What Is Poverty Reduction?

Abstract

There is a healthy debate about how to achieve poverty reduction in developing countries, but not enough discussion of what we mean by “poverty reduction.” “Poverty reduction” is often used as a short-hand for promoting economic growth that will permanently lift as many people as possible over a poverty line. But there are many different objectives that are consistent with “poverty reduction,” and we have to make choices between them. There are trade-offs between tackling current and future poverty, between helping as many poor people as possible and focusing on those in chronic poverty, and between measures that tackle the causes of poverty and those which deal with the symptoms. Because donors focus on just one dimension of poverty reduction (growth) they marginalise other legitimate objectives such as reducing chronic poverty or providing social services in countries that cannot otherwise afford them.

Because donor agencies do not recognize these different objectives explicitly, there are important negative consequences for the choice and management of individual aid programmers, and for donors’ ability to make transparent and evidence-baseddecisions about the composition of their portfolio. Aid could be more effective if there were greater recognition of the different dimensions of poverty reduction and if this was recognized in the objectives for and incentives in aid agencies.

There is an ethical case for a global system of social justice that provides long-term,re-distributional transfers of resources to the world’s poor, to enable them to lead better lives while their country is developing, even if there is no expectation that these transfers will accelerate economic development. Reasonable people can disagree about whether this is desirable but the existing hegemonic definition of poverty reduction does not sufficiently acknowledge this as a legitimate goal or permit a meaningful discourse about how it might be achieved.

The Center for Global Development is an independent, nonprofit policy research organization that is dedicated to reducing global poverty and inequality and to making globalization work for the poor.

Use and dissemination of this Working Paper is encouraged; however, reproduced copies may not be used for commercial purposes. Further usage is permitted under the terms of the Creative Commons License. The views expressed in this paper are those of the author and should not be attributed to the board of directors or funders of the Center for Global Development. This paper was made possible in part by financial support from the William and Flora Hewlett Foundation.

Foreword

From its inception the Center for Global Development has made its mark on issues of aid and aid effectiveness. Many of our staff and non-resident fellows – Owen Barder, Michael Clemens, William Easterly, Carol Lancaster, Ruth Levine, Todd Moss, Mead Over, Steve Radelet, David Roodman, Arvind Subramanian, and myself too—have been key contributors to a lively debateon the question of whether and how aid and the aid system work.1

Though we normally include in our working paper and other series only analyses by our own staff and non-resident fellows or analyses we commission ourselves for a particular program, in this special series we are pleased to publish from time to time at our discretion papers and essays prepared outside the Center. Our aim is to share more broadly otherwise unpublished work in which authors propose new thinking about aid and the aid system, and new approaches to operationalizing aid transfers. The focus will be on innovations—whether in ideas or operations.

Our goal is that the Innovations in Aid series speeds and broadens access to new ideas, and contributes to more effective aid programs—public and private, bilateral and multilateral, traditional and new donors.

In this paper Owen Barder raises fundamental questions about the purpose of aid transfers. For many donors the purpose is ―poverty reduction‖ but in the relatively narrow sense of growth that reduces poverty. In fact poverty reduction has other dimensions, including enabling the poor to live better lives through long-term,redistributional transfers while their country is developing, even with programs that might not contribute to growth. His point is not dissimilar to that of Severino and Ray2 in this series, who distinguish between aid aimed at ―accelerating convergence‖ vs. aim aimed at ―providing for basic human welfare as conceptualized in the MDGs‖, or between what I called in my remarks3 at a DfID conference (also included in this series) the development objectives of―transformation‖ and ―redistribution.‖ Barder‘s key concern is that the focus on poverty reduction through growth ignores such key tradeoffs as that between reducing current and future poverty, and between addressing the causes and symptoms of poverty. The reality of these tradeoffs stares us in the face; this is an important paper for practitioners as well as students of the way the aid system works.

 S U M M A R Y

1.Countries give foreign aid for many reasons. As well as moral and humanitarian motives, countries give aid for a variety of commercial, strategic, and political reasons. The development community has emphasized the importance of ―poverty reduction‖ as the main objective of foreign assistance, partly as a tactic to prevent aid from being diverted to other goals. This paper does not consider the other objectives of foreign aid. Instead it looks at poverty reduction and argues that by emphasizing this as a ―single‖ objective donors have glossed over the fact that poverty reduction actually encompasses many goals, some of which are contradictory.4

2.It is widely acknowledged that poverty is multi-dimensional, and that poverty reduction entails many different kinds of change. There is lively and healthy debate about the relative importance of different kinds of development interventions and their different contributions to the overarching objective poverty reduction.

3.But there is little recognition of the important trade-offs within the objective of poverty reduction. The development community agrees that there are many answers to how poverty reduction can be pursued, but sometimes behaves as if there is only one definition of what they are trying to achieve.

4.The reason that ―poverty reduction‖ is more than a one-dimensional objective is that there is no universally applicable way to ―add up‖ the reduction of poverty affecting different people in different circumstances in different places over time. In particular:

a.There is a trade-off between reducing poverty for as many people as possible, and focusing on a smaller number of people in chronic, long-lasting and deep poverty; (broad vs deep)

b.There is a trade-off between activities that reduce poverty today, and those that reduce poverty in the future; (today vs tomorrow)

c.There is a trade-off between programmed that provide immediate redistribution of income and provision of global public goods, but which require long-term funding to be sustained, and time-limited programmers that are intended to catalyze economic growth or social and political transformation so that long term funding is not required; (sustainable vs temporary)

5.The aid effectiveness research industry typically condensed the definition of poverty reduction into aone-dimensional measure: increases in GDP per capita, or a reduction in the poverty headcount. Models in the tradition of Burnside, Collier and Dollar measure the impact of aid on economic growth and then translate this into GDP per capita or an estimate of the number of people that will be taken to above an international poverty line. But defining the success of aid by its effect on GDP – now widely used as a default definition of aid effectiveness – is a considerable over-

4 I am grateful to David Roodman and to a number of UK and US government officials who had better remain anonymous for comments on an earlier draft of this paper. Many of the ideas in this paper were inspired by Michael Kremer, whoremarked in passing that we should be ―more willing to help people live better lives‖ while development is happening in their country.

