The dissertation has a methodological and a substantive contribution. First, it develops a method of measurement that serves to directly test specific hypothesis about the simultaneous interaction of classes in a multivariate regression environment. I show how compositional analysis can be applied to the same data that are used to construct the Gini coefficient serving to further investigate and adjudicate the class effects that are only suggested in the Gin results. I show that because identical Gini calculations result from drastically different land distributions it is wrong to narrow its interpretation to theories relating only the very rich and the very poor while ignoring the middle class. Second, it shows that municipalities in Colombia with relatively large numbers of middle-sized farmers versus large landowners provide larger proportions of their citizens with piped water, electricity, and other public services. I argue that middle-sized farmers play this role because they have the income and other resources needed to travel to local urban areas to gather relevant information and to lead local Communal Action Boards in bargaining with local elected officials, but they are poor enough to need the public provision of such services. On the other hand, large landowners live in urban areas, and find more efficient to pay privately for their own services than paying for everyone’s via local taxes.