POLITICS AND POVERTY: A BACKGROUND PAPER FOR THE WORLD

DEVELOPMENT REPORT 2000/1

Mick Moore1 and James Putzel2
September 1999

Contents
Introduction: Practical Political Analysis
Five Strategic Guidelines
1. Democracy has differential outcomes for the poor
2. States create and shape the political opportunities for the poor
3. There is no reason to expect that decentralisation will be pro-poor
4. There is a wide range of possibilities for pro-poor political alliances
5. Many of the policies needed to improve governance will benefit the poor
Implications for Aid Donors
Bibliography

Background:
This is a synthesis, for the World Bank Team working on the World Development
Report 2000/1, of the conclusions of a research project on the Responsiveness of
Political Systems to Poverty Reduction commissioned by the Governance Department
of the UK Department for International Development (DFID). It is based principally
on the work of the following people, who wrote papers that were discussed at a
meeting held on 16-17 August 1999: Michael Anderson, Richard Crook, George
Gray-Molina, John Harriss, Ron Herring, Peter Houtzager, Marcus Kurtz, Jennifer
Leavy, Mick Moore, Kimberly Niles, Alan Sverrisson, Ashutosh Varshney, Howard
White and Lawrence Whitehead. We are deeply grateful to them for the high quality
of the work they did within a short time. The papers are listed in the Bibliography.
Further, we owe a great deal to colleagues who, variously, attended that meeting as
discussants, contributed to the framing of the project at a preparatory meeting held on
26-27 February 1999, or provided helpful comments on earlier drafts: Catherine
Boone, David Booth, Teddy Brett, Kathryn Clarke, Monica Dasgupta, Garth
Glentworth, Merilee Grindle, E. Gyimah-Boadi, David Lehmann, Fernando Limongi,
Ian McKendry, Dele Olowu, Elisa Reis, Alice Sindzingre, Rehman Sobhan, Richard
Thomas, David Wood and Geof Wood. We greatly benefited from the opportunity to
attend the Summer Research Workshop at the World Bank on 6-9 July 1999, that was
oriented around the World Development Report 2000/1. Ravi Kanbur, the Task
Manager for the World Development Report 2000/1, provided invaluable guidance at
several points. Kathryn Clarke of DFID helped manage the project on a day to day
basis, and Julia Brown and Jenny Edwards of the Institute of Development Studies
organised our meetings most expeditiously. For the final shape of this report, we are
deeply indebted to Merilee Grindle for insightful advice, and to Roger Wilson, Head
of the Governance Department of DFID, who conceived the project and played a
major constructive role at every stage.

1
The Institute of Development Studies at the University of Sussex, Brighton BN1
9RE, UK (Email - MickM@ids.ac.uk)
2
Development Studies Institute, London School of Economics, Houghton Street,
London WC2A 2AE (Email - J.Putzel@lse.ac.uk)
 2

INTRODUCTION: PRACTICAL POLITICAL ANALYSIS

There is a tradition in aid and development agencies of bringing in political analysis,

if at all, in terms of problems and difficulties. 'Politics' is why desirable things may
not happen. Politics is messy. Political analysis is used only to explain and to try to
fix things that have already gone wrong. In reality, political analysis is not a gloomy
science. It has a great deal to offer in understanding what to do about poverty. First,
it provides policymakers with a more realistic perspective on their programs and
projects, and increases the chances that they will be successful. Second, political
realism is more optimistic than is much of our conventional wisdom about the scope
for effective pro-poor coalitions, policies and programs. Where the documents of aid
and development agencies make reference to politics and poverty, it is generally in a
depressing tone. The politics of poverty is assumed to be only about fighting to
divide up the cake. Someone's gain is always someone else's loss. We are told how
difficult it will be to do anything for the poor because this means depriving the rich
and powerful. The analysis centres on the conflict between the interests of the poor
and of the non-poor. This is a partial view of politics and a biased reading of
experience. The real world of politics is rarely a crude struggle of rich against poor; it
is also about accommodation, compromise and joint gains. And politics is not driven
solely by (material) interests: ideas and institutions also play a major role in shaping
how far people have political choices, what they want from politics, and how they go
about seeking it. By exploring what drives politics in poor countries, we can
identify: openings for political coalitions and alliances that benefit the poor; ways in
which 'friends of the poor' inside and outside government can shape public programs
and policies to increase the political capabilities of the poor; and areas where there are
common interests between poor and non-poor.

This kind of political analysis is not intrinsically difficult: politicians do it all the time.
It is more difficult for the staff of aid and development agencies, who have to deal
with many diverse situations to which they are outsiders. Angola is very different
from Bangladesh, and Bangladesh from Chile. There is no substitute for detailed
local, expert knowledge when it comes to country-level operations. Aid donors are
increasing their capacity to undertake operational political analysis. They need to do
more. They can however overcome some of their own bewilderment about the
complexity and diversity of developing country politics by using a few basic concepts
to map that diversity. Table 1 provides one such map in relation to the politics of
poverty. It is based around two fundamental dimensions of national political systems.
The first is the degree to which they are institutionalized, i.e. the extent to which
countries are ruled and administered through established, stable organisations and
procedures that are widely accepted as legitimate. The second is the extent to which,
within the more institutionalised states, rulers are chosen through genuinely open,
competitive processes - free elections of some kind. Electoral competition is
impossible without institutionalisation.

(Insert Table 1 - Types of state, politics and poverty - about here)

Table 1 provides us with guidelines about:

• The type of politics that are likely to dominate in different situations
 3

• The roles that poor people and the organisations representing them might play in
different situations
• The scope for both local political actors and external agencies to advance the
interests of the poor in the political process

Those people we label 'the poor' are diverse, even within localities. Much has been
said about that elsewhere. Table 1 illustrates the diversity of their political situations
and opportunities at national level. It also reflects something they have in common - a
critical feature of the politics of poverty that is explained in more detail below: the
scope for the poor to organise, and the ways in which they organise, are highly
interdependent with the character of the state or regime, the shape of public policy
and the behaviour of ruling elites. Small communities, groups or localities can
organise to tackle their immediate needs in the absence of effective government. But
large populations of the poor organise in relation to the forces that they are trying to
combat or influence. If the state is fragmented, unstable and incoherent, popular
organisations will develop along the same lines. The existence of effective
government is a condition for popular organisation that extends over large populations
and has the potential to influence public policy. The development of the political
capabilities of the poor should be an important objective of anti-poverty policy. That
is partly true by definition: empowerment is itself a dimension of poverty alleviation.
It is also true in a more instrumental sense: lasting, sustainable improvements in the
position of the poor will depend on their collective capacity to defend and build on
achievements.

There then emerges a new variant of a familiar paradox: development is often easier if
you are already halfway there than if you are still near the starting line. The capacity
for the poor to create effective organisations is greater in those - generally richer -
states where the quality of governance is higher. In designing policy for specific
countries, it is important to bear this paradox in mind. It is not however a cause for
despair. In countries afflicted by poor governance, the poor have much the same
primary interest as the great majority of the population: government that is less
repressive and arbitrary, more accountable, and more bound by law.

This report is organised around five key propositions about politics and poverty that
emerged from our extensive research and consultations. They are intended to help
policymakers think strategically about poverty reduction policy:

• Democracy has differential outcomes for the poor

• States create and shape the political opportunities for the poor
• There is no reason to expect that decentralisation will be pro-poor
• There are a wide range of possibilities for pro-poor political alliances
• Many of the policies needed to improve governance will benefit the poor

This list refers almost exclusively to the internal politics of developing countries. Are
we guilty, at this point in history when there is so much interest in the implications of
globalisation, of ignoring the international political dimensions of poverty? This is an
especially important issue because of the popularity of the view that globalisation
reduces the scope for governments to attend to poverty alleviation. There are several
variants of this argument. The core propositions are straightforward. They rest on the
notion that globalisation has produced a relative shift of power away from states to the
 4

people who make decisions about large scale capital movements (the 'controllers of
capital'). Capital has become more mobile internationally. The greater the efforts that
governments make to tax or redistribute income or capital, preserve or increase social
protection, or influence where and how the private sector invests, the more, it is
claimed, they will be ‘punished’ by the controllers of capital. The controllers of
capital will reduce investments in countries (and cities or regions) ruled by regimes
lacking in ‘realism’ - or simply threaten to do so, pointing out how easy it is for them
to move money and plant away to more favourable business environments. The
spectre or the reality of declining tax revenues, rising unemployment and falling
political support will be enough to persuade most governments to accept 'reality': to
reduce business taxes, cancel promises to redistribute land, dilute proposals to extend
employee rights, and postpone plans to provide a basic income to all destitute
households.

There is some truth in these arguments. However, we do not in general believe that
globalisation has reduced the scope for (developing country) governments to be pro-
poor. There are powerful countervailing forces. Two are particularly important:

• First, there is an alternative, contrary interpretation of the political consequences

of globalisation that is better supported by evidence: that the opening of national
economies (i.e. globalisation) consistently leads to a larger economic role for
governments, and more social protection, because openness generates economic
instability that governments are obliged to mitigate (Rieger and Liebfried 1998).
Political arrangements underpin the market. Without social and political
protection against the instabilities and uncertainties that it generates, the free
market itself is at risk.

