Randomized Control Trials have become the methodological benchmark in program evaluation, but they cannot always be applied. Ethical, political, institutional or fundamental reasons or simply the cost of data collection may render their implementation impossible. Today, there are also excellent alternatives available which do not always require a data collection for their sole purpose but make use of secondary data. The three essays which form this thesis all apply such methodological alternatives and make use of existing and rich sources of data, complemented with administrative information or data from other sources. Even though all essays can stand by themselves, their objects of study are linked in that they all examine issues which affect the long-term economic development of societies through their impact on human capital formation. The first essay studies the health impact and time savings of a rural water supply project in Guinea. In the second essay, together with fellow PhD student Marcio Cruz, we study how a Conditional Cash Transfer program affects the expenditure choices of its participants. The third essay examines what components in the variation of global food prices are transmitted to developing countries and how they affect household welfare. More detailed abstracts for each of the essays can be found on page iii.