Democracy and Poverty
Democracy and Poverty
Democracy and Poverty
Ashutosh Varshney
Associate Professor of Government and International Studies
University of Notre Dame
Notre Dame, IN 46556
Email: [email protected]
August 1999
South Korea, Taiwan and Singapore? There is a political, if not an economic, puzzle here, and it
is not often recognized. Unlike the US or Europe where less than 5 per cent of the population
lives in poverty, the poor in developing countries constitute a very large proportion of the
population. Their numbers being so small, the poor in the richer economies can hardly turn their
numbers into a electoral or political check on the government. What policies are developed
depend primarily on the conscience of the elite. In the developing world, however, the poor
constitute a majority or a large plurality. The numbers of the poor ought to exercise greater
pressure on the government, if the poor also have the right to vote and the political system is first-
past-the-post, in which a thirty per cent voting bloc can often be decisive.
accountable to the poor than authoritarian systems are. But no poor and relatively stable
democracy -- India, Sri Lanka, Malaysia, Costa Rica, Venezuela, Botswana, Jamaica, the
Philippines (democratic for much of its independence)1 -- has beaten the record of the three
countries mentioned above. Singapore, South Korea and Taiwan have mostly been authoritarian
in the last four or five decades,2 but poverty has virtually vanished from their lands. Democracies
in the developing world still have a large proportion of their populations stuck below the poverty
line.
1
Except a period of a decade and a half between the late 1960s and mid-1980s.
2
Singapore has been authoritarian since its birth in the mid-1960s, and South Korea and Taiwan conquered
poverty in the three decades they were dictatorial -- between the mid-1950s and mid-1980s. Both have become
democratic only in the last ten years.
2
It should, of course, be noted right away that South Korea, Taiwan and Singapore may
be exceptional. We know that many authoritarian polities, especially in Latin America, have
been very insensitive to the poor. I am, however, not asking why poor democracies have not
done as badly as poor Latin American dictatorships have. That question is not much of a puzzle
for, given the theoretical postulated effects of the size of the poor on democracies, we would in
any case not expect democracies to perform as badly. The more puzzling and interesting question
is why democracies have not done as well as East Asian authoritarian systems.3 Why is the
Surprising as it may seem, not enough is known about the relationship between
democracy and poverty.4 A great deal of literature is available on the relationship between
democracy and economic growth,5 but unless we assume that what is good for economic growth
hitherto well-known qualifications -- the implications of the literature on economic growth are
economic strategies do. Thus, the relevant questions are : Are there better or worse pro-poor
In one political respect, policies towards the poor can emerge from the same source in
democratic and authoritarian polities. Both can align themselves with the poor, if the political
elites have a consistent commitment to them, and they also force the state structure – especially
its bureaucratic institutions -- to translate that commitment into an equally consistent policy
3
Before its financial collapse came, it should further be noted, Indonesia too had made one of the most remarkable
attacks on poverty in an authoritarian political setting.
4
Among the few detailed arguments, and one very different from what I say below, is Kohli (1987).
5
For a summary and discussion, see Barro (1997).
6
The literature on his subject is voluminous. The earliest reservations go back to Bhagwati (1958).
3
implementation. Such pressure from above can mark both systems.7 In another and vital respect,
however, the two systems are very different. In a democratic polity, the poor can also exercise
their weight (“agency”) and push the government’s economic policy towards their interests in two
ways: political mobilization and/or voting. Both mobilization and voting can be viewed as forms
of pressure from below. This mechanism, unlike the pressure from above, is available in
democratic systems, but not in authoritarian polities in a regular and periodic manner.
Do poor democracies feel enough pressure from below? If they don't, why is that so? If
they do, what causes the gap between pressures from below and the actual outcomes? These are, I
My answers are two-fold. The first is that we not only need to make the standard
distinction between the direct and indirect methods of poverty-alleviation, but also between their
political and economic significance. The direct route covers public provision of income or assets
to the poor and the indirect method is growth-mediated -- not any kind of growth, but one that
aims at creating more opportunities for an increase in the incomes of the poor. That both
methods can make alleviate poverty is beyond doubt. But, as Bhagwati (1988) has convincingly
argued, the key economic issue is: which method is more productive (how resources are used)
and more sustainable in the long-run (how long the provision of public resources can be financed,
After decades of development experience with mass poverty, the mainstream economic
wisdom has veered round to the superiority of the indirect method and, I think, for good reasons.
For reasons of space and focus, I will not argue about the relative economic merits of direct and
indirect methods here. I will instead concentrate on the political issues, believing (and assuming
for this paper) that the growth-mediated, indirect approaches to poverty-alleviation are better.
7
The record of poor democracies, however, indicates that to expect state structures to act consistently in favor of
the poor in the absence of a sustained pressure from below is to expect what is relatively uncommon, not what is
typical.
4
The political implications of direct and indirect methods of poverty-alleviation are
remarkably different from their respective economic merits. Whether they are economically more
productive or sustainable in the long-run, direct methods -- asset transfers or income transfers -
have a logic that is clear, immediately obvious and have effects that can be quite tangible. Most
people, both politicians and others, can see the links proposed. Via land redistribution, land
reforms can give land to those who have too little of it, or none at all; and via tenancy reforms,
they can make poor tenants less dependent on the power and whims of landlords, potentially
imparting a more secure source of income. A similar directness marks the symbolism of income
In contrast, the utility and value of the indirect methods of poverty-removal may be
obvious to the scholars and specialists of development, but how exchange rate devaluations, trade
market-oriented economic strategy -- might more successfully attack poverty is not easy to
understand in political circles. However effective, the links proposed by the indirect method are
subtle, if not unclear and convoluted, and are also based on a long-run perspective. Such indirect
methods do not normally appeal to democratic politicians, for they not only have to understand
arguments, but also must try and carry constituencies with them and renew mandates in a
relatively short-run. That is why in no developing country has mass politics, in which large
numbers of average or poor citizens get involved (as opposed to elite politics confined mostly to
the Westernized upper and middle classes), pushed for trade liberalization, currency devaluation
and a market-oriented economic reform. A political constituency for economic reform may exist
in the middle and upper classes, but it is still to be built among the poor.
