THEORIES OF POVERTY

1
MODULE 1.1:
Introduction
The Old, Recurring Problem
Defined by modern standards, poverty had existed long before people became aware of this condition. As
Massey notes, "Poverty is old news. For thousands of years the great majority of human beings have lived and
labored at a low material standard of living." (Massey 1996:96). Awareness became increasingly acute with wars
of conquest and colonization and with the industrialization, capitalist development, and urbanization of recent
centuries. Great thinkers have pondered the causes as well as consequences of poverty and inequality.
Still, in the West after World War II, poverty came as something of a surprise; indeed, poverty was
"rediscovered” time and again (Polanyi 1944, 1957: 89-90, 103; Harrington 1962, 1967:140-141; Gordon 1972:
vii, 3; Rees 1998: 16, 19; Cotter 2002:535). Postwar poverty in underdeveloped countries came onto the global
policy agenda because of the Cold War (Myrdal 1971: 4-5).
Since the World Bank, IMF, and UN agencies focused their programs on poverty, a great deal more thought,
information and resources have been devoted to it. Yet, despite the material progress that many countries have
enjoyed in recent decades, poverty, inequality, and misery seem (from at least one viewpoint) to have persisted
and increased worldwide, in rich as well as in poor countries. Moreover, international aid agencies disagree
about the trends, the magnitudes, and the efficacy of the policies involved in the "wars” declared on poverty.
(Therien 1999).
Why Policy Needs Theory
This is a review of concepts and theories of poverty that may help us better understand the rationale for public
policies and programs designed to deal with the problem. Such policies and programs need to have congruent
theories about the relations between causes and effects if they are to succeed in implementing their objectives.
More or less explicit theories are also needed to provide some of the criteria and standards by which the
soundness of public measures, i.e. the effectiveness of the means used or proposed to achieve certain ends, can
be evaluated.
In practice, however, public policies on such difficult problems as poverty are too often based on faulty social
science premises, particularly those anchored on "strong" but erroneous causal reasoning (Rein and Winship:
1999). Incongruence may be rampant between manifest and "tacit” theories and between explicit and "implicit"
policies. Policies and practices may be as often running ahead, as falling behind, the theories intended to
rationalize them. (Gordon 1972: 4-5).
To ensure congruence, theories and policies have to be bridged by appropriate concepts and operational
measures. But a danger here is that these concepts and measures could be biased by the quantifiable indicators
chosen to represent the key variables (Streeten 1998). Empirical research could also diverge significantly from
theoretical work (Foster 1994: 365), making research results difficult to interpret.
Changing Concepts of Development
Theorizing about poverty has been at once enriched and complicated by the fact that the concepts and indices
used have been broadened beyond the income focus and made more multi-dimensional. This has been due to
changing ideas of development and underdevelopment, which has also been widened from economic, social, and
environmental development to "human development” in its manifold senses. (Wolfensohn & Bourguignon 2004:
3-4; Todaro 2001:13-20; Sen 1999; Streeten 1994; Srinivasan 1994). Yet the official literature and indicators may
remain focused on income (Satterthwaite 2001: 2; Satterthwaite 2003: 75), and differences persist about basic
anti-poverty strategy and policy, e.g., whether to "target" poverty directly and primarily or to deal with it
indirectly through broader approaches of economic growth.
Such issues have been further complicated by current processes of globalization and urbanization. Proponents
claim that globalization has improved national economies through freer and greater foreign trade and investment,
while opponents argue that it has worsened poverty and inequality within and between countries. Urbanization,
2
which is expected to mount in developing countries, is also seen as both a bane and a blessing, with poverty
becoming increasingly urbanized. (Buckley and Kalarickal 2004: 3). The spatial dimensions of poverty and
inequality have been receiving growing scholarly and technical attention, which hopefully could enhance
understanding and effective action. These problems, however, are multi-dimensional: economic, social, cultural,
and institutional as well as geographic.
