EC policy endorses local and regional procurement of food aid commodities (LRP), a practice that is believed to assist in the development of local agriculture and livelihoods in supplying countries. The research hypothesis for this study was that such procurement of food aid can make a much larger contribution to the economies of developing countries, and poor people in particular, and that policies can be put in place to increase such benefits.
Available literature was reviewed, and case studies were undertaken in Ethiopia and Uganda where combined LRP started in the 1990s and in recent years has been running at over 300,000 tonnes per annum. Available information permits certain conclusions about the impact of LRP, as follows:
- It provides much greater net benefits for rural and urban populations than equivalent expenditure on tied food aid (provided it is competently managed, and the local/regional supply base allows the activity to be developed without major adverse impacts on consumers – where this is not the case imported in-kind food aid from developed countries may sometimes be more appropriate).
- While no systematic assessment has been made of the overall impact of LRP on price stability, several cases can be cited where it has increased price instability. This does not call into question the value of LRP, but points to the need for more flexible funding arrangements, further developing market analysis capabilities, and untying food aid so that LRP, in-kind food and cash support may be combined in a way which has a benign effect on price movements. WFP’s Advance Financing Facility is a positive step in making funding arrangements more flexible.
- The evidence for impact on market efficiency is mixed. Local procurement has led to investment and improved practices among traders supplying food aid, but has not greatly impacted the conduct of the regular grain trade in source countries, or on the quality of grain in that trade. In Uganda, much of the LRP is logistically inefficient, and it has promoted investments in drying and storage plant that are excessively concentrated in Kampala, and poorly located to cope with a scenario of declining food aid quantities.
- LRP is having some positive impact on the development of Ethiopia’s export trade, but in Uganda, LRP has not been organised so as to equip traders – in terms of investments, financing and quality management – to develop an export market that would cushion them in the event of a major reduction in WFP purchases.
- In both countries, food aid grain sales are highly concentrated among a few suppliers. However there is no conclusive evidence that this market concentration is a source of market inefficiency.
- Notwithstanding the strong performance of certain groups, LRP aimed exclusively at producer organisations has proved an inefficient use of resources.
- Local and regional procurement have led to the development of industries for the manufacture of blended food commodities, and diversification into the production of soybeans, notably in Ethiopia.
Food aid agencies can adjust tendering procedures to mitigate certain problems, but in countries where they are involved for the medium or long term, they should consider more comprehensive approaches to market development which leave behind stronger and more efficient structures that will serve the host countries better. To this end, they should work closely with partners seeking to improve the performance of grain markets in areas such as contracting, warehouse receipt systems and exchange trading.
The findings confirm the research hypothesis and the following specific recommendations are made to those concerned with the distribution of food aid:to devote more resources to assessing impact of local purchase on supplying areas; to improve the information base for decision-making; for those still supplying in-kind food aid to untie a large proportion; to take all steps to facilitate timely and effective decisions by local food aid managers; to investigate the scope for ‘food aid pipelines’ in countries where food reserves still exist and function efficiently; to focus on developing existing marketing systems rather than exclusive relationships with producer organisations, and; where food aid is expected to continue at significant levels for several years, to work proactively with partner organisations to establish new market institutions with a view to improving the performance of local and regional markets.
EC Council Regulation No. 1292/96 dealing with food aid policy and food aid management and special operations in support of food security endorses the growing practice of food aid procurement within the benefiting country or from a neighbouring country, otherwise known as local and regional procurement (LRP). This practice is widely believed to assist in the development of local agriculture and livelihoods in the source countries. Moreover, LRP is commonly believed to result in multiple efficiencies in the provision of food aid vis à vis extra-continental procurement, due to:
- lower procurement costs, with large savings in transport and handling
- reduced delivery time and more timely emergency response
- provision of types of food preferred by the recipients, and
- allowing the staff of food aid agencies to procure more flexibly according to cost and need
There is a strong preference for LRP in the policies of many donors, particularly the EU and member states. However, the world’s largest donor of food aid (USA) does not support LRP, and focuses on providing in-kind food of American origin. In each of the last three yeas, the Bush Administration has proposed that one quarter of the food aid budget ($300m of $1. 2b) be made available for LRP, along the lines of Canada which has already untied 50%. Congress has not taken specific action on this, which means it has rejected the proposal by default. No LRP is provided for in the 2007 Farm Bill though it is possible that the Senate will include a small ($25m/year) pilot project (pers. comm., OXFAM America).
At the time that this study was commissioned, there had been little systematic and critical review of the contribution that locally procured food aid was making, directly and indirectly, as a development tool in the source countries. The study was financed by the DFID EC-PREP programme[ii], and was designed to test the hypothesis that LRP could make a much larger contribution to the economies of developing countries, and poor people in particular, and that polices and practice guidelines could be put in place to increase such benefits. It was also expected that the outputs of this research would inform and guide the EC and member states in maximising development impact through their implementation of LRP.
The research was not designed to address the case for or against the alternative of putting purchasing power in the hands of food insecure households by distributing cash or vouchers. We started from the premise that while this approach was likely to gain ground, physical distribution of food was likely to remain the dominant of source support for many years to come.
A review of available literature and secondary sources showed a general lack of information, few published papers, and little grey literature regarding developmental impact. However, we have been able to draw upon other studies which appeared after we completed our deskwork, including studies by OECD (Clay et al. , 2005) and Tschirley (2007), and reviews of local procurement in different countries, including Bolivia, Burkina Faso, Ethiopia, Nepal and Uganda. The latter were commissioned by the World Food Programme (WFP), and funded by the Governments of Belgium and Sweden.
Our preliminary work indicated that the focus of this study should be on grain, and that case studies in Ethiopia and Uganda would reflect contrasting situations. We therefore carried out two country case studies, each involving two week visits by a post-harvest technologist and an economist to Uganda (October 2004) and Ethiopia (January 2005), with a view to collecting and analysing locally available information. They identified and interviewed the main participants in local and regional procurement, and food aid in general, and prepared country study reports, quoted respectively as Wandschneider and Hodges (2005), and Walker and Wandschneider (2005).
