Quebec's young disadvantaged ­ poorly educated ­ men performed poorly in terms of employability on the Canadian labour market over the last decade. An analysis of the effects of the government employment schemes designed for them makes this phenomenon even more interesting, especially since the wealth of available data means that these paths can be tracked over a long time period and in terms of seven states (employment, unemployment, out of the labour force, welfare and three different employment assistance and training programmes). We consider a transition model used to estimate the conditional distributions of the time periods in the different labour market states. We take account of the endogenous nature of participating in these programmes. We study the sensitivity of the estimates by comparing a standard semi-parametric estimate with a series of parametric estimates of models with two to three factor loadings. The results show that participation in unemployment insurance schemes increases the frequency of transitions to employment. The young poorly educated men who participate in welfare programmes make the move to employment less frequently than those who do not participate in them. The lengths of time in the different labour market states are sensitive to variations in the welfare scale, variations in the minimum wage and the economic situation when this is represented by the unemployment rate.