• Second, unprecedented pressures and incentives for poor countries to become

more democratic are integral features of contemporary globalisation. Unlike the
first and second waves of democratisation, after World War 1 and World War 2
respectively, the wave that developed in the 1980s has not foundered on
international geo-political rivalries and the willingness of great powers to install
and support despotic but pliable regimes in the poorer parts of the world. This
third wave of democratisation has been unusually vigorous and sustained. It
shows no signs of receding. While the connection between democracy and pro-
poor policies is less close than many people believe (see below), there is a link.
By promoting democracy, contemporary globalisation discourages governments
from succumbing to pressures from internationally mobile capital to abandon
social concerns.

It is then quite appropriate to focus our strategic guidelines about poverty and politics
on the internal aspects of governance in poor countries.

FIVE STRATEGIC GUIDELINES

1. Democracy has differential outcomes for the poor

Democracy is desirable for many reasons. However, the democracies among

contemporary developing countries are no better than the non-democracies at poverty
reduction. This finding is not new. It was replicated in three of our papers, each of
 5

which examined the experiences of a large number of countries, and defined 'pro-
poor' in a different fashion. Ashutosh Varshney (1999) examined the record of
countries at reducing the numbers of people below the poverty line, defined in terms
of income or consumption. Kimberly Niles (1999) measured the effort governments
put into protecting the poor against the adverse effects of economic adjustment. Mick
Moore et al. (1999) explored the extent to which national political-economic systems
converted national income into longevity, literacy and education for the mass of
citizens. All concluded that there was no consistent connection between pro-poorness
and democracy. While the very worst performers tend not to be democracies -
democracy does provide some kind of safety net - there are non-democracies among
the best performers. Over relatively long periods of time some authoritarian regimes,
like that which existed in Indonesia for over thirty-five years, made faster progress in
reducing poverty than other states which have enjoyed long periods of democracy -
like the Philippines, where the rate of poverty reduction has been much more modest.
Table 2 illustrates the kind of inconclusive patterns that result when we classify poor
countries according to their poverty reduction performance and degree of democracy.

(Insert Table 2 - Distribution of Countries According to Poverty Reduction

Performance and Degree of Democracy - about here)

What is going on here? Democracy does offer more voice and influence to the poor
than most non-democratic systems. Why does this not result in a clear association
between democracy and good performance in poverty reduction? There are two main
answers to that question. The first is that some of the best performers in poverty
reduction over the past half century have been the (former) socialist states, that were
undemocratic by conventional criteria but highly focused on improving mass welfare
for reasons of ideology and politics. The positive impacts of these histories of pro-
poor mobilisation are still evident in the poverty and welfare statistics for China, Cuba
and Vietnam. The impacts have faded in the case of most of the countries of the
former Soviet Bloc after their transition to market economies, and have turned
grotesque in North Korea. The residual effects of this experience of non-democratic
but pro-poor socialist regimes is one reason why democracy is not associated with
poverty reduction in the cross-national statistics. The other reason, of more relevance
to the future, lies in wide variations in the substantive content of formal, electoral
democracy.

If democratic politics were mainly about organising people to vote according to their
broadly-defined economic interests, one would expect the poor to have considerably
more influence and voice than they do (Varshney, 1999). There are three important
reasons why democracy does not work in this way:

• The actual participation of the poor, particularly women, does not reflect their
numbers in society. Many poor people are excluded from, or do not participate
actively in, the political process.

• When the poor do participate, their 'class' identities - as poor people in general or,
specifically, as small farmers, landless, wage workers, tenants, recipients of food
subsidies, squatters etc.- are not the only influences on the way they vote or on the
politicians, parties or programs that they support. The forces that move them are
often more tangible, short-term, direct and local than relatively abstract notions
 6

about income group, class or occupational position: individual 'patronage' ties to

politicians who promise something immediate in return; the ethnic, linguistic or
regional identities that often prove so effective in creating emotional attachments
to political ideas and programs; or the opportunity to sell a vote (Putzel, 1995).

• Poverty in developing countries is especially acute in rural areas, where, except at

the very local level, it is particularly difficult to sustain effective organisations that
involve poor people on a continuous basis. Communication and travel are
difficult; information is scarce. Poor rural populations rarely sustain coherent,
encompassing political organisations. They tend to be responsive rather than pro-
active in politics.

In sum, in many democracies the poor are often badly organised and ill-served by the
organisations that mobilise their votes and claim to represent their interests.

Nevertheless some democracies in developing countries perform well in terms of

poverty reduction. This can be traced in part to differences in the quality of
organisations in political and civil society, and the ways in which the poor realise
voice within them. In every democratic system, different kinds of organisations
mediate between various poor sections of the population and the state. Civil society is
made up of a wide range of associations rooted in and cutting across class, gender,
income and other identities. Some of these emerge out of poor communities, or are
organised explicitly to interact with and promote the interests of those communities.
They include religious, community and non-governmental organisations (NGOs).
Groups within civil society, whether aligned with the poor or with elites, such as
business or professional associations, often serve as points of mobilisation and
advocacy in relation to the state.

While civil society organisations are important sources of assistance and mobilisation
for specific groups among the poor, it is the organisations of political society within
democratic systems that are crucial to the character and conduct of public policy.
Chief among these are political parties. Political parties can range from temporary
alliances of powerful individuals, through more stable organisations constructed
around regional, patronage, ethnic or religious networks, to organisations based on
clearly defined ideologies and programs, and run by committed voluntary members
operating through a democratic institutional structure. The more that parties are
located toward this latter end of the spectrum, the more likely they are to represent the
poor effectively.

(Insert Box 1 - Competitive politics and the poor in two Indian states - about here)

Contrasts between the states of India illustrate how variations in the pattern of
democratic political competition affect the extent to which governments are pro-poor.
India has enjoyed a relatively stable system of competitive elections and basic formal
democratic rights since independence in 1948. The southern state of Kerala has long
enjoyed very high levels of mass literacy, education, health and longevity in relation
to its average income levels. There is a large literature that attempts to explain this
'Kerala exceptionalism', and no consensus on the relative roles of the various
explanatory factors - a strong popular communist movement, a particular caste
structure, or a range of contingent historical factors. Even if one can explain
 7

exceptional cases, there may be no policy implications, because exceptional

combinations of historical conditions cannot be replicated. However, the Kerala
experience is becoming less exceptional as other South Indian states begin to
reproduce key elements: the mobilisation of the lower class/caste groups around
stable political parties that, once in government, redistribute resources to the poor in
ways that lead to permanent reductions in poverty. John Harriss (1999) demonstrates
that this kind of democratic politics does make a difference. Indian states that have
elected to office well-institutionalised populist or social democratic parties have
performed better in poverty reduction than underlying patterns of agricultural growth
would predict, and better than comparable states where clientelist patterns of politics
have persisted (see Box 1). Where the rule of upper class/caste alliances has been
challenged, the political parties have actively sought the support of the poor.

Where there are political parties actually competing for the votes of the poor, the poor
have a better chance of influencing policy or seeing policy formed that addresses their
needs. In the 1980s, competitive politics in Peru allowed the women’s movement,
organised through a vast network of community-based kitchens known as comedores,
to exercise limited but substantive influence over social policy, including the
enactment of new legislation (Houtzager, 1999). Kimberly Niles (1999) explains the
way in which the pattern of party competition affects the extent and way in which
parties compete for the votes of the poor. She compares countries in terms of the
extent to which parties are stable or fluid and fragmented. Where parties are stable,
elections are generally dominated by two to three competing programmatic parties
with a degree of party discipline. Each party needs to win a high proportion of votes
to enter government. Parties operate with relatively long time horizons. They
accumulate a great deal of information about potential voters and the potential pay-
offs to different political strategies. Each tries to appeal to a broad constituency of
voters. The system tends to produce parties and governments with high commitment
to the poor. In fragmented systems, politics are personalistic. There are many poorly
disciplined parties, that do not need to obtain a large proportion of votes to have a
chance of entering government. They have shorter time horizons and face higher
information costs because politicians are continually engaging in new activities, with
new allies or opponents, and seeking the support of different voters. These systems
tend to produce governments with low commitments to the poor.

