Market Liberalisation, Household Food Security and ...

Market Liberalisation, Household Food Security and the Rural Poor in Malawi ALASTAIR ORR, BLESSINGS MWALE and DONATA SAITI Rural poverty in Malawi is correctly identified with the absence of household food security. Household food security, however, is equated with household maize production. Poverty alleviation has thus focused on intensifying maize production through the adoption of hybrid seed and fertiliser. Successive devaluations have now made this technology package too expensive for the majority of smallholders, however. We argue that the focus on maize production overlooks livelihood diversity among the rural poor. Market liberalisation has encouraged this diversity by giving smallholders new opportunities for commercial cropping and off-farm income. As a result, average incomes have risen, despite the decline in per capita maize production. Thus, poverty alleviation in Malawi requires not only efforts to promote new maize technology but also policies that accelerate and widen commercialisation and micro enterprise. Recognising the potential of the market to provide income for the poor will require a broader view of household food security among policymakers in Malawi.

INTRODUCTION Malawi - a small, landlocked nation in east-central Africa - is one of the world's poorest countries. Four out of ten Malawians live below a poverty line based on the basic needs of food, clothing and shelter, unable to lead an active, healthy life [World Bank, 1996; GOM, 1996a]. Rural households make up 80 per cent of the poor. Poverty is deep as well as wide, with income inequality the highest recorded in sub-Saharan Africa [World Bank, 1996]. After the Alastair Orr, Natural Resources Institute, Chatham Maritime, Kent, United Kingdom; Blessings Mwale, Ministry of Agriculture and Irrigation, Lilongwe, Malawi; Donata Saiti, Kadale Consultants, Blantyre, Malawi. This research was funded by DFID and the Ministry of Agriculture and Irrigation (MOAI) of the Government of Malawi. We are grateful to two anonymous reviewers for helpful suggestions. The views expressed here are those of the authors and not necessarily those of DFID or the MOAI. The European Journal of Development Research, Vol.13, No.l, June 2001, pp.47-69 PUBLISHED BY FRANK CASS, LONDON

48

THE EUROPEAN JOURNAL OF DEVELOPMENT RESEARCH

introduction of multiparty rule in 1994, poverty alleviation has become a national priority. Aid flows have increased in consequence and currently average $250 million per year, equivalent to $50 for every man, woman, and child [World Bank, 1998). Malawi is now the second-largest recipient of British aid after India. Poverty in Malawi is closely identified with the absence of household food security. Food is usually defined in terms of maize, the staple food crop. Under its former Life President, Dr Hastings Kamuzu Banda, Malawi was largely self-sufficient in maize and poverty was officially denied to exist. This view was challenged in the 1980s by UNICEF and Malawian social scientists who exposed widespread levels of child malnutrition (Quinn, 1994). In 1987 a Food Security and Nutrition Unit was established in the Department of Economic Planning and Development under the Office of the President and Cabinet. This successfully pushed food security to the forefront of the development agenda [World Bank, 1989]. Today, household food security dominates thinking about poverty and the design of poverty - alleviation programmes in Malawi [Conroy, 1998; Devereux, 1996; Brown et. al., 1996]. Food security is defined as 'the ability of households to produce, purchase, or acquire through public or private transfer an adequate amount of food to meet biological requirements' [GOM/UN, 1993: 109]. In practice, however, food security in Malawi has been viewed primarily as the capacity of rural households to produce more food or to acquire it through public transfer. Relatively little attention has given to raising smallholders' cash incomes so that they can have access to maize through the market. This bias towards household maize production should be seen in historical perspective. Under Dr Banda's regime, smallholders were seen primarily as subsistence farmers and were systematically denied opportunities to earn cash income. They were subject to a wide array of controls on the production and marketing of crops [Ng'ong'ola, 1986]. They were legally barred from growing burley tobacco, Malawi's premier cash crop. The monopsony buying power of the Agricultural Development and Marketing Corporation (ADMARC) kept domestic prices for smallholder crops well below export-parity levels and stifled private trade [Kydd and Christiansen, 1982]. Low producer prices also foiled attempts to raise smallholder incomes through integrated rural development, since they made new technology packages unprofitable [Kinsey, 1984]. Income from wage employment was limited. Real wages in agriculture declined steadily throughout the 1970s [Kydd, 1985], ensuring a ready supply of cheap labour for the estate (plantation) sector. Overseas remittances dried up following the recall of migrant labour from South Africa in the mid-1970s [Christiansen and Kydd, 1983]. In these circumstances, the focus on improving household food security through household maize production seems understandable.

MARKET LIBERALISATION ... AND THE RURAL POOR IN MALAWI

49

Since the late 1980s, however, economic reforms have increased the scope for smallholders to improve their cash incomes. Smallholders now enjoy access to the world market for burley tobacco. ADMARC's monopoly on the purchase of maize and other smallholder crops has been abolished. Private traders now account for the bulk of maize purchases and fertiliser sales [Conroy, 1998]. With the exception of maize, for which there is a price band, the domestic prices for smallholder crops are close to world levels. Meanwhile, maize production has stagnated. Maize production per capita has fallen from 163 kg in the early 1990s to less than 160 kg in the late 1990s [Conroy, 1998]. In the early 1990s Malawi had seemed on the verge of a Green Revolution [Heisey and Smale, 1995]. But the removal of the subsidy in 1995 and currency devaluations in 1997 and 1998 sharply increased the price of inorganic fertiliser and have made hybrid maize unprofitable as a cash crop when devaluation creates lags between increases in input and output prices [Smale and Phiri, 1998; MOAI, 1998a]. Consequently, there was no increase in the share of the area planted to hybrid maize between 1994 and 1997, though it rose in 1998 thanks to the Starter Pack Scheme that distributed free fertiliser and seed sufficient for 0.1 hectares to every rural household [Longley et. al., 1999]. These developments have increased pressure on smallholders to earn cash income in order to buy maize. Despite the high cost of the new maize technology, household food security in Malawi is still equated with household maize production. Voices challenging this view and pointing to the growing relevance of the market for household food security are beginning to make themselves heard [Brown et al., 1996] but have met stiff resistance from government officials, for whom maize production remains paramount [Conroy, 1998], and from influential NonGovernmental Organisations that see commercialisation and market reforms as harming or excluding the rural poor [Devereux, 1996]. In this view, food security for the poor requires a 'safety net' that subsidises maize production, limits fluctuations in the price of maize, and transfers income directly to the poorest households. While the need for some form of safety net is generally accepted [World Bank, 1999], disagreement centres on how feasible it is for the poor to rely on maize purchases for household food security. In this article, we argue that market reforms are changing the relationship between household maize production and household food security, and between household maize production and household income. Specifically, we argue that: • The high cost of the new maize technology has led to a decline in household maize production. This has weakened the link between household maize production and food security and increased the need for market purchases.