During the second half of 2005, preliminary findings of this study were presented and discussed at a USAID ‘roundtable’ in Washington[iii] DC, at the First Regional Grain Summit[iv] in Nairobi, Kenya, at the head office of the World Food Programme, Rome, and at EC DG-Development, Brussels. Since April 2006, one of the authors has been stationed in Uganda and has been able to update certain information for that country.
Food aid procurement in Uganda
Food aid is required mainly to support internally displaced people (IDPs) sheltering from the armed insurgency movement (Lord’s Resistance Army) in the north of the country, and refugees and IDPs from conflicts in neighbouring countries. WFP is the main agency responsible for LRP. It started procuring in 1991 in support of feeding programmes in Burundi, Democratic Republic of Congo, Rwanda, Tanzania and Uganda. Annual procurement increased from 28,000 tonnes in 2000 to a peak of 171,000 tonnes in 2005, and then fell slightly to 161,000 tonnes in 2006. In the latter year 79% went to to IDPs in Uganda and 21% to neighbouring countries, mainly Burundi. Maize grain, beans, corn-soya blend (CSB) and maize meal have been the focus of local procurement in that order; 85% of purchases by weight are maize or maize based products.
The level of procurement was exceptionally high in the first half of 2007, and on 5 August, WFP reported it had bought 151,000 tonnes of maize and beans since the beginning of the year, of which 91,000 tonnes were for consumption in Uganda and the balance in the Region. Quantities procured are now falling sharply, due to IDPs returning home and, according to latest reports from WFP, a shortage of funds with which to carry out LRP. Procurement will be lower in future years, though they are likely to remain at substantial levels, due to continued political instability and natural calamities in parts of the Region, and delays in reaching a settlement with the Lord’s Resistance Army. Uganda has already proved itself highly convenient source of supply from which to meet these needs.
LRP has had a major impact on the development of maize production in Uganda where, unlike neighbouring Kenya and Tanzania, maize is not a leading staple food. Annual production has grown rapidly over the last 15 years to somewhere between 600,000 and 900,000 tonnes[v]. It is grown largely though not exclusively as a cash crop, and there are three main trade outlets:
- domestic hammer milling into ‘posho’ meal, a large percentage of which is used by both households and institutional outlets, especially schools, hospitals and prisons. The Chairman of the Kisenyi Maize Traders Association estimates that about 110,000 tonnes of maize pass through the market annually.
- exports, mainly to Kenya, but also to Rwanda, Burundi and Sudan. The quantity moving through the main border point to Kenya averaged 121,000 tonnes per annum from 2004-2006[vi].
- Purchases by WFP, which in 2005 and 2006 approached 150,000 tonnes per annum, in terms of the quantity of grain supplied directly and grain required to produce maize-based products. This represents between 17% and 25% of total estimated production of maize. Allowing for local consumption in producing areas, it represents a larger proportion of marketed production.
Private trade flows are illustrated in Figure 1, below, while food aid flows are shown later in the paper in Figure 2.
FIGURE 1: PRINCIPAL TRADE FLOWS IN MAIZE FOR PRIVATE USE
Most of Uganda’s maize trade, including exports, relies on informal trade intermediaries, and does not make use of mechanical drying and cleaning equipment. WFP’s intervention is of major significance not only in terms of the quantity procured but also because it has virtually ‘created a market’ for dried and graded maize. Prior to the arrival of WFP, Uganda had very limited mechanical drying and cleaning capacity, mainly former parastatal facilities privatised in the early 90s. However WFP’s procurement activities have induced traders to make considerable investment in high-capacity plants for drying and cleaning[vii] and stores, almost entirely in Kampala, with a view to supplying WFP warehouse in that city, and as a convenient collection point for supplying WFP warehouses in Burundi and Rwanda. In Uganda, it is evident that LRP has been the key determinant of an emerging trade structure and longer-term investment in supporting infrastructure.
WFP quality standards for LRP of grain and grain products are high in comparison with the quality marketed domestically, and in early 2006 it aligned itself with the demanding East African Community Maize Standard Grade 1, which requires 13. 5% moisture content and maximum total defects of 4%. Managing the quality of maize grain is somewhat problematic in Uganda where, due to the bimodal rainfall pattern, it is difficult to dry the crop in the field. Farmers do not have access to mechanical drying and cleaning facilities, and only a small minority have drying cribs. Some use drying mats, but they generally dry maize on the ground where it gets contaminated with sand. Drying is slow and difficult and farmers sell much of their grain to primary traders at high moisture levels (e. g. 18-20%), providing a fertile substrate for the development of moulds. Traders assemble lots out of smaller parcels of heterogeneous quality and deliver these to Kampala-based buyers who can put it through their dryers. They typically hold the grain for a week or more after purchase from the farmer, and during this period there is often considerable deterioration. Kampala-based traders sometimes have difficulty meeting WFP’s quality specifications, and in wetter harvests, WFP has found it must relax its standards in order to guarantee supply of the quantities of grain it requires.
About 95% of food aid procured locally is purchased from regular trading and processing companies. In 2006, WFP purchased maize and beans from 26 companies, but 74% of purchases were concentrated in four leading suppliers. Companies winning tenders can use their contract to secure bank financing to accumulate the necessary stocks to supply WFP. In a country where banks are rarely involved in financing the regular grain trade, this places WFP’s suppliers in a privileged position.
From 2000 onwards, WFP has also procured maize and beans directly from producer organisations (POs). Standard procurement guidelines applied to regular commercial procurement were significantly relaxed to encourage farmer participation, as follows:
- POs could delivery smaller quantities (50 tonnes minimum for maize, compared to 200 tonnes for commercial suppliers),
- POs are allowed to deliver in local warehouses rather than bear the cost of shipment to WFP’s warehouses. This is a major advantage given that WFP requires its commercial suppliers to deliver either to its central warehouse Kampala, or to locations in Burundi and Rwanda
- Delivery periods are extended
- POs are exempted from the requirement to provide bid or performance bonds
- The requirement to provide grain in marked bags is often waived after clearance from the relevant donor.