WHAT IS POVERY REDUC TION?

Simplification of what we mean by poverty reduction, and it embeds a number of implicit value judgments about the relative importance of many different and often competing objectives.

6. This emphasis on the one goal ―poverty reduction‖ defined as a permanent reduction in the global poverty headcount through economic growth has contributed both to poor programme selection and poor programme design and implementation, and it has thereby undermined the effectiveness of aid. Explicit recognition and endorsement of the diverse meanings of poverty reduction would allow different programmes to be selected and designed for those various objectives, rather than trying to distort every programme to meet aone-size-fits-all definition of success.

7.There is a strong institutional preference among donors (and their political leadership) for describing development assistance as a temporary measure aimed at catalyzing economic growth, rather than a long-term programme of redistribution.5 This leads to pressure to design programmes which can be presented as ―transformational‖ or ―financially sustainable‖either because they contribute to economic growth or because they will in time be self-financing.This limits willingness of donors to allocate resources to long-term programmes that redistribute income to the poor, deliver local public goods, and subsidize positive externalities but which are not ―financially sustainable‖ because they require continued external funding for as long as the benefits are delivered.

8.There is not much evidence so far of the success of foreign interventions in accelerating permanent economic, political and social change. It might be that donors should accept that, for at least some of their work, they are pursuing a less ambitious agenda of enabling more people to live better lives while the process of development is taking place. There are many possible programmes which could be supported by aid which make a huge impact on the lives of the poor– for example, by reducing hunger or by making it unnecessary for mothers to watch their child die of an avoidable disease – but which are not designed to have any impact on economic growth in the short or medium term and which can only be sustained by accepting the need for long-termtransfer of resources from rich to poor. At worst, such policies are not regarded as legitimate goals of aid policy and these programmes are not selected for funding; at best such programmes are undermined by ill-conceived attempts to make it appear that they can become financially sustainable within the lifetime of the project.

9.A new agenda for aid effectiveness would recognize the multidimensional nature of poverty reduction and the tradeoffs that it embodies. Development agencies would not target a single measure of poverty reduction but explicitly manage a portfolio of objectives that (a) promote long term and permanent changes in developing countries by investing resources and sharing knowledge;

(b) tackle the causes of poverty by changing the policies of rich countries and investing in global public goods; (c) transfer income and consumption from the world‘s rich to the world‘s poor to enable them to live better lives while development is taking place, as a matter of global social justice; and (d) target more assistance on those in chronic and deep poverty. The relative effort of each donor into these different dimensions of poverty reduction would depend on political priorities, the contributions being made by other donors, and evidence about the magnitude of the trade-offs and relative effectiveness of interventions.

10.A broader understanding of the nature of poverty reduction, and building that understanding into the objectives and incentives of donor agencies would thereby make aid more effective in future.

5 An exception to this is humanitarian aid which is provided without claims that it will lead to long-term development.

OWEN BARDER

T H E “ S IN GL E M IS S IO N ” O F U K FO R E I GN AS S IS TAN C E

11. When the Labour Government came to power in the UK in 1997, it committed itself to ―refocus our international development efforts on the elimination of poverty and encouragement of economic growth which benefits the poor.‖6 This was subsequently enshrined in the International Development Act (2002) 7 which, according to the Department for International Development (DFID), ―establishes poverty reduction as the over-arching purpose of British development assistance.‖8 (The Act also makes specific separate provision for humanitarian aid.)

Box 1: The objectives of UK Development Assistance

(1)The Secretary of State may provide any person or body with development assistance if he is satisfied that the provision of the assistance is likely to contribute to a reduction in poverty.

(2)In this Act “development assistance” means assistance provided for the purpose of—

(a)Furthering sustainable development in one or more countries outside the United Kingdom, or

(b)improving the welfare of the population of one or more such countries.

(3)For the purposes of subsection (2)(a) “sustainable development” includes any development that is, in the opinion of the Secretary of State, prudent having regard to the likelihood of its generating lasting benefits for the population of the country or countries in relation to which it is provided. source: International Development Act (2002)

12.UK Government ministers and senior officials have put a lot of emphasis on this single mission as a key driver of organizational effectiveness and as a bulwark against the diversion of aid funds to other objectives.9 This clarity of purpose has been an asset to DFID. It has made a significant contribution to the esteem with which DFID is held internationally;10 it has helped DFID to recruit, retain and motivate staff; and it has enabled DFID to organize itself around a clear mission statement.

13.The idea of poverty reduction as an overarching objective is not confined to the UK. Indeed, it predates the creation of DFID: in 1995 Jim Wolfensohn made poverty reduction the overarching objective of the World Bank. Other organizations, including the African Development Bank and the Asian Development Bank, have also identified poverty reduction as their overarching goal.11 Other countries have also defined poverty reduction to be the objective of their development cooperation efforts. For example, the German Federal Ministry for Economic Cooperation and Development says, ―The overarching objective of Financial Cooperation is to reduce poverty and improve living