50


• The decline in household maize production has not led to a decline in household income. Instead, market reforms have created new opportunities for poorer households to raise income through commercialisation, wage labour, and micro enterprise. Consequently, household maize production has become a less important share of total income. Thus, the argument challenges several aspects of the conventional wisdom about rural poverty in Malawi. Household food security is not synonymous with the level of household maize production. The fixation on maize fails to recognise the significance of diversified crop and income strategies, which provide maize-deficit households with a measure of income security. Finally, the tendency to reduce such strategies to purely defensive coping 'mechanisms' overlooks their complexity and the degree of planning, skill, and business acumen that they often require. Since these assumptions are also found in the wider literature on rural poverty and food security, they are relevant not just for Malawi but for other African countries that are now experiencing market liberalisation. Unfortunately, it is not possible to rigorously test our hypotheses. Malawi lacks a regular, national series of income-expenditure surveys that would allow an accurate estimate of changes in the level and composition of income among rural households during the period of market liberalization. (The Integrated Household Survey in 1997-98 provides a benchmark against which future changes in income can be measured [GOM, 2000]). Hence, our argument is based on three main sources. The first is a case study of the economics of the most common off-farm enterprises for a small sample of rural households in the Blantyre Shire Highlands, southern Malawi. The second is a case study of the composition of household income for a small sample of five representative household types in the same study area. Finally, we have used secondary, more broadly based studies that give insights into the dynamics of changes in the composition of household income in rural Malawi. The next two sections describe the study area and our methods of data collection. Section IV presents results of the analyses of off-farm enterprises and the level and composition of household income. In section V we examine other evidence on changes in income in relation to household food security. Section VI explores the links between income diversity and market liberalisation and draws some implications for policy. We then briefly summarise our conclusions. II. THE STUDY AREA AND DATA

The study was made in Blantyre Shire Highlands Rural Development Project (RDP), southern Malawi, an area with a high concentration of Malawi's rural


51

poor. Mombezi and Matapwata Extension Planning Areas (EPAs), where the research was conducted, are classified as among the poorest in Malawi [Moriniere et al., 1996]. Population density in the RDP reaches 300 persons/km2. Thirty-eight per cent of smallholder households in the RDP were headed by women [GOM, 1996b] and female-headed households (FHHs) form a disproportionate share of the poor [World Bank, 1996]. Six in ten households cultivated only half a hectare [GOM, 1996b]. On such farms, average maize production provides less than half the calorific needs for the household [Quinn et al., 1990]. The farming system is maize-based with numerous legume intercrops. Rainfall distribution is unimodal with one wet season between November and April, sporadic showers between May and July and a dry period between July and October. In terms of altitude, rainfall, and length of growing season the maize ecology is representative of 40 per cent of the area planted to maize in Malawi [Heisey and Smale, 1995]. The study area lies 30 km from the commercial centres of Blantyre and Limbe that provide both jobs and a market for smallholder crops. The data derives from two studies conducted in four villages in Mombezi and Matapwata Extension Planning Areas [Orr et al, 1999, 2000a]. The first study collected data on the economics of 22 common off-farm enterprises. Of these, seven enterprises originated from households that were visited regularly to collect data on sources of household income, while 15 derived from households that we encountered during the course of these visits. The second study collected data on sources of household income for a small sample of 15 households. These households were purposely selected as representative of five household types that had earlier been identified through cluster analysis [Orr and Jere, 1999]. These were: • Dimba households (a dimba is a field situated near a stream or well that can be irrigated in the dry season) producing high-value vegetables for sale, that were reasonably food secure; • Burley households that grew burley tobacco and that did not grow vegetables but that produced enough maize to be reasonably food-secure; • Vulnerable households producing neither vegetables nor burley tobacco, growing maize on marginal land, and without enough maize to be reasonably food-secure; • Stable MHHs producing neither vegetables nor burley tobacco, but enough maize to be reasonably food-secure, that were headed by a man; and • Stable FHHs producing neither burley nor vegetables but enough maize to be reasonably food secure, and that were headed by a woman;

TABLE 1 CHARACTERISTICS OF CASE STUDY HOUSEHOLDS Cluster

Burley

No.

1 2 3

Ditnba

4 5 6

Vulnerable Stable FHHa Stable MHHb

7 8 9 10 11 12 13 14 15

Area planted planted to maize, 1998/99 (ha)' 0.59 0.55 0.94 2.32 0.51 0.46 1.10 0.58 1.85 0.94 1.22 0.70 1.35 0.81 0.34

Age of household head (years)

42 46 36

42 33 43 60+ 50+ 60 + 44 50 46 50 + 40+ 28

Sex of household head

Male Female Female Male Male Male Female Female Female Female Female Female Male Male Male

Household size (no.)d

6.0 3.0 4.0 4.5 3.0 3.5 2.0 3.5 2.0 2.0 4.0 5.0 2.5 4.0 3.0

Maize selfsufficiency (months)

1996/97

1997/98

10 11 7 7 10 10 8 4

10 10 10 10 10 8 4 7 10 5 10 6 12 4 5

4

4 8 8

8 7 8

a

female-headed household, either dejure or where the husband is absent for six months of the year or more.

b

male-headed household. The agricultural year in Malawi runs from October-September Weighted 1.0 for adults, 0.5 for those below 15 years. In 1999, 45 Malawi Kwacha (MK)= $ US .1 The household income given here has been used to calculate the share of income from different sources shown in Figure 1.

c d e f

Source: Orr et al. [2000a].

