Maize Marketing and Trade Policy in a Pro-Poor Agricultural Growth Strategy: Insights from Household Surveys in

Eastern and Southern Africa

T.S. Jayne, B. Zulu, D. Mather, E. Mghenyi, E. Chirwa, and D. Tschirley

ZERO DRAFT June 14, 2005

Paper prepared for the Conference on “Toward Improved Maize Marketing and Trade Policies

in the Southern Africa Region,” Sponsored by the Food, Agriculture, and Natural Resources

Policy Analysis Network (FANRPAN)June 21-22, 2005,

Centurion Park Hotel, Centurion, South Africa

This paper draws on insights generated from previous and on-going research on agricultural marketing and policy issues in the Department of Agricultural Economics at Michigan State University. Support for this analysis was provided by the Rockefeller Foundation, USAID/EGAT and Africa Bureau, and the World Bank. Support from USAID missions in Kenya, Mozambique, and Zambia is acknowledged for their role in financing the collection and analysis of household survey data reported in this study.

T.S. Jayne and D. Tschirley are Professors, International Development, Department of Agricultural Economics at Michigan State University. B. Zulu is a Research Fellow at the Food Security Research Project, Lusaka, Zambia. E. Chirwa is a faculty member at the University of Malawi, Chancellor College. D. Mather is Assistant Professor, International Development, Michigan State University. E. Mghenyi is a graduate research assistant in the Department of Agricultural Economics at Michigan State University.

Maize Marketing and Trade Policy in a Pro-Poor Agricultural Growth Strategy: Insights from Household Surveys in

Eastern and Southern Africa

1. Introduction

There is widespread recognition that the existing food marketing policy environment in most of eastern and southern Africa is not generating the growth in farm productivity required to raise living standards and reduce poverty. One of the region’s major development challenges is to identify and put into place the policies, institutions, and investments that will enable agricultural marketing systems to catalyze productivity growth on the millions of smallholder farms in the region.

There is also widespread agreement on the broad ingredients of an agricultural marketing system capable of promoting farm productivity growth in eastern and southern Africa: these involve the three “I’s” of institutions, incentives, and investments (Gabre-Madhin, 2005; Coulter, 2005). Problems of uncertainty, price instability, access to markets, and weak coordination between the various stages in the food supply chain, pose major challenges for marketing actors and African policy makers. The role of institutions is to provide a certain amount of predictability about the behavior of other actors in the value chain, to provide incentives that channel behavior toward investing in cost-reducing technology throughout the entire agricultural supply chain, and to protect the interests of the poor. However, such discussions about the three I’s seldom offer much specific implementation guidance for policy makers. To be helpful to policy makers in the region, we need to know how to translate these three “I’s” into concrete and practical implementation (Omamo, 2001).

Of course, guidance can be gleaned from experience -- what has worked and not worked in the past. However, learning from experience requires an accurate characterization of the policy environment being studied. While the food marketing policy environment in eastern and southern Africa since 1990 has produced unimpressive results over the past decade, this paper argues that the empirical record has been largely mischaracterized, and that the policy implications arising from a more precise understanding of the food marketing policy environment over the past decade will sharpen our understanding of what is still needed to incorporate the three I’s in practice into the food marketing systems in the region.

The objective of this paper is to identify the major policy directions and some specific actions that governments might consider to enable domestic and regional food markets to kick-start pro-poor agricultural growth. We start by reviewing the policy environment in Malawi, Zambia, Zimbabwe, and Kenya since 1970. Then, based on recent nationwide surveys of small-scale farm households in Kenya, Malawi, Mozambique, and Zambia, we discuss the implications of smallholder production and marketing behavior for the design of food marketing and trade policies in the region. These nationwide survey findings

2

provide an updated empirical foundation for discussion of the role of food marketing and trade within the context of pro-poor agricultural growth strategies in Eastern and Southern Africa. The paper argues that some of the conventional wisdom guiding food policy debates in the region may warrant reconsideration as smallholder land allocation and food production and marketing patterns have changed over time.

Lastly, we attempt to spell out parts of a specific policy agenda for maize marketing systems to catalyze small farm productivity growth and food security. The intent of the paper is not to deal with implementation modalities, which would be premature since no government has yet to explicitly endorse the policy positions taken in the paper. Moving from this policy agenda to concrete implementation modalities – the “how” questions raised by Omamo (2003) – will require intensive and country-specific consultations between policy makers and analysts.

2. Maize Marketing Policy Environment since 1970

The 1970-1990 Period1

An important policy objective at Independence in Kenya, Malawi, Zambia, and Zimbabwe was to expand smallholder grain production. In each country, the newly independent governments responded to this objective in a fundamentally similar way. The key features were (a) expansion of state crop buying stations in smallholder areas that had been excluded from these benefits under the former colonial regime; (b) continuation of direct state control over grain supplies and pricing; (c) efforts to stabilize and often subsidize urban consumer prices without reliance on imports; (d) elimination of the dominant role of non-indigenous minorities; and (e) shifting the cost of the marketing system onto the Treasury. The expansion of state market infrastructure in smallholder areas facilitated the disbursement of credit and subsidized inputs to smallholders by allied state agencies designed to recoup loans through farmer sales to the marketing boards (Rohrbach 1989; Howard 1994; Putterman 1995). Management control of the grain boards was wrested from white commercial farmer interests. These attempts to achieve broader objectives occurred without fundamental changes to marketing institutions or to the policy framework. Land redistribution was accorded a limited role in the attempt to redress inherited dualistic agrarian structure. Agricultural dualism was in fact increased in Malawi, and to a lesser extent in Kenya and Zimbabwe, through state support to a newly created black-owned large farm sector.

Price stabilization policies were almost always implemented in ways designed to alter the mean as well as the variability of food prices. A key objective of pricing policy during this period was white maize self-sufficiency. Because the world market for white maize

1 This section draws heavily on Jayne and Jones (1997).

3

was thin,2 domestic production shortfalls often necessitated imports of yellow maize, a less-preferred substitute that had become connoted with an agricultural policy failure (Jayne and Rukuni 1993). Government stockpiles of white maize expanded both deliberately as a buffer against drought and as the unplanned result of increased state support to smallholder agriculture. The expansion of state buying stations in remote areas under pan-territorial pricing encouraged a notable increase in smallholder grain production, especially in Tanzania, Zambia, and Zimbabwe (Putterman 1995; Howard 1994; Rohrbach 1989). In most cases, the treasury and aid donors bore most of the cost of expanding marketing services to smallholders.

While the post-independence model of service provision to smallholders appears to have had important successes in boosting grain production and incomes in some rural areas, by the mid-1980s major problems had emerged in all the countries that propelled the grain marketing systems toward reform:

1. Marketing board costs escalated as the scale and complexity of their activities increased. The losses could be categorized into two types: those which government forced on the board by mandating it to carry out activities without allowing it to fully recover the costs, and those related to operational inefficiency: Of the first type, pan-territorial pricing was a particularly important source of losses in Tanzania, Zambia, and Zimbabwe, since this raised the share of grain delivered to the boards by smallholders in remote (but often high-potential) areas where transport costs were high (Bryceson 1993, GMB 1991). Also in this category were the losses associated with stockpiling white maize, which were largely a consequence of government food security objectives and pricing policy oriented toward maize self-sufficiency (Buccola and Sukume 1988; Pinckney 1993). Operational inefficiency varied across countries, but reached its nadir in Tanzania, Zambia, and Kenya (Amani and Maro 1992; Jansen and Muir, 1994; Bates 1989). Allegations of corruption were widespread, even in Zimbabwe’s relatively efficient Grain Marketing Board (CSM 1986). In some cases, the treasury costs of state marketing operations became so large as to affect the rates of inflation, interest, and currency exchange, especially in Zambia during the 1980s (Jansen and Muir 1994).

2. The state systems for input delivery and crop payment became increasingly unreliable, especially in Zambia, Tanzania and Kenya, (Howard 1994; Amani and Maro 1992; Westlake 1994). Smallholder credit repayment also became problematic in many cases. In Zimbabwe, almost 80% of smallholder recipients of state credit were in arrears in 1990 (Chimedza 1994).

3. Farmers near urban demand centers who were implicitly taxed through pan-territorial pricing increasingly resorted to parallel markets (as in Tanzania, Kenya, and Zambia) and/or switched to other, uncontrolled crops (in Zimbabwe and South Africa). Declining volumes through the state marketing channels further exacerbated the boards’ trading losses.

2 This is no longer the case as the US has dramatically increased its white maize production base in response to NAFTA. Since 1995, the US has exported an average of 1.2 million tons per year, about two-thirds of it to Mexico. Moreover, US white maize is 99% certified GMO free, an important consideration for eastern and southern Africa (Tschirley et al., 2004).

4

4. The increasing proportion of maize sales by smallholder farmers, generally on poorer land with less reliable rainfall, increased the instability of marketing board purchases and sales, and hence of the fiscal demands made by the marketing system.

5. Controls on private grain movement, imposed by the earlier colonial governments, were continued in all countries even after independence. Aside from continuing to block smallholders’ direct access to urban markets, the controls also prevented grain from flowing directly from surplus to deficit rural areas (Jayne and Chisvo, 1991; Odhiambo and Wilcock 1990). The system was based on the implicit assumption of rural grain self-sufficiency. On the surface, this assumption seemed plausible enough, because grain sales normally rose rapidly in smallholder areas where marketing board infrastructure was developed (Bryceson 1993; Rohrbach 1989). This provided some evidence of a "surplus" in excess of a particular area's consumption requirements. However, micro-level research had shown that grain deliveries to the marketing boards did not necessarily indicate a "surplus" from a given region, over and above consumption requirements, since sales from a small segment of well-equipped farmers often masked considerable grain deficits among a large proportion of the rural population. Official restrictions on private trade and weak market infrastructure often made it easier for surplus farmers to sell to the Boards rather than their deficit neighbors a few kilometers away (Jayne and Chisvo 1991).

6. A growing body of empirical evidence pointed to the controlled marketing systems as suppressing or imposing additional costs on parallel trading and processing channels that often served the interests of both producers and consumers more effectively than the official state apparatus (Odhiambo and Wilcock, 1990; Jayne and Jones, 1997).

The period since 1990

Fiscal crises and increased donor leverage over policy pushed the grain marketing systems of Eastern and Southern Africa toward liberalization in the mid-1980s. After first trying to strengthen the performance of state marketing boards in the 1960s and 1970s, donors and international lending agencies began promoting the reform of food marketing and pricing as a central component of structural adjustment programs in Africa. Several factors shaped this change. Donors lost patience with phased and partial reform programs that were increasingly seen as propping up costly and otherwise unsustainable pricing and marketing policies rather than facilitating reforms (Jones 1994). In addition, political economy models (e.g., Bates 1981) suggested that state interventions in agricultural markets, while ostensibly designed for rural development or to correct for market failures, were in fact designed to serve the interests of a dominant elite composed of bureaucrats, urban consumers, and industry. The framework of policy-based lending has strongly influenced the path of market reforms in each of these countries (except South Africa), and has expanded external leverage over domestic agricultural policy through aid conditionality. By the early 1990s, and despite much internal opposition, countries such as Kenya, Malawi, Zambia and Zimbabwe had embarked upon food market reform programs.

5

Up to this point, there is a general consensus on the effects of pre-reform food policies and on the pressures leading to market reform in Eastern and Southern Africa. After this point, however, there are divergent interpretations and implications of the reform experience during the 1990s and early 2000s. One version indicates that the reforms took place, the private sector was accorded a major role in responding to the opportunities opened up by liberalization, but the private sector’s response was weak because of market failures, high transaction costs, and coordination problems between different stages of marketing. As a result, most smallholder farmers were unable to access modern inputs and were cut off from stable and remunerative output markets that could have enabled them to raise farm productivity and initiate a transition out of poverty. The disappointing trends in African food production, input use, and food security over the past decade underscore the need for a reconsideration of the market reform / public goods investment model that drove most donors’ thinking in the 1980s and early 1990s (e.g., Dorward et al., 2004; Gabre-Madhin, Barrett, and Dorosh, 2003). In particular, some mix of market stabilization and subsidization of farm inputs and commodity markets may be necessary in Africa to kick-start the transformation process – policy tools that were identified as important factors contributing to the “green revolution” experiences in Asia.

