Agriculture, Aid, and Economic Growth in Africa

How can foreign aid to agriculture support economic growth in Africa? This paper constructs a geographically indexed applied general equilibrium model that considers pathways through which aid might affect growth and structural transformation of labor markets in the context of soil nutrient variation, minimum subsistence consumption requirements, domestic transport costs, labor mobility, and constraints to self-financing of agricultural inputs.Using plausible parameters, the model is presented for Uganda as an illustrative case.We present three stylized scenarios to demonstrate the potential economy-wide impacts of both soil nutrient loss and replenishment, and how foreign aid can be targeted to support agricultural inputs that boost rural productivity and shift labor to boost real wages. One simulation shows how a temporary program of targeted official development assistance (ODA) for agriculture could generate, contrary to traditional Dutch disease concerns, an expansion in the primary tradable sector and positive permanent productivity and welfare effects, leading to a steady decline in the need for complementary ODA for budget support.


Supplementary Online Appendix S3: Model Parameters
This paper focuses on the presentation of a general economic model rather than precise point estimates. Future research could usefully identify more refined subnational parameter values for Uganda. The parameters used for this model are presented below, and are set to a general baseline of the year 2002. Factor shares in production functions are assumed at standard values (e.g., see Eberhardt 2006). The cost of capital, depreciation rates, the FDI multiplier, share of services in consumption, and public sector service wage premiums are assumed. Capital stocks and productivity terms were estimated using the Uganda social accounting matrix for 2002 (Alarcon et al. 2006), and then scaled. Public sector import contents were set at one for simplicity. The labor force is set to match the approximate reality of 19.6 million workers as of 2002, and the population of origin from each region matches the approximate real distribution of populations. The distribution of roads and agricultural capital stocks across the four regions provides another approximate match of the reality in Uganda. The northern region is poorest, and therefore starts the first period with slightly higher transport losses, lower cash crop capital stock, and lower cash crop productivity.

(B) General
impcontc p Import content share in commodities for public sector p impcontc p = 1 impcontk p Import content share in capital for public sector p impcontk p = 1 ltot t Total labor ltot t = 19.6 pimpr t Price of imported goods in rural area pimpr t = 1 pimpu t Price of imported goods in urban area pimpu t = 1 (C) Regional

Supplementary Online Appendix S4: Overview of Uganda as Illustrative African Economy
Uganda faces many core challenges common across low-income African subsistence economies. Some key characteristics are described here. These draw from a range of sources, mainly published during the course of the early 2000s, and thus present a thematic overview rather than a precise snapshot at a single point in time. The data also predate Uganda's recent commencement of oil production. As of 2012, approximately 35 percent of Uganda's population still lived below the international extreme poverty line of $1.90 per day in 2011 purchasing power parity terms (World Bank 2017). The vast majority of the country's poverty is concentrated in rural areas, where most Ugandans are engaged in crop agriculture. Infrastructure is limited. Only approximately 10 percent of households had electricity as of the early 2000s (Okidi et al. 2005). For decades, gross domestic saving rates were extremely low, well below 10 percent of GDP, although they averaged a slightly higher 13.1 percent from 2006 to 2015 (World Bank 2017). A 1997 Bank of Uganda survey found that fewer than a quarter of rural Ugandans had ever saved and that 85 percent of the other three-quarters cited low income as the primary factor for not doing so (Musinguzi and Smith 2000).

Staple Food Agriculture
Uganda's staple agriculture sector has experienced general long-term stagnation. Figure S4.1 presents trend data for cereal production per capita from 1961 to 2014. From a peak of nearly 180 kilograms per person in 1969, output has been stagnant, at less than 100 kilograms per person, since the early 1980s. Figure S4.2 indicates similar trends for a broader index of food production per capita. Siriri, Bekunda, and Jama (2005) find that yields are typically one-quarter to one-tenth of current potential. This is significantly driven by the low usage of modern farm inputs. Fewer than a third of agricultural households use improved seeds, and only 8 percent use inorganic fertilizer (Okoboi and Barungi 2012). Yet despite the stagnation, Uganda has not become a marked food importer. Historically the coun- try has engaged in almost no staple crop trade; for many years nearly all its imported food was wheat and maize aid for conflict-affected areas in the north.