Maize fertilizer rate, 1998/99 (kg Nitrogen/ha)

Household income 1998/99 (MK)ef

39

27,524 26,308 11,700 14,292 4,208 4,886 6,166 10,167 9,500 16,940 22,739 10,952 19,706 8,319 5,715

34

45 17 59 83 5 32 5 29 11 105 57 34 56


53

This classification allows a more subtle analysis of household income than the classifications based on farm size or sex of household head that are conventionally used in Malawi [GOM, 1996b]. Farm size is a weak indicator of household income in the study area, where poor soil fertility means that fertiliser use, rather than the area cultivated, is the main determinant of household food security [Orr et al., 1997]. Similarly, FHHs are not homogeneous but range from households where three generations (grandmother, mother, and several adult daughters) may share fields and pool their maize harvests, to households headed by elderly widows living alone whose income relies heavily on gifts and remittances. III. METHODS In both studies, data were collected through case studies rather than large, formal surveys. The case study method was selected because staff time was limited and the collection of accurate data on household income required frequent visits throughout the year. Both studies were conducted between December 1998 (just after maize planting) and November 1999. The survey year was an average one for maize production. Off-farm income was defined as income that did not derive from the household's own production of crops, trees, poultry or livestock but included income from agricultural sources such as casual labour. Definitions of the economic indices used in analysis of off-farm enterprises are given in Table 2. Many studies use expenditure as a proxy for income because it is easier to measure. In this study, however, we measured income rather than expenditure in order to take account of subsistence income and income transfers received by the household. To capture livelihood diversity, household income was divided into four components: agriculture, casual labour (ganyu), microenterprise, and gifts. Income and expenditure surveys in Malawi do not normally distinguish between gifts and remittances as a share of income from transfers [GOM, 2000; Peters, 1998]. A separate study of household income in the survey area has shown that the bulk of transfer income derives from remittances from adult sons and daughters working in the towns of Blantyre and Limbe [Mwale et al., 2000]. Income from agriculture was further subdivided into subsistence and marketed production. Incomes from agriculture and off-farm enterprises are net of cash costs. Annual income data were divided into four quarters. For quarters II-IV, there were five field visits for each time-period. The recall period of two weeks gave coverage of ten weeks in every 12. For the first quarter, the months December and January were covered in full. To avoid a bias caused by the uneven distribution of household visits, the income data for quarters II-IV were adjusted upwards by 17 per cent.

54


Table 1 gives information about the 15 case study households. Only one household (HI2) was self-sufficient in maize during the 1997/98 agricultural year (1 October-30 September). All the remaining households had run out of their own maize by December, five months before the harvest of mature maize in May. Households that were classified as 'vulnerable' because of low food security in 1996/97, when heavy rainfall flooded the maize crop on low-lying fields, continued to have low food security in 1997/98. All three vulnerable households were headed by elderly women, and had low fertiliser rates for maize in 1998/99. IV. RESULTS

Off-farm Enterprises Off-farm enterprises differed widely in terms of markets, seasonality, level of skill, and profitability (Tables 2 and 3). Customers for micro-businesses (geni) consisted not only of distinct markets within the village (cattle-owners, schoolchildren) but also markets beyond the village boundary (traders, townsfolk, fish farms, tannery companies). By contrast, craft enterprises did not specialise for particular markets and produced directly to order from home or nearby villages. Of the 22 enterprises studied, ten (45 per cent) were annual rather than confined to the hungry months when most households run out of their own maize. Most geni enterprises were annual, implying that business and agriculture were complementary. Labouring (ganyu) was strongly seasonal and concentrated in the hungry months except when employment was available on nearby plantations. The most seasonal form of off-farm enterprise was crafts. Mat-making, basket-making, and making granaries were all confined to the dry season (June-October) when raw materials were available. Profit-and-loss accounts for each enterprise show that business had the highest average turnover (median, 868 MK/month). Four business enterprises (brewing gin, curing goat hides, selling fried fish, and tailoring) had a turnover of 2,900 MK/month or more. The business enterprises with the lowest turnover (trading maize bran in the wet season and trading ADMARC maize) were found in the hungry months when competition was high. In all but three geni enterprises, profit margins were above 25 per cent. By contrast, village shop keeping was a high-volume enterprise (turnover, 8000 MK/month) but with a profit margin of only 15 per cent. Monthly income from labour enterprises (ganyu) was low, reaching 1000 MK/month in only three of eight cases (carpentry, house-building, and permanent labour). Agricultural labour during the hungry months - land preparation and weeding - brought in relatively little income. Only one craft enterprise (basket-making) had a turnover of above 1000 MK/month and three craft enterprises had either zero or negative profits. Average returns to labour for crafts (median, 17 MK/day) were the lowest for

TABLE 2 CHARACTERISTICS OF OFF-FARM ENTERPRISES, BLANTYRE SHIRE HIGHLANDS R.D.P., SOUTHERN MALAWI So.

Category

Enterprise

Marital Status•

1 2

Business (geni) "

Brewing local gin Selling goat hides

FHH MHH

3 4

" "

5 6 7 8

" " " "

Selling fried fish Trading ADMARC maize Trading maize flour Selling cooked food Selling snuff Trading maize bran

9 10 11 12 13 14 15

" Tailoring Retailing (pkala) Village shop-keeping Labour (ganyu) Village carpentry " Building houses " Labouring: land preparation " Labouring: weeding " Permanent labour

Place of trade

Customers

Residence Residence company MHH Local villages FHH Local markets MHH Town FHH Village school FHH Residence FHH Local villages

MHH MHH MHH MHH Both

Local markets Home village Nearby village Local villages Local villages

Nov

Dec

Jan

Feb

Mar

Apr

Villagers Tannery Villagers Traders Townsfolk Schoolchildren Villages Cattle-owners, fish-farm Villagers Villagers Villagers Villagers Villagers

16 " 17 " 18 Helping (thandize) 19 "

MHH Nearby estate MHH Residence Both Home village MHH Residence

Villagers One village household Tea plantation Villagers Villagers Villagers

20 21 22 23

Crafts " " "

MHH Local markets MHH Residence MHH Local villages MHH Residence

Villagers Villagers Villagers Villagers

24

Professional

FHH

Villagers, townsfolk

Estate labour Selling firewood Moulding bricks Selling thatching grass Making baskets Making mats Making granaries Making hoe and axe handles Selling herbal medicine

Both Local villages FHH Nearby village

Oct

Residence

Notes:a FHH = female-headed {de facto or de jure); MHH = male-headed. Source: Onetal. [1999].

—

~.—

Hunger months C J = Months of off-farm enterprise

May

Jun

Jul

Aug

Sep

TABLE 3 THE ECONOMICS OF OFF-FARM ENTERPRISES, BLANTYRE SHIRE HIGHLANDS R.D.P., SOUTHERN MALAWI No.