6DFID (1997) – Eliminating World Poverty: A Challenge for the 21st Century

7 International Development Act (2002)

9 For example Sir John Vereker (2002) – ――This clarity of purpose, rapidly transmitted through the organisation, has been a powerful motivating, unifying and guiding force over the last five years.‖; Mark Lowcock (2007) " One of the effects of the 2002 act was to make us very explicit about our core objective. …, it has been our experience that clarity on the objective helps with efficiency and effectiveness."; Douglas Alexander (2008a) – ―… this government introduced the International Development Act in 2002, which legally requires all UK overseas aid to be spent on poverty reduction, moving us away from the scandals of “tied aid” seen under the previous Conservative government.‖; Douglas Alexander (2008b) ―This Labour Government is putting our values into action by untying aid and legally requiring all aid to be spent on poverty reduction.‖; Gillian Merron (2008) – ―Since this Government came to power in 1997 there has been a marked step change in the UK’s commitment to International Development – DFID became a Department in its own right, with a single mission of alleviating world poverty.‖

WHAT IS POVERY REDUC TION?

conditions, especially for the poor‖; the Danish development cooperation department, DANIDA, says,―Poverty reduction remains the fundamental challenge for Danish development cooperation‖;12 and the objective of Australia‘s aid programme is ―to assist developing countries reduce poverty and achieve sustainable development, in line with Australia's national interest.‖13

14. While the idea of a single overarching objective of poverty reduction is rhetorically attractive, the single objective masks a complex set of choices and trade-offs. As is often the case in welfare economics, there are important trade-offs for policy-makers who have to make choices among different kinds of improvements made in lives of different individuals. Those trade-offs are considered in the next section.

T H E TR A DE - O F FS IN P OV ERT Y R E D U C T IO N

15.This paper does not deal with the question of whether – despite the rhetoric – governments actually use development assistance to pursue broader goals such as commercial or strategic advantage. These issues are well documented elsewhere.14 For the purpose of this analysis we accept that the main purpose of development assistance is ―poverty reduction‖. Nor does this paper deal with the wide differences of opinion about which interventions are likely to be effective in promoting development, which give rise to diversity in how development assistance funds are used. Instead we focus on whether poverty reduction can sensibly be regarded as a single objective.

16.The trade-offs within the objective of reducing poverty are familiar from welfare economics which asks: what do we mean by “the greatest good to the greatest number”? How should we add up and compare different benefits, of different magnitudes, to different people, at different times?

17.For example, given the objective of reducing poverty, which of these should we prefer?

a.Causes vs symptoms:

temporarily alleviating the effects of poverty for 100 people for a single year or permanently lifting 10 people out of poverty forever?

b.Depth vs Breadth:

lifting 10 people out of extreme and persistent poverty, or lifting 15 people from just below to just above an arbitrarily-defined poverty line?

c.Today vs tomorrow:

lifting 10 people permanently out of poverty this year or lifting 20 people permanently out of poverty after five years?

18.Reasonable people may disagree about the answers to these questions, and their answers are likely to depend on the specific circumstances, yet all can fairly be considered as consistent with the objective of ―poverty reduction‖. Some donors would want to do all of these things, but with finite budgets they must make choices about relative priorities.

OWEN BARDER

19.In theory it would be possible to define a single composite measure of ―poverty reduction‖ which adds up the number of people, weighted according to the change in their income by some formula that captures our concern for alleviating deep poverty, aggregated across the indefinite future using an assumed discount rate. This composite measure would embody the Government‘s value judgments, and the donor could then seek to maximize this definition of aggregated poverty reduction. In practice, however, there are good reasons (discussed below) to want to pursue a portfolio of activities that is expected to achieve a variety of these objectives to differing degrees in different circumstances.

20.The aid-effectiveness research industry has, however, implicitly selected a composite measure of poverty reduction on our behalf. Cross-country aid-growth regressions have been used to estimate the impact of aid on economic growth over the subsequent 3-5 years.15 These have sometimes been combined with estimates of the impact of growth on the poverty headcount to establish a relationship between aid and poverty headcount.16 These models have been used, either explicitly (in the case of the World Bank and the UK Department for International Development resource allocation models) or implicitly as a measure of the effectiveness of aid. But these models (a) value getting more people over the poverty line rather than valuing lifting people out of deep and lasting poverty; (b) pay no attention to the trade off between current and future poverty; (c) value only economic growth and attach no weight to income redistribution or temporary welfare improvements. If an aid programme were extremely effective at reducing poverty and improving the lives of the poor, by redistributing income from rich to poor, but had no impact on economic growth, it would count as ineffective by this measure of aid effectiveness and poverty reduction. But some people might regard this as aid money well spent.

W H A T I S POV ERT Y R ED U C T IO N ? C AU S E S V S S Y M P TO MS

21.The history of development assistance can be characterised as a series of attempts to identify and address ever more fundamental causes of poverty.17

22.Following the Second World War, it was argued that long-run welfare depended on capitalinvestment, and helping countries raise savings through a ―big push‖ would launch them into self- sustaining growth, or ―take-off‖.18 As a result, donors funded infrastructure, such as dams and roads. However, by the 1980s the development community had concluded that capital accumulation and technological progress depended not only on the level of investment, but also on a better economic policy environment. The combination of policies that were thought desirable was subsequently dubbed ‗The Washington Consensus‘.19 By the 1990s, this approach too was in doubt, and it was argued that these policies could only have the impact intended if they were accompanied by more fundamental institutional reforms.20 A decade later, attention has shifted to even more fundamental causes of poverty such as conflict, rivalry between different social and economic groups, and lack of political accountability.21

 WHAT IS POVERY REDUC TION?

23.This search for ways to address the fundamental causes of poverty appears to be partly a consequence of worries about the prospect of having to finance aid programmes indefinitely. For example, UK Secretary of State for International Development Douglas Alexander emphasized in 2007 that aid should be seen as a temporary measure only: ―we in the development community should not forget that ultimately we exist to work ourselves out of business – and the best way to do that will be to encourage growth and trade.”22

24.The priority given to tackling the underlying causes of poverty is reflected in donors‘ preferencefor time-limited programmes to address the causes of poverty, so that temporary programmes will have permanent effects after the external support ends. (Such programmes are often called

―financially sustainable‖.)

