1 Introduction

Twenty years ago, the first round of RePEAT surveys was conducted in Uganda, Kenya, and Ethiopia. Before actual data collection started, Takashi Yamano, Keijiro Otsuka, and Frank Place visited these countries many times for preparation, and the name of the project ‘RePEAT’ came to mind as they visited the fields. The sampling in Uganda and Kenya was based on the surveys conducted before the project by the International Food Policy Research Institute (IFPRI) led by John Pender and the Smallholder Dairy Project (SDP).Footnote 1 Their approval to use their sample for the project was instrumental. For sampling in Ethiopia, Takashi’s student at Michigan State University (MSU), Berhanu Gebremedhin, provided a huge help. The project has benefited from experienced survey teams from Makerere University (Dick Sserunkuuma, George Sentumbwe); Tegemeo Institute; World Agroforestry Center; International Livestock Research Institute; and Ethiopian Development Research Institute (EDRI).

Research members based at the National Graduate Institute for Policy Studies at the time of surveys include Keijiro Otsuka, Takashi Yamano, Yoko Kijima, Tomoya Matsumoto, Chikako Yamauchi, and Alistair Munro, with the help of Rie Muraoka and Ryo Takahashi. Many graduate students that they supervised used RePEAT data for their dissertation.

RePEAT is a pioneering work in creating a panel dataset from Africa.Footnote 2 This is a research work whose brain and the initiator is Professor Keijiro Otsuka. To date, no other development economist has done extensive and deeply intensive work on household surveys in Africa.

This chapter lists up to 62 research outputs produced using RePEAT data and categorizes them by topics covered, countries, and survey rounds used. This catalog helps future users of RePEAT data have an overview of the project and identify which research areas are understudied.

Section 8.2 of this chapter describes the basic characteristics of the RePEAT survey and Sect. 8.3 reviews the existing literature, which used the data from the RePEAT survey. Finally, Sect. 8.4 discusses the future of the project.

2 Description of the Survey

The population of the survey is composed of farm households in rural areas. While the number of original sample households is 940, 934, and 420 in Uganda, Kenya, and Ethiopia, respectively, due to attrition in the long panel data (10–14 years), the number of households that are consistently present in all rounds is 647, 605, and 376. In the last survey round for each country, additional households were added to the sample. In Ethiopia, 100 communities were initially selected for the community survey, but due to budget constraints, households were selected from 40 communities instead. The 2014 survey, however, expanded the household survey into these 100 communities and added missed households in the originally selected communities (Yamano et al. 2011a, b). Thus, 1,366 households were interviewed in 2014. In the 2018 Kenya survey, sample households were added and replaced, resulting in the total number of households becoming 1,228. In the 2015 Uganda survey, the replacement and addition of households were done. The northern region was also included in the survey (the total number of sample households is 1,735).

The project aims to understand the problems and constraints rural farm households in East Africa face by constructing panel datasets. How they can move away from these constraints is also investigated. Questionnaires contained detailed questions to capture farm household income comparable among three countries.

3 Studies by Theme and Country

The references list 62 studies that used RePEAT data. In Table 8.1, they are categorized by theme and country. There are five sets of studies. We summarize their main objectives and findings and highlight their contributions to the literature.

Table 8.1 Research categorized by theme and country

The first set of research topics is related to agriculture, such as agricultural productivity, agricultural technology adoption, agricultural intensification, soil quality and soil management, demand for fertilizer, agricultural market participation, integrated farming system, and price analyses of agricultural outputs and inputs. Given that RePEAT sample households mainly depend on agriculture for their livelihood, enhancing agricultural productivity is crucial for their welfare. The adoption of modern technologies (hybrid seeds and chemical fertilizer) and the use of organic matter to manage soil quality are critical for enhancing production in these countries (Matumoto and Yamano 2013). For these analyses to be conducted, the survey collected soil samples from the field and created objective measures of soil quality, which is normally “unobserved” in the literature (Yamano and Kijima 2011).

The second set comprises studies about land and the environment, such as land markets, land conflicts, climate change, forest management, and communal grazing lands. As the population pressure increases, access to land becomes scarcer, which requires the land rental market to function better for better resource allocation. Detailed questions about land transactions and land tenure for all the parcels households revealed how the land rental market is used, specifically from whom and to whom (or the other way around), leads to different welfare implications. In Uganda and Kenya, land rental markets function to make land allocation more efficient and more equitable as the land is rented out from less productive to more productive farmers, and land is transferred from land-abundant to land-constrained households (Kijima and Tabetando 2020). Another feature of the RePEAT survey was that it was flexible enough to expand questionnaires and area coverage to examine issues faced by farmers. By conducting surveys just after the political violence in Kenya and expanding the sample to Northern Uganda, where many households were displaced due to armed conflict, the causes of land conflicts were examined. The formal land title may help when there are unexpected events. During the political violence in Kenya, those with no land titles were more likely to be victimized (Yamano and Tanaka 2014). Those displaced in Northern Uganda for a longer time were likely to have land conflicts after returning home.

