Assessing the Impact of Training on Lowland Rice Productivity in an African Setting: Evidence from Uganda
Introduction
In contrast to the dramatic success in increasing agricultural productivity in Asia since the late 1960s, agricultural productivity has been largely stagnant in sub-Saharan Africa (SSA) (Otsuka and Larson, 2012, Otsuka and Yamano, 2005). Due to the rapid population growth and urbanization in the region, the consumption of rice has been increasing far more rapidly than domestic rice production, thereby necessitating an increase in the net importation of rice (Africa Rice Center, 2008, Balasubramanian et al., 2007). In addition, the sharp rise in improved rice prices in 2007–08 has resulted in serious food insecurity among the poor in this region (Benson et al., 2008, Ivanic and Martin, 2008). Given that rice is a major cereal crop that has great potential for increases in productivity in SSA and is a tradable good unlike other staple crops commonly grown in many countries in SSA, strategic efforts to enhance rice production are urgently required not only for food security but also for income generation (Diao et al., 2008, Kijima et al., 2008, Kijima et al., 2010, Larson et al., 2010, Otsuka and Kijima, 2010). This is clearly the case in Uganda.
Among upland, irrigated, and rainfed lowland ecosystems, rice production in the rainfed lowlands is considered to have the highest potential for productivity growth in SSA. This is because upland ecosystems are not suitable for sustainable rice production (Balasubramanian et al., 2007) and because the cost of constructing irrigation facilities is generally high in SSA (Balasubramanian et al., 2007, Fujiie et al., 2005, Inocencio et al., 2007, Kajisa et al., 2007). Valley bottoms in SSA are wet and, hence, suitable for lowland rice production. Moreover, they have not been fully exploited, even though they can be used to produce rice sustainably with proper management practices. This is the main reason why the recent expansion of the area under rice cultivation in SSA has been concentrated in the rainfed lowlands (Dalton and Guei, 2003, Otsuka and Larson, 2012). Yet, the actual yields of the lowland rice are much lower than its potential (West Africa Rice Development Association (WARDA), 1999).
It is often argued that the low productivity of lowland rice in SSA is due to low input use and the low adoption of improved varieties. However, the fact that appropriate cultivation practices which are widely used in Asia are not commonly applied in SSA has been largely ignored (Balasubramanian et al., 2007). In some cases, for example, the seeds are broadcasted, which often results in a low germination rate and also makes it difficult to maintain the proper spacing.1 Even when transplanting is practiced, straight-row planting is not practiced, thereby failing to assure the required proper spacing. Constructing bunds and leveling are not applied or properly practiced so that the available water is not stored evenly in the paddy fields.2 Since inappropriate cultivation practices are common, the Japan International Cooperation Agency (JICA) has initiated a project on lowland rice production in Eastern Uganda that provides training to rice farmers on lowland rice cultivation practices based on the Asian experience.
This study attempts to assess the impact of this training program on the adoption of improved cultivation practices, the productivity of rice farming, and the income and profit of rice production by using ex-post non-experimental data. Since the coverage and the number of impact evaluation studies on agricultural training projects are limited, further research is clearly warranted (Winters, Maffioli, & Salazar, 2011). Whether training actually leads to the dissemination of better production knowledge is seldom examined, as is the consideration of whether what is learned is actually systematically put into practice. Note that the type of agricultural project we evaluate in this study does not suit the setting of a randomized control trial since the best cultivation practices are different across places with different agro-ecological and socioeconomic conditions, which results in different responses to interventions. We thus use propensity score matching to construct control groups by using non-experimental data. In addition to the estimation of the average treatment effect on the treated, this paper also examines how the production practices change due to the agricultural training projects by estimating the adoption function of new cultivation practices with weighted least squares (Hirano & Imbens, 2001).
The rest of this paper is structured as follows. Section 2 describes the present state of lowland rice production in Uganda and the characteristics of the training project. Section 3 explains the sample data collection in September 2009 and presents the descriptive statistics. In Section 4, the methodologies to evaluate the effect of the training program are explained. Section 5 presents the empirical results. The last section concludes the paper with a presentation of the policy implications.
Section snippets
Lowland rice production in Uganda
In Uganda, about 10% of the country is covered by wetlands or swamps in valley bottoms (FAO, 2006), which are particularly suitable for lowland rice production. In fact, rice is one of the few profitable cash crops that can be grown in the lowlands in this region. In Eastern Uganda which is a major lowland rice growing area, rice cultivation technology and modern rice varieties developed by the International Rice Research Institute in the Philippines were introduced in the 1970s by the Chinese
Sampling
Before expanding the project size, the pilot project was implemented in 2005 and 2006. To evaluate the effect of this pilot project on rice productivity and profitability, we match the data of such pilot project sites with those of households who live in areas where the training project will be implemented. A household survey was conducted from September to October 2009, and information on rice cultivation between September 2008 and August 2009 was collected. By then, JICA had selected the
Average treatment effect on the treated
This study analyzes an ex-post evaluation of the training provided in 2005 and 2006 by constructing the appropriate counterfactual. The average treatment effect on the treated (ATT) of outcome variable Y, conditional on X and T, is given bywhere Y1 and Y0 are the outcomes of farming activities after and before they receive treatment, respectively. X is a vector of covariates and T is the dummy variable taking unity if the farmer received the
Participation in training
The first column in Table 3 reports the estimation results of the training participation at the rice plot level by Probit model, from which propensity scores are estimated. The numbers demonstrate the marginal effects calculated at the sample means. Households with younger household head, membership of local organization (other than rice association), greater experience of lowland rice cultivation, less livestock, lower rainfall, and better access to the nearest district town are more likely to
Concluding remarks
This paper examined the effect of a training program on the adoption of improved cultivation practices, rice yields, and the income and profit of rice production by estimating the average treatment effect on the treated and the adoption function of the cultivation practices with distribution weights suggested by Hirano and Imbens (2001). We found that the training program increased the adoption of constructing bunds and planting in row on average. ATT estimates suggested that the training
Acknowledgment
We appreciate financial support from JICA Research Institute for collecting data. We thank the Makerere University research team led by Dick Sserunkuuma, George Omiat, Mpiriirwe Innocent, and Marion Iceduna for collecting the data and staff members in JICA Uganda office and Toshimasa Kobayashi for useful discussion.
References (27)
- et al.
Increasing rice production in sub-Saharan Africa: Challenges and opportunities
Advances in Agronomy
(2007) - et al.
Productivity gains from rice genetic enhancements in West Africa: Countries and ecologies
World Development
(2003) - et al.
Does piped water reduce diarrhea for children in rural India?
Journal of Econometrics
(2003) - et al.
Does the system of rice intensification outperform conventional best management? A synopsis of the empirical record
Field Crops Research
(2006) Africa rice trends 2007
(2008)- et al.
Impacts in Uganda of rising global food prices: The role of diversified staples and limited price transmission
Agricultural Economics
(2008) - et al.
How do agricultural programs alter crop production? Evidence from Ecuador
Journal of Agricultural Economics
(2011) - et al.
Social learning through networks: The adoption of new agricultural technologies in Ghana
American Journal of Agricultural Economics
(2001) - et al.
Modern rice technology and income distribution in Asia
(1994) - et al.
Toward a green revolution in Africa: What would it achieve, and what would it require?
Agricultural Economics
(2008)