POVERTY AND INEQUALITY: A LIFE CHANCES PERSPECTIVE

Norman Uphoff
Cornell University

One of Indias most significant accomplishments over the last five decades has been to change
the worlds image of its poverty. At independence, India was widely regarded as a poor
country" with almost all of its people living in poverty, albeit overlaid by a very thin but rich
upper crust of maharajahs and commercial moguls. The Nizam of Hyderabad and other
anachronistic remnants of the pre-colonial era particularly intrigued many people in other
countries, who considered such huge personal accumulations of gold and jewels as real wealth.
These extravagant riches and their accompanying regal lifestyles made even less tolerable the
contrasting miserable, constrained conditions of most Indians.
It represents progress for India to be known and respected now for its large middle class,
which is practically as large as the population of the United States. However, the number of
Indians who subsist below the poverty line, nearly 400 million in 1992 (World Bank, 2000), is
greater than India's total population was at independence. It is also progress that we understand
better that the wealth of nations, to use Adam Smith's term, lies not in precious metals and stones
but in the capacities of its people, to produce, to invent, and to organize.
It is a matter of debate -- and more a matter of values than of arithmetic --whether
inequality is greater in a country when the vast majority are poor and a few people are very
wealthy, with a small middle class -- or when there is a considerably larger middle class which
has distanced itself from the poor, with the number of rich persons, few of them opulently
wealthy, increased by several multiples. The latter situation has a lower Gini coefficient, but it

makes more visible to the poor on a continuous and very obvious basis how different are their
lives and their life chances from those of the rest of their society.

Some Thoughts on Indicators of Inequality and Its Justification

It is worthwhile to know what are the Gini coefficients and other such measures of inequality
discussed in the following chapter by Hitzhusen, so that we can assess differences between
countries, and over time, in specific and reasonably comparable terms. Having myself
undertaken to compare differences in income inequality among 16 Asian countries some years
ago as part of an evaluation of their rural development strategies and accomplishments (Uphoff
and Esman, 1974), I have come to favor the ratio of income that is accruing to the highest and
lowest quintiles of the population. Wealth, the stock of income-producing assets, is probably a
better indicator than is income, the flow therefrom, but measurement problems are daunting
enough to make such figures even less reliable than are income data.
This measure of inequality considers the income of the top 20 percent as a multiple of
that which goes to the bottom 20 percent. This is a comprehensible and meaningful indicator of
equality or, conversely, inequality. It is concrete, not abstract, and moreover, it highlights the
most extreme differentials, reflecting the non-linearity that goes with living very far below or
very far above the poverty line.
Our analysis as Asian experience, which was supported by USAID in the days when it
was more concerned with the substance and strategy of development than now, was undertaken
to make some objective comparisons among countries that ranged geographically, politically and
economically from China and Japan in East Asia to Turkey and Yugoslavia on the western edge
of Asia. The analysis was done in the positivist spirit of the time, emphasizing measurement.

While income distribution was not a central focus of our research project, it was something that
had to be considered. At that time, there was still a strong argument in the literature justifying
inequality as good for promoting economic growth.
We found, on the contrary, that those countries with the best records in rural development
across a wide range of measures (agricultural, nutritional, educational, public health, etc.) also
had definitely more equal distributions of income. Unfortunately, however, the relationships
between different development measures and income distribution were too many and too
complex for us to attribute which was cause and which effect. Since our research was not
intended to illuminate inequality, we reported our findings and left them for readers to consider
(Uphoff and Esman, 1983: 292-294).1
What could be concluded from our data, however, was that having a more equal income
distribution was not unfavorable for rural development. Analyses based on a simplistic
understanding of the Harrod-Domar model of economic growth had argued that unequal
distribution of income should promote economic growth and greater employment because -- it
seemed logical -- (a) richer people would save more than poor people; (b) a greater volume of
domestic savings will increase the supply of resources available for investment; and (c)
accelerated capital formation will raise gross domestic product and resulting incomes, a virtuous
cycle feeding back into greater savings. Thus, income inequality, even that reflecting widespread

It might have been expected that per capita income levels would explain both higher performance on measures
such as rate of agricultural production growth, adult literacy, life expectancy, and caloric intakes per capita, as well
as greater income equality, all measured in the early 1970s. In fact, the eight country cases classified as having more
kinds, and greater degrees, of local organization involved in rural development functions, the independent variable
in this analysis, had per capita incomes almost three times greater in 1973 than the average for the "less organized"
cases, $352 vs. $119. Just 20 years earlier, per capita income levels for the two sets of countries had been
practically the same, $74 and $78, respectively. There were good grounds for inferring that the functioning and
performance of local governments, cooperatives, farmer associations, etc. had contributed something to the general
advancement in agricultural, economic and human resource terms.

