Potential gap in Ethiopia: Wheat, maize and sorghum play an important role for Ethiopia’s agricultural production, but are grown in different agro-ecological zones. We defined areas of un-/underused agricultural potential from yield and area gaps of these crops. Yield gaps are areas which have good to very high agro-ecological suitability for the specific crop, but yields of this crop type stay below the national mean (z-scores < 0). Area gaps are areas with good to very high agro-ecological suitability for the specific crop, but the % area used for this crop type related to total cereal area is below the national mean (z-scores < 0). Green areas show yield and area gaps of maize, sorghum and wheat. A value >1 signifies the overlap of several gaps. Crop specific agro-ecological suitability was defined from the raster data set of Fischer et al. (2002) Data source: Fischer et al. (2002): Global Agro-ecological Assessment for Agriculture in the 21st Century: Methodology and Results. International Institute for Applied Systems Analysis, Laxenburg, Austria Yield and area information for the specific crops was obtained from the “Crop Production Forecast Sample Survey, 2011/12” (CSA, 2011)

Agro-ecological suitability for rainfed crops in Ethiopia based on the raster data set puplished by Fischer et al. (2002) Fischer et al. (2002) produced a worldwide classification raster of agronomic suitability based on crop modeling, including climate, soil and slope data. The mean suitability value of each woreda was calculated from the raster set and mapped on woreda level. Data source: Fischer et al. (2002): Global Agro-ecological Assessment for Agriculture in the 21st Century: Methodology and Results. International Institute for Applied Systems Analysis, Laxenburg, Austria

There is an increasing interest of acquiring farmland abroad, especially following the food price crisis in 2007/08. East Africa is a hotspot of activities, and given the high prevalence of poor people in the area, impacts on rural livelihoods are expected to be substantial. Following significant primary data collection in Ethiopia and Uganda, the study analyses the impact of two such large-scale land acquisitions on the rural economy and the local population’s livelihood, using Theory-based Impact Evaluations (Hemmer 2011) within an analytical framework of layered social analysis (Williamson 2000). Impact is assumed to manifest through five major channels: land, labour, natural resources, technological & organisational innovation and institutional change. The study consists of five chapters: The introduction surveys the global trend, reviews existing evidence and relevant theory to elaborate a conceptual and an analytical framework for the research. The second chapter takes stock of trend and types of large-scale land acquisitions in Eastern Africa, using national official data from Ethiopia and Uganda. While there is a clear increase in number of land transactions, media reports are only confirmed in a small fraction. Investors are coming from Europe, the Arabic peninsula as well as other emerging economies in the global South (South Africa and India, specifically). However, a surprisingly large number of acquisitions is done by domestic investors. The third chapter analyses the early stage impact of a large scale land acquisition in the far western lowlands of Ethiopia. A Saudi-Ethiopian investor tries to develop 10,000 ha for irrigated rice production. Building on primary household data and qualitative information gathered in the area in 2010, a mathematical programming model is calibrated to quantify likely impacts ex-ante. The investment is found to have poverty reducing potential, mainly due to employment creation and growth of the rural non-farm economy. However, the local population has to bear uncompensated costs of lost forestland and local inequalities are likely to widen in consequence of unequal participation on employment and business opportunities. The fourth chapter examines a forty year old large-scale investment in Uganda to understand long-term impacts, especially regarding technological and organisational innovation, as well as institutional change. Using an institutional economic analysis, changes at the organisational structure of the investment can be related to broader changes in the surrounding rural economy, indicating the significant impact a LSLAs can have on rural transformation. Again, the investment has overall contributed to poverty reduction, but organisational flaws and the collapse of a contract farming scheme indicate the difficulties to govern the large farm well. The emergence of a land market for wetlands, adoption of rice as a new crop and organisational improvements among smallholders can be considered as major outcomes of the investment’s activities. The fifth chapter synthesises the early three empirical chapters and locates the findings within a broader set of trends regarding the commercialisation of the agri-food system, the discussion on optimal farm size for production and poverty reduction, and the importance of functioning land and labour markets for poverty reduction and rural transformation in developing economies.