Targeted groups also received training in grain handling from WFP and partner organisations. However despite these advantages Wandschneider and Hodges (2005) found that POs were finding it difficult to handle the formal tendering procedures, to understand contracts and quality specifications, and to access sufficient quantity of grain and credit to allow them to hold the stock up to payment by WFP.
WFP’s objective was to fulfil a commitment it had made to the Government of Uganda, i. e. to purchase up to 10% of locally procured maize and beans from producer organisations (POs). However, the maximum it has achieved to date is 7,450 tonnes, representing 5. 3% of total purchases, in 2006[viii]. Supplies were heavily concentrated in a very small number of POs, and six accounted for about 80% of deliveries. This reinforces the findings of Wandschneider and Hodges (2005) that the maximum number of POs able to supply WFP in a particular year was 11, and only 5 had been involved in the activity more than once. The level of contractual default has been high, an average of 40% of quantities contracted between 2004 and 2006 (source: WFP).
Notwithstanding this, many POs supply WFP indirectly, i.e., they supply traders who themselves deliver to WFP; some have learnt to dry and clean the grain to WFP’s specification, and thereby earn remunerative premiums with the traders[ix]. These experiences suggest that the most important precondition to farmers benefiting from WFP procurement is for them to be effectively organised at the grass roots so that they can deal effectively with buyers, whether they sell directly to WFP or through a trader.
Food aid procurement in Ethiopia
Ethiopia is the largest regular recipient of food aid in the world. The quantity has tended to increase, and the annual average for 2000 to 2004 exceeded 900,000 tonnes, mainly cereals. Local procurement started with the Relief Society of Tigray (REST) in 1983, but became a major activity in many other areas of Ethiopia in 1996, when the EC began active support and funding. Since then, around a quarter of Ethiopia’s food aid has been procured locally in the form of maize, wheat and sorghum, and on average this is equivalent to about 12% of Ethiopia’s marketed surplus for these crops.
The main procurement agencies are the Ethiopian Government’s Disaster Preparedness and Prevention Commission (DPPC), WFP and EuronAid[x], all of which use tendering procedures, and purchase in lots of not normally less than 500 tonnes. Despite attempts to target farmer associations, most purchases are made from large traders specialising in food aid supply.
Annual cereal availability studies, following the model developed by REST, are designed to protect the market from adverse effects. However Walker and Wandschneider (2005) found that the quality of statistics for local and regional procurement of food aid was poor, there was disagreement between the different information sources and only limited co-ordination between the food aid agencies involved in local purchase. Additionally, there were various procedural sources of delay: donors’ slow and centralised procurement procedures (WFP require tenders over a certain size to be approved in Rome, EuronAid in Brussels), and the timing of cereal availability estimates (the survey was carried out after the harvest and estimates are not published until March).
These delays were frequently countered by the availability of food aid reserve stocks held and managed by the Ethiopian Food Reserve Administration. This is a unique ‘food aid pipeline’, which allows donors to draw down stocks for distribution and then procure locally or regionally to replenish the stocks. Thus, whilst the timing of distribution is normally appropriate, the timing of procurement tends to be later in the season than would be expected from the perspective of commodity availability and price.
Donors and Government are working on the implementation of a ‘Productive Safety Net Programme’, which involves a shift from food to cash distribution. However, various factors are likely to limit implementation, notably difficulties in ensuring security. Hence, food aid, including local procurement, will remain very important to Ethiopia in the foreseeable future.
Impact of local procurement
Given the general lack of prior impact studies, we mainly focused on interviewing key informants. However, in Ethiopia, there had been a review of the EC’s 1996 local purchase experience (Amha et al. , 1997), while the German agency GTZ, and REST, had carried out significant studies of impact in some supplying areas. The latter provided limited guidance re overall impact, given the small scale and atypical nature of their respective programmes.
Based on the fieldwork and literature mentioned, it is possible to make the following statements concerning impact.
Impact of local purchase vis à vis tied aid
One would expect LRP to have a beneficial development impact vis-à-vis a similar quantity of tied aid, since it makes a much larger net contribution to rural and urban livelihoods in the countries concerned. However, this assertion only holds in circumstances where local purchase is well managed and appropriate, i. e. there is an adequate local/regional supply base, so that it does not cause a price rise of such a proportion as to cause severe adverse impact on the food security of vulnerable populations. In general terms, the resource cost of LRP should be no more than the cost of importing in-kind food.
Clay et al. (2005) showed that the prospective benefit of untying US food aid supplied to Africa was likely to be very large; they estimated unit cost of the two main commodities maize and corn-soya blend (CSB), was only 61% and 52% respectively of in-kind food aid. The lower resource cost of LRP implied it was economically efficient in terms of supplies from the USA – indeed difference in cost is so large that it would still have been efficient if the US Government had waived the requirement to ship commodities in US flag vessels, at an estimated additional US$ 40 per tonne. A more recent study by Tschirley (2007) reinforces this conclusion. Comparing the costs of LRP in Kenya, Uganda and Zambia between 2001 and 2005 to estimated costs of in-kind food aid from the USA, he found that by using LRP, USA could provide 75% more food to beneficiaries. Not surprisingly LRP was most efficient in the more land locked of these countries, Zambia and Uganda.
Tschirley (2007) also comments favourably on the efficiency of WFP’s LRP activities in Africa, showing that the agency had consistently paid competitive prices in Zambia, and had done so in Uganda since late 2004, while paying about a 10% premium in Kenya. However, he did not consider the overall logistical efficiency of procurement, a subject we take up later in the case of Uganda.
LRP is normally much faster, assisting in timely response to food emergencies. WFP staff report this to be the case in Uganda, and that there are a lot of hiccups with imported food, resulting from red tape during transit through Kenya. However in many cases, aid bureaucracy nullifies this advantage of LRP[xi]. Such observations provide a case for improving the management of local purchase, rather than a justification for maintaining tied aid.