central of the democratic political process (Varshney, 1998). Pressured more by mass politics
than authoritarian systems, where people are not free to organize, democracies tend to be more
inclined towards direct methods of poverty alleviation. And since it can be shown that the direct
5
methods are less sustainable or less effective in removing poverty in the long run, one can also
understand why democracies in the developing world have been less successful than Singapore,
My second argument has to do with yet another distinction: that between class and
ethnicity. At its core, class is an economic category, but ethnicity is defined in terms of a birth-
based (ascriptive) group identity. Ethnic politics of subaltern groups is typically not couched in
terms of poverty, but in the language of dignity and social justice, in which poverty is only a
end to everyday degradations -- in schools, fields, places of work and worship, and on roads and
public transport. If the poor, irrespective of the ethnic group they come from, were to vote or
mobilize strictly on economic grounds, they would also press the decision-makers to attack
poverty a great deal more forcefully. However, it is becoming increasingly clear, at least in
multiethnic democracies, that it is not only easier to mobilize the poor as members of identity-
based communities, but that is also how they primarily vote – via ethnicity, not via class.
That being so, even with direct methods, a democratic polity is better able to attack
poverty a) if ethnicity and class roughly coincide for the poor, not if they clash, and b) if the
subaltern ethnic group(s) are also relatively large in size. If the poor belong to very different
ethnic groups (defined by caste, language, race or religion), the pressure on the political elite to
and b) because the poor are typically not from the same ethnic group. The former hurts the poor
because it can be shown that the indirect, market-based methods of poverty-alleviation are both
more sustainable and more effective; and the latter goes against them because a split between
ethnicity and class militates against the mobilization (and voting) of the poor as a class and
6
CLARIFYING TERMS
Before I proceed further, let me state the sense in which I am using the terms "poverty"
and "democracy". Since multiple definitions are available, we must pre-empt speaking across
intellectual registers, a problem especially common when interdisciplinary work is done and
presented.
By now, there are two ways in which the term "poverty" is used. The standard usage of
the term is narrower and consumption-based, focussing primarily, though not exclusively, on a
caloric floor that the human body, on average, minimally needs to function "normally". In this
narrower sense, hunger and endemic malnourishment more or less define poverty. In the richer
parts of the world, we typically try to reduce the number of calories our bodies consume every
day. In the developing world, the first challenge is not to reduce, but to provide a minimum of,
calories to millions. The $1/day yardstick used in PPP terms by the World Bank conforms to this,
The term "poverty" is also sometimes viewed more broadly. Going under the rubric of
human development, the broader usage includes education and health, in addition to consumption
(hence income) in its calculations. I will not use the term "poverty" in this sense. The reason is
not that education and health are not valuable. It is simply that while in the narrower sense we
can be more precise about what we are measuring for we have a "poverty line" clearly defined,
we don't yet have a composite index of poverty in terms of human development. What minima of
income, education and health -- combined with what weights -- would constitute the poverty line
There is no answer to this question yet. Nor is it clear there can be one, for it will require
specifying a certain life expectancy at birth and a certain level of education as vitally necessary
for the realization of human capabilities. That they are necessary for any worthwhile conception
7
of human life may be beyond doubt, but what level will constitute the minimum acceptable
The human development literature works better when we ask philosophical questions
about whether income alone can make human beings realize their capabilities. Its utility is
limited when we deal with the millions, who don't have minimum incomes, or minimum
nourishment, in the first place. Consider what a 1/day bought in India in 1993-4 (The World
Commodity Consumption/day
In 1993-4, after buying all this, the person would have about Rs 2/day (5 US cents) left to
purchase non-food items. About one third of India could not afford the meager bundle above. If
this is how a large part of the developing world lives, it will be best to conceptualize poverty in
Democracy as a term is also not clear-cut. In the seminal texts of democratic theory
(Dahl, 1989 and 1971; Huntington, 1983), two basic criteria have been used: contestation and
participation. The first principle, in effect, asks how freely does the political opposition contest
the rulers, and the second inquires how many groups participate in politics and determine who the
8
Where does this leave us with education? Now that we know that universalization of literacy was achieved in the
West and near-universalization in much of East Asia when those societies were poor (Weiner, 1991), we can indeed say
that primary education should, and can, be provided as a public good by the state at fairly low levels of GDP. But to
incorporate education as a measure of poverty, when nutrition is so abysmally low, is to complicate the matter
unnecessarily. One should defend universal literacy as a value and expect and push states to do more about it, but it is
less useful to make illiteracy constitutive of poverty line in the developing world.
8
rulers should be. The first principle is about liberalization; the second about inclusiveness (Dahl
1971).
In democratic theory, poverty does not figure prominently. Democratic theorists expect
that if socially or economically unequal citizens are politically equalized and if the deprived
constitute a majority of the electorate, their political preferences would, sooner or later, be
reflected in who the rulers are and what public policies they adopt. By giving everyone equal
it does not necessarily happen invites reflection on whether democratic politics is enough to
ostensibly deals with this point, but it muddies analytical waters immeasurably (Jalal, 1995). The
former emphasizes freedom of voting and mobilization and the latter stresses economic equality.