Concepts of Poverty
Income Concepts
Concepts of poverty have revolved around low income as the criterion. Income represents "command over goods
and services to meet minimum needs." So the lack of income also means poverty in terms of basic needs (such as
food, shelter, and clothing). (Ornati 1967: 168-169). An agreed-upon budget for basic needs for a society is
called a poverty line. The World Bank has adopted about $1 a day as a global poverty line below which people
"are struggling to survive" (WDR 1990), i.e. poor. Higher poverty lines from $2 to $14.40 a day have been used
for richer countries. (UNDP HDR 1997: 13).
An important distinction is that between absolute and relative poverty. Absolute poverty is the situation of all
those falling below the established poverty line. This threshold is fixed across all resource distributions, time,
and even countries, though its value is adjusted for variable price levels and exchange rates. Over time, the initial
line set may fall well below a more realistic current level in a growing economy, e.g., the US$3,000 a year set for
Americans in 1960 would understate a higher threshold in 1985 (of say $5,300). (Foster 1998: 3).
Relative poverty measures the income "gap” or "economic distance” between the poor and the non-poor. Instead
of measuring poverty according to an absolute standard (the poverty line),
the situation of the poor is compared to that of more affluent groups. "When most Americans have a great deal,
those who have much less are poor regardless of their absolute level of income." (Fuchs 1969, quoted by Gordon
1972: 4). Even when the poor move up income-wise, they remain poor if left "too far behind" by the richer
groups.
However, most current discussions suggest that the absolute poverty standard has not been abandoned, and that
the two definitions co-exist, interact, and can be switched. (Foster 1998: 4).
Two indices of absolute poverty are the head-count, i.e. the total number of people below the poverty line; and
poverty incidence or the proportion of the poor to the total population. Relative poverty is measured by the
"income gap” or "poverty gap", i.e. "the average income shortfall of all the poor as a proportion of the poverty
line," (Streeten 1998: 15) or "the additional income needed by the poor to rise above the poverty line." (Mills &
Pernia 1994: 4).
The distinction between absolute and relative poverty suggests that trends in income levels and distributions can
take different directions; the poverty gap can widen amid economic growth. The distinction also seems useful in
setting policy goals. Most international programs aim merely at "poverty alleviation", while absolute poverty can
be "eliminated." But does this mean that relative poverty will always be with us? That seems likely, given
the wider dimensions of poverty.
Basic Needs
Although income implies command over resources to meet needs, the income criterion has limitations that make
the distinction a practical one. Income may not adequately represent basic necessities such as food, shelter, and
clothing. On the other hand, government programs that address such basic needs directly may substantially
improve the health and welfare of the poor without necessarily raising their income.
Thus, basic needs received a great deal of attention from international and national poverty programs. However,
this concept has its share of limitations. Needs change over time. One reason for a shift to relative definitions of
poverty was that absolute standards "ignored the increasingly expensive requisites for daily 'subsistence' in
urbanizing, technologically more complicated societies. Telephones and automobiles become more and more
indispensable in the United States, for example, even for the poor...." (Gordon 1972: 4)
Another factor favoring the shift was the thought that "relative deprivation," economic isolation, and inequality
would remain although absolute poverty might disappear with economic growth and with "more munificent
income maintenance policies." (Gordon 1972: 4).
3
Capabilities
An ever wider perspective suggests the limits of both income and basic needs concepts. Income "gives only a
partial picture of the many ways human lives can be blighted. Someone can enjoy good health and live quite
long but be illiterate and thus cut off from learning, from communication and from interactions with others."
(HDI 1997: 16).
This thinking is in line with efforts to re-orient the concepts of economic and social development to "human
development." These efforts were marked by the idea of viewing the social as the end rather than merely the
means of economic development, i.e. of regarding health and education as ends rather than just "human capital"
resources. (Streeten 1994).