Another interpretation of events is that market reform -- in countries such as Kenya, Malawi, Zambia, and Zimbabwe -- was mainly a concept that donors thought they were achieving through aid-conditionality agreements rather than being a policy choice adopted by African governments (e.g., van de Walle, 2001; Jayne et al., 2002; McPherson, 2002; Bird, Booth and Pratt, 2003). Many governments, unable to continue implementing large-scale price subsidy cum stabilization programs, adopted the language of donor reform in their policy documents and official pronouncements in order to continue accessing external funds, which were then used to continue pursuing a similar set of policies and distributional objectives as before, albeit on a scaled-down manner commensurate with the lower level of funds available for marketing board activities. Private trade was indeed legalized, but was seldom supported (Meerman, 1997). Public investment to improve the functioning of markets was low, partly because public budgets to agriculture continued to be disproportionately channeled to marketing board activities. Marketing boards were only in a few cases eliminated; in many countries, the boards continued to operate as major actors in the market (Jayne et al., 2002). In some cases, successor agencies were formed with new names, ostensibly signifying a more truncated role confined to holding strategic stocks, but whose activities in practice may have exacerbated price instability while channeling resources to key constituents as part of governments’ overall goal of maintaining a coalition in the new multi-party political systems that emerged in the 1990s. Price stabilization policies were pursued in a highly discretionary way, typically altering both the level and variability of prices, thus affecting the risks and incentives facing private traders. Competing with government agencies releasing supplies unexpectedly has crowded out private arbitrage, both spatial and temporal, and has contributed to the perception that the market cannot be relied upon to ensure adequate food supplies when needed.

6

These two narratives are not, for the most part, mutually exclusive. They have sometimes consistent, and sometimes inconsistent, implications for future food marketing policy. Notwithstanding the lack of consensus on the causes, nature, and consequences of food marketing policies in the region during the 1990s and 2000s, it is clear that food price instability and weak coordination of activities in the food marketing system remain fundamental problems. Especially in the southern Africa region -- which has arguably transitioned from maize surplus to maize deficit as a consequence of turmoil in Zimbabwe and a reduction of exportable surpluses in South Africa, and where an increasing number of rural dwellers depend on the market for their food security -- ensuring adequate food supplies at tolerable prices for consumers has become a major problem. Redressing these problems will be a critical part of future attempts to raise rural livelihoods in the region.

The maize marketing policy environment since 1990

Many African countries have curtailed or eliminated direct government operations in food markets. However, in countries such as Kenya, Malawi, Zambia, and Zimbabwe, governments continue to pursue price stabilization and food security objectives through marketing board operations. The boards’ operations have generally been more modest in recent years than during the pre-control period. However, they continue to be major actors in the maize market.

Trends in Kenya’s National Cereals and Produce Board maize purchases and sales are presented in Table 1. The NCPB’s operations have trended downward since the early 1990s, after the country began its Cereal Sector Reform Programme in 1988. However, since the private trade was fully deregulated in 1995, the NCPB has continued to purchase about 26 percent of the marketed maize output, based on estimates that roughly a quarter of production is marketed. The NCPB generally has set maize purchase prices above market prices, and its operations are estimated to have raised domestic maize prices by about 15 percent in Nairobi and by about 16% in western Kenya since 1995 (Jayne, Myers, and Nyoro, 2005). The same study finds that NCPB activities have served to reduce the volatility of maize market prices over the same period. In addition, the Kenyan government has sought to influence maize price levels through a variable import tariff. Wholesale maize market prices in Kenya have been among the highest among Africa’s maize producing countries, averaging roughly $197 per ton in Nairobi between 1995 and 2004 (Jayne, Myers, and Nyoro, 2005).

The governments in Malawi and Zambia have strikingly similar agricultural policies. Both have withdrawn from direct intervention in most of their input and output markets, with the important exception of maize, fertilizer, and seed inputs. The governments coordinate maize imports, hold strategic maize reserves, and implement subsidized or free input delivery programs. In the past 10 years from 1994/95 to 2003/04, Malawi’s

7

Table 1. Marketing Board purchases and sales’ share of total sales from domestic production

Maize Estimated NCPB NCPB NCPB purchases NCPB sales as

Kenya output marketed

output from purchases sales as % of marketed

surplus % of marketed

surplus production

1990 2290

1991 2340

1992 2430

1993 2089

1994 3060

1995 2699

1996 2160

1997 2214

1998 2400

1999 2322

2000 2160

2001 2776

2002 2340

2003 2300

‘000 metric tons

573 585 608 522 765 675 540 554 600 581 540 694 585 575

233 316 488 463 535 100 62

150 35

175 308 255 88

160

663 728 255 508 67

110 54 14

122 144 73 23

194 135

41% 116% 54% 124% 80% 42% 89% 97% 70% 9% 15% 16% 12% 10% 27% 3% 6% 20%

30% 25% 57% 14% 37% 3% 15% 33% 28% 24%

Maize Estimated ADMARC ADMARC ADMARC ADMARC marketed purchases sales purchases as % sales as % of

Malawioutput

output from of marketed marketed production surplus surplus

1990 1343

1991 1589

1992 657

1993 2034

1994 1040

1995 1339

1996 1793

1997 1226

1998 1772

1999 2479

2000 2501

2001 1589

2002 1557

2003 1983

‘000 metric tons

336 397 164 508 260 335 448 307 443 620 625 397 389 496

190 600 26

442 67 61 87 96 13 58

198 11 21 42

68 90 26 40

115 39 25 34

146 67

57% 151%

16% 87% 26% 26%

18% 27%

19% 6%

31% 13%

3% 26%

9% 6%

32% 4%

3% 9%

5% 37%

8% 14%

8

Table 1, continued.

Zambia

Maize output

Estimated marketed

output from production

FRA purchases

FRA sales FRA purchases as % of marketed

surplus

FRA sales as % of marketed

surplus

‘000 metric tons 1995 1996 1409 352 16 na 5% na

1997 960 240 75 31% 1998 638 160 200 125% 1999 822 206 23 11% 2000 882 221 35 16% 2001 601 150 155 103% 2002 602 151 25 17% 2003 1161 290 98 34%

Notes: In all three countries, estimated maize sales are approximated as 25% of domestic maize production. ADMARC/NFRA sales from official imports are added to domestic ADMARC sales. Data on Zambian FRA maize sales not available. Sources: Malawi: (a) ADMARC purchases: World Bank, 2003 (for years 1995/96 to 2000/01); ADMARC data files supplied by Ministry of Agriculture for remaining years. (b) ADMARC sales from domestic production: ADMARC data files supplied by Ministry of Agriculture. ADMARC/NFRA sales from imports from Ministry of Agriculture. (c) maize production: FAOStat website. Kenya: (a) NCPB purchases and sales are from NCPB data files, 2005, reported in Jayne, Myers and Nyoro, 2005. (b) maize production: FAOStat website. Zambia: (a) Food Reserve Agency (FRA) maize purchases from Food Reserve Authority data files, reported in Tembo et al., 2005. (b) maize production: FAOStat website.

ADMARC has purchased a mean of 39,000 tons of maize per year (Table 1). Using FAO statistics on maize production and based on the assumption that 25 percent of production is marketed,3 then over the 1994/95 to 2003/04 period, ADMARC has accounted for 15% of marketed maize output. In the past 3 seasons, production has declined, and hence the government’s primary presence in the market has shifted from maize purchasing to the importation and release of maize stocks to ensure consumers’ access to maize. For example, when decision makers in Malawi were presented with a food balance sheet in May 2002 that forecast a deficit of 433,000 metric tons for the 2002/03 season, they imported 253,000 metric tons of maize entirely through government channels (NFRA), and arranged for 151,000 metric tons of food aid, for a total formal inflow of over 400,000 metric tons, nearly covering the forecasted deficit. However, these plans did not take into account the large informal trade of white maize from Mozambique into southern Malawi that had developed since the mid-1990s. Best estimates are that, during the 2002/03 season, 150,000-250,000 metric tons of maize entered Malawi informally from Mozambique, leaving the country with a maize surplus of about the same amount (Whiteside 2003). In March 2003, facing a good incoming harvest and the prospect of storing maize for over a year, the government sold some of its stock at very low prices.

3 Using the nationwide Integrated Household Survey of 1998, Chirwa (2005) reports that only 9.7 percent of smallholder maize production is marketed. This would imply that ADMARC’s maize purchases as a proportion of total marketed output is considerably greater than 15%.

9

This exacerbated the problem of already low maize prices, and most likely contributed to price instability and risks for farmers and traders (Tschirley et al., 2004).

In Zambia, although the food marketing parastatal, NAMBOARD, was abolished in 1990, the Food Reserve Agency (FRA) was formed in 1996. The FRA was originally conceived to hold buffer stocks to dampen price variability and, when necessary, provide liquidity in the maize market during the initial years of market liberalization while the private sector was establishing itself. Since 1997, the FRA has handled roughly 22 percent of the country’s domestically marketed maize. However, when maize imports arranged by the government are included, the governments’ control over maize supplies has occasionally exceeded total domestic supply (e.g., in 1998 and 2001, as indicated in Table 1). The government has also remained involved in arranging maize imports, subsidizing the price at which it offers maize imports to large millers (being in the range of $165 per ton in 2002, when the c.i.f. price of maize from South Africa was at least $230 per ton) (Nijhoff et al., 2002).

Although FRA’s original mandate did not include a price support function, soon after its creation, the agency was instructed to purchase maize in remote areas where maize production for the market is unlikely to be profitable under commercial conditions. In addition, fertilizer distribution was added to FRA’s activities, which further entrenched its maize purchase activities since maize was accepted as in-kind repayment for the fertilizer distributed by the FRA (Govereh et al, 2001). In 2004, the President of Zambia announced a floor price of $150 per ton (ZK 36000 per 50kg bag). In the most recent year for which data is available, 2003, the FRA is believed to have purchased roughly 34% of the country’s domestically marketed maize. Thus, the government has arguably been active in the maize market, and the combination of FRA maize operations and subsidized fertilizer distribution programs accounts for 63% of the 2005 national budget allocated to the Ministry of Agriculture (Shawa, 2005).

State maize operations in support of price stabilization and other objectives appear to have been equally if not more active in Zimbabwe, but recent data has been difficult to obtain.

A partial list of the ways in the governments directly intervene in domestic and regional maize markets and affect the development of the three I’s of market institutions, incentives, and investments includes the following:

1. Maize export bans: (Zambia: 1995/96 and 2005; Kenya: several times during the 1990s). These are normally implemented without prior notice, and can strand private firms that have already sourced maize internally for export. After the ostensible maize market reform process began in Zambia in the early 1990s, international firms such as Louis Dreyfus and Cargill established trading operations in Zambia. However, both firms closed their operations in Zambia in 1996 after an export ban was suddenly announced, causing these firms to incur large financial losses after having accumulated large maize supplies for export to the DRC.

10

2. Maize import tariffs: Zambia (continuous 10%); Mozambique (18% of maize grain) ; Kenya (ranging from zero to 35% since 1990). These reduce the flow of trade within the region and increase spatial price differences between countries.