Category

I 2 3 4 5 6 7 8

Business (geni)

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Retailing (okala) Labour (ganyu)

'Helping' (thandize) Crafts

Professional

Enterprise Brewing local gin Selling goat hides Selling fried fish Trading ADMARC maize Trading maize flour Selling cooked food Selling snuff Trading maize bran - wet season - dry season Tailoring Village shop-keeping Village carpentry Building houses Labouring: land preparation Labouring: weeding Permanent labour Estate labour Selling firewood Moulding bricks Selling thatching grass Making baskets Making mats Making maize granaries Making hoe and axe handles Selling herbal medicine

Turnover" (MK/month)b

Operating costs c (MK/month)

2947 2900 3600 100 662-805 868 284

2144 2259 3076 78 547-340 750 241

480 1400 3300 8000 675-1180 1200 _

292 419 2410 6771 263-402 546 -

263 _ 500 1170 144 195 20 667

0.75 — 31 841 196 195 48 171

'Profit margin' d (percent) 27 26 15 22 31-58 14 15 39 70 27 15 — — 0 -218 25 -59 74

Wini *

Net income f (MK/month)

820 1250 1076 57 350-530 498 97

1324 2435 1052 57 350-531 469 97

416 1152 2203 — _ — — — -

416 904 2203 1625 647-1152 1166 676 312 1024 526 262 469 1003 137 195 18 667

Returns to labour s(MK/day) 40 78 44 31

43-163 50 36 35 52 37 26

61-68 50

25-40

26 28 22 14 29 50 25 7 30 9 208

Notes a Turnover = the total sales revenue of an enterprise. b MK 45 = $ US 1 in 1998/99. c Operating costs = the cost of materials plus depreciation on fixed assets. Fixed assets were defined as tools, buildings, and any equipment necessary for the enterprise to function, d Profit margin = (operating profit/turnover) * 100. Operating profit was defined as turnover minus operating costs and the cost of labour. The cost of labour was measured as the wage rate (including the value of a midday meal) for male and female casual labour in the 1998/99 crop year that was paid by estates near our two research sites. The wage rate has been calculated for a working day of six hours duration and includes travel time. Because of the difficulty of calculating the cost of labour, the concept of a 'profit margin' is notional, e Local measure of 'profitability', defined as net returns minus the cost of investment and any associated costs, excluding depreciation on fixed assets and the cost of labour. Wini only applies to enterprises classed as geni. f Turnover — operating costs, g One manday @ six hours. Source: On et al. [1999].


57

any group of enterprises. Finally, professional skills like herbal medicine had a low monthly turnover (667 MK/month) but the highest return to labour (208 MK/day). Sources of Household Income Figure 1A shows the composition of net household income for five representative household groups between December 1998 and November 1999, while Figure IB shows the same for the 'hungry months' between December 1998 and February 1999. Weighted net income averaged 12,620 MK/household. The highest net incomes were found in two households (HI and H2) belonging to the burley cluster. Other households with relatively high incomes included a stable FHH (Hll) that had three active adult members and H13, where the household earned income from house rent and a regular salary. The lowest net incomes were found among households that grew dimba vegetables, not among households in the vulnerable cluster. Agricultural income was important for households that grew burley tobacco or high-value vegetables (Figure 1A). In nine households, however, agriculture accounted for less than half of net household income. It was least important for vulnerable households. During the hungry months, agriculture was an important source of income for households growing vegetables (Figure lb) but other households obtained only small amounts of cash from the sale of dimba maize (H14), mangoes (H7, H10, H12), or stored maize (Hll). Of 15 households, 13 derived income from casual labour (ganyu), mostly for land preparation and weeding. In only three cases, however, did ganyu account for more than a quarter of total household income (Figure 1 A). Only two households (HI and H14) used ganyu as a regular income strategy. Income from ganyu was significant only for one of the vulnerable households (H7), where sons returned home temporarily to help their mother during the hungry season. Of 15 households, 11 had some form of off-farm enterprise apart from ganyu. Four households received more than one-quarter of net income from off-farm enterprises (Figure 1A). Of these, two households (H8 and H9) were members of the vulnerable cluster. In the case of H9, off-farm income from selling snuff accounted for three-quarters of net income. The remaining two households received income from herbal medicine (HI) and from a combination of rents and salary (HI3). Off-farm enterprises was the most important source of income for vulnerable households during the hungry months, exceeding the value of income from gifts (Figure IB). With the exception of H13, which was self-sufficient in maize, all households received gifts. In four cases, gifts accounted for more than onequarter of total income (Figure 1A). Of these, one was a burley household

58

THE EUROPEAN JOURNAL OF DEVELOPMENT RESEARCH FIGURE 1A SOURCES OF HOUSEHOLD INCOME, DEC. 1998-NOV. 1999

100%

• Micro enterprise 0 Gifts • Ganyu 0 Subsistence crops H Marketed crops

H7

H8

H9

Vulnerable HH I

H10 H11 H12 H13 H14 H15 [Stable FHH

Stable MHH

FIGURE IB SOURCES OF HOUSEHOLD INCOME, DEC. 1998-FEB. 1999 100%

• Micro enterprise 0 Gifts • Ganyu S Marketed crops

H10

H11 H12 H13 H14 H15

I Stable FHH


59

(H2), one a vulnerable household (H8), and two were stable MHHs (HI3 and HI4). The first three obtained cash remittances from sons or married daughters in town while the last obtained free maize for malnourished children. Figure la shows that income from gifts was also important during the hungry months, except for two dimba households that received none. EVIDENCE FROM OTHER STUDIES While case studies are useful in exposing the fragility of existing generalisations, it is unwise to generalise from them. This section presents evidence from three more broadly based studies. A preliminary analysis of the national income-expenditure survey by the Agricultural Policy Research Unit (APRU) in 1997 gives an average household income in 1997 of MK 13,483, slightly above the level recorded in 1992 [Simons, 1999]. The study suggests that 'as an absolute minimum', per capita incomes in the southern region rose from MK 609 (US$24) in 1992 to about MK 685 (US$27) in 1997, or an increase of 12.5 per cent. Households had numerous sources of income. Although off-farm income accounted for only 27 per cent of household income in Malawi as a whole, in the povertystricken southern region the share was over 50 per cent. Income diversity was highest among poorer households, where income from wages, gifts, and microenterprise accounted for 69 per cent of household income. About half the rural households in the southern region engaged in some form of micro enterprise, usually trading. However, micro enterprise accounted for only five per cent of household income, not much higher than income from gifts. The main source of off-farm income in the southern region was ganyu, which almost equalled income from own-agriculture. Taken together, income from wages and micro enterprise exceeded income from crops and livestock. The Food and Nutrition Monitoring Survey provides a poverty profile for over 4,500 rural households in 1991 [Simler, 1993]. The analysis revealed that households with very small landholdings (0.5 ha or less) did not necessarily have the lowest incomes. Over 15 per cent of rural households with small landholdings had expenditure levels that put them in the top 20 per cent of households, and a third of these households belonged to the top 40 per cent. What was the secret of their success? The major factor that pushed these households into the top income groups was income from off-farm employment. The high-income households worked off-farm an average of one day a week more than the low-income households, and returns to their labour were three times higher for ganyu and over five times higher for self-employment. Unfortunately, the data do not specify the type of self-employment used by these households. Presumably, these included forms of trading, retailing, or

