Poverty and Social Impact Assessment - World Banksiteresources.worldbank.org/.../15009_malawi_admarc_psia.pdf · 2010-04-01 · ii Acknowledgements This report represents the outcome

i

Document of

THE WORLD BANK

FOR OFFICIAL USE ONLY

CONFIDENTIAL Report No.: 27512

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 This document has a restricted distribution and may be used by recipients only in the performance of their duties. Its contents may not otherwise be disclosed without World Bank authorization

ii

The findings, interpretations and conclusions reported in this document are preliminary and may be revised at a later stage. They represent solely the authors’ view and should not be attributed to the World Bank, its management, its Board of Directors, or the countries theyrepresent.

i

Abbreviations and Acronymns

ADMARC Agricultural Development and Marketing Corporation CPS Complementary Panel Survey CSR Centre for Social Research FEWS Famine Early Warning System GoM Government of Malawi GTZ Germany Agency for Technical Cooperation IFPRI International Food Policy Research Institute IHS Integrated Household Survey MK Malawi Kwacha MPRSP Malawi Poverty Reduction Strategy Paper NASFAM National Association of Smallholder Farmers MOEPD Ministry of Economic Planning and Development NEC National Economic Council NFRA National Food Reserve Agency PSIA Poverty and Social Impact Analysis TA Traditional Authority USAID United States Agency for International Development WFP World Food Programme

ii

Acknowledgements This report represents the outcome of a joint effort by the Government, GTZ and the World Bank, in consultation with other donors and civil society. The work is one of the pioneer studies using the Poverty and Social Impact Analisys (PSIA) approach. Clear practices on process to be followed for a PSIA were not yet available when the study was started, and have been improved over the last two years. The PSIA process has since become a more inclusive exercise involving a broader range of stakeholders from civil society and the donors’ community. The project was started by Hassan Zaman (AFTP1). The work was undertaken under the auspices of a secretariat chaired by Dr Milton Kutengule, Permanent Secretary, Ministry of Economic Planning and Development. Three background studies have been prepared by Maxton Tsoka (Centre for Social Research, University of Malawi, Zomba), Manohar Sharma, Ellen Panyongayong, and Todd Benson (International Food Policy Research Institute, IFPRI, Washington D.C); Peter Mvula, Ephraim Chirwa and John Kadzandira (Wadonda Consulting, Malawi, Zomba); and Janet Owens (consultant, World Bank). Funding was provided by GTZ and the Norwegian Trust Fund administered by the World Bank. This synthesis report has been written by Antonio Nucifora (AFTP1). The report was prepared under the guidance of Philippe Le Houerou, Sector Manager (AFTP1), and benefited from comments of peer reviewers Anis Dani (SDV) and Stefano Paternostro (PRMPV), and additional comments by Sudhir Chitale, Maxwell Mkezalamba (AFTP1), Louise Fox (AFTPM), Jean Paul Chausse, Stanley Hiwa, Francis M’Buka, Tijan Sallah, (AFTS1), Reiner Forster and Renata Hisrcher (SDV / GTZ). Rose Kumsinda provided excellent assistance in document preparation.

i

Reforming the Malawi Agricultural Development and Marketing Corporation (ADMARC):

Synthesis Report of the Poverty and Social Impact Analysis

TABLE OF CONTENTS

EXECUTIVE SUMMARY...................................................................................................... I

1. INTRODUCTION............................................................................................................1

2. THE AGRICULTURAL DEVELOPMENT AND MARKETING CORPORATION (ADMARC)...............................................................................................................................3

THE STRUCTURE AND PURPOSE OF ADMARC .......................................................................3 ADMARC’S ROLE IN AGRICULTURAL MARKETING AND FOOD SECURITY.............................5 THE IMPACT OF PREVIOUS ROUNDS OF ADMARC RESTRUCTURING .....................................6

3. QUANTITATIVE ANALYSIS OF DATA FROM THE 1997/98 INTEGRATED HOUSEHOLD SURVEY.........................................................................................................7

METHODOLOGY: AN ECONOMETRIC MODEL OF HOUSEHOLD EXPENDITURES IN 1997/98......7

4. QUANTITATIVE ANALYSIS OF DATA FROM THE 1997/98 INTEGRATED HOUSEHOLD SURVEY AND THE COMPLEMENTARY PANEL SURVEY. ...........10

METHODOLOGY: AN ECONOMETRIC MODEL OF CHANGES IN HOUSEHOLD EXPENDITURES BETWEEN 1997/98 AND 2002................................................................................................11 RESULTS OF THE QUANTITATIVE ANALYSIS OF DATA FROM THE IHS-1 AND CPS-4. ...........12

5. QUALITATIVE STUDY ON THE IMPACT OF ADMARC FACILITIES AND SERVICES, AND THE ROLE OF THE PRIVATE SECTOR IN AGRICULTURAL MARKETING ..........................................................................................................................14

METHODOLOGY: A QUALITATIVE STUDY OF THE IMPACT OF ADMARC CLOSURE IN 20 VILLAGES..............................................................................................................................14 BASIC FINDINGS OF QUALITATIVE STUDY: THE IMPACT OF ADMARC’S CLOSURE AND THE ROLE OF THE PRIVATE SECTOR IN AGRICULTURAL MARKETING IN RURAL AREAS...............16

6. CONCLUSIONS OF THE PSIA STUDY ON AGRICULTURAL MARKET REFORMS..............................................................................................................................19

REFERENCES.......................................................................................................................22

ANNEX A - QUANTITATIVE STUDY PREPARED BY IFPRI AND CSR .............................25

ANNEX B - QUANTITATIVE STUDY PREPARED BY J. OWENS (WB CONSULTANT) ...35

ANNEX C - QUALITATIVE STUDY PREPARED BY WADONDA CONSULT (ZOMBA) ..53

i

EXECUTIVE SUMMARY 1. This report assesses the impact of ADMARC marketing activities on household welfare in Malawi. The study is based on three background reports each using a different methodology. Two quantitative studies use econometric techniques to analyze data from the 1997/98 Integrated Household Survey (IHS-1) and the 2002 fourth round of the Complementary Panel Survey (follow-up panels on the IHS-1). These studies investigate the relationship between access to ADMARC’s services and changes in household welfare during 1997/98 and 2002. The third study uses a qualitative methodology to carry out field research in 20 rural communities on the effects of the decline in ADMARC’s marketing activities in recent years. 2. ADMARC has been a major concern of the government and the international financial institutions for several years because of its deteriorating financial position, and its role in agricultural markets and food security. These concerns are mainly due to the fact that ADMARC bail-outs in the past have derailed the macroeconomic framework (1.6 percent of GDP in 1997/98 and 1.9 percent of GDP in 2000/01). In addition, several analysts have argued that the presence of ADMARC markets and warehouses discourages private sector participation in maize trade, marketing, and storage, thus leading to a longer term inefficiency in marketing operations. Finally, the recent reports of mis-management of grain reserves at the onset of the 2001/02 food crisis have reinforced the doubts surrounding ADMARC’s role in food security. 3. As a result, ADMARC is under intense pressure from the government to improve its financial position partly through adoption of cost cutting measures. Several proposals have been advanced on how ADMARC can improve its economic performance, including (i) privatization/liquidation of the loss making subsidiaries; (ii) auctioning part of ADMARC’s marketing and storage infrastructure; and, (iii) repeal ADMARC Act and incorporate ADMARC as a limited liability company to be eventually privatized. In the Malawi Poverty Reduction Strategy Paper (MPRSP) privatization of ADMARC is one of the strategies towards liberalization of agricultural markets and the development of the agriculture sector (GoM, 2002). 4. However, significant controversy surrounds the decision to reduce ADMARC’s marketing role (and possibly to sell off a considerable of its marketing infrastructure). Firstly, there is concern that if remote markets are closed due to lack of efficiency, they are unlikely to be replaced by private traders because transportation costs are high relative to the return on maize sales. Secondly, in light of its substantial warehousing facilities, ADMARC is seen as a source of supply in times of scarcity. This is spite of the fact that recently ADMARC has had difficulties in meeting demand in times of low production, and has had to introduce restrictions on the quantities purchased. Nevertheless, the public perception of ADMARC’s importance in agricultural marketing and food security remains much higher than supported by the data. As a result, some members of government have suggested that the decision to close the markets is politically untenable and would give rise to considerable public resentment. 5. To inform the debate on the agricultural marketing reforms, the World Bank and the Government, in collaboration with other donors and stakeholders, have commissioned this Poverty and Social Impact Analysis (PSIA) focusing on the impact of the proposed closure of some of ADMARC’s agricultural markets and, more generally, to study the role of proximity to ADMARCs facilities (and services) for household welfare. 6. The results of the analysis indicate that access to ADMARC has a positive impact on household welfare. Specifically, the results of the analysis estimate that per capita consumption can be up to 20 percent higher for households living closer to ADMARC

ii

facilities. This can be taken as evidence that ADMARC facilities provide rural households with valuable access to a marketing channel for buying inputs, selling their output and purchasing maize for self-consumption. This finding also substantiates the complaint from rural communities, particularly in remote areas, that they feel they have been adversely affected by the reduction in ADMARC marketing activities. 7. The second major finding is that the beneficial impact of proximity to ADMARC markets is more important in remote rural areas, far from major roads. More remote areas have less developed private markets, and alternatives to ADMARC services are less likely to exist there. In other words, ADMARC’s impact depends on the availability of alternative marketing channels. In less remote areas, higher competition can be expected to lower marketing margins, hence reducing the benefit of access to ADMARC’s pan-territorial prices. Indeed this is confirmed by the finding that in areas where private market infrastructure is developed and market services function, ADMARC facilities do not have a significant impact on farm households. 8. The final finding is that the role of ADMARC is important in maize sales during the hungry season, but much less important during the harvesting season. This reflects the fact that small traders engage mainly in buying maize from farmers after harvest, and less on selling maize to consumers during the lean season. They transport maize out of remote areas and do not generally engage in storage activities. Similarly, small traders are not generally involved in agricultural input supply in remote areas. 9. A number of policy recommendations follow from the findings of the study:

i. Completely eliminating ADMARC facilities may ignore the fact that well-functioning market institutions and infrastructure do not exist uniformly throughout the country. In the short-run, the provision of ADMARC facilities may be beneficial in some areas that are under-served. The question then becomes whether ADMARC can play this useful function in a cost-effective manner, or if this social function could be done better through other market-based mechanisms.

ii. Maintaining ADMARC in its current form is wasteful. Its services are not

required in less remote areas of the country, where the activity of the private sector is already well established. Therefore, ADMARC’s facilities in developed parts of the country should be auctioned off to the private sector. All of ADMARC’s other non-marketing activities should also be sold to private operators. These activities are not needed for the operation of the ‘remote’ markets, and hence should be operated as a private company. Such a decision will prevent the possibility that unexpected financial losses originating from poor management of ADMARC’s non-social functions may again in the future have to be bailed out using public resources (from the budget).

iii. In the medium term, as private sector expands its marketing activities in

remote parts of the country, the role for the Government to facilitate food marketing in remote areas will diminish. The Government can accelerate this process by investing in market infrastructure in remote areas, and facilitating the development of institutions that bring greater transparency and competition in market transactions. These measures will enhance private sector participation in the agricultural sector and eventually eliminate the need for the Government to support marketing services in remote areas.

iii

iv. There is a need to identify and remove the constraints preventing private traders from entering the fertilizer supply chain. Similarly, obstacles to private investment in storage facilities and interregional and/or seasonal arbitrage need to be identified and tackled by appropriate policy intervention.

1

1. INTRODUCTION

1.1 The debate surrounding the restructuring of ADMARC has been revived in recent years because the deteriorating financial position of ADMARC has been a major concern of the government and the international financial institutions. These concerns are mainly due to the fact that ADMARC bail-outs in the past have derailed the macroeconomic framework (1.6 percent of GDP in 1997/98 and 1.9 percent of GDP in 2000/01). In addition, several analysts have argued that the presence of ADMARC markets and warehouses discourages private sector participation in maize trade, marketing, and storage, thus leading to a longer term inefficiency in marketing operations (see, inter alia, Abbott and Polin, 1996). Finally, the recent reports of mis-management of grain reserves at the onset of the 2001/02 food crisis have reinforced the doubts surrounding ADMARC’s role in food security. 1.2 As a result, ADMARC is under intense pressure from the government to improve its financial position partly through adoption of cost cutting measures. Several proposals have been advanced on how ADMARC can improve its economic performance. The proposals which have been discussed comprise (i) privatization/liquidation of the loss making subsidiaries; (ii) auctioning part of ADMARC’s marketing and storage infrastructure; and, (iii) repeal ADMARC Act and incorporate ADMARC as a limited liability company to be eventually privatized. 1.3 In early 2002, government transferred the four subsidiaries of ADMARC to Ministry of Finance as a stop-gap measure for eventual privatization. In the Malawi Poverty Reduction Strategy Paper (MPRSP) privatization of ADMARC is one of the strategies towards liberalization of agricultural markets and in the development of the agriculture sector (GoM, 2002). 1.4 However, significant controversy surrounds the decision to reduce ADMARC’s marketing role (and possibly to sell off a considerable of its marketing infrastructure). The main concerns focus on:

i. Market failures in remote areas: There is concern that if remote markets are closed due to lack of efficiency, they are unlikely to be replaced by private traders because transportation costs are high relative to the return on maize sales. This could be particularly true in the rainy season when the condition of rural roads worsens significantly.

ii. Food storage for lean seasons: ADMARC is seen as a source of supply in times of scarcity. By maintaining warehousing facilities across the country, ADMARC can store purchased maize in secure markets in the food-deficit areas or in a proximate central location for transportation during the rainy season. However, recently ADMARC has had difficulties in meeting demand in times of low production, especially as government support has dwindled, and has had to introduce restrictions on the quantities purchased.

iii. Political concerns: The public perception of ADMARC’s importance in

agricultural marketing and food security is much higher than supported by the available data. As a result, some members of government have suggested that the decision to close the markets is politically untenable. There is concern that rapid closure of markets would give rise to considerable public resentment and political difficulties.

1.5 To inform the continued debate regarding the sequencing of agricultural reforms and the extent that input and output markets ought to be restructured, the World Bank and the Government, in collaboration with other donors and stakeholders have commissioned a

2

Poverty and Social Impact Analysis (PSIA) focusing on the impact of the proposed closure of some of ADMARC’s agricultural markets and, more generally, to study the role of proximity to ADMARCs facilities (and services) for household welfare (Box 1). 1.6 The study is based on three background reports each using a different methodology (the full reports are attached as annexes to this synthesis paper). Two quantitative studies use econometric techniques to analyze data from the 1997/98 Integrated Household Survey (IHS-1) and the 2002 fourth round of the Complementary Panel Survey (follow-up panels on the IHS-1). These studies investigate the relationship between access to ADMARC’s services and changes in household welfare during 1997/98 and 2002. The third study uses a qualitative

Box 1. The PSIA of ADMARC’s restructuring: a brief note on the process followed The study on the Poverty and Social Impact Analysis of ADMARC’s privatization started in late 2001 asa result of concerns about the impact of ADMARC’s dismal financial performance on macroeconomic stability. The Government and the donors agreed it was critical to prevent ADMARC from causingfurther fiscal shocks to the economy, but were aware of the importance to proceed cautiously due toconcerns about the disruption of the social function provided by ADMARC’s markets in remote areas.1The goal was to minimize the impact on the budget while allowing the operation of social activities to be financed in a transparent manner. An agreement was reached between the World Bank and the National Economic Council (now Ministry of Economic Planning and Development) to carry out an in-depth study on the likely poverty and social impacts of the privatization of ADMARC marketing activities. It was agreed that before proceeding with a detailed proposal to restructure and/or privatize ADMARC, it was necessary to carry out such a study.It was also agreed that some of the heavily loss-making subsidiaries of ADMARC, which were not related to ADMARC agricultural marketing functions, could be earmarked for privatization (notably: David Whitehead & Sons, Cold Storage, Shire Bus Lines, and Cotton Ginning). The study was launched in early 2002. The host institution within the Government has been the NEC(now Ministry of Economic Planning and Development), in the person of Dr Milton Kutengule. Three background studies were commissioned, including a qualitative study and two quantitative analyses. The qualitative component of the PSIA was hosted by NEC and coordinated with the MoF. A local researchconsulting company was employed and worked with local researchers. The first of the quantitative component was prepared jointly by IFPRI (Washington, DC) and the Centre for Social Research(Zomba). The second quantitative report was prepared by a World Bank consultant. The individual pieces of the analysis were prepared during 2002 and, after several delays, completed in early 2003. This synthesis report, which summarizes the main results of the three background studies, was prepared byWorld Bank staff in mid-2003. Other stakeholders have also been consulted since the beginning. However, the process of consultation inthe design and execution of the study has not been has wide as it could have been. The Malawi ADMARC PSIA has been criticized by several NGOs for the lack of adequate consultation in its early stages. However, it should be noted that this study has been one of the pioneer studies carried out using the PSIA approach, and clear practices on process to be followed were not yet available when the study was started. Since then the Malawi ADMARC PSIA has become a much more inclusive exercise involving various stakeholders from civil society and the donors community. The dissemination of the results, originally scheduled for November 2002, is now planned for late 2003. The dissemination activities are being organized in collaboration with the MOEPD (ex-NEC), in line with the plans discussed with NEC/MOF representatives during the September 2002 annual meetings inWashington DC. Government officials, CSOs representatives and donors, will be invited to participate ina one-day workshop. OXFAM will be invited to present the findings of the ADMARC study they havecommissioned. Notes: 1. See, World Bank, 2001, Malawi: Public Expenditures Review, page 49, paragraph 5.32.

3

methodology to carry out field research in 20 rural communities on the combined effect of the closing of markets and the sharp decline of ADMARC’s marketing activities in recent years.1 1.7 The report is organized as follows. The next section provides a brief background on the structure and role of ADMARC, and its importance in domestic agricultural marketing. Sections 3, 4 and 5 present the results of the three studies. Section 6 provides a synthesis of the findings. 2. THE AGRICULTURAL DEVELOPMENT AND MARKETING CORPORATION (ADMARC)

The Structure and Purpose of ADMARC

2.1 The Agricultural Development and Marketing Corporation (ADMARC) is a parastatal organization created in 1971 with the mandate to market agricultural produce and inputs, and to assist the development of the smallholder agricultural sector through marketing activities and investments in agro-industry enterprises. In addition, ADMARC was mandated with a food security role in maize markets by acting as a buyer and seller in remote areas, providing grain storage across seasons and supporting a large marketing structure with distribution or market centers located throughout urban and rural areas. This function was especially critical at times of maize scarcity. The social role was reflected in the pan-territorial and pan-seasonal pricing system for smallholder farmers, particularly maize, and the establishment of markets in non-profitable areas. 2.2 To fulfill the marketing mandate ADMARC developed an extensive network and infrastructure of markets across the country (Figure 1). The market infrastructure includes regional offices, divisional offices, area offices, storage depots, parent markets, unit markets and seasonal markets.2 These markets are used to conduct sales of inputs, purchase commodities from smallholders, and sell food crops to net consumers. ADMARC currently employs about 3250 permanent employees, with an annual cost structure of around MK500 million.

1 A parallel study on the likely impact of ADMARC’s Privatization has been prepared by Dr Khwima Nthara (formerly at the Department of Economics, University of Malawi) on behalf of OXFAM malawi (Nthara, 2002). The main results of the OXFAM study are summarized below in Section 5, Box 2. 2 Seasonal markets are satellite markets of unit markets. They are mobile selling points that can be opened on demand with a temporary sales force. They operate mainly in the harvesting seasons and are mainly used to purchase produce from farmers. Unit markets have permanent structures such as small storage facility and office, with staff. They generally operate throughout the year and buy produce, sell produce and inputs. Parent markets combine permanent storage facilities with an administrative office that oversees unit and seasonal markets. The current ADMARC structure (as of September 2003) comprises a Head Office (448 employees), 3 Regional Offices (501 employees), 15 District Headquarter Offices with markets (344 employees), 400 Unit Markets (of which 24 parent markets; in total 1448 employees), 300 Seasonal Markets (or selling points; 271 employees, excluding casuals), 9 Storage and Selling Depots (206 employees). Until 2002 ADMARC’s total storage capacity used to reach approximately 468,000 metric tones, or 20 to 25 percent of an annual harvest, but has been reduced to about 180,000 MT with the transfer of storage infrastructures to the National Food Reserve Agency (NFRA). There has been a sharp reduction in the number of marketing establishments during the 1990s. In 1990, ADMARC operated through 1,300 seasonal markets, 217 unit markets, 80 area offices (parent markets), 18 storage depots, 12 divisional offices (district headquarters) and 3 regional offices (ADMARC, 1990).

tivities, ADMARC also invested heavily in
2.3 Apart from agricultural marketing ac
4

equities and loans in various enterprises and is directly involved in estate agriculture.3 By the mid-1980s ADMARC had equity investments in 34 commercial enterprises and owned numerous estates. Until 2002 ADMARC also used to run various subsidiaries (e.g. a cotton ginning company, a bus company, cold storage), many of which were loss-making.4 Recently the Ministry of Finance has assumed control over its four largest loss-making subsidiaries, in preparation for public sale. 2.4 As mentioned above, the financial performance of ADMARC has been increasingly poor in recent years (Table 1).5 ADMARC management indicates that the corporation continues to register operating losses mainly due to the high overhead costs. Such costs are subsidized by some of ADMARC’s other investments.

3 Harrigan (1991) argues that ADMARC used the surplus reaped in the trading activities to invest in industrial activities in various sectors of the economy. Kydd and Christiansen (1982) note that between 1971 and 1979 ADMARC extracted about MK182 million from the smallholder sector, of which 14 percent was used to cross-subsidize smallholder food production and consumption while the remainder was used for equity investments and loans to subsidiaries with only 4.3 percent of such investments related to the development of the smallholder agriculture sector. 4 ADMARC investments currently include: Mitco (100%), Tobacco Marketing Ltd (100%), Malawi Finance Company (100%), Manica Malawi Ltd (50%), Alexander Forbes (49%), Auction Holdings ltd (47%), Mateco, Indebank (27%), National Bank of Malawi (17%), Nedbank (2.7%), Illovo sugar (7%), Clark Cotton (49%), National Investment Trusts Ltd (15%). ADMARC also runs the following support departments: printing services, building services, carpentry services, tailoring services, hospital, football clubs, rice mill, groundnut grading factory, guest houses and cottages, urban housing. 5 ADMARC’s accounts are not regularly audited and, hence, reliable data on the financial performance of ADMARC is not readily available.

Table 1: ADMARC Financial Performance 1998/99 to 2002/03, MK thousands 1998/99 1999/00 2000/01 2001/02 2002/031 MARKETING ACTIVITIES Turnover 1,536,228 3,062,802 1,888,289 576,922 311,440Cost of sales, Contribution, Operational costs, Service costs 1,411,636 3,478,926 2,545,025 OPERATING PROFIT FROM MARKETING ACTIVITIES (Excl. Social costs) 124,592 -416,124 -656,736 Social Markets Costs 190,000 Maize Handling 30,463 63,231 70,000 OPERATING PROFIT FROM MARKETING ACTIVITIES (Incl. Social costs) 155,055 -352,893 -396,736 -771,314 -368,616 OTHER ACTIVITIES: OTHER REVENUES: Dividends receivables, Profit on disposal of fixed assets, and Transfer from deferred income 302,377 146,793 274,760 39,739 92,834OTHER EXPENDITURES: Interest payable on season finance and on LT borrowing, and provision for doubtful debts 80,687 144,643 493,435 166,944 113,576OPERATING PROFITS FROM OTHER ACTIVITIES 221,690 2,150 -218,675 -127,205 -20,742 NET PROFIT BEFORE TAXATION 376,645 -350,743 -615,411 -898,519 -389,358LESS: TAXATION 200 200 200 200 NET PROFIT AFTER TAXATION 376,545 -350,943 -615,611 -898,719 -389,358GDP at current market prices (million MWK) 445,552 916,389 1,090,597 1,445,149 1,984,181ADMARC profits / (losses) as a percentage of GDP 0.6 -0.4 -0.62 -0.6 -0.2Source: ADMARC Annual Financial Statements Notes: 1. Data for actual performance to 10/31/2002 (first half of financial year 2002/03).

2. This amount does not include the losses incurred by NFRA in 2000/01 resulting from ADMARC’s sale of NFRA’s stocks.

5

Figure 1: ADMARC market locations and main road network (Primary and Secondary roads)

Source: This map has been provided by Todd Benson, based on information collected in Benson (2002). A list of 363 markets was provided by ADMARC. Locations were found for 344.

ADMARC’s Role in Agricultural Marketing and Food Security 2.5 The degree of small farmer dependence on ADMARC for the purchase of inputs and marketing of crops has declined steadily since the liberalization of late 1980s. ADMARC’s role in maize trading increased rapidly between 1975 and the late 1980s, but after a few years of significant variability in traded volumes, ADMARC maize trading activities started to decline in the early 1990s (Figure 2). In fact, ADMARC began experiencing financial problems in the early 1980s due to a deterioration in terms of trade (arising from the 1980/81 economic crisis), the change in government pricing policies (the Government reduced ADMARC’s gross maize trading margin from 57 percent in the 1967-79 period to 25 percent in the 1980-86 period), and the and the liberalization of agricultural marketing to private traders in 1987 (Chirwa, 1998; Scarborough, 1990). 2.6 Since 1995, ADMARC purchased a share that varied from 1 to 9 percent of annual national maize production (Table 2).6 However, this corresponds to as much as 30 percent of domestic marketed surplus. ADMARC’s decline in market share applies across all commodities (tobacco purchased from 100 percent to 10 percent; cotton from 100 percent to 50 percent; and groundnuts from 20 percent to an insignificant share; see O&M Associates, 1999). Similarly, ADMARC’s supply of inputs has been reduced to 10 percent of total input supply, but still accounts for a much larger share of smallholders fertilizer and seed use (see O&M Associates, 1999; Kherallah et al., 2001). 2.7 As discussed above, ADMARC traditionally had a food security role as well. This role was executed mainly via the setting of a maize price band, and the import, marketing and storage of grain. Until

6 ADMARC purchases are lowest in marketing year 1997/98, as a result of the poor harvest of 1996/97 (and also 1997/98), and again in marketing year 2000/01 when ADMARC storage facilities where full of the grain purchased in 1999/00. ADMARC eventually sold the entire grain reserves at very low prices just before the poor harvest of 2001/02 (and was left with no stocks to mitigate the impact of the 2001/02 draught). ADMARC’s maize purchases have remained very low in 2001/02 and 2002/03 (not shown), due to the poor harvests and ADMARC’s precarious financial situation.

trategic grain reserve. After that the National
December 2000 ADMARC also managed the s
6

Food Reserve Agency (NFRA) was created to manage the strategic grain reserve and to act solely as a disaster and emergency relief agency, thus taking over the core of ADMARC’s social functions.7

The Impact of Previous Rounds of ADMARC Restructuring

2.8 In line with the Government strategy to facilitate private sector participation in marketing of agricultural commodities and inputs, ADMARC has been restructured several times in the past to progressively liberalize agricultural markets. Research conducted by IFPRI on the impact of these reforms suggests that liberalization has led to an increase private sector activity.8 This data is validated by smallholders surveyed, with 43 percent of 7 The government's strategy under IMF- and World Bank-supported programs in 1998 and 2000 was to end government interventions in the maize market while providing targeted food subsidies to the poor. ADMARC’s intervention distorted prices and other market signals and impeded the development of the market. These intervention were expensive, as the government bore the costs of these interventions by repeatedly bailing out ADMARC. In addition, they rarely were transparent and so raised governance issues. In 1998, the government agreed to eliminate price support operations for maize by ADMARC and prepare it to operate on a strictly commercial basis. It agreed to establish a National Food Reserve Agency (NFRA) to handle disaster relief involving the management of the strategic grain reserve in place of ADMARC, with a clear delineation of responsibilities between the two agencies. At present, however, most of the grain is still stored in ADMARC warehouses. This arrangement has been partially reviewed in 2003 to facilitate the management of relief efforts during the 2001/02 famine and in line with the government’s food security strategy. 8 The Institute for Food Policy Research (IFPRI) and Malawi’s Bunda College of Agriculture conducted a study covering approximately 800 households on the Impact of Market Reforms on Smallholder Farmers. Data collected during this study covered small holders, traders and local market

Table 2: ADMARC Participation in the Maize Markets, 1995-2001 1995 1996 1997 1998 1999 2000 2001

Smallholder Production (millions of metric tons) 1.33 1.79 1.22 1.53 2.48 2.50 1.71

Imports (millions of metric tons) 0.23 0.08 0.05 0.32 0.03 0.00 0.04

ADMARC Purchases as a percent of Smallholder Production 4.60 4.90 7.90 0.80 2.30 7.90 0.10

ADMARC Purchases as a percent of Marketable Surplus 18.40 19.60 31.60 3.20 9.20 31.60 0.40 Source: World Bank (2003) and Ministry of Agriculture, ADMARC, FEWS, NEC.

Figure 3: ADMARC Maize Sales and Purchases, 1970 – 2001 (Metric Tons)

0

200000

400000

600000

800000

1000000

70 75 80 85 90 95 00

Met

ric T

ons

the number of input sellers and 60 percent
households surveyed reporting an increase in
7

reporting an increase in the number of crop buyers. However, the impact of these reforms on different categories of smallholder farmers, and laborers is not clear. For example, some studies suggest that agricultural laborers below the legislated minimum wage are the most food insecure. In 1993, the importation and distribution of fertilizer were opened up to the private sector and in 1994/95 subsidies were reduced to 5 percent. The reduction in subsidies, combined with the devaluation of the Kwacha led to a doubling of the real cost of fertilizer and had a negative impact on fertilizer usage. In addition, inefficiencies in factor input markets, market information, credit delivery, and inadequate infrastructure have constrained sectoral growth, and private traders have not uniformly stepped in when state market presence was removed. 3. QUANTITATIVE ANALYSIS OF DATA FROM THE 1997/98 INTEGRATED HOUSEHOLD SURVEY 3.1 The importance of ADMARC services has been investigated using household data from the nationally representative 1997/98 Malawi Integrated Household Survey (IHS-1). The IHS-1 questionnaire was administered over a 12-month period to 12,960 households, divided in 29 primary sampling strata. Only 6,586 households were judged to have reliable expenditure and consumption information for use in the derivation of the consumption and expenditure based indicator of household welfare, and this is the sample used in this analysis.9

Methodology: An Econometric Model of Household Expenditures in 1997/98

3.2 The impact of ADMARC on household welfare has been assessed by testing for the existence of statistically significant differences in per capita household expenditures10 depending on proximity to ADMARC facilities.11 The general model used to assess the importance of ADMARC is:

)1(33221122110 ilADMARCADMARCADMARCiii eDDDXXyLog +++++++= µδδδβββ where iy is the log of per capita expenditures,

iX 1 is a vector of non-labor assets (in logs, including variables such as land), iX 2 is a vector describing characteristics of household labor (including the level of education and occupation of the household head), 1ADMARCD is a dummy variable that takes a value of one when time to

activity. In areas where survey research was conducted, the number of commodity wholesalers has increased in 59 percent of the locations and retailers in 81 percent of the locations since 1995. The number of input wholesalers and retailers has increased in 46 percent and 70 percent of the areas, respectively, since 1995 (Kherallah et al., 2001). 9 For a full report on the IHS-1 methodology and findings, see National Statistical Office, 2000 and Benson et al., 2002. 10 Household expenditure is a widely accepted proxy for overall household welfare. 11 The impact of ADMARC on demand for fertilizer, maize productivity, and total farm profits has also been investigated. In what follows we discuss only the results of the per capita household expenditures regressions. This is both because this is the most important regression and because it has the most robust specification. The full study is attached as Annex 1.

Table 3: Household distance to nearest ADMARC market (in IHS-1 sample)

Time taken to reach nearest ADMARC

Percent

half hour or less 21 over half hour to 1 hour 25 over 1 hour to 2 hours 29 over 2 hours 25 no response 1 Total 100

Source: IHS-1 data

2C is a dummy variable that takes a value of
nearest ADMARC is less that 30 min, ADMARD
8

one when time to nearest ADMARC is 30-60 minutes, and 3ADMARCD is a dummy variable that takes a value of one when time to nearest ADMARC is 60-120 minutes (Table 3 and Figure 3). The base scenario when distance to ADMARC is more that two hours is embedded in oβ . Finally, iµ is a fixed-effect term specified at the survey enumeration area (EA) level that controls for unobserved location characteristics, and ei is a stochastic error term. 3.3 The sδ thus provide an estimate of the impact of ADMARC. Distance to the nearest ADMARC facility may affect household consumption through several channels. Rural households may utilize ADMARC facilities for buying inputs, selling their output and purchasing maize for self-consumption. All of these functions can be expected to have a positive impact on per capita household consumption. Thus, our null hypothesis is that the ADMARC variables are jointly significant, indicating that access to ADMARC facilities is important. Consistency would also require that

321 δδδ ≥≥ , indicating that, ceteris paribus, the impact is greater for households nearer to ADMARC. 3.4 It is reasonable to expect that ADMARC’s impact on consumption levels will also depend on the availability of alternative marketing channels. The degree of commercialization surrounding the household determines both the opportunities for engaging in trade and the efficiency of those transactions.12 Thus, to test the hypothesis that the impact of ADMARC services is proportional to the degree of remoteness, we have split the sample of households on the basis of proximity to the nearest primary or secondary road13 (to proxy for market accessibility conditions), and have

12 This is because remoter areas have less developed private markets, and alternatives to ADMARC services are less likely to exist there. Higher competition can be expected to lower marketing margins, hence reducing the benefit of access to ADMARC’s pan-territorial prices. 13 In this paper, road access is measured as the straight-line distance in kilometres of the household’s enumeration area to the nearest primary or secondary road (primary roads are generally tarmac roads, while secondary roads are mostly not asphalted). In the first sample, on average households are located about 1.3 kilometres from the nearest primary road. In the second sample, on average households are located about 6.5 kilometres from the nearest primary road.

Figure 3: ADMARC market locations and mean distance from to Primary and Secondary roads

Source: This map has been provided by Todd Benson, based on information collected in Benson (2002). A list of 363 markets was provided by ADMARC. Locations were found for 344.

9

estimated two regressions separately for the two sub-samples.14

Basic Findings from the Quantitative Analysis of 1997/98 Household Survey Data 3.5 The results of the estimation are presented in Table 4. The results underline the importance of household education and access to land as major determinants of household welfare. This finding is in line with previous studies on the determinants of poverty in Malawi (National Statistical Office, 2000). Household consumption is higher by about 5 percent for households whose head has primary education (when compared to households where the head has received no formal education), 17 percent for households whose head has received secondary education, and 50 percent for households whose head has received higher education. The strong importance of household education is also confirmed by the strong statistical significance of the joint F-test on the three education regressors.15 Another major determinant of household income is access to land, with per capita household consumption increasing by 2 to 3 percent for every additional 1 percent of land available. 3.6 Other important factors include the gender of the household head, total household size, and the number of dependents. Per capita household consumption is 8 to 9 percent lower for female-headed households. Household size also has a negative impact on per capita consumption, with per capita consumption lower by about 9 percent for each additional household member. Finally, the dependency ratio of children to adults reduces per capita consumption by almost 5 percent for every 10 percent increase in the share of young children in the composition of the household. 3.7 Distance from ADMARC markets also appears to be significant. Although, only the ADMARC variable for distance less than 30 minutes are individually significant, the F-test of joint significance confirms that proximity to ADMARC has a significant positive impact on household consumption. Specifically, per capita consumption is estimated to be between 11 to 20 percent higher for households that reported living within 30 minutes to the nearest ADMARC facility (compared with households living over 2 hours away), and 4 to 6 percent higher for those living 30 to 60 minutes from the nearest ADMARC facility. Further, the relative magnitude of the three ADMARC variables is logically consistent. It is worth highlighting that ADMARC proximity appears to have a larger (positive) effect in remote areas, where households live far from major access roads.16

14 It should be noted that road-building decisions may be based on existence of other complementary infrastructure or on a region’s agro-climatic potential (Binswanger, et al. 1993). The use of fixed effects allows us to control for these as well as other unobservable factors (such as area-specific price levels, for inputs as well as outputs, and components of area-specific transactions costs 15 The F-test determines whether the estimated impact of several variables together is different from zero (with at least 95 percent probability). The probability that the real value of the coefficient is zero is also referred to as the level of significance. 16 The results of other regressions (not reported here, see Annex 2) indicate that the impact of ADMARC on maize yields, demand for fertilizer and farm profits is significant only for households located in remote areas. This finding suggests that traders engage mainly in buying maize from farmers after harvest and less on selling maize to consumers.

10

4. QUANTITATIVE ANALYSIS OF DATA FROM THE 1997/98 INTEGRATED HOUSEHOLD SURVEY AND THE COMPLEMENTARY PANEL SURVEY. 4.1 The importance of ADMARC for household welfare has also been assessed combining data from the 1997/98 IHS-1 and the fourth round of the Complementary Panel Survey (CPS). The CPS was implemented by the Center for Social Research whereby data

Table 4: Results of regressions to assess impact of ADMARC on household welfare using data from the 1997/98 Integrated Household Survey

Dependent Variable: ln (per capita expenditure)

Near to Main Road Far from Main Road

Coef. t-statistic Coef. t-statistic

Independent Variables:

Household demographic variables from IHS

Female-Headed HH -0.078 (3.01)** -0.093 (3.47)**

Share of children < 5 -0.481 (6.91)** -0.466 (6.78)**

HH size -0.089- (16.81)** -0.092 (17.34)**

Ln (land) 0.029 (1.600) 0.018 (1.030)

Age -0.005 (1.420) -0.000 (0.120)

Age2 0.000 (0.790) 0.000 (0.160)

Occupational Categories

administrative -0.011 (0.150) 0.032 (0.410)

clerical 0.018 (0.280) 0.002 (0.040)

sales 0.037 (0.800) 0.044 (1.010)

services 0.042 (0.840) -0.047 (0.920)

agriculture -0.000 (0.010) 0.051 (1.440)

other 0.006 (0.100) 0.090 (1.400)

School Categories

Primary 0.026 (0.880) 0.071 (2.410)*

Secondary 0.169 (5.510)** 0.171 (5.740)**

Higher 0.509 (11.860)** 0.459 (10.140)**

Time to ADMARC

Less than 30 minutes 0.111 (2.040)* 0.199 (3.650)**

30 – 60 minutes 0.041 (0.830) 0.063 (1.430)

60 – 120 minutes 0.003 (0.060) 0.003 (0.090)

Constant 2.797 (28.560)** 2.580 (26.690)**

Observations 2703 2655

R-squared 0.430 0.450

F-tests

ADMARC (3) 2.16^ 5.70**

Education (3) 52.94* 36.97**

Age (2) 4.87** 0.03

Enumeration area dummies 4.67** 6.50** Note: Absolute value of t-statistics in parentheses: ^ significant at 10 percent; * significant at 5 percent; ** significant at 1 percent

11

was repeatedly collected from a sub-set of the households that participated in the 1997/98 IHS-1. Four rounds of the CPS were conducted between January 2000 and September 2002. The fourth round of the CPS (CPS-4) collected data from 291 households. It was implemented primarily to collect information that would allow household welfare comparisons to be made between the 1997/98 IHS-1 and the 2002 CPS-4. Consequently, the CPS-4 questionnaire was designed to include questions on all of the components of the welfare indicator used in the IHS-1 poverty analysis. CPS-4 also collected information on a number of household transactions with ADMARC, including travel time to the nearest ADMARC facility, role of ADMARC in supplying fertilizers and improved maize seed, and role of ADMARC in providing staple goods for consumption. A detailed description of data is provided in Sharma, Tsoka, Payongayong and Benson (2002).

