Informal employment in Bolivia: A lost proposition?

We study participation and relative earnings in the formal, informal, and self-employed sectors in Bolivia. We estimate quantile earnings equations corrected for self-selectivity to address potential biases in the estimates of relative earnings gaps due to the endogeneity of sector participation. Selectivity is significant in all three sectors for all three years studied. The benefits of being more formal like at low quantiles of the informal sector vanish from 1997 to 2002 as the availability of formal jobs decreases. The human capital model is very well fit for 1993 and 1997. In 2002 it is best fit for the formal sector where education and experience explain much of a worker's earnings, and worst fit for the self-employed sector where education does not play a role and experience is only important at high quantiles. We exploit the semi-parametric nature of quantile regression to link the conditional returns to worker characteristics, obtained from the quantile regressions, with the poverty status of households to determine the extent to which unobserved earnings determinants interact with observed characteristics to penalize non-formal workers in poor households. We find that females in non-formal employment suffer the largest penalties. In unreported results (available from the authors upon request) we perform a counterfactual analysis of conditional earnings by sector, decomposing the earnings gaps into differences in endowments of skills and differences in returns to skills. The results suggest segmentation between the formal and informal sector at the lowest conditional quantiles, while higher productivity workers seem to have a choice of which sector to work in