25.Yet conventional public finance economics recognizes that governments should also indefinitely fund public goods and activities that generate positive externalities or which lead to redistribution of income. This means that there are many economically-sensible programmes in developing countries that would improve the welfare of the population but which will depend on government subsidies indefinitely. Using aid to finance these programmes is less attractive to donors because these programmes need to be externally funded externally at least until the government is able to support the programme from its domestic revenues, which may be many years in the future.

26.Although the UK often asserts that its legislation requires it to use aid money to pursue only the single goal of poverty reduction, the International Development Act actually identifies two separate objectives within that overall goal:

a.―Furthering sustainable development‖, which it defines as ―generating lasting benefits for the population‖, and

b.―Improving the welfare of the population‖.

27.The distinction that the Act draws between promoting sustainable development and improving the welfare of the population might be regarded as recognition that the UK Government can legally both search for ways to address the fundamental causes of poverty, andimplement programmes that have an immediate impact on the lives of the poor. 23

28.There are both moral and political reasons for emphasizing short term results as well as sustainability. Continued public support for aid is likely to be enhanced by showing that aid achieves results in a reasonable period of time and being able to connect those results directly to the aid spending. These tend to be more likely for interventions that address the symptoms of poverty and less easy for interventions that seek to address the underlying causes. But as Miguel and Kremer point out (below) there are also strong institutional incentives for aid agencies to to emphasize sustainability.

22 Alexander (2007)

23 The Act also makes provision for humanitarian aid in emergencies.

OWEN BARDER

ORGANISATIONAL INCENTIVES TO EMPHASIZE SUSTAINABILITY

“One possibility is that aid agencies are stuck in a rat race with each other for limited donor funds, and try to outdo each other in extravagant claims about what can be achieved through “sustainable” programs. … Claims about spectacular project “bang for the buck” typically remain unchallenged since aid agencies are not directly accountable to program beneficiaries through either political mechanisms (e.g., democratic elections) or the market mechanism, and rigorous program evaluations are rare. It is also worth noting that the sustainability approach may help aid agencies maximize their jobs and influence: teaching people to fish requires many more jobs for aid workers from the developed world than handing out fish, and it is more exciting for aid workers to launch new programs than simply administer a long standing subsidy program.”

Miguel & Kremer (2004)

29. An ideal aid-supported intervention would meet both objectives: it would deliver immediate and visible results that improve the lives of poor people, and it would result in permanent change after the programme has finished, for example by leaving behind more effective institutions, physical infrastructure, or by helping to bring about a realignment of political or other vested interests.

T H E N E G A T I V E C O N S E Q U EN C E S FO R S U S TAI NA B I L I T Y

O F T H E “C U LT O F R ES U LTS ”

30. Aid programmes which are designed to bring a direct and visible benefit to the population may achieve this immediate impact at the expense of undermining the development of long-termlocal capacity and sustainability. There are several different ways that the pursuit of impact and visibility can undermine domestic capacity:

a.The most egregious case in which aid reduces domestic capacity is the use of imported food aid. This directly benefits the recipients, but at the expense of undermining markets for local food producers and retailers, and so undermines the long-run food supply capacity of the recipient country.24

b.Individual aid projects, managed by donors or by NGOs on their behalf, may be reliable ways to deliver goods and services to the intended recipient, but they also often hire the most effective government administrators and drive up wages in the sector, so undermining public sector capacity. 25

24 Barrett & Maxwell (2005), Gelan (2006)

25 Brautigam and Knack (2004)

WHAT IS POVERY REDUC TION?

c.Support to the private sector may encourage firms to enter markets and provide services that would otherwise have been unprofitable for them, but it can also create subsidized incumbents that block entry to the market by potential innovative entrants that could provide better services at lower costs.26

d.Off-budget aid, separate sectoral budgets and project implementation units undermine government budgeting systems by preventing parliaments from establishing their role as holders of the purse-strings, and undermine the role of finance ministries in enforcing systems of budgeting, accounting and holding spending agencies to account.27

e.Delivery of services outside government (through projects and NGOs), and implicit accountability of governments to donors rather than their own population mayundermine the domestic accountability of government and dilute the “social contract” between the citizen and the state.28

31. The trade-off between sustainability and impact is illustrated in the diagram below. The characterization of programmes on this picture is inevitably subjective and there are variations among programmes in each of the categories shown here.

Figure 1: The Trade-Off Between Sustainability and Results

Builds capacity	Support	Global
and	civil	public
institutions.	society	goods
Effects are
sustained after		Reform trade
funding		system
finishes.

• Institutional reform projects and technical assistance
• Indirect impact less visible
• Benefits only while funding provided
• Likely to erode local capacity and institutions

• Sustainable
• Unsustainable

• Typical trade-offs
• in programme design
• Debt relief

General Budget support

Sector	Direct impact
Budget	more visible
support

 

OWEN BARDER

32.If a programme can be designed so that it is above and to the right of the trend line, then that is always preferable. Programmes below and to the left of the trend line – such as using imported food aid – should generally be avoided. On the trend line itself there is no universal basis for choosing between programmes – that depends on policy priorities of the decision-maker in thetrade-off between impact and sustainability.29 If sustainability is the priority, then programmes towards the top left will be preferred; if visible impact and demonstrable agency is needed, then programmes towards the bottom right are selected: even though many of these programmes will reduce local capacity and undermine sustainability.