More recently, Amartya Sen has sought to advance this view by looking at "development as freedom," i.e.
people's freedom to satisfy the ends of development. A component concept of freedom is that of "elementary
capabilities like being able to avoid such deprivations as starvation, undernourishment, escapable morbidity and
premature mortality, as well as the freedoms associated with being literate and numerate, enjoying political
participation and uncensored speech and so on." (Sen 1999: 36).
UNDP's Multi-Dimensional Concepts
Since 1990, the U.N. Development Programme has produced reports that define development as "'both the
process of widening people's choices and the level of their achieved well being.'" (UNDP 1990: 9, quoted byt
Srinivasan 1994: 238). UNDP defines poverty as meaning "that opportunities and choices most basic to human
development are denied - to lead a long, healthy, creative life and to enjoy a decent standard of living, freedom,
dignity, self-respect and the respect of others." (UNDP 1997: 15). It thus views poverty as multi-dimensional
involving three perspectives:
"Income perspective. A person is poor, if, and only if, her income level is below the defined poverty line...."
"Basic needs perspective. Poverty is deprivation of material requirements for minimally acceptable fulfillment of
human needs, including food ... health and education and essential services ... employment and participation."
"Capability perspective. Poverty represents the absence of some basic capabilities to function.... The
functionings relevant to this analysis can vary from physical ones [being well-nourished, adequately clothed and
sheltered, avoiding preventable morbidity] to more complex social achievements such as partaking in the life of
the community." (UNDP 1997: 16)
UNDP notes that "The capability approach reconciles the notions of absolute and relative poverty, since relative
deprivation in incomes and commodities can lead to an absolute deprivation of minimum capabilities." In
addition, it identifies another pair of perspectives: the "deprivational" one focusing on the way the poor and the
deprived fare, and the "conglomerative" one viewing the advances made by all groups in each community.
Accordingly, the U.N. agency has devised indices for development and poverty:
"Human development index [HDI] ... measures the average achievements in country in three basic dimensions of
human development - longevity, knowledge and a decent standard of living. A composite index, the HDI thus
contains three variables: life expectancy, educational attainment (adult literacy and combined primary, secondary
and tertiary enrolment) and real GDP per capita (in PPP$).
"Human poverty index [HPI] ... measures deprivation in basic human development in the same dimensions as the
HDI. The variables used are the percentage of people expected to die before the age 40, the percentage of adults
who are illiterate, and overall economic provisioning in terms of the percentage of people without access to
health services and safe water and the percentage of underweight children under five." (UNDP 1997: 14).
"Human poverty" is distinguished from "income poverty," now the dollar-a-day standard for developing
countries. Both indices are used to rate and rank countries' poverty status and trends. An application of both in
the Human Development Report 1997 (Fig. 1.1, pp. 21-22) shows that the HPI and the income poverty incidence
index can move in different directions.
In Egypt, for example, less than 10% of the people were income-poor but 35% were "affected by human
poverty." In China and the Philippines, about 30% were income-poor but less
Theories of Poverty 5
than 20% were "human-poor." Thus, some countries "have done better in reducing human poverty than income
poverty...." (UNDP 1997: 21-22).
4
Theories of Poverty
Contemporary Theories: The World Bank
For its part, the World Bank has defined poverty in almost identical terms with those of the UNDP. In addition,
in its 2001 "Attacking Poverty" report, it suggests some of the basic sources of the poor’s deprivations,
particularly in the notions of "vulnerability” and institutional "ill treatment." Aside from those "deprivations that
keep them from leading the kind of life that everyone values,” poor people
"also face extreme vulnerability to ill health, economic dislocation, and natural disasters. And they are often
exposed to ill treatment by institutions of the state and society and are powerless to influence key decisions
affecting their lives. These are all dimensions of poverty." (WB 2001: 1).
Contemporary concepts, including official ones, view poverty as a multi-dimensional, complex condition with
various causes and correlates, raising difficult theoretical and policy issues. As one author notes, poverty is
"over-determined", with many possible explanatory causes (Mokyr 1998: 70). The causes and the symptoms are
not easy to tell apart, and, where the distinction is possible, their relative priority in public policy is often
debatable.