However, if we want to know whether democracies tend to reduce poverty, such a distinction
simply cannot be made. For it ends up conflating the explanadum and explanan, or to put it
differently, the independent and dependent variables. We can not reasonably figure out the
A more coherent and alternative line of reasoning can be derived from democratic theory:
If poverty, despite democratic institutions, comes in the way of a free expression of political
preferences, it makes a polity less democratic, but it does not make it undemocratic. So long as
"more or less"), not a discrete or dichotomous variable (expressed as "yes or no"). Variations in
degree and dichotomies should be clearly distinguished. In the classic formulation of Dahl, the
United States was less of a democracy before the civil rights revolution of the mid-1960s, though
it can in future be even more democratic if economic inequalities come down further (Dahl, 1971,
9
p. 29).9 Given contestation and participation, greater economic equality and/or an absence of
poverty certainly make a polity more democratic, but greater equality or elimination of poverty,
poverty-alleviation in the developing world? There are three difficulties that make a relatively
First, the question about the impact of democracy on poverty is a question that requires
identifying which countries have been relatively stable democracies - i.e., democracies for a long
enough period. In the developing world, countries that meet the criterion of longevity are few
and far between: India, Sri Lanka, the Philippines between the late 1940s and 1960s and after the
mid-1980s, Botswana, the former British colonies in the Caribbean, Costa Rica and Venezuela
since 1959 (Huntington, 1983;Weiner, 1989). Some would add Malaysia to this list as well, but it
participation may be high but contestation is limited between political parties by consensus, and
political competition, by agreement, is designed around ethnic groups, not individuals (Lijphart,
1977).10 Malaysia, in other words, is a particular kind of democracy, not one in the standard
sense. For our purposes here, we can count it as a democracy, given that universal-franchise
elections are regularly held, so long as we remember the specific nature of electoral competition
9
Analogously, by allowing a great deal of contestation but restricting participation according to class (and also
gender), England in the 19th century was less democratic than it is today, but it was democratic nonetheless, certainly
by 19th century standards.
10
Political parties in India and Sri Lanka may also seek to represent specific ethnic groups, but there has been no
constitutional pact, or political requirement, that that should necessarily be so. Parties are free to build cross-ethnic
alliances, if that aids their political fortunes.
10
At any rate, the total number of stable democracies in the developing world is small.
Many more countries have become democratic in the last two decades than ever before, but if we
look at the entire post-1945 period,11 one will be hard-pressed to add countries to the small
number listed above. In contrast, the number of authoritarian countries, which remained so for
long periods after 1945, is large. This asymmetry means that we don't have a large number of
statistical analysis of their economic consequences. Until the current wave of democracies -- the
third wave (Huntington, 1991) -- has produced many more democratic observations for inclusion,
we will have to primarily settle for nuanced qualitative reasoning (King, Keohane and Verba,
1994).12
polities had roughly the same economic strategy, we could hold economic strategy constant, and
clearly identify what difference the nature of the polity makes. However, as is well-known, what
is different across countries is not simply the nature of political system, but also the strategy of
Taiwan, Singapore) have all come from countries that came upon a market- and trade-orientation
in economic policy sooner than most.13 If both the nature of polity and type of economic strategy
vary, we must find a way to account for which aspect -- political or economic -- led to what, and
why.
Third, whether democracy made a difference to poverty -- and to what extent -- is not a
cross-sectional question. If both country A and B, one democratic, the other authoritarian, had
about a tenth of their population below the poverty line at time T, we would still not be able to
11
For the most recent listing, see Freedom House 1998.
12
The analytic implication of such a small-n world, one might add, is very different from the one we encounter
when we examine the impact of democracy on economic growth globally. Inclusion of both developed and developing
countries makes the number of democracies sufficiently large, making the idea of a sophisticated statistical analysis
viable (Barro, 1997)
11
say that they were equally successful in attacking poverty, unless at time T-1 the proportions of
poor A and B were also the same. We need least two sufficiently distanced time-points for
analysis, if not an entire time-series. Such data on an inter-country basis do not exist. Based on
an international poverty line of $1/day in PPP terms, global figures for poverty were first
calculated for 1985 (The World Bank, 1990). On that basis, we can say that between 1987-93,
about 30 per cent of the world population remained more or less consistently below the $1/day
poverty line,14 and we also know something about the distribution of poverty across countries.
But we do not know the numbers of the poor, either globally or countrywise, in the
1960s. If, to put such numbers together, we go by the reports available on each country, we find
that the criteria used by different countries to define and measure poverty either do not match,
and often the criteria have not been consistently used within the same country. For some
individual countries such as India, however, such statistics do go back much further, and they
have also been consistently used to track progress on poverty reduction, but such is not the case
with most developing countries.15 A methodologically defensible time-series on poverty for the
entire developing world is neither available, nor can it be easily created for the pre-1985 period.
Luckily, even if we don't know as much as we ideally should, a few robust judgments are
possible on the basis of whatever we do know. Some conclusions do not require, or depend on,
but a certain consistency marks the way democracies have dealt with poverty. The authoritarian
countries have either exhibited spectacular success at attacking poverty (South Korea, Taiwan
13
Whether some of the colossal failures at poverty removal also came from authoritarian countries with trade-
oriented economic policies is unclear. At any rate, if there are such cases at all, they have not been studied.
14
Until more recent and detailed figures are released later this fall, the available global data cover the period up to
1993.
15
Indeed, the $1/day poverty line used by the World Bank is based on the work done in India in the 1960s,
when India began measuring the scale of the problem systematically.
