It is argued that the conventional wisdom which accords priority to investment at the primary level of education in developing countries may be based on methodologically flawed estimates. The problem arises on account of the "filtering down" of educated entrants to the labour market into lesser...

Output accounting methods are used to measure the effect of education on labor productivity. Some of the drawbacks of this approach to estimating the contribution of education are overcome by exploiting a "natural experiment" (Kenya and Tanzania are similar in relevant respects ex cept in their...