Methodology: An Econometric Model of Changes in Household Expenditures Between 1997/98 and 2002 4.2 The analysis in this section aims to explain household welfare in 2002 by access to ADMARC services, controlling for other major determinants of household welfare. The general model used to assess the importance of ADMARC is:

)2()(LNEXP97)(log(LNEXP02) ii ii eADMARCA ++++= βX iαδ where LNEXP02 is the (log of) per capita total household expenditure of household i in year 2002, and is a function of a constant term, A, the initial (log of) welfare position of the household in base year 1997, LNEXP97, access to ADMARC, ADMARC, a vector of household characteristics Xi,17 and stochastic error term, ei.18 The parameters δ,α, and ββββ describe the importance of base income in 1997, ADMARC access, and a vector of household characteristics, respectively. It is noted that a positive impact of ADMARC on expenditure levels implies that α<0 since ADMARC is measured as distance to the nearest facility. It is also reasonable to expect the impact of ADMARC on household welfare to be conditional on factors such as access to other types of marketing channels,19 reliability of the household’s own food supply,20 and access to various types of safety net services or social assistance policies.21 Several interactive dummy variables have been included to test these hypotheses. 4.3 Estimating (2) using ordinary least square technique cannot guarantee an unbiased estimate of α as placement of ADMARC across different locations cannot be assumed to be

17 Specifically, three sets of control variables (elements of Xββββ) are included in the regressions: (i) demographic characteristics that have an important bearing on households’ ability to manage economic change, including sector and type of employment; (ii) variables that control for access to resources and institutions that assist households to uphold or smooth consumption in face of changing incomes; and, (iii) changes in key asset variables that are likely to affect income earnings in the household. 18 As mentioned in Sharma, Tsoka, Payongayong and Benson (2002), it is likely that reported 2002 expenditure levels are higher simply because of improvements in survey administration. The systematic part of this conflation factor is also embedded in A. 19 Assuming other private sector market-related institutions are more concentrated around urban centers, then the impact of ADMARC is likely to increase with distance to urban center. 20 All else being the same, households that consistently carry stocks of staple are less likely to go to ADMARC to procure it. 21 The dummy variable for ‘participation in safety net or other social assistance programs’ provides an important control for availability of such programs in the area and as well as of the household’s access to them. Households that benefit from social safety nets are less dependent on marketing their agricultural product (compared to households with no access to such programs). Hence, ADMARC services are likely to be less useful for such households, and its impact, therefore, correspondingly lower.

12

random (see, for instance, Binswanger et al., 1993). It is reasonable to expect that placement decisions be related to local socio-economic-political conditions. For this reason, fixed effects at the traditional authority (TA) level have been specified to sweep out unobserved factors that are likely to be correlated with factors influencing placement. 4.4 In addition, the existence of underlying but unobserved factors that may affect both LNEXP02 and LNEXP97 would result in biased estimates of the regression parameters (see, inter alia, Deaton, 1998). In order to address this problem, a modified version of equation (2) has been estimated, where the term LNEXP97 is moved to the left hand side22:

)3()(logLNEXP97)((LNEXP02) ii iiADMARCA υα +++=− βX i

Results of the Quantitative Analysis of Data from the IHS-1 and CPS-4.

4.5 The results of the regressions indicate that improvement in household welfare between 1997 and 2002 is strongly related to changes in the size of the household, and changes in cultivated land area (Table 5). In a pre-dominantly low-technology agricultural economy like Malawi, increase in the access to cultivatable land is the principal determinant of increases in household welfare, while increases in household size chips away at per capita availability since land, not labor, is the binding constraint on increased production. 4.6 The results of the estimation suggest that proximity to ADMARC has no impact on household income. The ADMARC coefficient is not significant (and carries an unexpected positive sign, suggesting that the further away from ADMARC facilities, the better).23 The estimated value of the joint effect of proximity to ADMARC and distance from urban centers suggests that the impact of ADMARC decreases with distance from urban centers. This is contrary to our expectations (and also contrasts with the results presented in Section 3), as it suggests that proximity to ADMARC is less important in remote areas. The estimated impact is not statistically different from zero, however. The other interactive variables also suggest that ADMARC is less important for those households that are able to carry stocks of the basic staple and is higher for those households that are in the position to participate in social welfare programs. None of these estimated impacts is statistically different from zero, however. The results, therefore, do not provide any evidence of ADMARC’s impact on household welfare.24 4.7 However, it must be noted that CPS-4 survey consists of a fairly small sample size, and that it suffered from a high level of drop outs in remote rural areas, as the interviews had no resources for a second visit in case they could not meet the respondent during a first trip. Further, data collection coincided with an emerging income and food crisis in Malawi. In fact, the timing of the study appears to explain the counter-intuitive results. Since ADMARC 22 The dependent variable is thus defined as LOGDIFF = LNEXP02-LNEXP97 23 For the same reason, even though the F-test on the joint significance of all ADMARC coefficients rejects the null hypothesis that ADMARC proximity has no impact (at 10 percent probability level), interpretation remains problematic. 24 It should be noted that proximity to ADMARC appears to be weakly significant in the results of the regression in levels (Equation 2). The F-test of the joint impact of all ADMARC coefficients rejects the hypothesis of zero impact with a 10 percent probability. In terms of the individual impacts, the impact of proximity to ADMARC is not statistically different from zero but has the expected negative sign, implying that the impact of ADMARC decreases as distance to ADMARC increases. The interactive impact of proximity to ADMARC and distance from urban centres is statistically different from zero, and suggests that the impact of ADMARC decreases with distance from urban centres (reducing to zero as distance approaches about 26 kilometres). As discussed, this is contrary to our expectations. The results on the other interactive dummies are similar to those for the preferred model presented above.

T

lnex

adm

Hou

ageh

dpnd

hhhl

farm

man

prof

max

Acc

stoc

pclv

socc

snpa

Cha

hhch

chan

Inter

adm

adm

admc

distu

Note

13

did not have much to offer during the study season (due to its financial problems), it is not surprising that the impact of ADMARC decreases with distance. More generally, it could well be that the impact of ADMARC was simply drowned out by the far more overwhelming effects of the recent deep decline in agricultural production 4.8 In sum, the results of the analysis presented in this section (of the IHS-1 and CPS-4 data) do not appear to support a statistically significant impact of differential access to ADMARC facilities (and services) on changes in household welfare. However, put against the weaknesses of the CPS dataset and the timing of the study, it is not surprising that strong

able 5: Results of regressions to assess impact of ADMARC on household welfare using data from the IHS-1 and CPS-4.

Regression on logdiff:

difference between logged CPS4 and logged IHS welfare

indicators (ln MK)

Regression on lnexp02: natural log of CPS4 welfare

indicator (ln MK)

Coef. t-statistic Coef. t-statistic

p97 natural log of IHS welfare indicator (ln MK) 0.349 (5.06)** - -

arc distance to nearest ADMARC (km) -0.01 (0.43) 0.019 (0.75)

sehold demographic variables from IHS

hh age of head of household -0.001 (-0.46) -0.003 (-0.92)

rat dependency ratio: dependents/hh -0.035 (-0.20) 0.422 (2.17)*

it literate head of household (0/1) -0.027 (-0.24) -0.147 (-1.13)

er hh member with agricultural occupation 0.078 (0.62) 0.215 (1.48)

ufact hh engaged in manufacturing -0.260 (1.73)^ -0.208 (-1.18)

hh member with professional, ad -0.154 (-1.06) -0.356 (2.12)*

yrsed maximum years of education in hh 0.032 (2.27)* 0.020 (1.24)

ess to key consumption smoothing assets or institutions

k variable that equals one if hh owned stocks of staples at time of interview in all of the four CPS survey rounds (0/1) 0.304 (1.94)^ 0.131 (0.72)

sval per capita livestock value (MK) -0.00004 (-0.65) -0.00006 (-0.93)

ap hh reported in IHS giving or receiving income transfer (0/1) 0.043 (0.47) -0.202 (1.95)^

rtic hh participated in at least one safety net or social assistance program as reported in CPS1 (0/1) 0.106 (0.55) 0.297 (1.34)

nge in access to key human and physical assets

ange change in hh size betweeen IHS and CPS4 -0.095 (4.60)** -0.145 (6.21)**

ge_cult change in cultivated land area of hh betweeen IHS and CPS4 (acres) 0.05 (2.65)** 0.05 (2.30)*

acted variables

arc_disturb interaction of distance to ADMARC with distance to nearest urban center 0.0004 (1.93)^ 0.0002 (1.11)

arc_stock interaction of distance to ADMARC with whether hh usually possesses staple food stocks 0.014 (0.88) 0.025 (1.41)

arc_snparti interaction of distance to ADMARC with whether hh participated in safety net programs -0.021 (-1.17) -0.037 (1.75)^

rb distance to nearest urban centre (km) -0.0003 (-0.04) -0.005 (-0.69)

constant 1.621 (3.29)** 0.257 (0.47)

R2 0.244 0.213


Number of TA fixed effect variables 19 19

: t-statistics in parentheses: ^ significant at 10 percent; * significant at 5 percent; ** significant at 1 percent

14

evidence for or against the continued operation of ADMARC cannot be found in the analyses of the CPS data presented here. In light of these considerations, the results presented in this section should not be interpreted to indicate that there is no significant impact of ADMARC on household welfare, but rather simply an inability of the dataset to provide useful information on the role of ADMARC. 5. QUALITATIVE STUDY ON THE IMPACT OF ADMARC FACILITIES AND SERVICES, AND THE ROLE OF THE PRIVATE SECTOR IN AGRICULTURAL MARKETING 5. 1 The third study was designed to carry out an ex-post assessment of the impact of closing selected ADMARC markets in recent years on the various social groups. Participatory learning methods were used to explore well-being categories, poverty dynamics and the relation to ADMARC’s marketing services. The results also provide valuable insights into the operation of the private sector in maize markets. It should be highlighted, however, that the results of the study cannot distinguish between the combined impact of the closure of seasonal markets and the sharp decline in ADMARC’s business activities during the preceding years (see Figure 2 above).

Methodology: A Qualitative Study of the Impact of ADMARC Closure in 20 Villages 5. 2 The study was conducted within the catchment areas of 10 ADMARC unit markets (Table 6). A range of qualitative research methods were used to solicit households’ and other stakeholders’ reactions. Focus group discussions using variants of participatory learning methods were the main method for collecting data.25 Interviews with key informants in the selected sites were also conducted.26 Finally, semi-structured interviews with policy and decision-makers in stakeholder institutions were also carried out.27 5. 3 In each of the 10 ADMARC unit markets, participatory rural appraisal methodologies, focus group discussions and key informant interviews were conducted in two villages. This implies that a total of 20 villages were included in this study, resulting in a total of 40 focus group discussions (half with women groups), 54 semi-structured interviews and 44 key informant interviews (of which 9 interviews with ADMARC officials, 3 interviews with private traders and 32 interviews with other key informants). The fieldwork was carried out between April and June 2002. 25 The selection of interviewees and discussion partners accounted for differences with respect to social and economic stratification, ethnic and livelihood diversity and gender and this was done in consultation with local leaders and local extension staff. The discussions focused on perceptions on well-being and trends in the last five years, livelihood strategies and cropping patterns, the problem analysis, cause-effects diagrams, trend analysis, institutional analysis and analysis of opportunities, coping and survival strategies. 26 Different stakeholders existing or operating in the areas were interviewed, such as traditional leaders, private traders, shop-keepers, extension agents, food security, and non-governmental organizations. In each location at least 12 key informants’ interviews including key informants in the surroundings of the sites were conducted, with at least one informant for each well-being group identified by the community in focus group discussions. 27 The institutions and stakeholders visited included ADMARC management, National Association of Smallholder Farmers (NASFAM), National Economic Council, Grain and Milling Limited, Smallholder Farmer Fertilizer Revolving Fund (SFFRF), World Food Programme (WFP), Bharat Trading Limited, Ministry of Agriculture and Livestock Development. The purpose of the institutional interviews was to assess the objectives, motivations, interests, concerns and resistance of different stakeholders and institutions to the closure of ADMARC markets.

T

S

15

5. 4 The 10 ADMARC unit markets include 6 unit markets that had been proposed for closure and 4 unit markets that should continue to operate. In the 6 unit markets earmarked for closure, interviews were conducted in villages in the catchment area of closed seasonal markets, as it turned out that there were no unit markets that had been closed in the past five years. 5. 5 The location of the ADMARC markets was selected through purposive sampling on the basis of their remoteness and their dominant livelihood characteristics to cover the existing diversity in terms of regional and geographical criteria, farming and livelihood systems as well as other social and cultural criteria.28 The rationale for choosing remote markets underlies the assumption that ADMARC's marketing services would make the biggest difference in areas where infrastructure is poor, because private traders would be less likely to operate in those areas. As such, all ten sites are in remote rural areas and characterized by very poor road infrastructure, and therefore difficult to access particularly in the rainy season (except the Eswazini site). All except three sites had a primary school located within the community. However, other essential infrastructure such as a market place, health facilities, police posts and post offices were often located far from the communities, generally between five and fifteen kilometers away. All ten sites present a fairly diversified agricultural sector, in terms of both crops and livestock types. There are also a significant number of non-farm activities. Staple food production did not differ dramatically across the ten sites, as they all produced maize as well as other food crops such as groundnuts, cassava and sweet potatoes. However, there are considerable differences in cash crops, with three sites producing cotton and the rest being more involved in tobacco, but with some households also producing other cash crops such as soybeans and paprika.

28 Note however that the sample was limited to remote rural areas; no urban, semi-urban, or locations close to a main road, were included.

able 6: Location and basic infrastructure of sites selected for qualitative study

Site District Region Status Roads School ADMARC Market

Health Facility

Post Office

Police Post

Nthalire unit market (Kamphyongo seasonal) Chitipa North Closed Difficult to

access In village 3km 15km 14km 15km

Mpata unit market Karonga North Open Fairly

accessible 2km 15km 15km 15km 15km

Eswazini unit market (Emoneni seasonal) Mzimba North Closed Accessible In village 11km 11km 11km 62km

Chikho unit market (Mzandu seasonal) Ntchisi Centre Closed Difficult to

access In village In village 1km 34km 34km

Lipiri unit market (Kamtepa seasonal) Dowa Centre Closed Difficult to


Mikundi unit market Mchinji Centre Open Fairly


Chiondo unit market Nkhotakota Centre Open Fairly

accessible In village 5km 5km 5km 5km

Kasongo unit market Phalombe South Open Difficult to


Lundu unit market (Mkumba seasonal) Blantyre South Closed Difficult to


Kakoma (Tombondera seasonal market) Chikwawa South Closed Fairly


ource: Poverty and Social Impact Assessment 2002

16

Basic Findings of Qualitative Study: The Impact of ADMARC’s Closure and the Role of the Private Sector in Agricultural Marketing in Rural Areas

5. 6 The results of the interviews and focus group discussions indicate that private traders, both large-scale and small vendors, are progressively replacing ADMARC as buyers/sellers of crops, especially in markets with well-developed infrastructure. Large-scale traders are not many, and specialize mainly in the purchase of crops, particularly cash crops. Small-scale unlicensed traders (or small vendors), on the other hand, are very common and provide an accessible marketing channel for buying and selling of maize in the rural areas. It is these small vendors that to some extent have bridged the gap left by ADMARC’s inability to provide reliable and efficient marketing services. Small vendors tend to have more regular supplies of maize than ADMARC markets.29 Also, ADMARC often does not have adequate cash to buy produce from farmers. However, since most small traders work at the local level and almost exclusively engage in short-term arbitrage, they are unable to guarantee food security in times of shortage. Thus the function of inter-regional and inter-seasonal arbitrage performed by ADMARC has not been taken over by these smaller agents.30 5. 7 Compared to crop produce market, private sector involvement in the agricultural input market remains highly underdeveloped, with only a few private traders selling inputs in rural areas. Small vendors do not sell inputs to farmers, since most are mobile vendors who do not meet the capital and storage requirements of the input business. The agricultural input market in the rural areas is still dominated by ADMARC with a few large-scale wholesalers or limited companies in selected sample sites. Since most of the private traders that deal in inputs are located in urban and peri-urban centers, closure of markets has led to an increase in transport and transaction costs of input procurement for a large proportion of the rural population. 5. 8 Further, the withdrawal of ADMARC does not appear to be compensated by an equivalent increase in private sector activity, resulting in overall lower competition (and efficiency) in marketing institutions. On the contrary, the immediate effect of closure of a seasonal market or decline in ADMARC business is the reduction in marketing institutions and hence a more concentrated market structure. Since farmers are not organized in marketing cooperatives or associations, the few private traders that access remote rural markets experience an increase in their monopsony power in buying produce from smallholder farmers, and an increase in their monopoly power in the sale of maize and inputs.31 This finding is similar to the results of the Oxfam-financed study by Nthara (2002)

29 Small vendors appear to be active in all villages and now constitute the preferred marketing channel for selling of produce. During the recent famine, vendors operated in all villages and had maize to sell throughout the famine, albeit at unaffordable high prizes. Even after the government made available through ADMARC maize at subsidised prices, small vendors were able to undercut the subsidized price offered by ADMARC and continued to be the main provider of maize in rural areas (apart for humanitarian assistance). 30 It is important to highlight, however, that ADMARC has in fact been unable to buy or sell significant quantities of maize during the recent food crisis, due to its precarious financial situation. Hence, the perception of ADMARC role reflected in these findings may be based on ‘golden age’ memories. 31 The concentration of market power and/or the reduction in marketing channels resulting from the curtailment of ADMARC subsidized marketing activities has several negative implication for smallholder farmers. First, the closure of ADMARC seasonal markets leads to an increase in the distance to markets for smallholder farmers who continue to rely on ADMARC to purchase maize and farm inputs. This raises their transaction costs in terms of search and transport costs and opportunity cost of traveling to distant markets. Secondly, the monopsony power of private traders leads to lower produce prices. This is because smallholder farmers have low bargaining power owing to the absence of marketing associations. Thirdly, the lack of access to markets deprives the farmers of a reliable and affordable source of inputs. Fourthly, market closures lead to higher cost of inputs in the sense that

17

who observes that the number of private traders in particular markets is not positively correlated with the absence of ADMARC markets (Box 2).32 This finding is also confirmed by the observation that in many closed seasonal markets (and in the many non-functional unit markets), rural households now use the services of ADMARC in alternative active markets, and the distance traveled is longer than before seasonal markets were closed. This implies that immediate effect of the closure of markets is to reduce the profitability of smallholder agriculture, as reflected both in the increase in transaction costs for farmers and higher margins for the traders.33 5. 9 The price competitiveness of private traders is also an important indicator. In contrast to ADMARC’s stable prices, small vendors maize prices tend to be very volatile from one community to the next, or even within the same community. The high volatility of prices reflects the fact that most private traders face little local competition and are able to operate as monopsonists/monopolists in the local maize markets. This supports the findings of Fafchamps and Gabre-Madhin (2001) who find relatively high profit margins for Malawian traders, suggesting little competition amongst traders, and therefore low efficiency of the private marketing system in Malawi.34 Where active, therefore, ADMARC also plays an important role in benchmark prices through open access to information to smallholder farmers, particularly during the period when ADMARC markets are operation. 5. 10 Finally, another issue of concern in the existing marketing systems in the rural area is the lack of regulation and enforcement of fair trading practices of private traders. The poor quality of business practices negatively affects the efficiency of private marketing channels. It also leads to widespread claims of cheating on measurement and weights on the part of small vendors. Similarly, the lack of clear standard for quality grading gives rise to many commercial disputes. 5. 11 In sum, ADMARC markets in rural areas appear to play an important role as distribution networks for affordable maize in the lean season and in times of famine, in providing benchmark prices, in providing a reliable source of inputs, and in the purchase of crop produce from farmers. Even in markets where private traders are particularly active, notably border markets, the withdrawal of ADMARC markets may have negative consequences for food security and regular access to input.

smallholder farmers can either travel long distances to procure inputs at alternative ADMARC markets or buy inputs from private traders at higher prices. Finally, smallholder farmers also identified higher consumer prices of maize as one of the effects of ADMARC’s declining business. Private traders take advantage of the absence of ADMARC (or shortage of maize supplies at ADMARC) to charge higher prices than would have been the case if maize were available at ADMARC markets. 32 The study by Nthara (2002) also points to the important role ADMARC plays in food security particularly with respect to maize sales and input markets. The study finds that ADMARC’s role in purchasing agricultural produce from smallholder farmers is limited even in markets where it is still operational due its price uncompetitiveness, its late opening of markets in the season and its lack of cash for most of the buying season. However, with respect to selling food commodities to food deficit households and selling inputs to smallholder farmers, Nthara (op.cit.) finds that ADMARC still plays a critical role (and this is also confirmed by the findings of our qualitative study). 33 This is reflected in the negative views expressed in the focus group discussions, which suggest that a range of negative effects has been experienced by communities (and particularly smallholder farmers) as a result of marketing liberalisation (as experienced both through the closure of ADMARC seasonal markets and the decline in ADMARC’s business at unit markets). 34 Fafchamps and Gabre-Madhin (2001) show evidence that the private marketing system is dominated by petty traders with substantial financial and capacity constraints, and that inter-seasonal and inter-regional arbitrage does not form part of the activities of this class of traders. They note that the average private trader in Malawi operates within a radius of 53 kilometres. They also find evidence that the efficiency of private traders is constrained by the high transaction costs in form of search and transport costs.

18

5. 12 However, given the qualitative nature of the study, caution must be exercised in generalizing the results of the study. Firstly, it is important to note that the perceptions on closure of markets cannot be filtered out from the effects of other factors that are affecting ADMARC’s performance. Moreover, the study was conducted against the background of the famine that engulfed Malawi and southern Africa in the 2001/2002 season, and some of the views from the smallholder farmers could be influenced by the famine situation. Further, the study focuses on ten sites of the many markets ADMARC has throughout the country, and therefore, the results only allow some general insights in the operations of agricultural marketing systems in Malawi.35 In addition, there were no closed unit markets in the sample, and the impact of closure of unit markets is inferred from closed seasonal markets and the general situation of declining ADMARC business activities in most of its markets. Also, very few interviews were conducted with private traders due to the high mobility of private traders and the fact that most of the private traders were not resident in the villages where the interviews were conducted.

35 Purposeful sampling of a few villages was used to provide ‘ideal type’ situations and to explore typical situations in remote rural areas; therefore, the sample was not drawn to be statistically representative at the national level.

Box 2: Basic findings of OXFAM study on ADMARC’s Privatization “What needs to be done to improve the impact of ADMARC on the poor”. By Khwima Nthara (Dept of Economics, University of Malawi), July 2002. Prepared for Oxfam Malawi. The study involved a survey of rural poor (via a household questionnaire and focus groups) and aliterature review of other countries’ experiences with marketing boards. The survey targeted areas inall the three regions of the country where ADMARC is still operational and where ADMARC hasclosed its markets. A number of results emerge from the survey: � ADMARC no longer plays the crucial role that it used to play in the past when it comes to the

selling of agricultural produce. � ADMARC still plays a very crucial role with regard to the buying of food by deficit households.

Even in areas where ADMARC withdrew, people still depend on the next alternative ADMARCmarket. ADMARC sells maize at a much lower price than private traders, but usually runs out of stocks. As a result, people are forced to buy from private traders, though at exorbitant prices.

� ADMARC still plays a very important role in the provision of farm inputs in many areas,especially in those areas that are remote, despite the coming in of private traders. For those areas where agricultural activity is high and road conditions are good, there is evidence that privatetraders have displaced ADMARC. People in these areas prefer to buy from private traders becausetheir prices are lower.

� Based on people’s experiences with private traders so far, and the experience of other countries inthe region, full privatization of ADMARC would have adverse effects on the poor. In the absenceof ADMARC, only areas where agricultural activity is high, and where road conditions are good will be serviced. In remote areas, there will be very thin markets for produce, food, and inputs. Asa result, the level of prices there will be unaffordable for the poor.

The study acknowledges that maintaining the status quo would not be the best way forward. Although ADMARC has played an important social role, this has cost the government, and hence the poor, morethan would have been necessary. This is mainly because of political interference in the operation of ADMARC’s social activities. At the same time, the study contends that total privatization is not apanacea either. In fact, it is likely that a privatized ADMARC may decide to pull out of those areaswhere it not making a profit. The study’s proposed solution takes the form of a joint venture between the government and the private sector. The private sector would have a majority shareholding inADMARC, with the government having some direct influence through its board of trustees, in order tosafeguard the interests of the Malawian people. Although this arrangement amounts to privatization ofADMARC, the form is a qualified one.

19

6. CONCLUSIONS OF THE PSIA STUDY ON AGRICULTURAL MARKET REFORMS 6.1 Three studies have been carried out to investigate the impact of access to ADMARC on household welfare. Two studies use econometric techniques on data from the 1997/98 Integrated Household Survey and the 2002 Complementary Panel Survey (4th round), while the third study uses an array of qualitative research methods to solicit households’ and other stakeholders’ reactions during field interviews in 2002. 6.2 The results of the analysis indicate that access to ADMARC has a positive impact on household welfare. Specifically, the results of the quantitative analysis indicate that per capita consumption is estimated to be up to 20 percent higher for households living closer to ADMARC facilities. This reflects the fact that ADMARC facilities provide rural households with access to a fair market for buying inputs, selling their output and purchasing maize for self-consumption.36 Similarly, the qualitative study found that rural communities, particularly in remote areas, feel they have been adversely affected by the reduction in ADMARC marketing activities (which has resulted from both to the closure of ADMARC markets and the decline in the business activities of ADMARC due to its financial and operational problems). 6.3 The second major finding of the analysis is that the beneficial impact of proximity to ADMARC markets is more important in remote rural areas, far from major access roads. These benefits are attributed to the fact that presently agricultural market institutions and infrastructure are not developed in poor, remote areas. Households that are located in areas where infrastructure is the most readily accessible are less dependent on ADMARC. In areas where private market infrastructure is developed and market services function, ADMARC facilities do not have a significant impact on farm households. In other words, ADMARC’s impact depends on the availability of alternative marketing channels. The degree of commercialization surrounding the household determines both the opportunities for engaging in trade and the efficiency of those transactions. More remote areas have less developed private markets, and alternatives to ADMARC services are less likely to exist there. In less remote areas, higher competition can be expected to lower marketing margins, hence reducing the benefit of access to ADMARC’s pan-territorial prices. 6.4 The third finding is that the role of ADMARC is especially important in maize sales during the hungry season, but is much less important during the harvesting season. This reflects the fact that small traders engage mainly in buying maize from farmers after harvest, and less on selling maize to consumers during the lean season. They transport maize out of remote areas and do not generally engage in storage activities. Similarly, small traders are not generally involved in agricultural input supply in remote areas. Thus, although the private sector has emerged to take advantage of the opportunities created by the liberalization of agricultural markets, it has not been able to completely fill the gap left by ADMARC. 6.5 The above findings suggest a number of policy recommendations, some of which can be implemented in the short term, while others will require a longer time horizon: Short-Term Policies:

36 The analysis of the 2002 CPS-4 data, however, does not indicate any significant impact of access to ADMARC on household welfare. However, it must be noted that the CPS-4 survey consists of a fairly small sample size, and also that data collection coincided with an emerging income and food crisis in Malawi. It could well be that the impact of ADMARC was simply drowned out by the recent overwhelming decline in agricultural production. Put against this perspective, strong evidence for or against the continued operation of ADMARC cannot be found in the analyses of the CPS data.

20

• Completely eliminating ADMARC facilities may ignore the fact that well-functioning market institutions and infrastructure do not exist uniformly throughout the country. In the short-run, the provision of ADMARC facilities may be beneficial in some areas that are under-served. An analysis of the location of ADMARC’s markets will indicate which facilities are located in ‘remote’ areas and should continue to operate. The question then becomes designing a satisfactory scheme for dealing with government mandated ‘social marketing activities’ in remote markets (essentially the subsidized purchase and sale of maize). Whether ADMARC can play this useful function in a cost-effective manner, or if this social function could be done better through other market-based mechanisms needs to be assessed. In broad terms, two main options for providing the social services have been identified so far. However, neither their operational feasibility, nor the advantages and disadvantages of each, have yet been assessed: � Option A: The establishment of a specific (public) enterprise responsible for

undertaking ‘social activities’ on behalf of government. This organization would: (i) take over and manage ADMARC’s physical infrastructure located in clearly identified ‘remote/social markets’; (ii) have a lean work force; (iii) be managed transparently on the basis of clear policies and procedures; and, (iv) be funded through a subsidy specifically identified in the Government budget. Ideally, although the infrastructure may remain government property, the management of the ‘social ADMARC’ should be sub-contracted to a credible and independent third party. An analysis of ADMARC’s financial operations will provide some information on the extent of overheads and administration infrastructure (and personnel) needed to operate the ‘remote/social’ market facilities and the amount of public subsidy required.

� Option B: The tendering by government of specifically identified ‘social marketing programs’, on a least-cost basis, to any eligible enterprise (including ADMARC). This option would in theory be more transparent and efficient and it wouldn’t require any implementation capacity on government side. However, it may be more difficult to implement, essentially because of the difficulty for government to make sure that commercial operators carry out the agreed upon activities. Hence, it would require a heavy dose of monitoring by a credible neutral party.

• ADMARC services are not required in less remote areas of the country, where the activity

of the private sector is already well established and provides for fairly efficient transactions. Therefore, ADMARC’s facilities in developed parts of the country (notably in urban and semi-urban areas) should be auctioned off to the private sector. Similarly, all of ADMARC’s other non-marketing activities should also be sold to private operators. These activities are not needed for the operation of the ‘remote’ markets, and hence should be allowed to operate as private companies. Such a decision will prevent the possibility that unexpected financial losses originating from poor management of ADMARC’s non-social functions may again in the future have to be bailed out using public resources (from the budget).

Medium Term Policies: • Market failures in remote areas should be addressed by investment in rural infrastructure

and facilitating the development of private sector. There is a role for Government to invest in market infrastructure and the development of institutions that facilitate greater transparency and competition in market transactions. In the medium term, these measures will eventually eliminate the need for ADMARC services in remote areas as well. This would enhance private sector participation in the agricultural sector and eliminate the economic inefficiencies currently associated with ADMARC’s activities.

21

• There is a need to study the fertilizer supply chain to identify what policy or structural constraints prevent private traders from entering the market. Similarly, a separate analysis is needed to investigate the reasons behind the lack of private investment in storage facilities and interregional and/or seasonal arbitrage. Possible issues may include high interest rates, high capital requirements (e.g. for storage), and low repayment capacity of end consumer. Follow up actions would depend on the recommendations of the two studies.

Next steps required to implement the findings of the PSIA on the restructuring of ADMARC: Substantial preparation will be needed to implement the findings of this PSIA: �� Firstly, the criteria to be used in the selection of the ‘remote/social markets’ need to be

agreed and the list of the remote markets to be targeted finalized. �� Secondly, the various options for the provision of the ‘social marketing functions’ have to

be reviewed in detail, with technical assistance from expert consultants. The consultants should review the various operational options available and provide an assessment of the advantages and disadvantages of each, together with a reasoned recommendation as to which system should be adopted. The choice of operational system for the social functions should be based on a clear set of criteria to ensure the transparent and cost-efficient provision of the social services. The following principles provide the type of guidance needed to determine the operational aspects of the ‘social marketing functions’:

o The social marketing services are by nature temporary and will be phased out as soon as access to the selected remote locations is improved and marketing margins are adequately reduced.

o The target local markets will be selected on the basis of clearly established criteria to be identified.

o The coordination of social activities in these markets will be carried out by a suitable public institution, which may assume ownership of the limited infrastructure necessary at village level.

o Social activities will be carried out in a way that doesn’t impede the development of an efficient maize market and the development of private sector activities.

o Actual implementation of these activities will be carried out by eligible private sector operators and/or NGOs, including the management of the necessary infrastructure, equipment and staff, under contractual arrangements allocated through a competitive process.

o The public subsidy necessary to cover the cost of these social activities will be capped and explicitly included in the Government’s annual budget.

o An efficient mechanism will be established from the beginning of the scheme to monitor the justification of the selection of the ‘social markets’ and the transparency and efficiency of the social activities.

• Thirdly, all of ADMARC’s non-social-marketing activities should be restructured, and

sold to private operators. These activities should be allowed to operate as private companies. For the restructuring to be operationalized, a detailed business and financial plan needs to be prepared by ADMARC’s management with technical assistance from financial and management experts. Once Government has approved it, the plan will form the basis for the restructuing of ADMARC’s operations along commercial lines. The restructuring should aim to streamline and rationalize the non-social-marketing operations of ADMARC, to improve its efficiency, and with a view to prepare the groundwork for its prvatization.

22

REFERENCES

Abbott, R. D. and Poulin, R. J. (1996). The Future of ADMARC: A Policy Analysis, Agricultural Policy Analysis Project (APAP) Research Report No. 1012 (USAID) ADMARC (1990). ADMARC Annual Report and Accounts 1990, Blantyre: ADMARC ADMARC (2002). Strategy to turn around ADMARC. November 2002, Lilongwe. Benson T. (2002). Malawi: An Atlas of Social Statistics. National Statistical Office, Government of Malawi and the International Food Policy Research Institute. Benson T., Kaphuka J., Kanyanda S., and Chinula R. (2002). Malawi: An Atlas of Social Statistics. National Statistical Office, Government of Malawi and the International Food Policy Research Institute. Binswanger, H., S. R. Khandakar, and M. Rosenzweig (1993). How infrastructure and financial institutions affect agricultural output and investment in India. Journal of Development Economics, Vol. 41 (1993), pp. 337-366. Chirwa, E. W. (1998) ‘Fostering Private Food Marketing and Food Policies after Liberalisation – The Case of Malawi’, in Seppala, P. (ed.) Liberalized and Neglected? Food Marketing Policies in Eastern Africa. Helsinki: United Nations University/WIDER Deaton, A. (1998). The Analysis of Household Surveys. Published for the World Bank: The Johns Hopkins Press. Baltimore. Fafchamps, M. and Gabre-Madhin, E. (2001) Agricultural Markets in Benin and Malawi: Operation and Performance of Traders, World Bank Working Paper Series No. 2734. Washington, D.C. Government of Malawi (GoM) (2002). Malawi Poverty Reduction Strategy Paper, Lilongwe: Ministry of Finance and Economic Planning Harrigan, J. (1991) Malawi. In: P. Mosley, J. Harrigan and J. Toye (eds.) Aid and Power: The World Bank and Policy Based Lending, Volume 2 Case Studies. London, Routledge Kherallah M., Minot N., Kachule R., Soule B.G., Berry P. (2001). Impact of agricultural market reforms on smallholder farmers in Benin and Malawi. Final Report, Volumes 1 & 2. IFPRI, Washington D.C. Kydd, J. and Christiansen, R. (1982) Structural Change in Malawi since Independence: Consequences of a Development Strategy based on Large-Scale Agriculture. World Development, 10 (5), 355-75 National Statistical Office (NSO) (2000). The Poverty Analysis of the Integrated Household Survey 1997-98 - The State of Malawis Poor: The Incidence, Depth, and Severity of Poverty. Available at http://www.nso.malawi.net Nthara, K. (2002). What Needs to Be Done to Improve the Impact of ADMARC on the Poor, Phase II Final Report prepared for the Joint Oxfam Programme in Malawi, Blantyre. O&M Associates (1999). Commercialization and privatisation programme for ADMARC. Draft Final Report prepared for the Ministry of Finance, Lilongwe.

http://www.nso.malawi.net/

23

Scarborough, V. (1990) Domestic Food Marketing Liberalization in Malawi: A Preliminary Assessment, ADU Occasional Paper 13, London: Wye College, University of London Sharma, M., M. Tsoka, E. Payongayong, and T. Benson. 2002. Analysis of poverty dynamics in Malawi with an assessment of the impact of ADMARC on household welfare. Draft report submitted to the World Bank. World Bank (2003). Malawi Country Economic Memorandum: policies for accelerating growth. Report No. 25293-MAI. Africa Region, PREM 1. Washington D.C.

24

25

ANNEX A - QUANTITATIVE STUDY PREPARED BY IFPRI AND CSR

26

An Assessment of the Impact of ADMARC on welfare of Malawian

Households

Final Report Submitted to the World Bank

2033 K St. NW Washington, DC 20006

February 2003

By

Manohar Sharma, Maxton Tsoka, Ellen Payongayong, and Todd Benson1

ACKNOWLEDGEMENTS The first three rounds of the Malawi Complementary Panel Survey (CPS) were implemented as an activity of the Poverty Monitoring System of the Government of Malawi. The final round of the CPS and the analysis presented here were carried out as activities under a research contract between the International Food Policy Research Institute (IFPRI) and the World Bank and a research agreement between IFPRI and the Centre for Social Research (CSR) of the University of Malawi. Our thanks to all of the institutions involved. However, all views, interpretations, recommendations, and conclusions expressed in this report are those of the authors and not necessarily those of the Poverty Monitoring System, the Government of Malawi, the World Bank, CSR, or IFPRI.

(World Bank ref. 7120350; IFPRI ref. 2767-001)

1 Sharma and Benson are Research Fellows with the Food Consumption and Nutrition Division at IFPRI. Sharma is the research project manager. Tsoka is a Research Fellow at the Centre for Social Research of the University of Malawi; PO Box 278, Zomba, Malawi, and was responsible for implementing the four rounds of the panel survey. Payongayong is a Research Analyst with the Mozambique Food Security Project of Michigan State University, based in Maputo. Until recently, she was with IFPRI and was responsible for assisting CSR with the data processing of the CPS.

27

A1. Background

The Agricultural Development and Marketing Corporation (ADMARC) was established in 1971 in Malawi as an institution primarily responsible for providing marketing services in the country’s agricultural sector. Its main activities have included: providing reliable markets and price support for smallholder produce, provision of credit and subsidized inputs, and investing in agro-industry enterprises. ADMARC has established a nation-wide web of infrastructure consisting of offices, storage depots, and marketing centers. By the 1980s, it had also made direct investments in about 34 commercial enterprises, and a number of agricultural (mostly tobacco) estates.

Performance of ADMRC has been a source of concern for sometime now, however. ADMARC’s financial losses have mounted since the 1980s, and concerns have been raised both about ADMARC’s operational inefficiency and the extent to which ADMARC actually provides useful marketing services to smallholder farmers and poor consumers. The Agriculture (General Purpose) Act of 1987 liberalized agricultural markets by allowing licensed private traders to operate; markets were further liberalized in 1996 when the licensing system was abandoned. Sale and purchases of ADMARC of maize in recent years are considerably below levels in the periods before the mid-nineties. ADMARC has also divested its ownership in commercial investments and agricultural estates, even though some enterprises were later re-purchased and new investments in agricultural estates were again made.

As a part of its on-going reform program, the Government of Malawi (GOM) is considering the further liberalization of agricultural markets that are owned by ADMARC. The recent Malawi Poverty Reduction Strategy Paper (PRSP) explicitly mentions the privatization of ADMARC as one of the measures necessary to foster private sector led growth and reduce the parastatal’s fiscal burden. One sub-set of ADMARC’s privatization strategy comprises the intended withdrawal of ADMARC from around 200 unprofitable markets out of a total of around 350 rural markets throughout the country2. At the practical level, successful implementation of the envisaged reform program depends importantly on sound (and preferably prior) knowledge of the effects of the market liberalization on the poor and vulnerable. In the context of ADMARC, it is frequently put forth that, since private markets in Malawi remain underdeveloped, ADMARC services are important in providing agricultural inputs and output markets to the poor, and that dismantling ADMARC is likely to have a significant negative impact on the welfare of poor households, especially those living in remoter areas that lack access to alternative marketing channels. For example, in the wake of the recent food crisis in Malawi, distribution of imported maize through ADMARC’s extensive market infrastructure network has gained importance: about 73,960 metric tons were distributed between August 2001-May 2002 (Mvula, Chirwa, and Kadzandira 2002). It is contended that liquidation or privatization of ADMARC depots will significantly affect the welfare of households making use ADMARC services, and that this effect has to be clearly factored into the policy debate on the future of ADMARC.