The third set of studies focuses on poverty, ranging from typical measures (per capita consumption expenditure and income) to food security, livelihood diversification, and income from nonfarm employment and migration. As farm households tend to be affected by weather-related shocks, leading them to transitory poverty, diversifying their income source is an important livelihood strategy. This is even more essential in recent years as droughts and floods have occurred more often due to climate change. Using the RePEAT surveys can capture the kinds of diversification strategies households have taken in the long-term. In Kenya and Uganda, income and crop diversification positively impacted household per capita income (Tabetando and Kijima 2022). Another way to mitigate the negative impact of the shocks is to have coping strategies, such as saving, borrowing, and getting a loan. The recent development of information and communication technologies (ICTs) plays a critical role in reducing the transaction cost of accessing such financial services and increasing remittances received (Munyegera and Matsumoto 2016).

The fourth set of studies is on human capital investment. In Uganda and Kenya, a free primary education policy has been implemented. That allowed us to test if exogenous shocks of reduced education cost affected child education investment decisions, especially for girls in poorer households. It is hoped that later when these children become mothers, they may invest more in their children’s human capital. However, the quality of education in public primary schools have deteriorated due to increased enrollment, which affects school choice to private schools (Nishimura and Yamano 2013).

The final set of studies is about research methodologies. In Uganda, a randomized controlled trial (RCT) was implemented between 2009 and 2012 to investigate the price elasticity of demand for maize hybrid seeds and chemical fertilizer (Matsumoto 2014). Furthermore, in 2012 and 2015, lab-in-the-field experiments to elicit preferences were conducted in Uganda (Tanaka and Munro 2014; Masekesa and Munro 2020). This is another way of utilizing the panel data covering the entire country. In the last round of the Kenya survey, one section was used to test whether how questions are asked results in different responses or not (Munro 2020).

In sum, more studies are using Uganda data, followed by Kenya data. One reason is that Uganda’s survey team has been flexible and experienced in allowing the addition of more questions on themes that did not originally exist (such as health) and different research methodologies.

Table 8.2 shows the list of these studies by survey round used and country. Most of the studies use panel data from one country. However, as mentioned earlier, questions and research methodologies were added in the 2015 Uganda survey, which increased the number of studies using only fifth-round data. Studies with international comparisons were produced since the first and second rounds of questionnaires were almost the same among the three countries. So far, the datasets of the three countries collected in the first two rounds have been publicly available from the website (http://www3.grips.ac.jp/~21coe/e/data/content/main.html). But there is now website for the other rounds. Making data public and enhancing accessibility to the datasets is also an urgent issue.

Table 8.2 Research categorized by survey rounds and country

4 Concluding Remarks

This chapter reviewed studies that used data from the RePEAT project in three East African countries from 2003 to 2018. As of February 2022, 62 journal papers and book chapters have been published. Research topics have covered ones related to the project’s original purposes (poverty reduction, environment, and agricultural technology) and expanded to incorporate emerging issues in the survey villages, such as the introduction of mobile money, land conflicts, and climate change. Studies were also done to understand household welfare other than income and expenditure (education and health). The unique datasets provided Ph.D. students in the National Graduate Institute for Policy Studies (GRIPS) with rare and valuable opportunities to utilize the panel data, conduct lab-in-the-field experiments on the sample households, add some questions in questionnaires, and examine new research questions as a part of the project.

As initially intended by Professor Keijiro Otsuka, the RePEAT project will continue as it is aptly named ‘repeat.’ The project needs to examine new challenges faced by rural households. Furthermore, as the original sample households became older than the national average, national representativeness is no longer guaranteed though efforts to add extra households into the sample have been made in the last waves. Given limited funding and the increase in labor costs for conducting the survey, the project team must balance increasing the sample size with adding new questions.

5 Recollections of Professor Keijiro Otsuka

I first met Professor Otsuka in 2014 during my first year as a Ph.D. student at GRIPS. I was immediately attracted to his profound understanding of development issues in Sub-Saharan Africa. His course titled Strategies for Economic Development transformed my understanding of development economics. I am very thankful to him for all his effort to train a new breed of African academics and policymakers. We will not stop disseminating his ideas and values to our friends, colleagues, and students.

Rayner Tabetando.