Jump to ContentJump to Main Navigation
Agricultural Input SubsidiesThe Recent Malawi Experience$

Ephraim Chirwa and Andrew Dorward

Print publication date: 2013

Print ISBN-13: 9780199683529

Published to Oxford Scholarship Online: January 2014

DOI: 10.1093/acprof:oso/9780199683529.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2019. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see www.oxfordscholarship.com/page/privacy-policy).date: 18 June 2019

Introduction

Introduction

Chapter:
(p.1) 1 Introduction
Source:
Agricultural Input Subsidies
Author(s):

Ephraim Chirwa

Andrew Dorward

Publisher:
Oxford University Press
DOI:10.1093/acprof:oso/9780199683529.003.0001

Abstract and Keywords

This chapter sets out the objectives and contents of the book as a whole. Despite increased economic growth in many African countries, high rates of rural poverty continue, associated with low productivity in smallholder staple crops with very low fertilizer use. Higher than average fertilizer use in some countries is associated with large-scale agricultural input subsidy programmes. These are being adopted by a number of countries, but are controversial and new approaches have been the subject of limited study. Malawi’s experience with a large-scale programme since 2005 offers insights into the benefits and difficulties of such programmes. The book therefore aims to develop understanding of new forms of agricultural input subsidy programmes, to derive insights into this from Malawi’s experience, and to promote more informed debate on these issues. The chapter concludes with an outline of the book’s contents and of the data and analytical methods on which it draws.

Keywords:   input subsidies, agricultural development, Africa, Malawi, fertilisers

1.1. Background: challenges in African agricultural development

In the first decade of the twenty-first century sub-Saharan Africa moved, in the words of The Economist, from being ‘hopeless Africa’ to ‘the hopeful continent’.1 Economic growth indicators from 2000 to 2010 show an impressive recovery from the poor performance of previous decades, although growth rates vary considerably between regions and countries within Africa, with West and East African coastal countries growing faster than other regions (United Nations Economic Commission for Africa (UNECA) and African Union, 2012).

There are, however, concerns about the quality of economic growth in Africa. The United Nations Economic Commission for Africa (UNECA) and African Union (2012) note that Africa has witnessed jobless growth due partly to the fact that most of the growth has occurred in capital intensive extractive sectors with limited forward and backward linkages to the local economies. This is associated with high levels of unemployment and underemployment, particularly among the youth, with most of the youth trapped in less productive informal sectors. At the same time, not much structural transformation in African economies has taken place, implying that a large proportion of African people still depend on agriculture as a source of livelihood. These observations underlie concerns about growing inequity within African economies, and continuing high levels of poverty and particularly rural poverty (Africa Progress Panel, 2012). There are also major concerns about food insecurity in Africa, with adverse welfare and developmental effects of high national and international food prices on both the urban and rural poor, (p.2) despite the offsetting effects of economic growth (Headey, 2011b; Dorward, 2012b, 2013; Verpoorten et al., 2012). These concerns link into a set of longstanding but resurgent debates and controversies about agricultural development in Africa regarding

  • the importance and role of agriculture in development;

  • the extent and causes of African agriculture’s poor historic performance;

  • the relative advantages and disadvantages of large and small farms in agricultural development; and

  • the best means of promoting agricultural development in Africa.

Emphasis on agriculture as a critical sector for development has fluctuated over the last 50 years or so. At independence, most developing country governments saw agriculture as either a driver of growth in their economies or as a foreign exchange earner, with a large reserve of unutilized labour to be taxed to support industrial development for a modern economy. With a weak and/or mistrusted private sector, this led to large public investments in agricultural development. In many African countries these large investments were either ineffective or, where they were effective, very expensive and—in the context of over-extension of government budgets and activities—unsustainable without donor support. This was not forthcoming with both the emerging Washington Consensus (promoting structural adjustment and market liberalization) and disenchantment with agricultural investments—which were seen as ineffective and unnecessary in the context of increasing global food production and falling prices (although paradoxically these were partly the result of large and highly successful public investments in the Asian Green Revolution). Emphasis on agriculture was further undermined by empirical studies that revealed the widespread importance of non-farm incomes in the livelihoods of rural farming households (for example, Ellis, 2000; Haggblade et al., 2007b; Reardon, 1998).

