Household productivity and investment in soil conservation: Evidence from small-scale hillside farmers in Central America

Alleviating rural poverty and conserving natural resources are shared goals among governments in Central America. Even though several rural development programs have been established with those objectives as main driving forces, little empirical evidence exists regarding their effects on household productivity. Consequently, the main goals of this dissertation are to estimate the level of Technical Efficiency (TE), analyze the structure of the production technology and evaluate the determinants of inefficiency among hillside households associated with two natural resource projects in El Salvador and Honduras. This dissertation also compares the empirical performance of the three most commonly used methods to measure TE (SPF, SDF and DEA) for a sample of 639 households and examines the impact of different levels of investment in soil conservation on productivity. ^ The empirical results show a significant difference between the stochastic and determinist mean TE scores. The low level of TE obtained with the determinist approach could be explained by the fact this method do not account for stochastic noises when estimating TE. Thus, the stochastic techniques seem to be more appropriate to analyze TE for the group of households under analysis. However, the various TE rankings are all positively correlated suggesting that the methods do lead to some degree of consistency. The results also suggest that stochastic TE estimates are not significantly affected by the functional form or by the distribution of the one-sided error term. Despite the methodology used, the results of this study reveal substantial inefficiency for household production in El Salvador and Honduras. ^ The stochastic analyses show that off-farm income plays a significant role in determining the level of household efficiency. The determinants of inefficiency indicate that human capital is the most important element in defining household productivity. The findings also suggest the presence of market failures and gender inequality in the areas under analysis. ^ Finally, the adoption of soil conservation practices proves to be an influential element on determining TE. A switching regression model shows that farms with high levels of investment in soil conservation display higher average TE than farms with low levels of investments.