3. Un-announced and unpredictable waiving of maize import tariffs, sometimes for several days in Kenya’s case. An environment in which the import tariff can change suddenly stymies private traders from importing maize when the situation would otherwise warrant doing so. If traders suspect that the import tariff will be waived later in the year, this means that if they mobilize imports early (while the tariff is in place), they are likely to lose their market later when competing against other firms that can import more cheaply once the tariff is waived (J. Magnay, Ugandan Grain Traders Ltd, personal communication, 2005). The result is commonly a temporary under-provision of imports, which can produce a situation in which local prices exceed import parity levels for periods of time. Some political scientists have argued that unpredictability in import tariffs can be a deliberate tool of government policy to carry out patronage activities (e.g., van de Walle, 2001; Grosh, 1997) through waiving the import tariff for several days when a favored firms’s maize is being off-loaded at the port, and then subsequently reinstating the tariff. Analysts not familiar with the details of the situation often erroneously take such instances as evidence that markets fail and that the private sector is weak in Africa, leading to a rationale for continued direct government involvement in marketing.

4. Government importation and sale of subsidized maize to selected buyers as has occurred in Zambia (1999 and 2001); Malawi (in the 2001/02 crisis year). Similar to the foregoing example, government actions in the market can unintentionally lead to price surges well above import parity. For example, in 2001/02, the Malawi government imported maize from South Africa to distribute at subsidized prices, considerably lower than market levels, to protect poor rural consumers. However, the government imports arrived late and were not sufficient to meet demand. As a result, ADMARC depots began to experience stock-outs, and prices soared. Yet the private trade had not imported because they expected to be unable to compete against the low ADMARC official maize selling price. When it became clear that ADMARC’s supplies were insufficient to last the full season, private traders scrambled to import, but for several months much of rural Malawi experienced grain shortages and prices that were reportedly as high as $450 per ton (Rubey, 2004; Tschirley et al., 2004). The upshot from this is that well-intentioned but poorly implemented government actions can exacerbate food price instability rather than reduce it. Malawi had the highest magnitude of maize price instability of four countries analyzed in the region between 1994 and 2003 (Table 2).

Because Malawi’s ADMARC, Kenya’s NCPB, Zambia’s FRA, and Zimbabwe’s Grain Marketing Board frequently import maize (or use agents to handle the logistics of importing on their behalf) in quantities that are large compared to the size of the market, and sell at prices that can be substantially lower than the cost of commercial importation, the private sector is constantly at risk of watching its stocks lose value. In this environment, stockholding is extremely risky and there are no assurances that normal intra-seasonal price rises will occur due to the uncertainty over government action. The

11

worst possible scenario is that governments announce a ceiling price, state their intention to import maize to defend the ceiling price, and then not be able to do so. While some analysts point to the large intra-seasonal price variability observed in countries such as Malawi and Zambia as indicators of weak private trader capacity and market failure, we submit that the variability observed is from a policy environment in which government takes the initiative and primary responsibility for stabilizing food prices and is not reflective of a policy environment where the private trade is relied upon to stabilize prices.

Note that Malawi prices are much more unstable than those of Zambia, Mozambique or South Africa, even though price stabilization has been pursued most aggressively in Malawi over the sample period (see Nucifora, 2005). Mozambique has no direct government operations of any type in the maize market, although it has instituted a maize import levy at times on maize imported from South Africa. Zambia has over the sample period used import and export bans and taxes, importation of maize during apparent shortfalls, and some direct government involvement in maize purchasing and sale.

Table 2. Coefficient of Variation (a Measure of Instability) in Monthly White Maize Prices, Selected Markets in Southern Africa, January 1994 to August 2003.

CV* CV de-trended** Mozambique:

Beira .57 .45 Maputo .45 .26

Nampula .77 .58

Malawi: Karonga 1.62 1.59 Lilongwe 1.03 .71 Nkhata Bay .95 .62

Rumphi .79 .44

South Africa: Randfontein (SAFEX) .44 (.42) .34 (.34)

Zambia: Chipata .80 .55 Choma .91 .58 Lusaka .90 .59

Ndola .79 .42

Notes: * coefficient of variation (standard deviation/mean); ** Cuddy LaValle Index: CV(1-R2)0.5 where R2 is the coefficient of determination from linear time trend regression. n=114 for all markets except for Karonga (n=113), Beira (n=130) and Nampula (n=128). Figures in parentheses for South Africa are based exclusively on SAFEX prices April 1996 to August 2003. Sources: Mozambique and Zambia wholesale maize price data from Ministry of Agriculture; Malawi retail maize price data from Ministry of Agriculture; South Africa white maize wholesale from SAFEX from April 1996 to August 2003; from January 2003 to March 1996, Maize Board white maize selling price.

12

5. State monopoly on import and export of maize grain and maize meal (Zimbabwe, since 1990).

6. Internal levies on maize grain across district boundaries: Zambia (since 1992). Levies on the movement of maize across district boundaries have been implemented in Zambia, Ethiopia, and (sporadically) in a number of other countries. In Zambia’s case, Mwiinga et al (2005) found that the levies on maize grain ranged from 2 percent to 13 percent of the wholesale price and that farmers end up paying much of this tax. The levy on inter-district maize movement clearly reduces farmers’ revenue from growing maize for the market. Also, the levy exacerbates spatial price differences by dampening incentives for internal redistribution of grain from surplus to deficit areas.

7. Import licenses required for cross-border trade, yet licensing forms are commonly available only at the Ministries of Trade and Commerce in the capital cities (RATES, 2003).

8. Two-tiered selling price structure of the marketing board. e.g., Zimbabwe’s parastatal Grain Marketing Board sells maize at one price to selected registered millers, and another higher price to other buyers, thus providing market advantages to well-connected maize milling firms.

These examples, coupled with the data in Table 1, are intended to show that the governments of Kenya, Malawi, and Zambia remain important players in their maize markets. Market liberalization, as implemented in these countries, has not implied a greatly diminished role of public sector influence over food prices and supplies. Though the quantities they trade are smaller than during the controlled market era, food marketing boards in these countries still exert a dominant presence in the maize markets. State operations in the market are defended as being necessary to stabilize prices for producers, ensure the availability of food for low-income consumers when market supplies become tight, and other objectives related to food security. A key feature of this policy environment is the discretionary and unpredictable manner of state operations to influence prices and supplies (Jayne et al., 2002). Discretionary market operations without transparent and consistently enforced decision rules most likely exacerbate the coordination problems, opportunism, poor physical infrastructure, and other forms of risks and transaction costs that private traders already face in developing food markets in the region (Brunetti, Kisunko, and Weder, 1997). Empirical assessments of these countries’ food market performance since the 1990s reflects not the impacts of unfettered liberalized markets but rather the mixed policy environment of legalized private trade within the context of continued strong government operations in food markets.

Before leaving this section, we present trends in staple cereal production for these countries having pursued price support and stabilization objectives compared to cereal production trends for sub-Saharan Africa as a whole (Table 3). One cannot attribute differences in performance of these countries simply to the manner of government participation in food markets, yet it is perhaps noteworthy that none of the four countries pursuing food price stabilization and food security objectives through direct state

13

operations over the past decade have been able to match production growth for the continent as a whole. While cereal production in the Sub-Saharan Africa region as a whole has increased by roughly 60 percent over the past two decades, three of the four countries continuing to intervene heavily in their food markets are barely achieving cereal production levels of the 1980s.

Table 3. Cereal Production Trends in Kenya, Malawi, Zambia, Zimbabwe, and Sub-Saharan Africa overall, 1985 to 2004.

Sub-Saharan Africa

Kenya Malawi Zambia Zimbabwe

Production indices (1985 = 100)

1985 100 100 100 100 100 1986 106 115 96 110 90 1987 101 98 88 97 44 1988 119 113 105 172 92 1989 119 110 112 165 75 1990 112 93 99 103 76 1991 122 95 119 104 61 1992 117 97 47 53 13 1993 124 86 153 149 73 1994 129 126 78 102 80 1995 131 113 126 75 27 1996 146 94 139 134 91 1997 139 93 97 99 82 1998 146 102 136 70 55 1999 147 96 189 88 59 2000 140 89 187 91 73 2001 147 113 126 66 55 2002 145 97 124 65 22 2003 161 95 155 114 30 2004 159 95 131 114 35

Source: FAOStat website: http://faostat.fao.org/

14

3. Farm and Market Structure: Implications for State Efforts to Subsidize and Kick-Start Smallholder Maize Production

In our use of nationally representative smallholder farm survey data in eastern and southern Africa, we have noted several consistent aspects of farm structure that appear to hold major implications for the design of pro-poor agricultural growth strategies. These “empirical regularities” include 1) the inequality of landholding distribution within the smallholder sectors; 2) the consequent concentration of marketed food output; 3) the position of most rural households as purchasers of staple food rather than sellers; 4) the declining importance of maize in small farm income and in urban consumers’ diets; 5) the declining trend in farm size; and 6) the segmentation of “formal” and “informal” food marketing channels in the region. Details of the nationwide surveys are provided in Appendix 1.

3.1 Land distribution within the small farm sectors

Meaningful discussions of pro-poor agricultural growth strategies must be grounded within the context of prevailing asset distribution patterns. It is widely accepted that “pro-poor” agricultural growth is strongly associated the equity of rural asset distribution (Datt and Ravallion, 2002; Gugerty and Timmer, 1999; Chirwa, 2004). Contrary to the general assumption that Africa is generally land-abundant, severe land inequalities persist in many African countries between smallholder, large-scale, and state farms. In the Eastern and Southern Africa region, the smallholder farm sector is typically characterized as small but relatively “unimodal” and equitably distributed land holdings situated within a “bi-modal” distribution of land between large-scale and small-scale farming sectors.

However, even within the small farm sector, there are major disparities in landholding size. Using national household survey data in six countries in eastern and southern Africa, Jayne et al (2003)4 found average land holdings in the small farm sector ranging from 2.5 to 3.0 hectares in Kenya and Zambia to around one hectare in Rwanda and Ethiopia (Table 4). However, mean farm size figures mask great variations in farm size within the smallholder sector. After ranking all smallholders by household per capita land size, and dividing them into four equal quartiles, households in the highest per capita land quartile controlled between five to 15 times more land than households in the lowest quartile. In Kenya, for example, mean farm size for the top and bottom land quartiles were 6.69 and 0.58 hectares, respectively, including rented land. The computed Gini coefficients show land disparities within the smallholder sectors of these countries that are comparable to or higher than those estimated for much of Asia during the 1960s and 1970s (Haggblade and Hazell 1988).5

4 These six countries are Ethiopia, Kenya, Malawi, Mozambique, Rwanda, and Zambia. Each of these surveys were nationally representative of these countries’ small farm sectors except for Kenya and Mozambique. The Kenya survey was nevertheless a nationwide survey covering 1,578 households in 24 agricultural districts. Survey periods were 1991 for Rwanda, 1996 for Ethiopia and Mozambique, 1997 for Kenya, and 2000 for Zambia. 5 These findings are based only on the smallholder farm sectors. If these countries’ large-scale and/or state farming sectors were included, the inequality of landholdings would rise further.

15

In each country, the bottom 15 percent of small-scale farm households are approaching landlessness, controlling less than 0.5 hectares. In Ethiopia and Rwanda, the bottom land quartile controlled less than 0.20 and 0.32 hectares per capita. In Malawi, 80 percent of all smallholder households possess less than one hectare of land (Oygard et al., 2003). These surveys contain only households engaged in agricultural production; households not engaged in farming are not in these samples.