33.From time to time it is claimed that some programmes achieve both short-term gains and long- lasting benefits. For example, education and health advocates sometimes argue that spending on these services meets immediate needs and represent a long-term investment in human capital that will generate long-lasting benefits in the future.30 Unfortunately, there is scant evidence for this: statistically, increased public investment in health or education has little discernable effect on economic growth (the relationship seems to be the other way around).31Another example is that service delivery projects are often designed to include a component of technical assistance, which is aimed at enhancing capacity in parallel with short-term delivery, but there is a substantial body of evidence suggests that there are few, if any, long-lasting benefits from this sort of activity.32

T H E N E G A T I V E C O N S E Q U EN C E S FO R R ES U LT S

O F T H E “C U LT O F S U S TA INA B I LI T Y ”

34.Donors allocate money to projects and programmes which sit at many points along this trade- off. Additionally, they seek through programme design to push all interventions out towards the top right of the diagram. But (with the possible exception of humanitarian aid, which is accepted to be temporary) there is political and organisational dissonance with the idea that not all programmes can achieve all these goals at once. As a result of requiring every programme to be sustainable, aid effectiveness is reduced both by poor programme selection and poor programme design.

35.Poor programme selection occurs because donor staff are either instructed or reach the conclusion that they should not support projects and programmes that do not achieve both long- term sustained improvements in the causes of poverty and direct and visible results within the programme planning horizons. In particular, interventions – such as de-worming or vitamin supplements – which are demonstrably cost-effective but for which long-term public funding is required – tend to be underfunded compared to programmes which require only one-offinvestments.33 Similarly, investment in very long-term public sector reform programmes is hampered by a desire to demonstrate financial sustainability within the planning horizon of the project.

36.Poor programme design and implementation is a consequence of a valiant but doomed attempt to do too many things at once. There is evidence that many programmes are less effective because they have to achieve financial sustainability when they could produce significantly better results if the need for continued, long-term finance were accepted. For example, the introduction of

29 Technically, one could draw iso-preference curves of decision-makers showing their appetite for sustainability and speed of impact, and then choose projects at the tangent between the outermost iso-preference curve and the trade-off constraint.

• 30e.g. Commission on Macroeconomics and Health. (2001); Sachs (2005)
• 31Acemoglu & Johnson (2007); Boone & Johnson (2008)
• 32Most notably Berg (1993). See also IMF (2005); World Bank (2005); DFID (2006);
• 33Miguel & Kremer (2004)

WHAT IS POVERY REDUC TION?

Cost-recovery in health and education programmes, aimed at achieving financial sustainability, has lead to large drops in health care and education utilization.34 A study of microfinance suggests that pursuit of sustainability by microfinance organizations has led them to move away from serving the poor.35 In Kenya, a move away from donor support for water well maintenance to the establishment of community management resulted in the water infrastructure falling into terrible disrepair: in one large water project 43 percent of bore hole wells were useless ten years after the shift to a ―sustainable‖ local approach.36 Programmes to fill gaps in the staffing of medical services have to be redesigned so that they can be described as ―capacity building‖, with the result that they fulfil neither objective well.

37.But although interventions that require long-term funding can demonstrably deliver results –in that they improve the lives of the poor – and this appears to be popular with taxpayers, there is strong pressure to define development assistance as transformational rather than welfarist. This gives rise to a much more ambitious agenda within aid agencies to identify time-limitedprogrammes to address the underlying causes of poverty whose benefits will be sustained after the funding has finished. This pressure comes from an unholy alliance of economists (who emphasize the importance of economic growth) and from politicians (for whom it is politically attractive to define development assistance as temporary).

38.The current focus is of the transformational agenda is on accelerating the development of government institutions for the planning and execution of the budget, enhancing the capacity of government organizations to provide services to its citizens, and enhancing the accountability and responsiveness of government institutions.37

39.The transformation agenda is audacious and donors would do well to retain a degree of modesty about the evidence for their success so far. There may in fact be little that outside intervention can reliably do to accelerate the process of development and the causes of poverty. Donors would increase their measures of success if they accepted that some of their activities are aimed at a less ambitious, but more achievable, goal of enabling citizens of these countries to live better lives while those changes are taking place.

40.A group of interventions that appear to have the potential to address the underlying causes of poverty are measures to improve the policy environment of rich countries (e.g. trade policy, migration, corruption, intellectual property), reform international institutions, and invest in global public goods (such as climate change, research and development in agriculture, health or energy, and financial stability). There is growing evidence that these are both effective and cost-effectiveways to tackle the long-term causes of poverty. Investments in these programmes – including in the multilateral organizations that provide global public goods – tend to be relatively cheap in terms of programme spending, but require considerable staff and administrative resources. The transfer of staff from the administration of bilateral aid programmes to these activities would require either a reduction in total bilateral aid, or a reduction in the overhead administration of bilateral programme spending. The internal incentives of aid agencies, however, militate against voluntarily reducing their efforts on bilateral programmes to give more attention to these issues.

41.Reasonable people might think that aid can be also used as a long-term mechanism to transfer resources from the world‘s rich to the world‘s poor, even if that does not result in long-lasting

• 34Holla & Kremer (2009); Meuwissen (2002); Gertler, Locay & Sanderson (1987)
• 35Morduch (1999)
• 36Miguel and Gugerty (2002)

37DFID

OWEN BARDER

change in the country of the recipient. Programmes to deliver local public goods, subsidize positive externalities or redistribute income to the very poor are a legitimate goal of public policy, and supporting such programmes in developing countries is a reasonable use of foreign assistance. But this requires a change in attitude about the purpose of foreign assistance: we would have to accept that as well as promoting lasting change in developing countries, there is also a respectable case consistent with the principles of welfare economics for long term redistribution from rich to poor, even if they happen to be in different countries.

42. In their provocative paper in this series, Jean-Michel Severino and Oliver Ray argue that such a change in thinking is indeed under way (see box below).