Soil Nutrients
Like much of Africa, Uganda faces a major soil nutrient challenge. For many years, agricultural output was maintained through land clearing, but population pressures (see Figure S4.3) and a lack of fallowing mean that farmers are now mining nutrients at faster rates and decreasing long-term yields in the process. Data compiled by Ssali (2002) and Ruecker (2005) indicate that a large portion of Uganda's soil is now below the so-called critical 3 percent value for soil organic matter. Henao and Baanante (2006) estimate loss rates for nitrogen, phosphorous, and potassium to be among the highest in Africa, at more than 60 kilograms per hectare per year. The consequences are significant. Nkonya, Kaizzi, and Pender (2005) estimate a cost of approximately $153 per household per year to replenish mined soil nutrients at market prices, equivalent to nearly a fifth of GDP at the time of calculation.
Fallow periods have fallen from 10 to 15 years a century ago down to two and even zero years today (Nandwa and Bekunda 1998). In large parts of the country, fewer than 10 percent of farms have been estimated even to use fallows (Pender et al. 2001). Fertilizer is necessary, even if not sufficient, to stop and reverse the patterns of nutrient decline and address the soil nutrient challenge. However, cost is a barrier because staple crop farmers often face poor relative returns on fertilizer, often with a "value to cost ratio" of 1 or less (e.g., Wortmann and Kaizzi 1998;de Jager, Onduru, and Walaga 2003;Kaizzi 2002;Matsumoto and Yamano 2009).
Uganda's internal geographic heterogeneity underscores the disparate range of farming systems across Africa. Ruecker et al. (2003) identified seven categories of agricultural potential across the country based on permutations of four climate variables. First, annual precipitation cycles affect the extent of water availability throughout the year. The northeastern section of the country has unimodal rainfall, while the southern and central areas, which are closer to the equator, have bimodal rainfall. Second, the length of growing period is measured as the period over which mean monthly rainfall exceeds half the mean potential evapotranspiration. This ranges from less than five months in the northeastern districts to 10 or more months in the central region and in the southwestern highlands. Third, the actual level of annual precipitation varies tremendously throughout the country and changes at a steep gradient, particularly in the "crescent" around Lake Victoria. Fourth, extreme temperatures constrain agricultural productivity. The range of growing conditions results in significant variations in the concentration of staple crop by region.

Other Key Sectors
Cash crops, especially coffee, have historically been a major driver of Uganda's growth and poverty reduction. However, fewer than 10 percent of the country's farm households grow coffee, and the commodity's share of exports has declined significantly (Kappel, Lay, and Steiner 2005;Bussolo et al. 2006). As of the mid-2000s, cotton, tea, and tobacco had increased in volume and total value exported, while flower exports had also been introduced. Fisheries have overtaken coffee in overall export value, but only a small share of the population is engaged in that activity. In any case, all the export shares need to be considered in the context of Uganda's low overall export/GDP ratio, which ranged between 15 and 24 percent from 2006 to 2015 (World Bank 2017).
Manufacturing remains a small element of the economy, historically accounting for less than 10 percent of GDP, as indicated in table S4.1. Table S4.2 shows that the sector still employed less than 4 percent of the labor force as of 2002. The service sectors-including wholesale and retail trade, transport, communications, and construction-employed approximately 9 percent of the labor force, split fairly evenly between urban and rural areas, accounting for approximately one-third of GDP. Government accounted for the largest share of GDP, at approximately 20 percent, and 7 percent of the labor force. Most of the public labor force is situated in rural areas where the population lives, but approximately one-third is based in urban areas, especially the public administration hub of Kampala.

Transport Costs and Infrastructure
A key attribute of Uganda's economy is its limited infrastructure and high transport costs, which are among the highest in the world (Buys 2006). In multiple sectors, including food, these costs provide both implicit protection for domestic producers and implicit tax on exports (Milner, Morrissey, and Rudaheranwa 2000;Rudaheranwa 2006 (2004) and Alarcon et al. (2006).