poverty, was regarded as good for development. It would contribute to more savings, investment,
and growth of GNP.
Already by the mid-1960s, however, at the height of development thinking which equated
development with GNP growth, a few economists were already pointing out that this logical
construction was not empirically supported by evidence.2 Even though these empirical
challenges appeared in leading journals at the time, they were ignored. Why? They went against
the prevailing paradigm, which seemed so logical.3 Moreover, they went against predominant
economic interests. If development was regarded as depending almost entirely upon capital as
the scarcest, and thus as the most valuable factor of production, this justified the owners of
capital receiving the largest share of the benefits from development.
About this time (1967), President Julius Nyerere of Tanzania presented in "The Arusha
Declaration" a conceptual, not just empirical, challenge to the prevalent view. If poor countries
have little capital and an abundance of labor, he asked, why not use whatever capital is available
to make the most abundant resource, labor, more productive -- rather than use labor, often
wastefully and certainly with poor remuneration, to make the resource they had least of, capital,
particularly foreigners' capital, more productive? Why should the poor seek to fight their war

In an extensive review of the economic literature which documented the dominance of these savings and
investment theories, Hahn and Matthews (1964) noted that these theories failed to account adequately for the
phenomena under consideration. When Hamberg (1969) reviewed a number of cross-country studies that used UN,
OECD and other data sets, he found that the correlations between gross domestic investment and gross domestic
income growth seldom reached even .33, which would not account for more than 10 percent of variation, and were
anyway not statistically significant. Various other studies, summarized by Owens and Shaw (1972), showed that
persons with higher incomes did not necessarily save higher proportions of their income, and greater savings did not
necessarily lead to more investment. In addition, even though investment was correlated with national income
growth, the correlation was not a strong one, as returns to investment varied widely across countries. In my own
research on Ghana's economic development in the 1950s and 1960s, I found that gross domestic capital formation
reached 21 percent in 1965, and yet Ghana's economic growth the next year was negative. Per capita growth had
already been declining since 1961 despite GCDF rates over 15 percent. Even a modest incremental-capital-to-output
ratio (ICOR) of 5:1 should have given growth of 3 percent if capital is indeed a determinant of economic growth.
3
On the power of paradigms to restrict valid new ideas, see Krugman (1995) on how and why spatial relations have
been largely excluded from development economics analysis.

against poverty with the weapons of the rich? Nyerere asked pointedly. This was dismissed as
ideology rather a legitimate question.
There were some stirrings within the economics discipline during the 1960s that
questioned the dominant capital-favoring paradigm.4 But it took another 20 years before the case
for more equitable paths to development gained acceptance, though still not dominance. The
proponents of "meeting basic human needs" in the 1970s justified this more on grounds of equity
and fairness than as a way to raise productivity.
What finally seems to have gained the most ground for equitable distribution of income
and wealth was the success of the East Asian "Tigers" -- Japan, South Korea, Taiwan and
Singapore. By the 1980s, this success was too apparent to overlook, as was their more equal
distribution of income.5 In these countries, policies ranging from land reform and universal basic
education to public housing and primary health care had contributed to political-economic
systems that sought to contradict the Biblical admonition: "The poor you shall have always with
you." These countries considered poverty to be unacceptable and a drag on their economies.
There was also evidence accumulating such as that from Berry and Cline (1979) that more equal
distributions of land contributed to aggregate agricultural production as well as economic

Singer (1966) suggested that the problem of development was not the creation of wealth but the creation of the
capacity to produce wealth, and this favored investment in human resources. Kuznets (1966) showed that incomes
were almost always more equal in the more developed countries, though this did not resolve the question of what
was cause and what effect. A cross-national analysis by Leibenstein (1965) of sources of productivity found that
efficient allocation of the conventional factors of production could not account for more than 30 percent of variation
in productivity, attributing the rest to human, cultural, organizational and other sources. Lewis (1965) concluded that
efforts to accelerate the pace of development by devoting (or diverting) a larger share of increased output into
savings and capital investment, rather than to consumption, was self-defeating. This was partly because it could give
rise to political unrest, but also because output cannot be sustainably increased unless consumption also increases.
Capital formation in Lewis' view was more of an intervening or even resulting variable than an independent one. On
the fallacious dichotomy between consumption and investment expenditure, see also Morgan (1969).
5
Current data from the World Development Report 2000/2001, which unfortunately does not include data on income
distribution for Singapore and Taiwan, show income ratios (top 20 percent:bottom 20 percent) to be 3.4 in Japan and
5.2 in South Korea. For comparison, the ratios in China and U.S. are 7.9 and 8.9, respectively. In the four countries
of Scandinavia, on the other hand, the ratio is 3.6, less than half these latter figures. India's ratio is 5.7, as discussed
further below.