The impact of ADMARC services on the welfare of the poor is essentially an empirical issue, and the general lack of empirical evidence on this topic has led to a stalemate in policy action. This paper attempts to fill this gap using data collected in the 1997 Integrated Households Survey (IHS) and the various Complementary Panel Survey (CPS) rounds completed thereafter. The plan of the report is as follows. Section A2 provides a brief description of data sources and descriptive analysis of data is provided in Section A3. An econometric model for estimating the impact ADMARC is presented in Section A4, and results of the estimated models are discussed in Section 5. Final conclusions are presented in Section A6. 2 Detailed profile of ADMARC and description of envisaged reforms are contained in Mvula, Chirwa, and Kadzandira (2002) and Nthara (2002).

28

A2. Data sources

In 1997, the National Statistical Office (NSO) carried out the 1997-98 Malawi Integrated Household Survey (IHS), a comprehensive socio-economic survey of the living standards of households in all districts of Malawi, under the auspices of the Poverty Monitoring System. Later, the Complementary Panel Survey (CPS) was implemented by the Center for Social Research whereby data was repeatedly collected from a sub-set of the households that participated in the 1997-98 IHS. Four rounds of the CPS were conducted between January 2000 and September 2002. The fourth round of the CPS (CPS4) was implemented primarily to collect information that would allow household welfare comparisons to be made between the 1997-98 IHS and the 2002 CPS4. Consequently, the CPS4 questionnaire was designed to include questions on all of the components of the welfare indicator used in the IHS poverty analysis. Assessment of ADMARC’s impact on household welfare is based on 291 households which are found both in the 6,586 household data set of the IHS (households judged to have good consumption data) and in round 4 of the CPS. A detailed description of data is provided in Sharma, Tsoka, Payongayong and Benson (2002). A3. Household use of ADMARC services: descriptive evidence from CPS4

CPS4 collected information on a number of household transactions with ADMARC. Specifically, information was collected on: • Travel time to the nearest ADMARC depot,

• Role of ADMARC in supplying fertilizers and improved maize seed, and

• Role of ADMARC in providing staple goods for consumption

In the description of these variables below, data is used only for the 291 households for which comparable consumption data is available in the IHS.

Travel time to ADMARC. More than two-thirds (67.7 percent) of the CPS4 households indicated that the nearest ADMARC depot was reachable in one hour or less time. About 23 percent indicated travel time to be 1-2 hours, and less than 10 percent indicated that the nearest ADMARC depot was more than two hours away. However, because the mode of transportation was not standardized (i.e., households computed travel time based on varying modes of transportation), this information is not that revealing. For this reason, the regression analyses below uses data on actual distance to ADMARC collected in the first round of the CPS.

ADMARC’s role in fertilizer and seed supply. All households that reported fertilizer

purchases were asked the source of their purchase. Out of the 131 households that reported purchasing fertilizer, only 12 (9 percent) reported purchasing it from ADMARC. The most important source was the Starter Pack program (28 percent), closely followed by NASFAM or other cooperatives (26 percent) and traders (21 percent).

As a source of seed, ADMARC was equally unimportant. Only 14 out of the 171 households

(8 percent) reported purchasing seeds from ADMARC. The rest acquired seed from traders (22 percent), the Starter Pack program (16 percent), and other unspecified sources (20 percent).

ADMARC role in securing consumption of staple food. In the survey, all households that

had run out of their own stocks of their main staple (mostly maize) at the time of the interview were asked the current source of their staple, if any. At the time of CPS4, 39 percent of 278 responding households reported having no stock of the staple in households. Eighty-eight percent of such households (i.e., those that were without stocks) were still eating the staple. However, only nine percent of such households acquired their supply of the staple from ADMARC. Most - about 70 percent - purchased their supply from either traders or other individuals.

29

Caveat and conclusions from descriptive analysis. Because of the overall low reported use of ADMARC in making food, fertilizer, and seed purchases, use rates further disaggregated by expenditure groups are not discussed here, as cell sizes become exceedingly small and meaningful inferences cannot be drawn. More importantly, the very low usage of ADMARC services implies a proportionately small aggregate impact on household welfare. On the face value of reported information, therefore, there is little basis for committing public resources to ADMARC.

An important caveat exists, however. The reported low use of ADMARC as a source of

agricultural inputs and staple food in the CPS4 questionnaire is not consistent with the high ADMARC usage levels reported in the “Access to Facilities” section of the same questionnaire. In this section, all households were asked whether any of their members use ADMARC facilities. 235 out of 291 households – a very high 81 percent – reported making use of ADMARC facilities. Can the two sets of responses be reconciled? The answer is likely to be in the affirmative. While the fertilizer, seed, and staple food-related questions have a specific and narrow time reference for recall (e.g. “last crop season” or at “time of interview”), the facility-use question is open-ended and does not specify a recall period. The high rates of usage reported could very likely be due to the broader recall period used by the respondents when answering this question. If this is indeed so, average use rates over time of ADMARC services is much higher than that reported for the narrow recall period associated with CPS4. It also implies that the impact of ADMARC on household welfare may be proportionately larger that which the data presented above imply. For these reasons, an assessment of the impact of ADMARC in a broader setting and making use of multivariate analysis is done and reported in the next two sections. A4. Specification of an econometric model

The underlying idea behind the multivariate analysis is to explain household welfare in 2002, measured by per capita total household expenditure from the CPS4, by access to ADMARC services, controlling for other major determinants of household welfare. More specifically, it is posited that per capita total household expenditure of household i in year t, yi

t , is a function of the initial welfare position of the household in base year 0, and the growth rate of total per capita expenditure, r, between the two time periods as in (1) below

)1(0trii

ti

eAyy = where A is a constant term3.

It is reasonable to expect that the growth rate, r, is determined by both external conditions as well as by household characteristics, which includes its access to ADMARC. For estimation purposes, the endogeneity of r with respect to ADMARC and household characteristics is expressed linearly as r = α ADMARCi + Xiββββ where Xi is the vector of household characteristics, α describes the impact of ADMARC and ββββ is a vector describing the impact of other household characteristics. Substituting this into (1) leads to

)2()(0

tADMARCiit eAyy βX i+= α

In this analysis, the base and current year are 1997 and 2002 respectively. Hence the estimating equation is derived by taking logs on both sides of equation (2), renaming the logs of yt and y0 as LNEXP02 and LNEXP97 respectively, eliminating t (same for all households), and adding a stochastic error term, ei , thus yielding equation (3)4 below

)3()(LNEXP97)(log(LNEXP02) ii ii eADMARCA ++++= βX iαδ .

3 As mentioned in Sharma, Tsoka, Payongayong and Benson (2002), it is very likely that reported 2002 expenditure levels are higher simply because of better enumeration and across the board improvements in survey administration. The systematic part of this conflation factor is also embedded in A. 4 Note that this specification strictly applied implies the coefficient of LNEXP97 to be unity. However, this condition is not enforced here.

30

It is noted that a positive impact of ADMARC on expenditure levels or on expenditure growth implies that α<1 since ADMARC is measured as distance to the nearest depot. However, estimating (3) using ordinary least square technique cannot guarantee an unbiased estimate of α as placement of ADMARC across different locations cannot be assumed to be random. If factors affecting placement decisions are correlated with included regressors, then the estimate of α will be biased (Binswanger, Khandker, and Rosenzweig 1993). For example, if ADMARC depots are more likely to be placed in areas with less than average income levels, not accounting for this information will result in downwardly biased estimates of the effects of ADMARC. Of course, factors driving placement decisions are not known. However, it is reasonable to expect that these factors were related to some unobservable set of local socio-economic-political conditions. For this reason, TA level fixed effects are specified to sweep out unobserved factors that are likely to be correlated with factors influencing placement.

Another version of the equation (3) is also estimated in order to take into account the potential bias caused by underlying but unobserved factors that affect both LNEXP02 and LNEXP97. In this case LNEXP97 would be correlated with the error term and result in biased estimates of the regression parameters (Deaton1998). In order to address this problem, the term LNEXP97 is moved to the right hand side so that the estimated equation becomes:

)4()(logLNEXP97)((LNEXP02) ii iiADMARCA υα +++=− βX i Equation (4) is thus the preferred model. For exposition purposes, both (3) and (4) are estimated and discussed, however. Variables used in regression analyses are described below.

The dependent variable used in equation (3) is the log of per capita total household expenditure in 2002 as reported in CPS4, LNEXP02. In this regression, the log of per capita total household expenditure from the IHS, LNEXP97, is used as an independent variable that describes the initial position of the household. In Equation (4), LOGDIFF= LNEXP02-LNEXP97 is used as the dependent variable.

The variable used to describe access to ADMARC services is the reported distance to the

nearest ADMARC depot. Since distance to ADMARC depot was collected only in the first round of the complementary panel survey (CPS1), this was used. Information on the actual use of ADMARC was not considered since ADMARC use is likely to be endogenous in the equations specified. Use of the distance variable avoids this problem.

Three sets of control variables (elements of Xββββ in equation (3)) are also used in the regression:

1) Demographic characteristics that have an important bearing on households’ ability to manage economic change, including sector and type of employment5,

2) Variables that control for access to resources and institutions that assist households to uphold or smooth consumption in face of changing incomes (these variables are important since the welfare measure used is consumption based), and

3) Changes in key asset variables that are likely to affect income earnings in the household.

These variables are described in Table A1 grouped into these three sets.

Finally, it is reasonable to expect the impact of ADMARC on household welfare to be conditional on factors such as access to other types of market channels, reliability of the household’s own food supply, and access to various types of safety net services or social assistance policies. For this reason, the following interaction variables were used:

• ADMARC*DISTURB – Assuming other private sector market-related institutions are more concentrated around urban centers, then the impact of ADMARC is likely to increase with

5 Other basic human and capital assets are already embedded in the initial 1997income variable, LNEXP97.

31

distance to urban center. The inclusion of the interacted term ADMARC*DISTURB makes it possible to test this hypothesis.

• ADMARC*STOCK – All else being the same, households that consistently carry stocks of staple are less likely to go to ADMARC to procure it.

• ADMARC*SNPARTIC. Participation in safety net or other social assistance programs provides an important control for availability of such programs in the area as well as of the household’s capacity to access and use it. ADMARC services are likely to be less useful for such households and its impact, therefore, correspondingly lower.

When equation (3) is expanded to include the interaction terms, the impact of ADMARC on per capita expenditure in 2002 (EXP02) can then be given by

)5()(02/02 SNPARTICSTOCKDISTURBEXPADMARCEXP φγβα +++=∂∂

A5. Regression results

The results of the two regressions are presented in the rightmost columns of Table A1. The level specification regression (equation (3)) results are presented in the columns headed ‘Regression on lnexp02’. Since the estimated coefficients of this equation are most likely biased due to measurement errors, they may be termed “naïve” estimates. The ratio specification regression (equation (4)) results are presented in the columns headed ‘Regression on logdiff. As noted before, equation (4) is the preferred model. In interpreting the signs of the coefficient on the ADMARC and the interacted variables, note that the ADMARC variable is constructed as distance to the nearest ADMARC depot. Consequently, the larger the value, the less access the household has to ADMARC. Hence a negative sign for α indicates a positive effect on household welfare from access to ADMARC.

Results from the naïve model. When the level specification is used, the coefficient on ADMARC is negative implying that the impact of ADMARC decreases as distance to ADMARC increases, that is, as access to ADMARC decreases. However the coefficient is not statistically significant (p=0.67). Among the interaction variables, ADMARC*DISTURB is significant at the 5% level, but carries a positive sign – implying that the impact of ADMARC on household welfare is lower for households who live further away from urban centers. The coefficient of ADMARC*STOCK is positive implying that the impact of ADMARC is dampened for households that are able to retain sufficient stocks of a staple from their own sources. However, the coefficient is not statistically significant (p=0.38). The coefficient of ADMARC*SNPARTIC is negative, implying that, for any given level of access to ADMARC, its impact is higher for households that are in a position to participate in safety net or social assistance programs. However, again, the coefficient is not statistically significant (p=0.24).

A more meaningful statistical test, given the specification of the econometric model, are F-tests of the joint significance of the terms specified in equation (5), i.e. testing the joint significance of α, β, γ, and φ in various combinations and assessing whether the signs of these coefficients taken together make sense. Two sets of F-tests are carried out on the coefficients of the ADMARC variable with the interacted variables from the regression on LNEXP02. First, three F-tests of the joint significance of each of the interacting terms and the ADMARC term are carried out6. The result is that only the ADMARC*DISTURB and ADMARC coefficients are jointly significant (i.e. statistically different from zero) at the 10 percent level (F2, 241 = 2.49; p = 0.08), implying that ADMARC effects are not conditional upon either food stock at home or household involvement in social programs.

However, the second test, an F-test of the joint significance of all the coefficients taken together (i.e., testing the null hypothesis that α = β = γ = φ = 0) is rejected at the 10 percent significance level (F4, 241 = 1.51; p = 0.20). This implies the “larger” result that, while access to 6 That is, the following null hypotheses are tested: (1) α = β = 0, (2) α = γ = 0, and (3) α = φ = 0.

32

ADMARC has a positive impact on expenditure levels, the impact is higher in locations closer to the urban centers and for those households that are in the position to participate in social welfare programs, while it is lower for those households that are able to carry stocks of the basic staple. The relationship between ADMARC and proximity to urban centers appears to be especially strong and the estimated coefficients suggest that impact of ADMARC reduces to zero as distance to urban center approaches about 25.6 kilometers.

Results from the preferred model. ADMARC impacts are non-existent when the preferred ratio specification is used (columns headed ‘Regression on logdiff’ in Table A1). To begin with, the ADMARC coefficient carries an unexpected positive sign, even though this not significant (p=0.45). The set of F-tests of the joint significance of the ADMARC and each of the interacted terms indicates, as before, that only the ADMARC*DISTURB and ADMARC are jointly significant at the 10 percent level (F2, 242 = 2.52; p = 0.08), but because α>0, interpretation is not meaningful. For the same reason, even though the F-test rejects (at 10 percent significance level) the null hypothesis that α = β = γ = φ = 0 ((F4, 242 =2.01; p = 0.09), interpretation remains problematic.

The preferred model therefore does not suggest any evidence of ADMARC’s impact on the household welfare. However, the effects that do remain robust to changes in specification are those related to HHCHANGE (negative) and CHANGE_CULT (positive), both of which are significant at the 1% significance level in both equations (3) and (4). In a pre-dominantly low-technology agricultural economy like Malawi, increase in the access to cultivatable land is the principal determinant of increases in household welfare, while increases in household size chips away at per capita availability since land, not labor, is the binding constraint on increased production. A6. Conclusion

The IHS and CPS data were used to measure the effect of differential access to ADMARC depots on household welfare of sample households. The naïve model showed that access to ADMARC has a positive impact on expenditure levels and that it is higher in locations closer to the urban centers and for those households that participate in social welfare programs, and lower for those households that are able to carry stocks of the basic staple. However, the preferred model, which obtained a cleaner estimate of the impact of ADMARC by removing biases caused by common measurement errors affecting the computation of welfare indicators, did not suggest a statistically significant impact of ADMARC services on household welfare. In this context, it must be noted here that CPS4 coincided with an emerging income and food crisis in Malawi. It could well be that the impact of ADMARC was simply drowned out by the far more overwhelming effects of the deep decline in agricultural production that Malawi has witnessed recently. Put against this perspective, strong evidence for or against the continued operation of ADMARC cannot be found in the analyses of the CPS presented here. A7. References

Binswanger, H., S. R. Khandakar, and M. Rosenzweig, "How infrastructure and financial institutions affect agricultural output and investment in India," Journal of Development Economics, Vol. 41 (1993), pp. 337-366.

Deaton, Angus. 1998. The Analysis of Household Surveys. Published for the World Bank: The Johns Hopkins Press. Baltimore.

Mvula, P., E. Chirwa, and J. Kadzandira. 2002. Poverty and social impact assessment in Malawi: closure of ADMARC markets. Manuscript. Wadonda Consult, Malawi.

Nthara, K. 2002. What needs to be done to improve the impact of ADMARC on the poor. Phase 2 Report (final) prepared for OXFAM. Department of Economics, University of Malawi.

Sharma, M., M. Tsoka, E. Payongayong, and T. Benson. 2002. Analysis of poverty dynamics in Malawi with an assessment of the impact of ADMARC on household welfare. Draft report submitted to the World Bank.

33

Table A1: Variable Description and results of regressions to assess impact of ADMARC on household welfare.

Descriptive statistics

Regression on lnexp02

Regression on logdiff

Mean Stnd. Dev. Coef.

t-statistic Coef.

t-statistic

Dependent variables lnexp02 natural log of CPS4 welfare indicator (ln MK) 2.67 0.74 - - - -

logdiff difference between logged CPS4 and logged IHS welfare indicators (ln MK) 0.32 0.83 - - - -

Independent variables lnexp97 natural log of IHS welfare indicator (ln MK) 2.35 0.72 0.349 (5.06)** - - admarc distance to nearest ADMARC (km) 6.64 6.20 -0.010 -0.43 0.019 0.75

Household demographic variables from IHS agehhh age of head of household 43.73 15.67 -0.001 -0.46 -0.003 -0.92

dpndrat dependency ratio: dependents/hh 0.47 0.24 -0.035 -0.20 0.422 (2.17)*

hhhlit literate head of household (0/1) 0.73 0.44 -0.027 -0.24 -0.147 -1.13

farmer hh member with agricultural occupation 0.38 0.49 0.078 0.62 0.215 1.48

manufact hh engaged in manufacturing 0.07 0.25 -0.260 (1.73) + -0.208 -1.18

prof hh member with professional, ad 0.16 0.37 -0.154 -1.06 -0.356 (2.12)*

maxyrsed maximum years of education in hh 4.64 4.50 0.032 (2.27)* 0.020 1.24

Access to key consumption smoothing assets or institutions

stock variable that equals one if hh owned stocks of staples at time of interview in all of the four CPS survey rounds (0/1)

0.19 0.39 0.304 (1.94) + 0.131 0.72

pclvsval per capita livestock value (MK) 276 714 -0.00004 -0.65 -0.00006 -0.93

soccap hh reported in IHS giving or receiving income transfer (0/1) 0.33 0.47 0.043 0.47 -0.202 (1.95) +

snpartic hh participated in at least one safety net or social assistance program as reported in CPS1 (0/1) 0.86 0.35 0.106 0.55 0.297 1.34

Change in access to key human and physical assets hhchange change in hh size betweeen IHS and CPS4 -0.11 1.96 -0.095 (4.60)** -0.145 (6.21)**

change_cult change in cultivated land area of hh betweeen IHS and CPS4 (acres) 0.50 2.13 0.050 (2.65)** 0.050 (2.30)*

Interacted variables

admarc_disturb interaction of distance to ADMARC with distance to nearest urban center 498 620 0.0004 (1.93) + 0.0002 1.11

admarc_stock interaction of distance to ADMARC with whether hh usually possesses staple food stocks 1.31 3.86 0.014 0.88 0.025 1.41

admarc_snpartic interaction of distance to ADMARC with whether hh participated in safety net programs 5.65 6.19 -0.021 -1.17 -0.037 (1.75)+

disturb distance to nearest urban centre (km) 67.75 42.87 -0.0003 -0.04 -0.005 -0.69

constant 1.621 (3.29)** 0.257 0.47

R2 0.2444 0.2103


Number of TA fixed effect variables 19 19 + significant at 10%; * significant at 5%; ** significant at 1%

34

35

ANNEX B - QUANTITATIVE STUDY PREPARED BY J. OWENS (WB CONSULTANT)

36

How important is access to ADMARC markets in Malawi?7 B1. Introduction

Background. Following decades of state intervention in food marketing activities, Malawi, like other countries in Sub-Saharan Africa has, undertaken, since the early 1980s, reforms to liberalize agricultural markets (Jayne and Jones 1997). These reforms have come in the aftermath of the failure of many state-led marketing initiatives to either stabilize prices, guarantee food security for consumers, or income security for producers. More frequently than not, the reason behind such failures was unsustainably high fiscal cost of these programs (Chaudhury and Banerji 2001), made worse by widespread inefficiency within state-run institutions.

Even further liberalization of the agriculture sector is envisioned in Malawi at the current time, and restructuring of ADMARC remains high on the reform agenda. The rationale underlying this reform agenda is essentially three-fold (i) ADMARC bail-outs have burdened the central government budget in recent years; (ii) the presence of ADMARC markets and warehouses discourages private sector participation in input supply, marketing, and storage and (iii) recent reports of grain reserve mismanagement as documented by the 2002 Anti-Corruption Bureau investigation question whether ADMARC can perform an effective food security role. The ADMARC restructuring strategy involves the disposal of loss-making commercial subsidiaries as well as the auctioning of the loss-making markets to the private sector.

However, the enthusiasm for reforms is not universally shared. Those who do not counter that since private markets in Malawi are underdeveloped, ADMARC operations in providing agricultural inputs and output markets have a positive and significant effect on household income and welfare – especially of households living in remoter areas. They contend that liquidation or privatization of ADMARC depots will not necessarily trigger equivalent private sector responses and will therefore significantly hurt these households.

As a way of contributing to this debate, this paper assesses the potential impact of market restructuring on rural livelihoods by using household data from the nationally representative Integrated Household Survey (IHS) to examine the relationship between access to ADMARC services and household welfare.

Outline of the paper. This paper is structured as follows. Section B2 provides a brief summary of ADMARC involvement in Malawi’s agricultural sector. Section B3 describes the two data sources used in this study, namely the recent administrative data on market placement, and data from IHS, the last nationally representative household survey. Section B4 lays out the econometric approach used and estimating the impact of ADMARC on maize yields, crop profits, fertilizer demand, and household consumption levels. Finally, Section B5 presents the findings, and conclusions are summarized in Section B6. B2. ADMARC in the context of Malawi’s agricultural sector

The overwhelming importance of smallholder agriculture for growth and employment in

Malawi has prompted a long history of government involvement in the marketing of agricultural produce. In 1971, the Agricultural Development and Marketing Corporation (ADMARC) was established as a statutory corporation to supply agricultural inputs to farmers of customary land, market agricultural produce and expand agricultural exports. Until the mid-Eighties, ADMARC maintained a monopoly in the marketing of agricultural inputs, as well as food and cash crops. A gradual move was made thereafter in liberalizing input and output markets in the agricultural sector. 7 This draft report has been prepared by Janet Owens, with funding from the WB Norwegian trust fund. The findings, interpretation and conclusions reported in the report are preliminary and may be revised at a later stage. They represent solely the author’s view and should not be attributed to the World Bank.

37

This resulted in greater private sector participation and the removal of price controls. Since 1995, input and output prices have been further liberalized with private traders buying and selling maize within a price band established by ADMARC. More recently, the price band has also been removed. Despite this, the large network of ADMARC markets in rural Malawi has continued to play a role in ensuring food security in the countryside8. ADMARC currently operates 343 permanent markets (257 ‘unit’ markets, 73 parent markets and 14 depot markets)9. According to a recent assessment, the impact of agricultural reforms on small holders has been mixed in Malawi (Kherallah, et al., 2001). Farmers perceive that the marketing environment has improved significantly due to a larger presence of private traders but the input supply sources appear to have changed only marginally. A survey undertaken in 1997 to assess private trading activity and the marketing environment in light of recent agricultural reforms found that 95% of all markets experienced an increase in the number of agricultural traders.10 During this period, the government constructed new roads in about 70% of markets and reconditioned existing ones in almost all markets resulting in increased transport activity and an increase in the number of vehicles providing transport services.

Findings from a detailed study that tracked livelihood strategies in one region of Malawi concluded that the impact of market liberalization has been largely positive (Orr and Mwale, 2001). Liberalization encouraged diversified cropping patterns among small holders and many embraced inter-cropping strategies that led to increased maize productivity. Burley tobacco cultivation by some smallholders afforded them opportunities to obtain fertilizer which was also used on maize cultivation. Liberalization appeared to offer increased opportunities for the marketing an array of crops, and many of the households in the study concluded that liberalization had generated improvements in their economic status.

However, the positive perceptions associated with liberalization haven’t been ubiquitous. Some households experienced a decline in their economic mobility and attributed their predicament to liberalization. These households were less likely to have undertaken diversified cropping patterns, or improved maize productivity.

The question of whether and how ADMARC services impact household livelihoods has therefore become increasing relevant in evaluating the reform agenda in Malawi. B3. The data The primary data source for this analysis is the rural component of the nationally representative 1997-1998 Malawi Integrated Household Survey (IHS). The Malawi National Statistical Office administered the IHS questionnaire to 12,960 households over a 12-month period, November 1997 to October 1998. The then 25 administrative districts of Malawi, plus the four major urban centers of the country (Lilongwe, Blantyre, Zomba and Mzuzu), constituted the 29 primary sampling strata. The survey design was based on a national sample of 12,960 households to allow estimates on population characteristics to be generated at the district level.

8 ADMARC has also acquired a number of subsidiaries as commercial investments. These subsidiaries have imposed a major burden on ADMARC’s finances over the years and four of the largest loss-making ones are currently in the process of being divested. 9 Unit markets are permanent structures with a small holding capacity for storage. Parent markets combine permanent storage facilities with an administrative office that oversees unit and seasonal markets. Seasonal markets are mobile selling points that can be opened on demand with a temporary sales force. 10 The “Impact of Agricultural Market Reforms on Smallholder Farmers in Benin and Malawi” is a collaborative research project implemented by the International Food Policy Research Institute (IFPRI) to study the impact of agricultural market liberalization on input and output markets, and on agricultural households. Project activities were undertaken from June 1998 until April 2001 and included surveys of smallholder farmers, input and output traders, markets and community infrastructure.

38

The questionnaire was administered in two parts. The first was a large questionnaire that was administered to the respondent household in a single visit. This consisted of approximately a dozen modules on household composition, educational attainment, health and nutritional status, agriculture, home-produced and purchased consumption items, assets, and access to infrastructure, including ADMARC. The second part was a diary of expenditure. The diary was maintained over a minimum of 14 days through twice-weekly visits by the enumerator to the survey household to record any expenditure made since the previous visit.

Although administered to 12,960 households, the data set consisted of 10,698 households when the cleaned IHS data was released in early May 2000. However, as the diary of expenditure was not always consistently maintained by enumerators, upon additional assessment of the data, only 6,586 households were judged to have reliable expenditure and consumption information for use in the derivation of a consumption and expenditure based indicator of household welfare. As the household welfare indicator is used in this analysis, here the more restricted 6,586 household IHS data set is used.

Several months after the completion of the fieldwork of the IHS, a community level survey was carried out in all of the traditional authorities in which IHS sample households had been selected. Information on a range of community level variables and conditions was collected through interviewing key informants. Some of this data, particularly those on changes in local marketing infrastructure and opportunities has been used here. B4. Impact of ADMARC on household production and income: econometric specification

We assess the impact of ADMARC access on the following: a. Productivity in maize cultivation b. Total farm profits c. Demand for fertilizer d. Household per capita expenditure (proxy for overall household welfare)

Considerations in specifying regression equations that measure these impacts of ADMARC are discussed below.

Agricultural productivity and demand for fertilizer. Agricultural productivity is

determined by technology choice and the inherent characteristics of the inputs and fixed factors used to grow crops. The demand for factors of production (eg. fertilizer)is based principally on the profitability of its use in the production of the marketed commodity. Use of any input, therefore, depends both on the price of the final product being produced and its own price. Access to ADMARC services in marketing inputs and outputs is therefore likely to influence input intensity and, therefore productivity, on both counts. When reliable product markets are ensured, profitability is high and use of yield-enhancing inputs such as fertilizer is correspondingly high. Likewise when inputs are marketed at reasonable prices, profitability increases demand for the input.

Total farm profit. Total farm profit is a function of the technology adopted (which determines the yield rate), the price of the final produce, as well as the level and prices of various factors of production used in production. To the extent that fertilizer use and commodity prices are affected by ADMARC operations, access to ADMARC is also likely to have an impact on the level of farm profits.

Household expenditure. Household expenditure is a function of the level of household’s wealth, price levels, and household demographics. The effect of ADMARC on household food expenditure comes via the wealth effect (by affecting farm profits) as well as through price effects (by affecting the retail price of maize).

With these considerations in mind, the generic equation used to estimate impact is specified as

39

)1(33221122110 ilADMARCADMARCADMARCiii eDDDXXyLog +++++++= µδδδβββ

where iy is the outcome of interest, iX 1 is a vector (in natural logs) of non-labor inputs including land, fertilizer and seed, and iX 2 is a vector describing characteristics of household labor, including the level of education of its members. There are three ADMARC-related variables: 1ADMARCD is a dummy variable that takes a value of one when time to nearest ADMARC is less that 30 min;

2ADMARCD is dummy variable that takes a value of one when time to nearest ADMARC is 30-60 minutes; and 3ADMARCD is a dummy variable that takes a value of one when time to nearest ADMARC is 60-120 minutes. The base scenario when distance to ADMARC is more that two hours is embedded in 1β . The sδ , thus provide an estimate of the impact of ADMARC. Given the way the dummy variables are defined, consistency in impact would require 321 δδδ ≥≥ 11. Finally, iµ is a fixed-effect term that controls for unobserved location characteristics and ei is an stochastic error term. In the first set of estimates, iµ is specified at the EA level. In the second set, district level fixed effects are used; when this is done the estimated equation is further augmented by variables describing EA-level characteristics such as distance to market and population size. Both sets of estimates are presented and discussed.

The estimation procedure also makes the following additional considerations. First, we estimate the productivity-, fertilizer-, and farm profits-equations separately for local and hybrid maize varieties. This is because of our expectation that production parameters will significantly differ across the two varieties. Second, we take note that impact of ADMARC services is likely to be strongly proportional to the degree of remoteness. This is because remoter areas have less developed private markets, and alternatives to ADMARC services are less likely to exist there. We test this hypothesis by splitting the sample of households on the basis of distance to the nearest primary road measured at the community level12, and estimating regression equations separately for the two sub-samples. Of course, it may well be that road-building decisions are based on existence of other complementary infrastructure or on a region’s agro-climatic potential (Binswanger, et al. 1993). Thus, we use fixed effects, either at the Enumeration Area (EA) level or the district level to control for these as well as other unobservable factors such as area-specific price levels (for inputs as well as outputs), and components of area-specific transactions costs.

Finally, when estimating the determinants of household per capita expenditures, the X vectors in equation (1) are made to contain variables that measure the wealth of the household, namely land and labor assets. The equation is also augmented by one more right hand side variable: the imputed maize producer price. This is done to control for ADMARC’s transmission of impact through the production side. B5. Empirical Findings

Descriptives. In Table A4, we examine differences in sample means for key variables across two sub-samples differentiated by distance to the nearest main road. Hybrid maize productivity is slightly higher for remote households compared with households living in proximity to a major road, although fertilizer and seed inputs are applied more intensively by the latter. Remote households produced about 387 kilos of hybrid maize per acre compared with 365 kilos of hybrid maize per acre generated by households living in proximity to a major road. Local maize production across the two groups reveal even larger differences. Remote households generated 214 kilos of local maize per acre

11 This would ensuring that impact for households nearer to ADMARC is greater than impact for households further away, controlling for all other factors. 12 In this paper, road access is measured as the straight-line distance in kilometers of the household’s enumeration area to the nearest primary or secondary road.

40

compared with 194 kilograms of local maize per acre generated by local households having proximity to a major road. Remote households applied significantly higher quantities of local maize seed per acre (1.6 kilos) compared to households of proximity (1.3 kilos).

Remote households also generated higher profits per acre (2.6 kwachas) compared with households associated with greater road access (2.0 kwachas). However, remote households had significantly low per capita expenditures (9.7 kwachas) compared to households of enhanced road access (27.1 kwachas). On average, household size and the number of adult males per household are larger among households located within proximity to a major road. These differences suggest that households in this group have more labor assets that can be utilized for both on-farm and off-farm activities. Differences in labor endowments across these two groups may explain some of the variation in per capita expenditures.

Local maize regressions. In Table A5, we present regression equations that estimate the impact of ADMARC access on local maize yields according to proximity to the nearest primary road. In all specifications we include values of physical inputs, the gender and educational attainment of the household head to control for human capital and managerial efficiency.

Columns (1) and (2) provide regression estimates for households located in close proximity to a primary road. On average, households living in these communities are located 1.5 kilometers from the nearest road. We note that unobserved differences in soil fertility will affect the technical relationship between input usage and yield. Hence, in the equation reported in column (1), we included enumeration-area level dummies in order to control as much as possible variations in soil fertility across communities.. Of course, to the extent variations in the micro-nutrients in the soil varies within the enumeration area, these differences are not captured.

In column (2), we include district-level dummies to control for unobserved district level effects. In this case, we are also able to use additional community-level variables such as the population of the Traditional Authority and its distance to the nearest produce market13 as additional regressors. This specification therefore allows us to better control for the impact of commercialization. Results of the F-test on the joint significance of variable are reported at the bottom of the table.

As we might expect, the fertilizer coefficient and its squared term are jointly and highly significant. An additional 1% in the value of fertilizer applied to local maize leads to an increase in maize output of between .1 and .3 percent. The education variables are also jointly significant which suggests managerial efficiency has a significant and positive effect on productivity. Compared to household heads that considered themselves illiterate, those who received some secondary-level education produced more maize per area of land farmed. Across these two columns, we find that ADMARC access has no impact on local maize productivity, and that consistency rule fails to be preserved in case of (2). The F-tests reported for the set of ADMARC dummies confirms that distance to ADMARC is insignificant.

Columns (3) and (4) present local maize productivity equations for households located in remote communities, with EA level dummies and with district level dummies respectively. On average, these communities are located about 7 kilometers from the nearest primary road. In addition to the differences in coefficients observed for the sample of households in remote areas, it is also likely that soil fertility may vary by road proximity. It is plausible that households in remoter areas farm land that is substantially less fertile compared with those areas in close proximity to road infrastructure. Population density—and the corresponding network of nearby infrastructure—is likely to vary positively with soil fertility.

13 Traditional Area is a lower order geographic area compared to a district.

41

With the exception of seed usage in the first of these two specifications, land, fertilizer, and seed input usage are jointly significant. The combined marginal effect of these coefficients with their squared terms indicates that there is a negative impact on productivity for each of these inputs. Unobserved fertility may be negatively correlated with input usage which may drive down coefficients towards zero. Distance to ADMARC has a positive and significant effect on maize production in remote areas. Households located less than 30 minutes away from the nearest ADMARC facility produced between 17% and 28% more maize per acre than households located more than 2 hours away from the nearest facility. Additionally, the distance to ADMARC variable may also be picking up the impact of commercialization on household productivity. Households that are located in remote areas may have limited exposure to other types of marketing institutions. Although we control for variation within the enumeration area and the district in these specifications, it is plausible that distance to ADMARC is inversely correlated with the presence of other market institutions. It is noted, however, that the relative magnitudes of the three ADMARC dummy variables are not logically consistent.

Hybrid Maize regression. Hybrid maize equations are presented Table A6, with columns referring to the same cases as above. Households included in columns (1) and (2) were, on average, located about 1.4 kilometers from the nearest main road. Compared with local maize production, the usage of purchased seed and fertilizer is expected to generate a higher return to hybrid maize production. Therefore, marketing channels that facilitate access to inputs is likely to be more critical in hybrid maize production. A one percent increase in fertilizer usage leads to an increase of hybrid maize of about .13 percent per acre. The educational attainment variables are jointly significant in both columns. In particular, household heads who received some higher level of education produced about 27% more hybrid maize per acre than household heads who reported being illiterate. The educational attainment coefficients suggest that there is greater return to human capital and managerial capacity in the hybrid maize production technology. Proximity to ADMARC facilities has no impact on hybrid maize production for these households. Further the consistency rule is preserved only in the specification in column (1).

Columns (3) and (4) report hybrid maize productivity estimates for households located in remote areas. On average, these households are located about 7 kilometers from the nearest primary road. Fertilizer and seed usage coefficients are positive and significant across the two alternative specifications. A one percent increase in fertilizer usage leads to about 0.8 percent increase in hybrid maize productivity, while one percent increase in the amount of seed application leads to a 0.7 percent increase in productivity. The set of ADMARC coefficients are jointly significant in these two specifications. Households experienced a positive impact from being close to ADMARC facilities relative to households who reported living 2 hours away from the nearest facility.

Farm Profit regressions. In Table A7, we provide estimates of the determinants of total farm profits. Again, columns (1) and (2) report estimates of farm profits equation for households located within close proximity to a primary road. Female-headed households generated between 32% and 34% lower profits than male headed households while land holdings generated positive returns for total farm profits. An additional one percent of land increased household profits by about 0.70 percent. Access to ADMARC facilities had no impact on farm-level profits for households located within proximity to a primary road.

Columns (3) and (4) provide estimates on farm profits for households located in remote areas. Similar to above, female-headed households generated lower household profits than their male counterparts: on average, between 32%-44% lower. One possible explanation behind the gender-based disparity in profits is that female heads plant lower valued crops than other households, and they therefore have lower profit margins. Educational attainment has a positive impact on household profitability. Household heads that reported either secondary or higher education earned substantially more income than those households that reported being illiterate. When estimating the impact on profits with district-level dummies, we find that household profits are sensitive to ADMARC access. Households located within 30 minutes to the nearest facility generated about 37% higher profits

42

compared with households who reported living more than 2 hours away from the nearest ADMARC facility. The relative magnitudes of the three ADMARC variables are also logically consistent .

Fertilizer demand regressions. To further explore the relationship between input usage and ADMARC access, we provide estimates of fertilizer demand for local and hybrid maize in Table A8. We include land and household composition variables to control for the impact of household-level assets and household demographics on input usage. Under the assumption that rural households operate in an economy of functioning factor markets, the household decentralizes its production and consumption decisions (Singh, Squire and Strauss 1986). If these conditions hold, production decisions, such as input usage, are not sensitive to household parameters. However, households in Malawi are confronted with numerous incomplete markets (e.g., credit rationing and incomplete contingency markets) and therefore we assume that household characteristics may explain some of the variation in input usage.

Following the strategy above, we estimate the demand for fertilizer usage on local maize in columns (1) and (2). We incorporate only enumeration area dummies in this specification to control as much as possible for unobserved soil fertility and split the sample on the basis of distance to the nearest primary road. As we might expect, land holding is a significant determinant of fertilizer usage. An additional one percent in land holdings increases fertilizer expenditures on local maize by about 0.85 percent for households located within proximity to a primary road and about 1 percent for households located in remote areas. The education coefficients are jointly significant across both samples. Household heads who reported earning some higher level education spend considerably more resources on fertilizer usage as compared to those households that reported being illiterate. These households spent 100% more on fertilizer compared to households in which the head reported being illiterate. In addition to management capacity, these human capital variables could be also be picking up differences across households associated with income. Wealthier households can afford outlays of capital at the beginning of the agricultural season, while less wealthy households are more likely to face liquidity constraints.

The ADMARC variables do not explain fertilizer usage in local maize among households located in proximity to a primary road, but they are jointly significant at the 5% level for households located in relatively remote areas. Households located less than 30 minutes away from the nearest ADMARC market are likely to spend almost 100% more on fertilizer than households located over 2 hours travel time to the nearest ADMARC market. However, while the relative magnitudes of the ADMARC coefficients are logically consistent, they are negative for two of the three ADMARC variables. This makes interpretation of the results quite problematic.

In columns (3) and (4) we report the demand for fertilizer usage on hybrid maize. It is likely that fertilizer is more intensively applied on fields planted with hybrid maize compared with fields with local maize because of its production technology. Optimal hybrid maize production requires fertilizer usage; local maize production, in contrast, is less sensitive to the application of fertilizer. Household size positively affects fertilizer usage, possibly through the labor requirements associated with fertilizer application. An additional one percent increase in household size leads to a 12.3 % increase in the value of fertilizer purchased by the household in areas of close proximity to a primary road and a 9.7% increase by households located in remote areas. Across the entire sample, education positively influences fertilizer usage. Household heads that reported educational attainment beyond primary school applied over 100% more fertilizer to their hybrid maize fields compared with household heads who reported being illiterate. The education coefficients may be picking up both a wealth effect and a farm management effect.