12
and Singapore); or they have failed miserably (in much of Sub-Saharan Africa and Latin
America). In comparison, those developing countries that have gone through long democratic
tenures may not be colossal failures, but they can not boast of extraordinary successes either. On
poverty, the worst democratic performers have been India (35% of the population below the
poverty line in 1991), Botswana (33% in 1985/6), the Philippines (28.6 per cent in 1991), Costa
Rica (19% in 1989) and Venezuela (12% in 1991).16 The best performers have been Malaysia,
Sri Lanka and Jamaica, each having less than 6 per cent of its population below the poverty line
between 1989-93. But note that even the best do not compare with South Korea, Taiwan and
Singapore, where the percentage of population below the poverty line is zero. In the early to mid-
1960s, these countries were as desperately poor as the other Asian and African countries and
considerably poorer that Latin American nations (Morris and Adelman, 1973), but by now they
have become developed economies, with higher per capita incomes than some European
countries.
We should also note that no democracy has let absolute poverty worsen over a substantial
length of time, or dramatically; and no democracy, of course, has allowed famines to take place
(Sen, 1989). Authoritarian polities have seen both of the latter, and have for long periods of time
also managed to get away with it. Mobutu's Zaire, for example, economically declined for over
two decades, and Argentina's per capita income virtually did not grow between the late 1960s and
late 1980s, a period of more or less uninterrupted authoritarian rule. Democratic countries in the
In short, the wild authoritarian fluctuations contrast sharply with a certain middling
democratic consistency. Democracies may not be necessarily pro-poor, but authoritarian systems
can be viciously anti-poor. Democratic attacks on poverty have simply been slow but steady.
16
All percentages are based on the World Development Indicators published by the World Bank and shared by the
OECD countries.
13
Why Not Better?
Why is the record of democracies not any better? There are three theoretical
possibilities: 1) the poor either do not vote, or local coercion makes it difficult for them to vote
according to their true interests; 2) organizing the poor is difficult because of collective action
and an identification of whether those who abstain from voting are primarily the poor. While
turn-out rates can be put together for most democracies, their disaggregation in terms of class is
not available.
On India, where disaggregated data on turn-out rates are now available (Yadav and
Singh, 1997), the picture is exactly the opposite of what is implied above. The poor in India have
tended to vote more than the middle and rich classes, the villages more than the cities, the lower
castes more than the upper castes, at least in recent years. The incidence of coerced voting --
coercion exercised by the local patrons or local rich -- has also gone down significantly. If India
is any guide, our conclusion should be: even when the poor vote, poverty-alleviation can be slow,
or politicians do not necessarily make removal of poverty their prime goal. No Indian election --
of the twelve held for parliament since independence -- has turned on the performance of the
government on poverty-alleviation.17
Possibility 2) points to the political mobilization of the poor putting pressure on governments to
allocate resources to poverty-alleviation and also potentially changing the trajectory of politics,
reflected either in party competition for the vote of the poor, or in empowering a party that stands
17
However, one -- in 1971 -- was heavily influenced by the election slogan and promise of "garibi hatao"
("abolish poverty") by Mrs. Gandhi and her Congress party. Not much came of the promise. Moreover, when Mrs.
Gandhi was thrown out of power in 1977, the big issue was not poverty, but her suspension of democracy for 18
months (1975-77) and a draconian family planning program that her son initiated.
14
Possibility 3) also goes in the same direction, but in a different way. Mobilization may
run into difficulties of collective action, but voting constitutes individual, not collective, action.
One can express one's disapproval of a government by voting in a democracy. If all poor people,
as individuals, vote on grounds of whether or not the government is alleviating poverty, they can
Two reasons, I would like to suggest, ensure that voting and mobilization possibilities for
the poor do not lead to a removal of poverty, and do not become spurs for effective and sustained
First, for the poor, poverty-alleviation measures that are direct and short-run carry a great
deal more weight than measures that are indirect and have a long-run impact. Given the
subsistence, or the near-subsistence, levels of their existence, such reaction is rational and to be
expected.18 And second, the poor, like all of us, have multiple selves. They are not only
members of a class of poor, but also of linguistic, religious, tribal and caste communities. Often,
their voting, like those of many others, is identity-based, not class-based -- and so is their
mobilization. Multiple selves drive a wedge between the poor as a class and the poor as a
political collectivity, significantly reducing, if not eliminating, pressure on the government to act
Whether their impact on poverty is lasting or not, direct methods -- subsidies and hand-
outs and at a more radical level, land reforms -- have clearly demonstrable and short-run linkages
with the welfare of the poor. The relationship of indirect methods -- exchange rate devaluations,
18
This is also an argument shared by the otherwise divergent perspectives of moral and rational economy on poor
peasants (Popkin,1979; Scott, 1976).
15
strategy -- with poverty-removal is not so clear-cut, short-run, and intuitively obvious. This has
conditions would their link with mass welfare be clear and direct? Were a country's economy
heavily dependent on foreign trade, a lowering of tariff walls, a reduction in quantitative trade
restrictions and a devaluation of the currency would potentially be of great concern to the masses.
In 1996, trade constituted more than fifty per cent of the GDP of Singapore, Malaysia, Thailand,
the Philippines, Mexico, Hungary, South Korea, Poland and Venezuela, and between forty to fifty
percent of the GDP of Israel, Chile, China and Indonesia. Changes, especially dramatic changes,
in the trade and exchange rate regimes of these countries thus have a clear potential for mass
politics.
It would, however, be instructive to look at what the trade/GDP ratios for these
economies were in the 1970s. Without doubt, they were much smaller, certainly in Hungary,
Poland, China, Venezuela, if not others. The political implication should be obvious. If external
trade is a small part of the economy, changes in trade and exchange rate regimes will be of
peripheral importance to the masses.19 In order for trade and exchange rate regimes to become
part of mass politics, it may be necessary to make economies more trade-dependent in the future.