Such difficulties, however, have not fazed theorists. According to the World Bank's World Development Report
2000/2001, the causes of poverty are implied in the "dimensions highlighted by poor people," i.e. (1) lack of
income and assets to meet basic necessities, (2) sense of voicelessness and powerlessness, and (3) "vulnerability
to adverse shocks, linked to inability to cope with them." (WB 2001: 34).
Experience in the 1990s, the Report notes, showed that economic growth improves the poor’s income, and that
their expanded capabilities remain central, instrumentally and intrinsically, in poverty reduction strategies. But
new evidence also revealed the importance of institutional constraints on reforms, vulnerability factors, and
global forces of integration, as well as the renewed cogency of gender, ethnic, and racial inequalities. (WB 2001:
34).
As another departure from income, the World Bank Report takes the economic concept of "assets" as a starting
point in understanding the determinants of poverty - i.e. people's assets, "the returns to (or productivity of) these
assets, and the volatility of the returns." (WB 2001: 34). Assets include:
(1) Human assets (e.g. capacity for basic labor, skills, and good health;
(2) Natural assets (e.g. land);
(3) Physical assets (e.g. access to infrastructure);
(4) Financial (e.g. savings and access to credit); and
(5) Social assets (e.g. "networks of contacts and reciprocal obligations that can be called on in time of
need, and political influence over resources.")
The returns to these assets depend not only on access to markets but also on the institutions of state and society,
including those that define and enforce property rights and norms about common resources; on prevailing
patterns of gender, ethnic, racial or social discrimination; and on public policy and state interventions "which are
shaped by the political influence of different groups." (WB 2001: 34).
From the foregoing, it seems that poverty may be explained in terms of various kinds of factors, including
economic, social, political, and natural factors. Some of these may be categorized as institutional factors (the
Report cited here seems careful not place "market” in the same category as "state and society."). Poverty also has
geographic, technological, and cultural dimensions and variables.
As suggested by the multi-dimensionality of poverty, these various factors often work together to "cause” or
"determine" poverty or affluence. But theorists may emphasize some factors as more determinative than others,
raising issues that have animated policy controversies over the past two centuries or so, up to the present time.
Earlier Theories
In much earlier periods, even when food surpluses had allowed more complex, differentiated social groups to
develop, certain belief systems simply assumed that most people were destined for a life of want while a few
enjoyed a near-monopoly of wealth, status, and power. (Cf. Goldsmith 2001, who argues that most preindustrial
people were well until "development created poverty.") This fatalist belief is implied by the idea that accidents of
birth or geography could be "destiny" condemning some groups to poverty. Reflections about growing affluence
5
in the modernizing West, however, did throw some light on the question why some people were poor,
particularly as a result of the dislocations occasioned by the rise of capitalism and the agricultural and industrial
revolutions.
One reason for mass poverty was the narrow pre-modern concept of wealth as being due only to an elite. A
related one was the feudal ideology that set a ceiling on satisfaction and banned excellence in production as
excess. The Mercantilists sought to expand wealth but only for the emerging merchant class. The Physiocrats
favored production over mere acquisition, but only through agriculture. Adam Smith (Wealth of Nations, 1776)
extended this idea to industrial production in a "democratic, and hence radical, philosophy of wealth" as goods
produced for all. In Smith’s philosophy, " gone is the notion of gold, treasures, kingly hoards; gone the
prerogatives of merchants and farmers or working guilds...." (Heilbroner 1961: 34-35).
Smith observed that no society could be flourishing and happy if most of the people were poor and miserable.
With capitalist market society emerging, efficiently productive with its division and specialization of labor and
"invisible hand," he was optimistic about social progress. But in England and elsewhere, the changes occurring
were taking a different turn. From the 16th century onward, pauperism flourished in the rural and then in the
urban areas as farm laborers were uprooted by the enclosure of common lands, driven to the cities and then, for
many, driven back by the scarcity, low pay or uncertainty of urban employment.