With time, however, the pendulum began to swing back as a result of further empirical work showing the importance of the agricultural sector to poor rural economies and to the livelihoods of poor people within those economies (see, for example, Datt and Ravallion, 1996; Mellor, 2000; Thirtle et al., 2003; de Janvry and Sadoulet, 2010; Christiaensen et al., 2011). This coincided with growing concern among governments and donors about the lack of growth in African agriculture (particularly in staple crop production). In 2008 the World Development Report made a powerful case for the importance of agriculture in poverty reduction (World Bank, 2007b) and this was brought home by the 2008 global food price spike and recognition that the era of low and stable food prices was over, if it ever existed (Dorward, 2011, 2013). (p.3)

Table 1.1. Annual changes in cereal production from 1961 and 2000

East Asia & Pacific

Latin America & Caribbean

South Asia

Sub-Saharan Africa

Cereal prodn

1961–2009

2.80%

2.58%

2.27%

2.45%

Cereal land

1961–2009

0.39%

0.61%

0.31%

1.73%

Cereal yield

1961–2009

2.40%

1.95%

1.95%

0.95%

Cereal prodn

2000–09

1.93%

1.78%

1.01%

3.65%

Cereal land

2000–09

0.72%

0.41%

-0.01%

2.20%

Cereal yield

2000–09

1.20%

1.37%

1.01%

1.42%

Source: Author calculations from World Bank (2011).

As noted above, resurgent interest in agricultural development was in part stimulated by African agriculture’s poor performance and associated problems of food insecurity, lack of rural growth, and persistent rural poverty. Cereal production faced particular difficulties in that, while production grew (just keeping pace with population growth), most of this growth was the result of expansion in cereal areas, with very limited increases in yields. Table 1.1 shows that in contrast to Latin America and Asia, annual average growth in land under cereals was higher than growth in yields in sub-Saharan Africa from 1961 to 2009, and this applied both before and after 2000, though with higher rates of growth in both cereal areas and yields after 2000.

Wiggins and Leturque (2010) provide a helpful summary of different explanations for sub-Saharan Africa’s poor agricultural performance, while pointing to considerable variation in performance between regions within Africa. They identify core problems as limited production potential (due to geography, environmental degradation, and fertility decline which they link to lack of technical innovation), unfavourable external conditions (arising from OECD subsidies and trade rules and from limited demand for farm output), and government and market failures (the former involving a policy that deters investors and too little investment, the latter failing to deliver credit and input services and overcome poverty traps). These difficulties are of course interrelated. However, the lack of technical innovation is arguably the proximate cause of the lack of land and labour productivity growth in African agriculture and is the outcome of other difficulties—which reduce benefits, raise costs, or in other ways inhibit technical change—particularly on poor, small-scale farms.2 (p.4)

Table 1.2. Fertilizer use, cereal yields, and value of cereal production, 2002–9

Nitrogen application, kg/ha

Cereal yield, kg/ha

Value of cereal prodn as % agric. value added

Asia

106.0

3404

Northern America

58.8

5723

Europe

44.2

3563

Central America

38.6

2967

South America

36.6

3447

Northern Africa

37.8

1852

Sub Saharan Africa

5.9

1274

23%

Mauritius

96.3

0%

South Africa

27.2

41%

Malawi

22.8

55%

Zambia

17.6

18%

Zimbabwe

15.1

27%

Kenya

12.3

13%

Sources: Author calculations from FAO (2012), World Bank (2011), World Bank (2012).