growth. Certainly East Asian land reforms, including that in China, were important impetuses for
economic growth in a number of ways.
In recent years, the economic performance of some of the Tigers has flagged. Some
might want to attribute this to their relative income equality, since capital formation in Japan and
South Korea was only 1.1 percent and 1.6 percent during the 1990s. But this was in fact the
period in which income distribution in these countries became less equal, with fortunes made
(and later lost) in real estate, corporate and other dealings. Since economic behavior occurs "at
the margin" rather than being based on averages, it is difficult to draw definite causal inferences
with such complex relationships.
I had the good fortune to have W. Arthur Lewis as a teacher of development economics,
and I acquired from him a skepticism about capital formation as the cause of economic growth.
He considered it to be in general a consequence of growth, not being persuaded of the validity of
neoclassical economics' assumptions and preferring to think along the lines of more classical
economic theory. He did not regard market prices as an infallible equalizer of values, whereby
$100 worth of deodorants would be equal to $100 worth of productive land. I share Sir Arthur's
reservations, though the current theory and practice of economics is quite happy to equate
everything by market prices, even when it is acknowledged that these prices reflect very unequal
distributions of income which distort the forces of demand and supply. The price system, except
under unattainable conditions, is better able to maximize profits than to maximize human
welfare. It also fails to reflect adequately the needs and interests of future generations. But this is
not the time or place for a fundamental debate on economics.6

For critiques of contemporary economic theory and theorizing, offering well-considered alternative formulations,
see Leibenstein (1976), Daly (1990), and Ormerod (1998).

In any case, there is enough evidence now accumulated and analyzed to assure us that
relative income equality is not a necessary drag on growth, and we should know that there is
some significant evidence showing positive effects from relative equality. Reducing inequality
can thus be seen as a spur to economic growth, reducing poverty by that complex path rather
than by a direct process of income redistribution or transfer.

A "Life Chances" View of Poverty

As a social scientist who works on development, rather than an economist who tries to explain
all those things that can be denominated in terms of money (Alfred Marshall's definition of
economics), I would suggest the following perspective on poverty and inequality. If I were doing
today the kind of analysis I undertook with Cornell colleagues 25 years ago, I would still want to
look at income distribution data and to compare statistics such as top 20%: bottom 20% ratio to
assess magnitudes and trends.
If there are data detailed and extensive enough to use more refined indicators such as the
Foster-Greer-Thorbecke measure of inequality (Foster et al., 1984), this would be desirable
because it maps disparities in income distribution in more precise and meaningful ways. I would
also want to have some of the kind of qualitative (one might better say phenomenological)
assessments of poverty that were done for the World Bank by Deepa Narayan and her associates
to show the human face of poverty for its 2000-2001 World Development Report on poverty
(Narayan, 2000; and Narayan et al., 2000).
But increasingly the most meaningful measure of poverty, in my view, is one not found
in the literature. Assessing the lives of people -- their present living standards and conditions -- is
important, but I think poverty is most significant in terms of what it does to people's life chances

-- their opportunities to get educated, to have food, shelter and clothing that meet basic needs
(and more), to move not just a little way up the ladder of income distribution but to be able to
make some significant jumps and -- most important -- to give their children greater opportunities.
A life-chances indicator would tell us what is most significant and oppressive about
poverty: its stratification of society into relatively static as well as separate groups. This should
be of concern to almost everyone, not just those persons who bear the brunt of poverty. To be
sure, not everyone loses equally from a social arrangement of group stratification, but the losers,
who are more than just the poor, greatly outnumber the gainers from inequality. The poor
consume less because they produce fewer goods and services, which are consequently not
available to the rest of society.
What is the probability that someone who is born into poverty will in the course of his or
her life end up, reasonably stably, above the poverty line? Or put another way, what is the
probability that someone born into a family in that lowest 20 percent of households will
eventually head or co-head a household in the next higher 20 percent, so that her or his children
will have definitely "moved up the ladder," even if not out of poverty. One would like to know
this for persons born in the next higher quintile as well.7 Perhaps the worse thing about poverty is
the inescapability it creates from the problems, constraints and insults that are imposed upon the
poor, documented by Narayan and her collaborators in the recent World Bank studies cited
above. These deprivations and humiliations are of concern not just within a single generation, but
even more, from generation to generation.

In fact, there is often considerable transient movement into and out of a given quintile year to year as incomes rise
or fall. "Hard core" poverty occurs where persons remain within the bottom category for their entire lifetime. In
many countries, persons in the fourth quintile from the top qualify as "poor" because of their deprivation of income,
goods and services considered in absolute terms. In particularly impoverished countries, people even in the third
(middle) quintile may also be classified as poor. The fifth (botton) quintile is invariably "poor," and some or all of
them may be considered as "the poorest of the poor," so this is the focus of discussion here.