Notwithstanding the benefits of local purchase, 75-80% of annual food aid used in Ethiopia was tied (Walker and Wandschneider, 2005, p. 26). This constrains food aid managers in their ability to vary the proportions of local purchase and imported food aid in relation to the state of local harvests and markets. The problem was particularly evident in 2001 and 2002, when the country experienced an unusually large price fall for all its major cereals while simultaneously receiving large quantities of US imported grain. The tying of American aid made it difficult for the US in dealing cost-effectively with the famine which occurred in 2003[xii]. In Uganda, tied aid averaged 38% of the sum of locally procured and tied food aid to the country through WFP between 2004 and 2006 (based on information from WFP).
Impact of LRP on price stability
LRP is not explicitly designed to achieve price stability. However, in countries characterised by large price fluctuations it is desirable to make a net positive contribution to price stability, as this tends to work in favour of the livelihood of surplus producing farmers and of consumers. By varying quantities procured locally according to the size of the local crop and the timing of its availability, it should be possible to manage LRP so that it makes such a contribution.
No systematic assessment has been made of the overall impact of LRP on price stability in source countries. However, we may cite some cases where the timing and quantity of LRP appear to have increased price instability.
As regards Ethiopia, LRP was highest in 2003, the year when the estimated market surplus was lowest (Walker and Wandschneider, 2005, p. 26). Food aid donors seem to have reacted in a delayed and somewhat untimely fashion to very low and unremunerative cereal prices in 2001 and 2002. Gautam (forthcoming paper) finds a considerable degree of short-term (intra-annual) price volatility in Ethiopia, and it is likely that the pattern of local purchase is contributing to this. It peaks in the lean season (July), whereas it should logically peak earlier in the year, after the main harvest. Agridev Consult (2005) found that this benefits traders at the expense of farmers, who must sell their produce at poor prices to meet cash needs, and that it may have impacted adversely on food insecure consumers in the lean season. As noted above, centralised procurement procedures and the timing of the publication of the annual cereal availability estimates seem to have caused delayed procurement. More recently, Tschirley (2007) mentions protracted donor procurement which “may have contributed to disturbing the food security situation during 2006”.
In Uganda, WFP purchase is one of two factors, along with variability in Kenyan demand, which have a major impact on seasonal price patterns and contribute to price variability both between years and within years. There was a major maize price crash in 2001, resulting from the combined effect of a bumper crop in Kenya (which closed its borders to Ugandan maize) coupled with minimal WFP purchases for much of the year. By contrast, heavy WFP intervention in 2003 caused a major rise in Kampala prices in the April-August period, and maize prices peaked at about double those in Iganga, 119 km distant in Eastern Uganda (source: Wandschneider and Hodges, 2005). However, maize is not a leading staple food in Uganda, so this is unlikely to have affected the food security of nutritionally vulnerable populations.
To enhance the impact of LRP on price stability, food aid agencies need to have more cash available at times when they can procure most advantageously. WFP is making progress in this direction through the creation of an Advance Financing Facility, a sort of internal banking arrangement within WFP that allows it to borrow against donor commitments. The Head of Finance at the WFP Uganda Country Office, who describes it as “very effective”. Tschirley (2007) quotes WFP personnel saying the facility was working, though insufficient time had passed to systematically assess its impact on procurement cost. The same author (page 37) moreover recommends food aid agencies further develop their data bases and analysis capabilities and their understanding of commodity chains so as to better forecast the impact of LRP on prices and consumer welfare. Donors can further contribute towards price stability by delegating more authority to country offices and further untying food aid.
Impact of LRP on the efficiency of local grain markets
We consider efficiency from the standpoint of impact on: (a) logistical efficiency; (b) trade practices; (c) food quality, and; (d) market concentration.
Impact on logistical efficiency. Earlier in this paper we described how WFP procurement system had caused its commercial suppliers to concentrate investment in Kampala, far from the producing areas. The consequence is that much of the grain procured by WFP follows a logistically inefficient route, as shown in Figure 2. Much of it flows southwards towards Kampala, where it is first dried and cleaned, and then delivered to WFP’s warehouse, and after this shipped north again to ‘extended delivery points’ close to the refugee camps. This system is very costly to the donors who fund WFP[xiii]. It also leaves Uganda with infrastructure which is not well located for the post-WFP scenario, where there is likely to be continuing demand for quality maize. For example there is no drying and cleaning plant in Iganga, a major producing district in Eastern Uganda that is ideally placed to supply the Kenyan market as well as local markets in Uganda. We discuss the impact of LRP on exports further below.
FIGURE 2: PRINCIPAL FLOWS OF GRAIN FOR FEEDING IDPS & REFUGEES
Impact on trade practices. In both countries, LRP has encouraged significant investment on the part of traders supplying food aid, and has improved practices within the food aid supply channel. It has produced a cadre of relatively large companies whose main, and sometimes exclusive business, is to supply food aid agencies to order. They are reported to have improved their business skills and understanding of the grain trade through the food aid business.
However, it was difficult to find evidence that LRP was having impact on the performance of regular grain marketing channels, with regard to grain handling and trading practices. Companies supplying WFP were found to be almost exclusively concerned with the food aid market.
In Uganda, traders operating in Kisenyi market or selling to Kenya are in the habit of receiving and milling maize of variable moisture content (a Kisenyi miller described how the action of milling served to dry the grain!). Companies supplying WFP occasionally use Kisenyi to offload dried maize surplus to WFP requirements, but the working relationship is not very close. WFP suppliers sometimes mention improvement in handling practices among their own suppliers, which may have knock-on effects on quality of maize entering regular market channels, but we were unable to confirm this.
In Ethiopia, GTZ and REST claimed a gain in the number, size, and efficiency of grain traders in the locality of their procurement activities. These organisations now account for only a small percentage of total quantities procured, and their approach to procurement involves many field staff, and could be difficult to upscale. Notwithstanding this, aspects of their operation may be relevant to the larger food procurement agencies.