If globalization does proceed further, this scenario is quite possible. But it is important to recall
that in the protectionist era that lasted right until the late 1970s and early 1980s, such political
One can, of course, argue that even if trade-dependence of an economy is small, several
long-run or indirect linkages can be shown to exist between mass welfare on the one hand and
19
The overall size of the economy complicates the meaning of low trade/GDP ratios. Smaller economies tend
generally to have a high trade/GDP ratio, making trade very important to their political economies. With the striking
exception of China, however, the largest economies of the world -- the U.S., Japan, Germany -- are less trade
dependent. (Indeed, the trade/GDP ratio for India and the U.S. was roughly the same in 1996.) Still, trade politics, as
we know, has aroused a great deal of passion in the U.S. and Japan. The meaning of the same ratios can change if the
leading sectors (autos, computers) or "culturally significant" sectors (rice for Japan, agriculture in France) of the
economy are heavily affected by trade.
16
overvalued exchange rates or relatively closed trade regimes on the other. Anne Krueger (1991),
for example, has argued that by making "import competing" industrial goods dearer for the
countryside and also discouraging exports, an ISI-type trade and exchange rate regimes
systematically discriminated against the countryside all over the developing world. The
implication is that a majority, or large plurality, of a developing country's population was hurt,
At this point in our reasoning, the politics and economics of development significantly
diverge. Agrarian politicians have long existed in most countries, but they have rarely, if ever,
agitated for an open foreign trade regime, focusing instead on land reforms or on the unfavorable
urban-rural trade, which may indeed have, as Krueger argues, caused less overall damage. If
such indirect links were not even clear to the economists, who continued until the 1970s to look at
rural welfare primarily in term of internal terms of trade and direct benefits to the countryside,
can a politician be expected to mobilize peasants over the underlying and subtle, though hugely
important, links between foreign trade and mass welfare in a poor country? Long-run and
indirect links do not work well in democratic and mass politics: the effect has to be simple,
More direct evidence on how the masses look at market- and trade-oriented economic
reforms is also available. In the largest ever survey of mass political attitudes in India conducted
between April-July 1996, a full five years after reforms were initiated, 32 per cent of the urban
voters knew of reforms, but only 12 per cent of the rural electorate had heard of them (Yadav and
Singh, 1996), even though a change in trade-regime implied that the protection offered to
manufacturing relative to agriculture had gone down significantly and agriculture's terms of trade
had improved. Further, nearly 66 per cent of the graduates were aware of the dramatic changes in
economic policy, compared to only 7 per cent of the poor, who are illiterate and mostly residents
of the countryside. Thus, even as late as 1996, India economic reforms, toasted enthusiastically
in the domestic and international economic community, had barely registered as an important
17
event for the rural folk and the poor. An equally big initiative launched as a direct attack on
poverty, however economically unsound, would almost certainly have registered prominently.
Democratic preference for direct methods has on the whole -- and so far -- limited the
ability of democracies to eliminate poverty. It also signifies the outer limits of democratic
political action. For even if we could mobilize the poor as a cohesive force, their voting became
entirely economic, and their united pressure led to something as dramatic as land reforms, the
consequent effect on poverty, though positive, would not be as substantial, or as long-lasting and
Consider land reforms in some detail. They are commonly seen as the most effective
direct method of attacking poverty.20 For land reforms to remove poverty, two other conditions
must be satisfied: a) there should be enough land to go around, ensuring that a wholly suboptimal
distribution of small plots does not constrain the post-reform productivity of the poor; and b) new
agricultural technology, which is relatively capital-intensive and therefore not easily affordable
by the poor, is made available at a subsidized rate by government (Varshney, 1995). A better
utilization of labor inputs, possible and likely after the tenant is more secure or people work on
their own lands, will by itself not bring about sustained increases in productivity and therefore
income (Herring, 1989). If plots end up being suboptimal or new technology -- irrigation, HYV
seeds and chemical fertilizers -- can not be bought by the beneficiaries, land reforms may bring in
equality and introduce a measure of justice, but they can also leave most beneficiaries at a low-
level economic equilibrium, instead of facilitating a sustained increase in their incomes. Security
to tenants and allocation of land to the tiller do not necessarily take them above the threshold of
poverty line. The questions of productivity and access to new technology continue to be relevant.
This, I should emphasize, is not an argument against land reform or asset redistribution
per se. It is simply a call for pushing the inquiry towards asking how it compares with other
20
The argument below is economic, not political. For the political conditions under which land reforms can be
implemented and why these conditions are so rare, see Herring (1983).
18
methods and whether other methods are more productive and sustainable, while also being
poverty-alleviating simultaneously. Land reforms are defensible, if one can show that they
promote productivity and can be a source of sustained increases in income. In Latin America,
where the Gini coefficients of land distribution are considerably higher than elsewhere, the
land/man ratio is not entirely unfavorable, and there is enough land to go around, a case can still
be made for land reforms. By contrast, where landlessness is not typically high (for example,
Sub-Saharan Africa), the poor will not gain much from land redistribution. Other methods of
continent not known for high irrigation/total acreage ratios, or scientific research on seeds that are
If new titles to land, or security of tenancy, cannot guarantee an end to poverty, the
effectiveness of income transfers -- though credit, or producer subsidies -- to the poor is even
more open to question. Such direct transfers to the poor, when a large plurality of the population
is below the poverty line, are no panacea for the problem of poverty.
To conclude, the direct methods are less economically effective but politically more
appealing in the short-run, and the indirect methods are more economically effective but harder to
sell politically. A better alignment of the political and economic may be possible in authoritarian
countries, where politicians do not have to carry the masses with them and a preoccupation with
the long-run and indirect methods of poverty removal can simply be forced down the process, if a
political elite is committed to the poor. Unless the trade-dependence of an economy qualitatively
changes, making more and more people dependent on trade for their welfare, a trade-oriented
strategy will not ignite political passions in a democracy, and the economically desirable methods
19
The argument above underlines why direct methods of poverty-alleviation, even though
politically attractive in poor democracies, are not well-equipped to end mass poverty. The
argument does not imply that direct methods will have no impact. To repeat, both methods can
Within the parameters of direct action, however, the best results are obtained when class
and ethnicity coincide for the poor, not when class and ethnicity clash. The former are called
ranked ethnic systems in the literature, and the latter unranked ethnic systems (Horowitz, 1985).