60


occupations like tailoring that required some capital and skills. We are not told about differences in maize production between the income groups. The more frequent use of hybrid maize and fertiliser suggests that this was higher among the high-income group. Since this survey was made before liberalisation none of these households received income from burley tobacco. At the micro-level, a panel survey of 217 rural households in Zomba district, close to our study area, provides information on changes in household income over the period 1987-97 [Peters, 1998]. The sample is unusual because in 1997 80 percent of households grew burley tobacco. Between 1990 and 1997, average per capita income (as measured by expenditures) rose in real terms by 72 per cent. Among the poorest quartile, per capita incomes rose by 39 per cent. Sources of income also changed significantly. Among betteroff households, the main source of the increase in income was the sale of burley tobacco. Among poorer households, the rise in income was attributable partly to burley and partly to greater income from ganyu, which in turn reflected greater demand for hired labour from low- and middle-income households that had adopted burley. The share of household income from offfarm sources remained unchanged at about one-third. There was no marked increase in income from micro enterprise. It is hard to judge the trend in household maize production since yields in the survey year were exceptionally poor. Average household expenditure on fertiliser had not kept pace with the rise in fertiliser prices, however, suggesting that maize production has fallen. Among households with small landholdings that had adopted burley, however, maize production had increased slightly, presumably because they could now afford more fertiliser. Clearly, in this case study the engine of growth was the adoption of burley tobacco, which had reached a 'critical mass' capable of lifting the entire community to a higher standard of living. These studies support the view that livelihoods among rural households in Malawi are diverse and suggest that, among poorer households, the route to higher incomes lies largely through the commercialisation of farm production or through higher-paying forms of off-farm employment. VI. DISCUSSION AND POLICY IMPLICATIONS The case study findings, and some of the supporting data, relate to the southern region of Malawi, where high population density, access to urban markets, and opportunities for informal cross-border trade with Mozambique have created economic opportunities that may not exist to the same degree in other areas of the country. Nevertheless, the southern region also has the highest concentration of Malawi's rural poor. Our findings therefore have important implications for an understanding of the links between rural poverty and household food security in Malawi.


61

First and foremost, the findings show that while low maize production makes households vulnerable, it does not necessarily make them poor. Recent research has stressed the importance of livelihood diversity for household food security in sub-Saharan Africa [Ellis, 1998]. Off-farm income is a major source of income for food-insecure households [von Braun, 1989; Reardon, 1997] as is commercialisation or the income from cash crops, provided that markets are growing [Binswanger and von Braun, 1991]. Income diversity is particularly important in years of harvest failure [Pierce et ai, 1996]. The vulnerable households in our sample had the largest maize deficits, but the high share of household income from off-farm sources was effective in providing them with income security. As a result, they were not the poorest households. As we have seen, other evidence shows that households with small landholdings but high-paying self-employment may have high incomes [Simler, 1993]. It is misleading, therefore, to classify rural households as poor simply because they run out of maize more quickly than others. Food security is a function of income, not simply of household maize production. Except for casual labour (ganyu), very little is known about off-farm employment in Malawi [World Bank, 1996]. In the literature, off-farm enterprises are described as coping 'mechanisms', or unthinking reactions in response to immediate needs [GoM/UN, 1993]. Thus, a recent exercise on poverty monitoring declared that: The villagers suffering from poverty do not have any strategic approach, involving a long-term or even a medium-term perspective, to fight this situation. Instead, they rely on isolated short-term mechanisms to ease an immediate predicament ... people go as far as they can on maize and only react to food shortage the moment they encounter it. Thus, coping mechanisms are purely reactive measures and as such seasonal, based on the resources available at the time when food shortages hit the people [NEC, 1999: 39]. Agricultural ganyu may be described as a coping 'mechanism' because it is unskilled, and provides households with food or cash during the hungry period [Whiteside, 2000], though our evidence shows some vulnerable households may avoid it. The majority of off-farm enterprises did not fit this description, however. Several involved teamwork, foresight, and careful planning. Trading in maize flour and bran, for example, required a network of regular customers and suppliers that make this enterprise feasible only as a longer-term strategy. Similarly, craft enterprises use inherited skills. Making maize granaries, axe and hoe handles were all activities followed by elderly men who lacked the physical strength for ganyu labour. Interestingly, micro enterprise was also the preferred strategy for the vulnerable households in our sample. These enterprises - brewing local gin, making reed mats, and selling

62


snuff - all required minimal working capital, used materials that were available close to home, and used the homestead as the place of trade. Thus, while villagers are poor their strategies for dealing with poverty may be seen in a more positive light. The low returns from most business enterprises confirm the view that the non-farm sector is an unlikely 'engine of growth' for the rural economy in Malawi [Simons, 1999]. Although some enterprises have high profit margins, turnover is low because trading volumes are limited by a shortage of working capital. Under-capitalisation makes geni a high-risk activity since even small losses may destroy a business. This situation might be partly remedied by strengthening the enabling conditions for micro-enterprise in Malawi. For example, it is possible to develop credit programmes that are tailored for rural households and that make it easier to combine micro-enterprise with farming. It must be recognised, however, that the function of most off-farm enterprises is to give rural households some measure of income security, not to breed entrepreneurs. These enterprises represent the bottom rung of the business ladder, and few households have the vision, ability, or desire to climb higher. While the number of households that participate in micro-enterprise will grow, therefore, the scale of these enterprises is likely to remain small, limiting the scope for capital accumulation and employment. The high share of income from gifts is important in view of recent interest in social safety nets [World Bank, 1999]. Traditionally, gifts from relatives are an important source of income for the poor. One dictionary translates the ChiChewa word for poverty - umphawi - as 'the state of being without relations' [Illijfe, 1985]. Our results show that income transfers continue to provide an important source of income for vulnerable households during the hungry period and may also assist better-off households with the purchase of fertiliser. The location of the study area close to the cities of Blantyre and Limbe meant that several households received regular support from sons or married daughters living in town. This suggests that the growth of urban employment may strengthen informal support networks. Research in the same study area on income transfers between relatives living in the same village suggests that the level of support is not necessarily a function of wealth but of the quality of family relationships [Lawson-McDowa.il and Chiumia-Kaunda, 2000]. Given the depth of Malawian poverty, the share of household income from private transfers is remarkable and underlines their importance for household food security. A striking feature of our case study households was the high share of income from agriculture that is sold. This was true of all households and not only those growing burley or vegetables. Critics have argued that because most smallholders in Malawi have limited land and give priority to maize production, the scope for raising income through commercialisation or 'cash