One explanation for the lack of impact of LRP on general trade practices is that food aid suppliers have limited stock positions at the time of bidding, and mainly procure grain after winning the tender[xiv]. In the absence of a contract with WFP, they find it difficult to access bank finance, and they procure overwhelmingly to meet the needs of this one client. Consequently, they have little stock readily available for sale to non-food aid customers who require quality grain. In the case of Uganda, large scale Kenyan millers complain of their inability to procure quality grain at times when Kenyan supplies are short (pers. comm. , Sophie Walker, Kenagri Ltd. , Nairobi).
In Ethiopia by contrast, traders must have stock positions if they are to participate in tenders (as much as 75% in some cases), but it is likely that stocks are mainly procured after the tender is called and are not readily available for sale to regular trade customers. Requiring companies to bid from stock positions has the effect of smoothing the impact of local purchase and reducing performance risk, i. e. the risk that tendering companies will be unable to deliver on their contracts. However, as Agridev Consult (2005) and Sserunkuuma & Associates Consult (2005) point out, it tends to favour the stronger players and thereby reduces competition between suppliers.
Impact on food quality. The enforcement of strict quality procurement standards by food aid agencies has undoubtedly been of benefit to IDPs and other food aid recipients, but as indicated above does not appear to have greatly impacted the quality of grain traded in regular wholesale/retail channels.
Poor grain quality is a problem for consumers, both in Uganda and neighbouring countries. Kenya closed the borders for Ugandan maize in 2004, after 125 people died from suspected aflatoxin poisoning, which the Kenyan authorities blamed on Ugandan grain informally traded across the border. In reality however, the deaths were concentrated in districts to the east of Nairobi, and are unlikely to have been caused by Ugandan grain. Notwithstanding this, data for Uganda indicates that a significant proportion of maize passing through regular trade channels contains levels of aflatoxin well above the legal limit of 10 parts per billion (ppb). However, this does not appear to be the case with grain traders supplied to WFP. Analysis of 1,000 grain samples tested during the years 2003-2006 shows that in all years aflatoxin was either not detected or the average concentration was below 5ppb. 1% of samples exceeded 5 ppb. One can probably attribute this to the limited time between harvesting and drying – long enough to allow the development of moulds but not long for aflatoxin to reach dangerous levels.
Impact on market concentration. Agridev Consult (2005) and Amha et al. (1996), in the case of Ethiopia, and Sserunkuuma & Associates Consult (2005) have raised concerns that the high concentration of suppliers can lead to price manipulation. Between 2001 and 2004, the average shares for the three largest suppliers, in the two countries, were as shown in Table 1:
TABLE 1: CONCENTRATION OF PROCUREMENT AMONG SUPPLIERS, 2001 TO 2004
Sserunkuuma & Associates Consult (2005) quotes verbal evidence that leading suppliers in Uganda are developing monopolistic price-fixing powers. In 2006, the four largest suppliers accounted for 69% of total commodities delivered, but the proportion was higher for maize (78%) and lower for beans (61%). Data up to 24 May 2007 show a marked concentration in procurement. The top four companies supplied 80% of the maize and 88% of the beans, the leading company accounted for 45% and 55% respectively. This high concentration is not indicative of any decrease in market efficiency, per se, simply a position which might allow the leading supplier to exercise market power in its own favour. Morover, it has to be recognised that there are five suppliers with substantial drying capacity, and the position of the leading supplier can be easily be contested, particularly at a time of declining WFP procurement.
There is no conclusive evidence that this market concentration is resulting in excessive trading margins, and is a source of market inefficiency. Analysing the LRP performance of WFP in Zambia, Kenya and Uganda, Tschirley (2007, p. 28) finds a mixed picture: in Zambia it had consistently paid competitive prices, in Uganda it had done so since late 2004, and in Kenya it appeared to pay a premium of 10%. Other aspects of LRP, notably the round-about logistics inherent in the procurement system, and the lack of impact on trade practices and food quality in the regular trade should be of more immediate concern.
Impact on market access by producer organisations (POs)
Partly to avoid the effects of market concentration, most agencies have tried to include POs within their tendering systems, but with very limited success. There is now widespread consensus among seasoned practitioners that very small tenders (e. g. 50 to 200 tonnes) aimed at POs are costly to organise and an inefficient use of resources. Wandschneider and Hodges (2005) describe WFP’s Uganda scheme as resulting in a highly subsidised and artificial marketing channel for a few groups that were heavily dependent on the food aid market.
Similar findings emerged from the WFP-commissioned study of LRP in Burkina Faso (Institut du Sahel and CILSS, 2006), where producer organisations’ share of sales to WFP were found not to exceed 10%. The study attributes this in part to the restrictive tender requirements but concludes that even if procedures were more flexible, POs would be remain less competitive than traders in terms of price and compliance with quality standards. Nevertheless there were some professional POs upon which food aid agencies could rely.
A policy paper prepared by WFP HQ endorsed this view stating that “the experience has been generally negative resulting in higher prices paid, higher administrative costs because of having more contracts to monitor, and greater risk of default” and expressing concern that “support for farmers’ groups could lead to investments by these groups that they cannot maintain when WFP stops buying” (WFP, 2006). However this finding has not led to any change of approach in Uganda. Given the difficulties POs experience with paperwork and contractual requirements, one experienced WFP staff member advocates eliminating tender processes when dealing with POs and simply setting a market-related price at which they may deliver the grain (pers. comm. , Jack Keulemans, WFP).
Impact on the development of the export trade
Due to its favourable climate, Uganda is able to produce maize and beans competitively for export to Kenya and elsewhere. However, WFP’s high level of purchases of maize and products containing maize (maize meal and CSB) in Uganda have diverted some grain production away from exports. According to authors’ estimates, one needs a margin of at least $60 per tonne to provide a consistent incentive to export to Nairobi, but the Nairobi-Kampala price differential only exceeded this figure in 10 out of 36 months up to mid 2007[xv] This can be attributed to the high level of Kampala-centred LRP.
