Different channels of impact of education on poverty: an analysis for Colombia

This paper analyses pecuniary and non-pecuniary effects of education on poverty. Two are the main contributions: first, the pecuniary analysis employs the technique of instrumental variable quantile regression, very helpful method when one is interested in the lowest or highest extremes of the distribution function of the dependent variable. In fact, quantile regression offers coefficient estimations for any conditional quantile. The second contribution derives from our purpose to highlight the non-pecuniary returns to education: resources invested in education bring future returns to individuals, not only reflected in monetary earnings, but also in higher levels of satisfaction of basic needs (e.g. health and housing).