If ranked ethnic systems are also democratic, the poor can exert more effective pressure on
governments and the effect on poverty is greater than is normally possible in unranked ethnic
systems. Why should this be so? And what kind of evidence do we have to support the claim?
In generating collective action, the greater power of ethnicity vis-à-vis class can be
explained in two ways. One is to rely on theories of collective action, and investigate how the
Recall briefly why, according to rational choice theories, collective action is hard. The
key argument is that individual and class interests can often clash. (Elster, 1985; Przeworski,
1986). A landless peasant may be hurt by the decision of the landlord to keep wages stagnant, but
it does not follow that it is in his interests to join a union or party mobilizing agricultural workers.
From an individual perspective, class mobilization can have a serious cost: the peasant may be
dismissed altogether by the lord, and a dismissal may incalculably increase his misery which the
resources of an agricultural union are typically unable to alleviate. Moreover, and this is the
second part of the argument, if the mobilizing union did succeed in getting wages raised for the
landless class, the peasant would benefit anyway whether or not he participated in the agitation.
The landless worker has an interest in the benefits of class action, not in its costs, and class
benefits, if available, can't easily be denied to the landless peasant, but costs of class mobilization
must be individually borne. Behavior according to collective interest can take place, but it
requires setting up a incentive structure which counters the temptation to "free ride".
20
Developments in collective action theory seek to show why ethnicity solves the collective
action problem better than class does. Class action is bedeviled by free-riding (or, what would be
analogous, by problems encountered in a prisoner's dilemma), but the main strategic problem in
ethnic collective action is one of coordination, not free-riding (Hardin, 1995). Coordination
games are different from the prisoner's dilemma game. They rely on "focal points" to facilitate
convergence of individual expectations;21 hence they show how collective mobilization becomes
possible. Ethnicity can serve as a focal point; class can not, at least easily.
The idea of "focal points" comes from Schelling's seminal treatment of the coordination
"When a man loses his wife in a department store without any prior understanding on where to meet if they
get separated, the chances are good that they will find each other. It is likely that each will think of some
obvious place to meet, so obvious that each will be sure that the other is sure that it is obvious to both of
them." (Schelling, 1963, p. 54)
Schelling goes on to suggest that without having an intrinsic value for the couple, the
"lost and found" section of the department store could be one such place. It will, however, not be
a focal point if there are too many "lost and found" sections in the store. A focal point is
the "formation of mutually consistent expectations" (Schelling, p. 84). Ethnicity can be viewed
as one such focal point for mobilization. There is no equivalent in class action.
The second line of reasoning, not deployed in the political economy literature as the
reasoning above typically is, has emerged from the theories of ethnicity and nation-building.
Though the primary purpose of these theories is to show that ethnic and national groups are
imagined communities (Anderson, 1983), they have serious implications for the concept of class.
21
Coordination games take the following form. So long as others in the group are cooperating, it is rational for
me to cooperate -- for if all cooperate, the likelihood of the group gaining power (or the group realizing group
objectives) goes up tremendously. "(P)ower based in coordination is superadditive, it adds up to more than the sum of
individual contributions to it." (Hardin, 1995, p. 37). Third, all one needs to keep the coordination game going is a
"charismatic leader", a "focus", ( p. 36) and a mechanism through which information about others cooperating is
provided to me. "Coordination power is..a function of reinforcing expectations about the behavior of others." (p. 37)
21
imagined community for it does not allow face-to-face intimacy. Thus, classes, like ethnic
groups and nations, are also imagined communities. An individual does not "naturally" feel his
class; such consciousness depends on political mobilization, public policy, or other people's
Why, then, is caste/religion/nation more powerful than class? Because compared to class,
caste, ethnicity and religion are more likely to form a historically enduring bond and provide
common histories, heroes and villains. Moreover, the poor as a class rarely have leaders from
among the poor. In contrast, a poor ethnic community can witness the emergence of a small
The consequences for political action naturally follow. Given shared history, common
bonds, widely comprehended symbols and its own leaders, ethnicity is thus more easily
mobilizable than an economically constructed class, whose members may belong to a variety of
ethnic groups and which lacks such a repertoire facilitating convergence of individual
expectations. It also follows that if the poor as a class also form an ethnic group, they can act as a
Clearly, such a distinction between ethnicity and class may not be present everywhere. It
will certainly not mark the politics of monoethnic societies (such as South Korea, Taiwan and
many, though not all, Latin American countries); or societies where the subaltern ethnic group is
not only poor but small in size and yet to develop a middle class22. For all of these reasons, in the
literature on ethnicity, Latin America (Dominguez, 1996) and East Asia (Horowitz, 1985) are
normally considered outliers. They have seen a lot of class politics, but not enough ethnic
politics, at least not yet. In comparison, in South and Southeast Asia, Sub-Saharan Africa,
22
Or, sociologists have often reminded us, societies where the "hegemony" of the privileged groups is yet to be
broken.