63

cropping' is limited [MOAI, 1998a; Devereux, 1996]. Smallholders are regarded simply as 'subsistence farmers' unable to respond to price incentives [Lele, 1989] and with little role to play in the market economy except as a 'labour reserve' [Kydd and Christiansen, 1982]. In fact, poorer smallholders are closely integrated with markets, both as buyers and sellers. It is true that they may find it difficult to grow crops like burley tobacco that are labour-intensive and must be rotated with maize to preserve soil health [Orr, 2000]. They sell a wide variety of other crops, however. This may work to their disadvantage if they sell maize cheaply after harvest only to buy again during the hungry period [Peters, 1996]. But the highest proportion of households selling pigeon pea, sweet potato, cassava, millet and vegetables is found among those cultivating 0.5 hectares or less [Longley et al., 1999]. In our study area, the crops most widely sold were field pea, pigeon pea, and beans. Benefit-cost ratios for field pea and beans were higher than for fertilised composite maize [Orr et al, 2000b]. Unlike tobacco or groundnuts, these crops are intercropped with maize and therefore do not reduce household maize production. This explains their popularity with resource-poor households with smaller landholdings. Yet their importance in generating income for the poor has been neglected because they are conventionally defined as 'food crops'. Since 1994 the area planted to beans, pigeon pea, and soybean has increased significantly thanks to increases in relative prices [Mataya and Chulu, 1998]. Trade in grain legumes has grown thanks to higher profit margins [Peters, 1998]. A survey of four market centres showed that 30 per cent of traders (mostly women) had started trading since 1996 [Simons, 1999]. Regional and overseas markets have fuelled much of this increase in demand, with soybean exported to South Africa and pigeon pea to India [MOAI, 1998b]. The market potential for pigeon pea is particularly promising since exports from Malawi reach India in October before the harvest of the domestic crop, when prices are highest. Malawi already has the largest processing industry in the region and is well placed to take advantage of this market. New varieties of pigeon pea are needed that match the needs of the food processing industry and Indian consumers. Linking growers more closely with buyers would make production more profitable, increasing the share of income from these crops that accrued to smallholders. Developing the pigeon pea industry in Malawi along market lines will require a concerted effort by researchers, growers, and processors [Jones et al., 2000]. Commercialisation is also evident in the recent increase in the area planted to cassava and sweet potato. A survey of sweet potato growers in our study area showed that the crop was now being intercropped with maize or planted after the maize harvest [Mwale et. al., 1999]. Unlike pigeon pea, the market for sweet potato is domestic. Both cassava and sweet potato are in high demand

64


from urban consumers as a snack food [Peters, 1996]. Among households with low maize production, income from the sale of sweet potato was usually spent on maize. Moreover, sweet potato frequently replaced maize as a midday meal between June and September. Commercialisation has therefore served as a market strategy to generate cash to buy maize and also as a subsistence strategy that extends the period that the household is self-sufficient in maize and reduced the need for market purchases [Mwale et aL, 1999]. Our evidence shows that commercialisation is not necessarily a passport to riches. One-third of rural households in Malawi have access to dimba land suitable for horticulture [FEWS, 1996]. However, two of the dimba households in our sample had lower net incomes than households that were classed as 'vulnerable'. Production costs for cabbage and tomato are high because of expenditure on fertiliser and pesticides, while returns are variable because of damage from pests and diseases and fluctuations in market prices. Moreover, income from horticulture peaks in the hungry months between November and January, which limits the cash that dimba households may earn from casual labour and may also reduce the willingness of other households to make gifts. At present, therefore, the scope for poverty alleviation through horticulture is limited. Successful commercialisation will depend on reducing unit costs of production (through more rational use of pesticides and more efficient forms of manual irrigation) and, as in Kenya, creating an export market for highvalue vegetables [Jaffee, 1994]. In sum, the evidence suggests that recent market and price reforms in Malawi are changing the relationship between household maize production and household food security, and between household maize production and household income. Malawi's stalled Green Revolution has led to a decline in per capita maize production. Hybrid seed and fertiliser are too expensive for the majority of smallholders, and maize production now depends heavily on welfare measures such as Starter Packs that provide poorer smallholders with an extra three weeks' maize production [Longley et ah, 1999]. Smallholders have also reduced the amount of maize offered for sale [Peters, 1998] and rationed maize consumption by eating more cassava and sweet potato. It is not known whether these measures have fully compensated for the decline in maize production, but it seems unlikely. If not, then the rural poor have become more dependent on the market for household food security. Remarkably, this fall in household maize production does not seem to have been accompanied by a corresponding decline in household income. Average income over the period 1992-97 appears to have increased. Poorer households have shared in this increase in income. We cannot identify the source of this income growth precisely. The main sources seem to have been commercialisation of farm production (both burley and grain legumes) and increased participation in micro-enterprise, especially rural trade. Where burley


65

tobacco is widely grown, the demand for casual labour has risen, pushing up the share of household income from wage employment. Informal safety nets have also remained intact and are an important source of income for poorer households. They may have been strengthened by growth in incomes and in urban employment. Both commercialisation and micro-enterprise have been facilitated by economic reforms that have removed restrictions on the production of burley tobacco, on the marketing of crops, and on trade in agricultural commodities. If these trends continue, the share of maize production in household income seems likely to decline, and changes in household income will derive increasingly from cash cropping and off-farm employment. This argument must not be pushed too far. Household maize production will remain the key determinant of food security for the rural poor in Malawi. Smallholders will continue to limit their dependence on the market for maize purchases. Safety nets will be necessary to improve access to hybrid seed and fertiliser for the poor, to reduce fluctuations in market prices, and to provide employment during the off-season. In the longer term, however, the only sustainable solution to Malawi's present development impasse is to create more opportunities for poorer smallholders to earn cash incomes. This will give greater income security for the purchase of maize on the open market and allow higher investment in household maize production. Both will improve household food security and reduce rural poverty. The policy implications of these findings are that greater efforts are required to accelerate the processes of commercialisation and off-farm entrepreneurship that are now under way. Specifically, measures are needed to • Strengthen the enabling conditions for micro-enterprise. Innovative credit programmes are needed, tailored for rural households who combine microenterprise with farming. • Expand the market for the grain legume intercrops grown by poorer smallholders. The markets for pigeon pea, beans, and soybean are ripe for development provided that there is a coordinated effort by researchers, growers, and exporters to deliver the products that match the needs of consumers. • Develop the horticulture industry. High production costs and limited markets reduce the profits from horticulture. Successful commercialisation will require investment in extension services and in market infrastructure. CONCLUSION Rural poverty in Malawi is closely identified with self-sufficiency in maize, both at the national and the household level. Efforts to improve household food security have therefore focused largely on the promotion of hybrid seed and