Many developing country democracies remain characterised by just such fragmented

party systems. In some cases, social movements – or broad alliances of organisations
from within civil society – have emerged to give the poor a voice in national politics.
In other cases, civil society organisations are weak, or completely out of touch with
large poor populations, notably rural people. Large numbers of the rural poor in
South Africa, for instance, were never reached by political organisations during the
many years of apartheid rule. Their influence over national policy-making cannot be
expected to rise simply with the abolition of apartheid or the introduction of a
democratic constitution. Democratic transition, however, created new opportunities
for both the rural poor and their supporters to begin to organise, to become aware of
their rights and create new possibilities for voice. It may be that the first step in
creating political capabilities of the rural poor is simply providing a network of legal
aid services to inform the community of their rights under the law.
 8

Even when allied together in broad social movements, local community and non-
governmental organisations cannot play the same role in shaping public policy that is
played by well institutionalised programmatic democratic political parties. In
particular, they lack the mandate granted by the ballot box. These organisations can
help increase the political capabilities of the poor, influence national politics, and
perhaps lay the basis for more enduring, institutionalised and accountable political
parties.

While the poverty reduction record of democracies in developing countries is an

ambiguous one, there are reasons to believe that the opportunity for cohesive
competitive politics and the space for the poor to organise within civil society
permitted by democracy can contribute positively to poverty reduction. In the
absence of democratic politics, anti-poverty policy tends to be reactive and overly
determined by the legitimacy needs of a particular regime, or the preferences of those
who happen to be in power. It rarely contributes to increasing the political
capabilities of the poor. In Mexico, the long period of dominance by a single party
has led to an unstable anti-poverty policy dependent on the character of the sitting
president and marked by a lack of independent organisations among the poor (Kurtz,
1999). Some recently authoritarian regimes (South Korea, Taiwan, Indonesia) acted
positively to reduce poverty in order to gain and maintain legitimacy for their rule.
The challenge in developing countries today is how to make similar accomplishments
in poverty reduction a criterion for the legitimacy of modern democratic parties and
governments. This cannot be done through conditionality imposed from the outside,
but must come about as the product of domestic political mobilisation and education.

2. States create and shape the political opportunities for the poor

Empowerment of the poor is both an instrument to reduce poverty and, insofar as

powerlessness is part of the meaning of poverty, itself an aspect of poverty reduction.
It is useful to think of empowerment in terms of increasing the political capabilities of
the poor: personal political capabilities, self-confidence, capacity for community
organisation, recognition of dignity, and the collective ideas available to support
effective political action (Whitehead and Gray-Molina, 1999). It is the political
capabilities of the poor that will determine whether they can employ social capital
(the shared networks, norms and values created through social interaction)
constructively or create social capital where it is lacking. One of the problems in the
way social capital has been discussed to date is that scant attention has been paid to
the content and practice of politics within social networks (Putzel, 1997).

What circumstances lead to the enhancement of the political capabilities of the poor?
Citizens of developing and transitional economies have had so much experience of
oppressive and ineffective states over recent decades that the answer to this question
is often seen to lie in some kind of autonomous popular action. Citizens' movements,
NGOs, and civil society are seen as the alternatives to failing states. This is a
misleading myth. It is especially misleading in relation to movements of the poor. It
is certainly true that failing states can in some circumstances stimulate local level
alternatives. There are many accounts, for example, of how Ugandan rural
communities coped with a long period of civil war and then of a fragile peace by
taking control of and financing their own primary schools. But the operative word
 9

here is local. Effective, large scale organisation by poor people - the kind of
organisation that can make a consistent difference to public policy and affect a large
population - is dependent on the character of the state and the policies it pursues. One
can illustrate how states affect the scope for building the political capabilities of the
poor by examining three different levels of state capacity and action:

a) The macro-level: the character of states

People try to create the political organisations that will be most effective in the policy
arena where they wish to have influence. An important feature of that arena is the
pattern of organisation of the groups with whom they are likely to ally or compete.
National employers' organisations emerged in most industrial countries in response to
the creation of national trades union federations. National farmers movements
emerged only after governments began to intervene consistently in agricultural
markets, and there was a perceived need to deal with government at a national level.
Conversely, the mass-based civil rights movements that emerged to challenge so
many authoritarian Latin American and Central European regimes in the 1970s and
1980s fell apart once democracy triumphed and their focus - a consistently oppressive
state - had disappeared.

Poor peoples' movements respond to the same logic. Where the state is fragmented,
organisations of the poor hardly exist, or are the creation of external actors, like
international aid donors or non-governmental organisations (Houtzager, 1999). The
modern state eliminates rival centres of authority within society. This provides the
basis for social groups to organise on a national scale and to create collective
identities that cut across geographic regions. Whether social groups organise to
influence the state depends on whether they believe the state has the authority and
capacity to meet their demands. If the national state has little authority, why bother to
organise at the national level? Better to concentrate limited political resources - and
political resources are always limited - on exercising influence in different ways:
negotiating an acceptable level of informal taxes with the guerrilla movement in this
region; using ethnic linkages with a minister to remove oppressive policemen from
this town; building up connections with the aid donors who might provide money for
local NGOs in that district. Building large membership movements of the poor is
unlikely to be the most efficient means of exercising influence through these kinds of
channels. Global maps of effective states and effective social movements would look
very similar. Where the state is ineffective, social movements are rare, weak,
exclusive, localised and often closely connected with armed secessionists and
smugglers. Where states are ineffective, improving their capacity may be the best
way to stimulate effective organisations of the poor.

b) The meso-level: the pattern of public policy

Political resources are scarce. Political organisation is costly, especially for poor
people. They are more likely to organise when there is an incentive. Government
programs can provide powerful incentives. The mechanisms are sometimes obvious
and direct. If there is known to be a procedure for allocating public land to poor
people for house plots, there is an evident incentive to organise to try to secure access
to land. Similarly, the existence in the Philippines of a 'permanent land reform' -
legislation that enabled government to choose when and where to initiate the
implementation process - provided a standing incentive for potential beneficiaries
(and potential losers) to organise (Houtzager, 1999). However, many of the
 10

mechanisms through which public policy affects the mobilisation of the poor are
indirect and less obvious:

• 'Perverse' mobilisation takes place when, for example, governments succeed in

mobilising the poor by provoking them in some way - failing to truck in
emergency drinking water supplies into urban areas during drought, or forcibly
displacing people to build highways. This may be a powerful - but temporary,
localised and arbitrary - stimulus to the organisation of the poor.

• Unfulfilled promises may provide a more permanent stimulus. The success of

land reform in Kerala, India was precipitated by reform initiatives that were too
ambitious for the state to fulfill. Laws were passed but not implemented.
However, this encouraged the rural poor to mobilise to claim their legal rights
(Herring, 1999).

• Some of the most powerful incentives for the poor to organise can stem from a
sense of exclusion - the existence of a public program that benefits some people
but not other people who appear to have an equally valid claim. For example, one
reason why Sri Lanka became an early 'welfare state' and an exemplar of high
levels of human development at low levels of per capita income was the existence
of health and education services for the immigrant estate labour force. These
services were mandated by the British colonial government of Sri Lanka (then
Ceylon), at the insistence of the British colonial government of India, whence the
immigrants came. Their existence led to demands, from the 1930s, that similar
services be extended to the entire population (Wickremeratne, 1973).

Poor people are at a permanent political disadvantage. By providing incentives for

organisation, government programs can mitigate these disadvantages. If we are to
take a pro-poor perspective on public policy, the extent to which programs might
contribute to mobilising the poor should be one of the assessment criteria used. This
criterion may also lead us to think differently about the public-private divide in
general, and especially the issue of the role of NGOs in implementing public
programs. Like any other form of organisation, NGOs have advantages and
disadvantages. One of their disadvantages, from a perspective of those concerned
with the political capabilities of the poor, is that they are not governmental. They are
not legitimate objects of popular political mobilisation in the way that governments
are. Governments have widely understood obligations. It is legitimate in most
political systems to criticise governments for failing to provide for the poor, or for
performing badly. These concerns can be framed in terms of an accepted discourse of
rights. It is not possible to do the same thing in relation to NGOs. They are not
authorities against which rights can be asserted. All else being equal, the use of
NGOs as implementers of public programs is likely to de-mobilise the poor.

c) The micro-level: the shape of government programs

Public programs have political and economic costs as well as benefits. It is not being
suggested here that, because they help increase the political capabilities of the poor,
more government programs are better than fewer. Many other concerns enter into that
judgement. One of them is the extent to which particular programs are effective at
encouraging sustained political organisation by the poor. Little thought appears to
have gone into that question. It is one of the more obvious gaps in our knowledge.
 11

Joshi and Moore (1999) explore these issues on the basis of comparative case
material, stressing that the most important role for external agencies (government
agencies, NGOs) may not be directly to support the mobilisation of the poor, but to
create an enabling environment - an environment in which the poor have an incentive
to mobilise. At present, the environment in which poor people and external
organisations interact is frequently hostile to collective action by the poor, because
characterised by uncertainty, arbitrariness and inequality. External agencies should
focus more on creating incentives to collective action, above all by removing the
obstacles that they themselves create. Four dimensions of the performance or
behaviour of external agencies are cited:
• Tolerance - collective action on the part of the poor is more likely where the
political environment is not hostile and punitive.
• Credibility - the extent to which, in their relations with the poor, public officials
can be relied on to behave like good partners in an enterprise, i.e. to do their job
correctly, and to be reliable.
• Predictability - this refers to the form of external programs: the extent to which
they are stable over time in content, form, and procedural requirements.
• Rights - the extent to which (a) the benefits received under external programs are
recognised as moral or, better, legal entitlements, and (b) there are recognised
(preferably legal) mechanisms that the beneficiaries can access to ensure that
these entitlements are actually realised.