Table 4. Household Attributes by Landholding Size Per Capita Quartile, Various Countries

Mean Quartiles of Per Capita Farm Size Country Dimension farm size 1 2 3

(ha)

Kenya Landholding size (ha) Gross value of crop sales (1996 US$) Household income (1996 US$ per capita) Off-farm income share (%)

2.65 449.7 336.7 40.0

0.58 141.8 209.9 50.0

1.26 236.9 275.3 41.1

2.11 348.8 312.4 37.8

6.69 1575.7 550.3 30.6

Ethiopia Landholding size (ha) Gross value of crop sales (1996 US$) Household income (1996 US$ per capita) Off-farm income share (%)

1.17 145.8 71.6

8.1

0.20 33.7 53.1 13.7

0.67 82.3 52.1

9.0

1.15 120.6 88.3

5.4

2.58 265.2 91.0

4.6

Rwandaa Landholding size (ha) Gross value of crop sales (1991 US$ per hh) Household income (1991 US$ per capita) Off-farm income share (%)

.94 68.0 78.7 24.8

0.32 34.1 54.5 34.5

0.63 45.1 59.4 24.4

1.00 72.4 79.3 22.2

1.82 169.3 121.7 18.2

Mozambiqueb Landholding size (ha) Gross value of crop sales (1996 US$ per hh) Household income (1996 US$ per capita) Off-farm income share (%)

1.80 35.4 43.1 12.7

0.55 9.5

26.2 15.9

1.17 16.5 34.1 12.3

1.92 32.6 42.7 13.1

3.46 90.4 69.2

9.2

Zambia Landholding size (ha) Gross value of crop sales (2000 US$) Per capita income (2000 US$)

Off-farm income share (%)

2.81 68.8 62.9 28.4

0.79 20.3 48.2 38.5

1.61 32.8 53.3 25.7

2.68 73.7 65.9 26.6

6.16 180.2 84.2 22.9

Source: modified from Table 6 in Jayne et al., 2003.Notes: Samples include only “agricultural households” defined as households growing some crops orraising animals during the survey year. All numbers are weighted. Exchange rates: Kenya 58Ksh-1997US$; Ethiopia 6.2birr-1996US$; Rwanda 125.1FRW-1991 US$; Mozambique 11,294 Meticais-1996 US$;and Zambia 2,811Kw-2000 US$. Income figures include gross income derived from crop production onrented land.a For Rwanda: data is not available for land loaned out, only data on rented land is included.b North-Central Mozambique only where income data is available; no data available for land loaned out,figures only include land rented in.

16

4

Also shown in Table 4, the gross value of household crop sales from own production is highly correlated with landholding size. The gross revenue generated from crop sales among households in the top farm size quartile exceeded that of the bottom land size quartile by 8 to 11 times, except in the case of Rwanda, where the difference was only 5 times as great. These computations include all crops; when only cereals are examined, the relative disparities are larger.

3.2 Rural households’ position in maize markets

From these surveys, rural small-scale farm households generally fall into one of the following four categories with respect to major staple grain markets:

1. sellers of staple grains: Household surveys throughout eastern and southern Africa indicate that roughly 20 to 35 percent of the smallholder farms sell maize in a given year. Of course this figure will rise in good harvest years and fall in a drought year. However, there are two sub-groups within this category:

• a very small group of relatively large and well-equipped smallholder farmers with 4 to 20 hectares of land, usually in the most favorable agro-ecological areas (about 1 to 4 percent of the total rural farm population), accounting for 50% of the marketed output from the smallholder sector. These farms tend to sell between 10 and 1,000 tons of maize per farm in a given year.

• a much larger group of smallholder farms (20 to 30 percent of the total rural farm population) selling much smaller quantities of grain, between 0.1 and 5 tons per farm. These households tend to be slightly better off than households that buy grain, but the differences are not very great in absolute terms.

These households, especially the larger smallholder farmers, clearly benefit from higher grain prices, and have tended to advocate for the continuation of marketing boards procuring their crop at fixed support prices. They may also benefit from mean-preserving food price stabilization, although the benefits associated with price stabilization are likely to be much smaller than the benefits derived from raising mean prices (Myers, 2005).

2. buyers of staple grains: these rural households generally make up 50-70 percent of the rural population, higher in drought years and lower in good production years. These households are generally poorer and have smaller farm sizes and asset holdings than the median rural household. They are directly hurt by higher mean grain prices.

3. households buying and selling grain within the same year: In all of the nationwide surveys, relatively few households both buy and sell maize.6 Only about 5 to 15 percent of the rural population buys and sells maize in the same year. They comprise both relatively large farms that sell grain and buy back lesser amounts of processed meal, as

6 This empirical regularity contrasts with the common notion that, because of lack of credit, farmers typically sell at harvest at low prices and buy back latter at higher prices.

17

well as relatively poor households that make distress sales of grain after harvest only to buy back larger later in the season. However, this latter sub-group typically comprises less than 10 percent of the rural farm population.

4. households neither buying nor selling maize: these households make up a small proportion of the rural population in areas where maize is the dominant staple crop. However, in parts of northern Zambia and Mozambique, cassava is the main staple. Because of this, a sizable fraction of the rural population at the national level is autarkic with respect to maize.

Table 5. Distribution of small-scale farm population according to their position in the staple grain market, selected countries.

Household category with respect to main staple grain:

Sellers only: top 50% of total sales*

bottom 50% of total sales**

Buyers only

Buy and sell (net buyers)

Buy and sell (net sellers)

Neither buy nor sell

Zambia (maize)

Mozambique (maize)

Kenya (maize)

Malawi (maize)

Ethiopia (maize and teff)

---------------------------- % of rural farm population -------------------------­

19 2

21

13 2

10

18 2

16

5 1 4

13 2

11

33 51 55 na 60

3 7 na 13

5 12***

12 na 12

39 24 8 na 2

100% 100% 100% 100% 100% *Notes: after ranking all households by quantity sold, this row shows the percentage of households in the

**smallholder sector accounting for the first 50% of total maize sale; percentage of households accounting ***for the other 50% of total maize sales. The survey in Mozambique was not able to ascertain quantities

of maize purchased therefore whether these households are net buyers or net sellers in unknown. Sources: Zambia: Central Statistical Office Supplement to the Post Harvest Survey, 2000/01 marketing year. Includes small-scale (0.1-5 hectares) and medium-scale (5-20 hectare) farms, proportional to probability sampling. Kenya: Tegemeo Institute Household Survey, Egerton University, 1999/2000 season. Mozambique:Malawi: Chirwa (2005), based on analysis of Malawi Integrated Rural Household Survey, 1997-98, Government of Malawi. Data on maize purchases is limited to a 3-day recall period hence computation ofmaize purchases for the marketing year is not possible. Ethiopia: Central Statistical Authority, Government of Ethiopia, Food Security Survey, 1995/96 season.

Staple grain sales can be highly concentrated among a relatively small number of large and commericialized farmers in the smallholder sector. Table 6 disaggregates smallholder households included in the nationwide surveys into three groups: 1) the largest smallholder sellers of maize who accounted for 50% of the marketed maize output; 2) the remaining households that sold maize during the year who accounted for

18

the other 50 percent of the marketed output, and 3) those households that sold no maize during the 12-month marketing season.

As shown in Table 6, one or two percent of the farms account for 50% of the overall marketed maize surplus from the smallholder sector. These farm households appear to enjoy substantially better living standards, in terms of asset holdings, crop income, and non-farm income, than the rest of the rural population. The relatively “elite” smallholder farmers had roughly 2 to 5 times as much land and productive assets as the non-selling households, 2 to 7 times as much total household income, and 3 to 8 times more gross revenue from the sale of all crops.

These findings hold several important policy implications. First, maize producer price supports or stabilization policies that involve altering mean price levels over time (as they usually do), can have unanticipated income distributional effects that run counter to stated poverty alleviation goals. To the extent that the poor are net purchasers of staples such as maize, wheat, and rice, they are directly hurt by policies that raise prices of these commodities.7 Mean-neutral forms of price stabilization would most likely avoid these adverse distributional effects, and would also help to promote diversification toward higher-valued crops by maize purchasing households (Fafchamps, 1992).

A second implication of the substantial differentiation within the smallholder farm sector (not to mention the extreme differentiation between large-scale and small-scale farm sectors) is that the benefits of mean-raising food price stabilization policies are likely to be extremely concentrated. This was a major outcome of the price support and stabilization policies pursued during the pre-liberalization period. Jayne and Rukuni (1993), using data on maize purchases by Zimbabwe’s Grain Marketing Board (GMB) between 1985/86 and 1991/92, found that 1 percent of the nation’s smallholder households accounted for 44 percent of all the maize delivered to the Board by smallholder farmers (Table 7). These 9,000 households sold an average of 28.2 tons per year to the Board. Another 80,000 households (the next 9 percent of smallholder households in terms of maize sales) sold an average of 3.4 tons, accounting for 26 percent of the smallholder sector’s maize deliveries to the GMB. Of the remaining 800,000 smallholder households in the country, only 24,000 sold any maize, and those that did so accounted for 4 percent of the total maize delivered to the GMB by the smallholder sector. Of course, the total smallholder sector of 900,000 households received only 54 percent of the government outlays on maize purchases over this 7 year period, as 4,000 large-scale farmers received the rest.

7 Of course, a general equilibrium approach, taking into account indirect effects on welfare through labor market effects, would need to be undertaken before the welfare effects of mean-altering price policies could be fully understood.

19

Table 6. Characteristics of smallholder farmers classified by participation in the maize market, Zambia (2000/01), Mozambique (2002/03), and Kenya (1999/00).

Maize sellers Households not Farms accounting for top 50 percent

Rest of maize

selling maize

of total maize sales sellers

Number of households Zambia (weighted)

Mozambique (weighted) Kenya (unweighted)

Land holding size (hectares) Zambia

Mozambique Kenya

Value of farm assets (USD)a

Zambia Mozambique

Kenya

Total household income (USD) Zambia

Mozambique Kenya

Total crop income (USD) Zambia

Mozambique Kenya

Gross revenue, crop sales (USD) Zambia

Mozambique Kenya

Gross revenue, maize sales (USD) Zambia

Mozambique Kenya

Off-farm income (USD) Zambia

Mozambique Kenya

(1) (2) (3)

23,680 (2.2%) 234,988 (23%) 762,526 (75%) 4,654 (1.0%) 654,771 (15%) 2,466,572 (83%)

25 (1.7%) 535 (37%) 897 (61%)

------------------- Mean values ---------------------

6.00 3.91 2.79 3.46 1.70 1.60

11.09 2.77 1.56

1,558 541 373 205 47 62

6,168 1,107 617

2,282 514 291 2159 315 328

8,849 2,357 1,565

1,348 502 233 1247 176 114

5,479 1,147 628

823 135 36 715 47 20

5,318 831 419

690 74 0 509 20 0

3,474 162 0

1,350 354 291 542 52 77 666 778 638

Notes: alivestock plus farm equipment except for Mozambique, which is livestock assets only. Source: Supplemental Survey to the 1999/00 Post-Harvest Survey, Central Statistical Office, Lusaka, 2001.

20

Table 7. Concentration of Income from Maize Sales to GMB (1985/86 to 1991/92 marketing year)

Farmer category

Total number (approx.)

number of farmers that sell maize to GMB

tons

GMB maize purchases

(annual average) tons per farm

tons per all farms

% of total GMB expenditures on maize purchases

accruing to

(A) (B) (C)

selling maize (D)

within category (E) (F)

Large-scale farms 4,000 1,652a 490,902 297.2 122.7 46

54

9,000 9,000

81,000

24,000

28.2

3.4

2.0

24

26

4

Smallholder households 900,000 114,000 577,686 5.1 0.6 of which:

top 1% of maize sellers

next 9% of maize sellers

remaining households

81,000

810,000

254,182

275,556

47,948

28.2

3.4

0.06

All farms 904,000 115,652 1,068,588 9.3 1.18 100

Note: a based on 1985/86 to 1990/91 marketing year Sources: estimates of commercial and smallholder farm households in Row A are from CSO (1985); Rows B and C are from GMB (various years).

Row D= C/B Row E= C/A Row F=C/total GMB maize purchases

21

When a broader set of staples are aggregated together (maize, cassava, sweet potato, millet and sorghum) more than 55 percent of the sales of staples are still accounted for by 10 percent of the farmers with the largest sales. This concentration of surplus production and marketing by a relatively few farmers is one of the most important points to be borne in mind when thinking about the effects of policy instruments designed to alter the mean level of food prices.