22
Ranked Ethnic Systems and Poverty: Examples
Let me now turn from theoretical reasoning to the empirical world. What examples can
be cited for the claim that unless poverty is linked to identity politics, it does not necessarily
While we know a great deal about the ethnic profiles of most poor democracies, inter-
country comparisons on poverty, as already stated, are rendered difficult by the absence of a time-
series and lack of consistency in measurement criteria. Still, as we have already noted, of all poor
democracies, Sri Lanka and Malaysia have shown by far the best results on poverty-reduction. In
both cases, when democracy was instituted, the proportion of the population below the poverty
line is by now less than ten percent23, and in both, the majority ethnic group -- the Sinhalese and
the Malays -- was vastly more rural and poor than the major minority group -- the Tamils in Sri
Lanka24 and the Chinese in Malaysia. Moreover, once inaugurated, democratic politics got
ethnically structured, in the Malaysia case constitutionally (Milne, 1967). In each country, the
majority ethnic group, led by its small upper and middle class, came to power and once in power,
the elite undertook a large number of direct measures, both in the countryside and cities, to
intra-country comparisons,26 where a great many factors other than ethnicity can be controlled for
and the effect of ethnicity on poverty identified with greater certitude. In India, as reported
earlier, detailed and disaggregated statistics on statewise poverty, going back to the 1960s, are
23
Based on world development indicators of the World Bank.
24
With the exception of what have come to be known as Estate Tamils, who did not identify for a long time with
the politics of Jaffna Tamils.
25
For Sri Lanka, see Moore (1985); for Malaysia, Klitgaard and Katz (1983).
26
For Sri Lanka, it has been argued that compared to other countries, it had fewer inequalities right at the time of
independence. Its good performance thus is not simply a function of the policies pursued after independence. The
performance was path-dependent. See Bhalla and Glewwe (1986)
23
available. Patterns of state politics and policy can thus be clearly linked to the outcomes for
poverty.
The states of Punjab and Kerala have shown the best results.27 In Punjab, the green
revolution, an indirect and growth-based method, has been key to poverty alleviation. In Kerala,
the method was direct. Land reforms and extensive job reservations in government employment
Was the emphasis on direct methods in Kerala a result of the poor organizing themselves
as a class? On the face of it, this would appear to be the case, primarily because a Communist
party, repeatedly elected to power after 1957, led the campaign for land reforms and social
However, both social history as well as electoral data make it clear that there was a
remarkable merging of caste and class in Kerala, the former defined ethnically, the latter
economically. At the center of this coincidence is the Ezhava caste, estimated to constitute a little
and were therefore considered "polluting" by the upper castes. They were not only considered
"untouchable" but also "unseeable". The catalogue of everyday humiliations for the Ezhavas was
painfully long :
"They were not allowed to walk on public roads. .. They were Hindus, but they could not enter temples.
While their pigs and cattle could frequent the premises of the temple, they were not allowed to go even there.
Ezhavas could not use public wells or public places. .. .
..An Ezhava should keep himself, at least thirty six feet away from a Namboodiri (Brahmin) …...... . He must
address a caste Hindu man, as Thampuran (My Lord) and woman as Thampurati (My Lady). ….. He should
never dress himself up like a caste Hindu; never construct a house on the upper caste model. .. T)he women
folk of the community...were required, young and old, to appear before caste Hindus, always topless. About
the ornaments also, there were restrictions." (Rajendran, 1974, pp. 23-4)
At the turn of the century, experiencing some mobility and developing a small middle
class, the Ezhavas rebelled against the indignities of Hindu social order and started fighting for
their civil rights. Led by a famous Ezhava saint, Sri Narain Guru, sometimes called the Gandhi of
24
Kerala, their protest movement aimed at self-respect and education. Self-respect entailed
withdrawal from toddy tapping, a movement into modern trades and professions, and a
nonviolent attack on the symbolic order. Since they were denied entry to temples and were only
allowed to worship "lower gods and spirits", the Ezhavas, the Guru said, would have their own
temples, in which they would worship "higher gods" to whom they would offer flowers and
sweets, not animals and liquor reserved for the "lower gods". Meanwhile, to improve their
economic and social status, they would educate themselves. And to facilitate all of these
These issues, all caste-based, decisively restructured the politics of Kerala in the 1930s.
Entry into temples, an attack on the social deference system concerning dress and access to public
roads and a more equal access to education drove the civil rights campaign. It is only
subsequently that tenancy rights and land reforms spurred the mobilization for economic rights.
If the fit between the Ezhava caste and the rural poor had not been so good between the 1930s
and 1940s, class mobilization would have made little headway. Class politics was inserted into
the campaign for caste-based social justice.30 To this day, the Ezhava caste continues to be the
principal base of the CPM. People of similar class-positions, if Nair, have gone on the whole
with the Congress; if Christian, with Kerala Congress; if Muslim, with the Muslim League
27
For a quick overview of all states, see Ravallion and Dutt (1996).
28
For a detailed analysis, see Rao (1979).
29
In a disarmingly candid statement, E.M.S. Namboodiripad, the greatest Communist mobilizer of 20th
century India and a Kerala-based politician admitted before this death that the inability of the decades-long class
mobilization in Kerala to overwhelm the religious divisions of the state might be rather more rooted in historical
realities than Marxists had expected. See Namboodiripad (1994)
30
For a compelling argument that this merger facilitated the emergence of a Communist movement, Menon,
1994. While talking about the peasants and workers, the Communists could repeatedly use caste issues, which had
great resonance in Kerala.
25
Concluding Observations
Instead of summarizing the argument in this concluding section, let me push the line of
alleviation, reforms have been implicitly present in the paper, not explicitly analyzed. The
context of poverty-alleviation has remarkably changed in the last decade, as more and more
countries have adopted for a market-based and trade-oriented economic strategy (Sachs and
Warner, 1995). In what ways are the political arguments about poverty above, summarizing on
the whole the post-1945 development experience, relevant to the era of economic reforms, if at
all?
well. Politically speaking, market-oriented economic reforms in the developing world, though
highly economically desirable, have thus far not emerged from below, which is where the masses,
and therefore the poor, are located. Reforms have emerged from above -- through economic
bureaucracies, both national and international such as the World Bank -- in moments of economic
crisis that have demonstrated the exhaustion of statist and inward-looking economic strategies.