One of their case studies is the Employment Guarantee Scheme in Maharashtra, India.
Over the 23 years from 1975/6 to 1998/9, this massive scheme has provided an annual
average of 132 million work days, on 341,661 separate work sites - soil and water
conservation, small scale irrigation, reforestation, and local roads. When first
introduced, the Employment Guarantee Scheme appeared innovative and received
considerable attention from the outside world. It has received much less attention
over the past decade, and is gradually shrinking in scale, in large part because
economic growth has reduced the demand for off-season manual work at minimum
wages. Despite its many problems, the Employment Guarantee Scheme has been a
success. It continues to provide relatively cost effective and reliable income support
for significant sections of the rural poor of Maharashtra. Joshi and Moore
demonstrate that a major reason is that jobseekers, via political representatives of
various kinds, have continuously been mobilised to demand their rights. And that in
turn stems from the ways in which the framing and implementation of the Scheme
contribute to creating an enabling environment for mobilisation. These include:
• Incentives to collective action are built into the Scheme: a minimum number of
people need to be in search of employment before work sites can be opened.
• The public agencies involved enjoy a degree of credibility, and the Scheme is
predictable in important respects. It has been in place a long time, is likely to
continue because it is legally mandated, and is implemented effectively partly
because, unlike many (emergency) public works programs, it is being
implemented continuously. Public officials are skilled in managing public works
projects.
• Above all, there is a legal right to employment provided certain conditions are
met. These rights can be pursued through the courts, and have a great deal of
moral force.
 12

The Employment Guarantee Scheme is far from perfect. It is unlikely to be directly

replicable elsewhere. It does however illustrate how clever political and institutional
design of public programs can encourage the mobilisation of the poor while helping to
make those programs more effective. And that in turn suggests the importance of
examining these kinds of political considerations when assessing the potential poverty
impact of public programs.

3. There is no reason to expect that decentralisation will be pro-poor

Decentralisation is a popular prescription for the governance problems affecting poor

countries in particular. It is widely believed that decentralisation will also have pro-
poor impacts. The most common argument is that, because decentralisation by
definition involves bringing government closer to the governed in both the spatial and
institutional senses, government will be more knowledgeable about and hence more
responsive to the needs of the people. This is expected to lead to pro-poor policies
and outcomes. It is difficult to evaluate these kinds of arguments, because
'decentralisation' covers a very diverse range of phenomena. Two people can be
arguing about 'decentralisation' and have very different things in mind. The following
are some of the main questions one has to bear in mind about the character of
decentralisation:
• Is it political and administrative devolution - the assignment of more power to
lower levels within a nested hierarchy of territorial administration - or
deconcentration - the shifting of functions and personnel to a lower level unit
within a centralised administrative hierarchy?
• Does it also involve (more) democracy? This is often a tricky question, since a
programme may promise more democracy but, by empowering local elites may
actually reduce the possibilities for the voices of poor people to influence policy
outcomes.
• What level are we talking of? The implications of devolving authority from
central to state governments in Nigeria may be very different from those of
devolving from the township to the village level in China.
• Is decentralisation part of a broader change in the allocation of administrative
and political authority that also has centralising components? Many
contemporary public sector reforms involve shifting authority for operational
decisions to lower levels while granting higher levels increased power to
establish performance targets, monitor performance, and allocate resources. An
example would be giving school principals direct control of school budgets and
of personnel decisions, but requiring them to follow a detailed curriculum and an
intensive program of student and teacher performance assessment, with rewards
for improved performance and penalties for bad. This kind of 'decentralisation'
is often employed to give central government more control over rural areas and
outlying provinces and cities. The proponents of decentralisation sometimes see
only one aspect of a two-way relationship.

(Insert Table 3 - Review of Studies of the Impact of Decentralisation - about here)

Given the great diversity of processes labelled decentralisation, it is not surprising that
an intensive study of all available evidence by Crook and Sverrisson (1999) yielded
no support for the contention that decentralisation is intrinsically or generally pro-
 13

poor. Indeed, they found little hard evidence of any kind about impact. The
conclusions from twelve cases on which there was the most reliable information are
summarised in Table 3. Crook and Sverrisson assessed the programs along two main
axes: (1) impact on participation by, representation of, and responsiveness to, the
poor; and (2) impact on the social and economic position of the poor measured by
growth, equity, human development and spatial equity (between regions/localities).
They found an unambiguously positive impact of decentralisation only in West
Bengal state, India. The study suggested that four key sets of variables determine
performance in terms of both responsiveness and pro-poor social and economic
outcomes:

a) The politics of the relations between newly empowered local government and
central government
Central government needs to support the decentralised system with financial and
administrative resources and legal powers and also needs to have the capacity to
control and monitor their use. It also needs to have both an ideological commitment
to pro-poor policy and an active engagement with local politics to challenge local elite
resistance. In West Bengal and some parts of Brazil, pro-poor outcomes were the
product of this kind of synergy between action at the local and central/state
government level. Conservative local elites were challenged by local groups who had
the support of central authorities. In Bangladesh, Nigeria, and Kenya, although
substantial resources were transferred, central governments failed to ensure their
proper use. Conservative elites were empowered through decentralisation because the
main concern of central government was to create a power base in the countryside.
This eventually happened in Ghana as well. In Cote d’Ivoire and Mexico,
decentralisation was used by ruling parties intent on renewing party support without
any commitment to pro-poor policies, which might have threatened those networks.
These cases stood in contrast to apparently more successful programs in parts of
Colombia, where a strong political party or local elites seeking legitimacy were
committed to poverty reduction. In Bolivia, decentralisation has meant not the
abolition of long-established clientelist networks from the centre to the local areas,
but the introduction of competition in the form of newly empowered local networks
that contain some poor people (Whitehead and Grey-Molina, 1999). Evidence from
comparative work on Indian states reinforces conclusions about the pivotal role of
central-local relations in making decentralisation more favourable to the poor. In
India, the federal system has allowed enough room for state governments controlled
by alliances favourable to the poor to implement programs with more positive poverty
reduction outcomes than elsewhere. At the same time, the existence of central
government programs has often proved crucial to providing both the resources and
political leverage necessary for local authorities to implement policies favourable to
the poor (Harriss, 1999; Herring, 1999). Only after a national constitutional
amendment in 1993 did four states adopt a local government reform that reserved
33% of seats for women and seats for Scheduled Castes and Tribes according to their
proportions in the population.

(Insert Box 2 - Decentralisation in Colombia - about here)

b) The extent to which enhanced participation established accountability of local

governments
 14

Fair and competitive elections were crucial to establishing accountability in the more
successful cases. In Colombia, the most successful mayors in terms of responsiveness
and pro-poor outcomes were those who relied on a public constituency for their
office. More negative outcomes were seen in Cote d’Ivoire, where mayors had almost
no connection with the electorate. At the institutional level, it was important that,
even where the interests of poor people were represented, there was a proper balance
between political control and the legal accountability provided by a well-established
and reasonably autonomous administration.

c) Successful decentralisation required central authorities to provide some key inputs

These included in particular: secure and adequate earmarked funding; targeted
poverty reduction programs with built-in accountability; and support for a hierarchy
of authorities where intermediate levels of government (regional, district or state
level) could handle resources, raise additional revenues and provide administrative
support to local governments. Experiments in decentralisation in Ghana, Bangladesh
and Nigeria all suffered from inadequate staffing and poor management in newly
established local authorities.

d) Decentralisation required long term central support

West Bengal state, India had twenty years in which to develop its decentralisation
program, and achieved considerable success as a result.

What emerges clearly is that the success of decentralisation programs in securing pro-
poor outcomes depends in important ways on the role of central governments. In the
history of the United States, the role of the federal government has often been pivotal
in programs aimed at poverty reduction. Its failure to act on behalf of the poor at key
moments had devastating effects. The failure to break local landed power in the
South after the Civil War reinforced the long-term deprivation of the former slave
population (Herring, 1999). A wide range of evidence seems to warn against what
Herring calls ‘a premature celebration of the local’ in development strategies that aim
to reduce poverty.

4. There are a wide range of possibilities for pro-poor political alliances

When international aid and development agencies produce reports on poverty, they
rarely say much about the political dimensions. What they do say typically focuses
on conflicts of interests between poor and non-poor, and the need somehow to
overcome the resistance of the rich if governments are to be more committed to
poverty reduction. Some reference may be made to the need for political movements
that mobilise large numbers of poor people in order to counteract the influence of the
rich.

This perspective on the politics of poverty alleviation is especially attractive to

economists. Let us label it 'interest group economism'. It is not merely simplistic, but
also consistently misleading and pessimistic. It exaggerates potential political
obstacles to public anti-poverty action, and ignores some of the opportunities (Moore,
1999). Let us first illustrate the point with a case study.