A critical and widely recognized difference between agriculture in sub-Saharan Africa and in other regions is the low rate of fertilizer use in sub-Saharan Africa. The extent of this is illustrated in the upper part of Table 1.2, which shows FAO estimates of mean rates of inorganic fertilizer application (measured in kg nitrogen per ha arable and permanent crop land) from 2002–9 in different regions of the world. Even allowing for the difficulties of gathering and interpreting such data, the table shows a striking contrast between sub-Saharan Africa and other parts of the world. Although significant amounts of fertilizer use is for non-cereal crops (and the importance of this varies between countries), a similar contrast is evident for cereal yields.

The lower part of Table 1.2 shows estimated nitrogen application rates for the six sub-Saharan African countries with the highest fertilizer rates. It also provides an indicator of the importance of cereal production in the agricultural sector.3 Mauritius has a very high rate of fertilizer use but negligible cereals production, so the high rates of fertilizer use are largely on other crops. Fertilizer use in South Africa is spread across both cereals and other crops but is affected by substantial maize production by the large-scale commercial sector.4 Malawi then stands out for its relatively high rate of fertilizer use (by African but not global standards), large share of cereals in the agricultural (p.5) sector, and (not shown in Table 1.2) the overwhelming importance of smallholder agriculture in cereal production.5 Many would argue that a major factor in Malawi’s high rate of fertilizer use in a poor and smallholder maize-based agricultural economy has been its longstanding use of agricultural fertilizer subsidies. In Table 1.2, data for Zambia, Zimbabwe, and Kenya are presented below data for Malawi. Smallholder cereal production accounts for a smaller share of agriculture in all three of these economies as compared with Malawi. Zambia has, however, also been implementing an agricultural input subsidy programme. The basis for the relatively high rates of fertilizer use reported for Zimbabwe is not clear, but contributors to higher fertilizer use in Kenya without subsidies have been explored by Minde et al. (2008) and Ariga and Jayne (2011) and these include: good transport links to and through Mombasa, high export volumes reducing back-load costs, and high fertilizer demand for use on smallholder cash crops alongside food crops (stimulating market development and lowering retail unit costs as well as supporting a variety of mechanisms for easing cash flow constraints on purchases of fertilizer for food crops).

In this book we examine the often controversial roles and impacts of agricultural input subsidies (generally dominated by fertilizer subsidies) in promoting technical change in agricultural development, with particular attention to lessons and insights from the large agricultural input subsidy programme which Malawi embarked on in 2005.

The topics addressed are important for many countries in sub-Saharan Africa, as well as for Malawi. As discussed above, agricultural production has been stagnant in many parts of sub-Saharan Africa, and associated with high incidence and severity of rural poverty and food insecurity. The challenges in ‘getting agriculture moving’ are exacerbated by local resource pressures from rapid population growth, the threat of climate change leading to increasingly uncertain rainfall in many parts of the region, high and volatile world food prices, and uncertainties about the global economy. As we shall discuss, the number of African countries implementing large-scale agricultural input subsidies has been growing, and these programmes are costly—in terms of fiscal costs, lost benefits from investments of these resources in alternative uses (such as in education, health, infrastructure, or agricultural research), and the long term distortions they can foster in political, financial, social, and economic structures. Failure will not only blight the lives of millions of poor rural people and their children, it may also prejudice policy makers against future investments in agriculture. (p.6)