This raises, of course, "the Lake Woebegone problem." Not everyone within a country or
community can be above average -- by definition, a fifth of people must always be in the lowest
quintile. So whenever incomes or standards of living are compared, some inequality is
unavoidable, though this can be a greater or lesser degree. If one can move from a zero-sum to a
positive-sum framework for thinking about wealth and poverty, of course, it is easier to address
this problem, both analytically and psychologically.
It is an important question practically and ethically whether persons in the lowest bracket
are those who have the least physical and/or mental capacity -- or whether they are persons who
have, through no action or failing of their own, been deprived of effective opportunities to
develop their productive capacities to the fullest and to attain concurrent status and security. The
latter situation represents a loss not only for the persons who are so constrained by economic,
social, cultural or political circumstances, but also for the whole society. Its aggregate loss may
be even greater than that for the poor.
All in society remain somewhat poorer when others' productive potential is unfulfilled.
Not only are there fewer goods and services to be enjoyed, but there are fewer contributions to
the life and culture of a country, fewer songs and poems, less self-respecting friends to enrich
social relations, fewer persons of talent and integrity to hold political office, etc. Poverty
reduction is thus not something to be done just to benefit the poor. It is good for everyone except
for those persons who derive their wealth from extractive relations that are zero-sum, or worse,
negative-sum.8

Many economic relationships, such a certain share-cropping arrangements or petty trading sustained by obligations
of indebtedness, are negative-sum for society in that they impoverish a large number (or class) of people, by holding
down the productive potential of large numbers, while only a few gain from these exploitative relationships. The
gains of the latter are less than the aggregate imputable loss, but since the latter never gets considered, the social cost
to society of such relationships is invisible.

If there is one core process that underlies development it is that of creating positive-sum
relationships, such as through the production of value-added, through creation of consumer
surplus, through economies of scale from market integration and trade, or through broader
friendship networks (Uphoff and Ilchman, 1972: 75-121; Uphoff, 1996: 284-289). Economic
relations that are only zero-sum contribute little to development, even though they may add to
GNP as conventionally measured. Ironically, some negative-sum transactions, such as waste
disposal and pollution abatement, also add to GNP. What truly accelerates development are
positive-sum effects.
Such an understanding makes issues of poverty and inequality more central to
development theory, policy and practice. Poverty and inequality are not just "a blemish" on the
development record of a country, nor are they just "unfinished business" to be taken care of once
development has progressed fairly far. Where poverty is of the locked-in variety with stagnant
life chances for the poor, it reflects a pattern of development that is not basically driven by
positive-sum dynamics. It is a stunted form of development.
There are thus some strong practical as well as ethical reasons for "attacking poverty," to
use the subtitle of the World Development Report 2000/2001. The conditions of life for the poor
can be improved in various ways, directly through assistance, or indirectly but more sustainably,
by enhancing people's productivity. The latter can be accomplished: (a) by upgrading the factor
endowments of the poor (their human resources as well as physical ones), (b) by ensuring them
greater access to opportunities through general or specific processes of market integration that
enable them to employ their factors more productively, (c) by enhancing bargaining power
(where it is weak) to get more return for factors of production or goods and services, usually
through organization, or (d) by innovative initiatives of entrepreneurship and leadership that

10

alter structures of economic, social and political production in more productive directions9
Returns to factors of production are affected more by bargaining power than by intrinsic value
since the market by itself offers no means to appraise the latter.
This "dynamic" view of poverty and inequality should be of interest both to individuals -especially those within categories of "the poor" -- and to society as a whole. Living in poverty
has myriad degradations and debilitations, well documented in Narayan (2000), but being locked
into this status, with its attendant diminutions of life quality, makes a bad situation worse. The
prospect that one's children will, through no fault of their own, have not better chances of living
a more productive and fulfilled life, adds greatly to the psychological burden of poverty. (This
was not adequately addressed, in my view, in the surveys that Narayan and associates drew on
for their two volumes.) From a societal point of view, to the extent that more people and more
talent are locked into poverty, their contributions to GNP but also to cultural creations and to
political and social life are diminished.
"Life chances" can be measured fairly precisely at any point in time, at least
retrospectively, by tracking inter-generational mobility in economic and social terms -- through
interviews with persons according to some appropriate simple classification or scaling of
economic and social status (class). The implications for policy are that steps should be taken, and
investments made, which most surely increase the probability that people can move to a higher
rank, level or category in the future, and particularly that their children will be able to live stably
in a higher one.

These generic processes contributing to attaining higher levels of productivity were addressed analytically in
Uphoff and Ilchman (1972), esp. pp. 82-86.

