The Education Production Function reviewed in the framework of Education For All in Sub-Saharan Africa: from theoretical and methodological bounds to the contribution to educational policies

Abstract

The Coleman report (1966) was the first of numerous researches exploring quality of learning determinants analysis. With the Education Production Function (EPF), economists have greatly contributed to the extension of this research domain. Actually, empirical works are leading the field with important methodological debates which have fuelled controversies (Hanushek, 2003; Krueger, 2003). Thus, methodological issues have particularly rallied researchers' interests on empirical estimations, while academic researches on the renewal of the theoretical framework have remained scarce despite obvious bounds. Without contesting the validity of miscellaneous estimation problems and the necessity of other kind of researches as experimental approach, it seems that their consequences are not insurmountable if using a cautious interpretation of the EPF results.<br />The challenges of Education for All (EFA) in Africa and particularly primary universal education stress out the teacher issue. The need of a massive recruitment of teachers till 2015 raises tricky issues as financial sustainability, linked to socio-political problems, but also the profile of these new teachers. On this last point, a selected researches survey shows a relative convergence of results. According to the analyses, teachers' characteristics, such as training, length of service, status, etc. have limited effect on pupils' learning achievement. Teachers' status is a controversial subject in francophone Africa where recruitment of contractual teachers has increased during the last years. The results of empirical studies, as PASEC studies in primary education, do not confirm their responsibility in the weakening education quality. At the opposite, our analysis on Mali shows, in this particular case, a positive effect. Nevertheless, according to this analysis, status and other teachers' characteristics do not play a key role in quality of learning whereas disparities between classes