THE GROWING COMMITMENT TO GLOBAL REDISTRIBUTION

In the Millennium Declaration for example, the international community took the resolute decision to embody the principles of the United Nations Charter in concrete and operational programs. In this truly cosmopolitan logic, each citizen of the planet, by virtue of his or her humanity, is given the right to a minimum living standard. Figures, however, show that many states will not have the macroeconomic capacity to guarantee these basic standards for many decades to come. By aiming for targets that are out of reach from the neediest countries‘ public authorities, the ‗international community‘ (i.e. donor nations) has therefore accepted to substitute itself to some states in the provision of basic social services through long-term financial transfers. The consequence is that the concerns of financialself-sustainability have been dampened: no one asks whether projects funding the education of Mali‘s children or the access to clean water for urban dwellers of Haiti are ‗economically viable‘ in themselves. This change of philosophy implies a real revolution for the development community, one which very few states have fully apprehended: theefficiency of a programme is no longer evaluated on the basis of its recipients‘ capacity to emancipate themselves from international transfers through economic growth, but through the sole improvement of the targeted populations‘ basic living standards. In a way, official development flows have moved beyond a logic of economic investment to include one oflong-term social redistribution.‖

Jean-Michel Severino and Olivier Ray, ―The End of ODA: Death and Rebirth of a Global Public Policy‖,

Center for Global Development Innovations in Aid Series, Working Paper 167, March 2009

WHAT IS POVERY REDUC TION?

W H A T I S POV ERT Y R ED U C T IO N ? D E P T H VS . B R E AD T H

43.A second trade-off within the ―single‖ objective of reducing poverty is the choice between lifting as many people as possible out of poverty, and focusing attention on a smaller number of people who are in the most long-lasting and deep poverty.

44.This choice is exemplified by a comparison of two very influential contributions to development thinking in the last decade. The work by Burnside, Collier and Dollar on the most ―poverty efficient‖ allocation of aid focuses on lifting as many people out of poverty as possible for a given aid budget.38 By contrast, more recently Collier has argued that donors should focus their resources on the eponymous ―bottom billion‖ – on the grounds that these are the people suffering from long- lasting poverty from which they are unlikely to escape without deep and sustained external assistance.39

45.There are an estimated 300-500 million people trapped in chronic poverty – people who will remain poor for much or all of their lives and whose children are likely to inherit their poverty. These chronically poor experience multiple deprivations, including hunger, under-nutrition,illiteracy, lack of access to safe drinking water and basic health services, social discrimination, physical insecurity and political exclusion. Many will die prematurely of easily preventable deaths.40

46.There is a powerful ethical case for focusing more attention on people in chronic poverty.41There is also an increasingly well-made case based on the huge human development costs of the effects of volatility on the most vulnerable. Because there are few cushions from the state or from financial markets, a negative shock can lead people to have to eat their seed, take their children out of school, sell their livestock or their cooking pot. This leaves them in chronic poverty not just for one generation but for the next generation as well. Welfare schemes that reduce the vulnerability of the very poorest to this kind of shock not only protect people who would fall below a minimum standard of consumption, it also promote long-term human development.42

47.A donor‘s attitude to the trade-off between helping as many people as possible, and targeting the chronic poor, can have profound implications for its allocation of aid, for its selection of programmes and for the design of those programmes. Focusing on shifting the largest number of people across the poverty line would lead agencies to allocate substantially more aid to India; and to the fast-growing developing countries such as Mozambique and Tanzania. It would focus on generating economic growth in the most promising regions and sectors within countries to create jobs.

48.Giving more priority to those in chronic poverty would lead to quite different choices. Donors would increase their activities in fragile states and post-conflict environments. Greater emphasis would be given to the most marginalized groups in those and in other societies, such as pastoralists, minority ethnic groups or disabled people. Interventions would target increasing political voice and rights for these groups, social safety nets, effective anti-discrimination action, gender empowerment, and programmes to increase transparency and economic and social participation. These programmes

• 38 Dollar & Burnside (1997); Collier & Dollar (1999)
• 39Collier (2007)
• 40Chronic Poverty Report 2008-09 (http://www.chronicpoverty.org)

41In economic terms, the argument is that if there are diminishing marginal welfare benefits to consumption, increasing the consumption of the very poorest will have the greatest positive impact on total welfare.

 

 

OWEN BARDER

would be much more expensive (in terms of cost per person), and more uncertain, but where they succeeded they would lead to much more profound improvements in the lives of the beneficiaries who, without such help, are likely to remain in poverty for generations to come.

49.There are some complementarities between these objectives. Broad economic growth can havebenefits for the chronic poor, in a form of ―trickle down‖ effect. The emergence of a middle class can lead to establishment of institutions that guarantee rights or reduce corruption from which the poorest members of the community will also benefit. Service delivery programmes can be designed to include special measures that target marginalized communities. Investments can be made to improve the impact of migration and urbanization on the most marginalized communities.

50.But the trade-offs are more pronounced than the complementary. In particular, most donors make choices within twin budget constraints for programme spending and administrative costs. An increase in aid for a fast growing, successful developing country necessarily reduces aid available for a post-conflict society. Increasing investments in economic growth and infrastructure come at the expense of investments in social protection programmes. Furthermore a rush for growth that ignores inequality and social marginalization may exacerbate the problems of those in most poverty: for example, market liberalization that accelerates growth in urban areas may result in greater inequality and social exclusion that further limits the life chances of the chronic poor.

51.One reason that many donors do not choose to focus only on people in chronic poverty or the bottom billion is that these investments are more risky, and it is more likely that there will be little evidence of success and progress for several years. The public ―case for aid‖ is based on aggregate numbers of people lifted out of poverty, and on associating donors with individual success stories.43 If investments were made mainly in the hardest-to-help communities, the overall number of people helped would be smaller, the successes fewer and the risk of failure higher.