11

Relative vs. Absolute Measures

This approach to assessing and attacking poverty leads into some sticky analytical and evaluative
terrain. It also argues against my preferred measure of inequality -- comparing as a ratio the
income going to the top 20 percent and that to the bottom 20 percent of households (or
individuals). Such a measure is zero-sum in that it uses a fixed proportion. The ratio can
improve, i.e., move lower, but it can only approach, never reaching, zero; 2:1 or 3:1 ratios would
represent a great victory in reducing inequality and alleviating poverty, when the ratio can
exceed 25:1 as in Brazil or El Salvador.
To assess progress in improving life chances, one would use appropriate poverty line
measures, secondarily looking at movement between quintiles of distributed income. There could
be considerable poverty alleviation if all households simply moved up in income level without
any change in rank-ordering (seen from an analysis of which households are in which quintiles).
However, the creation of greater opportunities for achievement and mobility based on merit will
not have been achieved, since one of the few things we know with some certainty in the social
sciences is that there is, in inter-generational terms, invariably some "regression toward the
mean" in terms of intelligence and other talents. A "rising tide that lifts all boats" should be
welcomed as an unprecedented policy achievement, but it would not represent a full-fledged
victory in the war against poverty. While more individuals would be better off, society as a
whole would not have gained as much as it could by opening up more opportunities for
leadership and responsibility based on talent and innovativeness.
Whether or not persons are in poverty can thus be viewed in either absolute terms
(poverty lines) or relative terms (ratios). Having a high degree of equality in a situation where
everyone is poor in terms of their possibilities for consumption and living a good life is hardly

12

satisfactory. For this reason, we are concerned with both poverty and inequality together, even
though they can be and should be analyzed separately.10
There is always some tension between the absolute and relative concepts of poverty.
"Poverty lines" get conceived and drawn as something absolute, producing certain numbers of
persons below them who thus belong in the category of "the poor." Even such lines are, however,
relative to some conception of human needs or social acceptability; and the data on which such
calculations are based are themselves often very debatable, the products of sampling and surveys
that can be contested.11 So one should not regard the numbers as being true or real in any
absolute sense. Rather, they are constructs, worth knowing, and of special value when they are
tracked over time or compared across regions or social groups, using the same standards for
derivation.

Implications of Social Mobility for Societal Efficiency and Equity

For assessing life chances, one needs to ascertain how much socio-economic mobility there is in
a society through surveys and observations that do not rely so much on measurement as on
simple categorizations, such as job classifications or possession of certain kinds of assets, which
are not very ambiguous. Comparing that status (category) of persons with that for their parents
can be reasonably objective even if recall must be used because the things being recalled are
simple and discrete. (The information solicited in surveys that enable analysts to report or
10

This dual concern is the focus of a new interdisciplinary program at Cornell on poverty and inequality. It is
directed by Prof. Ravi Kanbur, previously the World Bank's chief economist for Africa and then the head of its task
force for preparation of the World Development Report 2000/2001. Inequality is seen as both a cause and
consequence of poverty.
11
Anyone who places much confidence in such data should read the evaluation of such data in rural Nepal that has
been done by Gabriel Campbell and associates (1979). They found 50 to 200 percent discrepancies between the
results of surveys carried out conventionally and the reality that could be ascertained through in-depth
anthropological methods. Income data were indeed some of the most difficult to obtain accurately.

13

estimate current incomes is more subject to error). One can put aside the fact that there will
always be some persons below average, even way below average; the important question in this
kind of analysis is whether they are always the same persons, or always persons from the same
families.
This "life chances" approach to understanding and evaluating poverty and inequality can
be justified by efficiency as well as equity concerns. As noted above, one of the few things about
human beings known with reasonable certainty is that intelligence, or at least potential
intelligence, as well as other talents are distributed quite evenly across all populations, all races,
both genders, etc. In a country with a high degree of access to positions of higher income, status
and authority based on merit, the offspring of the families in the upper quintile, biologically
speaking, have some greater chance of being in that quintile in the next generation simply based
on natural talent. This will be augmented by various acquired, as distinguished from innate,
characteristics.
But this is only a chance, not a certainty. If all of those persons in the top quintile come
from parents who themselves have had that status, the country's economic, political and other
institutions are being directed by persons who have less than the greatest natural, innate
capability. They may have certain advantages of education and social connections that make
them effective in such positions, and this is not to be neglected. But the very highest intelligence
and other talents will not be among their endowments.
The law of "regression toward the mean" means that most children of the most privileged
group in a society will be less capable than their parents were and will deserve to end up in a
lower quintile than they were born into. The converse implication of this law -- that persons of
highest intelligence can and will be born into any and all social categories -- means that, on the