Most members of the Brazilian political, business and governmental elite are in
favour of land reform (Reis, 1999). We can assume that the supporters do not include
 15

the minority of that elite who themselves are landowners. Even so, this finding will
surprise people who believe that political divisions will generally reflect economic
self interest, and that rich and poor will generally find themselves on different sides.
Why is the idea of land reform so appealing to rich Brazilians? Part of the
explanation does lie in perceived self-interest: '…. the optimistic expectation that land
redistribution would improve living conditions in the large cities where members of
the elite live. They appear to have a dream of exporting the poor to the countryside
where they could not only produce for their own consumption but even generate a
marketable surplus. …. (elite) respondents would systematically mention high
criminality rates and pressure on the provision of public goods in large cities as the
major consequences of poverty and inequality. They would blame poverty and
inequality for the lack of personal security, dirty and dangerous public spaces, and
related problems' (Reis, 1999: 131-2). But surely the Brazilian elite are deluded? No
feasible land reform is actually going to reduce the numbers of the poor in Brazil's big
cities. It will at best reduce the rate of rural-urban migration. True. But the fact that
members of the elite are willing to believe something different reflects the skill of
many politicians, especially those associated with the influential Landless Workers
Movement, in constructing a case for land reform that is both plausible and congruent
with what elites like to believe about themselves and the world. In this case, an
important element of the story is that ('feudal') landlordism is seen as old-fashioned, a
constraint on the modernisation of Brazil, and a potential stain on the image of Brazil
as a fast developing industrial nation. Elite Brazilians are open to arguments for land
reform that have little to do with their self interest in any direct sense of the term.

The Brazil case helps illustrate some points about the politics of anti-poverty policy
that are easily overlooked by people who begin from those assumptions about the
political process that we have labelled interest group economism:

a) Poverty alleviation can be a public good

The non-poor often will have a better life when there is less poverty, will recognise
that, and be prepared to support public action against poverty. The non-poor may
suffer from poverty because of the links to crime, disease, social unrest and poor
national economic performance through an inefficient (because unhealthy,
uneducated) labour force (Toye, 1999; Hossain and Moore, 1999). In addition, as the
Brazilian case illustrates, the wealthy may find poverty disturbing in less tangible
ways.

b) A sympathetic and politically-intelligent presentation of the 'facts' of poverty can

mobilise a great deal of political support for public anti-poverty action
Specialists debate endlessly whether poverty is better defined in terms of
income/consumption or more diffuse concepts of 'deprivation'. This is an important
issue from a political perspective. Presenting poverty in terms of deprivation (or 'lack
of capabilities') may often elicit the sympathy of the non-poor more effectively than
defining it purely in terms of low income or consumption levels. Publicising the fact
that X% of the population live below some poverty line of $1 or $2 per day is very
attractive to economists, statisticians and people involved in making international
comparisons. However, it may be an ineffective way of obtaining the support of non-
poor groups for public action against poverty. First, this poverty line measure lacks
moral content. There does not appear to be anything intrinsically wrong with $1 or $2
per day. Second, it may signal to the non-poor that poverty reduction is primarily a
 16

matter of dividing up a fixed cake: more dollars for the poor clearly implies fewer
dollars for someone else. By contrast, presenting poverty as deprivation - the inability
to feed children, to send them to school, or to give young women an alternative to
prostitution; powerlessness in the face of policemen extorting protection money; or
the helplessness of sick people who simply have no access to medical treatment - has
much more constructive political implications. These deprivations are intrinsically
moral. People who are not poor may imagine or perceive themselves as sharing
similar problems, and therefore empathise. In thinking of solutions, they may believe
that their interests overlap with those of the poor, rather than compete: better
schooling/security/public services may be a platform around which all can unite.

This contrast between deprivation and income concepts of poverty illustrates wider
points about poverty and politics. First, 'as politicians know only too well but social
scientists often forget, public policy is made of language' (Majone, 1989:1). The
terms in which issues are presented for public debate can greatly influence the
outcome of that debate. Second, there is a wide range of choice about how one
presents poverty issues. Neither the character, the causes of, nor the solutions to
poverty are immutable facts, clear to and agreed by all parties concerned. On the
contrary, they are all malleable concepts, which can be reinterpreted and represented
in a variety of ways, many of them conducive to public action against the poor. In
late Victorian Britain, Charles Booth helped effect a major opinion shift in favour of
public action to support the majority, ordinary working poor, by presenting the middle
and upper classes with a new interpretation of 'poverty'. Although he invented the
concept of the 'poverty line', Booth defined it more in social than in material terms.
This was the line below which even decent working families were likely to lose the
battle for respectability and fall into the ranks of paupers - the exotic, feckless,
threatening, immoral class against whom public 'poverty' policy had traditionally been
directed. The notion of respectability at risk was something to which the British
middle and upper classes could relate in a positive manner. 'Booth "re-moralized" the
poor by separating them from the "very poor" and freeing them from the stigma of
pauperism and degradation' (Himmelfarb, 1991: 11). In the immediate aftermath of
Booth's work, 'Instead of yet another revision of the Poor Law dealing with paupers,
Parliament enacted laws for the benefit of the laboring poor; housing, workmen's
compensation, extended schooling, school meals, old age pensions, cheap trains, labor
exchanges, unemployment and health insurance' (Himmelfarb, 1991: 12). In the early
twentieth century, American state governments were persuaded, mainly by national
middle class women's organisations, to spend public money on supporting poor
families on the grounds that this was the only way to protect the moral and physical
integrity of the nation (Skocpol, 1992).

c) The poor and the 'middle strata' often have common interests
The extent to which different fractions of a population have a common interest varies
widely. It depends in part on the policy issue involved and in part on the way in which
populations are divided into social or political groupings. If ethnic and religious
distinctions cut across income categories, the more prosperous groups may actively
support pro-poor policies that they believe will benefit poor members of their ethnic
group. Indeed, it is those aspects of shared identity cutting across class that have
achieved many gains for women. If income is relatively equally distributed, there will
be more similarities of income and lifestyle between different strata, and thus more
common interests, than in situations where income distribution is very unequal. If
 17

poor and non-poor live in the same localities, they are likely to work together to press
for better roads, water supply or health services. If they are physically segregated,
each group may seek to solve its own problems separately. The complexity and
uncertainty of the possibilities for political alliances involving the poor may be
frustrating for people searching for general truths. However, it can be a source of
inspiration for politicians seeking widespread support. They have incentives to
identify public programs that will benefit many people, alienate few and, if possible,
remain affordable. We can learn from their insights.

In 1980, the Government of Sri Lanka introduced a program to provide

schoolchildren with free textbooks in three major subjects (mathematics, first
language and religion). The education professionals and aid donors have accepted
that the program has a positive educational impact, and have considered extending it.
In a country where school attendance is almost universal, the program has been
popular and has benefited poor households significantly. After a decade, a later
government added in the provision of free material for school uniforms - a significant
bonus to most of the population, but above all to the poor, in a country where school
uniforms are obligatory. There were no popular demands for these programs. They
were conceived by politicians in power who were seeking to consolidate their
electoral support by demonstrating their understanding of the problems of ordinary
people, and providing visible benefits to large numbers of people at low cost and in
ways that would pass the scrutiny of aid donors in terms of educational and poverty
impact. The programs became popular only after people experienced them (Sunil
Bastian, private communication).

(Insert Box 3 - Agrarian Reform and Poverty Reduction: new Political Possibilities)

d) Governments have scope to take major initiatives in anti-poverty policies

The case of school textbooks in Sri Lanka illustrates two important, related
dimensions of the politics of poverty alleviation. The first is that governments are not
passive instruments of voters, pressure groups, party coalitions or socio-economic
forces. Especially in developing countries, where civil society groups are often
poorly organised and fragmented, governments have considerable scope to take policy
initiatives, to shape political debate, and to define what kinds of action are appropriate
for public agencies and technically feasible (Grindle and Thomas, 1991). The issue of
free school textbooks was a Sri Lankan government initiative. It was not a popular or
party demand, and had not been raised in election campaigns. The second point is
that, especially where government takes initiatives, the political feasibility of anti-
poverty strategies may depend more on avoiding opposition than on mobilising
societal support. It is easy to make the assumption that pro-poor policies require the
political mobilisation of the poor, to overcome the (potential) opposition of the non-
poor. In some contexts, notably most land reform, this is the right assumption. Even
if legislation is passed, the mobilisation of potential beneficiaries is often needed to
ensure implementation in the face of opposition at local level from property owners
(Box 3). These landowners are being deprived of something they wish to cling on to.
They will naturally use their influence to resist. Even in the case of 'market-led land
reform', the mobilisation of potential beneficiaries is generally needed to ensure that
they get a reasonable deal, and that landowners' considerable informal local power is
offset. But this kind of direct asset redistribution is but one of many kinds of pro-
poor policies. The nature of politics varies widely according to the context. In the
 18

case of school textbooks in Sri Lanka, the main potential sources of opposition lay
within the government itself, particularly the Ministry of Finance. In the absence of
opposition from this source, the program was quickly implemented and accepted as
routine. It is not a cure for poverty but makes a significant contribution at low cost.