Trade diversion is not a problem per se. If Uganda can produce maize competitively and to acceptable standards, it can quickly recover foregone export markets when WFP reduces its purchases. However, this raises a question as to whether WFP has taken steps within its power to prepare Uganda for this transition. Currently Ugandan exports are mainly in informal sector, they are heavily discounted on quality grounds, and are vulnerable to the application of the East African Maize Standard if and when Kenya decides to apply it strictly. A recent COMESA meeting has recommended harmonised standards that, if gazetted in Kenya, would increase Uganda’s vulnerability in this regard. If WFP had promoted a more decentralised grain drying infrastructure it could have improved grain quality and rendered Ugandan products more food safe in domestic and regional markets, and reduce the risk that its entry into Kenya would be barred on quality grounds.
Due to geographical and logistical factors, the impact of local procurement on regional trade is less of an issue in Ethiopia than in Uganda. However, it is undoubtedly assisting in the development of trade in blended foods and sorghum between Ethiopia and Sudan. The EU/EuronAid initiated cross border movement of over 24,000 tonnes of sorghum from Sudan to Ethiopia in 2003 that had additional development impact by stimulating significant commercial trade over the same route (Smalbruch and Walker, 2004). This transaction was facilitated by use of the grain reserve stocks in each country. There is scope for increasing such operations, but this will require a determined effort to overcome administrative and trade barriers, and the compartmentalisation within food aid agencies themselves.
Development of agro-industries and crop diversification
Local and regional procurement in Ethiopia has led to the manufacturing of blended food aid commodities (corn-soya blends) such as Faffa, Famix and Unimix. The four companies involved have diversified from baby food production to be largely suppliers to the food aid sector; for one of them the business was worth around US$3 million per annum. These companies employ many hundreds of employees, albeit some casual due to the variable nature of food aid demand, and create further employment in industries supplying transport, packaging, printing and other supporting services (Walker and Wandschneider, 2005). There is also a small blended food aid manufacturing sector in Uganda.
Soya is not a common crop in Ethiopia and whilst there is a limited market for soya as a constituent of poultry feed and as an export commodity it appears that its main use is as a constituent of blended food-aid products. Ethiopian blended food manufacturers believe that their supply of the local and regional procurement markets has initiated and maintained significant increases in national soya bean cultivation.
Alternative approaches to procurement
In countries like Ethiopia and Uganda where they are active in LRP over many years, food aid agencies have considerable market power which they can use to fashion their own supply channels, and contribute materially to the overall efficiency of agricultural markets. The case for this is particularly strong in Africa where the transition from single-channel to free-market systems has encountered many difficulties (see for example Jayne, 2005). While local and regional procurement of food aid cannot, by itself, make good all the flaws in existing agricultural marketing sectors, it can provide a clear lead. Notably, it can:
- help formal sector traders develop stock positions from which they can supply both private and food aid customers;
- help them improve grain quality and make it easier for them to sell the grain to formal sector buyers both domestically and regionally, and earn quality premiums accordingly;
- encourage the trade to invest in warehouses, grain handling and drying equipment in producing areas
- encourage producers to bulk up their product and sell further down the marketing chain, and;
- promote greater competition within food aid marketing channels.
In order to do this, food aid agencies need to work closely with other parties seeking to improve the performance of grain markets, with a view to identifying and implementing the most promising options. Among development practitioners and some trade bodies, there is considerable interest in the development of new market institutions, including contracting and dispute settlement, warehouse receipt systems and commodity exchanges. These were a major focus for participants in two Regional Grain Summits organised in 2005 and 2007 with support of USAID and COMESA[xvi], and which led to the organisation of a regional trade body, the Eastern African Grain Council (EAGC)[xvii]. In the following paragraphs we discuss how food aid agencies might make use of two of these institutional innovations: warehouse receipt systems and commodity exchanges.
Warehouse receipt systems. With such systems, Governmental or non-Governmental agencies license (or certify) private companies (including traders, millers, cooperatives and inspection companies) to hold stock on behalf of others. Upon receipt into storage, the stock has to be standardised in terms of grade, and this encourages sellers to improve the quality of produce and earn grade premiums accordingly. The system also allows food aid agencies to establish multiple storage locations closer to the producers, and from which they can lift grain for delivery to the ultimate beneficiaries. One of the main reasons advanced by WFP for requiring delivery to its own warehouses is lack of trust in third party warehouse operators – the purpose of the licensing agency is to address this problem, i.e. to build up trust in such warehouses, rendering them more reliable and bankable
The system also allows food aid agencies to procure from existing stock positions, rather than the present tendering system which exposes them to traders defaulting or seeking extensions when they cannot deliver within the terms of the tenders. As warehouse receipts are good collateral for banks, traders can more easily constitute their stocks in advance of tenders.
WFP considered the warehouse receipt option in a policy issues paper presented to its Executive Board (WFP, 2006), describing it as a “private sector response to manage risk and increase credit flows”, going on to say that “WFP does not have a direct role in the creation of such systems”, though “WFP can then use warehouse receipts as a means to secure food suppliers”. This is not altogether correct because warehouse receipt systems are as much a ‘public good’ as they are a ‘private sector response’. In the case of the USA, WFP’s largest donor, the agricultural price support agency (the Commodity Credit Corporation) is a key pillar of the system established under the US Warehousing Act of 1916. Public interest makes it quite legitimate for WFP to take a proactive stance in establishing such systems.
It is sometimes objected that WFP is “not a reliable buyer in local markets, even where it has a long-term presence” (WFP, 2006). However, the impermanence of WFP in a given country and the variability of its level of procurement do not preclude it from contributing to the development of new market institutions. The long-term viability of the systems will depend upon buy-in from private players and notably the banks, but WFP can play an important role in kick-starting the system and inducing the other players to get involved. This catalytic role can be particularly important in African countries where, following structural adjustment, banks have largely lost contact with the agricultural sector, particularly with field crops produced by small farmers. For banks to re-involve themselves in this sector requires them to ascend a steep learning curve; they are more likely to do so if dominant commodity buyers, like WFP, provide a lead[xviii].