3.3 Declining importance of maize in small farm income and in urban consumers’ diets

In the 1980s, fueled by pan-territorial support prices, and fertilizer subsidies, white maize accounted for half or more of area cultivated on smallholder farms (Byerlee and Eicher, 1997). However, since at least the late 1990s, cropping and consumption patterns in eastern and southern Africa appear to have become more diversified. In Kenya, Mozambique, and Zambia, for example, white maize now accounts for only 13, 14, and 28 percent of gross revenue from crop sales on smallholder farms nationwide (Table 8). Gradually shrinking landholding sizes over the past decades have led to shifts in cultivation toward crops that provide greater caloric value per unit of land (e.g, cassava) and higher valued crops (e.g., horticultural crops). The elimination of pan-territorial pricing on maize in the region has also diversified cultivation patterns. Cassava and sweet potato now account for about half of the value of maize production in Zambia (Govereh et al., forthcoming). Gross revenue from horticultural crop sales are roughly equal to that of maize in Zambia. In Kenya gross revenue from tea, horticulture and dairy sales all exceed that of maize.

Recent surveys of urban households also reveal some new trends, notably the rapid rise in wheat and rice consumption in urban areas of Africa. The rising importance of wheat and rice has probably moderated the effects of variability in coarse grain prices. Recent consumption survey data from Nairobi, Kenya indicates that wheat expenditures now exceed those of white maize, while rice has further diversified consumption patterns. Maize expenditure shares range from 1% among high-income urban groups to only 10% among the poorest segment of Nairobi (Muyanga et al, 2005). A recent survey of urban and rural households in South Africa’s Eastern Cape Province indicates that maize accounts for only 23% of staple carbohydrate expenditures, which is about half that of wheat (Traub and Jayne, 2005). Rice has also made major inroads into the diets in consumers along the coastal areas of eastern and southern Africa. In urban areas of southern Mozambique, for example, rice and wheat expenditures both exceed that of maize (Ministry of Planning and Finance, 2002). The rising diversification of diets, particularly in urban areas, can make it easier for households to stabilize their food expenditures through substitution effects. Importation of one commodity can help stabilize prices of other food commodities, even those that are thinly traded on world markets. For the past few decades, the dominant role of maize in rural consumption and incomes and in urban consumption has resulted in ag policy in southern and eastern Africa to a large extent being equated with and driven by maize policy. However, while maize remains a very important crop in the region, the recent rural and urban survey evidence suggests that a dominant role for maize in ag policy is perhaps outdated.

22

Table 8. Farm Production and Sales Patterns, Smallholder Sector, Kenya (1999/00), Malawi (1998/99),a Mozambique (2001/02), and Zambia (1999/00 and 2003/04).

Crop % of households cultivating

Total National Value

% of total value of farm output

% of total value of farm gross revenue

from sales Maize

Kenya 98 20.2 13.3 Malawi 51 na 2.3

Mozambique 80 31.6 13.8 Zambia 79 45.5 28.2

Sorghum/millet Kenya 50 5.5 6.0 Malawi 4 na 0.1

Mozambique 38 3.3 0.3 Zambia 14 3.7 0.1

Beans/groundnut/oilseeds Kenya 96 4.2 1.9 Malawi 27 na 4.7

Mozambique 72 10.4 9.0 Zambia 52 10.5 7.6

Roots and tubers Kenya 81 5.0 3.0 Malawi 7 na 0.1

Mozambique 85 29.6 6.2 Zambia 54 16.8 5.4

Non-food cash crops Kenya 52 24.2 34.0 Malawi 19 na 88.9

Mozambique 12 6.2 16.9 Zambia 7 5.5 16.1

Fruits and vegetables Kenya 98 17.1 14.7 Malawi na na na

Mozambique 79 13.5d 30.4 Zambia na 17.9d 27.5

Livestock products Kenya 97 23.1b 26.7c

Malawi na na na Mozambique 29 5.6 23.4 Zambia na na 14.7

Notes: na means data not collected. Crops included in surveys listed in Appendix 1. adoes not include fruits, vegetables, or livestock products. b includes production of milk only. cincludes milk, eggs, and live animals. dbased on assumption that production value is double the value of gross sales (only sales data was available in this survey. Data on livestock production and sales is not available.

23

3.4 The Challenge of Increasing Labor Productivity when Farm Sizes are Declining

One of the most salient trends in African agriculture is declining land holding size. Between 1960 and 2000, according to FAO data shown in Table 9, the amount of arable land under cultivation (including permanent crops) has risen marginally, but the population of households engaged in agriculture has tripled. This has caused a steady decline in the ratio of arable land to agricultural population. In Kenya, Ethiopia, and Zambia, for example, this ratio is about half as large as it was in the 1960s. These trends suggest that farming will be increasingly unable to sustain the livelihoods of many rural households without substantial productivity growth on available land as well as shifts in labor from agriculture to non-farm sectors. Education, which played an important role in Asia by allowing households to exit agriculture into more lucrative off-farm jobs, is relatively low in most areas of rural Africa by world standards. Investments in rural education and communications are likely to become increasingly important to facilitate structural transformation. The challenge is how to raise labor productivity given the fact that land available per worker has been declining significantly.3

Table 9. Land to Person Ratio (10 year average) in Selected Countries 1960-69 1970-79 1980-89 1990-99

Ethiopia 0.508 0.450 0.363 0.252

Kenya 0.459 0.350 0.280 0.229

Mozambique 0.389 0.367 0.298 0.249

Rwanda 0.215 0.211 0.197 0.161

Zambia 1.367 1.073 0.896 0.779

Zimbabwe 0.726 0.664 0.583 0.525

Sources: FAOStat website: http://faostat.fao.org/Note: Land to person ration = (land cultivated to annual and permanent crops) / (population in agriculture).

By identity, the monetary value of crop output per unit of labor (Y/L = labor productivity in agriculture) can be partitioned into the ratio of land to labor (A/L), which is a proxy for per capita farm size, and the monetary value of crop output per unit of land (Y/A):

(1) Y/L ≡ A/L × Y/A

Trends in subdivision of land and limited potential for area expansion are causing acute land pressures in many parts of Eastern and Southern Africa. With most arable land already under production, population growth is causing a decline in land/labor ratios (A/L). From Equation (1), it is clear that labor productivity growth, in the face of declining

Labor productivity closely tracks per capita incomes and is probably the most relevant indicator of productivity from a welfare perspective. Output per hectare “is important only as a vehicle for raising labor productivity" (Timmer 1988).

24

3

land/person ratios, will require an even greater rate of growth in the value of output per unit of land.

There are two sources of this potential growth in Y/A. One comes from productivity-enhancing technology, which requires, among other things, well-funded and well-organized agricultural research and extension systems. The other source of growth in Y/A comes from reallocating land from activities with relatively low values per acre to those with relatively high values. Small farmers’ ability to do this is contingent not only on the performance of input, output, and credit markets for high-return crops, but also the reliability of food markets. Survey evidence indicates that smallholder farmers in the region are seeking opportunities to raise Y/A by shifting into higher-valued activities – to the extent that they have the ability to do so. Crops such as tea and horticultural crops are accounting for an increasing share of crop income in many parts of Kenya. Cross-sectional village fixed-effects model results in Jayne et al (2005) show that the proportion of crop area and crop sales from horticulture is significantly inversely related to household landholding size, suggesting that land constraints are altering crop choice, presumably toward crops with higher returns to scarce land. However, these opportunities are impeded by measures that raise the costs and/or risks of acquiring staple food through markets. The higher the price of food, and the greater the price variability during the lean season, the greater the incentives to revert to self-provisioning of food staples (Fafchamps, 1992; Jayne, 1994), with consequences for the growth in land productivity (Y/A), and ultimately labor productivity (Y/L).

An important role for government in supporting agricultural land and labor productivity growth is to provide a policy environment conducive for investment in such high-value, high-return commodities, support R&D and extension, and nurture the development of a reliable and low-cost food marketing system to serve rural areas. Given that government resources are scarce, policies to raise maize price levels have an opportunity cost, and these costs would need to be weighed carefully against the anticipated payoffs from other public investments. Unfortunately little empirical evidence exists (that we know of) that assesses the cost-effectiveness of price stabilization in Africa vs. other approaches to achieve the same productivity growth and food security objectives, such as investing in transportation infrastructure to reduce the costs of input delivery and output marketing, seed research and extension services to raise the marginal value product of using fertilizer, basic education, and the like. However, in light of patterns of concentration of marketed surplus presented in the previous section, there are strong reasons to believe that expenditures on agricultural technology, farmer organizations, credit for small farms, policies to raise smallholders’ access to land, and market infrastructure would more directly benefit smallholder farmers in the bottom half of the income distribution and contribute more to rural poverty reduction objectives than output price supports (Hazell, 2005). Implementing this broader agenda of public investments to support pro-poor agricultural growth will, by themselves, stretch government and donor resources to the limit. But, as Hazell warns, the future for small farms will depend on mobilizing the support for such investments. Thus, the question for state maize price stabilization or price support is not whether these policies can generate positive benefits, but rather whether such benefits could reasonably be expected to exceed the payoffs to alternative forms of using limited donor and government resources.

25

3.5 Segmentation between formal markets and informal food markets

The maize marketing systems in much of eastern and southern Africa appear to be increasingly segmented into two channels that are poorly coordinated with each other. On the one hand, we see “formal” marketing channels linking commercial farmers (mainly in South Africa) and international suppliers to large grain trading, processing, and retailing firms with subsidiary distribution networks throughout Southern Africa. This marketing system is characterized by:

• commodity exchanges, including futures and options markets, enabling farmers and marketing agents to reduce risks of current and future investments;

• a network of integrated silos, millers, and supermarket retailers, often with transnational firm ownership;

• market information accessible on a daily basis, some of which is public, and some which is proprietary, providing asymmetric information advantages for those willing to pay;

• large transaction volumes, which enables transaction costs to be spread over greater quantities traded, hence reducing per unit marketing costs;

• well-specified grades and standards to allow for remote contracting by commodity specification rather than by visual inspection;

• legal systems to accommodate more sophisticated contracting arrangements and facilitation of contract disputes;

• organized lobbies representing firms widely perceived as having a legitimate interest and voice in the determination of regulations governing agricultural markets.

By contrast, the “informal” marketing systems in the region, on which most small-scale farmers rely, are generally characterized by:

• spot market transactions with weak mechanisms for market-based risk management; • small percentages of production sold off the farm, resulting in relatively thin markets

and high transaction costs per unit traded; • weak road and communications infrastructure, resulting in high transportation costs; • weak information systems for reporting local market conditions • processing of maize, either at home by consumers, or by low-cost small-scale mills

not integrated with other stages of the marketing system; • limited coordination between input delivery, farm finance, and crop sale, resulting in

part from poorly functioning input credit systems; • Small businesses with relatively little political influence or voice in the determination

of regulations governing the agricultural sector.

The future of the small-scale farming sector’s ability to prosper from maize production and marketing will depend on strengthening the performance of the marketing system serving small-scale farmers, and on integrating the informal marketing system with the more

26

developed “formal” marketing channels that are rapidly expanding in the region.8 Meeting this market development challenge is crucial not only for small-scale farmers’ as sellers, but also as purchasers of food.

Applied research in the region has shown the increasingly adverse impacts on rural and urban consumers’ food insecurity resulting from the segmentation of these marketing channels. For example, the Zambian government has in the past frustrated private imports during food shortages by sending confusing signals to markets. During the 2001/02 food crisis in southern Africa, the government announced its intention to import 200,000 metric tons of maize grain to cover a national deficit, and to sell that grain at below market prices directly to a small number of large formal sector millers. Given this announcement, other potential private importers, including informal traders from Mozambique, held off (Mano et al., 2003).9 When government instead imported only 130,000 metric tons and did so very late in the season, prices rose steeply, since this amount was insufficient to meet demand (Nijhoff et al, 2002). Moreover, because grain was channeled only to large millers (rather than released onto informal public markets), consumers had to pay the high price of refined meal instead of being able to source grain in informal markets and mill it more cheaply through the network of informal sector hammer mills.10

Another example of how informal marketing channels serving smallholder interests are disadvantaged by government behavior concerns the Malawian government’s tendency to arrange imports through government contracts with South African suppliers. The sourcing of grain from South Africa and subsequent release onto local markets has frequently depressed the informal trade from Mozambique. Since Mozambican smallholder farmers are the source of informal market trade to southern Malawi, the Malawi government’s preference for arranging imports through South Africa has almost certainly added greater risks and price instability for smallholders relying on informal markets for their incomes (Nijhoff et al., 2003).