52.It is not necessary for a donor to choose either breadth or depth – different approaches can be pursued in different circumstances. But it is important to be clear which objectives are being pursued by particular programmes and why, to avoid thinking that all programmes should achieve all of these objectives and to manage the portfolio to deliver the desired combination of results.

53.The increased focus of some donors on ―fragile states‖ is an indication that they are willing to devote resources to more difficult challenges and not merely to cherry-pick the easiest and most certain investments. But the existing incentives within many agencies – to meet the numerical goal of halving the proportion of people living in poverty and to demonstrate rapid and visible results in each programme – all create pressure against targeting aid on measures to assist the chronic poor.

W H A T I S POV ERT Y R ED U C T IO N ? TO DAY V S . TO M O R ROW

54.There is an obvious trade-off between achieving results quickly today or securing benefits, which may be larger, in the future.

55.The increasing trend towards identifying and reporting measurable results of aid programmes may be shortening time-horizons for development interventions. Despite promoting policies that emphasize long-term results, institutional incentives tend to favour projects that can produce results

43 See H M Treasury and DFID (2002). See also http://www.dfid.gov.uk/aboutdfid/dfidsuccesses.asp

 

WHAT IS POVERY REDUC TION?

quickly over projects that are time-consuming to design and implement, and whose benefits are expected to be longer-term.

56.For example, there is an acknowledged gap in amount of good quality evaluations of aid programmes. The reasons are clear: rigorous evaluation adds to the expense of a programme and slows down its implementation; the benefits (which may be large) accrue to future generations of development policy-makers.44 The lack of investment in good quality evaluation reflects incentives within aid agencies which favour short-term results over the long-termbenefits of evaluation.

57.There are other examples of the trade-off between long-term and short-term effects which suggest a systematic bias towards the short-term. Establishing separate project implementation units is preferred to reforming the local procurement and accounting systems; agricultural research and investment in agricultural productivity have been consistently under-funded relative to food aid; technical assistance in the form of flying in foreign experts is preferred to investing in the development of domestic capacity; and money has flooded to importing drugs and health commodities but not to longer-term investment in health systems or health research.45

58.This bias towards the short-term is hard to explain in the context of contemporary development thinking and the stated policies of many agencies which emphasize the long-termnature of the processes of institutional and economic changes that underpin the development process. It is easier to understand through the lens of political economy, given the need for any particular administration to demonstrate the impact of their own programmes, and through the lens of internal institutional incentives within donor agencies, given that individuals are more likely to prosper if they can show that the projects they have sponsored have achieved an impact.46 But in practice the achievements of which the development community is most proud –and justly so – have been the results of longer-term investments, such as the ―green revolution‖ in agricultural productivity, the eradication of smallpox, and the economic transformation over decades of recipients of large and sustained quantities of aid such as Korea and Taiwan. The short-term gains are more easily won but quickly forgotten.

59.The trade-off between short- and long-term objectives should also inform the allocation of aid resources. The conventional Collier-Dollar model for aid allocation makes no allowance for expected changes in poverty levels in each country over time. This leads to inefficient and inconsistent judgments about the allocation of aid.47 Alternative models which take account of expected future poverty levels result in higher aid allocations to those regions that are expected to endure poverty for a longer period of time.48 The extent of this effect depends on the discount rate, and on the weight that is placed distributive justice and depth of poverty as well as reductions of the numbers in absolute poverty.49 Using an aid allocation model that does not take account of expected changes in poverty over time tends to promote allocating funds towards countries such as India and China, which are home to very large numbers of poor people but which are growing rapidly (and so expected to reduce poverty irrespective of aid) rather than the most fragile states and investment in the chronic poor. Similarly, the measures of allocative efficiency that inform choices about the allocation of funds to multilateral organizations make no allowances for expected changes in poverty over time.

44 Savedoff, Levine & Birdsall When Will We Ever Learn? (2006)

45SIDA (2002)

46SIDA (2002)

47Wood (2008)

48Cogneau & Naudet (2007); Wood (2008)

49Cogneau & Naudet (2007)

OWEN BARDER

60. There is an obvious correlation between these choices and the trade-off described above between causes and symptoms: the most sustainable (but least direct) interventions typically deliver benefits over a longer timescale; the interventions in which a donor can demonstrate direct impact tend to be short-term. There is also a correlation with the choice between helping the largest number of people (which tends to be possible with shorter-term interventions) and helping those in deepest poverty (which tends to require longer-term programmes).

W H A T I S POV ERT Y R ED U C T IO N ? A P O RT FO LI O A PP ROAC H

61.Donors maintain a portfolio of activities in many different countries and many different sectors. This partly reflects the fact that poverty reduction is multi-dimensional and encompasses many different objectives; and that donor agencies are not risk-neutral. It may also reflect the reality that donors actually have a wider set of objectives than just poverty reduction, even if they prefer not to acknowledge this.

62.Organisations with explicit aid allocation models, such as the World Bank and the UK Department for International Development, claim to be allocating their aid broadly to have the largest impact on poverty reduction. They both use a one-dimensional measure of poverty reduction (GDP per capita) and use models based on the Collier-Dollar analysis to derive a ―poverty efficient‖ allocation of aid across countries.

63.For these models to recommend a portfolio of aid allocations there must be convexity (ie diminishing returns) in the relationship between aid and the objective being maximized. Without diminishing returns, the models would suggest that all aid should be allocated to the single country in which, according to the model, it will do most good. The models recommend spreading the aid more widely than this because they assume aid promotes economic growth but with diminishing returns. But if there are diminishing returns from aid to growth, this is not in fact the reason that donors allocate aid to many different countries. (If donors took diminishing returns to aid seriously they would need to use predictions of global aid flows in their models, which they don‘t.) In the real world, the convexity comes from diminishing returns in donors‘ utility from the achievement of different kinds of poverty reduction in different places at different times. Most donors would prefer to lift 100,000 people out of poverty in both Ethiopia and Sudan than to lift 200,000 people out of poverty in Ethiopia alone. This reflects convexity in donor‘s preferences, not diminishing returns to aid. With convex preferences for different kinds of poverty reduction, they should manage a portfolio of activities which includes a combination of different kinds of poverty reduction and a combination of associated risks.