Making use of commodity exchanges. Food aid agencies usually procure through competitive tender procedures, but it is worthwhile investigating alternative competitive systems of procurement. Commodity exchanges can be highly transparent and low cost, can greatly increase market transparency and facilitate sale of grain in quantities much smaller than possible under the existing ‘restricted tender’ procedures. For example, as an alternative to opening single tenders for 3,000 tonnes, an agency could make six bids for 500 tonnes at different trading sessions. Each 500 tonne lot could be assembled from smaller quantities offered by different suppliers; hence it possible to make the system directly accessible to smaller traders and include farmers’ groups who have deposited grains in warehouses overseen by the exchange. Box 1 further discusses the advantages of procuring through exchanges.
The warehouse receipt system and the commodity exchange are complementary:the former ensures availability and quality of produce, while the latter provides a means of valuing the product and trading it competitively. However it needs to be recognised that in some parts of the World there are few viable commodity exchanges. In Africa, the Republic of South Africa is alone in having a vibrant commodity exchange[xix], and this is a futures exchange mainly used for hedging. The only successful case of a non-futures exchange is the Zimbabwe Agricultural Commodity Exchange (ZIMACE) which the Government of that country caused to close in 2001, for political reasons. However, various countries, including Malawi, Zambia, Ethiopia and Uganda, are attempting to organise exchanges, and partnership with a Food Aid agency in the early stages can enhance prospects for success.
Hence, there is scope for food aid agencies to enter into a win-win relationship with new market institutions such as warehouse receipt systems and commodity exchanges. Box 2 shows how such an arrangement could be beneficial to these parties and the public in general. However, we stress that this can only be achieved if the institutions are well organised and regulated, and public policy supports their development (Coulter and Onumah, 2002). Food aid agencies need to investigate the soundness of these emerging institutions and the supporting policy environment.
At the same time, we believe that if food aid agencies are to engage with these options then they will need to make a cultural shift. In the case of WFP, top management expresses strong interest in partnering others in testing innovative approaches[xx]. However at other levels, ideas are frequently opposed at out of conservatism, and fear of changing established procedures, rather than by systematic analysis of pros and cons, or of risks and rewards. A leading speaker (Tom Carr) at the Second African Grain Summit echoed these concerns when he spoke of WFP needing to consider itself as “part of the trade” instead of being different and separate from it[xxi].
Conclusion and recommendations
The above findings confirm the research hypothesis that LRP can make a much larger contribution to the economies of developing countries, and poor people in particular, and that polices and practice guidelines can be put in place to increase such benefits. The main challenges are in the timeliness of purchase, development of more efficient supply chains and more generally enhancing the performance of food markets in the source countries.
The authors make the following specific recommendations to those concerned with the distribution of food aid:
- Food aid donors should devote more resources to assessing the impact of local procurement on supplying areas, particularly to determine the benefits from market development and identify issues requiring attention. The following are a few of the topics that might be investigated: (a) lessons from GTZ and REST’s local purchase programmes in Ethiopia; (b) the role of regional procurement in developing long-term trade relationships between neighbouring countries and; (c) the role of local procurement in kick-starting agri-business development – particularly increasing soybean production with a view to developing efficient manufacture of animal feedstuffs.
- Governments and donors should strive to improve the information basis for food aid distribution, including local purchase. This will involve enhancing the accuracy of crop forecasts, food balance sheets and estimates of marketed surpluses, and as recommended by Tschirley (2007), developing data bases and understanding commodity chains. Food balance sheets should take account of the substitutability across food crops (including root crops) and informal regional trade[xxii]. Uganda needs reliable production statistics, while Ethiopia needs to establish a single authoritative source of the same, to improve statistics for local and regional procurement, and to improve co-ordination among food aid agencies. Government should also consider the possibility of bringing forward the cereal availability survey to the time of harvest (November-December), rather than January-February as at present.
- Donors that supply food aid in kind, and particularly the USA, should follow the example of the Canadian Government and untie a large proportion of its food aid, so as to contribute to the development of local agriculture, trade and livelihoods.
- Donors should take all possible steps to facilitate timely and effective decisions by local food aid managers. Those already supporting local procurement should make multi-annual cash commitments, which will reduce uncertainties concerning the availability of funding. There is also a need for higher levels of co-ordination and information sharing between the leading donors, and considerable delegation to their local offices.
- Where food reserve agencies still exist and function effectively, donors should investigate ways of entering into arrangements whereby the food reserves provide stocks for timely distribution in return for pledges to replace through local purchase. The food reserve agency gains by finding it easier to rotate its stocks; the food aid agency gains ready and timely access to stocks.
- Food aid agencies should attempt to improve the performance of existing marketing systems, rather than create new structures involving direct and exclusive relationships with producer organisations. Present attempts to procure from farmer groups and other small suppliers appear to be ineffective and do not represent value for money.
- Where local circumstances are favourable and food aid is expected to continue at significant levels over several years, food aid agencies should support the development of local market institutions, such as warehouse receipt systems and commodity exchanges. As the leading food aid agency, WFP should be much more proactive in this area.
Agridev Consult (2005)Local and Regional Food Procurement – an Analytical Review (Ethiopian Case Study); Final Report. Study report for WFP, Mimeo
Amha W. , Stepanek J. , Jayne T. S. and Negassa, A. (1997)Meeting Food Aid and Price Stabilisation Objectives through Local Grain Purchase: a Review of the 1996 Experience. Working paper 7, Grain Market Research Project, Ministry of Economic Development and Co-operation, Addis Ababa.
Barrett C. B. and Maxwell D. G. (2005)Food Aid after Fifty Years; Recasting its Role. Routledge, London
Coulter J. P. and Onumah G. E. (2002)The role of warehouse receipt systems in enhanced commodity marketing and rural livelihoods in Africa. Food Policy (27) 4 pp 319 337
Deloitte Emerging Markets and NRI (2004)Vulnerability, Access and Market Price Stabilisation in Ethiopia. Consultancy report for USAID and the Netherlands Embassy, Addis Ababa, Ethiopia
Gautam, M. (forthcoming paper)Ethiopia: Policy Options for Pro-Poor Agricultural Growth. Mimeo
Institut du Sahel and CILSS (2006)Impact of WFP’s Local and Regional Food Purchases (a Study Case on Burkina Faso). Final Report submitted to World Food Programme.
