We analyze the targeting and additionality of the Chilean afforestation program reforms implemented in the mid-1990s. Propensity score matching estimates are obtained by potential erosion categories using random area sample data. The additional afforestation percentage is estimated to be larger for none or low and very severe categories of potential erosion parcels. Even though the program helped maintain forest cover to highly erodible land, afforested parcels with high opportunity costs and possible negative amenity benefits are also attributed to the program. To improve the cost-effectiveness and to avoid misallocation of land and water resources, more effective targeting is required. (JEL Q23, C14)