While there is widespread acceptance of the need to make food markets work better for smallholders (including those who sell grain as well as those who are dependent on the market for their food requirements), policy makers and donors alike need a greater empirical understanding of the interplay between formal and informal markets and their implications for policies designed to stabilize markets and promote smallholder welfare.

8 Our premise is that while developing markets for higher-valued crops is crucial for improving smallholder income and food security, this approach should be viewed as a complement rather than an alternative to the development of reliable food marketing systems to serve smallholders. Research has shown that smallholders’ ability to diversify into high-valued non-food crops depends crucially on the ability of food markets to ensure reliable supplies at tolerable prices (e.g., Omamo, 1998; Jayne 1994). 9 In the 1997/98 and 2000/01 seasons, Tshirley et al (2004) estimate that informal maize imports from Mozambique were on the scale of 150,000 to 200,000 tons. 10 Jayne et al (1996), studied the cost differences of refined maize meal supplied through formal sector channels and the less-refined meal available through hammer mills in informal markets in five countries in Southern Africa during the mid-1990s. They find that the hammer-milled meal could generally be obtained at 65-80% the cost of meal provided through formal sector outlets. However, when public markets become thinly traded and informal traders are not able to continue supplying these markets, consumers lose the option of hammer milled meal and have become reliant on the more expensive formal sector channels for their maize meal.

27

4. Conclusions and Potential Options for the Future

Agricultural market reform in Africa over the past twenty years has been undertaken in the context of aid conditionality. Yet in much of eastern and southern Africa, where white maize remains the primary staple food, the maize marketing systems retain fundamental similarities to the controlled marketing systems of their earlier histories. Some aspects of policy change have been implemented, primarily the legalization of domestic private trading, and marketing board activities have been downsized in response to the unavailability of funds to continue trading at levels during their controlled marketing periods. Instead of purchasing the entire marketed surplus, as was the goal during the former control period, these boards now attempt to influence market prices through their operations in the market, ostensibly for food security and/or price stabilization purposes. Since the reforms began, marketing boards in Kenya, Malawi, Zambia, and Zimbabwe have handled between 10-50 percent of the marketed maize from domestic production in most years. In countries where marketed surpluses are falling and national food security relies increasingly on imports (e.g., Malawi), marketing board’s sales of maize become a more important indicator of their presence in the market than purchases. Despite the quite significant role that marketing boards in these countries continue to play up to the present, maize price volatility and its potential effects on production incentives and food security remain critical concerns.

Perhaps the greatest irony of the aid conditionality process in the region is the widespread perception that the World Bank has forced these African governments to implement orthodox agricultural policy reform, and that the lack of clear economic turnaround in the region casts doubt on the technical logic of the Bank’s model. The weight of the empirical evidence indicates that many countries in eastern and southern Africa have continued food price stabilization cum subsidy programs of various types, and hence an empirical assessment of these countries’ food market performance since the 1990s reflects not the impacts of unfettered market forces but rather the mixed policy environment of legalized private trade within the context of continued strong government operations in food markets.

Although price stabilization could have important benefits for producers and poor consumers, these benefits do not appear to have been successfully achieved by the existing mix of import tariffs, sporadic export bans, and marketing board operations to influence producer and consumer prices. Maize price instability in countries like Malawi and Zambia are extremely high despite the persistence of these government operations. While it is extremely difficult analytically to estimate the counterfactual – i.e., the level and instability of maize prices that would have prevailed over the past 15 years in the absence of these government operations – there are strong indications that at least some aspects of government interventions in the market have exacerbated rather than reduced price instability for both producers and consumers. In some countries in the region, government policy has tended to raise maize market prices, generating distributional effects that were most likely anti-poor. While the full general equilibrium effects of government price policy, including their effects on the labor market, have rarely, if ever been estimated for this region (primarily because the data is not available to adequately do so), the

28

information on small farm production and marketing patterns presented earlier in this paper suggest that mean-raising price policies are likely to have very concentrated benefits among relatively large farmers and would constitute a direct tax on consumers, many of whom are small farmers living in rural areas. Knowing the potential distributional consequences of any policy influencing staple prices would need to take into account knowledge of land allocation patterns in smallholder farming systems, the concentration of marketed food output, and trends toward smaller farm sizes and increased land constraints as population pressures intensify. This information will seriously affect the costs and benefits of alternative approaches to price stabilization, particularly those that are likely to alter the mean as well as the variance of food prices.

4.1 Potential Approaches to Make Markets Work Better for Smallholder Farmers

Improving government market interventions as part of the transition to private sector-based marketing

If government market interventions are to remain in some form, as part of the transition to private sector-based marketing, it seems clear that they would be more effective and more supportive of market development if they followed well-defined, consistent, and transparent rules. In those countries where government involvement in the staple food market is seen as part of a transitional phase towards full market liberalization (e.g. Zambia and Malawi), predictable and transparent rules governing state involvement in the markets would reduce the risks facing private traders, which would facilitate greater coordination between private and public decisions in the market. Also, government interventions need to be consistent with the resources that are available. Overstating government import intentions has in the past led the private sector to conclude that it had no role to play in importation, which contributed to the situation where prices rose above import parity levels as they did in Malawi in 2001/02 and Zambia in 1999/00 (Rubey, 2004; Nijhoff et al, 2003).

Institutional/organizational innovations to improve vertical coordination

Production costs advantages that smallholder farmers typically have in food production are often offset by weak market access – for credit, inputs, and commodity sales. Because most smallholders have small quantities to sell, they involve high transaction costs for traders unless farmers can bulk up their sales and sell as a group. Forms of group credit for purchasing agricultural inputs have also shown some success (Tschirley, Zulu, and Shaffer, 2004). Outgrower arrangements, involving the interlocking of credit, inputs, and output sale, have sometimes been successful in improving vertical coordination in commodity supply chains where the output market could be controlled, which seldom includes food crops (Dorward, Kydd and Poulton, 1998). Cash crop outgrower schemes providing credit to participating farmers have also generated important synergies and spillover benefits for smallholder food crop production (Dione, 1989; Jayne, Yamano and Nyoro, 2004). While farmer organizations may provide only limited scope for stabilizing commodity prices for farmers, they have shown great potential to provide a stable stream of income, help

29

smallholders engage in higher-return crop activities, and promote farm productivity and food security, which may be of greater importance that price stabilization per se. There would appear to be significant unexploited potential to expand the role of outgrower arrangements and farmer organizations in the region, both as a means to reduce the risks and transaction costs between small farmers and trading firms and to promote smallholder incomes from an integrated credit-input-crop marketing system that would benefit maize production as well as selected cash crops.

Public goods investments

Many agricultural market failure problems in Africa reflect an under-provision of public goods investments to drive down the costs of marketing and contracting. The existence of market failures should not imply that direct government operations in food markets is necessarily required to solve these market failures. Rather, ameliorating market failure may have more to do with increased commitment to investing in public goods (e.g., market infrastructure, R&D, education, health, agricultural extension systems) and institutional change to promote the functioning of market-oriented trading systems.

Yet there is sometimes a tendency to downgrade this approach to promoting market stability because of fears that such investments would not provide a visible enough demonstration of governments’ commitment to smallholder constituents and because of uncertainty over the time frame of future benefits. For example, there is little information to estimate how long it would take for improved systems of contract enforcement and collateral recovery in a given country to raise the supply of commercial agricultural credit for small farmers and traders. Likewise, little analysis has been devoted to estimating the stream of benefits over time that would be generated from greater investment in market infrastructure to reduce the costs of delivering fertilizer to remote smallholder areas, or to the payoffs generated from improved mobile phone communication in rural areas and its effects on investment in crop marketing and efficiency in spatial arbitrage.

In a pluralistic political system, policy makers need to show that they are taking immediate action to promote rural constituents’ welfare. A key issue for donors is to determine what can be done to reduce policy makers’ preferences for immediate and visible signs of “doing something” to benefit their constituents in the short run (which may have little lasting impact on development and rural welfare) vs. making investments that may have lasting positive development impacts but which might not clearly manifest until after the leaders in power leave office.

Market risk shifting mechanisms

This issue is not discussed here because it is the topic of several companion articles to be presented at the workshop. However, so as not to suggest that market risk-shifting tools (such as warehouse receipt systems, commodity exchanges offering spot and forward contracts, and ultimately perhaps futures and options markets) are not part of the tool kit to

30

help stabilize food markets in the region, we simply highlight this class of policy instrument here. However, we caution that viable market-oriented risk transfer mechanisms would be unlikely to develop in an environment in a market environment where one actor (e.g., the government) had the power and proclivity to influence price levels in a discretionary way, as this would mean that certain actors would have an information advantage in the purchase or sale of commodity instruments and could exercise that advantage to the disadvantage of others. This would have obvious implications for a strategic food reserve or public buffer stock program. As stated by Oygard et al. (2003), “unless some very predictable and credible management rules can be established for the reserve, private agents will be reluctant to hold stocks, out of a fear that the reserve will be sold out at unpredictable times at subsidized prices, undercutting the value of their stored commodity.”

4.2 Specific actions for governments to consider to enable maize marketing and trade system to facilitate pro-poor agricultural growth

One specific proposal would be for governments in the region to organize a regional body to coordinate with Africa-wide initiatives such as NEPAD/CAADP, the Partnership to Cut Hunger in Africa, and IHEA to lobby for pro-poor investments that international donors and lenders could feasibly support. Such a regional body could encourage donors to step up financial support for the following:

1. performance contracts with international seed companies to work with national and regional agricultural organizations to develop improved maize seed technology relevant for the semi-arid areas that characterize much of eastern and southern Africa (Lipton, 2005; Bagwati, 2005).

2. rehabilitation and expansion of port, rail, and road infrastructure within the region (e.g., the proposed Sudan-Kenya-Uganda railway system currently under discussion). Because much of the maize price instability problem, and its associated effects on smallholder production incentives and food insecurity, is related to high costs of transport within the region and between the ports and major production and consumption areas, the reduction of transport costs would go a long way to improving the stability of maize prices and supplies in the region. While such investments would take years to put into place, it is clear that such investments must be part of an overall pro-poor productivity growth strategy for the region. Donor development assistance for physical infrastructural development could play a major supportive role to the future of smallholder agriculture (for maize as well as many other crops).

3. support for public extension systems. Household survey data indicates that, within a given community or district, maize yields and productivity are highly variable across households (e.g., Nyoro et al., 2004). The variation in maize production costs, even controlling for production technologies, tends to be very high. This suggests that variations in management practices and husbandry skills are probably very great. This result underscores the importance of appropriate extension messages. Simply by bringing

31

the relatively high-cost producers to the mean in a particular area, the overall production costs of maize production could decline significantly and improve smallholder incomes and food security. Donors could once again play an important supportive role in this regard. Even indirect support, e.g., funding for soil testing, developing recommendations for fertilizer application rates that take into account the micro-variability in soil, rainfall, and market conditions, could be a big help.