"Where do the poor come from?" According to Karl Polanyi (The Great Transformation, 1957), this was a
question that exercised many pamphleteers and philosophers as the 18th century wore on. They blamed
pauperism on various proximate causes, from wage disparities, to type or number of animals used, to "bigoted
diets" and to the eating habits of the poor. To Polanyi, the more basic causes of high unemployment rates were
the excessive fluctuations in the growing foreign trade, the growth of manufacturing, and the dislocating
"territorial division of labor” created by the enclosure movement and the conversion and consolidation of farm
holdings. (Polanyi 1957: 90-93).
Theories of Poverty 7
When no satisfactory answer could be found, some thinkers simply decided that "pauperism and progress were
inseparable." Most poor people were found, not "in barren countries or amidst barbarous nations, but in those
which are the most fertile and the most civilized," and it was axiomatic "that the wealth of a nation corresponds
with its population; and its misery corresponds with its wealth." (Polanyi 1957: 103). As the English economy
expanded, the "unprofitable poor" were expected to increase further. Many cures were proposed, but the thought
persisted: " ... were not the wants of the poor ('... prudence to relieve, but folly to cure’...) essential for the
welfare of the state?" (Heilbroner 1961: 45).
If poverty was attributable to underproduction, a more dismal idea was that it was due to population growth
overtaking production. This was the specter that Thomas Malthus raised (Essay on the Principle of Population,
1798) and ascribed to "the human reproductive urge...." (Heilbroner 1961: 45; Ross 1998). For their part, utopian
thinkers blamed poverty on a defect in the organization of labor, and proposed ways to enable the poor to find
employers or to pool their labor for their own production and benefit. The best known was Jeremy Bentham's
Panopticon plan to replace prison labor with the poor and to turn the latter into industrial shareholders. This plan
suggested that one basic cause of unemployment was the substitution of machinery for "superseded hands."
(Polanyi 1957: 107).
Karl Marx dismissed both Malthus' and utopian ideas but confirmed that poverty was inevitable due to capitalist
exploitation. Capitalists extract "surplus value" and profit by making workers work longer hours than they
should (for their own reproduction) without extra compensation. Their wages and working conditions are kept
low through a "reserve army" of the unemployed and through the substitution of labor-saving machines when
wages threaten to rise. Only when the proletariat overthrows the ruling class would they be liberated from
"misery, oppression, slavery, degradation, exploitation ...." Marx was sure of the eventual demise of capitalism,
but did not say if poverty would be eradicated in the socialist and communist future. (Heilbroner 1961: 132-136).
More Recent Perspectives
In the early 1960s, when the U.S. was enjoying growing affluence, an economist declared that both Malthus and
Marx had been proven wrong:
"Most economists have long since given up the idea that a progressive society needs the threat of poverty to
induce work and sobriety in the lower classes. Similarly, one can consign to folklore the ideas that some are rich
only because others are poor and exploited, that if none were poor then necessary but unpleasant jobs would go
6
undone, that the middle class has a psychological need to exclude a minority from above-poverty living
standards, and that poverty is a necessary concomitant of the unemployment which necessarily accompanies
economic growth." (Lampman 1964, in Budd 1967: 177).
Yet another, more recent synthesis argues that the theoretical-ideological divide has persisted. According to
Michael Todaro (2000: 181-182), the "traditional argument” is that "highly unequal distributions are necessary
conditions for generating rapid growth." The rich save and invest significant proportions of their income while
the poor spend all their income in consumption. GNP growth derives directly from the national income saved, so
that highly unequal income distributions would lead to more savings and faster growth.