Distance to the nearest reported ADMARC has no impact on households located nearby to a primary road while this set of dummy variables is jointly significant for the remote sample of households. Further the signs and relative magnitudes of the ADMARC coefficients are also logically consistent.

43

The differential impact of ADMARC distance on fertilizer usage across the sample is an interesting finding. While it has been recognized that maize trading by the private sector proliferated with agricultural liberalization in Malawi, there is little consensus about the impact of liberalization on the fertilizer trade. The purchase and importation of fertilizer requires relatively large outlays of capital, thus imposing formidable barriers on business entry. If the availability of fertilizer had not expanded beyond ADMARC facilities, we would expect to see little variation on the impact of ADMARC access across these two types of households.

Hence the overall findings on production activities is as follows: for households that have access to market institutions and live in relatively more commercialized areas, ADMARC services do not seem to be important in determining yields or profits. In contrast, households located in remote areas may rely heavily upon ADMARC, especially when access to other market infrastructure is limited.

Consumption Expenditures regression. According to Kherallah et al (2002), home production of maize represents over two-thirds of the household consumption with about 70% of the households supplementing their own production with maize purchases. Regression equation relating household per capita consumption with access to ADMARC is reported in Table A9.

Distance to the nearest ADMARC facility may affect household consumption through several channels. As indicated before, rural households may utilize ADMARC both for points of sale for their commodities and as a source for purchasing maize for their own consumption. Increased farm income that results from selling to ADMARC will translate into higher per capita consumption. Additionally, low consumer maize prices offered by ADMARC will have a positive impact on the household budget.

However, ADMARC’s impact on consumption levels will be tempered by the accessibility of alternative markets. The degree of commercialization that a household experiences will determine the opportunities for engaging in trade with other institutions and the outcomes of those trade, such as the negotiated price. Thus, similar to the strategy outlined above, we split the sample according to proximity to the nearest primary road to understand how market accessibility conditions the effect of ADMARC access.

In columns (1) and (2) we report expenditure estimates for the sample of households located nearby to a major primary road. On average, these households are located about 1.3 kilometers from the nearest road. Across both specifications, female-headed households consume about 7-8% less than households headed by males. The dependency ratio of children to adults has a negative impact on per capita consumption. Per capita consumption drops by almost 5% for every additional 10 percent share of young children associated with the household. Household size also has a negative impact of dependents on per capita consumption. For every additional household member, per capita consumption is likely to fall by between 8 and 9 percent.

In addition to the household-level variables included in the production-related equations, we include occupation related variables to control for the effect of occupational choice on earnings and preferences. We also include the imputed area-level maize producer price as an additional regressor. This price has an overall positive effect on expenditures. ADMARC accessibility positively affects household expenditures regardless of proximity to markets and transportation. Per capita consumption is between 11 and 18% higher for households that reported living within 30 minutes to the nearest ADMARC facility compared with those households that reported living over 2 hours away from the nearest facility. The relative magnitudes of the three ADMARC variables are logically consistent in all the equations, even though it is negative in the regression with district dummies.

Columns (3) and (4) report parameter estimates for households located in remote areas. On average, households are located about 6.5 kilometers from the nearest primary road . These estimates are similar to those reported for the sample of households located in close proximity to a

44

main road. The household variables that capture productive labor capacity and the dependency ratio of children to adults have a negative and significant impact on household consumption. Schooling attainment at all levels has a positive impact on consumption.

ADMARC proximity has a positive effect on consumption in areas where households live far from major access roads. Households that reported living within 30 minutes to an ADMARC facility consume about 20-22% more per capita than households located in excess of 2 hours travel to the nearest ADMARC facility. Relative magnitudes are logically consistent even though one ADMARC coefficient turns out to be negative when district dummies are used.

Generally, in contrast to production related findings, the relationship between household consumption and ADMARC proximity does not seem to be conditioned by the availability of alternative markets. This finding suggests that traders engage mainly in buying of maize from farmers after harvest, less on selling maize to consumers. They transport maize out of remote areas and don’t engage in storage activities that would potentially increase the availability of maize for consumption after harvest. B6. Conclusions

This paper utilized data on smallholder production activities and consumption expenditures to assess the relative importance of ADMARC markets on rural welfare in Malawi14. The ADMARC effect is proxied by the distance to ADMARC since household level data on ADMARC transactions is limited.

This paper used both 1997/98 household survey data and geographic information to show that after controlling for household and area effects, the average importance of ADMARC in maize production in Malawi is differentiated according to infrastructure access. Specifically, we stratify the sample of households according to distance to the nearest primary road and find that ADMARC access has a positive effect on smallholder production outcomes in areas where exposure to developed market institutions and infrastructure is limited. These benefits are attributed to the fact that alternative agricultural market institutions do not generally develop in poor, remote areas.

Households that are located in areas where infrastructure is the most readily accessible appear to be less dependent on ADMARC. In areas where private market services and market infrastructure function, ADMARC markets may no longer be necessary for farm households15. Not withstanding this, the development of complementary institutions that facilitate greater transparency and competition in market transactions would further enhance private sector participation in the agricultural sector.

These findings suggest that some careful thought be given to the restructuring ADMARC. First, to leave ADMARC in its current form would at best be inefficient and wasteful use of government resources. In a worse-case scenario, leaving ADMARC in its original form may be harmful to both traders who are crowded out of the market by ADMARC’s erratic signals and actions, and also to producers whose market transactions are affected adversely by ADMARC’s presence. Second, to remove ADMARC completely may ignore the fact that well-functioning market institutions and infrastructure do not exist uniformly throughout the country. In the short-run, the provision of ADMARC facilities may be necessary in some areas that are under-served. This provision should be made with the caveat that government investment in market infrastructure would eventually eliminate the need for ADMARC support.

14 Clearly, this paper is limited by the fact that more precise information about agricultural market transactions was not available. Soil fertility indicators or suitable proxies were also not available. Moreover without panel data, it was also not possible to assess whether the presence of ADMARC markets contributed to consumption-smoothing or reducing household vulnerability. 15 It would be of interest to examine whether the presence of the parastatals in such areas actually crowd out the private sector.

45

Table B1 Maize Production and ADMARC Purchase Volumes Year

National Production

(metric tons)

ADMARC Purchases

(metric tons)

Percentage 1995-6 1,327,865 61,280 4.61% 1996-7 1,619,307 87,228 5.39% 1997-8 1,226,478 96,489 7.87% 1998-9 1,772,392 12,586 0.71% 1999-0 2,245,824 57,739 2.57% 2000-1 2,211,859 196,656 8.89% Source: ADMARC Table B2 ADMARC Proposed Closures Region

Capacity (metric tons)

Percentage (of total)

Proposed Disposal (metric tons)

Percentage

Central

123,910

66.1

22,370

18.05

North 19,310 10.3 1720 8.90 South 44,180 23.6 2000 4.53 Table B3 - Household Distance to Nearest ADMARC Market Time taken to reach | nearest ADMARC | Percent --------------------------------------------------- half hour or less | 20.72 over half hour to 1 hour | 24.61 over 1 hour to 2 hours | 29.05 over 2 hours | 25.08 no response | 0.55 ------------------------------------------------------------------------------- Total | 100.00

46

Table B4 - Comparison of Household Means by Road Access Close to Road Far from Road T-Statistic Uc Uf (Ho: Uc = Uf ) ln (hybrid maize yield) 5.90 5.96 1.4 [0.08] kilos per acre ln (hybrid fertilizer) 3.00 2.74 1.98 [0.02] kilos per acre ln (hybrid seed) 3.2 2.9 3.15 [0.00] kilos per acre ln (local maize yield) 5.27 5.37 2.73 [0.00] kilos per acre ln (local fertilizer) 0.78 0.83 0.51 [0.31] kilos per acre ln (local seed) 0.24 0.47 4.27 [0.00] kilos per acre ln (total profit) 0.67 0.94 4.75 [0.00] kwachas per acre ln (total land) 0.66 0.67 0.60 [0.27] acres ln (expenditures) 3.30 2.27 1.63 [0.05] kwachas Household size 4.55 4.27 4.57 [0.00] Number of Adult Males 1.04 0.99 2.39 [0.02]

47

Table B5 - Local Maize Productivity with TA- and District-level Dummies Dependent VariableLn (local maize yield)

Independent VariablesClose to Main Road Far from Main Road(1) (2) (3) (4)

ln(land) -0.054 -0.100 -0.166 -0.213(1.35) (2.38)** (2.69)** (3.26)**

ln(land)2 0.000 -0.012 0.061 0.102(0.00) (0.55) (1.81)* (2.76)**

ln(fertilizer) -0.025 -0.001 -0.125 -0.115(0.44) (0.01) (2.43)** (2.02)*

ln(fertilizer)2 0.018 0.014 0.029 0.027(2.03)* (1.45) (3.62)** (3.08)**

ln(seed) -0.060 -0.167 -0.091 -0.193(1.10) (2.77)** (1.33) (2.68)**

ln(seed)2 0.011 0.030 0.015 0.030(0.97) (2.51)** (1.12) (2.11)*

Female-Headed HH -0.057 -0.063 -0.194 -0.182(1.19) (1.15) (3.83)** (3.27)**

Age -0.001 -0.001 -0.003 -0.003(1.10) (0.75) (2.10)* (2.35)**

Schooling CategoriesPrimary -0.045 0.036 -0.074 -0.060

(0.81) (0.57) (1.32) (0.97)Secondary 0.098 0.162 -0.017 0.041

(1.70)* (2.52)** (0.29) (0.66)Higher 0.060 0.161 -0.043 0.003

(0.60) (1.49) (0.41) (0.03)Time to ADMARCLess than 30 minutes 0.076 0.074 0.253 0.158

(0.77) (0.97) (2.20)* (1.86)*30- 60 minutes 0.059 0.020 0.235 0.124

(0.67) (0.27) (2.68)** (1.74)*60 –120 minutes 0.016 0.080 0.237 0.186

(0.21) (1.14) (3.24)** (3.07)**TA-level variablesDistance to Market -0.006 0.000

(1.38) (0.08)Population of TA/Ward 0.000 0.000

(0.43) (4.71)**Constant 5.196 5.242 5.387 5.119

(51.82)** (35.69)** (57.01)** (36.44)**

Observations 1575 1447 1592 1431

F-TestsTime to ADMARC (3) 0.26 0.63 3.78** 3.24*Land (2) 1.05 4.21** 3.63* 5.41**Fertilizer (2) 46.24** 43.53** 30.72** 25.38**Seed (2) 0.67 4.12* 1.14 6.01**Education (3) 2.13^ 2.41* 0.66 0.96Absolute value of z-statistics in parentheses ^ significant at 10%; * significant at 5% level;** significant at 1% level

48

Table B6 - Hybrid Maize Productivity with TA- and District-level Dummies Dependent VariableLn (Hybrid Maize Yield)

Independent Variables

Close to Main Road Far from Main Road(1) (2) (3) (4)

ln(land) 0.080 0.149 0.149 0.182(1.91) (3.43)** (3.60)** (4.03)**

ln(fertilizer) 0.134 0.132 0.082 0.087(12.73)** (12.28)** (8.37)** (8.36)**

ln(seeds) 0.010 -0.003 0.072 0.070(0.76) (0.19) (5.70)** (5.34)**

Female-headed HH -0.102 -0.127 -0.073 -0.011(1.50) (1.70) (1.08) (0.14)

Age 0.005 0.004 0.006 0.005(3.00)** (1.87) (3.35)** (2.79)**

Schooling CategoriesPrimary 0.033 0.001 0.043 0.165

(0.41) (0.01) (0.57) (1.92)Secondary 0.083 0.001 0.159 0.265

(1.08) (0.01) (2.24)* (3.30)**Higher 0.242 0.229 0.287 0.356

(2.52)* (2.26)* (2.72)** (3.02)**Distance to ADMARCLess than 30 minutes 0.116 -0.000 0.243 -0.159

(0.92) (0.00) (1.88) (1.58)30 – 60 minutes 0.043 0.035 0.260 0.046

(0.37) (0.39) (2.45)* (0.52)60 – 120 minutes 0.003 0.075 0.125 -0.069

(0.03) (0.86) (1.41) (0.90)Distance to Market 0.003 0.005

(0.69) (1.04)Population of TA/Ward -0.000 0.000

(0.50) (2.01)*Constant 5.069 5.306 4.971 4.813

(36.39)** (26.61)** (40.57)** (25.15)**

Observations 1203 999 1186 993R-squared 0.52 0.37 0.55 0.35

F-TestsDistance to ADMARC (3) 0.44 0.36 3.37** 1.76Education (3) 2.41^ 3.04* 2.08^ 4.45**

Absolute value of t-statistics in parentheses ^ significant at 10% ;* significant at 5%; ** significant at 1%

49

Table B7 - Total Farm Profits with TA- and District-level Dummies Dependent VariableLn (Total Profits)

Independent VariablesClose to Main Road Far from Main Road(1) (2) (3) (4)

ln(land) 0.702 0.720 1.002 0.907(10.93)** (11.96)** (15.70)** (14.44)**

# male adults 0.046 0.090 0.092 0.115(0.74) (1.44) (1.63) (1.92)

# female adults 0.003 0.023 0.047 0.047(0.05) (0.37) (0.85) (0.79)

female-headed household -0.419 -0.395 -0.411 -0.582(3.77)** (3.53)** (4.00)** (5.27)**

Age 0.030 0.037 0.025 0.023(2.47)* (3.02)** (2.25)* (1.86)

(Age)2 -0.000 -0.000 -0.000 -0.000(2.90)** (3.86)** (2.89)** (2.47)*

School CategoriesPrimary -0.103 -0.118 0.073 0.058

(0.98) (1.12) (0.81) (0.59)Secondary 0.090 0.036 0.272 0.258

(0.83) (0.33) (2.95)** (2.58)**Higher 0.180 0.175 0.417 0.482

(1.24) (1.22) (2.94)** (3.13)**Distance to ADMARC

Less than 30 minutes 0.124 -0.215 0.322 0.323(0.72) (1.68) (1.93) (2.56)*

30 – 60 minutes 0.099 0.008 0.046 0.258(0.65) (0.07) (0.35) (2.28)*

60 – 120 minutes 0.068 0.074 0.108 0.177(0.52) (0.68) (0.97) (1.86)

Distance to Market -0.001 0.016(0.14) (2.67)**

Population of TA/ ward 0.000 0.000(0.63) (4.90)**

Constant 5.450 5.343 5.347 4.540(17.92)** (16.17)** (20.34)** (13.84)**


F-TestsAge (2) 5.78** 1.79 8.68** 7.04**Education (3) 1.94 13.97** 4.84** 4.76**Distance to ADMARC (3) 0.19 2.31^ 1.82 2.61*

Absolute value of t-statistics in parentheses* significant at 5%; ** significant at 1%

50

Table B8 - Fertilizer Demand with TA- and District-level Dummies Dependent VariablesFertilizer Demand – local maize and hybrid maize

Independent Variablesln (fertilizer) ln(hybrid fertilizer)(1) (2) (3) (4)Near to Rd Far to Rd Near to Rd Far to Rd

ln(land) 0.853 1.024 0.530 0.073(8.41)** (7.79)** (4.16)** (0.52)

Household size 0.040 -0.016 0.123 0.097(1.27) (0.44) (3.20)** (2.18)*

share male adults 0.490 -0.070 -0.723 0.810(1.45) (0.19) (1.59) (1.67)

Female-Headed HH 0.179 -0.346 -0.200 -0.190(1.07) (1.73) (0.81) (0.70)

Age -0.006 0.000 0.004 -0.001(1.53) (0.06) (0.77) (0.13)

EducationPrimary 0.132 0.208 0.030 0.099

(0.84) (1.22) (0.12) (0.39)Secondary 0.341 0.332 0.712 0.890

(2.12)* (1.91) (3.03)** (3.78)**Higher 1.091 0.717 1.649 1.534

(3.94)** (2.26)* (5.63)** (4.39)**Time to ADMARC

Less than 30 minutes 0.525 0.687 0.273 1.492(1.89) (1.96) (0.70) (3.49)**

30 – 60 minutes 0.348 -0.000 -0.007 0.824(1.42) (0.00) (0.02) (2.32)*

60- 120 minutes 0.259 -0.293 0.233 0.284(1.21) (1.31) (0.73) (0.96)

Constant 0.475 0.934 1.944 1.424(1.45) (2.72)** (4.24)** (2.93)**


F-TestsEducation (3) 5.45* 2.19^ 13.50** 10.58**Distance to ADMARC (3) 1.21 3.42* 0.61 4.54**

Absolute value of t-statistics in parentheses ^significant 1t 10%;

* significant at 5%; ** significant at 1%

51

Table B9 - Household Per capita Expenditures with TA- and District-level Dummies Dependent Variable: ln (per capita expenditure)

Independent Variables:

Near to Main Road Far from Main Road(1) (2) (3) (4)

Female-Headed HH -0.078 -0.087 -0.093 -0.114(3.01)** (3.01)** (3.47)** (3.67)**

Share of children < 5 -0.481 -0.485 -0.466 -0.416(6.91)** (6.19)** (6.78)** (5.16)**

HH size -0.089 -0.084 -0.092 -0.083(16.81)** (14.42)** (17.34)** (13.56)**

Ln(land) 0.029 0.029 0.018 0.006(1.60) (1.58) (1.03) (0.33)

Occupational Categoriesadministrative -0.011 -0.001 0.032 0.032

(0.15) (0.02) (0.41) (0.38)clerical 0.018 -0.020 0.002 -0.031

(0.28) (0.28) (0.04) (0.44)sales 0.037 0.027 0.044 0.035

(0.80) (0.55) (1.01) (0.71)services 0.042 0.032 -0.047 -0.012

(0.84) (0.62) (0.92) (0.21)agriculture -0.000 0.034 0.051 0.055

(0.01) (0.95) (1.44) (1.47)other 0.006 -0.00 0.090 0.108

(0.10) (0.01) (1.40) (1.36)Age -0.005 -0.007 -0.000 -0.006

(1.42) (1.77) (0.12) (1.29)Age2 0.000 0.000 0.000 0.000

(0.79) (1.48) (0.16) (1.19)

School CategoriesPrimary 0.026 0.044 0.071 0.081

(0.88) (1.29) (2.41)* (2.36)*Secondary 0.169 0.161 0.171 0.182

(5.51)** (4.75)** (5.74)** (5.26)**Higher 0.509 0.511 0.459 0.431

(11.86)** (11.05)** (10.14)** (8.32)**

Time to ADMARCLess than 30 minutes 0.111 0.177 0.199 0.183

(2.04)* (4.39)** (3.65)** (4.18)**30 – 60 minutes 0.041 0.050 0.063 0.018

(0.83) (1.25) (1.43) (0.48)60 – 120 minutes 0.003 -0.022 0.003 -0.023

(0.06) (0.56) (0.09) (0.72)

Market Distance -0.004 0.006(2.14)* (2.56)*

Population of ta/ ward -0.000 -0.000(1.55) (1.64)

Maize producer price 0.064 0.142(1.10) (3.04)**

Constant 2.797 2.710 2.580 2.299(28.56)** (12.11)** (26.69)** (11.41)**

Observations 2703 2335 2655 2269

52

R-squared 0.43 0.25 0.45 0.23

F-testsADMARC (3) 2.16^ 11.36** 5.70** 8.95**Education (3) 52.94* 45.17** 36.97** 25.22**Age (2) 4.87** 2.20 0.03 0.89Enumeration areadummies 4.67** 6.50**

District area dummies 6.95** 9.67**

Absolute value of t statistics in parentheses* significant at 5%; ** significant at 1%

53

ANNEX C - QUALITATIVE STUDY PREPARED BY WADONDA CONSULT (ZOMBA)

54

POVERTY AND SOCIAL IMPACT ASSESSMENT IN

MALAWI: CLOSURE OF ADMARC MARKETS

DRAFT FINAL REPORT

Peter M. Mvula

Ephraim W. Chirwa John Kadzandira

Wadonda ConsultWadonda ConsultWadonda ConsultWadonda Consult PO Box 669, Zomba Malawi, Central Africa Web: http://www.geocities.com/wacoafrica

Tel/Fax: (265) 01 527 399 / 01 527 747 Mobile: (265) 08 839 296 / 09 930 200 Mobile: (265) 08 827 933 /09 951 788 Email: [email protected] Email: [email protected]

January 2003

WACO

55

TABLE OF CONTENTS Executive Summary 1. Introduction 1.1 Background to the Study 1.2 State Intervention in Agricultural Produce Marketing in Malawi 1.3 Scope of the Study 1.4 Organisation of the Report 2. Methodology and the Study Process 2.1 Introduction 2.2 Selection of Field Sites 2.3 Data Collection Methods 2.4 The Study Process 2.5 Limitations of the Study 3. Description of Study Sites and Livelihood Features 3.1 Introduction 3.2 Description of Study Sites

3.3 Perceptions on Well-being 3.4 Concluding Remarks 4. The Current Roles of ADMARC and Private Traders in Agricultural

Marketing Activities 4.1 Introduction 4.2 Current Roles of ADMARC in Sample Sites 4.3 The Marketing Operations of Private Traders

4.4 Comparative Performance of ADMARC and Private Traders 4.5 Concluding Remarks 5. Stakeholders’ Response to ADMARC’s Declining Business and

Marketing Liberalization 5.1 Introduction 5.2 The Effects of ADMARC’s Declining Business and Closure

of Seasonal Markets 5.3 Households’ Reactions in Closed and Active Markets 5.4 Response by Other Stakeholders 5.5 Concluding Remarks 6. Main Findings, Implications and Recommendations 6.1 Introduction 6.2 Main Findings 6.3 Implications 6.4 Recommendations References Annex 1 List of Key Informants for Institutional Interviews Annex 2 Infrastructure of the Study Sites Annex 3 Guide for Focus Group Discussion in ADMARC’s Active Markets Annex 4 Guide for Focus Group Discussion in ADMARC’s Closed Markets Annex 5 Guide for Semi-Structured Interviews in ADMARC’s Active Markets Annex 6 Guide for Semi-Structured Interviews in ADMARC’s Closed Markets Annex 7 Guide for Key Informants Interviews in ADMARC’s Active Markets Annex 8 Guide for Key Informants Interviews in ADMARC’s Closed Markets Annex 9 Guide for Private Traders Annex 10 Guide for Stakeholders Interviews

56

List of Abbreviations ADMARC Agricultural Development and Marketing Corporation ADD Agricultural Development Division AHL Auction Holdings Limited AIDS Acquired Immuno-deficiency Syndrome CADECOM Catholic Development Commission CGL Cotton Ginnery Limited ELDP Evangelical Lutheran Development Programme EPD Economic Planning and development EU European Union FGD Focus Group Discussion FSG Food Studies Group GoM Government of Malawi GTZ German Technical Cooperation HGD Hardware and General Dealers HIV Human Immuno-deficiency Virus LLTC Limbe Leaf Tobacco Company MASAF Malawi Social Action Fund MK Malawi Kwacha MPRSP Malawi Poverty Reduction Strategy Paper MRFC Malawi Rural Finance Company NASFAM National Association of Smallholder Framers NEC National Economic Council NGO Non Governmental Organization PSIA Poverty and Social Impact Assessment PTC People’s Trading Centre QIM Qualitative Impact Monitoring WFP World Food Programme SFFRF Smallholder Farmer Fertilizer Revolving Fund SSI Semi Structured Interviews TAMA Tobacco Association of Malawi

57

Acknowledgements

This report benefited from advice, comments and suggestions from different individuals and institutions. The authors of this report wish to thank Reiner Foster and Janet Owens of the World Bank for the guidance and suggestions throughout the study period and the many constructive comments made on earlier drafts of the report. We also acknowledge the contribution made by the Department of Economic Planning and Development (the then National Economic Council) through Dr. Milton Kutengule and Mrs Lana Chikhungu. Special thanks also go to Mr Chipungu of the Agricultural Development and Marketing Cooperation. We also thank Sabina Schnell and Renate Kirsch for the comments made on earlier drafts of the report.

We also thank our team of research assistants who showed dedication and willingness to be trained in new ways conducting research and for their contribution and active participation in a research design workshop in developing the survey instruments and their participation in the data analysis workshop. These research assistants include Vincent Gondwe, James Mwera, Veronica Mbulaje, Bessey Chirwa and Titus Nyirenda. We also acknowledge the contributions made by Lilian Saka, Francis Saka and drivers Messrs S. Mattaka and L. Mizimbe.

Last but not least, we acknowledge the excellent secretarial services offered by Mrs Eluby Kafumbata and Miss Patricia Gadi.

58

Executive Summary Introduction and Background

Over the years, agricultural marketing systems have undergone several reforms in developing country particularly in the advent of structural adjustment programmes implemented by many countries since the 1980s. These changes have ranged from liberalisation of marketing services to privatization of state marketing agencies, and the experiences in achieving intended objectives of such reforms have rather been mixed.

Within this policy context, and within the framework of structural adjustment programmes, agricultural markets in Malawi have undergone several reforms. Notably, the reform process has included the liberalisation of agricultural produce and input pricing, liberalization of agricultural marketing services that opened up the activities to the private sector, restructuring of the state marketing agency through portfolio divestment and closure of unprofitable markets in order to improve its operational and financial position. The expectation from these reforms was that the private sector would efficiently replace ADMARC activities in markets where the state marketing agency had withdrawn. The first round of closure of ADMARC markets occurred in 1988, in which 125 seasonal markets were closed due to their insignificance in the volume of trade.

The operational and financial performance of ADMARC has continued to deteriorate in spite of the reforms that have been implemented. This has exerted a lot of pressure to hasten the privatization of ADMARC as encapsulated in the Malawi Poverty Reduction Strategy Paper (MPRSP). Several proposals have been advanced on how ADMARC can improve its economic performance. One of the proposals is to restructure ADMARC by rationalising its operations through re-organisation of markets and regional offices and closure of 204 unit markets and converting others into temporary markets throughout the country. The proposed closure of some ADMARC markets is against the background of ADMARC’s declining business activities since 1985 due to liquidity problems and operational inefficiency.

It is within the context of the proposed closure of some ADMARC markets that this study was commissioned to investigate the likely impact of such a policy on the livelihood of affected rural households and understand how households are likely to cope and the identify opportunities that may arise from such a policy. The main objectives of the study are to learn from poor rural households and other main stakeholders about the effects of closing ADMARC markets on their well-being and economic activity; to explore people’s reactions to the closure of the markets including their adjustment and coping strategies; and to extrapolate the findings of the qualitative ex-post field research to the proposed closure of further markets and draw draft policy implications regarding a socially responsible design, implementation arrangements and sequencing of the reform. Methodology of the Study

The study was conducted in the catchment areas of 10 ADMARC unit markets. These markets were purposively selected on the basis of their remoteness, the dominant livelihood characteristics. The 10 ADMARC unit markets include 6 unit markets that were earmarked for closure and 4 unit markets that were still active. In the 6 unit markets earmarked for possible closure, interviews were conducted in villages in the catchment area of closed seasonal markets, as it turned out that there were no unit markets that had been closed in the past five years. In each of the 10 ADMARC unit markets, participatory rural appraisal methodologies, focus group discussions and key informant interviews were conducted in two villages. This implies that a total of 20 villages were included in this study, resulting in a total of 40 focus group discussions (half with women groups), 54 semi-structured interviews, 9 interviews with ADMARC officials, 3 interviews with private traders and 32 interviews with other key informants. Given the qualitative nature of the study, caution must be exercised in generalizing the results of the study. Moreover, the study was conducted against the background of the famine that engulfed Malawi and southern Africa in the 2001/2002 season, and

59

some of the views from the smallholder farmers could be influenced by the famine situation. In addition, there were no closed unit markets, and the impact of closure of unit markets is inferred from closed seasonal markets and the general situation of declining business activities in most of its markets. Main Findings From The Study

The major finding of the study is that rural communities particularly in remote areas have been adversely affected by the withdrawal of ADMARC markets or the decline in the business activities of ADMARC due to its financial and operational problems. The poor lack reliable markets for the little surplus produce that they market and affordable maize in the lean season or when they run out of food 1 to 2 months after harvest. Although the private sector, in form of large private traders and vendors, has emerged to take advantage of the opportunities that have created by the liberalisation of agricultural markets and the weak position of ADMARC, they have not been able to efficiently close the gap that ADMARC has left particularly in the sale of maize at affordable prices during the lean season and in the supply of agricultural inputs to the moderate poor and the poor that use such inputs. Private traders that deal in inputs are located in urban and peri-urban centres, thereby increasing the transport and transaction costs of procurement for the larger proportion of the rural population. The following are the specific findings of the study.

• Communities in the study areas mostly define poverty in terms of household food security, housing characteristics, possession of assets and access to financial markets. Food security appears to be the most important distinguishing feature of poverty, with the ‘poor’ reported only having food stocks from own production that last within 1 to 2 months after harvesting while the ‘well-to-do’ have adequate food throughout the year. The poor are a larger proportion of the communities and the share of the poor has increased over the last decade.

• The business activities of ADMARC in agricultural marketing, particularly the purchase of

agricultural produce from smallholder farmers, have been declining since the introduction of economic reforms in the early 1980s. Smallholder farmers now sell their produce in alternative market channels, and ADMARC has lost its market share in maize and tobacco marketing.

• ADMARC markets have an important role as distribution networks for affordable maize in

the lean season and in times of famine, in providing the benchmark prices, in providing a reliable source of inputs and in the purchase of crop produce from farmers as last resort institution in rural Malawi. These roles of ADMARC in the recent famine were minimal and sporadic. ADMARC did not purchase maize from farmers because of liquidity problems and did not sell much maize and inputs to rural households.

• There has been proliferation of unlicensed retailers and vendors in areas where accessibility is

not an important constraint following the liberalisation of agricultural markets. The extent to which large-scale to medium scale licensed private traders operate in the study sites remains decimal and scanty. Although most of the unlicensed small retailers and vendors actively participate in the produce market, they do not participate in the marketing of agricultural inputs and the input supply role of ADMARC has hardly been substituted by the private sector. The large-scale private traders that deal in inputs are usually located in the peri-urban and urban areas.

• Since the private market is dominated by small-scale local vendors, the function of inter-

regional and inter-seasonal arbitrage which is performed by ADMARC could not effectively be taken over by another actor. The reduction in unit markets is likely to affect the extent to which ADMARC could market maize to deficit areas. This is particularly important in times

60

of food shortage in one region which could be compensated by the surpluses of another region, if there were a functioning national market.

• The business practices of private traders are less favourable to smallholder farmers. Private

traders tend to announce higher prices for crop produce and their timing of trade early in the season coincides with the cash needs of smallholder farmers, but the effective producer prices may not be substantially greater than those offered by ADMARC because of the widespread cheating on measurements, their false quality assessment of farmers produce and through high degree of price discrimination partly due to the weak bargaining power of individual farmers.

• The liberalization of agricultural marketing has benefited more the richer rural households

and vendors compared to the poorer segment of the rural population. This has strengthened the position of the rural elite and ‘well-to-do’ households as the poor tend to do more ganyu as a survival strategy. A large proportion of poorer farmers are less able to access inputs and staple food through market channels and have increased their reliance on the ‘well-to-do’ for paid labour (ganyu).

• The survival and coping strategies that have been adopted by rural households and

smallholder farmers in reaction to the closure of seasonal markets and declining role of business to sustain their livelihoods and food security include increasingly relying on private traders, particularly vendors, in the produce market and maize procurement; resorting to more to ganyu to earn income or to obtain maize as in-kind payment; and expansion of cropping areas for the main staple food and diversification towards drought resistant crops.

• Many stakeholders believe that the problem of ADMARC does not lie in the operations of

rural market, but rather at the institutional structure and the efficiency of the head office and the burden of unprofitable subsidiaries and other portfolio investments.

Implications of the Study

The findings of this study based on data from 20 villages have implications for some of the 204 markets that are targeted for closure or those where ADMARC activities have substantially declined over the years with similar socio-economic and remoteness characteristics.

• Many smallholder farmers, particularly the ‘poor’ and the ‘a-bit-well-to-do’, will have to pay more to access the same amount of maize to private traders/vendors without ADMARC than with ADMARC markets in the short and medium terms. In addition, transaction costs due to travelling longer distances to alternative active markets and search costs for suppliers of maize during the lean periods will increase.

• The problem of input availability, combined with the high cost of inputs from the private

traders and the high transportation costs in the production of maize and other food products will lead to a decline in the utilization of fertilizers and improved seeds leading to lower productivity and cultivation of marginal land by many smallholder farmers.

• The closure of ADMARC markets will remove the benchmarking pricing role that ADMARC

currently plays in the rural areas in which most private traders already act as discriminating monopolist. There may be a reduction in the effective producer prices offered by private traders and vendors resulting in strengthened countervailing power of private traders vis-à-vis the individual poor farmer.

61

• Many smallholder producers are likely to lose incomes through cheating practices and monopsonistic power of private traders/vendors if appropriate institutions and regulations are not put in place to protect the interests of farmers and consumers.

• The risk of food insecurity for many rural households will increase even in normal years due

to the behaviour of large scale private traders of only purchasing maize from the rural areas and not reselling the maize in the lean period. In addition, the risk of food insecurity will also be perpetuated because some of the vendors who buy maize from ADMARC unit markets for sale to the remote areas may find difficult to do so or will experience higher transaction costs in accessing distant active ADMARC unit markets.

• With the poor infrastructure in many rural markets the further increase in the number of

private traders will be marginal and the mobility of private traders is less likely to create competitive trading activities necessary to discipline private traders.

• The closure of markets may aggravate food insecurity situation of rural communities in times

of hunger and famine because relief food is likely to move slowly to the areas due to the poor road infrastructure and due to lack of storage facilities and lack of a reliable distribution network.

• The producer and consumer price volatility will remain high, and this will work to the

advantage of the private traders who can generate arbitrage and monopsony or monopoly profits from price variability but smallholder farmers will be disadvantaged with the unpredictable prices continuing to play a very little role in the farmers planning decisions for food crops. Consumer food prices are likely to increase in the short-run, but the extent to which prices will fall with closure of some markets in the medium or long term will depend on the competitiveness of the private food marketing system.

Recommendations

The marketing services of ADMARC in the rural areas tend to benefit the rural population in Malawi and it is not surprising that rural communities have recently linked the increase in poverty to some extent to the closure of ADMARC markets. Liberalisation of markets is not necessarily a bad strategy for agricultural markets. However, it is important to ensure that a conducive environment exists for transition from a state-dominated marketing system to a private sector led marketing system. If the government is to proceed with the proposed closure of ADMARC markets, it is important to condition such a process on the development of infrastructure and appropriate institutions to facilitate competitive private sector activities and increase the bargaining power of smallholder farmers. We make the following specific recommendations based on the results of the study:

• The government, donors, multilateral institutions and ADMARC should be careful in advocating the proposed closure of 204 ADMARC markets by taking into account the framework of poverty alleviation and the state of supporting infrastructures and the need to have appropriate food distribution network in the frequently occurring droughts or food crises in Malawi.

• Concrete decisions to close ADMARC markets or to privatise ADMARC should be done on

case by case basis, supported by sound empirical studies on how markets function that involve extensive consultations with the main stakeholders, the smallholder farmers and rural poor households that attach more value to the marketing services provided by ADMARC. More in-depth research on the capacity of the private sector in performing these market functions is necessary for informed decision making or policies.

62

• In cases where closure of markets is unavoidable, the government and ADMARC should ensure that appropriate institutions and infrastructure are in place before such markets are closed. The government should ensure that the proposed closed markets are linked through good all-weather roads to facilitate the movement of private traders and better communication facilities.

• There is need to strengthen the organization of farmers into marketing co-operatives or

marketing associations or farmers price and business practices monitoring committees to promote good business practices and provide countervailing power to private traders.

• There should be a comprehensive review of ADMARC as an institution, particularly focusing

on the management structure and management efficiency and the role of its commercial subsidiaries in the performance of its marketing functions, its mandate and the viability of the dual social and economic functions and the extent to which political interference can be minimized or eliminated.

• While in the short or medium term the social and developmental functions of ADMARC will

still be desirable because some of the social functions it performs can not be substituted through the private marketing system in the short-run, there is the need to spell out clearly and explicitly the so-called ‘social function’ of ADMARC. The costs related to such a social function need to be made transparent in order to allow for a sensible discussion of how these social costs should be financed through cross-subsidization across regions or business lines, national budget, consumer prices. For instance, the social function of ADMARC could imply that ADMARC should only withdraw from markets where the private marketing system is working efficiently and concentrate its activities in areas where the private marketing system is not working efficiently through organization of farmer co-operatives or community managed commodity exchanges facilitated by ADMARC within an agreed time frame.

• If the long-term objective in agricultural marketing is the development of the private

marketing system, the government needs to properly sequence the privatization of ADMARC. The first step in such a sequence is for ADMARC to be restructured to exclusively concentrate on its marketing function and to be allowed to operate commercially in all crops produced by smallholder farmers and act as an effective competitor to private traders. In the short-term it should continue to perform its social function with respect to food security and development of alternative community-based marketing institutions, but it should be given adequate flexibility (reasonable margin) in the pricing of maize. This implies that ADMARC should divest ownership in all enterprises. The second step should be the privatization of ADMARC unit markets infrastructure in areas where the private sector is actively substituting the marketing services of ADMARC. Concurrently, ADMARC should embark on a ‘develop, operate and privatise (DOP)’ programme of markets that are less accessible by private traders or less serviced by private traders.

• If the main reason for closing the proposed unit markets is to cut costs (particularly staff

costs) in the short-term, other alternative ways should be explored giving each market some years of observation before such a market is closed subject to the proposed privatization sequence. Short-term cost cutting measures would include organising some unit markets around one core unit market and the surrounding unit markets could open on specific days similar to the ‘salaula’ and ‘tiwonewone’ concept. Three markets could be operated by one officer on bi-weekly basis, but the frequency of opening such markets could be increased during times of famine.

• In circumstances that ADMARC has excess storage space such space should be sub-let at

commercial fees to private traders operating in the areas in the short-run. This would promote

63

the activities of private traders and vendors while at the same time ensuring that the poor have access to alternative market channels.

• Under each scenario, whether ADMARC is to withdraw from the proposed 204 markets or

not, critical measures would be required to enable ADMARC to effectively operate and fulfil the marketing functions. These measures comprise

o Privatising ADMARC’s subsidiaries and divesting its share ownership in commercial enterprises.

o Strengthening of ADMARC’s operational performance and financial viability through restructuring of management.

o Responding to the needs of the poor by opening markets early in the season and providing the inputs on time and repackaging inputs into smaller quantities where possible.

• Regardless of what will happen to the 204 ADMARC markets proposed for closure, measures

of an active market development would be required in future to improve the competition, market structure and regulation. This should comprise

o measures to improve market and price information, o market regulation and supervision functions, o improved access to credit and assets to encourage the private sector to engage in input

marketing and storage, and o promotion of producer associations or marketing cooperatives to increase the

bargaining power of rural producers.

64

Chapter 1 C1. Introduction C1.1 Background to the Study

Many developing countries have sought to liberalize the marketing of agricultural produce or inputs within the framework of structural adjustment programs. The justification has been that state marketing agencies crippled the operation of the market mechanism by restricting competition resulting in inefficiency of state marketing agencies and unrealistic prices to producers. Liberalization of agricultural markets has ranged from liberalization of marketing services to privatization of state marketing agencies. However, there has been considerable debate on whether the private marketing system can efficiently operate in developing countries, particularly in food marketing, in ensuring food security for the poor. There is evidence that the private marketing system is dominated by petty traders with substantial financial and capacity constraints and inter-seasonal and inter-regional arbitrage does not form part of the activities of this class of traders (Fafchamps and Gabre-Madhin, 2001).