Malawi is, unfortunately, no exception to this. However, its post-2005 subsidy experience provides a good case study for examining the potential strengths and weaknesses of agricultural input subsidies in addressing these issues. The programme follows and builds on a long history of different forms of subsidy in Malawi, with fertilizer price subsidies to smallholder farmers from the 1960s to 1980s, which were then removed and reinstated in the 1990s, and then replaced by an initially universal Starter Pack but later Targeted Input Programme (TIP) of free distribution of small fertilizer and seed packs to smallholder farmers. This programme (thoroughly documented in Levy, 2005) adopted increasingly innovative systems involving private distributors of seed and fertilizer and was continued until the 2004/5 season. Following the 2004 elections and food shortages in 2005, however, the increasing political significance of fertilizer subsidies led to the introduction of a much larger programme providing approximately 50% of Malawian smallholder farmers with much larger packs of inputs at highly subsidized prices. This new programme, the Agricultural or Farm Input Subsidy Programme (AISP or FISP),6 attracted immediate controversy, from both supporters (for example, Dugger, 2007; Denning et al., 2009) and detractors (for example, The Economist, 2008) but was very popular in Malawi and has since continued.7 The programme has been held up as an example for other countries to follow, and large-scale input subsidies are now being implemented in a large number of African countries. Many of these are both drawing on Malawi’s experience and introducing their own innovations to address perceived opportunities for improvement. There are, however, also significant concerns among many economists, development analysts, and policy makers both in Malawi and elsewhere about the effects and cost of Malawi’s programme.

The FISP has also been the subject of a range of different studies, of varying scope and quality, and advocates and sceptics, supporters and detractors of the programme often draw on contradictory evidence to support their positions.8 There is therefore a need to bring these different perspectives and studies together and to set these in the context of wider debates and experience (p.7) to draw robust conclusions where these are possible, to recognize areas of disagreement and ambiguity, and to identify outstanding questions for agricultural policy makers, programme implementers, and researchers not only in Malawi but across Africa. That is the purpose of this book.

1.2. Objectives and outline

In this book we aim to contribute to a greater understanding of the roles, contributions, and pitfalls of agricultural input subsidies as instruments for promoting food security, poverty reduction, social protection, and wider economic growth in poor agrarian economies. The specific objectives are

  • to update and develop theoretical understanding of agricultural input subsidies’ impacts, allowing for new delivery systems and instruments and specific constraints inhibiting the livelihoods of poor subsistence farmers and the economies of which they are a major part;

  • to derive from Malawi’s experience lessons about the implementation and impacts of a large-scale agricultural input subsidy programme, with specific focus on the contextual, design, and implementation determinants of economy-wide, beneficiary, and market impacts; and

  • to promote debate about strategic policy decisions in the design of large-scale agricultural input subsidies in contemporary low income agrarian economies, including targeting and graduation, to foster their sustainable contribution to agricultural development and poverty reduction.

In order to achieve these objectives the book is divided into three parts following this introduction. The first part provides the theoretical and empirical context for the rest of the book. It is consists of three chapters. Chapter 2 sets out the longer standing empirical evidence and theories on the roles of agricultural input subsidies in poor agrarian economies. It then extends conventional theories to provide a richer account of the potential contributions of innovative delivery systems and instruments to microeconomic, mesoeconomic, and economy-wide processes promoting poverty reduction, food security, economic diversification, and wider economic growth. This provides the basis for a broad understanding of the potential roles and impacts, positive and negative, of a large-scale subsidy programme in poor agrarian economies with different characteristics. Chapter 3 follows with a review of the limited information available on twenty-first century agricultural input subsidy programmes in sub-Saharan Africa—but leaves to later chapters any discussion of Malawi’s post-2005 programme. It identifies a (p.8) number of commonalities across different programmes, against which the Malawi programme is compared in later chapters. Chapter 4 completes the first part of the book with a review of Malawi’s political, livelihood, market, and agricultural policy history.

The second part of the book draws on panel household surveys, market surveys, monitoring and implementation reports, close engagement with a range of stakeholders, and the authors’ detailed studies of the Malawi subsidy programme from 2006/7 to 2011/12. Chapter 5 describes in detail its evolving implementation. Chapters 6, 7, and 8 analyse various potential impacts—direct impacts on beneficiaries and on production, indirect impacts on the wider economy, and direct and indirect impacts on input markets. Returns to investment are considered in Chapter 9.