MODULE 1 8
The "counter-argument," Todaro continues, is that (1) widespread inequality and poverty deprive the poor of
opportunities to invest and drive them to have many children as financial security; (2) the rich in poor countries
squander their incomes on luxuries or stash them abroad and do not necessarily save and invest more than the
poor; (3) low incomes and low levels of living can lower the poor's economic productivity and slow down
growth; (4) raising the poor’s income will raise demand for basic domestic products like food and clothing,
stimulating investment and growth more broadly; and (5) reduced mass poverty and income disparities can
stimulate economic expansion through wider public participation in development. (Todaro 2000: 182-183).
International institutions like the UNDP and even the World Bank probably would agree with Todaro that the
oft-discussed "trade-off" between growth and poverty is not critical and that improved equity is consistent with
growth. However, "traditionalists", particularly those of a neoliberal or neoclassical persuasion, also die hard.
After 150 years since the Communist Manifesto issued by Marx and Engels, Nancy Birdsall finds that globally,
"the old saw is still correct: The rich get richer and the poor get children." But she accepts inequality as a
justifiable result of prosperity and, with Latin American experience in view, blames history and politics, the
poor's "rational decisions" to beget more children, elite capture of prosperity in poor countries, and bad
(protectionist and populist) economic policies. Democracy and open markets, global market integration, and
technology have also been creating a new divide among workers. In the end, Birdsall suggests accepting
inequality as "the natural outcome of prosperity" and gives pointers on learning to live with it as Western
countries did. (Birdsall 1998: 4-7).
While agreeing that prospertiy could produce poverty and inequality, Charles M.A. Clark blames capitalist
society for creating artificial scarcity and underproducing for the sake of profit, so that the rich could get richer
at the expense of the poor. To this end, neoclassical economics switched from a science of abundance to one of
scarcity, and to a "demand-constrained" business system that must maintain a rate of return on wealth and "the
social power that attaches to 'scarce' wealth." (Clark 2002: 416)
John Maynard Keynes thus noted that capital was being kept scarce and concluded that "'in contemporary
conditions the growth of wealth, so far from being dependent on the abstinence of the rich ... is more likely to be
impeded by it [i.e. abstinence from consumption, by saving for investment]. One of the chief social justifications
of great inequality of wealth is, therefore, removed." (Keynes 1936, quoted by Clark 2002: 417). But to remain
rich, or grow more so, the rich must maintain a high rate of return, through high unemployment and lower
economic growth. "In a very real sense, just as in Plato, wealth can cause poverty." (Clark 2002: 417).
Neoclassical economics, Clark says, tries to explain both wealth and poverty with the same theory and
individualist paradigm. Wealth derives from productivity and its rewards for effort, "while poverty is ... due to
the absence of productivity and the inability or unwillingness to work and wait." The theory errs in assuming that
the key determinants are individuals' characteristics and not social structures and institutions. This absolves
society and the rich of any responsibility for creating poverty and having a small role in its alleviation.
Instead, Clark offers an "institutionalist theory" positing three central processes:
(1) Rather than stemming from the "niggardliness of nature" or unlimited wants of human nature, scarcity is
socially created through "conspicuous consumption" and "industrial sabotage," as Thorstein Veblen theorized
(Theory of the Leisure Class, 1899). The leisure class keeps raising consumption standards to maintain the
scarcity value of the
Theories of Poverty 9
goods consumed, including the basic needs of the poor for land, housing, etc. The productivity of modern
technology and industry is kept in check through industrial concentration to limit competition and by keeping
interest rates "too high".
(2) Wealth, as well as poverty, is created through social exclusion, starting with the institution of private property,
7
which assigns ownership rights to a person or entity at the expense of society. Redundant, easy-to-refill jobs
generate low pay and low or no benefits due to strong competition. While exclusionary devices like unions,
minimum wages, and tenure reduce labor competition, higher-level assets and incomes require higher barriers to
protect their scarcity value. Such barriers keep the poor from participating in the economy and society.
(3) Economic costs, which are not allocated through normal market operations, are assigned or shifted to
disadvantaged groups. These include externalities like pollution and the full cost of the worker often borne by his
family and others outside the workforce. Wealth is created through a utility owner’s ability to shift "system
losses", for example, to his workers, neighbors, and society as a whole, who are correspondingly impoverished
thereby. (Clark 2002: 417-419).