Within the spirit of structural adjustment programs adopted by the Malawi government since 1981, reforming the agricultural sector which is the mainstay of the economy has been at the centre of adjustment policy (Chirwa, 1998). Agricultural sector reforms in Malawi began with the periodic upward adjustments in producer prices on major crops including maize and partial removal of subsidies on fertilizers. This was followed by restructuring of the state marketing agency, the Agricultural Development and Marketing Corporation (ADMARC), through rationalization and divestiture of its investment portfolio away from sectors unrelated to agricultural marketing activities since 1984. In 1987, the marketing of smallholder agricultural produce was liberalized through licensing of private traders, followed by removal of all licensing and registration procedures for private traders in produce and input markets by 1996.

In addition to the divestiture of its investment portfolio, ADMARC also restructured its marketing activities by closing down markets that were unable to attain an annual throughput of 60 tons. The first round of closure of such markets occurred in 1988, in which 125 seasonal markets (15 percent of the total) were closed based on their insignificance in the volume of trade. Unfortunately, this rationalization of markets affected markets that were largely in remote rural areas with poor infrastructure facilities (Chirwa, 1998). The expectation from the reforms was that private traders would take up the role of ADMARC even in remote areas where ADMARC pulled out, following the liberalisation of agricultural produce marketing in 1987.

However, many studies cast doubt on the capacity of private traders to reach the very remote areas that are deemed unprofitable by a state marketing agency. In earlier studies on performance of private traders in Malawi, Mkwezalamba (1989) and Kaluwa (1992) reveal that most private traders are small-scale entrepreneurs with rural-based enterprises and usually face constraints such as transport facilities, storage facilities, processing facilities, financing and credit facilities. Fafchamps and Gabre-Madhin (2001) also note that the average private trader in Malawi operates within a radius of 53 kilometres. Other recent studies have shown that while private traders have expanded their activities in agricultural produce since 1987, they are heavily concentrated in the Southern region and to some extent in the Central region while the Northern region is generally unserved due to the problem of inaccessibility (Mthindi et al., 1999). Fafchamps and Gabre-Madhin (2001) using case studies from Benin and Malawi also find evidence that the efficiency of private traders is constrained by the high transaction costs in form of search and transport costs. Others have argued that the continued existence of ADMARC as a state marketing agency is hindering the development of the private marketing system (Abbott and Poulin, 1996). Nonetheless, others identify ADMARC as a very important institution during crisis situations and that the closure of ADMARC markets has the potential to affect the livelihood systems of the poor (Khaila et al., 1999).

65

The privatization of ADMARC, one of the state-owned enterprises and a strategic institution in the food security area, has been a contentious issue in Malawi. Those that are sympathetic to the plight of poor farmers are calling for a cautious and well-informed approach while the pure capitalists or neoclassical economists interested in the commercial viability of ADMARC believe that the market will favourably take care of the poor smallholder farmers. The view from the business community is that

“There is need to conduct a survey to find out how viable all the ADMARC markets are in terms of business and assess how to compensate for markets that are not viable. Some markets may not be viable but government has a duty to provide services.” [Malawi Confederation of Chambers of Commerce and Industry President quoted in the Nation Business Review, 15 August 2002]

C1.2 State Intervention in Agricultural Produce Marketing in Malawi State Intervention and the Institutional Structure

State marketing of agricultural crops in Malawi dates from the colonial era with the establishment of the Native Tobacco Board in 1926, the Maize Control Board and the Cotton Control Board in 1949, the renaming of the Maize Control Board to the Produce Marketing Board in the early 1950s, and the incorporation of all the three boards into the Agricultural Production and Marketing Board in 1956 (Chirwa, 1998). The board was mandated to provide a stable and efficient marketing system for cash crops produced on trust land while ensuring that growers were paid a price that provided a reasonable rate of return. However, in 1957, government policy shifted to providing the minimum uniform pre-planting prices.

In 1962 the Agricultural Production and Marketing Board was superseded by the Farmers Marketing Board, which had more extensive functions including responsibility for marketing, processing and disposal of agricultural products; provision of adequate price stability; provision of storage facilities on behalf of the government and subsidization of agricultural inputs. The Agricultural Development and Marketing Corporation (ADMARC) was created in 1971 and replaced the Farmers Marketing Board. According to ADMARC (1974) the main objectives of ADMARC are:

• to promote the economic interests of Malawi by increasing the volume of exportable crops and improving the standards of agricultural production;

• to provide an efficient system for supplying the agricultural requirement of the farmers of

customary land;

• to purchase, for cash, the crops produced by these farmers, and to process them where necessary;

• to market agricultural produce for export, and to promote its consumption in Malawi and

abroad;

• to establish and develop agro-industry enterprises that fully exploited the agricultural potential of the country.

It is clear from the objectives of ADMARC that it had broadly two mandates: marketing of

agricultural produce and inputs and development of the smallholder agricultural sector through marketing activities and investments in agro-industry enterprises. To fulfill the marketing mandate ADMARC’s developed an extensive network and infrastructure of markets across the country. The market infrastructure included regional offices, divisional offices, storage depots, area offices, unit markets and seasonal markets. Unit markets have permanent structures such as storage facility and

66

office, with staff and ideally operate throughout the year and buy produce, sell produce and inputs. Seasonal markets are satellite markets of unit markets and have temporary structures and operate mainly in the harvesting seasons and are mainly used to purchase produce from farmers.

Figure C1.1 shows the organization of ADMARC’s marketing network. In 1990, ADMARC operated through 1,300 seasonal markets, 217 unit markets, 80 area offices (parent markets), 18 storage depots, 12 divisional offices (district headquarters) and 3 regional offices (ADMARC, 1990). However, as can be seen in Figure C1.1 there has been a sharp reduction in the number of marketing establishments. In 2001, ADMARC only operated through 441 seasonal markets, 343 unit markets, 24 parent markets, 10 depots and 14 district headquarters markets. In 2002, the regional office in the north and the centre were combined and the regional office in Lilongwe caters for both the central and northern regions.

ADMARC also pursued a social function which was reflected in the pricing system for smallholder farmers particularly maize and the establishment of markets in non-economic areas especially prior to structural adjustment programs. The pan-territorial and pan-seasonal pricing system that was implemented by ADMARC were a reflection of ADMARC’s de facto social function of providing maize at affordable prices to the poor and urban workers.

Apart from agricultural marketing activities, ADMARC also invested heavily in form of equity investments and loans in various enterprises in the economy and was directly involved in estate agriculture. According to Chirwa (2000), by the mid-1980s ADMARC had equity investments in 34 commercial enterprises and owned numerous estates. Harrigan (1991) argues that ADMARC used the surplus reaped in the trading activities to invest in industrial activities in various sectors of the economy.1

1 Kydd and Christiansen (1982) note that between 1971 and 1979 ADMARC extracted about MK181.9 million from the smallholder sector, of which 14 percent was used to cross-subsidize smallholder food production and consumption while the remainder was used for equity investments and loans to subsidiaries with only 4.3 percent of such investments related to the development of the smallholder agriculture sector.

67

Figure C1.1 ADMARC’s Marketing Establishments in 2001

Head Office [Limbe]

Regional Office – North [Mzuzu]

(1)

Regional Office – Centre [Lilongwe]

(1)

Regional Office – South [Limbe]

(1)

Divisional Offices / District

Headquarters (3)


Headquarters (5)


Headquarters (6)

Depots

(2) Depots

(3) Depots

(5)

Parent Markets (6)

Parent Markets (8)

Parent Markets (10)

Unit Markets

(69) Unit Markets

(146) Unit Markets

(128)

Seasonal Markets (110)



Note: Figures in parentheses are number of establishments as of November 2001

The Role and Performance of ADMARC

The 1970s were hay days for ADMARC like for many other state-owned enterprises during this period - a decade that recorded consistent positive economic growth. Scarborough (1990) asserts that until the late 1970s and early 1980s ADMARC provided vital services to smallholder farmers including assured and reliable markets for smallholder output, payment of farmers with cash on delivery, provision of credit and subsidized inputs, stabilization of prices seasonally and annually, payment of higher maize prices to producers than the export parity prices and provision of subsidized maize and rice to consumers.

However, ADMARC began experiencing operational problems beginning 1980, like many other state owned enterprises, following an economic crisis that led to negative growth rates of gross domestic product in 1980 and 1981.2 The deterioration of terms of trade adversely affected ADMARC’s profitability of the crop trading account. Scarborough (1990) argues that various external factors in combination with the illiquid nature of ADMARC’s investments, its work force and its operational inflexibility weakened its capacity to sustain the quality of its marketing activities. The major factors that substantially affected the operational efficiency of ADMARC include deteriorating

2 The economy was in a crisis due to several external factors including the intensification of the civil war in Mozambique that exacerbated the transport problems, deteriorating terms of trade, rising fuel costs, adverse weather conditions and weakening internal demand. The general decline in economic activities implied adversely on most sectors of the economy, including the state-owned enterprise sector.

68

prices for the country’s exports, change in government pricing policies, adverse weather conditions, a directive to establish and manage a strategic grain reserve, increasing external transport and finance costs (Chirwa, 1998).

Consequently, the financial position of ADMARC’s crop trading account began to deteriorate, resulting in the decline of crop purchases from smallholder farmers. The performance of the maize and rice trading accounts which were heavily cross-subsidized by cash crops even in the 1970s worsened particularly between 1979 and 1986 and the profitable cash crops started recording losses. Chirwa (1998) argues that the poor performance of the maize trading account was partly a manifestation of the maize pricing policy that reduced the gross margin from 57 percent in the 1967-79 period to 25 percent in the 1980-86 period, ADMARC increased involvement in portfolio investments, responsibility to manage strategic grain reserves and increased subsidies on inputs amid rising transport and handling costs. The Liberalization of State Marketing Activities

The problems that confronted ADMARC in the early 1980s led to the liberalization of the marketing of smallholder agricultural produce in 1987 through the licensing of private traders within the framework of structural adjustment programs. Chirwa (1998) argues that the liberalization of agricultural markets were largely due to the push factors following the adoption of structural adjustment programs in which the deterioration of state-owned enterprises finances, high subsidies via increasing budget deficits and inflexible government-determined pricing systems were identified as some of the structural factors leading to weakness in economic growth (Harrigan, 1991).3 The deregulation of agricultural markets was legislated in the Agriculture (General Purpose) Act of 1987 in which private traders were required to be licensed annually to operate in specific markets, maize exports were controlled, minimum producer prices were to be announced annually and ADMARC would buy at these guaranteed prices, and private traders were to submit monthly statements of their transactions. In effect, there were no significant barriers to entry in the marketing of agricultural produce in the post-1987 period and the licensing procedure was highly decentralized with each Agricultural Development Division (ADD) being responsible for issuing licenses.

The number of licensed traders increased from 387 in 1987/88 to 917 private traders in 1988/89 season and tampered off in 1989/90 season to 543 private traders (Kaluwa, 1991). Due to the capacity constraints in the Ministry of Agriculture and the lack of enforcement of licensing requirements by 1992 many traders never bothered to renew or/and obtain licenses and the private marketing system was completely unregulated. Consequently, by 1996 licensing was no longer a requirement for the marketing of smallholder agricultural crops. However, it is important to note that liberalization of marketing activities largely formalized small-scale private traders’ activities that were hitherto informal. The only significant new entries in agricultural produce and input marketing were the large-scale private traders (institutional companies) and manufacturers.

Several previous studies provide evidence that small-scale private traders existed prior to liberalization and accounted for a significant proportion of the marketed food surpluses. According to Mkwezalamba (1989), 50 percent of private traders started their business in agricultural produce marketing before liberalization of markets. Similarly, in a survey of private traders in Lilongwe and Blantyre ADDs, only 19 percent and 9 percent of private traders started operating after liberalization, respectively (Kaluwa, 1991). This underscores earlier observations that ADMARC accounted for a small proportion of the marketed maize output prior to liberalization (13 percent in 1975). According to Kandoole et al. (1988), traditionally most of the maize surplus was purchased by small-scale

3 Harrigan (1988) notes that the Malawi government was under heavy pressure from the World Bank, International Monetary Fund and other bilateral donors to implement reforms in the pricing and marketing of smallholder agriculture crops with emphasis on rapid commercialisation and diversification of smallholder agriculture.

69

private traders (licensed or otherwise) in various market places including farm-gate, local markets, district council markets and urban markets.

The liberalization of agricultural markets took place at the time when ADMARC’s pivotal role was already dwindling as shown in Figure C1.2. There is an increasing trend in the volume of maize trading between 1975 and 1984 with maize purchases being a larger share of maize trading. However, the decline in ADMARC maize trading activities started in 1985 and since then its role in maize trading has not been systematic. The volume of maize purchases and sales by ADMARC substantially declined in 1987 before picking up between 1988 and 1991. FSG (1991) attributes this sharp fall in maize trading after liberalization to ADMARC’s increasing financial problems and the declining producer incentives. Interestingly, one also observes ADMARC’s declining role in the purchase of maize since 1985, except in 1988-89, 1991 and 1993. It is also clear from the figure that ADMARC’s maize trading activity has been higher in the period after marketing liberalization than the period before liberalization. FSG (1991) also observe that ADMARC purchases as a proportion of production stabilized at 15 percent of production in the early 1990s, levels which were broadly in line with those prevailing in the early 1980s.

Figure C1.2 also shows ADMARC’s increasing role in the sale of maize since 1994 and its decreasing role in the purchase of maize. Recent data on the distribution of imported maize during the famine period show the importance of ADMARC in supplying maize to households in difficult times using its extensive market infrastructure network. According to NEC (2002b), between August 2001 and May 2002, ADMARC sold 73,960 metric tons of imported maize compared to 31,100 metric tons sold by private traders and 7,330 metric tons sold by the European Union (EU) and World Food Programme (WFP). Due to the underdeveloped nature of the private marketing system, the government suspended selling imported maize to private traders due to the exorbitant prices they were charging consumers (NEC, 2002b), a reflection of the inefficiency evident in the private marketing system in Malawi (Fafchamps and Gabre-Madhin, 2001). Figure C1.2 ADMARC Maize Sales and Purchases, 1970 – 2001 (Metric Tons)

0

200000

400000

600000

800000

1000000

70 75 80 85 90 95 00

Sales PurchasesYear

Met

ric T

ons

ADE (2000) also asserts the important role ADMARC has played in stabilizing the supply of maize, but argue that this stabilization effort has been done at a high cost. Another recent study on agricultural marketing (Nthara, 2002) also points to the important role ADMARC plays in food security particularly with respect to maize sales and input markets. The study finds that ADMARC’s role in purchasing agricultural produce from smallholder farmers is limited even in markets where it is still operational due its price uncompetitiveness, its late opening of markets in the season and its lack

70

of cash for most of the buying season. However, with respect to selling food commodities to food deficit households and selling inputs to smallholder farmers, Nthara (2002) finds that ADMARC still plays a critical role. There is also evidence that even in closed markets rural households use next alternative operational markets to purchase maize at lower prices than those charged by private traders, although at most ADMARC markets maize was out of stock resulting in households procuring maize from private traders at exorbitant prices.

Overall, the role of ADMARC in agricultural produce and marketing services has been declining. Nonetheless, this ought to be understood within the overall organizational framework and the structure of its investment portifolio. ADMARC began rationalizing its investment portfolio beginning 1984 and divested its ownership in commercial investments in which it was a minority share holder and sold all its estates, such that by 1998 it had shareholding only 8 enterprises with only two wholly owned subsidiaries (Chirwa, 2000). Unfortunately two major enterprises Grain and Milling Limited and Shire Bus Lines Limited which were partially privatized were sold back to ADMARC after failing even to perform under private ownership. Grain and Milling, Shire Bus Lines, Cold Storage and David Whitehead and Sons, the four sick babies of ADMARC, have been the greatest burden of ADMARC in the new millennium due to their persistent loss making records. Most government subsidies made to ADMARC were therefore used to bail out and sustain these loss making enterprises. Recently, ADMARC also moved back to direct farming activities, particularly production of tobacco. In 2001, ADMARC had 23 estates, 10 in the southern region, 7 in the central region and 6 in the northern region. The introduction of direct agricultural production has also shifted the focus of ADMARC from agricultural marketing to a production enterprise.

The deteriorating financial position of ADMARC has been a major concern of the government and the international financial institutions, and there is a lot of pressure to hasten the privatization of ADMARC. In early 2002, government transferred the four subsidiaries of ADMARC to Ministry of Finance as a stop-gap measure for eventual privatization. In the Malawi Poverty Reduction Strategy Paper (MPRSP) privatization of ADMARC is one of the strategies towards liberalization of agricultural markets and in the development of the agriculture sector (GoM, 2002a). This is reflected in the Malawi Government letter of intent and memorandum of economic policy to the International Monetary Fund according to which the commercialization of non-profitable agricultural outlets is planned to be completed by March 2003 (GoM, 2002b).

ADMARC is under intense pressure from the government to improve its financial position partly through adoption of cost cutting measures. Several proposals have been advanced on how ADMARC can improve its economic performance. According to ADMARC (2001) the high staff levels and high operational and administrative expenses make crop marketing uneconomical. One proposal is to restructure the state marketing agency and embark on cost cutting measures. The proposed restructuring programme has four elements, some of which have been implemented. First, ADMARC (2001) proposes to re-organise markets and regional structures based on those that existed in the 1960s and 1970s with ADMARC operating more seasonal than permanent markets. Secondly, ADMARC plans to close about 204 extraneous markets throughout the country or to operate them on seasonal basis. Thirdly, following the closure of the 204 markets the labour force will be reduced by more than 50 percent. Finally, the direct crop production (mainly tobacco) that was re-started in the 2000/2001 season is part of the survival strategy, justified on the basis that ADMARC no longer has a guaranteed market to satisfy its requirements.

Other proposals particularly from the government and other stakeholders include the privatization of the four subsidiaries of ADMARC that are persistently under performing and divestiture of ownership in other enterprises. The study by Abbott and Poulin (1996) recommends creation of separate companies for the trading of cotton and tobacco for eventual privatization, and also recommends the privatization of ADMARC’s extensive network of marketing infrastructure over a five year period. O&M Associates (2001) in a review of ADMARC’s marketing infrastructure recommend alternative uses of the permanent infrastructure such as conversion to social service provision units and leasing the units to private sector operators. These recommendations are subject to

71

substantial costs of alterations and it is even questionable how private sector operators may find business more profitable in remote areas to motivate them lease the units. Most of these recommendations from previous studies on ADMARC are made without due considerations to the social impact of such measures for rural households in the remote areas where private traders are not that active for competitive operations. A recent quantitative study based on consultations with smallholder farmers in selected markets cautions of the social consequences of wholesale closure of ADMARC unit markets (Nthara, 2002), particularly with regard to consumer markets for maize and input markets. C1.3 Scope of the Study

One of the proposed restructuring activities towards the privatization of ADMARC is further closure of seasonal and unit markets that are deemed ‘unprofitable’. Seasonal markets are temporary markets that are mainly used for buying produce from smallholder farmers and serve as satellite markets of unit markets. Unit markets have permanent infrastructure and staff and operate throughout the year offering all three services of purchasing and selling produce and selling inputs. ADMARC proposes to close another 204 ‘extraneous’ unit markets throughout the country and convert others into temporary markets. The market infrastructure in some of the closed unit markets is being proposed for privatization.

While previous studies have discussed the likely impact of closure of ADMARC markets, there has been no systematic study to understand the impact of such policies on the livelihood of affected rural households and understand the coping strategies and opportunities that may emerged following the eventual closure of unit markets. This Poverty and Social Impact Assessment (PSIA) study attempts to address the research gap, by qualitatively assessing the likely impact of the closure of some ADMARC unit markets on the poor. This qualitative study was intended as an ex-post assessment of the impact of closing some ADMARC markets and ADMARC’s declining business activities in recent years by soliciting households’ and stakeholders’ reactions before the implementation of such a policy. Thus, the PSIA study may be an important input for stakeholders in Malawi when designing and implementing the reform policies being currently formulated in the Malawi Poverty Reduction Strategy Paper (MPRSP). This study was commissioned jointly by the World Bank and GTZ and hosted in the National Economic Council. Purpose of Qualitative Research

The main objectives of the qualitative PSIA with respect to the closure of ADMARC markets in Malawi are:

(a) to learn from poor rural households and other main stakeholders about the effects of closing ADMARC markets on their well-being and economic activity;

(b) to explore people’s reactions to the closure of the markets including their adjustment and

coping strategies; and (c) to extrapolate the findings of the qualitative ex-post field research to the proposed closure of

further markets and draw draft policy implications regarding a socially responsible design, implementation arrangements and sequencing of the reform.

Research Questions

The following are the key questions, which guided this qualitative PSIA on closure of ADMARC markets:

(a) Who are the main stakeholders and institutions related to the closure of ADMARC markets in rural areas in Malawi and what are their interests, motivations and their potentials and whether any of these stakeholders are losers or winners of the reform?

72

(b) What are the perceived effects of closing ADMARC markets on the well-being and livelihood

strategies of different segments of rural households? (c) What are the patterns of reaction of rural households related to the closure of ADMARC

markets in their vicinity (for example, changes in coping mechanism, changes in livelihood, farming, input demand and marketing strategies of rural smallholders and households)?

(d) How do other main stakeholders such as private traders and shopkeepers react to or how are

they affected by the closure of ADMARC markets? (e) How does the closure of ADMARC markets influence the accessibility of maize and of inputs

in rural areas and their pricing? (f) Are rural households better off or worse off with the closure of ADMARC markets? (g) What solutions do rural households have on the issue of maize and input marketing?

C1.4 Organisation of the Report

This report is organized in six chapters. Chapter 2 outlines the methodology employed in the poverty and social impact assessment and the research process. The study mainly used qualitative techniques in collecting the data with special emphasis on focus group discussions and semi-structured interviews with key informants and institutions. The research process included two workshops: one for the development of study guide and the other on data analysis at the end of the field work involving principal researchers and some research assistants. Our research findings begin in Chapter 3, in which we provide a description of the study sites focusing on the physical features and socio-economic characteristics and the livelihood patterns, including rural households’ perceptions on definitions of well-being and then determine the trends in livelihoods in the context of the declining role of ADMARC and marketing liberalization. Chapter 4 discusses the current role of ADMARC and private traders, in agricultural marketing activities; identifying the market functions that have been taken up and the private traders’ role in famine situations. The functional review of ADMARC and private traders provides the basis for a comparative performance of these market institutions in terms of accessibility and availability, pricing and other business practices. In Chapter 5, we analyze the impact of closure of ADMARC markets or the declining role of ADMARC in crop purchase, crop sales and input sales on food security and examine whether such impacts have gender dimensions. Chapter 6 offers conclusions, lessons and recommendations for strategic design and implementation of agricultural marketing reforms in Malawi.

73

Chapter 2 Methodology and the Study Process

C2.1 Introduction

The study aims at understanding the consequences of closing some of ADMARC markets on the well-being and livelihoods of the primary beneficiaries of ADMARC services in the rural areas. Given the competing arguments for and against privatization of ADMARC activities, and government’s willingness to privatize seem to be irreversible in the long term, it is important to assess the likely consequences of some of the proposed strategies on producer and consumer welfare. However, as noted earlier, the proposed closure of ADMARC markets will take place against a background of ADMARC’s declining business due to its continued liquidity problems and the impact of marketing liberalisation, although the later does not seem to be substantiated empirically. It is therefore worthwhile to note that the perceptions on closure of markets can not be filtered out from the effects of privatization and other factors that are affecting ADMARC’s performance. This chapter details the methodology and the study process that was followed. It first gives the sites that were selected and the criteria used to select the sites. It then spells out the various components of the research and methods that were used. Finally, we elaborate on some of the limitations of the study that warrant some cautious interpretation of the qualitative research results. C2.2 Selection of Field Sites

The unit for sampling sites ADMARC’s unit markets categorized into two groups: closed markets and active markets. A purposive sample of about 10 locations affected by ADMARC market closure across the country were selected to match the pattern of the 204 ADMARC markets lined up for closure (6 sites were places where ADMARC used to operate but has since ceased, while 4 were sites where ADMARC was still operating). The sample of unit markets was stratified by geographic distribution, the period after closure and the livelihood systems. To this effect 3 markets were sampled from the northern region, 4 from the central region and 3 from the southern region. However, all 10 selected unit markets were found to be operational during the study and for those unit markets that were selected on the basis of closure the study team interviewed communities in closed seasonal markets. It is important to note that the sample of closed markets in this study imply closed seasonal markets in the catchment area of the selected market. Table C2.1 below shows the sites that were selected for interviews with smallholder farmers, private traders and other key informants. Table C2.1 Sites Selected for the Study Site District Region Villages Status Nthalire (Kamphyongo seasonal) unit market

Chitipa North Kamphyongo Mutale

Closed

Mpata unit market Karonga North Mwakikwinga Luweni

Open

Eswazini (Emoneni seasonal) unit market

Mzimba North Mateyu Ng’oma Mholonondo

Closed

Chikho (Mzandu seasonal) unit market

Ntchisi Centre Mzandu Ngwewa

Closed

Lipiri (Kamtepa seasonal) unit market

Dowa Centre Kamtepa Nyengere

Closed

Mikundi unit market Mchinji Centre Laisi Nzama

Open

Chiondo unit market Nkhotakota Centre Kajaliza Matchipisa

Open

Kasongo unit market Phalombe South Kasongo Mkhwayi

Open

Lundu (Mkumba seasonal) unit market

Blantyre South Mkumba Jorodani

Closed

74

Kakoma (Tombondera seasonal market)

Chikwawa South Tombondera Nsaliva

Closed

Source: Poverty and Social Impact Assessment 2002 C2.3 Data Collection Methods

The purpose of the field research was to contextualise the impact of policy by looking more for specific explanations, attributions, dynamics and perceived options of behavior and risks from the context of specific social groups than for statistical representative information. For this purpose a range of qualitative research methods were used, including selected participatory learning methods. Focus Group Discussions

The selection of interviewees and discussion partners accounted for differences with respect to social and economic stratification, ethnic and livelihood diversity and gender and this was done in consultation with local leaders and local extension staff. Focus Group Discussions (FGDs) using variants of Participatory Learning Methods were the main method for collecting data. The discussions focused on perceptions on well-being and trends in the last five years, livelihood strategies and cropping patterns, the problem analysis, cause-effects diagrams, trend analysis, institutional analysis and analysis of opportunities, coping and survival strategies. Key Informants/Semi-Structured Interviews

Interviews with key informants in the selected sites were also conducted. Different stakeholders existing or operating in the areas such as traditional leaders, private traders, shop-keepers, extension agents, food security, and non-governmental organizations. In each location at least 12 key informants’ interviews including key informants in the surroundings of the sites, at least one in each well-being group identified by the community in focus group discussions. Institutional Interviews

The survey team also conducted institutional interviews at policy or decision making levels in stakeholder institutions. The institutions and stakeholders visited included ADMARC management, National Association of Smallholder Farmers (NASFAM), National Economic Council, Grain and Milling Limited, Smallholder Farmer Fertilizer Revolving Fund (SFFRF), World Food Programme (WFP), Bharat Trading Limited, Ministry of Agriculture and Livestock Development. The purpose of the institutional interviews was to assess the objectives, motivations, interests, concerns and resistance of different stakeholders and institutions to the closure of ADMARC markets. Desk Research

In addition to the primary data that were collected, the study team also conducted desk research on food marketing, particularly on institutional analysis from several previous studies on the operations of private traders and on perceptions on the relative importance of ADMARC from previous qualitative studies such as the Qualitative Impact Monitoring (QIM). C2.4 The Study Process Training Workshop

The research was conducted by researchers from Wadonda Consult and their assistants. The group underwent training for two weeks under the guidance of colleagues from World Bank and GTZ (respectively), based in Washington. In attendance at this work workshop were the National Economic Council official and a senior officer from ADMARC headquarters. The training workshop

75

took a week and half from 18th March 2002 to the 28th of March 2002. The rest of the period was devoted to pre-testing and refining the instruments. On 4th April 2002 the team was deployed for data collection, which lasted until 27th May 2002. After putting together the site reports, the team with the GTZ and NEC members that were involved during training held a preliminary analysis workshop that has guided the writing of this report. Data Collection

Field work commenced from the first week of April 2002 through to the end of June 2002 involving Focus Group Discussions (FGD), Key Informants Interviews and Semi-Structured Interviews (SSI) in the 10 sample sites and selected institutional interviews. The FGD was the main method for data collection. In each market location, 4 FGDs were conducted two for women and two for men. Based on the information from the focus group discussions individuals were identified from each well-being category for case studies or semi-structured interviews. Interviews were also conducted with key informants in all the selected sites. These included different stakeholders existing or operating in the areas such as traditional leaders, private traders, shopkeepers and extension agents. Table C2.2 shows the activities that were undertaken in each site and the number of each activity and Annex 1 presents a list of key informants consulted in institutional interviews. In total, 40 focus group discussions were conducted, 54 semi-structured interviews were done and 44 key informant interviews were conducted. Table C2.2 Activities Conducted During the Fieldwork

Key Informant Interviews Site (District)

FGD

SSI ADMARC Private

Traders Other

Ntharile (Chitipa) 4 5 0 0 2 Mpata (Karonga) 4 5 1 0 2 Eswazini (Mzimba) 4 6 1 0 3 Chikho (Ntchisi) 4 5 1 0 2 Lipiri (Dowa) 4 6 1 1 3 Mikundi (Mchinji) 4 6 1 0 3 Chiondo (Nkhotakota) 4 4 1 0 3 Kasongo (Phalombe) 4 5 1 1 4 Lundu (Blantyre) 4 6 1 1 6 Kakoma (Chikwawa) 4 6 1 0 3 Total 40 54 9 3 32 Source: Poverty and Social Impact Assessment 2002 Data Analysis and Report Writing

Field work was followed by a four-day data analysis workshop in Zomba in the first week of July 2002. The workshop was attended by the research team (principal researchers and research assistants) and colleagues from GTZ-World Bank and National Economic Council. The purpose of the data analysis workshop was to generate main research findings and tentative recommendations that would guide the research team in data analysis and report writing. C2.5 Limitations of the Study

This study used a qualitative approach to assessing the implications for the closure of ADMARC markets or its declining business within the framework of agricultural marketing liberalisation in Malawi. Naturally, the study only deals with specific issues and is limited in its coverage of the sample markets and issues that are equally important but beyond the scope of this study. It is important therefore that the results of the study should be interpreted with caution subject to the following limitations.

76

First, the study was undertaken during the famine period in Malawi and some of the views from smallholder farmers could be influenced by the prevailing conditions. Second, although the ADMARC list used for sampling indicated closed unit markets, the study team found that all the selected unit markets (closed and active) were active. For ‘closed’ unit markets, the study team therefore decided to hold interviews in villages where seasonal markets were closed. Third, the effects of closure of markets does not only reflect closure of markets, since seasonal markets have been closed during the period of agricultural marketing liberalisation and during the time ADMARC business has been declining partly due to its liquidity problems. Thus, the interpretation of the effects of closure encompasses the other trends in the post-marketing liberalisation period.

Fourth, the study team was unable to conduct more interviews with private traders in the sampled villages or surrounding villages due to the high mobility of private traders. Most of the private traders were not resident in the villages where the interviews were conducted as they operate mobile markets and in multiple markets. It was therefore difficult to do gender analysis of private traders in this study.

Fifth, the study is weak in the gender analysis particularly on the socio-economic effects of closure of markets or declining ADMARC business. We, however, conducted focus group discussions with men and women separately, and where possible we only report whether there were differences on the responses and perceptions from the gender groups.

Sixth, it was also difficult to secure appointments with various institutions because of commitments of the officers. For instance, during the time the team visited Lilongwe most officers in the Ministry of Finance were busy with the budget session of Parliament and given the limited resource constraints it was difficult to arrange for the second visits.

Finally, the study focuses on ten sites of the many markets ADMARC has throughout the country. The sample is not nationally representative and therefore the results can not be generalized as characterising the operations of the agricultural marketing systems in Malawi. This study is a first attempt to capture the likely effects of a policy instrument prior to its implementation through a qualitative study.

77

Chapter 3 Description Of Study Sites And Perceptions On Well-Being

C3.1 Introduction

This chapter begins to present the findings of the study and forms the basis of analysis in subsequent chapters. In this chapter we provide a brief description of the study sites and analyze the livelihoods patterns and diversification and perceptions on well-being or ill-being. The next section provides a general description of study sites focusing on the remoteness, infrastructure development and livelihood patterns. Section C3.3 is devoted to the analysis of well-being, particularly communities own understanding of poverty, the critical factors that form the classification of poverty, the changes that have taken place in well-being over the last five years and factors that have initiated such changes. The concept and classification of well-being is important in understanding the consequences of policies in rural Malawi, as different well-being groups may react or may be affected differently by changes in government policy. Section C3.4 provides concluding remarks. C3.2 Description of the Study Sites

In this section a brief description of the study sites is given paying particular attention to the infrastructure that is available in the study sites and the economic activities that take place in the ten sites. The section also describes the main sources of livelihoods in these areas. Physical Characteristics and Infrastructure

All the ten sites were characteristically rural with very poor road infrastructure and therefore difficult to access particularly in the rainy season except for the Eswazini site in Mzimba. In terms of access to other social services, all the sites (except three) had at least a primary school located within the studied communities whereas in the three sites (one in each region), the schools were less than two kilometres away. As expected, health facilities, police posts and mail post offices were very far from most of the studied communities with the Phalombe site being 27km away from the nearest facilities whereas the rest were between five and fifteen kilometres away.

Regarding their agricultural patterns, the ten sites were not very different from each other in their food crops as they all produced maize as their main staple as well as producing other food crops such as groundnuts, cassava and sweet potatoes. However, considerable differences appeared regarding their cash crops with three of the sites producing cotton (the Kakoma (Chikwawa) site, Chiondo (Nkhotakota) site and Mpata (Karonga) site) and the rest being more involved in tobacco farming even though some households were also producing other cash crops such as soy beans and paprika. Annex 2 provides detailed description of the study sites regarding their infrastructure and livelihood patterns.

78

Table C3.1 Accessibility and Condition of Available Infrastructure Unit Market

(District)

Roads

School ADMARC

Market Health Facility

Post Office Police Post

Nthalire (Chitipa)

Difficult to access

In village 3km 15km 14km 15km

Mpata (Karonga)

Fairly accessible

2km 15km 15km 15km 15km

Eswazini (Mzimba)

Accessible In village 11km 11km 11km 62km

Chikho (Ntchisi)

Difficult to access

In village In village 1km 34km 34km

Lipiri (Dowa)

Difficult to access


Mikundi (Mchinji)

Fairly accessible


Chiondo (Nkhotakota)

Fairly accessible


Kasongo (Phalombe)

Difficult to access


Lundu (Blantyre)

Difficult to access


Kakoma (Chikwawa)

Fairly accessible


Source: Poverty and Social Impact Assessment 2002 Livelihoods Patterns and Diversification

Livelihoods systems in the ten areas were looked into through an analysis of the economic activities that preoccupy people in the villages. Table C3.2 shows the economic activities in the ten sites. Cropping Pattern

From Table C3.2 it is evident that in five sites (Nthalire, Mpata, Eswazini, Chikho and Lupiri) there is little crop diversification compared to the other five sites. In these sites, maize is the predominant food crop and with the exception of Mpata which has cotton for cash crop, the other four sites have tobacco as the main cash crop. The other five sites i.e. Mikundi, Chiondo, Kasongo, Lundu and Kakoma have a lot more crop diversification as seen from the number of crops that are grown in the areas. Maize is still the predominant food crop in these areas but there are a lot of the draught resistance crops like cassava and sorghum being grown in these areas. Chiondo and Kakoma grow cotton for cash while Mikundi, Kasongo and Lundu grow tobacco as a cash crop.

79

Table C3.2 Economic Activities in the Area Site (District) Crops grown Livestock Non-farm activities

Nthalire (Chitipa)

Maize, beans, paprika, tobacco, groundnuts, soya.

Cattle, goats, chickens, guinea fowl.

Trading in farm and non-farm products at day markets, maize vending and cross border trade, ganyu.

Mpata (Karonga)

Maize, groundnuts, cotton and rice.

Cattle, goats, chicken and guinea fowl.

Trading with people from Tanzania, non-farm ganyu, handcrafts, maize vending..

Eswazini (Mzimba)

Maize, groundnuts, oriental tobacco, soya beans and paprika.

Cattle, goats, chickens and pigeons.

Off-farm and on-farm ganyu, grocery shops, handcrafts, maize vending.

Chikho (Ntchisi)

Tobacco, groundnuts, maize, soya beans, sweet potatoes.

Sheep, goats, cattle, chickens, ducks, hare, pigeons, guinea fowl.

Selling foodstuffs, grocery shops, fish mongering, and maize vending.

Lipiri (Dowa)

Tobacco, maize, groundnuts, sweet potatoes, soya beans

Cattle, chickens, pigs, goats, sheep, ducks

Grocery shops, ganyu, land letting, maize vending, selling firewood, sawing timber and beer brewing.

Mikundi (Mchinji)

Tobacco, maize, soya beans, cassava, groundnuts, sweet potatoes, vegetables, rice, beans, paprika.

Pigs, chickens, goats, cattle

Small-scale business (hawkers), ganyu, felling people on vehicles, fishing, shops, maize vending, letting land and cross border trade.


Tobacco, maize, groundnuts, cotton, rice, sorghum, sweet potatoes, cassava, millet, vegetables, soya beans.

Goats, chicken, ducks, guinea fowl, cattle, hare.

Shops, maize mills, maize vending, ganyu, fish mongering and carpentry.

Kasongo (Phalombe)

Maize, tobacco, sunflower, rice, sweet potato, cassava, cow peas, groundnuts, chick peas, grams, sorghum, pigeon peas, mucun.

Cattle, goats, guinea fowl, turkey, chicken, pigs.

Ganyu, maize vending, Groceries, butcheries, fish mongering, fishing, beer brewing and selling.

Lundu (Blantyre)

Maize, rice, sweet potato, groundnuts, tobacco, pigeon peas, vegetables, rice, grams, paprika, sorghum, millet, ground beans, cotton, bananas, sugar cane.

Cattle, goat, guinea fowl, turkey, chickens.

Ganyu, maize vending, business like tailoring, fish mongering, selling foodstuff, shops, running hawkers and selling charcoal.

Kakoma (Chikwawa)

Maize, cotton, rice, chitowe, sweet potatoes, groundnuts, sorghum, millet, cassava, pigeon peas, beans, mucun, ground beans, fruits and vegetables, sugarcane.

Goats, cattle, pigs, chicken, guinea fowl, ducks.

Maize vending, farm and off-farm ganyu, retail shops, selling of crop produce.

Source: Poverty and Social Impact Assessment 2002 Livestock

Own crop production or crop purchases tell but only part of the story of people's sources of livelihoods. Livestock is another important part of the livelihoods equation. In all the 10 ten sites

80

(Table C3.2) there is the presence of both big and small animals. Nthalire, Mpata, Eswazini, Lipiri, Mikundi and Kakoma have more cattle than the other four sites. The most common livestock though was the chicken. Pigs were found at Kakoma, Kasongo, Mikundi and Lipiri. Sheep on the other hand were found at Lipiri and Chikho. In general though, despite the fact that most of the households kept small animals and birds, there was a high incidence of keeping livestock in all the sites. Non-Farm Activities

Another side of the livelihoods story that this research looked at are the off-farm activities that households engaged in. Table C3.2 summarises the non-farm economic activities that some households in the ten sites had recourse to. Vending in maize and its products seems to be the most common activity. Closely associated with this is the trading in farm and non -farm products (small-scale businesses). Examples of the small-scale businesses included handcrafts in Mpata and Eswazini, fish mongering in Kasongo, Lundu, Chikho and Chiondo, and beer brewing and selling in Kasongo and Lipiri. For sites in districts bordering other countries like Nthalire (Zambia), Mpata (Tanzania), Mikundi (Zambia) and Kasongo (Mozambique) there is a proliferation of cross border trade. This trade involves exchanging farm products with manufactured wares from across and visa versa. The production and selling of charcoal was only reported at Lundu.