14

basis of merit, there should be many persons, indeed a majority in any generation, in the top
category who were born into lower quintiles and on the basis of their talent were able to rise up
the socio-economic ladder.
In fact, it is unlikely that any social policy aiming to end poverty and inequality can ever
succeed fully. The chances of the "first" really becoming "last" are negligible, even though it
can be fruitful to think about the implications of this (Chambers, 1997). The advantages of being
brought up in an advantaged family, with social contacts, psychological confidence, role models,
etc., cannot be redistributed except by heinous measures that are destructive for everyone in
society, as seen from the Khmer Rouge experience in Cambodia which tried to expunge all past
privileges by force.
What is possible, however, is to have an active policy of investment in developing human
capabilities including universal and high quality education and health care, with effective
programs of prenatal maternal as well as childhood nutrition. A progressive inheritance tax that
levels the economic playing field between generations could finance a good part of this, offering
at the same time the social utility of it becoming easier for persons with talent, imagination,
energy and social skills to rise, their way not blocked by less capable persons who had extrinsic
inherited advantages.
In India, there is the special problem, which few people are willing to talk about. Even
after 50 years, there is still strong residual discrimination against persons born into scheduledcaste or scheduled-tribe families. There are some exceptions, as some of these households have
been able to climb up some rungs on the socio-economic ladder. But this is one of the most
glaring sources of poverty and inequality in India: the continuing effect of a caste system several

15

thousand years old. A life-chances approach to evaluating poverty is particularly relevant where
we know that there are certain socio-cultural impediments to upward mobility.

Issues for India Today

The good news is that income distribution in India appears to have become more equal over the
past 35 years. When calculating the ratio of incomes in India going to the top 20 percent and the
bottom 20 percent, we found two sets of figures; one was from 1964-65 (National Sample
Survey) analyzed by Pranab Bardhan, and the other from 1967-68 (National Council of Applied
Economic Research) analyzed by K. R. Ranadive. These data sets produced quite different ratios,
6.0:1 and 10.9:1, which we averaged, to consider 8.5:1 as a representative figure for India
(Uphoff and Esman 1974: 147). The most current figures on income distribution in India (World
Bank 2000: Table 5) give a ratio of 5.7:1, as a result of 46.1 percent of income going to the top
20 percent, while 8.1 percent of income goes to the bottom 20 percent. This suggests that India
has made some progress in reducing inequality compared to earlier NCAER data, though not
with regard to NSS surveys.
But what is the vision and strategy of development that the Indian government and its
citizens will pursue? Will it be purely incremental, being satisfied to have moved annually some
number of individuals or household above the poverty line? Will there be longitudinal tracking to
know how this number compares with those who have, in this same time period, fallen below the
line? Will we know what kinds of persons are moving out of poverty, and what kind are sinking
into it? Aggregate numbers that balance these two groups out, perhaps with little net change, are
not very informative. A life-chances conception of poverty will focus on such data, and on what

16

can be done to create "one-way tickets" out of poverty because that is what "reducing poverty" is
taken to mean.
As suggested above, poverty should be seen as bad for everyone, not just for the poor.
Looking for ways to help people get themselves out of poverty -- note that I did not say, looking
for ways to get people out of poverty -- would focus on the obstacles for different categories of
persons defined as being among the poor. Often these will derive from socioeconomic and
sometimes political relationships that are extractive and exploitative, i.e., negative-sum, where
the gains of the few are in total less than the losses of the many. If improvements in life chances
are the measure and criterion of success, these relationships become unavoidable focuses of
concern, whereas with conventional poverty or inequality measures, any net incremental changes
are interpreted as positive, and there is no need to address structural impediments or resistances.
What will most improve life chances of the poor in India? Education and health care are
the two most obvious measures, which have the advantage of being positive-sum and not
requiring anyone else to lose thereby -- except those who have been exploiting cheap labor.
Having a more educated population is good for the large majority in a country, and having better
health has positive payoffs by reducing disease that can harm the better-off. Programs for fair
hiring and promotion are more difficult to install because they involve some reallocation of
opportunities, from less-qualified to more-qualified. But they are not impossible to promote, a
kind of fair employment practices system that would benefit employers because they are
supported in hiring and promoting on the basis of merit, which should improve the efficiency and
profitability of enterprises.

