To train or to educate? Evidence from Tanzania.

In this paper we ask how the returns to academic education compare with the return to two types of training drawing on labour force data from Tanzanias manufacturing sector. The first is vocational training or attending a technical college as part of schooling, the second is on-the-job training in a firm. There has been much dispute in the literature as to whether the returns to vocational or academic training are higher. We show that in addressing this question in a schooling system where entry occurs at differing levels it is necessary to allow both for the entry level into vocation or technical school and for the characteristics of the firm in which the worker is employed. If the firm fixed effect captures a substantial element of unobserved worker quality then the return to vocational education, at the level at which it occurs, exceeds that on academic education. However as the return to education rises with its level the return to any form of vocational training is less than that achieved by those who reach A-Level and above. While those with current training earn more this effect disappears once we allow for firm fixed effects. One interpretation of this result is that the effects of the training get embodied in the quality of the workforce. The paper highlights the importance of panel data which enables the effects of such unobservables to be identified in assessing returns to both vocational education and training