Primary school enrolment rates are very low in francophone Africa. In order to enhance education supply, many countries have launched large teacher recruitment programmes in recent years, whereby teachers are no longer engaged on civil servant positions, but on the basis of (fixed-term) contracts typically implying considerably lower salaries and a sharply reduced duration of professional training. While this policy has led to a boost of primary enrolment, there is a concern about a loss in the quality of education. In this paper we analyse the impact on educational quality, by estimating nonparametrically the quantile treatment effects for Niger, Togo and Mali, based on very informative data, comparable across these countries. We find that contract teachers do relatively better for low ability children in low grades than for high ability children in higher grades. When positive treatment effects were found, they tended to be more positive at the low to medium quantiles; when negative effects were found they tended to be more pronounced at the high ability quantiles. Hence, overall it seems that contract teachers do a relatively better job for teaching students with learning difficulties than for teaching the ‘more advanced’ children. This implies that contract teachers tend to reduce inequalities in student outcomes. At the same time, we also observe clear differences between the countries. We find that, overall, effects are positive in Mali, somewhat mixed in Togo (with positive effects in 2nd and negative effects in 5th grade) and negative in Niger. This ordering is consistent with theoretical expectations derived from a closer examination of the different ways of implementation of the contract teacher programme in the three countries. In Mali and, to some extent, in Togo, the contract teacher system works more through the local communities. This may have led to closer monitoring and more effective hiring of contract teachers. In Niger, the system was changed in a centralized way with all contract teachers being public employees, so that there is no reason to expect much impact on local monitoring. In addition, the extremely fat hiring of huge numbers of contract teachers may also have contributed to relatively poor performance in Niger. These results are expected to be relevant for other sub-Saharan African countries, too, as well as for the design of new contract teacher programmes in the future.