Adoption of Soil Conservation Technologies in El Salvador: A cross-Section and Over-Time Analysis

This paper examines the determinants of adoption of soil conservation technologies among farmers participating in the Environmental Program of El Salvador (PAES). A sample of PAES participants was surveyed in 2002, from which a sub-sample was re-surveyed in 2005 along with a control group of non-participating households. This is an ex-post evaluation of PAES and there is no baseline and no randomized control group. Therefore, quasi-experimental techniques were applied to select the treatment and comparison groups after the intervention occurred. Matching techniques were employed to construct a comparison group that resembles the treatment group based on observed characteristics while statistical controls were applied to measure differences on adoption of soil conservation for the treatment group at two points in time (2002 and 2005) and between the treatment and comparison groups at a given point in time (2005), allowing for before-after and with-without comparisons. Two adoption models (Conservation Practices and Conservation Structures) were estimated using two separate samples (PAES beneficiaries 2002-2005 and beneficiaries-control group 2005). We then analyzed the factors associated with disadoption of the technologies in 2005 relative to 2002. Overall, the adoption of practices and structures is positively associated with schooling, off-farm earnings, crop diversification, technical assistance, participation in social organizations, frequency of extension visits and access to local markets. More importantly, PAES beneficiaries have increased significantly the area treated with conservation between 2002 and 2005. The results confirm that the project intervention (frequency of extension visits and years with project) is positively associated with adoption. Our findings also indicate that access to agricultural markets and rural infrastructure and institutions is crucial for adoption and diffusion of new technologies. Two probit models are estimated to examine the disadoption of practices and structures using the 2002-2005 data for PAES beneficiaries. These models indicate that off-farm income, education, frequency of extension visits, years with PAES and participation in social organizations are significant contributors to reducing the probability of disadoption of conservation technologies. In turn, disadoption is significantly associated with farm size. Disadopters of practices tend to be younger than adopters and have more diversified and larger farms with more access to input and output markets. In turn, disadopters of structures tend to have larger and more steeply farms, are more involved in social organizations and have less access to input and output markets and local infrastructure.