The selling of one’s labour in exchange for cash or in kind payment (ganyu) is another activity that happens in all the sites. The letting out of a household's productive piece of land was reported only at Lipiri and Mikundi. These are places where land is still relatively in abundance. The letting out of land could signal stress on the part of the household in that they did not have enough resources to cultivate all their land. In some cases it was found that the households did not have enough resources (labour or/and cash) for their livelihoods. Hence, they were forced into letting out some of their productive assets. C3.3 Perceptions on Well-being

This section provides an overall picture of what it means to belong to a particular well-being classes as discussed by some participants from the sampled communities. The section also discusses how and why the proportions of households in each category have been fluctuating in the past decade and the role that ADMARC or access to markets has had in their livelihoods with particular reference to access to inputs, produce and food markets. Classification of Well-being

The communities in the sample sites discussed the concept of well-being and how they differentiate well-being groups. In the process three main groups were identified using various attributes. The first group of well-being is the ‘the haves/ well-to-do’ or the rich locally known as opeza bwino,wochita bwino, odya bwino, wasambazgi, wakusanga or olemera. The second group is that of the ‘the ‘a bit-well-to-do’ or moderately well-off known as wopezako bwino pang’ono, odya bwino pang’ono, wakusangako pachoko. The third group is that of the ‘the have-nots/poor’ known as osauka, ovutika or wakusuzgika. In Kakoma and Mikundi site the poor the communities also differentiated the poor from the poorest (osaukitsi or ovutikitsitsa). These groups are similar to those found in the QIM 2 study (NEC, 2002a). In the study we use three categories of well-being: ‘the haves’ or the rich, ‘the ‘a bit-well-to-do’ or the moderately rich and ‘the have-nots’ or the poor. Several distinguishing features were discussed and these have been clustered into eight categories namely: level of food security and self-sufficiency, farming systems and characteristics, housing and household characteristics, motivation to grow crops, asset ownership, participation in labour market and off-farm activities, agricultural marketing behaviour and participation in the financial market (Table C3.3).

81

Level of Food Self-sufficiency and Food Security

Household’s level of food self-sufficiency and food security is a very good well-being classifier and this criterion is increasingly being used in poverty assessment studies as it measures the position of households on the basis of their access to one of the most basic human rights: ‘the right to adequate and nutritious food’. Three indicators [namely: length of maize self-sufficiency, meal frequency, dietary diversification and capability to procure maize (food) from the market] were commonly discussed in the ten sites as being different across well-being classes. In terms of food self sufficiency, ‘the haves’ produce enough food to feed their household members throughout the year (they are self-sufficient). Their meal pattern does not vary across the year (at least, three meals in a day) and they also eat meat and other animal products more frequently compared to households belonging to the other well-being classes. In times of food shortage, such as when there is drought or floods, these households afford to procure maize from the market and even from private traders. Due to their sound asset base, these households are able to cushion the effects of such droughts/floods through selling or pawning some of their assets (such as livestock, chairs, beds etc.).

On the contrary, ‘the have-nots’ are perpetually food insecure, sometimes they deplete their maize before harvesting or within 1-2 months of harvesting and since they cannot afford to buy meat and other animal proteins, they eat a lot of vegetables (except for the cheap small dry fish). Whereas the well-to-do eat vegetables as a matter of dietary diversification, the ‘have-nots’ or the poor consume a lot of vegetables because they cannot afford to buy meat or other animal products more frequently. With the exception of the harvesting period, the ‘have-nots’ usually eat one meal in a day. Ganyu work provides a very good survival strategy for these households. Because the ‘have-nots’/poor deplete their maize stocks early, they are the main buyers of maize from the market (such as ADMARC). Although they purchase in small quantities, they collectively purchase a large proportion of the maize that is offered for sale because they form a high proportion of rural households, the high frequency of visiting the market and the length of period without own food in any given year. ADMARC is preferred by the ‘have-nots’/poor because they are able to buy in small quantities and at a lower price compared to what is offered by the private traders. The ‘a-bit-well-to-do’ group falls in between the two groups. These households are partially food self-sufficient in that they usually deplete their maize from own production within 4-6 months of harvesting as opposed to within 8-12 months among the ‘haves’ and 1-2 months among the ‘have-nots’.

In terms of defining well-being classes regarding length of maize self-sufficiency, meal frequency and dietary diversification, the findings from the ten sites provided similar criteria for the three well-being classes. However, a difference was emerging when poor households run out of their maize stocks in that the poor in places where ADMARC was still operating indicated to have relatively more access to maize at a competitively low price than their counterparts who indicated that they walked long distances to the nearest ADMARC or to flea markets in search of maize. The only exception to this observation was last year when ADMARC sold maize irregularly.

82

Tab

le C

3.3

Dis

tingu

ishi

ng C

hara

cter

istic

s am

ong

Diff

eren

t Wel

l-bei

ng C

lass

es in

the

Sam

pled

Com

mun

ities

C

hara

cter

istic

s

‘The

hav

es o

r th

e w

ell-t

o-do

’

‘A b

it-w

ell-t

o-do

’

‘the

hav

e-no

ts, t

he p

oor’

Le

vel o

f foo

d se

curi

ty a

nd

cons

umpt

ion

patt

erns

• Th

eir

food

sto

cks

last

thr

ough

out

the

year

or

at l

east

up

to

8-9

mon

ths

from

har

vest

•

Eat a

t lea

st th

ree

mea

ls in

a d

ay

• A

ble

to b

uy f

ood

(mai

ze)

in l

ean

harv

ests

e.g

. dur

ing

envi

ronm

enta

l stre

ss

• A

ble

to b

uy a

nim

al p

rote

in o

r to

sl

augh

ter t

heir

lives

tock

•

Som

etim

es b

uy c

heap

mai

ze f

rom

A

DM

AR

C

to

pay

labo

urer

s/

gany

u (p

iece

wor

k) in

kin

d

• Th

eir

food

sto

cks

from

ow

n pr

oduc

tion

last

ing

4-6

mon

ths

from

ha

rves

t •

Eat 2

-3 m

eals

in a

day

•

Dep

end

on A

DM

AR

C t

o bu

y ch

eap

food

and

in sm

all q

uant

ities

•

Dep

end

on

gany

u (p

iece

w

ork)

to

buy

food

or

be p

aid

with

fo

od

• A

ble

to b

uy a

nim

al p

rote

in o

r to

slau

ghte

r the

ir liv

esto

ck

• O

wn

prod

uced

mai

ze la

stin

g in

the

field

or

1-2

mon

ths f

rom

har

vest

•

Eat f

lour

mad

e fr

om m

aize

hus

ks/b

ran

• Ea

ting

one

mai

n m

eal i

n a

day

• D

epen

d on

AD

MA

RC

to b

uy c

heap

food

and

in

smal

l qua

ntiti

es

• C

ore

buye

rs o

f AD

MA

RC

’s m

aize

•

May

buy

foo

d fr

om n

eigh

bour

s or

paw

n w

ith

asse

ts

• D

epen

d on

gan

yu (

piec

e w

ork)

to

buy

food

or

be p

aid

with

food

•

Eat

a lo

t of

ve

geta

bles

an

d pl

ant

prot

eins

(b

eans

) or s

mal

l dry

fish

•

Som

e ar

e m

alno

uris

hed

and

freq

uent

ly il

l Fa

rmin

g sy

stem

s an

d ch

arac

teri

stic

s •

Prac

tice

mon

ocro

ppin

g be

caus

e th

ey h

ave

labo

ur a

nd fe

rtiliz

er

• C

ultiv

ate

bigg

er l

and

(>3

acre

s),

som

e of

whi

ch m

ay b

e re

nted

•

Use

an

d ca

n af

ford

in

orga

nic

ferti

lizer

s in

larg

e qu

antit

ies

• Pr

oduc

e m

ore

of h

igh

valu

ed c

ash

crop

s suc

h as

toba

cco

and

cotto

n

• Pr

actic

e in

terc

ropp

ing

to

redu

ce la

bour

cos

ts

• C

ultiv

ate

1-2

acre

s of l

and

• So

me

may

pro

cure

inpu

ts (

up

to th

ree

50kg

bag

s max

imum

) •

Gro

w t

obac

co,

cotto

n bu

t at

lo

wer

leve

l tha

n ‘th

e ha

ves’

• Pr

actic

e in

ter-

crop

ping

to re

duce

labo

ur c

osts

•

Cul

tivat

e on

sm

all

land

of

less

tha

n 1

acre

(th

ough

som

e ha

ve b

igge

r par

cels

) •

Una

ble

to te

nder

/car

e fo

r cro

ps

• M

ost

can’

t af

ford

in

puts

or

bu

y in

lo

ose

quan

titie

s

Hou

sing

and

hous

ehol

d ch

arac

teri

stic

s

• D

ecen

t ho

uses

roo

fed

with

iro

n sh

eets

•

Hav

e ac

cess

to s

afe

wat

er (

or a

ble

to tr

eat)

• H

ave

dece

nt l

atrin

es,

dish

rac

k,

kitc

hen,

bat

hroo

m a

nd w

ell

vent

ilate

d ho

uses

•

Peac

e of

min

d •

Abl

e to

pay

seco

ndar

y sc

hool

fees

• So

me

have

dec

ent h

ouse

s w

ith

iron

shee

ts

• H

ave

acce

ss to

safe

wat

er

• So

me

have

de

cent

la

trine

s, di

sh r

ack,

kitc

hen,

bat

hroo

m a

nd

wel

l ven

tilat

ed h

ouse

s •

Abl

e to

pay

sec

onda

ry s

choo

l fe

es

• H

ave

poor

cl

othe

s an

d be

ddin

gs

• H

ave

smal

l, sh

abby

loo

king

dw

ellin

gs,

gras

s th

atch

ed th

at m

ay le

ak d

urin

g ra

ins

• M

ay h

ave

acce

ss t

o sa

fe w

ater

but

mos

t ju

st

drin

k w

hate

ver i

s ava

ilabl

e •

No

latri

nes

or d

ilapi

date

d, c

ook

outs

ide,

poo

r ve

ntila

tion

• Fa

il to

pay

seco

ndar

y sc

hool

fees

•

Diff

icul

ties t

o bu

y so

ap a

nd n

ew c

loth

es

• H

ave

poor

bed

ding

s

83

Cha

ract

erist

ics

‘T

he h

aves

or

the

wel

l-to-

do’

‘A

bit-

wel

l-to-

do’

‘t

he h

ave-

nots

, the

poo

r’

Mot

ivat

ion

to

grow

som

e cr

ops

e.g.

mai

ze,

toba

cco,

cot

ton

• M

aize

for

foo

d as

wel

l as

pay

ing

gany

u w

orke

rs

• To

bacc

o an

d co

tton

for i

ncom

e •

Legu

mes

fo

r fo

od,

mon

ey

and

soil

impr

ovem

ent

• Se

eing

th

e w

ell-t

o-do

as

‘m

odel

s’

and

tryin

g to

imita

te th

em

• M

aize

for f

ood

• So

me

toba

cco/

cotto

n fo

r mon

ey

• M

aize

for s

ubsi

sten

ce

• So

me

toba

cco/

cotto

n ju

st f

or ‘

trial

and

er

ror’

and

pet

ty c

ash

– a

few

ridg

es

Ass

et o

wne

rshi

p •

Poss

ess

mor

e hi

gh v

alue

liv

esto

ck l

ike

cattl

e, p

igs a

nd m

ore

of th

e ot

her

• Po

sses

s oxe

n-dr

awn

carts

• H

ave

smal

l nu

mbe

rs

of

lives

tock

m

ainl

y go

ats,

duck

s, ch

icke

ns

• H

ave

acce

ss to

ani

mal

man

ure

• H

ire o

xcar

ts

• So

me

poss

ess

smal

l nu

mbe

rs

of

lives

tock

whi

ch th

ey se

ll-of

f qui

ckly

•

Use

hea

ds to

tran

spor

t pro

duce

Part

icip

atio

n in

th

e la

bour

mar

ket

and

off-

farm

ac

tiviti

es

• Em

ploy

ers o

f the

poo

r and

the

‘a b

it w

ell-

to-d

o’ o

n th

eir f

arm

s •

Do

not

parti

cipa

te i

n ca

sual

or

gany

u em

ploy

men

t •

Self-

empl

oyed

th

roug

h sm

all

busi

ness

pr

oprie

tors

hip

• R

un s

mal

l bu

sine

sses

lik

e fis

h ve

ndin

g,

tailo

ring,

buy

ing

and

selli

ng p

rodu

ce

• So

me

parti

cipa

te

as

casu

al

labo

ur

empl

oyee

s on

far

ms

or b

usin

esse

s of

the

‘w

ell-t

o-do

’ to

a le

sser

ext

ent

• Se

lf em

ploy

ed

to

a le

sser

ex

tent

th

roug

h sm

all b

usin

ess p

ropr

ieto

rshi

p •

Som

e ru

n bu

sine

sses

but

at

a lo

wer

le

vel

• M

ost p

artic

ipat

e in

the

labo

ur m

arke

t as

casu

al l

abou

rers

or

gany

u w

orke

rs o

n fa

rms

or b

usin

esse

s of

the

‘wel

l-to-

do’ a

nd s

ome

of

the

‘a b

it-w

ell-o

-do’

•

Dep

end

on g

anyu

to

get

food

or

som

e in

puts

•

Rar

ely

in

self-

empl

oym

ent

thro

ugh

busi

ness

pro

prie

tors

hip

• N

o bu

sine

ss o

r se

ll fir

ewoo

d as

a p

etty

bu

sine

ss

Agr

icul

tura

l m

arke

ting

beha

viou

r

• A

ble

to p

rocu

re in

puts

(in

bulk

qua

ntiti

es)

and

food

from

priv

ate

trade

rs a

nd A

DM

AR

C

• Se

ll th

eir

prod

uce

to p

rivat

e tra

ders

and

ot

her d

ista

nt m

arke

ts su

ch a

s auc

tion

floor

s •

Buy

cas

h cr

ops

from

the

‘a-

bit-w

ell-t

o-do

’ and

the

poor

• Se

ll th

eir

mai

ze a

nd o

ther

pro

duce

ea

rly t

o pr

ivat

e tra

ders

, th

e ‘w

ell-t

o-do

’ an

d to

a

less

er

exte

nt

AD

MA

RC

pa

rticu

larly

late

r in

the

mar

ketin

g se

ason

•

Buy

mai

ze f

rom

AD

MA

RC

in

lean

m

onth

s •

Sell

thei

r cas

h cr

ops t

o pr

ivat

e tra

ders

or

the

‘wel

l-to-

do’

• Se

ll th

eir m

aize

and

oth

er p

rodu

ce e

arly

to

priv

ate

trade

rs,

the

‘wel

l-to-

do’

and

sell

rem

aini

ng A

DM

AR

C l

ater

in

the

mar

ketin

g se

ason

•

Cor

e bu

yers

of

AD

MA

RC

mai

ze i

n sm

all q

uant

ities

Part

icip

atio

n in

fin

anci

al m

arke

t

• H

ave

acce

ss to

indi

vidu

al c

redi

t •

Act

as

mon

ey l

ende

rs t

o th

e ‘a

bit-

wel

l-to

-do’

and

the

poor

• So

me

have

acc

ess

to i

ndiv

idua

l or

gr

oup-

base

d cr

edit

• La

ck a

cces

s to

form

al c

redi

t fac

ilitie

s •

Usu

ally

ex

clud

ed

from

cr

edit

and

exte

nsio

n

So

urce

: Pov

erty

and

Soc

ial I

mpa

ct A

sses

smen

t 200

2

84

Farming Systems and Characteristics

The ‘haves’ also cultivate on considerably bigger land (mostly more than three acres) some of which may be rented from the ‘have-nots’/poor or from other sources and they possess high valued assets such as ox-carts and large numbers of high valued livestock such as cattle and goats (for example, over 10 herds of cattle). The poor also spend more time on ganyu, they cannot afford inputs such as seed and fertiliser, and end up cultivating less land. The poor also have labour allocation dilemmas (perceived immediate opportunity costs) between own farming and ganyu work. The majority of the ‘have-nots’/poor households though, have less than one hectare of land and they also usually fail to pay for secondary school fees for their children if they happen to proceed with their education.

In terms of cropping patterns, all well-being classes grow similar crops but what differs is that the ‘well-to-do’ households apply fertilizer to their crops and they also cultivate larger parcels of land compared to their counterparts who may cultivate tobacco or cotton on countable ridges. The other striking difference among the three well-being categories is that the ‘haves/well-to-do’ households are role models for the other categories but the ‘a bit-well-to-do’ are more optimistic about their future. As such, they try several initiatives to catch up with the well-to-do (including running businesses) whereas the ‘have-nots’/poor appear to be more pessimistic and resigned although some efforts are taken to adopt better ways of farming. On the other hand, most of the ‘have-nots’/poor, rarely use inputs because they cannot afford inputs, and those who can afford buy in loose quantities, sometimes they obtain inputs through ganyu. Housing and Household Characteristics

With respect to household characteristics the ‘haves’ live in decent houses which have corrugated iron sheets on the roof and are cemented on the floor, they are spacious and well ventilated and they have decent latrines, bathrooms and kitchens. The ‘have-nots’/poor, on the other hand, live in shabby looking dwellings that are usually grass thatched with poor ventilation and lighting, houses that may leak when it is raining. Since these people spend most of their time looking for or doing ganyu, they do not tender their home environment such that they look filthy and resembling uninhabited or abandoned dwellings. The ‘haves’ have access to safe water or/and are able to treat the water while the ‘have-nots’/poor. In addition, the poor can hardly afford other basic products such as food and to pay for secondary school education for their children. Motivation to Grow Crops We also find differences in the motivation to grow crops, particularly cash crops between the well-being classes. The ‘haves’ grow maize for food and to use it as payment in-kind for ganyu workers, and grow tobacco and cotton for income generation, and legumes for food, income generation and soil improvement. The ‘a-bit-well-to-do’ tend to imitate the ‘haves’ and grow some cash crops for income generation. On the other hand, the ‘have-nots’/poor, grow maize mainly for subsistence needs and only sell the little they harvest out of distress. Some of the ‘have-nots’/poor grow cash crops such as tobacco and cotton just on ‘trial and error’ for some petty cash. Asset Ownership

The other distinguishing feature of well-being classes is the possession of durable assets. The ‘halves’ possess more high-value livestock such as cattle, pigs and goats in addition to the type of livestock typically kept by the ‘have-nots’ and the ‘a-bit-well-to-do’. In addition, the ‘haves’ also own oxen-drawn carts. The ‘a-bit-well-to-do’ own a small number of livestock particularly goats, ducks and chicken, and they hire ox-carts. On the

85

other extreme, some of the ‘have-nots’/poor do own some livestock which are easily sold off in distress times. Participation in the Labour Market and Off-Farm Activities

The ‘haves’ do not participate in ganyu employment as employees, but are the providers of ganyu employment to the other well-being categories on their farms or small business enterprises. On the other hand, most of the ‘have-nots’ participate in the labour market as ganyu employees to get food or inputs. With respect to the ‘a-bit-well-to-do’, some do engage in ganyu to a lesser extent.

In terms of participation off-farm activities, some of the ‘haves’/rich are engaged in the buying and reselling of agricultural produce, they run small scale businesses such as tailoring and fish vending and they are able to procure inputs (in bulk quantities) from either ADMARC or private traders. Although the culture of business exists among the ‘have-nots’/poor, they do not run such businesses because of capital constraints, if anything, their business activities are degradatory to the environment such as selling fuel wood and charcoal. Agricultural Marketing Behaviour

The ‘haves’ are able to buy inputs in bulk from private traders and ADMARC and tend to have a wider choice on where they sell their produce because they have means of transporting produce to distant markets. Sometimes, the ‘haves’ tend to buy cash crops, particularly tobacco, from the ‘a-bit-well-to-do’ and the poor for sale at the auction floors, and buy maize from ADMARC to pay ganyu workers.

The ‘a-bit-well-to-do’ and the ‘have-nots’/poor sell their produce early in the season at a very low price mainly to private traders and vendors before ADMARC starts buying 1 - 3 months later. However, these groups also sell maize and other crops at ADMARC later in the marketing season. While both the ‘a-bit-well-to-do’ and the ‘have-nots’/poor buy maize from ADMARC when they run out of own production stock, the poor are the major buyers of maize from ADMARC. Participation in the Financial Market

The ‘haves’/rich have access to credit, if needed, which enable them to grow and sell a lot of tobacco and cotton - high valued crops in the visited communities. Due to the poor asset ownership status, ‘the have-nots’ fail to qualify for credit as they have nothing which they can declare as chattel or collateral. Some of the ‘a-bit-well-to-do’ have access to credit, particularly to group based credit when available in the communities. Nature of Relationships across the Well-being Classes

Another important distinguishing element of the three well-being classes is the nature of relationships that exist across the classes. The ‘haves’/rich employ people belonging to the ‘have-nots’/poor class (and sometimes from the ‘a bit-well-to-do’/moderately rich class) to work in their farms and homes. They are also buyers of maize and other produce from the ‘have-nots’/poor and sellers of food in the lean months either in cash or through ganyu, a function which saves the ‘have-nots’/poor from starvation. One poor participant in a FGD in the Blantyre site likened the ganyu which they accessed to a bridge across a dangerous river, basically implying that the ganyu helps them to offset some of the impacts of hunger in their households. Table C3.4 Relationships between Well-being Classes

86

Characteristics

‘The haves or the well-to-do’

‘A bit-well-to-do’

‘the have-nots, the poor’

‘the haves’ to • Trading partners • Employers/paymasters • Buyers of produce • Sellers of food in lean period

• Employers/paymasters • Buyers of produce • Sellers of food in lean period / providers of food for ganyu

• Lenders of transport (oxcarts and bicycles)

‘a bit-well-to-do’to • Casual employees/hired labourers

• Buyers of their stored food • Transporters (on their heads)

• Partners • Employers/paymasters

‘the have-nots’ to • Casual employees/hired labourers

• Children herding their livestock

• Buyers of their stored food • Transporters (on their heads)

• Casual employees/labourers

• Partners

Source: Poverty and Social Impact Assessment 2002 Trends in Well-being

Recent poverty statistics released by the government of Malawi show that 65 percent of the population lives in poverty and that about 30 percent of the population are absolutely poor (NSO, 2000). The variance between these statistics and what the various participants reported is decimal. The participants reported figures ranging from 2-10 percent of households in the study sites as being in ‘the haves/well-to-do’ class against 60-75 percent in the ‘have-nots’/poor class and 10-20 percent in the middle group, the ‘a bit-well-to-do’.

The findings in this study suggests that poverty has worsened in the last decade, reinforcing the evidence in QIM 2 in which 50 percent of the communities reported worsening of the poverty situation against the 25 percent that reported improvements in well-being (NEC, 2002a). The discussions with the communities in the sample sites revealed 5-15 percentage drops in the proportion of households that were in the ‘haves’/’well-to-do’ and the ‘a bit-well-to-do’ categories in the last decade and a consequent expansion of the ‘have-nots’/poor class within the same reference period. While some households have dropped from their categories in the period, participants also said that the definition of the well-being categories have also changed though not very much. For example, a decade ago, the ‘have-nots’/poor households used to have maize for 6-8 months, they had more livestock which have been sold off to cushion effects of hunger and they had more ganyu opportunities compared to the current situation where the demand has highly superseded its supply. One commonly reported difference which the participants discussed was the absence of reliable produce and food markets after ADMARC withdrew or reduced its operations. They contended that the closure or scaling down of ADMARC activities made some households to stop producing other crops such as cotton and tobacco because they had no markets to sell them or that they had no access to input credits or cheap inputs. Determinants of Changes in Well-being

Several factors were cited to have pushed many households down the league table towards the poorest pole and these have been put into four groups namely: policy changes, natural calamities and health/personal shocks, market factors and land constraints. Effects of Policy Changes

87

Among the issues which were raised in relation to policy changes included the stoppage of the free cotton seed distribution in cotton growing areas, a situation which made some people to stop or reduce growing the crop; the stoppage of open farm credit; escalating input prices and those of other commodities, unreliable markets after closure of ADMARC because private traders are mobile and can not be readily found within the communities and devaluation of currency. ADMARC had another peculiar relationship with local people in that it was organizing annual competitions as a way of rewarding high producing farmers who produced high quality tobacco and groundnuts. This, according to the participants, promoted production of some crops in some areas of the country. Effects of Market Factors

The market factors cited were price fluctuations and their unpredictability, private traders (including Auction Holding Limited), estate owners and vendors buying crops at lower prices prior to ADMARC opening its markets than prices offered when ADMARC is operational, rural-urban migration in search of jobs and property confiscation after failing to repay loans. It was also worth noting that insecurity was one of the factors that has contributed to the declining lifestyles as some people lost their livestock, property and crops while others suffered bodily injuries. As such, some people sold off their livestock or never invested again in these items for fear of the same. Land and Farming Constraints

Land constraints that were cited included declining landholding sizes and soil infertility which has resulted from over tillage of the land and other environmental degradation issues. Effects of Health and other Personal Shocks

Among the health and personal shocks cited were chronic illnesses (such as HIV/AIDS), deaths of breadwinners, property grabbing and the practices of polygamy and reckless spending patterns. C3.4 Concluding Remarks

This chapter has provided a description of the study sites and categorized the communities into three well-being groups. The areas studied were all relatively remote and with the exception of one or two, difficult to access. Some were deprived of infrastructure like post office, produce market, health facility and police post. Yet, according to livelihood literature improved infrastructure in a place can go a long way towards improving rural livelihoods by stimulating growth of non-farm activities (Anderson and Leiserson, 1979). These non-farm activities would include marketing activities. All the same, the absence of essential infrastructure serves as an indicator of remoteness, and ADMARC in part compensates for this lack of infrastructure to improve market access by the smallholder farmers.

In all the ten sites there is a variety of crops that are grown and a variety of livestock that is reared. There are also a diverse number of non-farm activities. Hence, apart from farming, there are other means of livelihoods. The high incidence of selling household labour in exchange for food, cash and farm inputs and the emergence of the practice of letting out land signals reflects high levels of livelihood diversification to mitigate the consequences of poverty.

88

The perceptions on well-being reveal that food security is an important distinguishing feature of well-being. The well-to-do have adequate maize through out the year or their food stock lasts for at least 8 to 9 months after harvest, the ‘a-bit-to-do’ have food stocks that only last 4 to 6 months from harvest while the poor have maize stocks that last before harvesting or for 1 to 2 months. It turns out that the poor, followed by the ‘a-bit-well-to-do’ are the core buyers of maize from ADMARC and private traders. The other distinguishing features relate to land ownership and housing characteristics, motivation to grow particular crops, asset ownership and participation in the financial market. The analysis has also shown that the proportion of households that is poor in the communities has increased in the last decade mainly due to unfavourable policy changes (closure of markets, removal of subsidies and devaluation of currency leading to high input prices), price fluctuations and their unpredictability, land constraints and the effects of HIV/AIDS.

89

Chapter 4 The Current Roles Of Admarc And Private Traders In Agricultural Marketing

Activities C4.1 Introduction

In the previous chapter we have discussed how rural households categorize well-being or ill-being within the context of the rural livelihoods framework and how some of the categorization are linked to access to inputs and use of the market system. This chapter discusses the current roles of ADMARC and private traders in the marketing of smallholder agricultural produce and inputs given the liberalized environment since 1987. As observed in Chapter 1 the role of ADMARC in crop marketing started to diminish prior to the liberalization of agricultural marketing particularly when ADMARC intensified investments in industrial activities. However, the deterioration of ADMARC’s central place in agricultural produce and inputs marketing is very much associated with agricultural liberalization and the continued cash flow problems the institution has been experiencing since the economic crisis in 1981. Private traders were allowed to compete with ADMARC from 1987, with the expectation that they will efficiently replace ADMARC in agricultural produce and input marketing. We therefore also look at the current role of private traders of various sizes in crop and input marketing in Malawi after 15 years of liberalization of marketing services.

The next section provides evidence on the current role of ADMARC in unit markets targeted for closure (closed seasonal markets) and in unit markets that are proposed to be retained (active markets) in the planned wave of ADMARC closure of markets. One objective of ADMARC was the maintenance of price stability or guaranteeing maize prices for smallholder farmers and consumers. We therefore evaluate whether ADMARC still plays a role in the pricing of maize in rural areas and the role ADMARC played in the recent famine (2001/2002 agricultural season). Section C4.3 focuses on the current role of private traders. We begin with a brief analysis of their evolution, the categories that were found in the sample areas and the roles in the recent famine. Section C4.4 compares the service characteristics of ADMARC and private traders on the basis of types of marketing functions provided, accessibility, pricing and payments systems and general business practices. We distinguish between service characteristics in produce markets and input markets. In Section C4.5 we offer concluding remarks. C4.2 Current Roles of ADMARC in Sample Sites

This section presents the role of ADMARC in the past five years and during the recent famine in the 2001/2002 agricultural season from the perspective of rural households and other stakeholders. We attempt to distinguish the role of ADMARC in closed (seasonal) markets and active markets. We find that ADMARC still plays an important role in the rural areas through its marketing infrastructure particularly in the food supply and input provision functions both in closed and active markets. Roles of ADMARC in Closed (Seasonal) Markets

The evidence from the qualitative assessment in 6 ADMARC unit markets in which seasonal markets were closed shows that ADMARC still plays an important role. Thus, although the seasonal markets in the villages were no longer operational, some rural households switch to alternative ADMARC seasonal markets and unit markets for available marketing services. Table C4.1 shows that, in closed seasonal markets, nearby ADMARC markets still perform the three marketing functions of buying produce including maize when it has the financial resources, selling maize to food insecure households and selling inputs and fertilizers to smallholder farmers in rural areas. However, in recent years ADMARC has been

90

unable to perform these market functions in the rural areas effectively due to its illiquidity and competition from private traders. ADMARC’s role in the marketing of agricultural produce has been declining over time.

Table C4.1 Services offered by ADMARC in Closed Markets in the Past Five Years

What crops and inputs are marketed at ADMARC markets?

Unit Market

(District)

Do households still access ADMARC markets?

Purchase of Produce

Sale of Produce Sale of Inputs

Kakoma (Chikwawa)

Yes Cotton, maize Maize Pesticides and seeds

Lundu (Blantyre)

Yes Maize, Cotton

Maize Various

Chikho (Ntchisi)

Yes Maize, groundnuts and tobacco

Maize Seeds

Lipiri (Dowa)

Yes Beans, groundnuts, maize

Maize Various

Eswazini (Mzimba)

Yes Maize, groundnuts, beans, soya beans

Maize Fertilizer, pesticides

Nthalire (Chitipa)

Yes Maize, tobacco, beans, paprika

Maize Various

Source: Poverty and Social Impact Assessment 2002

ADMARC still performs the function of purchasing produce from farmers in closed seasonal markets albeit to a very small extent. The use of smallholder farmers to dispose off their produce depends on the well-being group. In most cases it is the poor and the moderate poor that sell maize to ADMARC since they sell small quantities, but they also sell crops to private traders particularly soon after harvest before ADMARC opens its market. In Chikwawa, it is the poor and the moderate poor that sell maize to ADMARC. Cotton is another important crop that is sold to ADMARC by all well-being groups particularly the rich and the moderate poor who have the money and land to cultivate in Kakoma (Chikwawa) and Lundu (Blanytre) sites. In other markets, such as Blantyre and Dowa, in the past five years the well-to-do or rich and the moderate poor used to sell tobacco, maize, cotton and groundnuts at ADMARC. In Chikho (Ntchisi) site, the focus group discussions revealed that all well-being groups were selling produce including maize at ADMARC although the poor sold maize in small quantities.

Similarly, ADMARC’s role in selling produce particularly maize has not diminished as much as that of purchasing maize from smallholder farmers. However, the supply of maize at ADMARC’s markets was irregular and erratic particularly in the famine season and this has lead to households to use alternative sources of maize. When maize is available at ADMARC market, it is the most reliable and affordable source of maize for rural households particularly the poor and food insecure that could afford and the rich who buy the maize to pay ‘ganyu’ workers or tenants. The purchase of maize from ADMARC by the well-to-do was observed in Lipiri (Dowa) and Nthalire (Chitipa) sites as a form of payment to the poor who are usually contracted for ‘ganyu’ labour and to their tenants. In Nthalire (Chitipa) it was indicated that the moderate poor and the poor are the groups that mainly purchase maize from ADMARC using proceeds from ‘ganyu’. The evidence in the closed seasonal market sites suggests that ADMARC is the first-choice market for the supply of maize whenever maize is available due mainly to the low and stable price charge for maize. However, the erratic supply has led

91

households to depend on vendors within the village or at ‘day markets’ but purchase maize at much higher prices than those charged by ADMARC.

ADMARC also plays a significant role in the input market for most rural households across all well-being groups, although the rich with transport facilities or who can afford transport expenses are able to access alternative suppliers. In all the closed seasonal markets, households still use nearby ADMARC markets to source inputs such as fertilizers, seeds, insecticides and farm implements when available because they stock good quality inputs and ADMARC markets are relatively closer to the rural households than private traders that sell similar inputs such as Farmers World. However, the problem is that most of the inputs are not available on time. The input market seem to benefit mainly the ‘a bit well to do’ and the ‘well to do’, since most of the poor do not buy inputs particularly fertilizers because they can not afford. The poor reported having used fertilizers when it was provided on credit or through grants. Roles of ADMARC in Active Markets

The role of ADMARC in sites with active markets is similar to that in sites with closed seasonal markets, in that the ADMARC’s purchase of produce has substantially declined while it still plays an important role in the sale of maize and agricultural inputs (Table C4.2). However, it appears ADMARC has been a major buyer of cotton from smallholder farmers in the Chiondo (Nkhotakota) and Mpata (Karonga) sites in the past five years. In Chiondo (Nkhotakota) site, ADMARC and Cotton Ginnery Limited are the only buyers of cotton. While in Mpata (Karonga) site ADMARC is the sole buyer of cotton. Maize is the main food crop that ADMARC sells to rural households and particularly to the poor whenever maize is available at the market. If maize is available both at ADMARC and private traders or vendors, rural households prefer buying maize from ADMARC because of lower, stable and affordable prices and the low transaction costs because of the proximity of markets. Table C4.2 Services offered by ADMARC in Active Markets in the Past Five Years

What crops and inputs are marketed at ADMARC markets?

Unit Market

(District)

Do households still access ADMARC markets?

Purchase of Produce

Sale of Produce Sale of Inputs

Kasongo (Phalombe)

Yes Maize, beans, chicken peas

Maize, groundnuts,

beans

Seeds, pesticides and implements


Yes Cotton, groundnuts and maize

Maize Fertilizer, cotton seeds, pesticides

Mikundi (Mchinji)

Yes Maize and groundnuts

Maize Various

Mpata (Karonga)

Yes Maize and cotton Maize Various


Rural households in sites with an active ADMARC market largely depend on ADMARC to purchase fertilizers and other inputs and it is seen as a reliable source of inputs as it is closer to the communities than the large private traders that engage in input sale. All the well-being groups tend to rely on ADMARC for inputs, although the ‘well-to-do’ also afford to purchase inputs from large-scale private traders.

92

Role of ADMARC in Produce Pricing

One of the objectives of the reform in the pricing system of agricultural produce, particularly the pricing of maize was that ADMARC’s marketing activities in the maize would be reduced to buyer of last resort and supplier of maize in difficult times based on the guaranteed price band. We find evidence in this study that ADMARC tends to set the benchmark price below which smallholder farmers cannot sell their products to private traders. There is a tendency for maize prices offered by private traders to smallholder farmers before ADMARC opens its purchasing activity in the season to be lower than the equivalent prices that private traders offer when ADMARC is also buying maize. ADMARC usually opens its markets late in the season while private traders provide an early market for smallholder produce in the season. Nonetheless, smallholder farmers in most sites reported that the announced prices offered by private traders earlier in the buying season are higher than the prices offered by ADMARC.1 This benchmarking pricing role of ADMARC may not exist in a purely private marketing system.

Similarly, when maize is available at ADMARC markets the private traders tend to charge lower prices than when maize is not available at ADMARC markets. For instance in one of the focus group in Lundu (Blantyre) site the importance of benchmark pricing role of ADMARC is captured in the following:

“Last year the private trader was selling maize initially using a No. 10 plate at MK1.00 per plate. But when maize was available at ADMARC at K17.00 per kilogram, the private trader started using a weighing scale and sold maize at MK17.50 per kilogram.” [Female Group in Lundu Market]

ADMARC’s Role in Recent Famine

The role of ADMARC in the famine season (2001/2002) has been marginal both in active markets and in areas where some seasonal markets were closed. ADMARC hardly purchased crop produce in most of the markets included in this study, mostly due to its cash flow problems. The sentiments from the rural households about the inactive role of ADMARC in all its three marketing functions support the low volume of agricultural produce trade recorded in official statistics. Nonetheless, ADMARC still sold maize to rural households whenever it was available. In all the sample markets, the supply of maize at ADMARC markets was rather irregular and erratic, such that most of the demand for maize was not satisfied and households resorted to private traders particularly vendors to procure maize at very high prices (Table C4.3). 1 The effective prices offered by private traders may be lower or no better than ADMARC prices due to the widespread cheating of private traders on measurements and the quality of smallholder farmers’ produce as observed below.

93

Table C4.3 ADMARC Sale of Maize in Closed and Active Markets in 2001/2002 Season

Unit Market

(District)

Did ADMARC Sale Maize in the

Famine season?

Price per Kilogram (Kwacha)

Regularity of Supply

When was Maize Available?

A. Closed Markets Kakoma (Chikwawa)

Yes/No 17 Once/none -

Lundu (Blantyre)

Yes 17 Irregular

June – September 2001

Chikho (Ntchisi)

Yes 17 Irregular

January end, February end and April 2002

Lipiri (Dowa)

Yes 17 Irregular

January

Eswazini (Mzimba)

Yes 17 Irregular

November 2001 and April 2002

Nthalire (Chitipa)

Yes 17 Irregular

-

B. Active Markets Kasongo (Phalombe)

Yes 17 Very Irregular

-


Yes 17 Only twice

January and February 2002

Mikundi (Mchinji)

Yes 17 Irregular

February 2002 – April 2002

Mpata (Karonga)

Yes 17 Irregular, but also sold rice

-


The short supply of maize led to bribery and nepotism in the selling of the available maize at some ADMARC markets. ADMARC markets were characterized by long queues of starving people who sometimes spent nights at the markets but failed to buy maize in the end or managed to buy little maize because of the rationing regulation which was instituted. In some places, ADMARC officers were reportedly serving people from the surrounding villages first or vendors, implying that people from closed ADMARC markets had to wait until their counterparts were served first. In all the sites, ADMARC sold maize at MK17.00 per kilogram. The maize was procured from the National Food Reserve Agency at the MK15.00 per kilogram and ADMARC sold the maize at the government recommended price. The long queues at ADMARC markets during the famine season suggest the relative importance of ADMARC in the supply of maize in difficult situations.

In addition, ADMARC did not perform its benchmark pricing effectively in the famine season due to the fact that ADMARC hardly purchased produce and the maize supply was not available at all times during the farming season. This implied that most rural household relied on private traders and vendors for their maize purchases at much higher prices than what would have ruled if maize were available at ADMARC markets. In all the sites, the poor who are usually food insecure based on their own production could hardly afford maize at prices offered by the vendors, but did rely on ganyu and gifts from richer households. C4.3 The Marketing Operations of Private Traders

The liberalization of agricultural produce marketing was undertaken in order to complement the marketing services of ADMARC and to assist in the creation of private trade

94

structures operated by Malawians, thus stimulating employment opportunities and reformulation of private enterprises (EPD, 1990).2 Over the years, a number of private traders, both small and large scale, have emerged and are trading in several crops including the controlled barley tobacco and cotton. As noted by Chilowa (1998), some of the structural adjustment initiatives, notably agricultural marketing liberalization, have not impacted favourably on the livelihoods, food security and the general welfare of the poor in Malawi. There are losers and winners in the process. The losers are mainly the smallholder farmers in the category of net food buyers, low income or wage earners in urban and remote areas of the country whereas the winners are smallholder farmers in the category of net food sellers, private traders and institutional traders.