64.For example, an investment portfolio might include a combination of bilateral programmes and funding of multilateral programmes that achieve:

a.visible, deliver short term benefits, and which are likely to succeed, to build public confidence in the aid programme;

b.far-reaching and permanent changes in developing countries which tackle the causes of poverty, but which are more risky, longer term and less visible;

WHAT IS POVERY REDUC TION?

c.changes to the policies of rich countries, reforming international institutions, and investing in global public goods.

d.significant, visible and rapid impact on the lives of the poor, possibly requiringlong-term government funding either from donors or (when the country becomes rich enough) from internal revenues, and which can be designed in ways that avoid undermining domestic capacity.

e.targeted improvements for fragile states, the chronic poor, and other―hard-to-reach‖ poverty; these programmed are relatively expensive, risky and requirelong-term engagement but address long lasting poverty that is otherwise unlikely to be addressed.

65.It is comforting to observe that this is what many donors do in practice, at least to some degree. Many official donors have some combination of different types of activity – sometimes within one aid agency, sometimes spread among several different organizations. But for many agencies, such a portfolio is maintained despite the pressures of an analytical framework which focuses on a one- dimensional measure of poverty reduction. For example, within the allocation frameworks used by these donors it is difficult to justify interventions that have little impact on overall economic growth but which target the chronic poor; or which support fragile states where the probability of failure is higher. This in turn leads to the poor choice of interventions, and poor implementation, described above.

66.The extent to which each donor invests in these different activities should depend on ethical preferences for different kinds of poverty reduction, the overall size of the programmer, the priorities and activities of other donors, and evidence about the magnitude of the trade-offs, risks and relative effectiveness of interventions. Organisational comparative advantages may indicate that some of these investments should be delivered through multilateral bodies rather than bilateral programmer.

67.Proper management of this portfolio requires more explicit identification, and quantification, of these different objectives and associated levels of risk, and continued review and assessment of the results. The portfolio should be reviewed taking into account investments made directly through the bilateral programmer, indirectly through the allocation of bilateral funds to the activities of multilateral agencies and the central allocation of resources to multilateral agencies. This more sophisticated portfolio analysis will validate a variety of different kinds of development assistance, and avoid trying to squeeze everything into a one-size-fits-all definition of success.

C O N C L U S IO N

68.In practice, foreign assistance has many purposes, not all of them directly linked to the reduction of poverty. Partly to protect itself from diversion of funds from non-development objectives, the development community has sought to emphasize poverty reduction as the ―overarching objective‖. By simplifying its objectives in this way, we have ignored important trade-offs and choices.

69.Each individual project and programme cannot meet all the objectives of poverty reduction at the same time and to the same extent. If the many different objectives are not recognized and valued in their own right, staff in aid agencies will endeavor to design and select programmed that reflect what they perceive to be the dominant narrative of the day; or, even worse, they will attempt to meet a number of incompatible objectives at once. The resulting homogenization and

 

OWEN BARDER

Compromises lead to ineffective project selection, and poor project design, resulting in unsatisfactory progress towards all the objectives and a reduction in the effectiveness of aid.

70. There are significant incentive effects in project selection and design which donors should seek to overcome: there is pressure to promote ―sustainability‖; this can lead to underinvestment in proven, effective, long-term interventions to improve the lives of the poor but which require long- term funding; there is a tendency towards designing programmes to achieve a variety of incompatible objectives which leads to insufficient specialization in programme design; if the focus of a programme is to achieve direct, short-term impact then it will achieve more if it does not also try to demonstrate long-term sustainability (though it should seek to avoid undermining local capacity); despite the evidence that they are among the most cost-effective and most sustainable interventions, there are institutional biases against investing in promoting changes in the policies of rich countries, reform of international institutions, and in global public goods. Many of these require increased investment in multilateral organizations; current measures of success emphasize breadth – that is, lifting the greatest number out of poverty – at the expense of depth – the harder, longer-term and more expensive task of building pathways for the 300-500 million people living in chronic poverty; aid agencies have incentives towards designing projects to deliver short-term results, as a consequence of political and institutional incentives, which can come at the expense of investments in long-term change.

71.Donors should:

a.Recognize and validate the multiple objectives encompassed by, and trade-offs inherent in, the single mission of ―poverty reduction‖;

b.Aim to build an explicit portfolio of bilateral programmes and contributions to multilateral organisations that, overall, makes the most effective contribution within the international system to their different objectives;

c.Value the specific contributions that different programmes can make to these diverse objectives; this entails recognizing that it is the portfolio as a whole, not each programme, that has to balance these objectives, so allowing each programme to be chosen and designed to have the largest possible impact on a subset of the organisation‘s objectives;

d.Encourage a discourse that is explicit about the contribution that programmes make to these different objectives to be talked about and, where possible, to be quantified and measured;

WHAT IS POVERY REDUC TION?

72. Over the last forty years aid has come to be viewed as needing to be either ―humanitarian‖ – a temporary response to a disaster – or ―transformation‖ – a temporary injection of resources that will lead to economic growth and financial independence. In my own view there is a compelling ethical case for a global system of social justice and protection of minimum standards of human welfare that provides long-term, re-distributional transfers of resources to the world‘s poor, even if there is no expectation that these transfers will accelerate economic development, to enable the citizens of a country to live better lives while the process of development takes its course.