Provided they can avoid stirring up opposition, governments can often take positive
pro-poor initiatives without the support of the poor. Indeed, the support of the poor
can sometimes be a problem, because it does threaten to stir up the non-poor, and to
mobilise opposition that might otherwise lie dormant. Let us conclude with some
quotations from the most extensive study ever undertaken of the politics of 'reformist'
pro-poor redistribution in poor countries (of Latin America) - William Ascher's
Scheming for the Poor (1984):

“The case studies show that the image of the poor confronting the rich is
profoundly misleading. On the one hand, the most effective strategy for securing
the political viability of a redistributive policy often is to gain the backing of a
selected part of the higher-income population. .... On the other hand, one of the
most serious problems of carrying out redistributive programs is that the already-
benefited poor often resist the spread of benefits to other segments of the needy”
(Ascher 1984: 34).

“.... the support of beneficiaries is of limited political value: those who have
already benefited from redistribution are not likely to behave in grateful ways, and
those who may benefit from contemplated redistributive policies are often
incapable of being mobilized sufficiently to help the government vis-à-vis typical
opposition tactics” (Ascher 1984: 310).

“There are many sub-sets of ‘the poor’, often with opposing interests with respect
to a given economic policy. Food subsidies for the urban poor, generally
engineered to the detriment of the rural poor, are the classic example. The
implication is that mobilizing ‘the poor,’ far from being a straightforward task,
turns out to be a complicated and often unrewarding exercise insofar as the
differences among lower-income segments are likely to be substantial and
politically divisive ” (Ascher 1984: 310-11).

Contrary to the pessimistic 'interest group economism' discussed above, there are
many reasons why non-poor groups and actors are willing and able to support pro-
poor policies, and a wide range of potential pro-poor alliances. We have not explored
the full range of possibilities. In particular, we have not discussed the important
contribution that religious organisations have often made to the creation of pro-poor
alliances and movements. To emphasise the role of non-poor groups is not to lose
sight of the importance of building up the political capabilities of the poor. That only
happens when organisations representing the poor develop some autonomy. In this
respect, non-poor allies can be both a help and a hindrance. Social science throws
light on the subtleties of such relationships, but warns us against drawing simple
conclusions (Houtzager and Kurtz, forthcoming).

5. Many of the policies needed to improve governance will benefit the poor
 19

Improved governance has been a development policy priority for some years. The
current focus of aid donors on poverty reduction might be thought to compete or
conflict with this. In fact, there is no competition. The two are in most respects
consistent with one another. Indeed, they are to a large degree complementary: better
governance is good for the poor, and many pro-poor policies imply or promote better
governance. One aspect of this complementarity has already been discussed above:
the fact that effective states are a condition for the creation of effective poor people's
movements. The basis for what is potentially the widest and most significant pro-
poor political alliance lies in the high incidence of bad governance in much of the
developing world: governments that are simultaneously oppressive and unaccountable
and lack the capacity and resources to provide basic services, including law and order,
for the mass of their populations. Almost everyone has an interest in better
governance; the poor feel this most strongly.

(Insert Box 4 - Lawlessness Contributes to Poverty - about here)

At one end of a combined poverty-governance agenda are tangible and relatively

straightforward measures in the fields of law and policing. Examples of how
lawlessness exacerbates poverty are given in Box 4. Lawlessness and oppressive
policing are not marginal or 'luxury' concerns as far as poor people are concerned.
The World Bank has recently sponsored a participatory assessment of poor people's
problems at 468 sites in 23 developing countries. Oppressive policing was identified
as a problem at the majority of sites. Oppressive policing is of course not an easily
soluble problem. But a start can be made by changing statutes that give the police
(prison officers, other public officials) almost unrestricted power over people who
have no effective access to courts, i.e. the poor. Unrestricted power breeds oppression
and extortion.

Providing police and other public officials with coercive legal authority can often be a
significant obstacle to the functioning of poor people's political organisations and
development of the political capabilities of the poor. Even in democratic India,
magistrates retain the statutory power to ban public assemblies - a power that dates
from colonial times and appears blatantly to violate the constitutional right to freedom
of assembly.

At the other, more complex, end of the combined poverty-governance agenda lie a set
of issues about the interactions between public expenditure, taxation, citizenship and
accountable government that are certainly not amenable to quick or simple solutions,
but are fundamental to both poverty reduction and good government, and merit
attention because they are so frequently overlooked.

Public spending does not necessarily contribute to poverty reduction. But it clearly
plays an important role in effective public action. And public spending, as a
percentage of GDP, is not high in developing countries. Indeed, it is low by any
comparative standard. It is low partly because overall levels of taxation are low. And
they are low in large part because the better-off people - those who could afford to
pay income, consumption, turnover or property taxes - actually pay relatively little tax
in most developing countries. The reasons for this are various. They include the fact
that structural adjustment programmes have generally involved a reduction in the
export and import taxes that have been a major source of public revenue in poor
 20

countries. While this has been balanced by the gradual spread of value-added taxes,
that are often effective at raising resources, aid and development agencies have not in
practice placed sufficient emphasis on increasing governments' tax take to pay for the
programs that are needed to reduce poverty and meet other urgent needs. The OECD-
based culture of capping total taxation and public expenditure at current levels has in
effect been exported to poor countries where levels are too low. In some cases at
least, this is because aid and development agencies are wary of upsetting the elites
with whom they do business - who would be visible losers from, for example, the
introduction and enforcement of reasonable levels of property and corporate taxation.

The link from taxation and public expenditure to poverty reduction is in principle
clear. There is however another causal link - from taxation and public expenditure to
citizenship and governance - that is less evident but no less important. There is a long
tradition in political science of viewing taxation as one side of a social contract
between citizens and states. The other side of the contract comprises the services that
states provide to citizens in return. This contract is more easily enforced by citizens
in a democracy. But the underlying dynamics still operate in non-democratic
environments. Where states are dependent for their incomes on taxing their citizens,
they face incentives (a) to treat their citizens reasonably and indeed to help enrich
them, in order to increase the tax base; and (b) establish a competent public
bureaucracy to raise tax revenues, with positive spill-over effects on the quality of
public services generally. The poor governance of many developing states stems in
part from the fact that their governments are largely independent of most citizens for
revenue: they are funded by large revenue inflows from direct control over oil wells
or other sources of mineral revenue and, to a lesser extent, large aid inflows.

This notion of a tax-mediated social contract leads to the expectation that the
developing country governments that are most dependent on their own citizens - for
revenue and other resources - would tend to perform best in terms of poverty
alleviation. This was the finding of one of our papers, that used cross-national
statistical analysis to examine the relative performance of national political-economic
systems at converting national income into human development (longevity, adult
literacy, school enrolment) for the mass of citizens (Moore et. al., 1999). Where
governments are (a) dependent on their citizens for revenue and (b) not propped up
politically and militarily by a former colonial power, levels of human development
are high in relation to average levels of income. Where governments have access to
substantial mineral revenues and enjoy strong political and military support from an
external power, levels of human development are low in relation to income levels.

In sum, better governance and poverty reduction to a large degree go together.

Measures to tackle one objective will tend also to contribute to the other. However, a
note of warning is needed: that conclusion depends in part on how one conceives and
defines 'good governance'. Different groups do have different needs in relation to
government. The type of governance that meets the urgent needs of the poor may not
much overlap with the needs of other groups. This was strikingly confirmed in the
statistical analysis by Moore et. al. mentioned above. There was a strong statistical
correlation between poverty reduction performance and the scores that countries were
allocated for 'quality of government institutions' by a reputed agency that works for
international investors. But the connection was perverse: countries with high ratings
performed badly at poverty reduction! Good governance as defined by international
 21

investors - basically security of investments and reliability of commercial contracts -

appears to be distinctively different from the more fundamental type of good
governance that we have identified as pro-poor.

LESSONS FOR THE DONORS

Those working in international development have long known that domestic politics
often determine whether an aid program or project can be effective in contributing to
poverty reduction. Nevertheless, little attention has been paid to increasing the
capacity within the agencies to carry out political analysis. It has often been easier to
approach development intervention as a series of technical problems abstracted from
the ‘messy’ world of politics, which in any case, is a domestic affair. In addition to
the strategic guidelines presented above, a number of other lessons emerge for aid
donors out of the study of political systems and poverty reduction:

• Donors need to undertake political analysis to understand the actual workings of

the political systems in the countries where they operate. The economic logic of a
particular project, program or policy intervention may be flawless, but political
analysis may reveal an entirely different set of concerns and conclusions.

• Donors must also recognise that they are, in fact, political actors. The choices
they make about the countries in which to work, about whom they consult and
with whom they deal are all highly political. Understanding who the potential
participants in reform coalitions are in any given situation will allow donors to
provide encouragement and support.