The final part of the book examines two major issues that emerge from Malawi’s recent subsidy experience, focusing on targeting (in Chapter 10) and graduation (in Chapter 11). The concluding chapter summarizes the main arguments and evidence presented in Chapters 2 and 3, draws out the major lessons from the Malawi experience, and considers the question of agro-ecological, fiscal, and political sustainability. It concludes with a brief discussion of possible ways forward for agricultural input subsidies in sub-Saharan Africa.

Although parts of the book are written from an economist’s perspective, most of the book should be of much wider interest, addressing general policy and implementation issues concerned both with agricultural input subsidies and wider problems of development in poor agrarian economies. There is also explicit consideration of the political influences on policy and its implementation: these considerations have wider relevance beyond policies concerned only with input subsidies.

1.3. Data and methods

We conclude this introductory chapter with a brief discussion on the main sources of information used in analysis of the Malawi subsidy programme. We draw on four main sources of information:

  • implementation records on the subsidy programme;

  • household and input supplier surveys conducted in 2006/7, 2008/9, and 2010/11 as part of the evaluation of the programme;

  • official statistics;

  • other studies on the subsidy programme.

We discuss each of these in turn. (p.9)

1.3.1. Implementation records on the Malawi subsidy programme

Since 2006/7 the logistics of subsidized fertilizer distribution and payments to fertilizer suppliers, fertilizer transporters, and seed suppliers have been managed by the Logistics Unit, working in close cooperation with the Ministry of Agriculture and Food Security (MoAFS), donors, the two parastatals involved in subsidized fertilizer and seed distribution (Agricultural Marketing and Development Corporation and Smallholder Farmer Fertilizer Revolving Fund of Malawi, ADMARC and SFFRFM), and contracted transporters and seed and fertilizer suppliers. In Chapter 5 extensive use is made of information from the Logistics Unit’s weekly and annual reports, supported by minutes of weekly task force meetings and information supplied directly by the MoAFS.

1.3.2. Programme evaluation studies

Much of the information and analysis on implementation is also contained in various reports of FISP evaluations led by the authors (for example, SOAS et al., 2008; Dorward et al., 2010; Dorward and Chirwa, 2011). Since 2006/7 the authors have led annual evaluations of the subsidy programme, with more intensive and ‘light touch’ evaluations in alternate years. More intensive evaluations of the 2006/7, 2008/9, and 2010/11 programmes involved household surveys with focus group discussions and a community survey and in 2006/7 and 2008/9 an input supplier survey. ‘Light touch’ evaluations of the 2007/8, 2009/10, and 2011/12 programme years have drawn mainly on implementation records as outlined above, together with information from other studies and official statistics, and analysis of data from more intensive evaluations.

The 2006/7 survey used a sub-sample of households sampled in the National Statistical Office (NSO) 2004/5 second Integrated Household Survey (IHS2) in order to provide panel data for analysis of programme impacts on beneficiaries. A total of 3298 households were sampled across all districts in Malawi. After data cleaning this gave 2431 balanced matched panel households also sampled in the IHS2. Agro-economic livelihood zones defined by the Malawi Vulnerability Assessment Committee (MVAC) were used to stratify the sample (Malawi National Vulnerability Assessment Committee, 2005). Urban, peri-urban, and protected areas (national parks and reserves) were omitted from the sample. Data collection and entry were conducted by the National Statistical Office. The survey provided very valuable information on household access to subsidized and unsubsidized inputs and on many aspects of programme implementation. Unfortunately it was less successful as regards plot areas and production reported by farmers: these were not found to give (p.10) reliable and consistent results, and this prevented estimation of production impacts of the programme. The 2006/7 input supplier survey involved focus group discussions, key informant interviews, and a survey of 271 retail outlets in 6 districts. This was supplemented by information from fertilizer and seed importers and sellers. The findings were reported in School of Oriental and African Studies et al. (2008).