Jayne, T. S. (2005)Food Marketing and Price Stabilisation Policies in Eastern and Southern Africa: a Review of Experience and Emerging Policy Options. Paper presented at Price Risk Management Workshop, Washington, D. C., 28 February-1 March 2005.
OECD (2005)The Development Effectiveness of Food Aid: Does Tying Matter?Pre-print edition. OECD publishing
Smalbruch, G and Walker D. J. (2004)Fewer adverse side effects. D+C Development and Co-operation, 31, June, pp 239-241
Sserunkuuma & Associates Consult (2005)Local and Regional Food Procurement in Uganda; an Analytical Review. Study report for WFP, Mimeo
Thurlow R. and Kilman S. (2005)Meal Ticket: Farmers, Charities Join Forces to Block Famine-Relief Revamp. Wall Street Journal, 26 October 2005; Page A1
Tschirley, D. (2007)Local and Regional Food Aid Procurement: an Assessment of Experience in Africa and Elements of Good Donor Practice. MSU International Development Working Paper No. 91. http://www.aec.msu.edu/fs2/index.htm. East Lansing; Michigan State University.
Wandschneider, T. and Hodges, R. J. (2005)Local Food Aid Procurement in Ethiopia. Case Study Report by Natural Resources Institute for EC-PREP (UK Department for International Development). http://www.ec-prep.org
WFP (2006)Food Procurement in Developing Countries. Policy Issues, Agenda Item 5, Executive Board First Regular Session, 20-23 Feb. 2006. WFP/EB. 1/2006/5-C. Available on http://www.wfp.org/eb
Walker, D. J. and Wandschneider, T. (2005)Local Food Aid Procurement in Ethiopia. Case Study Report by Natural Resources Institute for EC-PREP (UK Department for International Development).
[i]Corresponding author email address: email@example.com
[ii]EC-PREP is a UK-funded policy orientated research programme to enhance DFID/European Commission collaboration.
[iii]This was organised by USAID Africa Bureau and Food for Peace.
[iv]This was organised by COMESA, the East African Community, the RATES Project, USAID, and the Nairobi Stock Exchange on 12-13 October 2005.
[v]Development projects and seed companies provide estimates in this range. Due to the lack of reliable production statistics we cannot provide a better estimate
[vi] Source of data:RATIN, Regional Agricultural Trade Intelligence Network, Nairobi.
[vii] Leading players typically install equipment with capacity of 200 tonnes of maize per 24 hours.
[viii] Some anecdotal indicate that private traders sometimes manage to masquerade as farmer groups. However, we are unable to assess the quantitative significance of this.
[ix] Source:Jonathan Coulter and Alex Rwego, Ugandan Commodity Exchange, 2006: unpublished notes from fieldwork.
[x] EuronAid is a European NGO Association and Network whose members are active in Food Aid and Food Security programmes
[xi] Barrett and Maxwell (2005) state that while in principle, local procurement allows donors to drastically reduce time between call and delivery, vis-à-vis the alternative of transoceanic shipment, the limited available evidence on delivery delays is somewhat mixed. They recommend that several major donors, “perhaps most especially the EC”, should expedite their approval and disbursement procedures. The EC has earned a reputation for often being slower to deliver resources than the USA, despite the EC’s more flexible and frequent use of local purchases, triangular transactions and complementary cash resources.
[xii] Thurlow and Kilman (2005) quote Andrew S. Natsios, USAID Administrator commenting upon the US Government’s proposal to use a quarter of the Food for Peace Program to buy food as close as possible to countries stricken by famine. According to Natsios, the proposal could have saved lives in Ethiopia during the 2003 famine. However, it was successfully opposed by Congress. A bumper wheat harvest the previous year depressed local prices so sharply that farmers were discouraged from planting. When drought hit in 2003, production fell further and a famine resulted. If the USA had bought Ethiopian wheat in 2002 to use as food relief, such a move would have stabilized prices and supported the local farm economy. Instead, in 2003, the USA shipped $500 million of US food. The US food travelled past warehouses filled with the 2002 Ethiopian harvest.
[xiii] A budget of Feb 2007 for WFP to provide Protracted Relief and Recovery in Uganda over a three year period suggests that savings could be substantial. The total budget presented is for US$ 411 million, of which food cost is US$ 187 million; external transport $ 51 million, and; landside transport, storage and handling $105 million. Some of the ‘landside transport’ relates to the shipment of in-kind donations from Mombasa, but given the scale of local purchase much of it corresponds to internal transport within Uganda.
[xiv] At the time of the fieldwork, traders overwhelmingly supplied from a non-stock position. WFP claims a large part of the grain it procures now comes from a stock position, but we were not able to get a percentage breakdown.
[xv] Price data from RATIN.
[xviii]Richard Pelrine, terminating his stint with the USAID-funded Rural Speed financial services project in Uganda states as follows: “the donor community seeks to promote new financial products of interest to rural areas but habitually underestimates the learning-curve for banks, and the difficulties of tempting them away from alternatives, e. g. financing investment in urban real estate. The response of the international banks is constrained by the fact that non-traditional loans must normally be approved by headquarters outside of Uganda”.
[xix] This is the SAFEX Division of the Johannesburg Stock Exchange
[xx]In her opening remarks at the Global Procurement Meeting in Rome, 30 May, Catherine Scheeran, Executive Director of WFP stated that she hoped the meeting would come up with solutions on how WFP could use commodity exchanges and warehouse receipts initiatives to procure food. WFP would look into these marketing initiatives and try and pilot such systems with a view to expanding them.
[xxi] www.grainstradesummit. com
[xxii] Recommendation by T. S. Jayne in presentation at Food Aid Roundtable, USAID, Washington, 15 August 2005