Degree of socio-economic marginality in Ethiopia, using the following economic, health and educational positional indicators: 1. Economy 1.1 Access to markets (travel time to next city > 50.000 capitals) 1.2 Transportation infrastructure (average all-weather road density m/km2) 1.3 Information infrastructure (% of households without telephone) 2. Health: 2.1 Access to education (no. of primary schools related to population density) 2.2 Access to information (% of households without radio) 2.3 Access to information (% of households without television) 3. Education: 3.1 Access to education (no. of primary schools related to population density) 3.2 Access to information (% of households without radio) 3.3 Access to information (% of households without television) Data source: 1.1/1.2/1.3/2.2/2.3/3.2/3.3: Central Statistical Agency (CSA) (2007): Population and Housing Census. Atlas of Ethiopia 2007. Washington DC, USA 2.1/3.1 Central Statistical Agency (CSA), Ethiopian Development Research Institute (EDRI), International Food Policy Research Institute (IFPRI) (2006): Atlas of the Ethiopian Rural Economy. Washington DC, USA

Degree of socio-economic marginality in Ethiopia, using the following conditional indicators of the economic sector: 1. Regional poverty headcount indices (% of population whose income/consumption is below the poverty line = 3781 birr) 2. Food poverty headcount indices (% of population whose income/consumption for food is below the cost of 2.200 kcal/day per adult food consumption) 3. Wealth index (% of population being part of the lowest/2.lowest wealth quintile) Data source: 1.+ 2. Ministry of Finance and Economy Development (2012): Ethiopia‘s Progress Towards Eradicating Poverty: An Interim Report on Poverty Analysis Study (2010/11). Addis Ababa, Ethiopia 3. Central Statistical Agency(CSA), ICF International (2012): Ethiopia Demographic and Health Survey 2011. Addis Ababa, Ethiopia; Calverton, USA

This marginality hotspot map of Ethiopia uses the lowest quartile as thresholds for the dimensions of marginality. Again, this map shows how many dimension of marginality - as defined by Gatzweiler et al. (2011) - overlap.

Poverty and inequality in Bolivia have reduced to a great extent in the last 20 years in Bolivia. There are mixed opinions regarding the role of the state in this overall positive result, and consistent evidence of state intervention is still missing. This dissertation aims to explore the topic of the impact of government intervention on inequality and poverty from three different perspectives. In the first chapter, I frame the theoretical framework and set the research questions of each of the chapters of the dissertation. In the second chapter, I examine the impact of a policy experiment in Bolivia in 2007/2008, in which the payment method of a cash transfer changed from a yearly lump sum to monthly installments. Both amounts do not differ if we take them in full, but the change in the payment method could have an impact given inherent behavior-specific constraints like lack of control of expenditures, propensity to overspend and inability to save regularly. I am interested in the effects that this policy change might have had on educational outcomes when the outcomes of those affected by the policy change are compared with those who were not affected. Results show an increase in attendance (around three percentage points) and a decrease in child labor (by eight percentage points) for older children (attending secondary school). The results are fairly robust to the use of different specifications. This suggests that a smaller but more regular, constant in time and predictable flow of cash transfers can be preferable to a once-a-year significant lump-sum transfer. In Chapter 3, I evaluate the impact of increased fiscal decentralization on outcomes as nutrition, access to safe water and sanitation in Bolivia during the 2000s decade. The results show that fiscal decentralization has not increased the access of the population to safe water or sanitation. Meanwhile, nutritional status of children less than five years old has slightly improved during the study period, suggesting a positive impact of increased decentralization on nourishment indicators. The inclusion of other dimensions of decentralization policy (like administrative decentralization and the role of political institutions) are also analyzed, showing important interactions with fiscal decentralization. On the other hand, decentralization does not appear to be pro-poor, as the results show that the progress on nourishment indicators was more considerable in non-poor municipalities versus poor municipalities. These results are robust to different thresholds and deprivation measures. In the last chapter of the dissertation, I study the topic of horizontal inequality. Horizontal inequality refers to the difference in income (or another welfare indicator) due to membership in a specific group (e.g., determined, by race, gender, location, etc.). This difference could be relevant in a context in which particular groups have been historically excluded, as the case of indigenous people in Bolivia. In this chapter, a tax-benefit incidence analysis model is used to assess the role of net public transfers on horizontal inequality in Bolivia for the year 2015. The group categories that are subject of the analysis are defined by ethnic status, gender, and location, besides a combination of these categories. Results show that the most significant group inequality is observed when the indigenous status is defined using an ethnolinguistic metric. However, the role of self-identification in determining indigenous status is less important in explaining the income gap. While the fiscal system seems to be progressive for indigenous and urban/rural categories, this progressivity is not present when the gender dimension is assessed.