66


fertiliser to increase maize production. Relatively less attention has been given to the scope for increasing smallholder incomes and achieving household food security through the market by the purchase of food. With the high cost of the new maize technology, however, household maize production has declined and the market has become a more important source of household food security for the rural poor. Our evidence suggests that the decline in household maize production has not been accompanied by a decline in household income, however. This suggests the need for changes in widely held assumptions about the relationship between household maize production and rural poverty in Malawi. First, poverty is not synonymous with household maize production. Obviously, households require a minimum level of food security from their own maize production. At the margin, however, rural households with maize deficits make decisions about the relative costs and benefits of different income strategies. These strategies are often effective in providing maizedeficit households with some measure of income security. Our evidence suggests that vulnerable households with high maize deficits are not necessarily the poorest households. Second, it is misleading to regard off-farm enterprises as simply coping 'mechanisms'. Off-farm enterprises are better seen as calculated attempts to diversify household income and create the economic basis for a sustainable livelihood. Many are not confined to the hungry season but encompass the entire year, and often require initiative, business acumen, and skill. Income diversity is important for income security that, in turn, is an important source of household food security. Market liberalisation has increased the scope for income diversity among the rural poor. While liberalisation of burley tobacco has primarily benefited richer households, where burley production has reached a 'critical mass' incomes for non burley-growers have also risen through the increase in demand for hired labour. Poorer smallholders have also benefited from rising prices for beans, soybean, and pigeon pea. Unlike traditional cash crops, commercialisation of these intercrops is possible without reducing maize production. In addition, the relaxation of trade restrictions has resulted in an increase in micro enterprise, such as brewing, trading in crops, and wholesaling. The policy implications of these findings are that greater efforts are required to accelerate the processes of commercialisation and off-farm entrepreneurship that are now underway. In particular, measures are needed to expand the markets for grain legumes and for high-value vegetable crops, and to expand credit for micro enterprise. These policy prescriptions are hardly new, but they apply with particular force in Malawi, where opportunities for the poor to earn cash income have for so long been suppressed and where the cost of hybrid maize technology means that smallholders must urgently search for new sources of cash income to safeguard household food security.


67

REFERENCES Binswanger, H.P. and J. von Braun, 1991, 'Technological Change and Commercialisation in Agriculture: The Effect on the Poor', World Bank Research Observer, Vol.6, No.2, pp.57-80. Brown, D.G., Reutlinger, S. and A.S. Thompson, 1996, Malawi: Food Security in a Market-Oriented Economy, Bethseda, MD: Development Alternatives Incorporated. Christiansen, R.E. and. J.G. Kydd, 1983, 'The Return of Malawian Labour from South Africa and Zimbabwe', Journal of Modern African Studies, 21, No.2, pp.311-26. Conroy, A.C., 1998, 'Draft Discussion Document Prepared for the National Food Security Steering Committee', mimeo. Food Security Issues Paper, Lilongwe: Ministry of Finance. Devereux, S., 1996, Food Security in Malawi: Issues, Strategies, and Policies, Lilongwe: ActionAid. Ellis, F., 1998, 'Household Strategies and Rural Livelihood Diversification', The Journal of Development Studies, Vol.35, No.1, pp. 1-38. Famine Early Warning System (FEWS), 1996, Rapid Food Security Assessment, NovemberDecember 1995, Lilongwe: Famine Early Warning System. Government of Malawi/United Nations in Malawi, 1993, Malawi: Situation Analysis of Poverty, Lilongwe: United Nations in Malawi. Government of Malawi, 1996a, Malawi Social Indicators Survey 1995, Ministry of Economic Planning and Development, National Statistical Office, and Centre for Social Research. Government of Malawi, 1996b, National Sample Survey of Agriculture 1992/93. Vol. 1: Smallholder Household Composition Survey Report, Zomba: National Statistical Office. Government of Malawi, 2000, Integrated Household Survey 1997-98. Draft Report. Statistical Abstract, June, Zomba: National Statistical Office. Heisey, P.W., and M. Smale, 1995, Maize Technology in Malawi: A Green Revolution in the Making? Research Report No.4, Mexico: International Maize and Wheat Improvement Centre (CIMMYT). Illiffe, J., 1985, 'The Poor in the Modern History of Malawi', in Centre for African Studies, Malawi: An Alternative Pattern of Development. Proceedings of a Seminar at the Centre for African Studies, University of Edinburgh, 24-25 May, 1984, Edinburgh: Centre for African Studies, pp.245-92. Jaffee, S., 1994, 'Contract Farming in the Shadow of Competitive Markets: The Experience of Kenyan Horticulture', in P.D. Little and M.J. Watts (eds.), Contract Farming and Agrarian Transformation in Sub-Saharan Africa, Madison, WI: University of Wisconsin Press, pp.97-137. Jones, R.B., Likoswe, A., and A. Freeman, 2000,. 'Improving Poor Farmers' Access to Technologies and Markets for Pigeonpea in Malawi', in Ritchie [2000: 150-63]. Kinsey, B.H., 1984, 'Integrated Rural Development, Agricultural Growth, and Equity in Malawi: Redefining Development Strategies and Tactics for the 1980s', Agricultural Administration, Vol.13, pp.45-63. Kydd, J., 1985, 'Malawi in the 1970s: Development Policies and Economic Change', in Centre for African Studies, Malawi: An Alternative Pattern of Development, Centre of African Studies, University of Edinburgh, pp.295-380. Kydd, J. and R. Christiansen, 1982, 'Structural Change in Malawi since Independence: Consequences of a Development Strategy Based on Large-Scale Agriculture', World Development, Vol.10, No.5, pp.355-75. Lawson-McDowall, J. and C. Chiumia-Kaunda, 2000, 'Village Level Safety Nets: A Study of Resource Flows Between Closely Related Households', mimeo. Farming Systems Integrated Pest Management Project, Department of Agricultural Research and Technical Services, Ministry of Agriculture, Malawi. Lele, U., 1989, Structural Adjustment, Agricultural Development, and the Poor: Lessons from the Malawian Experience, MADIA Discussion Paper No.9. Washington, DC: World Bank. Longley, C., Coulter, J. and R. Thompson, 1999, Malawi Rural Livelihoods Starter Pack Scheme, 1998-99: Evaluation Report, London: Overseas Development Institute. Mataya, C. and O. Chulu, 1998, Socio-Economic Effects of Policy Changes on Rural Agricultural Households, Bunda College of Agriculture, University of Malawi. Ministry of Agriculture and Irrigation (MOAI), 1998a, Malawi: Soil Fertility Options and Issues: A Discussion Paper, Lilongwe: Ministry of Agriculture and Irrigation.