As noted in chapter 1, in the initial period following announcement of the agricultural marketing liberalization, the response of private traders was somehow encouraging. However, Chirwa (1998) notes that such an increase was short-lived, as 1990 saw a decline of 43 percent in the numbers of private traders, which would imply that either the private traders lost interest in the agricultural marketing and/or that with more liberalization, some private traders saw no regulatory requirement to be registered, a situation which would make monitoring and evaluation of their activities to be very difficult. Types of Private Traders Operating in the Sample Sites

The results from this study suggest that four categories of private traders operate in the ten sample sites as follows: institutional companies and manufacturers, large wholesalers and other large traders, retailers/vendors, and estate/large scale farmers.3 Table C4.4 provides a summary of the private trading institutions and non-governmental organizations that operate in the sample sites. Institutional Companies and Manufacturers

Institutional companies or large scale traders identified in the sites include Auction Holdings Limited (AHL), Tobacco Association of Malawi (TAMA) and Limbe Leaf Tobacco Company (LLTC) which specialize in tobacco trading. Manufacturing firms include Cotton Ginnery Limited (CGL) that specializes in cotton trading, Cargil Inc that specializes in groundnuts and rice purchases for manufacturing and Rab Processors specializing in maize purchases for manufacturing. These companies are based in urban areas but they have satellite depots and/or offices in rural areas. CGL was identified in the cotton growing areas such as Kasongo (Chikwawa) and Mpata (Karonga) whereas AHL and TAMA4 were identified in the areas where farmers grow tobacco. In all the sites where these were identified participants indicated that these companies were at trading centres or other places of high economic activities where people took their produce for sale or where they went to procure and/or obtain inputs.

2 Complementing the services of ADMARC by private traders was envisaged to have such advantages as reducing the financial requirements of ADMARC, thus assisting it to overcome its liquidity problems; introducing more competition in marketing operations thereby improving the services and increasing the sale options for the farmers; and lastly, permitting ADMARC to reduce uneconomic operations and substitute them by the initiatives of private traders. 3 The categorization of private traders is based on the information from focus group discussions, in which some of the categories appearing in the literature did not come out such as the private traders with pick-up trucks. Farmers categorize all non-institutional traders as vendors. 4 TAMA is an association for tobacco farmers (particularly smallholder farmers) whose functions include organizing and assisting farmers with group transportation of their tobacco as well as bargaining for better prices at the auction floors and assisting the registered farmers with some input credit.

95

All the institutional companies and manufacturers are highly specialized in their

particular crops, as such; they only deal with particular types of farmers. However, an observation was made that CGL used to provide free seed to cotton farmers until the mid-90’s and that thereafter, these seeds were provided on credit until in the last 2-3 years when the scale of these services declined considerably, thereby making some farmers to reduce their cotton output or to abandon the crop completely. Table C4.4 Availability of Private Traders and other Organizations in the Study Sites

Closed Markets Active Markets Market (District) Institutions Market (District) Institutions • • Nthalire • (Chitipa)

• Private Traders • AHL/TAMA • Chipiku • Phwezi school • Malawian vendors • Zambian vendors • Other Organisations • MRFC

• Mpata • (Karonga)

• Private Traders • HGD • PTC • Malawian vendors • Tanzanian Vendors • Secondary school

• • Eswazini • (Mzimba)

• PrivateTraders • Cheetah • LLTC • Farmers World

• Chiondo • (Nkhotakota)

• Private Traders • CGL • CARGIL • AHL/TAMA • Farmers World • Vendors • Other Organisations • MRFC

• Lipiri • (Dowa)

• Private Traders • Farmers World • Chipiku • PTC • Vendors • Other Organisations • ELDP

• Mikundi • (Mchinji)

• Private Traders • Chipiku • Cheetah • AHL/TAMA • Mc Connell & Co. • Malawian vendors • Zambian Vendors • Other Organisations • NASFAM • MRFC

• Chikho • (Ntchisi)

• Private Traders • AHL/TAMA • Farmers World • Vendors

• Kasongo • (Phalombe)

• Private Traders • Vendors • Cheetah

• Kakoma • (Chikwawa)

• Private Traders • CGL • Vendors • Other Organisations • CADECOM • ELDP

• Lundu • (Blantyre)

• Private Traders • Vendors

Source: Poverty and Social Impact Assessment 2002 Large Wholesalers and other Large Traders

Large wholesalers and other large traders are limited companies whose main business is trading. In this category we include the following institutions identified in the sample sites: Chipiku Stores, Farmers World, Cheetah, Peoples Trading Centre (PTC), Hardware and General Dealers (HGD), McConnell and Company and various secondary schools and educational institutions. The educational institutions involved in purchase of maize and food crops include Phwezi Secondary School in the Nthalire site and Chaminade Secondary School

96

in the Mpata site. With the exception of the secondary schools (which were over 20km away), the rest were located at trading centres which sometimes were very close to the study sites and/or very far (over 20km).

The other limited firms such as Chipiku Stores, PTC, HGD, Farmers World and McConnell and Company sell inputs (tobacco and maize seed and fertilizer) but they do not procure any produce from farmers although in one case (Lipiri (Dowa) site), Chipiku bought soya beans from farmers5. Farmers World was only identified in four of the 10 sites. The problems associated with these firms are that they are located at the district headquarters or at trading centres which are very far from remote areas and that sometimes they do not stock adequate quantities of the inputs. It is apparent that many of the large wholesalers particularly those in the input market are not accessible to the majority of the rural population, as they choose to locate in areas where the infrastructure is good. Vendors and Other Retailers

Retailers and vendors are small and often unregistered traders. Vendors in the context of this study and from the perspective of smallholder farmers include those resident in the village and those traders with pick-up trucks who operate mobile markets. These were reported to operate in all the sample sites. In the sites bordering other countries, some of the vendors/retailers came from these countries such as Mozambique, Tanzania and Zambia while at the same time some Malawian vendors and farmers were also reported to venture into these countries for the purposes of either selling or procuring maize and other tradable products such as clothes. In the Lundu (Blantyre) site, the retailers included Indian shop owners residing in Blantyre town but had agents in the rural areas.

The vendors specialize in the marketing of crop produce and do not deal in agricultural inputs. The vendors are either middle men (purchasing tobacco, maize, soya and cotton from smallholders in loose quantities for resale to either ADMARC, AHL, CGL or other big companies) or they are direct traders, who buy from one area or country and resale that produce either in loose or bulk quantities at a later date to consumers in another area or (luckily) in the same area. The vendors and retailers move from door to door, village to village and sometimes from market to market following the schedules of flea markets such as Tiwonewone markets in the Nthalire (Chitipa) site or Salaula in the Eswazini (Mzimba) site. Large-Scale Farmers and Estates

The other group of private traders in the rural areas is that of the large-scale farmers and estates. Apart from growing their own maize, large farmers were also said to procure maize from the smallholders, stock it and resell later or use it to pay for hired labour or for feeding tenants. They were also reported to buy cotton and tobacco for re-grading and selling at higher markets. However, the scale of these operations are minimal bearing in mind they there are very few estates in most rural areas and most of these do not want to buy some crops from outside their production for quality maintenance. Private Traders and Distribution of Farm Inputs

The discussion above points out to the fact that private traders have not taken up the distribution of farm inputs in the same way that they have done with the purchasing of produce. CGL has been providing free cotton seeds to cotton farmers until mid-90’s when it started providing the inputs on a soft credit, a scheme which has been gradually declining to the extent that it now provides the credit at commercial value. On the other hand, licensed tobacco farmers obtain input loans from AHL/TAMA through commercial banks and are 5 PTC sometimes procures vegetables and fruits from large scale farms which are sold in its shops

97

deducted at the time of selling their tobacco. These schemes have existed even during the pre-independence times and will not be negatively affected by the withdrawal of ADMARC markets.

The experiences of the visited communities with vendors and retailers were that this calibre of players is not interested in selling inputs and/or that there are very few vendors/retailers who sell inputs in loose quantities during flea markets but at exhaustive prices. The only exceptions here are firstly, for the sites bordering Zambia where there is a lot of informal (usually illegal) importation and selling of fertilisers. Secondly, is the case where some vendors pre-arrange with some farmers to provide them with inputs and bind the farmers to sell their produce to them after harvesting. This is usually the case in tobacco and cotton growing areas and other areas growing other cash crops.

As such, when one is talking about input distribution in rural areas by private traders, one is therefore left with the larger wholesalers and companies such as Farmers World, Chipiku Stores, SFFRFM and larger shops. These wholesalers are unevenly distributed and some have either reduced their operations or have completely shutdown. Roles of Private Traders in the Recent Famine

In some areas private traders (mainly vendors) played significant roles in the recent famine as they sold maize at various markets around the country. Unlike in the produce procurement times when they move around in the villages, at this time the vendors/retailers sell the maize at markets (daily or mobile) or at designated places in the community. Comparatively, border communities had more vendors/retailers than in-land communities.

However, the vendors were reportedly overcharging for their maize and tampering with measuring equipment such as loosening scales and/or flattening measuring plates. While ADMARC was selling its maize at K17/kg throughout the country (whenever it had maize in its depots), the vendors and some private traders were reportedly selling their maize at prices ranging from K22/kg to K45/kg. As expected, the vendors/retailers had no ceiling on the amount of maize that people had to buy. The Role of Other Governmental and Non-Governmental Organisations

The other institutions found in the 10 sample sits are governmental and non-governmental organizations (NGOs) working with the communities in the rural areas. These include the Catholic Development Commission (CADECOM) in the Kakoma; the Evangelical Lutheran Development Programme (ELDP) in the Kakoma and Lipiri sites; the National Association of Smallholder Farmers (NASFAM) in the Mikundi site; and Sasakawa Globe 2000 in the Kasongo site. These NGOs were implementing several food security interventions in the stated areas including seed multiplication and input credit projects as well facilitating produce transportation and fair market practices (the case of NASFAM is a good example).

Farmers Financing Company6 and Malawi Rural Finance Company (MRFC) were institutions that provide input loans to smallholder farmers. According to the participants, MRFC provided input loans to farmers through their clubs, either Tilime clubs or Tikolole clubs. Tilime clubs were for the a-bit-well-to-do and the well-to-do farmers whereas Tikolole clubs were the poorest households. The two clubs differed in the amount of inputs which were 6 FFC was a sister company to Farmers World but was specializing in providing farm inputs (fertilizer and maize hybrid seed) to maize smallholder farmers on credit. FFC has since 2000/01 been closed indefinitely.

98

provided in that Tilime farmers had a higher loan ceiling compared to the Tikolole club members. The common observation regarding the operations of MRFC though was that it had scaled down its credit provision services in the last 2-5 years basically because of the repayment crisis which had rocked the company as some borrowers could not repay because of poor harvests, low produce prices and the high interest rates which were being charged on the principal loans. The same observations were also made for the FFC though in some sites, it was reported to have completely pulled out (closed down). C4.4 Comparative Performance of ADMARC and Private Traders

In the previous sections we have discussed the roles that ADMARC and private traders play in the marketing of smallholder agricultural produce and agricultural inputs. While the role of ADMARC is declining particularly in the purchase of agricultural produce, it could still play a potentially important role as a provider of food (maize) and inputs at stable and affordable prices. Private traders, both large-scale and vendors, are progressively replacing ADMARC in its market function of purchasing produce in markets that have well developed infrastructure and to a lesser extent vendors have attempted to engage in the supply of food crops, particularly maize in the last few years. However, the private agricultural input market is under-developed and only a few private traders sell inputs to rural households and tend to locate in urban and peri-urban areas that are far from the rural areas. We have also noted that in markets where private traders are active particularly in border markets, the withdrawal of ADMARC markets may not have significant negative consequences on access to produce markets, but food security and regular access to inputs can not be guaranteed. The Service Characteristics in the Crop Produce Market

The comparative analysis in the service characteristics of ADMARC and private traders focuses on four service elements: accessibility to the rural households, reliability of traders, pricing and prevailing payment systems and general business practices. We look at the crop produce market particularly focusing on the sales and purchases of maize and the agricultural input market. Table C4.5 presents the service characteristics of ADMARC and private traders in maize marketing based on information from the study sites. Both ADMARC and vendors purchase and sell maize, while large-scale traders only buy maize from farmers. One aspect that distinguishes the aspects marketing aspects of ADMARC and private traders is accessibility to markets by smallholder farmer.

Respondents were asked to compare the relative accessibility of markets provided by ADMARC and private traders. The results from the study indicate that ADMARC sells and buys maize at fixed locations (seasonal or unit markets) but such locations are more accessible in active markets as they are within short distances than in closed seasonal markets. In many closed seasonal markets, rural households now use the services of ADMARC in alternative active markets and the distance travelled is longer than before seasonal markets were closed. This implies that the closure of markets can not guarantee the growth of the private sector, but its immediate effect would be to raise transaction costs for farmers. This finding is similar to the results from Nthara (2002) who observes that the number of private traders in particular markets is not positively correlated with the absence of ADMARC markets. Large-scale traders are not many and were reported to be less accessible to most farmers. These large-scale traders mainly specialize in the purchase of crops, particularly cash crops. However, when they come to buy produce some provide door-to-door service. Vendors, on the other hand, were reported to be the most accessible as most of them live in the same or surrounding villages. In all sites, vendors offer a readily market for maize both purchases and sales in the rural areas.

99

Table C4.5 Service Characteristics in the Purchase and Sale of Maize Private Traders Service

Characteristic

ADMARC Large-Scale Vendors

Services offered

o Buy maize o Sell maize

o Buy maize o Do not sell maize

o Buy maize o Sell maize

Accessibility

o Fixed locations o Relatively short distance in active markets and long distance closed markets

o Less accessible – a few just come to buy maize o Come to buy specific crops at specific points o Long distance to markets, but others door-to-door purchases

o Very accessible – tend to live in the same village o Door-to-door services for both purchase and sale of maize

Reliability

o Reliable but lack funds most of the times o Have irregular maize stocks o Keep maize in the market for resale later in the lean season when available

o Less reliable - not sure that they will come again o Buy maize from the areas and do not resale in the lean season

o Very reliable because most are located in the village o Buy maize and also sell maize in the lean season o Sell maize in rural ‘day markets’

Pricing and Payments

o Low maize producer prices o Stable producer and consumer maize prices o No price discrimination o Prices not negotiable o Late payments and sometimes buy on credit

o High maize producer prices o Volatile producer maize prices o Price discrimination with smaller traders discriminating more o Dictated prices o Instant cash payment o Some do barter trade – especially in border areas

o Outrageous consumer maize prices o High producer maize prices but less than large-scale traders o Volatile maize prices o High price discrimination o Negotiable prices but vendors tend to dictate producer prices o Most pay cash instantly, a few buy on credit o Some do barter trade – especially in border areas

Business Practices

o Buy small quantities o Tampers with scales to less extent o Farmers can verify weighing scales o Start buying maize late in the season o Sometimes sell maize mixed with sunflower o Favours selling maize to vendors o Price information is displayed

o Announces better prices but actually pay less o Tampers with scale to a large extent and some use volume o Do not allow farmers to verify their weighing scales

o Mostly use volume measures o Less cheating when using volume measures o Tamper with scales to a large extent – those that use weights o Deliberately undervalue crops o Do not allow farmers to verify their weighing scales o Sell poor quality maize


On the issue of reliability of various marketing institutions, vendors were reported to be the most reliable because of their proximity to farmers, on daily basis or through day markets such as ‘Tiwonewone’ in Nthalire (Chitipa) and ‘Salaula’ in Eswazini (Mzimba). In addition, vendors are also most reliable because they buy and sell maize in the areas in contrast to large-scale traders who only purchase maize. Although, on this count ADMARC also performs both purchasing and selling functions in the food market, vendors tend to have maize supplies more regularly while the supply of maize is irregular. In Dowa, respondents indicated that when vendors run out of maize stock sometimes they buy maize from distant ADMARC markets to sell in the remote areas. ADMARC in most sites was ranked second in terms of reliability because of their location at fixed points, although the volume of business has gone down substantially in recent years. Respondents reported that in most cases ADMARC does not have adequate cash to buy produce from farmers, but when they have funds the fact that they purchase maize and keep some for resale was a positive contribution to food security in difficult times. The large-scale traders are the least reliable because they just come to buy the maize and do not come back to sell in the lean season. In addition, since

100

large scale private traders are not located in the rural areas, it is difficult to trace them when farmers want to sell products.

The price competitiveness of various marketing institutions is also an important distinguishing feature in rural markets. Generally, in all sample sites ADMARC offers lowest maize producer and consumer prices and such prices are fixed and stable throughout the season and are not discriminatory across sellers and buyers. Large-scale private traders offer the highest producer prices for produce including maize. However, these prices are volatile and some traders tend to price discriminate and negotiate for lower prices and in most cases the traders’ prices were reported to prevail. Except in one case, in Chitipa one of the main institutional trader Phwezi Secondary School buy maize at prices offered by the farmers and sometimes buy maize at a price higher than that offered by the farmers.7 Vendors tend to offer intermediate maize producer prices between those offered by ADMARC and those offered by large-scale private traders, but charge substantially higher maize consumer prices than ADMARC. In all the study sites it was reported that vendors bargain more over prices than large-scale traders. In the Mzimba site vendors tend to offer lower prices than those at which ADMARC buy produce and the communities feel exploited by the vendors. Prices therefore tend to be very volatile and there is high discrimination in the purchase price. As one focus group discussion in Dowa (closed seasonal market site) revealed:

“These vendors do not have a uniform price. They go at one house they buy a pail [approximately 20 kilograms] of maize at MK100.00, at the next house they offer MK50.00, then the next is MK30.00 and then MK70.00 per pail. Their prices vary from one house to the next per pail.”

It is apparent that most private traders operate as discriminating monopsonists in the

purchase of maize and other agricultural produce and as discriminating monopolists in the sale of maize. This casts doubts on the efficiency of the private marketing system in Malawi, but may also explain why the profit margins are quite high among Malawian traders (Fafchamps and Gabre-Madhin, 2001). In addition, the pricing behaviour of private traders creates a lot of price uncertainties for farmers in rural areas and in most cases the uncertainty in prices does not seem to motivate farmers to positively respond to price developments. For instance, across all well-being groups particularly the ‘a bit-well-to-do’ and the ‘poor’ mainly grow maize to meet their subsistence needs other than to generate income. Selling maize for income generation is usually out of distress and as such the decision to sell part of the maize is residual for these categories of farmers.

In terms of mode of payment, ADMARC pays cash when it has funds available but often buys produce without instantly paying the farmers. On the contrary, most private traders whether large-scale or vendors pay cash instantly, although in very exceptional cases vendors were reported to have purchased produce on credit and pay after they sell the products. This further supports the evidence in Fafchamps and Gabre-Madhin (2001) that 99.2 percent of private traders in Malawi pay cash for crop purchases. We also found high incidence of barter trade between private traders and farmers in markets that are located close to the border such as Mchinji, Karonga and Chitipa. NEC (2002) attributes this phenomenon to the problem of currency conversions particularly for foreign traders.

The most problematic area with the private marketing system that adversely affects the incentive structure of smallholder farmers is the quality of business practices. The fact that private marketing is completely unregulated, combined with monopsonistic or monopolistic tendencies of private traders in most sites, fair trading practices have not 7 The respondents reported that the founders of Phwezi Secondary School come from this are and raise the price at which they buy from farmers to promote agricultural development in the area, as part of the return to social capital.

101

evolved over time. While it is true that in many crops private traders tend to offer higher producer prices, some of the business practices have the potential to off-set the positive gains and the incomes of rural farmers. First, the most cited problem associated with the private traders is the cheating on measurement and weights, although ADMARC officers in isolated cases are engaged in cheating practices. Respondents in all sites believe that they are cheated more when they are conducting business with private traders than when doing business with ADMARC. Private traders, particularly vendors tamper with their weighing scales more than ADMARC officials. From a focus group discussion in Ntchisi (closed seasonal market site) one participant had this to say to demonstrate the extent of cheating by private traders:

“One day I wanted to sell my Soya at ADMARC market. On my way I found one of the vendors along the road which was leading to the ADMARC market. I went to him to sell my produce. However, I was surprised that the Soya weighed only 11 kilograms and decided not to sell to him then took my Soya to ADMARC where it weighed 25 kilograms.”

The cases of cheating on measurements tend to be limited where private traders use

volume other than weight measures in the purchase of produce. The reported cheating in volume measures occur because some private traders tend to use larger pails than those used by the farmers or by deliberately undervaluing the produce to buy at a lower price or use smaller plates when selling the maize to farmers. The assertion that there is more cheating among private traders is supported by the fact that farmers are not allowed to verify private traders’ weighing scales compared to ADMARC where farmers can verify the weighing scales and the weight of their produce. The other form of cheating is the announced prices versus the actual prices paid to farmers. Vendors tend to announce better producer prices, but they actually pay the farmer less than the announced price when trade takes place. On the contrary, ADMARC displays the prices for all the farmers and abide by the announced prices.

However, in other cash crops such as cotton ADMARC offers better prices than vendors. For instance, in Chikwawa (closed market site) respondents noted that ADMARC buys cotton at better prices than private traders (Cotton Ginnery Limited) because officers do not tamper with weighing instruments. In Nkhotakota (active market) ADMARC offers better prices than venders who sometimes buy cotton early from farmers and later sell to ADMARC.

The second aspect that emerges from the study is the quality of products. ADMARC buys produce by grades while private traders, particularly vendors, tend to buy without grades. There are mixed observations, however, on the quality of maize sold to rural households by ADMARC and private traders. For example, in the Chikwawa and Blanytre sites, ADMARC was reported to have sold poor quality maize mixed with sunflower while in the Mchinji, Phalombe, Dowa, Ntchisi, Mzimba, Chitipa, Karonga and Nkhotakota sites the maize sold was reported to be of good quality. The vendors were in most sites reported to be frequently selling poor quality maize.

The third business practice that emerges is the timeliness of trade. Private traders particularly vendors ranked highly on providing markets for produce earlier in the selling season than large scale traders and ADMARC. The early entry of vendors in purchasing produce is viewed as a positive aspect of the private marketing system because it provides opportunities for farmers to earn income when they need it most. However, on balance there may be losses made by farmers because vendors tend to offer lower prices than the prices that are offered when ADMARC and large scale traders are also purchasing produce from the farmers. The Service Characteristics in the Agricultural Input Market

102

The agricultural input market in the rural areas is still dominated by ADMARC with a few large-scale wholesalers or limited companies in selected sample sites. Vendors do not sell inputs to farmers; they are so specialized in the crop produce market. Since, most private traders found in the sites are mobile vendors the capital and storage requirements of the input business may be prohibitive for many private traders. Table C4.6 summarizes the service characteristics of ADMARC and large-scale wholesale companies involved in agricultural input sales. Both ADMARC and institutional companies sell inputs in fixed locations, but ADMARC markets are more accessible to a larger proportion of farmers than large-scale wholesalers or limited companies. Large-scale wholesalers or limited companies such as Farmers World, Chipiku Stores, McConnell and PTC were all found in urban and peri-urban areas. For instance, Farmers World was only found to be a supplier of agricultural inputs in 4 of the 10 sample sites. Similarly, the Smallholder Farmer Fertilizer Revolving Fund of Malawi (SFFRFM) has only eight selling points in the central region in 4 districts and it was only mentioned in the Dowa site of the 10 sample sites. In the sites where these institutional companies were operating, it was the ‘well-to-do’ group that has more access because they can afford the transport costs while ADMARC is accessible to all well-being groups.

ADMARC was also reported to be more reliable in normal years when inputs are available although often inputs are not available on time. It was reported that in some cases, ADMARC bring the inputs to the markets late in the farming season, at a time when such inputs are not required. The large-scale wholesalers have inputs available most of the times, but only a few farmers have access. In terms of pricing, ADMARC was reported to offer inputs at lower prices. However, in sites where private traders are also engaged in the sale of inputs, large-scale wholesalers do offer lower shelf prices than ADMARC but the transport costs tend to wipe the price differential that exists. Nonetheless, in the Lipiri (Dowa) site, prices of inputs at Farmers World were reported to be higher than those offered by ADMARC. Table C4.6 Service Characteristics in the Sale of Inputs

Service Characteristic

ADMARC Large-scale Wholesalers and Limited Companies

Type of Service o Sell various farm inputs o Sell various farm inputs

Accessibility

o Fixed locations but within shorter distances o Located in rural areas o Accessible to even ‘poor’ and ‘a-bit-to-do’ farmers

o Fixed locations from rural areas o Located in urban and peri-urban areas o More accessible to the ‘well-to-do’ farmers o High transport costs

Reliability

o More reliable when inputs are available o Do not deliver inputs on time

o Reliable to those farmers who have access o Usually have inputs always in stock and at appropriate time

Pricing and Payments

o Sell inputs at lower prices in some sites o If the only reliable source; sells inputs at high price o Otherwise, offers lower price than ADMARC

o Business Practices

o Sometimes sell inputs of poor quality o Sometimes sell farmers inappropriate type of fertilizer

o Mix fertilizer with sand o Sell better quality inputs most of the times


The private traders are less reliable suppliers of agricultural inputs. In the Eswazini (Mzimba) site for example, a private institution collected what it called ‘deposits’ for input credit qualification but it vanished after it collected the deposits and was never traceable again. In the same Eswazini site, some farmers who teamed up to buy large amounts of fertilizer so that they may qualify for free transportation by the selling company got disappointed when their fertilisers were not delivered to their villages and were left stranded

103

halfway before reaching their destination. The general picture therefore that one would draw from this discussion is that despite its structural weaknesses, ADMARC still perceived as a reliable supplier of farm inputs to most rural areas.

The quality of inputs was reported to be poor in most sites for both ADMARC and institutional companies. ADMARC in some cases supplies expired inputs and sometimes sell inappropriate types of fertilizers to farmers. In the Kakoma (Chikwawa) site, respondents indicated that ADMARC sold expired cotton seeds and in the Lundu (Blantyre) site respondents noted that most of the seeds do not germinate. The views in the other sites are different and ADMARC was reported to sell good quality inputs. On the other hand, in sites where institutional companies operate they tend to sell better quality inputs. C4.5 Concluding Remarks

The analysis in this chapter reveals that the activities of ADMARC have gone down substantially in the past decade both in the closed and active markets, particularly in the past two years. In all the sites in most cases ADMARC does not have adequate cash to buy produce from smallholder farmers, they do not have regular supplies of maize and agricultural inputs are usually supplied late. However, ADMARC still performs the three marketing functions of buying smallholder produce, selling food products particularly maize to rural households and selling agricultural inputs from smallholder farmers although its business has declined in recent years. There are several reasons that can be attributed to the declining business of ADMARC including inefficiency in its investment portfolio and marketing activities, liquidity problems, competition from private traders following liberalisation of agricultural markets.

Only a few large-scale private traders were found to operate in the 10 sample sites. However, the private marketing system in the sample sites is dominated by individual small-scale unlicensed traders or vendors engaged in short-term arbitrage. It is the vendors, other than the large-scale private traders, that to some extent seem to have bridged the gap left by ADMARC’s inability to provide services efficiently when it comes to buying of smallholder produce both maize and cash crops and the vendors have particularly provided vital in selling maize to the rural households. Unfortunately, the closure of seasonal markets and the irregular supply of maize and inputs and the lack of funds by ADMARC has increased the distance which the ‘a-bit-well-to-do’ and the ‘poor’, who are the main users of ADMARC’s marketing services, travel to markets for purchase of cheaper maize and inputs.

In the produce market, vendors and some private traders are associated with no direct transportation costs for the farmers as traders tend to provide door-to-door services compared to ADMARC markets that are at fixed locations and tend to be far from the remote areas. The situation in the input market is different. For more remote villages and particularly in closed seasonal ADMARC markets, ADMARC and institutional companies and large-scale wholesalers are associated with high transaction costs in the input markets. The cost of transportation is however higher when farmers purchase inputs from large-scale wholesalers because they are located in urban and peri-urban areas. However, the ‘well-to-do’ have the means of transportation and usually buy inputs from the institutional companies when such inputs are not available at ADMARC markets. In any case, despite ADMARC’s decline in business, it is still the most accessible market for the ‘a-bit-well-to-do’ and the ‘poor’ for agricultural inputs. Compared to crop produce market, private traders and vendors have hardly entered into the provision of inputs.

The other important issue that emerges from the comparative analysis of ADMARC and private traders is the pricing behaviour of different marketing agents. Private traders tend to offer higher producer prices for maize and other cash crops from the perspective of smallholder farmers. However, these higher prices are associated with higher price volatility

104

that create a lot of uncertainty in the decision making process of the farmers. The high volatility of prices is partly due to the lack of countervailing power of smallholder farmers in bargaining for better prices for their produce with individual private traders. The incidence of price discrimination is quite high particularly among vendors. Due to lack of information and the mobility of private traders and the declining business of ADMARC, most private traders behave as discriminating monopsonists in the purchase of maize and other cash crops and as discriminating monopolists in the selling maize to rural households. However, ADMARC plays an important role in benchmark prices through open access to information to smallholder farmers, particularly during the period when ADMARC markets are operation. Private traders offer relatively better prices when ADMARC markets are open latter in the marketing season compared to prices offered earlier in the season before ADMARC markets open for crop purchases.

Another issue of concern in the existing marketing systems in the rural area is the lack of regulation and enforcement of fair trading practices of private traders. The analysis in this chapter has revealed that many private traders engage in cheating practices on measurements and prices, practices that have the potential of reducing farmers’ incomes. While cases of cheating were also reported on some of ADMARC officers, the extent of cheating is widespread among private traders particularly the vendors who take advantage of the illiteracy of many smallholder farmers in the rural areas. The case of the Eswazini (Mzimba) site where few private traders operate, communities feel exploited by vendors and would prefer to sell produce to ADMARC if they were to open their buying service early in the season. If one is to filter the extent of cheating in the prices offered by private traders, it is more likely than not that smallholder farmers actually are no better-off with the high announced prices of private traders than the fixed and stable prices that ADMARC offers.

105

Chapter 5 Stakeholders’ Response To Admarc’s Declining Business And Marketing Liberalization C5.1 Introduction

In the previous chapter we have reviewed the relative roles and service characteristics of ADMARC and private traders in agricultural produce and input marketing services. We noted that ADMARC’s role in crop purchase has been declining and private traders particularly vendors are to some extent bridging the gap as intermediaries servicing large scale traders or directly involved in short-term arbitrage. In the maize sale and input markets, ADMARC is still viewed as an important institution in ensuring food security under the condition that maize and inputs are available. With the declining business of ADMARC, vendors have been active in selling maize to rural households but the input market has largely been abandoned. The few private traders that sell agricultural inputs are less accessible to the poor as they are mainly found in the urban or peri-urban areas.

This chapter presents perceptions of various stakeholders to the declining role of ADMARC and to the liberalization of markets. The reactions are looked at in terms of food security, the selling of produce, the purchasing of inputs and the purchasing of food commodities. The next section presents evidence on the effects of withdrawal of ADMARC markets and general market liberalisation based on information using the cause-effect analysis in six sites with closed seasonal markets and based on the problem analysis of the four sites with active ADMARC markets.8 This will enable us to extrapolate the likely effects of closing further markets from the perspectives of smallholder farmers. Thus, the closure of seasonal markets and the declining ADMARC business may offer some lessons on the likely impact of closing some of the unit markets that are currently accessible to smallholder farmers. It should be noted, however, that smallholder farmers in closed markets still access ADMARC services in nearby unit markets, although business has declined in these markets. Since seasonal market were not selling inputs, some of the experiences articulated in the cause-effect analysis capture the impact of declining ADMARC business at their unit markets. This implies that we can realistically extrapolate the effects of closure of markets from focus group discussions in closed seasonal markets and the problems of marketing experienced in active ADMARC unit markets. Section C5.3 focuses on the reactions by households to food security, changes in produce, food purchases and input markets and the price incentives. Section C5.4 discusses reactions by other stakeholders in the areas of food security and policy-making. Section C5.5 presents concluding remarks. C5.2 The Effects of ADMARC’s Declining Business and Closure of Seasonal Markets

In the six sites where ADMARC closed seasonal markets participants in the focus group discussion were asked to elaborate on the problems that emerged after the closure of markets or due to the declining business at the nearby unit markets. Figure C5.1 presents an analytical framework of the causes and effects of closure of seasonal markets and declining ADMARC business at unit markets from the point of view of rural households and key informants.9 While the cause-effect analysis was only conducted in unit markets in closed seasonal markets sites, similar issued emerged in the problem analysis in the sites with active

8 The unit markets in which we sampled sites in closed seasonal markets are actually among the markets that are being proposed for closure and the farmers perceptions closely reflect what is likely to happen if such unit markets are eventually closed. 9 This framework is based on the views from focus group discussions and reflects the experiences of smallholder farmers during marketing liberalisation including closure of seasonal markets. The lack of positive effects in the cause-effect analysis reflects the relative dominance of problems over the gains that have been brought by marketing liberalisation from the perspective of the communities.

106

ADMARC markets. The cause-effects analysis in closed seasonal markets therefore reflects all aspects of the effects of marketing liberalisation including the declining business activities of ADMARC. The declining business of ADMARC implies that the perceptions of the smallholder farmers can easily be extrapolated to the eventual closure of ADMARC markets.

The withdrawal of ADMARC markets was mainly attributed to the changes in government policy in agricultural marketing while the declining business was mainly attributed to the lack of resources available to ADMARC to purchase crop produce from smallholder farmers. The lack of resources at ADMARC markets led to declining business in unit markets. The immediate effect of closure of a seasonal market or reduction in ADMARC business is the reduction in marketing institutions and hence the structure of the market or lack of access to markets. It was argued that since farmers are unorganised, the few private traders that access such remote rural markets increase their monopsony power in buying produce from smallholder farmers and substantially increase their monopoly power in the sale of maize and inputs.

The concentration of market power or/and the lack of access to markets create five main problems for smallholder farmers. First, the closure of ADMARC seasonal markets and declining business of ADMARC increases the distance to markets particularly for smallholder farmers who still rely on ADMARC to purchase maize and farm inputs. This raises the transaction costs in terms of search and transport costs and opportunity cost of travelling to distant markets. Some groups argued that the long distance to markets is associated with insecurity particularly for women who can easily be robbed of their earnings or commodities. The problem of distance to markets and high transport costs were also identified main marketing problems in sites with active ADMARC markets.

Secondly, the monopsony power of private traders would lead to lower produce prices due to the lower bargaining power of smallholder farmers owing to the absence of marketing cooperatives or associations. The problem of low produce prices is not peculiar to sites in closed seasonal market; it was also reported in all sites with active ADMARC markets. While we find that producer prices offered by private traders are higher than prices offered by ADMARC, the cheating practices by private traders implies that the effective prices received by farmers might be lower than those offered by ADMARC. This means that the farmers in the poor and moderate poor categories, who are usually desperate for cash income, accept the low prices dictated by the private traders.

107

Figure C5.1 Typical Causes and Effects of ADMARC Closure of Seasonal Markets and Declining Business

Thirdly, the lack of access to markets deprives the farmers of a reliable and affordable

source of inputs. In many sites, including the two sites with active ADMARC markets, the declining business of ADMARC is associated with lack of access to inputs through credit programs particularly for the poor farmers who used to get inputs on credit.

Fourthly, lack of access to markets also leads to high prices for inputs in the sense that smallholder farmers have to travel long distances to procure inputs at alternative ADMARC markets or buy inputs from private traders or shop owners at higher prices. This increases the transaction costs to farmers. The problem of high input prices was also evident in all the four sites with an active ADMARC market.

Government Policies

ADMARC’s Lack of Cash

Reductions in Marketing Institutions / Lack of Access to Markets

Long Distance to Markets

High Prices of Food / Maize

High Prices of Inputs

Low Produce Prices

Food Insecurity and Poor Diet

High Transaction Costs

Low Returns from Farming

Low Production (Low Yields)

Poverty

Lack of Inputs

Lack of Basic Needs

Disease and Poor Health

Closure of Seasonal Market and

ADMARC’s Declining Business


108

Finally, smallholder farmers in all ten sites also identified high consumer prices of

maize as one of the effects of ADMARC’s declining business. The argument here is that ADMARC offers better and stable prices for maize throughout the year when it is available and its absence in the area forces the food deficit households to buy maize from private traders at exorbitant prices especially during famine situations like in the 2001/2002 agricultural season. Private traders therefore take advantage of the absence of ADMARC or shortage of maize supplies at ADMARC to charge higher prices than would have been the case if maize were available at ADMARC markets.

The high transaction costs due to long distances to markets combined with the low produce prices imply that the farming becomes less profitable. The low returns from farming combined with high input prices leads to less use of agricultural inputs either because inputs are not readily available or because smallholder farmers can no longer afford to buy the inputs that are adequate for efficient production. As observed in the previous chapter very few private traders sell inputs to smallholder farmers in rural areas and usually operate in urban and peri-urban areas.

Low returns from farming activities imply that rural farming households have low incomes leading to limited access to inputs, difficulties in meeting the basic needs of quality life and provides less incentives for production.10 Lack of production incentives and lack of inputs lead to low production, this consequently leads to high maize prices and food insecurity and poor diet. Food insecurity in turn leads to disease and poor health which in turn may lead to low production due to the limited family labour expended on farming activities. The lack of basic needs and low production lead to a situation of poverty that may in turn lead to disease and poor health, which further constrain the production capacity of rural households and perpetuating food insecurity. C5.3 Households’ Reaction in Closed and Active Markets

In the previous section, we have seen how closure of ADMARC markets and a decline in ADMARC’s business activities affect the well-being and livelihoods of rural households particularly the poor and the ‘a bit-well-to-do’ groups. We look at the adaptive strategies employed by rural households in the sample sites in times of difficulties; reactions in changes in the marketing systems for crop purchases and sales and input markets and their reactions to price changes. Reactions to Food Insecurity and Adaptive Strategies

The ultimate impact of closure of ADMARC markets or its declining business activities in rural areas is increase in food insecurity among the poor, through lack of access to markets, low produce prices and high price of maize (Figure C5.1). The communities in the sample sites have responded to the closure of the markets or ADMARC’s erratic nature of operating by employing a number of strategies to help them maintain their food security status. Table C5.1 shows the ray of activities that households, both in closed markets and where markets are still operating but the business activity declined, engage in so as to be food secure.

Four main strategies were mentioned in all the ten sites. The first of these strategies is selling of individual or family labour in return for food. The second response common to all sites is the abandoning of the cultivation of the crop in instances where ADMARC was the 10 The low incomes may have multiplier effects on the other sectors of the rural economy such as capital constraints for non-farm business activities, low effective demand for farm and non-farm products, low demand for farm and non-farm ‘ganyu’ or formal employment.