Data on investment licenses from Ethiopia's Licensing office in Addis Abeba (Feb 2011). For each investment, country of origin, requested land size, some information on planned production, as well as employment and capital is contained (N=2814). Data quality: This data-set was received from the Ethiopian Investment Authority during a field visit in 2011. It contains data from investors that invested in the agricultural sector and requested some land. Unfortunately the land size information was very noisy (Units not clearly specified and missing values). Some land was requested in ha, some in sq.m. Processing: The data set attached was filtered based on size-capital assumptions, leaving out those investments that are purely processing (less land intensive). A cut-off point of 100 ha was chosen and 2814 observations for the period 1992 to Dec 2010 remained. This data set was used for Baumgartner (2012) Large-scale investments in Ethiopia, in (eds.) Allan et al.: Handbook of Land and Water Grabs in Africa, Routledge; and in Baumgartner et al. (2015) Poverty impacts of Large-scale land acquisitions, WDev.

Degree of socio-economic marginality in areas with capabilitiy gaps Socio-economic marginality: Socio-economic marginality in Ethiopia was defined by the following economic, health and educational conditional indicators: 1. Economy: 1.1 Regional poverty headcount indices (% of population whose income/consumption is below the poverty line = 3781 birr) 1.2 Food poverty headcount indices (% of population whose income/consumption for food is below the cost of 2.200 kcal/day per adult food consumption) 1.3 Wealth index (% of population being part of the lowest/2.lowest wealth quintile) 2. Health: 2.1 Child mortality rate (no. of deaths out of 1000 live births <5 years) 2.2 Nutritional status of children (% of children <5 years being stunted) 2.3 Nutritional status of adults (% of men/women age 15-49 with BMI <18.5 = acute under nutrition) 3. Education: 3.1 Illiteracy rate (% of population not being able to read/write in their native language) 3.2 Net enrolment ratio primary school 3.3 Net enrolement ratio high school Data source: 1.1/1.2: Ministry of Finance and Economy Development (2012): Ethiopia‘s Progress Towards Eradicating Poverty: An Interim Report on Poverty Analysis Study (2010/11). Addis Ababa, Ethiopia 1.3/2.1/2.2/2.3: Central Statistical Agency(CSA), ICF International (2012): Ethiopia Demographic and Health Survey 2011. Addis Ababa, Ethiopia; Calverton, USA Capability gap: Areas with good agro-ecological suitability, but limited socio-economic capabilities of farmers to make use of this suitability. Agro-ecological suitability in Ethiopia was defined from the raster data set of agro-ecological suitability for rainfed crops (Fischer et al. 2002) Data source: Fischer et al. (2002): Global Agro-ecological Assessment for Agriculture in the 21st Century: Methodology and Results. International Institute for Applied Systems Analysis, Laxenburg, Austria The socio-economic capabilities of farmers were defined by the following indicators: 1. Access to technology (% of holders applying inorganic fertilizer to any crop during Meher season) 2. Access to credit (% of holders utilizing credit services) 3. Access to knowledge (% of holders utilizing advisory services) Data source: Central Statistical Agency (CSA) (2002): Ethiopian Agricultural Sample Enumeration. Addis Ababa, Ethiopia

Using the methodology developed by Graw and Husmann (2014), the map overlays three indicators (high agricultural potential, high poverty mass and high yield gaps) to identify areas with high potential for agricultural development and poverty reduction in Kenya. Data sources and thresholds: Agricultural potential: Suitability of currently available land area for rainfed crops, using maximising crop and technology mix, FGGD map 6.61 (2005), High: top 3 suitability classes (medium high, high and very high) Poverty mass: Number of poor people in Kenya (by district), KIHBS (2005/06), High: >300,000 per district Yield gap: Yield gap for a combination of major crops, FAO/IIASA - GAEZ (2000/05), High: < 0.25 (on a scale from 0-1, with the highest value in Kenya ca. 0.44) District boundaries: Kenya Central Bureau of Statistics (2003)