4. HIV/AIDS mitigation, and the need for a better understanding of effective responses to the crisis that will support smallholder agriculture. The extent to which the AIDS crisis is hindering smallholder maize production in particular, and livelihoods in general, is not fully understood, but it is likely to be great. Donors are responding in many ways, but the effects of some programs are not clear. For example, while some donor programs are designed to inject food aid into communities hard-hit by the disease to help afflicted households, it is not clear what effect this will have over time on maize production incentives or price instability. Coming up with clear guidelines for devoting scarce resources to effectively combat HIV/AIDS while minimizing potential disruptions to local livelihood patterns is an important contribution that donors can make to smallholder agriculture.

5. land redistribution and/or new settlement of remote but productive land. As indicated earlier in Section 3, landholding sizes in the smallholder sector are declining over time in many areas as population pressures intensify. Fully one-third of smallholder farms in the region are now under one hectare. Opportunities for pro-poor agricultural growth, especially with food crops such as maize as the foundation, will be highly constrained for most small farmers. Supply response potential for farmers cultivating less than one hectare are indeed limited, even if new technology were able to substantially raise maize productivity. Shifts to higher-valued crops will afford some opportunities for growth, but inadequate access to land is likely to remain an overriding constraint on pathways out of poverty for a large part of the rural population unless efforts are made to deal with the land issues. In some cases where the distribution of land is highly concentrated and where little opportunity for land expansion exists, land redistribution may be critical.

Where there currently exist large tracts of “remote” but productive land, the public sector can play an important role in raising the returns to settlement in such areas through service provision and infrastructural development to better link currently isolated areas with existing road and rail infrastructure. This may help reduce the current population pressure in areas of relatively good access to markets through migration to areas of low population density but which have experienced an increase in the economic value of land through public investments in infrastructure and service provision. Such an approach was pursued successfully by Southern Rhodesia and Zimbabwe starting in the 1960s with its “growth point” strategy in the Gokwe area, once cleared of tse tse fly. Key public investments in this once desolate but agro-ecologically productive area induced rapid migration into Gokwe from heavily populated rural areas, leading to the “white gold rush” of smallholder cotton production in the 1970s and 1980s (Govereh, 1999).

32

These five specific areas for government action constitute a tall agenda. Implementing it will require close dialogue and coordination with international lenders and donors, not only to help with financial support, but also to help in working out the details of implementation, including the detailed “how” questions. By taking the initiative and engaging donors through African-driven initiatives like NEPAD and CAADP, governments in the region can show real commitment to this agenda. This agenda can go forward with implementation without necessarily needing to resolve at the same time the more thorny issues such as fertilizer subsidies and marketing board price stabilization policies. Getting consensus and action on part of the agenda would most likely be strongly preferable to having the whole process stalled over a few contentious issues.

4.3 Concluding remarks

Maize will remain a crucial part of the food security equation in two ways: first, as a purchased commodity for satisfying the food requirements of a more diversified rural economy, and second, as a cash crop in areas where it is agro-ecologically suited to provide high returns. Rising land constraints will progressively encourage farmers to shift toward crops providing high returns to scarce land. Because most parts of Africa are experiencing increased land pressure and limited potential for area expansion, population growth is causing a decline in land/labor ratios and farm sizes are declining. Maize is a relatively low value-to-bulk crop, providing consistently high returns per unit of land in only a relatively small proportion of smallholder farming areas in the case study region (e.g., Kenya’s North Rift, Zimbabwe’s Mashonaland maize belt, parts of Zambia’s Central and Southern Provinces). Given reasonable assumptions about future productivity improvements, it is unlikely that maize can provide the net revenue on the millions of farms that are 0.5-1.0 hectares or smaller to generate substantial income growth, especially in the semi-arid areas.

States will therefore need to invest in more efficient private systems of input delivery, finance, and commodity marketing, not only for maize but for a range of crops that offer higher returns to farming in the changing environment of Africa’s rural areas. These include sugar in the lowlands of western Kenya, tea in the highlands, cotton in semi-arid parts of Zimbabwe, Zambia, and Kenya, and horticulture where water control is possible and where transport infrastructure to export markets can be developed. Investments in education and skill development, market infrastructure, and agricultural research programs for a more diversified set of crops will be necessary.

Such investments would represent a shift from the strategy of food self-sufficiency to one of comparative advantage. While such a shift will be crucial for poverty alleviation for millions of small farms in these countries, it is not assured. Governments have an important role to play in building the institutions to reduce the transaction costs and physical costs of trade so that households can benefit from the higher incomes afforded by a well-functioning commercialized agricultural sector while still ensuring their access to food.

33

The issue of how to stabilize food markets and prices are transcended by issues of governance. The aims of promoting producer and consumer welfare can be promoted – in principle – through either direct government operations or through private trade. In actual experience, neither approach has worked very well. Effective governance is central to the effective operation of both state enterprises and markets. Marketing boards have a mixed track record in Africa. But attempts to rely on markets, given a chronic under-provision of public goods investments, often fail too. The evidence seems clear that, without increased government and donor support for public goods investments to drive down the costs of production and marketing, the future for smallholder farms in Africa is not good. But such investments will also need to be complemented by a policy environment that nurtures the development of historically neglected informal marketing systems on which most smallholder farmers depend, including the integration of these markets with the vertical channels of the formal marketing systems that are rising in importance in the region. Governance will make or break these possibilities.

The most vital question is: can a local constituency be formed that can stake a claim on public resources in support of agricultural research, marketing institutions, and other kinds of growth-promoting public goods? There is an obvious connection between agricultural development and governance. The early success of the maize industry in Zimbabwe and Kenya can be largely attributed to the strength of the institutions built by settler farmers, which mobilized a constituency to support public and private investments. Today, farm lobbies are generally weaker and more fragmented. Representation has always been weak for smallholder farmers, particularly when their welfare is closely tied to the reliability and efficiency of maize markets where they purchase maize as consumers. How will growth- and equity-promoting investments in agricultural research, infrastructure, and market institutions be financed? Where will the domestic political pressure for these public investments originate?

34

References

Amani, H.K.R. and W. Maro, 1992. Policy to Promote an Effective Private Trading System in Farm Products and Farm Inputs in Tanzania, in J. Wyckoff and M. Rukuni (Eds.), Food Security Research in Southern Africa: Policy Implications, Harare: University of Zimbabwe.

Bates, R.,1981. Markets and States in Tropical Africa: The Political Basis of Agricultural Policies. Berkeley: University of California Press.

Bates, R.H, 1989. Beyond the Miracle of the Market: The Political Economy of Agrarian Development in Kenya, Cambridge: Cambridge University Press, 1989.

Bird, K., D. Booth, and N. Pratt, 2002. The Contribution of Politics, Policy Failures, and Bad Governance to the Food Security Crisis in Southern Africa. Forum for Food Security in Southern Africa, Poverty and Public Policy Group, Overseas Development Institute, London.

Brunetti, A., G. Kisunko, and B. Weder, 1997. Institutional Obstacles to Doing Business: Region-by-Region Results from a Worldwide Survey of the Private Sector. Policy Research Working Paper 1759, Washington D.C., World Bank.

Bryceson, D., 1993. “Urban Bias Revisited: Staple Food Pricing in Tanzania,” in Real Markets: Social and Political Issues of Food Policy Reform, Hewitt de Alcantara, C. (ed.), London: Frank Cass.

Buccola, S., and C. Sukume, "Optimal Grain Pricing and Storage Policy in Controlled Agricultural Economies: Application to Zimbabwe," World Development, Vol 16, No. 3 (1988): pp. 361-71.

Chimedza, R. 1994. Rural Financial Markets, in M. Rukuni and C. Eicher (Eds.), Zimbabwe's Agricultural Revolution, Harare: University of Zimbabwe Press, pp. 139-152.

Datt, G. and M. Ravallion. 1998. Farm Productivity and Rural Poverty in India. Journal of Development Studies, 34 (4): 62-95.

Dione, J., 1989. Informing Food Security Policy in Mali: Interactions Between Technology, Institutions and Market Reforms. Unpublished PhD Dissertation, Michigan State University, East Lansing, MI.

Dorward. A; J. Kydd and C. Poulton.eds. 1998. “Smallholder Cash Crop Production Under Market Liberalization.” Oxon; New York, NY: CAB International.

Dorward, A., J. Kydd, J. Morrisson, and I. Urey, 2004. A Policy Agenda for Pro-Poor Agricultural Growth. World Development 32(1): 73-89.

35

Gabre-Madhin, E., C. B. Barrett, and P. Dorosh, 2003. Technological Change and Price Effects in Agriculture: Conceptual and Comparative Perspectives. MTID Discussion Paper 62, Washington, D.C.: International Food Policy Research Institute. Available from http://www.cgiar.org/ifpri/divs/mssd/dp.htm.

Gugerty, M. and C. P. Timmer, 1999. “Growth, inequality and poverty alleviation: Implications for development assistance”. Mimeo.

Haggblade, S. and P. Hazell. 1988. Prospects for Equitable Growth in Rural Sub-Saharan Africa, AGRAP Economic Discussion Paper 3, Washington D.C., World Bank.

Hazell, P. 2005. Is there a Future for Small Farms? In Reshaing Agriculture’s Contribution to Society, D. Colman and N. Vink (eds.), Oxford, Blackwell Publishing.

Howard, J. 1994. The Economic Impact of Improved Maize Varieties in Zambia," unpublished Ph.D. dissertation, East Lansing: Michigan State University.

Jayne, T.S., and M. Chisvo, 1991. Unravelling Zimbabwe's Food Insecurity Paradox, Food Policy, Vol. 16, No. 5: pp. 319-329.

Jayne, T.S., and M. Rukuni, 1993. Distributional Effects of Maize Self-Sufficiency in Zimbabwe: Implications for Pricing and Trade Policy, Food Policy, Vol. 18, No. 4: pp. 334-41.

Jayne, T.S., 1994. "Do High Food Marketing Costs Constrain Cash Crop Production?," Economic Development and Cultural Change, Vol. 42, No. 2: 387-402.

Jayne, T.S., L. Rubey, D. Tschirley, M. Mukumbu, M. Chisvo, A. Santos, M. Weber,and P. Diskin (1996). "Effects of Market Reform on Access to Food by Low-Income Households," International Development Paper 19, Michigan State University, East Lansing.

Jayne, T.S. and S. Jones, 1997. Food Marketing and Pricing Policy in Eastern and Southern Africa: A Survey. World Development 25(9): 1505-1527.

Jayne, T.S., J. Govereh, A. Mwanaumo, J. Nyoro, and A. Chapoto. 2002. False Promise or False Premise: The Experience of Food and Input Market Reform in Eastern and Southern Africa. World Development 30(11): 1967-1986.

Jayne, T.S. T. Yamano, M. Weber, D. Tschirley, R. Benfica, A. Chapoto, and B. Zulu. 2003. Smallholder Income and Land Distribution in Africa: Implications for Poverty Reduction Strategies. Food Policy, Vol. 28 (3): 253-275.

Jayne, T.S., T. Yamano, and J. Nyoro. 2004. Interlinked Credit and Farm Intensification Evidence from Kenya. Agricultural Economics 31(6): 209-218.

36

Jayne, T.S., R. Myers, and J. Nyoro. 2005. Effects of Government Maize Marketing and Trade Policies on Maize Market Prices in Kenya, report prepared for the World Bank/Africa Region, Washington, D.C.

Jones, S.P., 1994. Privatization and Policy Reform: Agricultural Marketing in Africa. Food Studies Group, University of Oxford, Oxford

Kherallah, M., C. Delgado, E. Gabre-Madhin, N. Minot, and M. Johnson, 2002. Reforming Agricultural Markets in Africa. Baltimore: Johns Hopkins University Press.

Mano, R., B. Isaacson, and P. Dardel. 2002. “Identifying Policy Determinants of Food Security Response and Recovery in the SADC Region: The Case of the 2002 Food Emergency”. Policy Paper prepared for the FANRPAN Regional Dialogue on Agricultural Recovery, Food Security and Trade Policies in Southern Africa, Gaborone, Botswana, 26 – 27 March 2003.

