The rapid expansion of the education sector in developing countries an
d the scarcity of public funds have increased the need for an accurate
evaluation of educational policies. Estimates of rates of return to e
ducation have often been used as an integral part of cost-benefit stud
ies and programme evaluation efforts. This study uses household survey
data from Honduras for 1990 to estimate returns to education that all
ow for worker heterogeneity and individual self-selection in the educa
tion process. A sequential estimation procedure is used that enables s
tudy of the interaction between educational attainment and earnings de
termination. It is found that accounting for endogenous educational at
tainment leads to substantially higher estimated returns to education
(compared to traditional least squares estimates). The possible magnit
ude of the bias underscores the importance of recognizing the role of
self-selection and comparative advantage in the education process.