• Just as awareness of the social costs and environmental effects of alternative

projects, programs and strategies has led both governments and donors to
undertake their own social and environmental impact assessments, so donors
should undertake political impact assessments before embarking on major projects
and programs. Corporations often carry out political risk analysis and donor
agencies could learn from this practice.

• Simplistic assumptions about formal democratic institutions or transplanting

political standards of rich developed countries to developing country situations
through conditionality will not advance the cause of poverty reduction. Donors
need to acquire specific and expert knowledge about the political systems in
which they intervene and the dynamics that drive them if they are to understand
both the political determinants and consequences of their intervention.

• Developing the political capacities of the poor is a long-term enterprise which cuts
against the grain of the donor mentality (and domestic politics in the donor
countries) that wants to assess the ‘return on investment in aid’ over relatively
short periods of time. In terms of policy intervention, donors might be best
advised to adopt as a first principle, ‘to do no harm’. This implies that in some
cases, rather than trying to promote pro-poor policy, they should instead
encourage governments, in the first instance, to at least eliminate those aspects of
current practice and policy that are clearly ‘anti-poor’.
 22

Table 1: Types of State, Politics and Poverty

Types of Character- Types of Role of poor What can be Potential role
political istics policies in in politics done to of official
systems place promote the external
interests of the agencies
poor?
Collapsed There is no No policies No distinct role; Establishing Concentrate on
states effective central politics is effective central helping re-
government dominated by government. establish
force central rule
Personal Rule through Policies are No distinct role, More Improved
rule personalities unstable; a unless poverty institutionalised governance,
and personal major objective overlaps with rule will benefit including
connections. If is to enrich political most poor effective central
political parties those in power; networks based people and is a rule and human
exist, they are few basic public on ethnicity, condition for and civil rights
based on services are language, effective
personalities. provided religion or organisation of
locality poor above local
level.
Minimally An unstable There exist As above, There is some As above, with
institution mixture of organisations to except that scope to more scope
-alized personal and provide a range there is more organise around directly to
states impersonal of basic public scope for demands for support pro-
rule, with and welfare political parties improved public poor
varying services; or social services and a organisations
degrees of coverage is movements to less repressive and to
legitimacy. patchy and often have influence system of condition aid on
Parties are based on governance pro-poor
based partly on patronage policies and
personalities absence of
repression
Institution Rule through A wide range of Regimes try to There is As above, with
-alized stable and basic and suppress or considerable a case for
non- legitimate welfare services control all scope to stronger
compet- organisations may be organisations, organise around conditionality.
itive and provided, but but may be demands for More potential
states procedures; no citizens have responsive to both improved to help develop
open little influence popular public services civil service
competition for over the range concerns. and a less expertise
power. Political and type of Conflicts within repressive
parties serve provision ruling elites system of
the regime or may provide governance
are hindered opportunities
and controlled for popular
by it organisations
Institution Rule through A wide range of The poor have Potentially high. More potential
-alized, stable and basic and scope to Very to support
compet- legitimate welfare services. organise; but disadvantaged appropriate
itive organisations The range and other axes of groups (e.g. decentral-
states and type of provision political physically isation. Non-
procedures; are major mobilisation, handicapped, official
open themes in such as small ethnic channels are
competition for politics ethnicity and minorities) may appropriate for
power through language, may find a voice much
programmatic dominate governance
parties support
 23

Box 1: Competitive Politics and the Poor in Two Indian States

The Indian states of Karnataka and Andhra Pradesh have a number of

significant features in common. Both are middle-income states in the south,
generally considered to be both economically more dynamic, and socially more progressive
than the states of the north. Yet research has shown that, while Andhra Pradesh has a
good record in poverty reduction, Karnataka has not. Amongst the reasons for this contrast
are that the local political dominance of particular landed castes persists in Karnataka, while
party competition in the state is fragmented and involves different factions headed by
members of those same castes relying on clientelistic recruitment. The structure of local
power in Karnataka means that agricultural growth is less pro-poor than it might otherwise
be. The nature of party competition makes the political system less responsive to poorer
people. In Andhra Pradesh, on the other hand, local-level political dominance has been
reduced to a greater extent, and agricultural growth has been more pro-poor (having
created more productive employment, with rising real wages) while stable two-party
competition has created conditions in which political leaders are more responsive to poor
people. A populist party, relying initially on charismatic leadership, has become fairly well
institutionalised, and in competing with the Congress party, has successfully delivered pro-
poor programmes. Through the extension of the public distribution system into rural areas
in the later 1980s rural poverty continued to decline in Andhra Pradesh, in spite of a down-
turn in the rate of agricultural growth in the state. Though there are legitimate concerns
about the extent to which the costs of the public distribution system have eroded public
investment, and hence about the sustainability of poverty reduction, the growth rate of state
domestic product has been higher in Andhra than in most other states, including Karnataka.
It does not seem to be the case, therefore, that poverty reduction has been achieved at the
expense of economic growth, though the state government does now face difficult choices
in the allocation of public finance.
 24

Table 2: Distribution of Countries According to Poverty Reduction

Performance and Degree of Democracy
No. of countries More democratic Less Undemocratic Total
countries democratic countries
countries
High poverty
reduction 5 5 9 19
performance
Modest poverty
reduction 2 10 12 24
performance
Low poverty
reduction 4 3 11 18
performance
Total 11 18 32 61
Notes: Poverty reduction performance is conceived as the efficiency of national political-
economic systems at converting national income into longevity, literacy and education for
the mass of the population. For more details on the measures used here, and the sample
of countries, see Moore et.al. (1999). The measures of poverty reduction performance
relate to 1995.
The measures of the degree of democracy are from the POLITY III database. An average
was calculated for the period 1980-89, on the basis of annual scores on a scale of 0-10.
'More democratic countries' scored 8 or more, and 'undemocratic countries' scored 0.
 25

Table 3: Review of Studies on the Impact of Decentralisation

Outcome
Country Participation by & responsiveness to the Impact on social and economic poverty
poor
Particip- Represent- Responsive- Growth Equity Human Spatial
ation ation ness Develop- Equity
ment
W. Bengal, Improved Improved Improved Improved Improved Improved No
India evidence
Karnataka, Fair Improved Low Low Low Fair Fair
India
Colombia Ambiguous Ambiguous Improved Inadequate evidence Improved Improved
Philippines Improved Improved Ambiguous Inadequate evidence
Brazil Low but limited evidence Little Mixed* Mixed* Low
evidence
Chile Inadequate evidence Improved Improved Ambiguous to Low
Cote d’Ivoire Low Low Very Low No evidence Improved
Bangladesh Fair Very Low Low Very Low
Ghana Improved Fair Low Low but little evidence Fair
Kenya Very Low Low Fair
Nigeria Low Low Very Low Low Very Low Low
Mexico No evidence n.a. Low Low Low
* = Improved only in exceptional states
 26

Box 2: Decentralisation in Colombia

The decentralisation reforms in Colombia after 1982 attempted a democratic opening in local
government and reallocated sectoral functions to municipalities, away from quasi-autonomous agencies.
Mayors were to be elected, plebiscites on particular issues allowed, and consumer and local voices were
to be represented on local development agencies. The reforms can be interpreted either as a significant
democratic opening, or as an attempt to devolve conflict to the local level, making it more manageable
for the national government. The major motivation behind the reforms was an attempt by the two
traditional parties to restore legitimacy and rebuild local power bases in the face of growing opposition
from sub-municipal juntas.

There is some evidence to suggest that democratic decentralisation has increased the representation of
non-elites in Colombia. In 1988, the Liberal and Conservative parties – an effective oligarchy at the
municipal level for a century - had 80% of the popular vote, increasing to 90% in 1990, and controlled
almost 90% of municipalities. By 1992, this had decreased to 65% of the popular vote and non-
traditional parties controlled about 300 of Colombia’s 1,007 municipalities.

Some municipalities have adopted a participatory approach to local governance. In Valledupar, local
government staff wear badges which proclaim “we govern with your participation”, and the mayor has
established various means of dissemination through local media. In other municipalities, community
participation occurs in just one sector, or independently of the municipal administration. Some of these
participatory practices pre-date decentralisation and in many places traditional elite run patronage
politics remain strong.

