MEASUREMENT ERROR BIAS IN RETURNS TO EDUCATION: EVIDENCE FROM A DEVELOPING COUNTRY - SRI LANKA

There is a continuing debate about the size and direction of the bias in estimates of returns to education. Evidence from developing countries is particularly scarce. This paper addresses the problem of measurement error bias in returns to schooling for Sri Lanka, by exploiting dual measurements of reported schooling for a sub-sample of the data and deriving a reliability estimate of schooling. This is used to obtain measurement error corrected fixed effects estimates of the proportionate increase associated with an additional level of schooling. This corrected measure is 5.5%, which is less than the OLS estimate of 7.8% for two person households.