109

sole buyer of the crop or was the most trusted buyer. This then meant that households became more food insecure. Cases in point are those of groundnuts in the Lipiri site and cotton in the Mpata and Kakoma sites where most poor people reported to have stopped growing these crops due to the withdrawal of ADMARC. This worsened people’s food security problems because they lost a main source of income that used to supplement maize deficits. Table C5.1 Household Food Security Strategies by Site in the Past Five Years

Site Food Security Strategies

Closed Markets (District)

Gan

yu

Div

ersi

fy

crop

sSa

fety

net

pr

ogra

mm

eA

band

on

crop

Buy

mai

ze

from

priv

ate

td

Eat w

ild

frui

ts, t

uber

s d

Buy

from

ac

ross

the

bd

Hec

tara

ge

expa

nsio

n

Wee

kly

mar

kets

Prod

uce

mar

ket

AD

MA

RC

Nthalire (Chitipa) Υ Υ Υ Υ Υ Υ Υ Υ Eswazini (Mzimba) Υ Υ Υ Υ Υ Υ Υ Chikho (Ntchisi) Υ Υ Υ Υ Υ Υ Lipiri (Dowa) Υ Υ Υ Υ Υ Υ Υ Lundu (Blantyre) Υ Υ Υ Υ Υ Υ Υ Kakoma (Chikwawa) Υ Υ Υ Υ Υ Υ Υ Υ Active Markets Mpata (Karonga) Υ Υ Υ Υ Υ Υ Υ Υ Mikundi (Mchinji) Υ Υ Υ Υ Υ Υ Υ Υ Υ Chiondo (Nkhotakota) Υ Υ Υ Υ Υ Υ Υ Υ Kasongo (Phalombe) Υ Υ Υ Υ Υ Υ Υ Υ Υ Source: Poverty and Social Impact Assessment

On the other extreme, there is a tendency for rural communities to expand the production of a crop (increasing hectarage) that was considered a main staple to cushion themselves against a bad season. Most households expanded hectarage for growing maize since in most places it is the main staple. In times of serious food shortages one of the strategies used is the buying from private traders. As we have noted above, most of these private traders it must be said are small time traders, commonly known as vendors. They move around the villages with a bag or half bag and sell in small quantities. The poor of the poor, it was reported prefer to transact with these because they cannot afford a bag. The problem with these other markets is that maize may not be available all the time at these markets.

Five sites (Nthalire, Kakoma, Mpata, Mikundi and Kasongo) were in districts that bordered neighbouring countries. Households from these sites in cases of food shortages benefited a lot from the cross border trade. Two sites reported the existence of Non-Governmental Organizations (at Mpata there was the joint GTZ/MASAF food for work programme and at Kakoma, CADECOM was providing maize and soya flour rations to households with orphans) that were implementing programmes aimed at improving household food security. Two sites, Nthalire and Mpata also benefited from the weekly markets that take place in these areas. Since the survey was done the year following the famine, in 6 of the 10 sites mention was made of the eating of wild fruits, tubers and grass (especially by the poor of the poorest) simply because the maize was not available. It was reported that some people were eating Nadanga grass, which is a kind of grass that produces tassels that bear grains. The people used to process it into flour. The participants from the group discussion indicated that its consumption had serious side effects:

110

“Mauzu amenewa munthu ukadya umakhala ndi mavuto akulu ukamapanga chimbudzi. Munthu umamva kupweteka kwambiri. Ambiri aife tili ndi mabala kumatako kwathu chifukwa chakumva kupweteka tikamafuna kuchita chimbudzi” (This grass used to cause problems when defecting and people had scars on their anus as a result of eating this grass).

Some people used to eat tubers known as mpira. This was soaked in water for a week

to remove its poison and bitter taste. It was then ground and made into flour.

Some households in six of the ten sites also reacted by diversifying into other crops. This could either be food or cash crops. The crop that most households reported taking on was paprika and the reason behind this move was that this crop does not require the use of fertilizer. Hence, it was cheap to grow and at the same time if market was available, a profitable crop.

Even in instances where an ADMARC market had closed, people still go to ADMARC markets that may be at a distance. The reason behind this is that ADMARC often stocks good quality maize and their prices are fairer compared to private traders. All this serves to underline the importance and value that people, especially the poor in the villages attach to the existence of the ADMARC structures.

Most of the responses and adaptive strategies were equally utilized by both men and women. The eating of wild fruits and, tubers and grass was said to involve mainly women and children. Women often used the weekly markets and produce markets. Men because of the distances and risks involved often did buying from across the border. Reactions to Changes in Food Crop Purchase Markets

Investigations into where people buy crops in cases where they have deficits indicated that the majority buy from vendors (Table C5.2). This is one of the most attractive avenues for rural poor households because vendors move from door to door and they sell in small quantities which are affordable to the poor. The same applies to fellow villagers. People do sell to each other on plates in small quantities and at reasonable prices. ADMARC still features as one of the avenues for purchasing crops because it is said that in normal years ADMARC often has maize. The problem as alluded to earlier is the distance to these places and where ADMARC is still operating the problem is the erratic supply of the commodities.

Private traders and produce markets are some of the prominent avenues for rural households to purchase maize. Unfortunately, private traders are not always in the areas when the people need them most. In the case of produce markets, the very poor find it hard to access them due to distances. Other places where rural households purchase their commodities are the weekly markets that were reported in two sites. These are informal markets that move from place to place. Depending on where the market is held, these are sometimes easily accessible to the poor, but sometimes they could be out of reach. Last but not least, neighbouring countries for border districts provide a good source of some produce.

111

Table C5.2 Sources of Households’ Purchase of Crops by Site in the Past Five Years Site Household’s Responses

Closed Markets (District)

Vendors Villagers ADMARC

Across the

border

Weekly

market

Produce

market

Private Traders

Nthalire (Chitipa) Υ Υ Υ Υ Υ Υ Eswazini (Mzimba) Υ Υ Υ Υ Υ Chikho (Ntchisi) Υ Υ Υ Υ Lipiri (Dowa) Υ Υ Υ Υ Υ Lundu (Blantyre) Υ Υ Υ Υ Υ Kakoma (Chikwawa) Υ Υ Υ Υ Υ Υ Active Markets Mpata (Karonga) Υ Υ Υ Υ Υ Mikundi (Mchinji) Υ Υ Υ Υ Υ Υ Chiondo (Nkhotakota) Υ Υ Υ Υ Υ Kasongo (Phalombe) Υ Υ Υ Υ Υ Υ Source: Poverty and Social Impact Assessment 2002

Both men and women purchased food commodities from ADMARC markets and private traders. It was only men that were reported to go across the border and buy produce for consumption. As stated before, this was mainly attributed to long distances that were involved to get to these places. On the other hand, women mostly used vendors, villagers, weekly and produce markets to acquire the foodstuff for their households. Reactions to Changes in Crop Produce Markets

In the past while as ADMARC provided a ready market for most of the crops that people grew, that ready market is no longer there or it is less reliable because of the decrease in the business activity. For example people at Chikho site, the participants in the men focus group discussion stated that:

“With tobacco and paprika, we do not have readily available traders. Whenever the vendors come and whatever price they offer, we just accept. The reason is that we cannot keep this crop produce in the house as if we are going to consume. We therefore let our produce go at a give away price”.

In this respect, in all the sites visited (closed and not closed), there were four main

outlets for the produce (Table C5.3). These four were: the door-to-door maize buyers (vendors), fellow villagers, small private traders (from outside the community and would buy a bag or two), and ADMARC itself. For areas where ADMARC has withdrawn the ADMARC markets mentioned here are those in distant places. For instance, people at Lipiri had to go to distant places like Nambuma, Mponela and Madisi. All these places are more than 20km away. In other places like Lundu people had problems to sell in distant markets because they got robbed on the way.

112

Table C5.3 Outlets used by Households to Sell their Produce by Site in the Past Five Years

Site Households’ Alternative Produce Markets

Closed Markets (District) V

endo

rs

Vill

ager

s

CA

RG

IL

Priv

ate

trade

rs

Acr

oss

the

bord

er

Inst

itutio

ns

Lim

be

Leaf

Nal

i Ltd

Che

etah

Day

/wee

kly

mar

kets

AD

MA

RC

Prod

uce

Mar

ket

NA

SFA

M

Auc

tion

Hol

ding

s

Nthalire (Chitipa) Υ Υ Υ Υ Υ Υ Υ Eswazini (Mzimba) Υ Υ Υ Υ Υ Υ Υ Υ Chikho (Ntchisi) Υ Υ Υ Υ Υ Lipiri (Dowa) Υ Υ Υ Υ Υ Υ Lundu (Blantyre) Υ Υ Υ Υ Υ Υ Kakoma (Chikwawa) Υ Υ Υ Υ Υ Υ Active Markets Mpata (Karonga) Υ Υ Υ Υ Υ Υ Υ Mikundi (Mchinji) Υ Υ Υ Υ Υ Υ Υ Υ Chiondo (Nkhotakota) Υ Υ Υ Υ Υ Υ Υ Kasongo (Phalombe) Υ Υ Υ Υ Υ Υ Υ Source: Poverty and Social Impact Assessment 2002

In some areas traders came in to purchase special crops. These included CGL in Kakoma and Chiondo, Limbe Leaf Tobacco Company in Eswazini and Cheetah in Eswazini, Kasongo and Mikundi. Then there are people from across the border in Nthalire, Mpata and Mikundi that purchase substantial amounts of produce from these areas. In Mpata and Nthalire as well, institutions like secondary schools are major buyers of some of the produce. At Mikundi, NASFAM was in operation and it helped organize people when selling their produce. Nali Limited, a company that is located in the southern region bought chillies in the Kasongo area.

In some places, weekly markets serve as major outlet for people to sell their crops. Access to these places depends on where they take place and if households can manage to transport their things to these places. Burley tobacco growers mainly sold their produce at the Auction Floors through their organization, the Tobacco Association of Malawi. Last but not least, people also sell at produce markets. The problem though is that most of these places are distant and those without the means to transport find it hard to access these places, particularly for the ‘a-bit-well-to do’ and ‘poor’ groups.

The analysis based on gender of the focus groups revealed that both men and women do the selling of produce to private traders, limited companies/institutions and ADMARC. However, men mainly deal with the specialized markets like CGL and Cheetah. Only in the case of Limbe Leaf Company are women the main stakeholders that deal with the company. This is because the tobacco that Limbe Leaf Company is involved with is in that site grown mainly by women. Women also to a greater extent deal with vendors, villagers, produce markets and the weekly markets. This is because these usually deal with small quantities. At Mpata the non-involvement of men in selling produce at weekly markets was attributed to them being busy herding cattle. In the case of Auction Holdings it deals with both men and women but with a bias towards men because men grow the bulk of the burley tobacco. NASFAM also deals with both men and women. Reactions to Changes in Input Markets

113

The purchase of inputs is a major challenge for the rural households in the advent of

ADMARC withdrawing or reducing its business activity. ADMARC remains the most common source of inputs in all the ten sites (Table C5.4). In closed sites, people reported walking long distances (people from Lundu went to Lirangwe and Lunzu) to buy inputs from ADMARC. This is because ADMARC is reputed to stock good quality inputs. Its prices were also reported to be affordable and people could get fertilizer in loose quantities. Farmers World and Chipiku Stores were identified in four of the 10 sites as sources of marketed inputs. Their depots though are located in major trading centres and they often are beyond the reach of most poor rural households. Other shops like the People’s Trading Centre also stock inputs but just like Farmers World; these are only in major town centres. Hence, only people with some means of transport can get inputs from the sources located in urban or peri-urban centres. Table C5.4 Input Sources used by Households by Site in the Past Five Years

Site Households’ Input Source

Closed Markets (District) A

DM

AR

C

Chi

piku

Farm

ers

Wor

ldA

gora

Oth

er

shop

s Li

mbe

Le

afM

RFC

TAM

A

Acr

oss

the

bord

er

CG

L

Asi

an S

hops

Nthalire (Chitipa) Υ Υ Υ Υ Υ Eswazini (Mzimba) Υ Υ Υ Υ Υ Υ Chikho (Ntchisi) Υ Υ Υ Υ Lipiri (Dowa) Υ Υ Υ Υ Υ Υ Lundu (Blantyre) Υ Υ Υ Υ Υ Υ Kakoma (Chikwawa) Υ Υ Υ Υ Active Markets Mpata (Karonga) Υ Υ Υ Υ Mikundi (Mchinji) Υ Υ Υ Υ Chiondo (Nkhotakota) Υ Υ Υ Υ Kasongo (Phalombe) Υ Υ Υ Υ Υ Υ Source: Poverty and Social Impact Assessment 2002

With the exception of Kasongo, in all sites the poor who cannot afford to buy from ADMARC and other outlets sell their labour to the well to do so that they get seeds for planting in their fields. Thus, fellow villagers act as suppliers of the inputs. At Eswazini, Limbe Leaf Tobacco Company supplies inputs to farmers that grow the type of tobacco that the company is interested in. The same is true of CGL in Kakoma where it provides seeds to cotton growers only. The Tobacco Association of Malawi also helps its members to acquire inputs that they use in growing burley tobacco. Last but not least, households in districts bordering Tanzania, Zambia and Mozambique get some of their inputs from these countries. The inputs were relatively cheaper than those from Malawi. The problem here is the distances to these places. It is only people with oxcarts and bicycles that do access these places and buy in bulk.

Both men and women focus group discussions reported that they accessed inputs from ADMARC and Chipiku Stores. Men often accessed inputs from Farmers World, Agora, and other shops, MRFC, TAMA, across the border and from Asian shops. With the exception of MRFC and TAMA, access to the other sources was a monopoly of the men because of the long distances. In the case of MRFC, the collateral used was a cow, which is often kept by

114

men. In the case of TAMA they interact more with men because most of the burley tobacco growers are men. On the other hand almost exclusively women practiced the acquisition of inputs through ganyu. Most of these inputs were open pollinated varieties of maize seed. Reactions to Changes in Prices

The private marketing system as noted in Chapter 4 is associated with very volatile, unpredictable and discriminating prices of produce and inputs compared to the stable prices that ADMARC offers at its existing markets. ADE (2000) findings provide inconclusive evidence on the impact of maize prices using data between 1989 and 1997, in which they found some positive relationship between price and area planted to maize but also noted that the opposite pattern was evident in several years. Although, we did not inquire directly on the effects of prices, the evidence from the ten sample sites does not suggest that the previous period prices play a critical role in the farming decisions of farmers. In all the sites, the motivation to grow maize among the poor and the ‘a-bit-well-to-do’ was to satisfy their subsistence needs.

With respect to high input prices, we find little evidence that farmers resort to use of manure. Smallholder farmers with small land holdings such as the ‘well-to-do’ and the ‘poor’ who used to apply fertilizers have completely abandoned the technology. The abandonment of the fertilizer technology is particularly compounded by the lack of input credit under which many smaller farmers who could not buy on cash had access to new technology. The application of animal manure particularly for those households with livestock was only mentioned as a coping mechanism to this problem on in Lundu, Chikho and Mikundi sites. C5.4 Reaction by Other Stakeholders

Different view points were also sought from private traders and other stakeholders in the areas of food security and policy making. For private traders, the closure of ADMARC markets and declining business of ADMARC creates more business opportunities. However, the business practices of private traders undermine the gains that the private marketing system promises. In addition, private traders and vendors are mobile and generate uncertainties on the availability of markets. The general view from institutional interviews at policy level was that liberalization of markets without proper development of infrastructure and institutions may not ensure fair trading practices among private traders, and consequently adversely affect the poor.

The institutional interviews raised several concerns about the proposed closure of ADMARC markets and privatisation of ADMARC. First, there was concern on whether the private traders were capable of providing a reliable food distribution network in times of a food crisis given the documented evidence of the inefficiency of the private marketing system and the infrastructure constraints that exist. Secondly, there was concern that closure of markets and privatisation of ADMARC in its totality may not be appropriate instruments to address the problem. Most contend that the problem with ADMARC is the way it manages its resources and the efficiency at the head office level. Others argue that the main problem of ADMARC is not that of agricultural markets, but the troubled and loss making subsidiaries. C5.5 Concluding Remarks

The main stakeholders in agricultural marketing services are smallholder farmers in different categories of well-being. The analysis of the effects of closure of ADMARC markets can be derived from the results of the closure of some of its seasonal markets and the declining ADMARC businesses in all the ten sites. It is apparent that closure of markets, declining ADMARC business activities and liberalisation of markets lead to reductions in marketing institutions or lack of access to markets, both in the short- and medium term. This

115

results into long distances to markets (high transaction costs), low producer prices (through cheating practices), high input prices and lack of inputs and high maize consumer prices. These in turn lead to low production levels and food insecurity, consequently perpetuating poverty in the rural areas.

We have also observed that smallholder farmers and households adopt several strategies to reduce the problems of food insecurity associated with ADMARC’s declining role and liberalisation of markets. The most important strategies include engaging themselves in ganyu to earn income to buy maize or to obtain maize in cases where payment is in-kind. Other strategies include abandoning crops that ADMARC used to purchase that are not purchased by private traders, buying maize from private traders (mainly mobile vendors and those found at produce markets) at exorbitant prices, buying maize from active ADMARC markets with available stocks and expansion of cropping areas for staple foods. The findings also suggest that the presence of an ADMARC market in all the ten sites was perceived as an important repository of food security for rural poor households. This is not only because ADMARC markets used to be closer to the people (in cases where they have closed) or are closer (in places where they are still operational), but also because they often had maize in stock and their prices were fair and uniform. The withdrawal or decrease in the business activity therefore creates a food security problem for most poor households in rural areas because it is hard for them to get maize for consumption at low prices.

The food insecurity problem is further aggravated by the inability of poor households to easily access inputs at reasonable quantities, prices and distance. Most private traders in the input market are located in urban or peri-urban areas, leading to high transaction costs thereby creating problems for the poor and the ‘a-bit-well-to do’ in accessing inputs.

The declining ADMARC’s business, the closure of markets and liberalisation have benefited the well-to-do farmers and the private traders more than the poor and the ‘a-bit-well-to-do’ households in the rural areas. Closure of ADMARC market or/and its declining business has led to the increase in the number and use of private traders in marketing activities particularly in the crop produce marketing and the food marketing activities. The private traders tend to charge high prices taking advantage of the absence of the effectiveness of ADMARC in marketing of agricultural products. Their profit margins tend to be high, through high consumer prices and cheating on measurements and quality of products. Thus, closure of ADMARC markets has created more business opportunities for private traders.

The sentiments of the other stakeholders in the food security and policy area suggests that the real concern is whether the private marketing system as undeveloped and inefficient as it is in Malawi is capable of handling the distribution of food supplies in rural areas during difficult times. In addition, most tend to argue that the ‘ADMARC problem’ has very little to do with its marketing activities, but rather the burden imposed by its loss making subsidiaries and the management at the head office are the issues that have to be sorted out before questioning the efficiency of ADMARC in serving the poor.

116

Chapter 6 Main Findings, Implications And Recommandations

C6.1 Introduction

This study set out to investigate the likely impact of further closure of the proposed ADMARC unprofitable unit markets across Malawi by understanding the current trends and extrapolating reactions of stakeholders in the closed seasonal markets and the declining business of ADMARC activities. We used mainly qualitative and participatory rural appraisal methodologies and focused group discussions with the rural communities in 10 ADMARC unit market sites, of which 6 sites had closed seasonal markets and 4 sites were active markets. These sites were selected on the basis of their remoteness and the livelihood characteristics. Most of the sites had poor road infrastructure and difficult to access by motorized transport particularly during the rainy seasons and most villages were far from the basic infrastructure such as produce market, health facility, post office and police post.

It must be appreciated that this study presents the situations that are typical in the 10 sites (20 villages) included in our sample within the context of a study period that coincided with the famine of 2001/2002 season. It is therefore important to be cautious in generalizing the results to all the 2004 unit markets proposed for closure. The results in this study are indicative of the likely impact of closing of some unit markets with similar remoteness and socio-economic characteristics. Notwithstanding the foregoing, this chapter presents the main findings from the study, the policy implications and recommendations based on data from the sampled communities. C6.2 Main Findings

The following are the main conclusions of the study: (a) Poverty in the sample communities is mainly defined in terms of food security,

housing characteristics, asset possession and access to financial markets. Food security appears to be the most important distinguishing feature of poverty, with the ‘poor’ reported only having food stocks that last within 1 to 2 months after harvesting while the ‘well-to-do’ have adequate food throughout the year. The poor are a larger proportion of the communities and the share of the poor has increased over the last decade.

(b) The business activities of ADMARC in agricultural marketing have been going down

since 1985 and as a result smallholder farmers sell their produce to alternative marketing channels. The market function for which ADMARC has been most affected in recent years is the purchase of produce, particularly maize and tobacco from rural households, in which it has substantially lost the market share in many markets. However, it is difficult to determine the extent to which the declining role of ADMARC in the purchase of agricultural produce can be attributed to the competition from the private traders, the inefficient operations of ADMARC or the specific business strategies implemented by ADMARC. In order to isolate such effects there is need to understand the whole institutional set-up of ADMARC, its financial position and the relative burden exerted by its portfolio investments.

(c) ADMARC markets have an important role as distribution networks for affordable

maize in the lean season and in times of famine, in providing the benchmark prices, in providing a reliable source of inputs and in the purchase of crop produce from farmers as last resort institution in rural Malawi. These roles of ADMARC in the recent famine were minimal and sporadic. ADMARC did not purchase maize from

117

farmers because of liquidity problems and did not sell much maize and inputs to rural households.

(d) Liberalization of agricultural marketing has brought about the proliferation of private

traders in areas where accessibility is not a problem. This is particularly true for the unlicensed retailers and vendors who operate in most of the study sites but the situation for the bigger private traders remain decimal and scanty. The large-scale traders operate very far, usually at trading centres or at the district headquarters.

(e) Over the years private traders have to some extent tended to substitute ADMARC in

the three marketing functions in rural areas with varying degrees and scale of operation. Private traders, particularly vendors, have taken on the role of ADMARC in purchasing crop produce and selling maize both under a normal season and during famine while private traders and large-scale trading limited companies do not evidently participate in the sale of maize to the rural areas. These vendors some times buy maize from ADMARC when available and sell the maize in the remote areas where ADMARC is not available or has no stocks of maize. The large scale wholesalers only participate in the buying of agricultural produce from farmers, and there is no evidence that they do supply the maize to vendors for sale to rural households. The marketing of agricultural inputs is a marketing function that has hardly been infiltrated by private traders. The large scale private traders that compete with ADMARC are usually less accessible to the poor and ‘a-bit-well-to-do’ groups due to their geographic location in urban and peri-urban areas. Vendors do not seem to participate in input marketing. The poor, particularly those that can afford, purchase inputs from distant ADMARC markets whenever such inputs are available on time.

(f) While the private traders, particularly vendors, are substituting ADMARC in some of

the market functions, their business practices are less favourable to smallholder farmers. Private traders tend to announce higher prices for crop produce and their timing of trade early in the season coincides with the cash needs of smallholder farmers. However, the effective producer prices may not be substantially greater than those offered by ADMARC because of the widespread cheating on measurements, their false quality assessment of farmers produce and through high degree of price discrimination partly due to the weak bargaining power of individual farmers. Vendors who operate within a narrow geographic area tend to charge exorbitant maize consumer prices, the margins that cannot be justified given that an average trader conducts trade within a radius of 53 kilometres.

(g) The liberalization of agricultural marketing has benefited more the richer rural

households and vendors compared to the poorer segment of the rural population. This has strengthened the position of the rural elite and ‘well-to-do’ households as the poor tend to do more ganyu as a survival strategy. Thus, an increasing share of poorer farmers are less able to access inputs and staple food through market channels (ADMARC’s unreliable supply, higher prices of private traders, higher transaction cost, no credit schemes) and have increased their reliance on the ‘well-to-do’ for paid labour (ganyu), share cropping arrangements, accessing of farm inputs and transport as survival strategies.

(h) The closure of seasonal markets and ADMARC’s declining business has reduced the

number of market institutions or has created the problem of access to markets by poor households. The closure of markets has increased the transaction costs for farmers particularly in input markets and food procurement, reduction in effective producer prices due to cheating and monopsonistic power, increase in input and maize consumer prices. These problems are likely to result in lower incomes from farming

118

activities, lower production due to lack of inputs and perpetuate food insecurity and poverty in the rural areas.

(i) Rural households and smallholder farmers have reacted to the closure of seasonal

markets and declining role of business in different ways to sustain their livelihoods and food security. The smallholder farmers are increasingly relying on private traders, particularly vendors, in the produce market and maize procurement. The poor households resort more to ganyu to earn income or to obtain maize as in-kind payment. The procurement of maize from vendors is a second best strategy and people prefer buying maize from ADMARC when available because of fair, non-discriminatory and stable prices relative to the exorbitant prices charged by vendors. Another strategy employed by rural households to improve food security is to expand cropping areas for the main staple food and to diversify towards some drought resistant crops.

(j) There is no evidence in the study that the decision to grow particular food crops such

as maize is motivated by the price information available to farmers in the previous season. Although, prices of maize have risen, maize (particularly among the poor and ‘a-bit-well-to-do’) is grown mainly to meet the subsistence needs. The problem with the private marketing system is that prices are unpredictable and volatile and based on information from the communities individual private traders tend to act as discriminating monopsonists.

(k) The liberalization of markets and the declining business of ADMARC have created

more opportunities for private traders and particularly vendors through creation of self-employment and potentially higher profit margins from arbitrage opportunities and price discrimination behaviour of private traders and vendors. The absence of effective ADMARC operations has enabled private traders or vendors to extract monopoly rents from poor households who sell the little they have early in the season and buy more of the food requirements in the lean season.

(l) The reactions to liberalization and declining ADMARC markets from other

stakeholders in the food security and policy making institutions has been more cautious. Most policy makers believe that the problem of ADMARC does not lie in the operations of rural market, but rather at the institutional structure and the efficiency of the head office and the burden of unprofitable subsidiaries and other portfolio investments.

(m) Since the private market is dominated by small-scale local vendors, the function of

inter-regional and inter-seasonal arbitrage which is performed by ADMARC could not effectively be taken over by another actor. The reduction in unit markets is likely to affect the extent to which ADMARC could market maize to deficit areas. This is particularly important in times of food shortage in one region which could be compensated by the surpluses of another region, if there were a functioning national market.

C6.3 Implications

Although the period of this study coincided with the famine and the fact that only ten markets (20 villages) were sampled, it is very likely that the problems that are experienced in sites where ADMARC seasonal markets were closed and the declining business of ADMARC in markets to which communities switched to after closure of seasonal markets will be replicated in the proposed closure of the 204 unit markets. The poor, who are a large proportion of the communities in the proposed closed markets, are likely to be the main losers while the rich and vendors will continue to extract the undue surplus from the poor. The

119

proposed closure of ADMARC markets may have the following implications in areas that have similar socio-economic and remoteness characteristics as our sample sites: (a) Many smallholder farmers, particularly the ‘poor’ and the ‘a-bit-well-to-do’, will

have to pay more to access the same amount of maize to private traders/vendors without ADMARC than with ADMARC markets in the short and medium terms. In addition, transaction costs due to travelling longer distances to alternative active markets and search costs for suppliers of maize during the lean periods will increase.

(b) The problem of input availability, combined with the high cost of inputs from the

private traders and the high transportation costs in the production of maize and other food products will lead to a decline in the utilization of fertilizers and improved seeds leading to lower productivity and cultivation of marginal land by many smallholder farmers.

(c) The closure of ADMARC markets will remove the benchmarking pricing role that

ADMARC currently plays in the rural areas in which most private traders already act as discriminating monopolist. There may be a reduction in the effective producer prices offered by private traders and vendors resulting in strengthened countervailing power of private traders vis-à-vis the individual poor farmer.

(d) Many smallholder producers are likely to lose incomes through cheating practices and

monopsonistic power of private traders/vendors if appropriate institutions and regulations are not put in place to protect the interests of farmers and consumers.

(e) The risk of food insecurity for many rural households will increase even in normal

years due to the behaviour of large scale private traders of only purchasing maize from the rural areas and not reselling the maize in the lean period. In addition, the risk of food insecurity will also be perpetuated because some of the vendors who buy maize from ADMARC unit markets for sale to the remote areas may find difficult to do so or will experience higher transaction costs in accessing distant active ADMARC unit markets.

(f) With the poor infrastructure in many rural markets the further increase in the number

of private traders will be marginal and the mobility of private traders is less likely to create competitive trading activities necessary to discipline private traders.

(g) The closure of markets may aggravate food insecurity situation of rural communities

in times of hunger and famine because relief food is likely to move slowly to the areas due to the poor road infrastructure and due to lack of storage facilities and lack of a reliable distribution network.

(h) The producer and consumer price volatility will remain high, and this will work to the

advantage of the private traders who can generate arbitrage and monopsony or monopoly profits from price variability but smallholder farmers will be disadvantaged with the unpredictable prices continuing to play a very little role in the farmers planning decisions for food crops. Consumer food prices are likely to increase in the short-run, but the extent to which prices will fall with closure of some markets in the medium or long term will depend on the competitiveness of the private food marketing system.

C6.4 Recommendations

The marketing services of ADMARC in the rural areas tend to benefit the rural population in Malawi and it is not surprising that rural communities have recently linked the

120

increase in poverty to some extent to the closure of ADMARC markets (NEC, 2002a). While the motivation for privatising ADMARC is mainly from a profit and efficiency point of view, those who advocate private market systems need to understand the activities of private traders in closed markets or in areas that are far from ADMARC unit markets and consider the plight of the poor. The necessity of maintaining a reliable distribution network should be viewed in the context of the recurrent droughts and famines in Southern Africa and Malawi in particular. Liberalisation of markets is not necessarily a bad strategy for agricultural markets. However, it is important to ensure that a conducive environment exists for transition from a state-dominated marketing system to a private sector led marketing system. If the government is to proceed with the proposed closure of ADMARC markets, it is important to condition such a process on the development of infrastructure and appropriate institutions to facilitate competitive private sector activities and increase the bargaining power of smallholder farmers. We make the following specific recommendations based on the results of the study: (a) The government, donors, multilateral institutions and ADMARC should be careful in

advocating the proposed closure of 204 ADMARC markets by taking into account the framework of poverty alleviation and the state of supporting infrastructures and the need to have appropriate food distribution network in the frequently occurring droughts or food crises in Malawi.

(b) Concrete decisions to close ADMARC markets or to privatise ADMARC should be

done on case by case basis, supported by sound empirical studies on how markets function that involve extensive consultations with the main stakeholders, the smallholder farmers and rural poor households that attach more value to the marketing services provided by ADMARC. More in-depth research on the capacity of the private sector in performing these market functions is necessary for informed decision making or policies.

(c) In cases where closure of markets is unavoidable, the government and ADMARC

should ensure that appropriate institutions and infrastructure are in place before such markets are closed. The government should ensure that the proposed closed markets are linked through good all-weather roads to facilitate the movement of private traders and better communication facilities.

(d) There is need to strengthen the organization of farmers into marketing co-operatives

or marketing associations or farmers price and business practices monitoring committees to promote good business practices and provide countervailing power to private traders.

(e) There should be a comprehensive review of ADMARC as an institution, particularly

focusing on the management structure and management efficiency and the role of its commercial subsidiaries in the performance of its marketing functions, its mandate and the viability of the dual social and economic functions and the extent to which political interference can be minimized or eliminated.

(f) In the short or medium term the social and developmental functions of ADMARC

will still be desirable because some of the social functions it performs can not be substituted through the private marketing system in the short-run. These functions include food security and maize distribution, input supply and benchmark pricing or price stabilization in favour of producers and consumers in the rural areas. However, there is the need to spell out clearly and explicitly the so-called ‘social function’ of ADMARC. The costs related to such a social function need to be made transparent in order to allow for a sensible discussion of how these social costs should be financed through cross-subsidization across regions or business lines, national budget, consumer prices. For instance, the social function of ADMARC could imply that

121

ADMARC should only withdraw from markets where the private marketing system is working efficiently and concentrate its activities in areas where the private marketing system is not working efficiently through organization of farmer co-operatives or community managed commodity exchanges facilitated by ADMARC within an agreed time frame.

(g) If the long-term objective in agricultural marketing is the development of the private

marketing system, the government needs to properly sequence the privatization of ADMARC. The first step in such a sequence is for ADMARC to be restructured to exclusively concentrate on its marketing function and to be allowed to operate commercially in all crops produced by smallholder farmers and act as an effective competitor to private traders. In the short-term it should continue to perform its social function with respect to food security and development of alternative community-based marketing institutions, but it should be given adequate flexibility (reasonable margin) in the pricing of maize. This implies that ADMARC should divest ownership in all enterprises. The second step should be the privatization of ADMARC unit markets infrastructure in areas where the private sector is actively substituting the marketing services of ADMARC. Concurrently, ADMARC should embark on a ‘develop, operate and privatise (DOP)’ programme of markets that are less accessible by private traders or less serviced by private traders.

(h) If the main reason for closing the proposed unit markets is to cut costs (particularly

staff costs) in the short-term, other alternative ways should be explored giving each market some years of observation before such a market is closed subject to the proposed privatization sequence. Short-term cost cutting measures would include organising some unit markets around one core unit market and the surrounding unit markets could open on specific days similar to the ‘salaula’ and ‘tiwonewone’ concept. Three markets could be operated by one officer on bi-weekly basis, but the frequency of opening such markets could be increased during times of famine.

(i) In circumstances that ADMARC has excess storage space such space should be sub-

let at commercial fees to private traders operating in the areas in the short-run. This would promote the activities of private traders and vendors while at the same time ensuring that the poor have access to alternative market channels.

(j) Under each scenario, whether ADMARC is to withdraw from the proposed 204

markets or not, critical measures would be required to enable ADMARC to effectively operate and fulfil the marketing functions. These measures comprise (1) Privatising ADMARC’s subsidiaries and divesting its share ownership in

commercial enterprises. (2) Strengthening of ADMARC’s operational performance and financial viability

through restructuring of management. (3) Responding to the needs of the poor by opening markets early in the season

and providing the inputs on time and repackaging inputs into smaller quantities where possible.

(k) Regardless of what will happen to the 204 ADMARC markets proposed for closure,

measures of an active market development would be required in future to improve the competition, market structure and regulation. This should comprise (1) measures to improve market and price information, (2) market regulation and supervision functions, (3) improved access to credit and assets to encourage the private sector to engage

in input marketing and storage, and (4) promotion of producer associations or marketing cooperatives to increase the

bargaining power of rural producers.

122

While this study has shed some light on the likely implications of the proposed

closure of 204 ADMARC markets, it is only based on a sample of 10 markets, six of which are proposed for closure. In addition, the study focussed on specific issues and did not examine the operations of ADMARC and private traders in a holistic manner. The foregoing imply that there are several areas of further research in understanding the performance of various institutions and the degree to which competitive market outcomes could be expected in agricultural marketing in Malawi. First, there is need to undertake an empirical analysis of each of the 204 ADMARC unit and seasonal markets proposed for closure in order to identify the gaps that would not be filled by private traders and non-governmental organisations. Secondly, there is need to review and evaluate the institutional framework, management structure and decision-making processes, business strategy and operational efficiency of ADMARC.

123

References Abbott, R. D. and Poulin, R. J. (1996) The Future of ADMARC: A Policy Analysis,

Agricultural Policy Analysis Project (APAP) Research Report No. 1012 (USAID) ADE (2000) Malawi: Inception Technical Assistance to the National Food Reserve Agency,

Draft Report ADMARC (1974) ADMARC: 10 Years of Progress, 10th Anniversary of Independence,

Blantyre: ADMARC ADMARC (1990) ADMARC Annual Report and Accounts 1990, Blantyre: ADMARC Anderson, D. and Leiserson, M.W. (1979) ‘Rural Nonfarm Employment in Developing

Countries’, Economic Development and Change, Vol. 28: 227-245. Chirwa, E. W. (1998) ‘Fostering Private Food Marketing and Food Policies after

Liberalisation – The Case of Malawi’, in Seppala, P. (ed.) Liberalized and Neglected? Food Marketing Policies in Eastern Africa, Helsink: United Nations University/WIDER

Chirwa, E. W. (2000) Privatization and Economic Efficiency in Malawi Manufacturing:

Mixed Enterprises in Oligopolistic Industries, Unpublished PhD Thesis (University of East Anglia, Norwich)

Economic Planning and Development (EPD) (1990) Economic Report 1990, Zomba:

Government Printer Fafchamps, M. and Gabre-Madhin, E. (2001) Agricultural Markets in Benin and Malawi:

Operation and Performance of Traders, World Bank Working Paper Series No. 2734 (Washington, D.C)

Food Studies Group (FGS) (1991) The Impact of Pricing and Marketing Policies on the

Development of Private Sector Trading, A Report on the Study of the Impact of the Liberalization of Smallholder Agricultural Produce Trade in Malawi, International Development Centre, University of Oxford

Government of Malawi (GoM) (2002a) Malawi Poverty Reduction Strategy Paper,

Lilongwe: Ministry of Finance and Economic Planning Government of Malawi (GoM) (2002b) Malawi – Letter of Intent, Memorandum of Economic

Policy and Technical Memorandum of Understanding, Lilongwe: Ministry of Finance Harrigan, J. (1988) Malawi: the Impact of Pricing Policy on Smallholder Agriculture, 1971-

1988, Development Policy Review, 6 Harrigan, J. (1991) Malawi, in P. Mosley, J. Harrigan and J. Toye (eds.) Aid and Power: The

World Bank and Policy Based Lending, Volume 2 Case Studies, London: Routledge

Kaluwa, B. (1991) Study of the Impact of the Liberalization of Smallholder Agricultural

Produce Trade in Malawi, Draft Final Report prepared for Food Studies Group, International Development Centre, University of Oxford

124

Kaluwa, B. M. (1992) ‘Malawi Food Marketing: Private Trader Operation and state Intervention’, in J. B. Wyckoff and M. Rukuni (eds) Food Security Research in Southern Africa: Policy Implications, proceedings of the Sixth Annual Conference on Food Security Research in Southern Africa, University of Zimbabwe/ Michigan state University Food Security Research Project.

Kandoole, B. F., Kaluwa, B. M. and Buccola, S. (1988) Market Liberalization and Food

Security in Malawi, in M. Rukuni and R. H. Bernsten (eds.) Proceedings of the Third Annual Conference on Food Security Research in Southern Africa, University of Zimbabwe/Michigan State University Food Security Research Project.

Khaila S., Mvula P. and Kadzandira J. (1999) ‘Consultations with the Poor – Country Report

for Malawi’. A qualitative study done on behalf of the World Bank, Poverty Reduction Group, Washington D.C.

Kydd, J. and Christiansen, R. (1982) Structural Change in Malawi since Independence:

Consequences of a Development Strategy based on Large-Scale Agriculture, World Development, 10 (5), 355-75

M&O Associates Limited (2001) Identification of Alternative Uses of ADMARC’s Excess

Market Infrastructure, Final Report prepared for the Ministry of Finance and Economic Planning

Mthindi, G.B., Chilowa, W. R., Milner, J. A. G. and Ntata, P. (1999) Social Policy in the

Context of Economic Reforms in Malawi: The Survival and Adaptive Strategy of Vulnerable Groups, First, Second and Third Monitoring Survey Reports, Southern African Regional Institute for Policy Studies.

Mkwezalamba, M. (1989) ‘The Impact of Liberalisation of Smallholder Agricultural Produce

Pricing and Marketing in Malawi’, Report submitted to Ministry of Agriculture, Lilongwe

National Economic Council (NEC) (2002a) Qualitative Impact Monitoring (QIM) of the

Poverty Alleviation Policies and Programmes in Malawi Volume 1: Study Findings, Lilongwe: National Economic Council

National Economic Council (NEC) (2002b) Economic Report, Budget Document No. 2,

Zomba: Government Printer National Statistical Office (NSO) (2000) The Poverty Analysis of the Integrated Household

Survey 1997-98 - The State of Malawi’s Poor: The Incidence, Depth, and Severity of Poverty. Available at http://www.nso.malawi.net

Nthara, K. (2002) What Needs to Be Done to Improve the Impact of ADMARC on the Poor,

Phase II Final Report prepared for the Joint Oxfam Programme in Malawi, Blantyre. Scarborough, V. (1990) Domestic Food Marketing Liberalization in Malawi: A Preliminary

Assessment, ADU Occasional Paper 13, London: Wye College, University of London

http://www.nso.malawi.net/

125

Documents

Poverty and Social Impact Assessment - World Banksiteresources.worldbank.org/.../15009_malawi_admarc_psia.pdf · 2010-04-01 · ii Acknowledgements This report represents the outcome