17

The Special Issue of Land Distribution and Access

A controversial but sound policy would be pursue a kind of land reform or redistribution that is
different from the classical "land to the tiller" program. I call this "universal access to land." It
would not try to give every household in the agricultural sector a size of holding sufficient to
produce a subsistence income as has been the usual policy objective when such redistribution has
been contemplated. In many places, there is not enough arable land to set up every household
wanting to practice agriculture with a so-called "economic unit." This constraint has been a
sufficient argument to get land distribution kept off the development agenda for the past several
decades.12
But the image of agriculture that underlies -- and is used to discredit -- the classical form
of land reform is an outmoded one. In most countries, including India, an increasing share of
rural incomes come from non-farm and non-agricultural sources. In part this represents a high
degree of desperation as poor rural households find that they must turn to other sources of
income to meet their basic needs. But it can also represent modernization and diversification of a
rural economy which is no longer solely dependent on agricultural and own-enterprise activities
for output and employment.
Two lines of argument support this suggestion, one emphasizing agricultural productivity
and the other human productivity. First, as arable land becomes relatively scarcer with
population growth, and demand for production continues to rise for this same reason, higher
productivity per unit of land becomes critical for further development. In almost all situations,
smaller holdings are more productive per hectare than larger ones because smaller ones are more
intensively farmed, while larger ones are farmed more extensively (Berry and Cline 1979).

12

I am pleased that it has been resurfacing recently in discussions of development strategy, e.g., Binswanger and
Deininger (1997).

18

Mechanized production which substitutes machines for labor raises profits more than it raises
production. Only where mechanization increases intensification, as with plowing that permits
cultivation of an extra crop, does it increase output. It is true that larger units of production
produce higher incomes, but not because of higher output per unit of land. Most of the gains are
due to economies of size rather than to technical economies of scale. Gains are based on
advantages of bargaining power rather than on real gains in efficiency.
Second, there can be very real gains in welfare that contribute to the productivity
obtainable from providing poor households with even small holdings. These units may be
considered "sub-economic" by analysts if one expects households to get all their income from
agricultural and own-farm pursuits; but they can add to the health, productivity and security
(bargaining power) which can help households begin moving up out of poverty. This can be seen
from two theses done some years ago at Cornell whose findings were ignored because they
pointed in a direction that prevailing economic development thinking was not prepared to go.
Research in India by Kumar (1977) found that -- other things being equal, in other words,
for the same level of household income -- children's nutritional status was higher if the
household owned some, even a small piece of land.13 This could be easily explained. If a
household had an opportunity to produce even a small share of the food that it needed, it had
more control over its food supply and would not be as vulnerable to hunger periods. The land did
not even need to be high quality since good management of the soil could improve it sufficiently
for growing vegetables and fruits and maybe some staple crop. If a household had only a small
plot, it was worthwhile investing labor in raising its productivity.

13

She also found that nutritional status was higher -- for any given level of household income -- the larger was the
share of this contributed by the mother. This is not counterintuitive once the relationship is pointed out: the more a
mother contributes to income, the more influence she can have over how income is used, and she is in a stronger
position to insist that a larger share be devoted to nourishing small children.

19

Research in Indonesia by Hart (1978) showed that -- other things being equal, including
controlling for level of education and thus for the inferred level of "human capital" -- households
that owned even a small amount of land had higher returns for their labor, i.e., they had higher
net wages per hour. If a family had even one-quarter acre (a tenth of a hectare) on which to grow
some food to meet its subsistence needs, its workers could hold out for more than the very lowest
wages being offered. Those completely landless workers who had no land to fall back on had to
accept whatever work was available. These were often jobs to which they had to travel several
hours in both directions, being desperate. The hourly returns for such employment were thus
pitifully low. Poor families with even some small amount of land were considered to be more
desirable "clients" by the more powerful "patrons" in the village, those persons who had larger
landholdings, so these poor families were better able to find employment locally and received
better wages for their labor. Also, they were more likely to get benefits like gleaning rights on
larger farmers' fields after harvest.14
The issue of "land access" should be put on the development agenda, even for a country
like India where formal land reform efforts have been mostly a failure (Herring, 1982), and
where person:land ratios are in many parts of the country quite high. Two generations of
population growth and resulting subdivision of land have accomplished at the upper ends of the
land tenure system part of what land reforms intended: the breakup of very huge landholdings.

14

This analysis empirically contradicted the claims of "the new household economics" which was in vogue at the
time. It claimed that returns to labor are in general efficiently and equitably determined because they reflect
productivity as embodied in human resources, particularly as improved through education. In this study of rural
realities that are surely not very dissimilar from those in rural India, it could be shown econometrically that the
returns to education were less than those to owning even a small amount of land. People got more for their labor
from owning some small amount of land, or having secure usufruct rights, than from having small increments in
education, mostly because by not owning land they were so vulnerable that they could not, in a labor-abundant
market, capture the value of their labor productivity. With some land, they could afford to get a larger share of the
value they added to production. This research has been elaborated and published in Hart (1986).

