Soil fertility assessment, principal-agent problem and farmer willingness to pay for soil testing in Uganda

Abstract

Despite the increasing investments to enhance the adoption of soil testing in Uganda, its uptake has remained limited. Farmers have remained reliant on indigenous soil fertility indicators such as soil color, crop vigor, weeds, previous yield, erosion history, presence of soil organisms, and observable organic matter content. However, the low adoption of new agricultural technologies and innovations in Uganda is highly attributed to institutional and market imperfections. Therefore, using data from the Agriculture Cluster Development Project, this study aims to assess indigenous soil fertility indicators, principal-agent problem in soil testing extension, its drivers, and impact on farmer willingness to pay for soil testing among maize and coffee farmers in Uganda. The specific objectives were: to compare indigenous and soil testing soil fertility assessment approaches; To assess the principal-agent problem in soil testing extension service delivery and its drivers; and lastly to determine the impact of the principal-agent problem in soil testing service delivery on farmer willingness to pay for soil testing in Uganda. This dissertation consists of six chapters: Chapter one presents the background of the study, the statement of the problem, the study objectives, research hypothesis, the scope of the study, the significance, justification, and structure of the dissertation. Chapter two presents the literature and the conceptual framework. Chapter three highlights the inaccuracies in indigenous soil fertility assessment and its correlates. The fourth chapter presents the principal-agent problem in the extension system and its drivers. Chapter five highlights the impact of the principal-agent problem on famer willingness to pay for soil testing in Uganda. While the last chapter discusses the key findings, policy recommendations, limitations and areas for further research. Chapter 3 used correlations, a binary probit, and a multivariate probit model to determine the level of misclassification of soil fertility from the indigenous assessment and its correlates. The findings revealed that vegetation cover, weeds, soil colour, and previous yield indicators were associated with 71%, 67%, 66%, and 61% levels of misclassification, respectively. According to the probit model estimates, these misclassifications were significant and positively associated with the plot size and tenure system. Therefore, the results suggest that as far as fertilizer use is concerned, soil testing is the only efficient approach for soil fertility assessment before fertilizer use. To assess the extent of the principal-agent problem in the extension system of Uganda and its drivers in chapter 4, descriptive and inferential statistics and an ordered probit model were used. The findings revealed shirking (moral hazard) and deception (adverse selection) behaviours in soil test results’ dissemination. The Ordered Probit model estimates showed that shirking of soil test results’ dissemination without deception was positive and significantly influenced by larger household sizes and higher education level of the extension worker. However, shirking of the soil test results’ dissemination with deception was negative and significantly associated with larger household size and higher education level of the extension worker as well as farmer’s higher education level and group membership. These results therefore suggest that for the principal-agent problems to reduce in the extension system, highly educated extension officers should be recruited and farmers should be educated and encouraged to join farmer groups for collective action. In chapter 5, the effect of the principal-agent behaviours on farmer willingness to pay for soil testing was evaluated using the Inverse Probability Weighted Regression Adjustment. The Inverse Probability Weighted Regression Adjustment model estimates show that the Principal-agent problem behaviours generally reduced the farmer willingness to pay for soil testing by 14.8%. These results suggest that the principal-agent problem in the agricultural extension system of Uganda is one of the factors significantly limiting the adoption of new farm technologies including precision farming agriculture. Yet, an Ordered Probit model estimates showed that the dissemination of sample soil test results would significantly increase farmer willingness to pay for soil testing by 48.5% other factors held constant. However, an Ordered Probit model estimates also reveal that access to agricultural credits, education, increase in farming experience, and training on soil fertility management can also increase farmer willingness to pay for soil testing. These results suggest that although the principal-agent problem in extension system is one of the issue that needs to be addressed to increase farmer willingness to pay for soil testing but access to agricultural credit, farmer education and trainings need to be improved. Therefore, the following policy reccomendations were made; (1) Before fertilizer use, farmers should be encouraged to complement their indigenous soil fertility knowledge with soil testing. (2) To address the principal-agent problem in extension service delivery in Uganda and enhance trust and accountability between stakeholders, we recommend close supervision, monitoring, and development of a validation system on extension service delivery. (3) In addition, refresher and continuous training of the extension officers should be considered to reduce shirking from extension workers. (4) To enhace adoption of soil testing in Uganda, the government and its development partners should consider easing access to agricultural credits, increasing farmer education and training on soil fertility management as well as subsidizing soil testing services for smallholder farmers.