The wheat supply chain in Ethiopia: Patterns, trends, and policy options
Wheat is one of the four most important food grains in Ethiopia. As a source of calories in the diet, wheat is second to maize. In terms of the area of production, wheat is fourth, after teff, maize, and sorghum. In terms of the value of production, it is 4th or 5th, after teff, enset, and maize, and approximately tied with sorghum.
Wheat production has expanded rapidly in the past decade. According to the CSA, wheat production has grown at 7.5% per year since 1995-96 and at 9.3% over the past decade.
The expansion of wheat area and higher yields have each contributed roughly equally to this growth. Since 1994-95, wheat yields have doubled, rising from 1.2 tonnes/hectare to 2.4 tonnes/hectare. This represents an average growth rate of 3.9%, slightly more than the rate of wheat area expansion over this period.
However, large discrepancies between wheat production and consumption estimates need to be resolved. One discrepancy is the estimated volume of wheat production and imports is 1.6 million tonnes larger than what can be accounted for by human consumption, seed, feed, industrial uses, and losses. Another issue is that official wheat yield estimates are 15-30% greater than yield estimates from other sources.
Wheat production occurs throughout the central highlands of Ethiopia, but is concentrated in a few zones. Just six zones account for more than half of Ethiopian wheat production: Arsi, Bale, West Arsi, East Gojjam, East Shewa, and South Wello.
Wheat farms are numerous, but most are quite small. The Agricultural Sample Survey (AgSS) estimates that there are 4.7 million wheat farms, but the average size is just 0.34 hectares.
Almost three-quarters of wheat area is fertilized. According to the AgSS, 73% of wheat area is fertilized, up from 54% ten years ago and more than other major cereals. The application rate has increased to 140 kg/fertilized hectare, though this is still less than recommended levels.
Less than 6% of the wheat area is planted with purchased, improved seed. Although it is not necessary to buy new wheat seed every year, this is a very low rate of replacement. Shortage of certified seed is a contributing factor.
One of the most common priorities identified by farmers is to increase the quantity and quality of improved wheat seed available. This is a particular source of concern given the threats posed by yellow rust and stem rust, to which many of the most widely used varieties are susceptible.
The use of mechanization and irrigation in wheat production is very rare. Less than 1% of wheat area in Ethiopia is irrigated, and less than 1% of the wheat plots are cultivated with tractors. Almost all wheat plots are ploughed using animal traction.
Total grain storage capacity in Ethiopia is estimated to be 29 million tonnes. The quality of storage ranges widely, from traditional 10-quintal goteras to large-scale warehouses operated by the EGTE.
The capacity of on-farm grain storage is almost 26 million tonnes, accounting for 89% of the total. Although on-farm storage facilities are small and very simple, there are more than 10 million on them. Surprisingly, farmers report very low (2-4%) storage losses.
A large majority of wheat farmers engage in on-farm storage. According to both the AgSS and the 2012 Baseline Survey, large majorities (80-90%) of wheat farmers had some wheat in storage at the time of the survey. Wheat is retained both for later consumption and for later sale. This contradicts the view that farmers are forced to sell their entire surplus at harvest to meet cash needs.
The role of cooperatives in wheat storage is negligible. According to the 2012 IFPRI-ATA Baseline Survey, just 0.5% of wheat sales went through a cooperative. Almost all wheat sales were directly to traders.
Grain wholesalers report very little long-term wheat storage. During the 2007-08 spike in grain prices, government officials accused traders of speculation and hoarding, and in a few cases stocks were confiscated. As a result, traders remain reluctant to engage in seasonal grain storage or reluctant to report it.
The share of wheat production that is marketed is in the range of 18-25%. The share varies from year to year, being higher after a good harvest.
Most of the marketed surplus is produced by a relatively small number of medium and large farmers. Household surveys suggest that 60% of wheat farmers do not sell any wheat. The top 20% of wheat sellers account for 60% of wheat sales. The marketed surplus ratio is higher for higher-income farmers, farmers with more land, and farmers in SNNP.
The share of wheat sold by growers varies significantly across different types of households. The marketed surplus ratio is positively related to farm size and ownership of equipment and livestock. It is negatively related to distance from roads and the nearest cooperative. And holding other factors constant, it tends to be greatest in SNNP and lowest in Tigray.
The lack of increase in the marketed surplus ratio is mostly (but not entirely) explained by rural population growth and growing demand for wheat by wheat growers. Based on these factors alone, wheat demand among wheat farmers should grow at 7.1% per year. Estimated annual growth in wheat production over the last 10 years is 9.3%.
Wheat surpluses are geographically concentrated. We estimate zone-level market surpluses using HICE consumption data and AgSS production data. Two-thirds of the zone-level surpluses come from just four zones: Bale, Arsi, West Arsi, and East Gojam.
Addis Ababa is represents a relatively small proportion of national wheat demand. Addis Ababa represents the largest deficit area, but accounts for just 12% of the zone-level deficits. Fafan (Somali) and Sidama (SNNP) are the second and third largest “demand sinks”, respectively.
Most of the wheat flows are rural-rural rather than rural-urban. A simplified least-cost analysis of how to distribute surpluses among deficit zones suggests that most of the inter-zonal flows are from one rural area to another rather than from rural areas to cities.