68


Ministry of Agriculture and Irrigation, 1998b, Grain Legumes: Issues and Options for Research, Production, Marketing, and Utilization in Malawi, Lilongwe: Ministry of Agriculture and Irrrigation. Moriniere, L., Chimwaza, S. and E. Weiss, 1996, A Quest for Causality. Vulnerability Assessment and Mapping (VAM) Malawi Baseline 1996, Lilongwe: Famine Early Warning System. Mwale, B., Orr, A. and D. Saiti, 1999, 'Sweet Potato and Smallholder Food Security in Blantyre Shire Highlands RDP', Farming Systems Integrated Pest Management Project, Department of Agricultural Research and Technical Services, Ministry of Agriculture, Malawi. Mwale, B., Orr, A and D. Saiti, 2000, Adapting to Adjustment: Smallholder Livelihood Strategies in the Blantyre Shire Highlands, Lilongwe: Rural Economic Policy Centre. National Economic Council (NEC), 1999, Qualitative Impact Monitoring of Poverty Alleviation Policies and Programmes in Malawi. Vol.1: Research Findings, mimeo, Lilongwe: National Economic Council. Ng'ong'ola, C., 1986, 'Malawi's Agricultural Economy and the Evolution of Legislation on the Production and Marketing of Peasant Economic Crops', Journal of Southern African Studies, Vol.12, No.2, pp.240-62. Orr, A., 2000, '"Green Gold?" Burley Tobacco, Smallholder Agriculture, and Poverty Alleviation in Malawi', World Development, Vol.28, No.2, pp.347-63. Orr, A., Jere, P. and A. Koloko, 1997, 'Baseline Survey, 1996/97', Farming Systems Integrated Pest Management Project, Department of Agricultural Research and Technical Services, Ministry of Agriculture, Malawi, Mimeo, 102 pp. Orr, A. and P. Jere, 1999, 'Identifying Smallholder Target Groups for IPM in Southern Malawi', International Journal of Pest Management, Vol.45, No.3, pp.179-87. Orr, A., Saiti, D. and B. Mwale, 1999, 'Osauka satopa: Case Studies of Off-Farm Income in Blantyre Shire Highlands RDP', mimeo, Farming Systems Integrated Pest Management Project, Department of Agricultural Research and Technical Services, Ministry of Agriculture, Malawi. Orr, A., Saiti, D. and B. Mwale, 2000a, 'Off-Farm Income and Smallholder Livelihoods in Blantyre Shire Highlands RDP', mimeo, Farming Systems Integrated Pest Management Project, Department of Agricultural Research and Technical Services, Ministry of Agriculture, Malawi. Orr, A., Mwale, B. and D. Saiti, 2000b. 'Markets and Livelihoods: Lessons From FSR in Blantyre Shire Highlands RDP', in Ritchie (ed.) [2000: 279-91]. Pierce, J., Ngwira, A. and G. Chimseu, 1996, Living on the Edge: A Study of the Rural Food Economy in the Machinji and Salima Districts of Malawi, Lilongwe: Save the Children (UK). Peters, P., 1996, Conceptual Quagmires, Old Problems, and New Questions: Rethinking Policy Assumptions about Malawi's Rural Economy, Harvard, MA: Harvard Institute for International Development. Peters, P., 1998, Agricultural commercialisation, rural economy and household livelihoods, 1990-1997, Harvard, MA: Harvard Institute for International Development. Quinn, V, 1994, 'A History of The Politics of Food and Nutrition in Malawi: The Context for Food and Nutritional Surveillance', Food Policy, Vol.19, No.3, pp.255-71. Quinn, V., Chiligo, M. and J.P. Gittinger, 1990, 'Malnutrition, Household Income, and Food Security in Rural Malawi', Health Policy and Planning, Vol.5, No.2, pp.139-48. Reardon, T., 1997, 'Using Evidence of Household Income Diversification to Inform Study of the Rural Nonfarm Labor Market in Africa', World Development, Vol.25, No.5, pp.735-47. Ritchie, J.M. (ed.), Integrated Crop Management Research in Malawi: Developing Technologies with Farmers. Proceedings of the Final Workshop of the Farming Systems Integrated Post Management Project, Club Makokola, Malawi, November 29—December 3, Chatham, UK: Resources Institute. Simons, G., 1999, 'Malawi Rural Sector Assessment', Development Alternatives Incorporated, Bethseda, Maryland. Simler, K., 1993, Poverty Profile of Rural Households in Malawi: A Summary of Recent Findings', Agricultural Sector Memorandum, Working Paper No.4, Lilongwe: World Bank. Smale, M. and A. Phiri, (with contributions from G.A. Chikafa, P.W. Heisey, E. Mahatta, M.N.S. Msowoya, E.B.K. Mwanyongo, H.C. Sagawa and H.A.C. Selemani), 1998, 'Institutional Change and Discontinuities in Farmers' Use of Hybrid Maize Seed and Fertilizer in Malawi:


69

Findings from the 1996-97 CIMMYT/MoALD Survey', Economics Working Paper 98-01, Mexico, DF: CIMMYT. von Braun, J., 1989, The Importance of Non-Agricultural Income Sources for the Rural Poor in Africa and Implications for Food and Nutrition Policy, Reprint No.189, Washington, DC: International Food Policy Research Institute. Whiteside, M., 2000, 'Ganyu Labour in Malawi and its Implications for Livelihood Security Interventions — An Analysis of Recent Literature and Implications for Poverty Alleviation', Agricultural Research and Extension Network, Network Paper No.99, London: Overseas Development Institute. World Bank, 1989, Malawi Food Security Report, Report No.8151 - MAI. Washington, DC: World Bank. World Bank, 1996, Malawi: Human Resources and Poverty. Profile and Priorities for Action, Report No.15437 - MAI, March, Washington, DC: World Bank. World Bank, 1997, Accelerating Malawi's Growth: Long-term Prospects and Transitional Problems, Sept., Washington DC: World Bank. World Bank, 1998, World Development Indicators 1998, Washington, DC: World Bank. World Bank, 1999, Malawi: A Safety-Net Strategy for the Poorest, Report No.19974-MAI. Washington, DC, World Bank.