The masses may have felt an acute dissatisfaction against the excessive and abusive powers of the
state, especially in the former Communist world, but a desire for lesser state interference in
everyday life does not necessarily translate into a support for the market mechanism. Reliance on
markets has simply emerged from the ruins of a discredited statist ideology of economic
development, whose capacity to deliver mass welfare had seriously declined and which had
interpret the turn towards economic reforms in the developing world as a sign of vigorous mass
support for them. As of now, reforms have simply been a welcome diversion from the excesses
of bureaucrats and state functionaries. This diversion may cease to be an indirect source of
support for reforms, if markets fail to deliver mass welfare. In short, as of now, democratic
26
politics and market-oriented economic reforms are quite awkwardly aligned. It is not a happy and
permanent marriage.
as well as suggests a possible danger. The challenge is simply that reformers must produce a
political language for their arguments that can be understood by "mass politicians", so that a mass
constituency for reforms is created, and think of an incentive structure that can make reforms
politically attractive for politicians (Sachs, Varshney and Bajpai, 1999). The danger is also
equally simple. In the absence of an incentive structure or a clearer understanding of, if not
ideological convictions about, what economic reforms can do for mass welfare in the long-run,
any significant short-to medium-run downturn that makes the masses precipitously worse off can
take the shine off reforms in democratic politics, and create a constituency for disciplining the
The onus is on those who believe that economic reforms are not only about economic
efficiency, but also about enhancing mass welfare, especially that of the poor. If reforms must be
pursued in a democratic framework, an imaginative integration of the political and the economic
27
REFERENCES
Irma Adelman and Cynthia Morris, 1973, Economic Growth and Social Equity in Developing
Countries, Stanford University Press.
Jagdish Bhagwati, 1988, "Poverty and Public Policy", World Development, Vol. 16, No. 5.
Surjit S. Bhalla and Paul Glewwe, 1986, "Growth and Equity in Developing Countries: A
Reinterpretation of the Sri Lankan Experience", World Bank Economic Review, September.
Robert Dahl 1989, Democracy and its Critics, Yale University Press.
Jorge Dominguez, 1994, ed, Race and Ethnicity in Latin America, Garland Press.
Jon Elster, 1986, "Three Objections to Class", in John Roemer, ed, Analytical Marxism,
Cambridge University Press.
Russell Hardin, 1995, One for All: The Logic of Group Conflict, Princeton University Press.
Ronald J Herring, 1989, "Dilemmas of Agrarian Communism", Third World Quarterly, January.
Samuel P Huntington, 1991, The Third Wave: Democratization in the Late Twentieth Century,
University of Oklahoma Press.
Samuel P Huntington, 1983, "Will More Countries Become Democratic?", Political Science
Quarterly, Fall.
Ayesha Jalal, 1995, Democracy and Authoritarianism in South Asia, Cambridge University Press.
Gary King, Robert Keohane and Sidney Verba, 1994, Designing Social Inquiry, Princeton
University Press.
Atul Kohli, 1987, The State and Poverty in India, Cambridge University Press.
Anne Krueger, 1990, The Political Economy of Agricultural Pricing Policy, Vol. 5, The Johns Hopkins
University for the World Bank.
28
Arend Lijphart, 1977, Democracy and Plural Societies: A Comparative Exploration, Yale University Press.
Dilip Menon, 1994, Caste, Nationalism and Communism: Malabar, 1900-1948, Cambridge University
Press.
Mick Moore, 1985, The State and Peasant Politics in Sri Lanka, Cambridge University Press.
E.M.S. Namboodiripad, 1994, "Presidential address", International Congress of Kerala Studies, 27-29
August 1994, Addresses and Abstracts, Vol. 1, AKG Centre for Research and Studies.
Adam Przeworski, 1985, Capitalism and Social Democracy, Cambridge University Press.
G. Rajendran, 1974, The Ezhava Community and Kerala Politics, Academy of Political Science,
Trivandrum, 1974.
M.S.A. Rao, 1979, Social Movements and Social Transformation, MacMillan Company of India
Limited.
Martin Ravallion and Gaurav Dutt, 1996, "India's Checkered History on Poverty Alleviation",
Economic and Political Weekly, September.
Jeffrey Sachs, Ashutosh Varshney and Nirupam Bajpai, 1999, "Introduction", in Sachs, Varshney
and Bajpai, eds, India in the Era of Economic Reforms, Oxford University Press.
Jeffrey Sachs and Andrew Warner, 1995, "Economic Reform and the Process of Global
Integration", Brookings Papers on Economic Activity, No. 1.
James Scott, 1976, The Moral Economy of the Peasant, Yale University Press.
Thomas Schelling, 1963, The Strategy of Conflict, Oxford University Press, 1963.
Ashutosh Varshney, 1998, "Mass Politics of Elite Politics? India's Economic Reform in
Comparative Perspective", Journal of Policy Reform, December.
Ashutosh Varshney, 1995, Democracy, Development and the Countryside, Cambridge University
Press.
Myron Weiner, 1991, The Child and the State, Princeton University Press.
Myron Weiner, 1989, "Institution Building in South Asia", in The Indian Paradox, Sage
Publications, 1989.
The World Bank, 1997, India: Achievements and Challenges in Reducing Poverty, Report no.
16483-IN, Washington D.C., May 27.
The World Bank, 1990, World Development Report 1990, Oxford University Press.
29
Yogendra Yadav and V.B. Singh, 1997, "The Maturing of a Democracy", India Today, August
31.
30