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 artefact 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.