The household survey in 2008/9 was conducted by the evaluation team with a sample of 1982 households across 14 districts and represented almost all livelihood zones. The sample was a sub-set of the 2006/7 sample and therefore provided a panel data set across three surveys going back to the IHS2. The input supply retailer survey sampled 230 retailers in 6 districts. Both surveys were again supplemented by focus group discussions, key informant interviews, and a community survey, but detailed fertilizer import information was not available. Findings were presented in a portfolio of reports focusing on different aspects of the programme (for example, Dorward et al., 2010a, b; Kelly et al., 2010). The survey again provided valuable information on programme implementation and outputs. However, the introduction of innovative approaches to production and yield measurement (such as yield sub-plots with enumerator and farmer harvests), plot areas, and production data did not give reliable and consistent results. This not only precluded estimation of production impacts of the programme, it also raised questions about the reliability and consistency of area, production, and yield estimates from other studies which rely mainly on farmer estimates and recall of production (see Dorward and Chirwa, 2010b).

The 2010/11 study did not include an input supplier survey, and the sample size of the household survey was reduced further to 760 households across 8 districts in the 3 regions. The sample represented 8 major maize growing livelihood zones covering 77% of all rural households and was again a sub-set of the previous survey (this time the 2008/9 survey). The 2010/11 survey replaced attrition households with younger and newly formed households. The IHS2 and three FISP evaluation surveys generated a matched panel of 461 households. As for 2008/9, findings were reported in a portfolio of topic-specific reports (for example, Chirwa et al., 2011d; Dorward and Chirwa, 2011a; Mvula et al., 2011).

A number of specific studies were conducted and reported within the programme evaluation—for example, on programme impacts, benefit–cost analysis, targeting, and graduation. These are explained and cited where appropriate. It is, however, necessary to provide a little more information here on the development and use of the partial equilibrium Informal Rural Economy or IRE model to explore some of the economy-wide impacts of the programme. This model is fully described in Dorward and Chirwa (2012b). It is based on detailed programming models of different farm household types (p.11) in the two most populous livelihood zones in Malawi. These models allow for seasonal constraints affecting farm household activities and the direct impacts of the subsidy are investigated by simulating the livelihood effects of specific households’ access to subsidized inputs. These effects are then aggregated in order to estimate impacts on supply and demand of seasonal labour and maize. Wage rates and maize prices are then adjusted iteratively to find new equilibrium wages and prices and to derive estimates of economy-wide impacts on both subsidy recipients and non-recipients. The nature of the data available and of the models means that results should be interpreted as indicative of possible effects rather than predictive of actual effects. Nevertheless, when taken together with other information they provide useful insights into possible economy-wide impacts.

A full set of evaluation reports from 2006/7 to 2011/12 can be found at 〈http://www.wadonda.com/〉. These provide further documentation of analytical methods and references are provided whenever their findings are drawn upon.

1.3.3. Official statistics

Malawi has an extensive set of agricultural and other national statistics.

The Ministry of Agriculture publishes valuable monthly information on market prices for major crops, with data collected on a weekly basis from a large number of markets around the country. The Ministry also publishes annual estimates of crop areas, production, and yields, and reports annual estimates of the number of farm families. The annual production estimates, with large increases in estimated maize production following the introduction of the subsidy programme in 2005/6, have been widely cited as evidence of the impact of the subsidy programme. However, there are inconsistencies between the large estimated production increases from 2005/6 and the very high domestic maize prices experienced in some years, notably in 2008/9. These inconsistencies are discussed in Chapter 7. There are also inconsistencies between national maize areas, production, and yield estimates from the Ministry of Agriculture, and from different reports by the National Statistical Office (National Statistical Office, 2005a, 2010a). These discrepancies are discussed in more detail in Dorward and Chirwa (2010b) and summarized in Chapter 6. Another set of discrepancies between Ministry of Agriculture and NSO data concerns the number of farm families (reported annually by the Ministry of Agriculture) and the number of rural households enumerated in the 2008 census (National Statistical Office, 2008a). This discrepancy and the difficulties it raises are discussed in Chapter 5.

