Estimación de frontera de producción para el sector de agua y saneamiento en América Latina

This article has the aim to estimate stochastic frontiers to determine relative efficiency over a sample of Latin American water and sanitation providers, with panel data for the period 2003-2008. The study searches as a byproduct information which helps a regulatory use. It has as novelties, the more recent regional database, the first study on a panel for the region, a discussion on the variables and the better functional specification to describe the technology in use. The regulatory employment of the results can follow a far reach approach, or more limited ones. In the first case ("yardstick competition"), the estimates are instruments to discuss tariffs and performance goals; a second, more limited approach, specially in the international arena, could yield consensus on some standards and practices which can be set looking to better situated peers. In the middle, the approach known as "sunshine competition" has as aim showing best practices and exposing good and bad cases ("naming and shaming"). The production frontier which best fits is a Translog, of the type time varying decay (TVD) model, which point out that the inefficiencies are variable in time and also that they are posible to control by the providers (random effects).