) over the period 1983-2004 made it possible to analyse the determinants for their earnings and the impact on their careers of fluctuations in the number of medical school places. Location and increases in rates of pay had a considerable impact on earnings. The career paths of doctors are very different from those of salaried employees: all things being equal, their earnings rise quickly at the start of their careers and then fall, on average from the twelfth year of practicing onwards. It is as if doctors preferred to concentrate their efforts at the start of their working lives and then reduced their workload. Doctors’ earnings depend heavily on the demography of the medical profession when they start practicing. The gap between the permanent earnings of the different cohorts can be up to 20%, all things being equal. The cohorts who started practicing in the 1980s suffered the combined impact of the baby boom and a high number of medical school places: they earned the least. The fall in the number of medical school places improved the situation for subsequent cohorts. A stochastic dominance comparison of earnings distribution shows that the disparities connected with unobserved heterogeneity do not compensate for the averag