Small-scale mills account for an estimated 65% of the total milling capacity in Ethiopia. We estimate that there are roughly 29 thousand small-scale mills in Ethiopia with a total capacity of about 15 million tonnes. By contrast, there are 682 large-scale flour factories with a total capacity of 7.9 million tonnes.
Small-scale mills and large-scale flour factories cater to different markets. Small-scale mills serve wheat farmers who wish to mill their own wheat, rural households that buy wheat, and a good number of urban households. Large-scale flour mills purchase imported wheat from EGTE and domestic wheat from traders in order to sell to bakeries and urban wholesalers and retailers.
Ethiopian imports about one million tonnes of wheat each year. This includes food aid and commercial imports by the EGTE, the proportions of each vary widely from year to year.
Food aid, much of which is in the form of wheat, is used for emergency relief and the Productive Safety Net Programme (PSNP). Emergency relief stocks are maintained by the Emergency Food Security Reserve Administration, which can be used by national and international agencies for assistance programs provided they replace the grain later. The PSNP distributes both food and cash to targeted households in selected woredas on a sustained basis.
Commercial wheat imports consist almost entirely of EGTE imports to supply subsidized wheat to selected mills. The EGTE purchases wheat on the international market and distributes it to 205 mills in different cities at subsidized prices. The mills are required to sell the flour to designated bakeries at a controlled price, and the bakeries are required to sell bread at a fixed, below-market price.
Most large-scale mills do not receive subsidized wheat from the EGTE. Data from the Bureau of Investment indicate that there are 682 large-scale flour mills, of which 205 receive EGTE wheat. The other large-scale mills rely entirely on locally-produced wheat.
Wheat plays an important role in the Ethiopian diet. Wheat and wheat products account for 14% of the caloric intake in Ethiopia, making it the second-most important food item after maize.
Wheat consumption varies significantly across types of households. Per capita wheat consumption is greater in cities than rural areas, greater in Tigray than the other main regions, and greater among high-income households than low-income households. In fact, the richest quintile of Ethiopians consumes three times as much on a per capita basis than the poorest quintile.
The share of the budget allocated to wheat and wheat products rises with income among poor households, but begins to fall at middle and higher income levels. This inverted U-shape is also found in other countries. It implies that as Ethiopia gets richer, it will eventually diversify its diet away from wheat and other grains.
At the average income in Ethiopia, the income elasticity of demand for wheat is 0.83. Taking population growth and income growth into account, this suggests that total wheat demand will grow at 7-8% per year. The income elasticity of pasta is higher, suggesting demand growth of more than 10% per year.
In 2014, the EGTE subsidies of imported wheat are equivalent to 32% of the import parity price in Addis. According to EGTE data, the cost of imported wheat delivered to Addis was US$ 413/tonne, and the price at which EGTE sold the wheat to millers was US$ 280/tonne.
Without the subsidy, commercial wheat imports would not be viable in 2014 and wheat prices would have been 22% higher. Making plausible assumptions about the price elasticity of supply and demand, we estimate that domestic production and consumption would equilibrate at a price of US$ 340/tonne.
The costs of the wheat import subsidy to government and farmers is eight times greater than the benefits that accrue to consumers. The fiscal cost of the subsidy is about US$ 66 million/year, the costs to farmers in the form of lower prices is almost US$ 7 million/year, and the benefits to consumers is less than US$ 9 million/year. One reason for the low benefit-cost ratio is that the subsidy costs US$ 133/tonne (the difference between the import parity price and the subsidized price), but it only reduces the domestic price by US$ 62/tonne (the difference the no-subsidy market price and the current price).
The tool is one of a series of commodity studies to be carried out by the International Food Policy Research Institute (IFPRI) at the request of the Ethiopian Agricultural Transformation Agency (ATA). The general goal of these studies is to provide a comprehensive description of the commodity marketing channel from production to consumption in order to assist the ATA in its work of raising crop productivity and improving market efficiency. The terms of reference, jointly prepared by IFPRI and ATA, identifies nine questions to be addressed with these commodity channel tool: 1) What proportion of production is marketed by type of farm and by location? 2) What are the major routes to market, from surplus producers to consumers? 3) What is the volume and value of these channels and how does it vary by season? 4) What are the margins for smallholders and other value chain actors? 5) What is the status of the market infrastructure in terms of storage, processing, wholesaling, and retailing? 6) Who are the major market actors in the marketing of the commodity? 7) What are the main challenges in increasing marketable surpluses as well as expanding the market infrastructure to handle larger volumes? 8) What are the main challenges to achieving competitive markets and becoming competitive on international markets?
The tool uses a variety of methods to analyze the data collected, including econometrics, linear programming, spatial analysis, and cost-benefit analysis. More specifically, the analysis of the spatial patterns of production and marketed surplus makes use of geographic information systems (GIS) to map the results. Econometric analysis is used to evaluate the determinants of the marketed surplus of wheat and the factors influencing demand for wheat and wheat products. In addition, we use linear programming to approximate the flows from surplus wheat zones to deficit zones.
Photo credit: UN