McPherson, M. (ed.), 2002. Restarting and Sustaining Growth and Development in Africa. Equity and Growth through Economic Research, Cambridge, Harvard Institute for International Development.

Meerman, J. P., 1997. Reforming Agriculture: The World Bank Goes to Market. Operations Evaluation Study, The World Bank, Washington, D. C.

Muyanga, M., T.S. Jayne, G. Argwings-Kodhek, and Joshua Ariga. 2005. Staple Food Consumption Patterns in Urban Kenya: Trends and Policy Implications. Working Paper, Tegemeo Institute of Agricultural Policy and Development, Egerton University, Online at : http://www.tegemeo.org

National Economic Council, 2000. Profile of Poverty in Malawi: Poverty analysis of the Malawi Integrated Household Survey, 1997-98, Poverty Monitoring System, National Economic Council, Lilongwe.

Nijhoff, J.J., T. Jayne, B. Mwiinga, and J. Shaffer. 2002. “Markets Need Predictable Government Actions to Function Effectively: the Case of Importing Maize in Times of Deficit”. Policy Synthesis 6, Food Security Research Project. Lusaka. http://www.aec.msu.edu/agecon/fs2/zambia/polsyn6zambia.pdf

Nijhoff JJ., D. Tschirley, T. Jayne, G. Tembo, P. Arlindo, B. Mwiinga, J. Shaffer, M. Weber, C. Donovan, and D. Boughton. 2003. Coordination for Long-term Food Security by Government, Private Sector and Donors: Issues and Challenges. Policy Synthesis No. 65. Michigan State University, Department of Agricultural Economics.

37

Nyirongo, C., H. Gondwe, F. Msiska, H. Mdyetseni, and F. Kamanga, 2002. 2001-02 Targeted Inputs Programme (TIP). Evaluation Module 1: Food Production and Security in Rural Malawi. Submitted to: The Ministry for Agriculture and Irrigation (Malawi) and the Department for International Development (UK).

Nyoro, J., M. Kiiru, and T.S. Jayne, 1999. “Evolution of Kenya’s Maize Marketing System in the Post-Liberalization Era,” paper presented at the Fourth Agricultural Transformation in Africa Workshop, Nairobi Serena Hotel, Nairobi, Kenya, June 27-30, 1999, Tegemeo Institute, Nairobi.

Odhiambo, Mark, and David Wilcock, "Reform of Maize Marketing in Kenya." In M. Rukuni, G. Mudimu, and T. Jayne (eds.), Food Security Policies in the SADCC Region (Harare: Univ. Zimbabwe, 1990).

Omamo, S.W. 1998. Transport Costs and Smallholder Cropping Choices: An Application to Siaya District, Kenya. American Journal of Agricultural Economics, 80 (1): 116-123.

Øygard, R., R. Garcia, A. Guttormsen, R. Kachule, A. Mwanaumo, I. Mwanawina, E. Sjaastad, and M. Wik. 2003. The Maze of Maize: Improving Input and Output Market Access for Poor Smallholders in Southern African Region, the Experience of Zambia and Malawi, Agricultural University of Norway Department of Economics and Resource Management Report No. 26, ISSN 0802-9210.

Pinckney, Thomas C., "Is Market Liberalization Compatible with Food Security? Food Policy, Vol. 18, No. 4 (1993), pp. 325-333.

Putterman, L., “Economic Reform and Smallholder Agriculture in Tanzania: A Discussion of Recent Market Liberalization, Road Rehabilitation, and Technology Dissemination Efforts,” World Development, Vol. 23, No. 2 (1995): pp. 311-326.

Republic of Zambia, "A Review of Maize Marketing Liberalization During 1994: The Transition Programme 1995/96" (Lusaka: Food Security Division, Ministry of Agriculture, Food and Fisheries, 1995).

Rohrbach, D.D., "The Economics of Smallholder Maize Production in Zimbabwe: Implications for Food Security," International Development Paper No. 11 (E. Lansing: Michigan State Univ., 1989).

Rubey, L., 2004. How Well-Intentioned Government Actions Exacerbate Food Insecurity: Two Case Studies from Malawi. report, USAID/Malawi Agricultural Office, Lilongwe.

Sahn, D., P. Dorosh, and S. Younger, 1997. Structural Adjustment Reconsidered: Economic Policy and Poverty in Africa. Cambridge: Cambridge University Press.

38

Sibale, P. K., A. M. Chirembo, A. R. Saka, and V. O. Lungu, 2001. TIP 2000-01. Food Production and Security. Module 1. Submitted to: The Ministry for Agriculture and Irrigation (Malawi) and the Department for International Development (UK).

Tembo, G., J. Shaffer, T.S. Jayne and J.J Nijhoff, 2005, forthcoming. Staple Crop Marketing Patterns in Zambia’s Small- and Medium-scale Farm Sector. Working Paper, Food Security Research Project, Zambia, Lusaka.

Timmer, P. 1988. "The Agricultural Transformation." In Handbook of Development Economics, Vol. 1, edited by H. Chenery and T.N Srinivasan, Elsevier Science Publishers, Oxford.

Toye, J.,1992. Interest Group Politics and the Implementation of Adjustment in Sub-Saharan Africa. Journal of International Development 4(2): 183-198.

Tschirley, D., Nijhoff, J. Arlindo, P., Mwinga, B., Weber, M. and Jayne, T. 2004. Anticipating and Responding to Drought Emergencies in Southern Africa: Lessons from the 2002-2003 Experience. Prepared for the NEPAD Conference on Successes in African Agriculture, 22-25 November 2004, in Nairobi, Kenya. Online at: http://www.aec.msu.edu/agecon/maizemarket/index.htm

Tschirley, D. B. Zulu, and J. Shaffer. 2003. Cotton in Zambia: An Assessment of its Organization, Performance, Current Policy Initiatives, and Options for the Future, Working Paper 10, Zambia Food Security Research Project, Lusaka. Online at: http://www.aec.msu.edu/agecon/fs2/zambia/index.htm

Walker, T., D. Tschirley, J. Low, M. Pequenino Tanque, D. Boughton, E. Payongayong, and M. Weber, 2004. Determinants of rural Income, Poverty, and Perceived Well-Being in Mozambique in 2001-02. Research Report 57e, Ministry of Agriculture and Rural Development, Maputo. Online at: http://www.aec.msu.edu/agecon/fs2/mozambique/wps57e.pdf

van de Walle, N., 2001. African Economies and the Politics of Permanent Crisis, 1979­1999. Cambridge: Cambridge University Press.

Whiteside, M. 2003. Enhancing the Role of Informal Maize Imports in Malawi Food Security. A consultancy report for DfID.

World Bank, 2003. Reforming The Malawi Agricultural Development And Marketing Corporation (ADMARC): Synthesis Report of the Poverty and Social Impact Analysis December 1, 2003, Poverty Reduction and Economic Management 1 Africa Region.

39

Appendix 1: Notes on the National Rural Household Datasets

Kenya

Analysis is based on survey of 1,422 small-scale farming households surveyed in May-June 2000. The survey was designed and implemented under the Tegemeo Agricultural Monitoring and Policy Analysis Project (TAMPA), implemented by Egerton University/Tegemeo Institute, with support from Michigan State University. The sampling frame for the survey was prepared in consultation with the Central Bureau of Statistics, although its agricultural sample frame was not made available. First, 24 districts were purposively chosen to represent the broad range of agro-ecological zones and agricultural production systems in Kenya. Next, all non-urban divisions were assigned to one or more AEZs based on secondary data. Third, proportionally to population across AEZs, divisions were chosen purposively from each AEZ. Fourth, within each division, villages and households within selected villages were randomly selected. As a result, a total of 1,578 households were chosen from 24 districts within the eight agriculturally-oriented provinces of the country.

We excluded two pastoral districts (40 households) that differed substantially from other zones and had high rates of attrition. Of the 1,538 remaining households that we attempted to revisit in the 2000 survey, 1,460 households in six provinces were locatedand re-interviewed, and these households form the basis for analysis in this paper. The crop categories and crops specifically asked about in the survey included the following:

Cereals category: sorghum, millet, wheat, rice, oats, barley, simsim, and bulrush millet Beans: beans, cowpeas, greengrams, snowpeas, pigeon peas, njahi-dolichos, groundnuts, soybeans, greenpeas, and runnerbeans. Roots/tubers: irish potatoes, cassava, sweetpotatoes, arrowroots, yams, and tangawizi Non food cash crops: commercial trees, cotton, sisal, pyrethrum, flowers, tobacco, sunflower, and miraa. Fruits and Vegetables: tamarind, bananas, cowpea leaves, coconut, cashewnuts, frenchbeans, tomatoes, onions, kale, capsicum, etc. Livestock: milk, eggs, live animals.

Mozambique

In 2002, the Mozambican Ministry of Agriculture and Rural Development (MADER) in collaboration with the National Institute of Statistics (INE) conducted a survey of rural households known as the Trabalho do Inquerto Agricola (TIA) 2002. The sampling frame was based on that developed for the 'sample' Census of Agriculture and Livestock 2000 (CAP). The sample was stratified by province (10 provinces) and agroecological zones. Eighty of the country's 128 districts were included in the sample. A total of 4,908 small and medium-sized farms were interviewed in 559 communities that were the primary sampling units. The sample is nationally representative of rural households and

40

representative to the provincial level. Sampled households were interviewed over a period of several weeks in August-September 2002 concerning a range of household farm and non-farm income and activities from the 2001/02 agricultural year. These data on small and medium-sized households were complemented by group interviews at the community level and by area measurements of about 2,500 of the 12,000 fields operated by respondent households.

Zambia

Analysis is drawn from Post Harvest Surveys (PHS) conducted annually by the Central Statistical Office (CSO) and the 1999/2000 Supplementary Survey (SS) to the Post Harvest Survey, also conducted by CSO. The PHS is the still the most comprehensive and statistically valid source of information for the small- and medium-scale farm sector in Zambia (Zulu et al 2000). The SS involved revisiting the same rural households that were interviewed in the 1999/00 PHS with a set of “supplementary” questions which are not normally asked in the regular post harvest surveys. These questions pertained to access to land, information on non-farm income and household socio-demographic characteristics. The SS was conducted in May 2001.

The PHS/SS is based on a sample frame of about 8,000 small-scale (0.1 – 5 hectares) and medium-scale farm households, defined as those cultivating areas between 5 to 20 hectares. Large-scale farmers are not included in this survey. Households were included in the sample only if they were found through initial screening questions to cultivate crops or raise livestock. Because the PHS is an agricultural household survey, by definition, the sample contains no landless households. However, initial village listings to prepare the sample frames for these surveys enumerated all households in these villages. These listings were made available, and the percentage of households who engaged in neither crop nor animal production on their own land was found to be low, less than 4%. Landlessness is somewhat higher in areas closer to towns, where a higher proportion of households are engaged in exclusively off-farm activities.

Malawi

Malawi: Ephraim Chirwa, University of Malawi, Chancellors College, analysis of Malawi Integrated Rural Household Survey, 1997-98, Government of Malawi. Recall data on maize purchases pertains to a 3-day recall period hence computation of maize purchases is not possible.

Ethiopia

The data come from two sources: the 1995/6 Annual Agricultural Sample Survey (ASS), fielded by the Ethiopian Central Statistical Authority (CSA) and the Food Security Survey (FSS), fielded on a subset of ASS households in 1996 by the CSA and the Grain Marketing Research Project. The 1995/6 Agricultural Sample Survey uses the same frame of enumeration areas (EAs) as used to conduct the 1994 Population Census. Some

41

615 rural EAs in 373 weredas are sampled out of roughly 60,000, with probability proportional to population size. In each of the EAs, 25 households are randomly selected, for a total of 15,374 households. Out of these, 7 are randomly sampled to be in the Food Security Survey, some 4,112 households total. The Food Security Survey collected detailed information regarding amounts of food aid received by each household in 606 EAs and 366 weredas, plus other information.

42