20

But there has been concentration in lower-upper and upper-middle echelons, and the number of
landless has continued to grow.
Exclusion from access to land in rural areas, coupled with poor or inaccessible schools,
no or non-functioning medical facilities, and social discrimination, means that there are several
hundred million Indians who are now -- or will in the next generation be -- denied the kinds of
"life chances" that ought to be a human right. Such life chances are essential for the progress of
an economy that is prosperous and dynamic in the modern world. The absence of life chances
will slow an economy due to the inertia of millions of persons who have been marginalized and
made dispensable by the economic system. They nevertheless need to meet their survival needs,
and they can be adversely affect the economy by becoming, in small or large numbers, strongly
negative social and political forces.
This perspective on poverty and inequality has various measurement and normative
aspects that can be addressed with more or less elegance, but it also has very practical and
political implications that need to be addressed with some sense of urgency. A danger of
preoccupation with the measurement aspects of poverty and inequality, especially if divorced
from normative considerations, is that analysis will have nothing to contribute to the redress of
practical needs and political pressures.

21

REFERENCES
Berry, R. A., and W. Cline. 1979. Agrarian Structure and Productivity in Developing
Countries. Baltimore, MD: Johns Hopkins University Press.
Binswanger, H. and K. Deininger. 1997. Explaining Agricultural and Agrarian Policies in
Developing Countries. Policy Research Working Paper 1764. Washington: World Bank.
Campbell, J. G., R. Shrestha, and L. Stone. 1979. The Use and Misuse of Social Science in
Nepal. Kathmandu: Centre for Nepal and Asian Studies, Tribhuvan University.
Chambers, Robert. 1997. Whose Reality Counts? Putting the First Last. London: Intermediate
Technology Publicatiions.
Daly, Herman. 1990. Steady-State Economics. Washington, DC: Island Press.
Foster, J., J. Greer, and E. Thorbecke. 1984. A class of decomposable poverty measures.
Econometrica, 52: 761-766.
Hahn, F. H., and R. C. O. Matthews. 1964. The theory of economic growth. Economic Journal,
December, 779-902.
Hamberg, David. 1966. Savings and economic growth. Economic Development and Cultural
Change, 17:4, 760-782.
Hart, Gillian. 1978. Labor Allocation Strategies of Rural Javanese Households. Ph.D thesis,
Department of Agricultural Economics, Cornell University.
_________. 1986. Power, Land and Livelihoods: Processes of Change in Rural Java. Berkeley:
University of California Press.
Herring, Ronald J. 1983. Land to the Tiller: The Political Economy of Agrarian Reform in South
Asia. New Haven: Yale University Press.
Krugman, Paul. 1995. Development, Geography and Economic Theory. Cambridge, MA:
Massachusetts Institute of Technology Press.
Kumar, Shubh. 1977. Composition of Economic Constraints on Child Nutrition: Impact of
Maternal Incomes and Employment in Low Income Households. Ph.D. thesis, Division of
Nutritional Sciences, Cornell University.
Kuznets, Simon. 1966. Modern Economic Growth: Rate, Structure and Spread. New Haven:
Yale University Press.
Leibenstein, H. 1966. Allocation Efficiency vs. 'X-Efficiency.' American Economic Review,
June, 392-415.

22

_________. 1976. Beyond Economic Man: A New Foundation for Microeconomics. Cambridge,
MA: Harvard University Press.
Lewis, W. Arthur. 1965. A review of economic development. American Economic Review, May,
1-16.
Morgan, T. 1969. Investment vs. economic growth. Economic Development and Cultural
Change, 17:3, 392-414.
Narayan, D., with R. Patel, K. Schafft, A. Rademacher, and S. Koch-Schulte. 2000. Voices of
the Poor: Can Anyone Hear Us? New York: Oxford University Press.
Narayan, D., R. Chambers, M. K. Shah, and P. Petesch. 2000. Voices of the Poor: Crying Out
for Change. New York: Oxford University Press.
Ormerod, Paul. 1998. Butterfly Economics: A New General Theory of Social and Economic
Behavior. New York: Pantheon.
Singer, Hans. 1966. The Notion of Human Investment. Review of Social Economy, March, 1-14.
Uphoff, Norman. 1996. Learning from Gal Oya: Possibilities for Participatory Development
and Post-Newtonian Social Science. London: Intermediate Technology Publications.
Uphoff, Norman, and Milton J. Esman. 1974. Local Organization for Rural Development:
Analysis of Asian Experience. Rural Development Committee, Cornell University, Ithaca, NY.
Report of comparative research project for the Asia Bureau, USAID.
_________. 1983. Comparative Analysis of Asian Experience with Local Organization and
Rural Development. In Rural Development and Local Organization in Asia, Volume III: South
East Asia, ed. Norman Uphoff, 263-399. New Delhi: Macmillan.
Uphoff, Norman, and Warren F. Ilchman. 1972. The Political Economy of Development.
Berkeley: University of California Press.
World Bank. 2000. World Development Report 2000-2001: Attacking Poverty. New York:
Oxford University Press.

23