In this paper, we have used the standard Human Capital model to describe the post-compulsory schooling behaviour of Sri Lankans. We assumed that there is no uncertainty in the education system or in the labour market. Therefore, inthe steady-state, the earnings profile of one generation is a replica of the earnings of the next generation. Then, we modeled and estimated the school enrolment and the length of schooling decisions.Our results show a very clear positive association between the family background and the education decision. Children of more affluent families seem to derive more benefits from the free education policy. In particular, mother's education has a very strong effect on the education of the child. This casts doubt on the effectiveness of the free education policy as a poverty alleviation instrument and its role in social mobility.We also observe that the ability effect on the return to education is greater than the ability effect on the cost of education. Therefore, more able children stay longer in full-time education than the less able children. We further found some negative evidence on the Behrman and Taubman observation of the negative birth order effect on ability. We observe that younger siblings are more able than older siblings of the same family.By decomposing the total variance of the schooling length, we observed that the rate of return variation is more important in explaining the schooling behaviour than the variation in the cost of education.Analysis of the residuals has given us an impression that there is a family fixed effect which is not explained by the model. Most probably, the poor specification of the ability and the excluded school quality variables would be the main reason for this unexplained fixed effect.