Labour Market Outcome of 1976 Universal Primary Education in Nigeria

Abstract

In this study attempt to estimate the impact of schooling attainment on an important indicator of labour market performance: wealth. OLS and IV regressions produced economically and statistically significant estimates, with OLS estimate of about 18 percent and IV estimate of about 30 percent when pooled DHS is used and about 56 percent when HNLSS data are employed. We have no evidence that OLS estimates are an artifact of the way the dependent variable is constructed or influential observations are driving observed outcome. In our IV regression specification, econometric tests prove that instrument is strong. Indicative and formal tests of instrument validity such as addition of new relevant variables, falsification tests, plausibly exogenous test and over-identification test are proofs of instrument validity. A number of econometric strategies implemented indicate that influential observations and selective migration are not biasing our results.