In terms of responsiveness, a key element of the Colombian reform programme is the move towards a
‘demand-driven’ approach to public services, involving extensive participation. Competition for political
office acted as a catalyst for responsible leadership, which in turn became the driving force behind
capacity building, but performance still depends heavily on the leadership qualities of individual
administrations. There has been an increase in ‘voice’, with protests leading to local government action
and some local governments establishing channels for systematic expression of needs and problems by
the community. The World Bank team studying this case found quite reliable evidence of improved
performance in the fields of education, water supply and local roads. (Crook and Sverrison, 1999)
 27

Box 3: Agrarian Reform and Poverty Reduction:

New Political Possibilities
In recent years there has been a revival of interest in redistributive agrarian reform programmes
in many developing countries due to both changes in agriculture and the possibilities for new
coalitions in support of reform as one component of a fight against rural poverty. Agrarian reforms
were central to poverty reduction in countries and regions such as China (and Taiwan), Vietnam,
South Korea, and the Indian states of West Bengal and Kerala and in recent years reform has
been on the agenda in the Philippines, Colombia and South Africa. New demands for agrarian
reform have animated politics in Brazil, Thailand and other developing countries. Agrarian reform
constantly reappears on the public policy agenda due to persistent rural poverty and landlessness
in countries where a substantial number of people still depend on agriculture for a sizeable
proportion of their income. With overwhelming evidence of the positive economic potential of
small farming across a wide range of agricultural and ecological environments and the removal of
Cold War ideologically inspired opposition to property rights reforms, new possibilities have
emerged to employ agrarian reform in the fight against poverty.

In a study of the political conditions for agrarian reform, Ron Herring (1999) identified seven
reasons why the prospect of using agrarian reform to fight poverty has become politically more
possible:

• Both the nature of agriculture and its place in most developing economies has changed and
landownership is less central as a site of accumulation than it once was, which has reduced
its political salience;
• Land is more clearly understood as an anchor of natural systems and redistributing
ownership and stewardship rights has become important to those concerned with
environmental sustainability;
• Increased recognition of the gender differences in patterns of poverty has opened up a new
public policy debate about the distribution of rural assets and brought women more clearly
into the possible alliances backing redistributive agrarian reform;
• New forms of organisation of indigenous peoples and claims for ancestral domain and world-
wide interest in their cause form yet another potential political ally backing redistributive
reforms;
• Democratic transitions have offered new possibilities for putting reform on the political
agenda and there is some grounds for understanding that redistributing rural assets is crucial
to developing a sense of citizenship among the rural population;
• Technological changes in agriculture have made it increasingly possible to farm productively,
whether on a part-time or full-time basis, on smaller areas of land in many agricultural
systems;
• The proliferation of NGOs and human rights movements has created a more conducive
environment for the development of rural agrarian movements.

Redistributive reforms have known and proven results in reducing poverty which makes the
ethical case for supporting them much more clear cut than many indirect approaches to poverty
reduction. Where thorough going reforms have been carried out, the political power of the
minority who might oppose other poverty reduction measures has been significantly reduced.
Even in sites of partial reform, like in the Philippines and Brazil, new possibilities for the
mobilisation and organisation of rural poor people and their allies have been created (Putzel,
1999; Houtzager, 1999).
 28

Box 4: Lawlessness Contributes to Poverty

• At a hospital in Babati district in Tanzania, a new delivery of essential medical supplies
purchased with foreign currency disappears from the public dispensary within hours but is
available for purchase at a doctor’s home that evening. The poor do not receive the free
medical care promised by the government, but those with the right connections who are able
to pay are able to secure pharmaceuticals in abundance.
• In Muzaffarpur, Bihar, a man found innocent at a criminal trial is illegally imprisoned without
explanation. He is held in Bihar’s notoriously harsh prison conditions for 9 years before the
government enquires into his condition, but he is still not released for another five years.
When he is finally freed after 14 years of wrongful imprisonment, the Bihar government is
unable to come up with an explanation for his detention, although his innocence is in no
doubt. During his time in jail he has not only suffered irreparable injury to his physical and
mental health, he has also lost wages from fourteen of the most productive years of his life.
• In Johannesburg, South Africa, rates of theft and violent crime are some of the highest in the
world. Wealthy residents are able to employ sophisticated alarms, security guards, and other
forms of private policing to protect their property and persons, while the poor are stuck in
poorly-built homes sometimes without even simple locks, and are left vulnerable to theft,
assault, murder, and other violent crimes.
• A man who is too old to work is left without assets or income after his son is murdered. In
order to gain access to his son’s estate, he requires a ‘succession certificate’ from the civil
court of first instance in Lahore, Pakistan – over 160 kilometres to the south of his village.
The cost of the train journey and the bribe demanded by the clerk of the court send him
deeper into debt, but after five separate trips to the court over as many months, he has still
not been given the stamped piece of paper to which he is legally entitled. Each trip to court
send him further into debt, but the clerk refuses to produce the certificate while the authorities
in his home village refuse to give him access to his son’s assets until the certificate is
produced.
• Government-promoted oil development in the Oriente region of Ecuador has disrupted the
lives of the Huaorani people who have lived in the area for centuries. The oil operations have
driven away wildlife, disrupted food supplies, polluted drinking and bathing water, and
introduced new diseases. Both oil company security men and police have been involved in
killing members of the free clans of Huaorani living in the forest. In these circumstances,
malnutrition and disease have increased among the Huaorani. While most of these activities
clearly plainly amount to violations of the right to life as well as other rights guaranteed under
the 1983 Ecuador Constitution, there are no effective means of redress before the Tribunal of
Constitutional Guarantees.
(Anderson, 1999)
 29

REFERENCES
(Papers prepared for the purposes of this project are starred -*)

*Anderson, M. R., 1999, ‘Access to justice and legal process: making legal institutions responsive to
poor people in LDCs’.
Ascher, W., 1984, Scheming for the Poor. The Politics of Redistribution in Latin America,
Cambridge MA and London: Harvard University Press
*Crook, R. C., and Sverrisson, A. S., 1999, ‘To what extent can decentralised forms of government
enhance the development of pro-poor policies and improve poverty-alleviation outcomes?’.
Grindle, M. S., and Thomas, J. W., 1991, Public Choice and Policy Change. The Political Economy
of Reform in Developing Countries, Baltimore and London: The Johns Hopkins University
Press.
*Harriss, J., 1999, ‘How much difference does politics make? Regime differences across Indian States
and rural poverty reduction’.
*Herring, R. J., 1999, ‘Political conditions for agrarian reform and poverty alleviation’.
Himmelfarb, G., 1991, Poverty and Compassion: The Moral Imagination of the Late Victorians,
New York: Alfred Knopf.
Hossain, N., and Moore, M., 1999, ‘Elite perceptions of poverty: Bangladesh’, IDS Bulletin, Vol. 30,
No. 2, pp. 106-16.
Houtzager, P. P. and Kurtz, M. (forthcoming), 'The Institutional Roots of Popular Mobilization: State
Transformation and Rural Politics in Brazil and Chile, 1960-95', Comparative Studies in History
and Society.
*Houtzager, P. P., and Pattenden, J., 1999, ‘Finding the shape of the mountain: when ‘the poor’ set the
agenda’.
Joshi, A., and Moore, M., forthcoming, ‘Rights, institutions and poverty’ in A. Booth and P. Mosley
(eds), The New Poverty Strategies: What Have they Achieved?, Macmillan: London.
*Kurtz, M., 1999, ‘The political economy of pro-poor policies in Chile and Mexico’.
Majone, G., 1989, Evidence, Argument and Persuasion in the Policy Process, New Haven and
London: Yale University Press.
Moore, M., 1999, ‘Politics Against Poverty? Global Pessimism and National Optimism’, IDS Bulletin,
Vol. 30, No. 2, pp. 33-46.
*Moore, M., Leavy, J., Houtzager, P. and White, H., 1999, ‘Polity qualities: how governance affects
poverty’.
*Niles, K. J., 1999, ‘Economic adjustment and targeted social spending: the role of political institutions
(Indonesia, Mexico and Ghana)’.
Putzel, J., 1997, ‘Accounting for the dark side of social capital’, Journal of International
Development, Vol. 9, No. 7, pp. 939-49.
Putzel, J., 1995, ‘Managing the main force: the Communist Party of the Philippines and the peasantry’,
Journal of Peasant Studies, July, pp. 645-71.
Reis, E., 1999, ‘Elite perceptions of poverty: Brazil’, IDS Bulletin, Vol. 30, No. 2, pp. 127-36.
Rieger, E., and Liebfried, S., 1998, ‘Welfare State limits to globalization’, Politics and Society, Vol.
26, No. 3, pp. 363-90.
Rodrik, D., 1998, ‘Why do open economies have bigger governments?’, Journal of Political
Economy, Vol. 106, No. 5, pp. 997-1032.
Skocpol, T., 1992, Protecting Soldiers and Mothers. The Political Origins of Social Policy in the
United States, Cambridge MA and London: The Belknap Press of Harvard University Press.
Toye, J., 1999, ‘Nationalising the Anti-Poverty Agenda’, IDS Bulletin, Vol. 30, No. 2, pp. 6-12.
*Varshney, A., 1999, ‘Democracy and poverty’.
*Whitehead, L., and Gray-Molina, G., 1999, ‘The long term politics of pro-poor policies’.
Wickremeratne, 1973, ‘The emergence of a Welfare Policy, 1931-1948’, in K. M. de Silva (ed.),
History of Ceylon, Volume 3: From the Beginning of the Nineteenth Century to 1948,
Peradeniya: University of Ceylon.