Abstract

In developing countries, public agricultural advisory services (AASs) have been one of the key components of local development in terms of enhancing production and productivity, achieving food security, contributing to poverty reduction, and improving livelihoods in the face of economic, social, and environmental challenges. However, the benefit that farmers obtain out of participating in AASs and the resulting impact depend, to a great extent, by the (intensity of) farmers’ direct and indirect access to these services. The issue of intensity of farmers’ involvement in AASs is also especially important considering the various stages involved in farming – land preparation to post-harvest management. In this regard, the extent of farmers’ participation in such services and the determinants of their intensity of participation are not very well documented in the context of many developing countries. On the other hand, recent studies have shown that the impact of AASs is mixed and that some empirical investigations lack methodological rigor. Furthermore, there is a dearth of context-specific empirical evidence on the impact of new approaches to AAS provision. One such approach is the Farmers’ Training Centers (FTCs) implemented by the government of Ethiopia to improve smallholder farming systems. Although public AASs have been hailed to be a key constituent of agricultural development and poverty reduction strategies in many developing countries, due to declining public budget allocation, alternative ways of providing AASs have been sought and implemented. One way of diversifying AASs is through a fee-for-service arrangement, which proved useful in many developed countries. In the context of developing countries, however, the potential for payment for AASs has yet to be explored in full. This is especially relevant considering the current economic crisis, large number of farmers relying on agriculture for their livelihoods, inefficiencies of public AASs, and specialized needs of market-oriented producers. In order to address these issues that relate to intensity of farmers’ access to on-farm AASs, impact of FTC-based AASs on outcome indicators, as well as willingness to pay for improved FTC-based AASs, this study is carried out in Haramaya district of eastern Ethiopia. More specifically, a field-based household survey was conducted from May to October 2013 on a total sample of 340 farm households. In relation to the first objective of the investigation, i.e., analyzing the predictors of farmers’ intensity of participation in on-farm training and demonstration, due to the nature of the outcome variables (i.e., count outcome data) both the Poisson regression and the negative binomial regression models are employed on data collected from the full sample. The findings of the investigation indicate that a host of factors – relating to human capital, financial capital, physical capital, social capital, and access to infrastructure and services – influence the farmers’ differential involvement in these services. With regard to the impact evaluation, the study made reference to a sub-sample within the full sample (i.e., 90 households trained at FTCs in 2009 and 160 control respondents selected for this purpose). Using the same data collection instrument in both treatment and comparison areas, data were generated specifically for the purposes of impact evaluation. Due to the non-random allocation of FTCs and self-selection issues as well as the cross-sectional nature of data collected, the propensity score matching (PSM) procedure was employed to estimate the causal effect of an FTC-based modular training on farm income. The results of the analysis show a positive and statistically highly significant gain of farm income (excluding chat) by the participants of the modular training, which is between Birr 9,557.47 and Birr 10,387.53 per household, on average. Lastly, in order to analyze whether a payment for improved AAS scheme is a possible way towards diversifying the institutional options for providing AASs, a choice experiment was conducted to estimate the farmers’ willingness to pay (WTP) for services describing improved FTC-based provision of AASs. By using the best-worst approach to elicit preference data from 120 households selected from treatment areas, applying the solemn oath procedure to mitigate hypothetical bias, and estimating the random parameters logit (RPL) model, the WTP estimates are derived for advice, training, and demonstration provided by the FTCs. In doing so, the study shows that the largest WTP values are associated with demonstration, followed by training. A small premium price is also associated with advice in the group that is not exposed to the solemn oath (i.e., the control treatment). Moreover, it is found that signing on a solemn oath form reduces the mean WTP values compared to not signing (but agreeing to tell the truth) and the control treatment. This dissertation consists of three essays: Essay 1 “Predictors of Smallholder Farmers’ Intensity of Participation in On-farm Agricultural Advisory Service Provision in Haramaya District, Eastern Ethiopia”; Essay 2 “Improving Smallholder Farmers’ Income through Farmers’ Training Centers: an Impact Evaluation in Haramaya District, Eastern Ethiopia”; and, Essay 3 “Smallholder Farmers’ Willingness to Pay for Agricultural Advisory Services: Investigating the Effect of Solemn Oath in Mitigating Hypothetical Bias in Best-Worst Choice Experiments in Haramaya District, Eastern Ethiopia”. Each of the three essays begins with an introductory section where the background to the problem and gaps in previous empirical investigations are discussed. In Essays I & III, there is also an additional discussion on this under the section ‘literature review’. The ‘research methodology’ section in each essay presents a detailed account of the sampling procedure, survey design and methods, as well as empirical strategy and estimation. Immediately following this is the result and discussion section where the main findings of the study are presented and discussed. Finally, the conclusion and recommendation section provides some policy implications of the studies.