Apparent discrepancies also arise between maize prices reported by the Ministry of Agriculture and the consumer price index reported by the NSO. (p.12) No detailed analysis of this has been published, to our knowledge, but the high maize prices observed in 2008/9 and in 2011 do not appear to be consistent with consumer price index figures for the same period, given the high weighting of food and particularly maize in the consumer price index. We also note that the NSO itself refers to revised ‘CPI data’ with ‘overall inflation between the IHS2 and IHS3 periods of 128.9 per cent’ (National Statistical Office, 2012: p. 207), when official CPI estimates for the same period suggest a considerably lower rate of inflation. This also raises wider questions about the deflator used in recent years’ GDP estimates, and hence about these GDP estimates themselves.

Integrated Household Surveys (IHS) conducted by the NSO in 2004/5 (IHS2) and 2010/11 (IHS3) provide national estimates on a wide range of variables. We refer to these in Chapters 4 and 7. However, we also note some apparent discrepancies between and within some of the results presented, and—with the publication of the first report on the 2010/11 survey (National Statistical Office, 2012) as the manuscript for this book was being finalized—it has not been possible to resolve these.

1.3.4. Other studies on the subsidy programme

A number of other studies have been made of different aspects of the subsidy programme. Due to their varied nature and focus we do not discuss these here but refer to them at relevant points in the following chapters. In broad terms the main focus of most other work has been to use survey data to compare observations on recipients and non-recipients in order to examine targeting of and direct outcomes and impacts of subsidy receipt—in terms of differences in wealth, gender, and other household characteristics affecting access to and use of subsidized and unsubsidized inputs, and subsequent differences in changes in wealth and other household characteristics. There has been much less examination of indirect and economy-wide impacts, of impacts on market development, of benefits relative to costs, and of the important question of graduation, though we discuss notable exceptions where appropriate.

Notes:

(1) The Economist (13 May 2000). ‘Hopeless Africa’, 〈http://www.economist.com/node/333429〉and The Economist (3 December 2011). ‘The Hopeful continent: Africa Rising’ 〈http://www.economist.com/node/21541015〉 (18 September 2012).

(2) We do not address here the longstanding debate on the relative merits of investment and support for large- and small-scale farms (see, for example, Collier and Dercon, 2009; Hazell et al., 2010; Hazell, 2012). We favour the complementary approach to large and small farms advocated by Hazell (2012), but our focus is on the potential multiple benefits of overcoming problems faced by poor, small-scale farmers.

(3) Due to difficulties in sourcing better data the indicator used is value of cereal production at international grain prices (estimated with prices and grain index weights taken from World Bank, 2012) as a percentage of agricultural value added (World Bank, 2011) in current US$.

(4) The 2002–9 average rate of nitrogen application per ha arable and permanent crop land in sub-Saharan Africa, excluding Mauritius and Swaziland, is only 4.0 kg/ha.

(5) Smallholder maize production is estimated to account for 97% of the maize and total cereal areas in Malawi in 2009/10 (Ministry of Agriculture and Food Security, 2010).

(6) The names Agricultural Input Subsidy Programme and Farm Inputs Subsidy Programme (AISP and FISP) are often used interchangeably. We generally use the former in discussion of the earlier years of the programme, when AISP was its official title, and the latter when discussing the later years of the programme or the programme as a whole.

(7) Chinsinga (2006) provides a detailed analysis of the political narratives of the farm input subsidy programmes including broad agreement across political parties on the need for farm subsidies in varying form, and the sceptical views of development partners. We discuss these issues in Chapter 4.

(8) Indeed Ricker-Gilbert and Jayne (2012) suggest that debates on agricultural input subsidies are addressing a ‘wicked problem’ that is difficult or impossible to resolve because of contested framings of the problem, incomplete and contested information, and absence of agreement on the core issues.