Clark concludes by observing that poverty in the U.S. has not declined appreciably despite significant growth
because this growth has been attained at the expense of the poor. "When the amount of wealth has grown much
faster than the output of the economy, part of its has to come from either a redistribution of existing wealth and
incomes ... or in a shifting of costs away from capital.... In a very real sense, the last thing the poor need is more
accumulation of wealth." (Clark 2002: 419-420)
Birdsall tends to confirm the trend of growing inequality with economic growth in the U.S. and elsewhere.
Income inequality in the U.S. increased due to losses at the bottom as well as gains at the top (1973-1993), with
educational gains possibly reinforcing rather than offsetting income inequality. "Elsewhere, the forces of change
- whether the spread of capitalism and global integration, or simply the march of technological progress - have at
best reinforced, or at worst exacerbated, high inequality." Only 8 out of 45 countries showed any improvement in
income distribution. (Birdsall 1998: 2).
Cultural and Other Theories
Poverty, along with wealth and inequality, has been viewed from other perspectives or other variants of
economic and non-economic theories. The American economist Robert Lampman sought to explain the
persistence of involuntary poverty in the U.S. in terms of certain external events (e.g. accidents of birth), social
stratification, limited ability or motivations, and a "subculture" that develops and "sustains values and attitudes
that are hostile to escape from poverty." (Lampman 1964: 177-178).
Cultural as well as sociological theories of poverty have abounded since the 1950s. Cultural theories attribute
poverty to features of a community’s distinctive values, beliefs, traditions, techniques, and ways of life that
prevent it or its members from aspiring for and achieving material progress and social mobility. One of the best
known was the "culture of poverty" concept propounded by Oscar Lewis, who characterized the poor as isolated,
inward-looking, and weighed down by strong feelings of marginality, dependency, alienation, inferiority, and
powerlessness against existing institutions. (Lewis: 1998).
MODULE 1 10
Cultural explanations have been used for broader patterns of economic development and decline. Max Weber’s
association of Calvinist frugality and work ethic with the rise of capitalism is a familiar one. Another is the
Confucian ethic of discipline, which has been cited as responsible for both the economic "miracle” in East Asia
and the financial crisis of the late 1990s there. A similarly alternating role has been ascribed to the "Asian
values" that Lee Kwan Yew of Singapore claims to be a crucial source of economic success in the region but
which critics cite as a source of authoritarianism and corruption. (Pieterse: 2004: 48).
Rao and Walton (2004) review the literature to provide an "overview of the role of culture in reproducing or
alleviating poverty." One extreme view is that "‘culture matters’ because societies steeped in traditional cultures
are unsuited to market-oriented development and are thus fundamentally hampered in their pursuit of growth." In
Huntington's "clash of civilizations" hypothesis (Huntington 1998), "poverty and low rates of growth are deeply
affected by adverse rules and norms that reduce incentives for mobility and investment." Weber's "Protestant
ethic” is wrongly invoked here as the opposite of "toxic cultures." (Rao and Walton 2004: 1-2, 12-13).
At the other end, Rao and Walton continue, are "the cultural critics of the discourse of Development," who see
culture as "a system of control that creates and extends existing... inequalities” between and within rich and poor
countries. To these critics, this system is perpetrated by economists who have shaped the dominant neocolonial
modernization perspective on development since the 1950s by constructing the "Third World" and reifying,
through the IMF and the World Bank, the distinctions between the North and the South. (Rao and Walton 2004:
13)
As a middle ground, Rao and Walton develop "constraining preferences” as a concept to depict the strength of
8
cultural influences on people's chances of becoming poor or wealthy. By that concept they "refer to culturally
produced dispositions, beliefs and behaviors" that help determine one’s life chances. "For instance, preferences
derived from the Hindu caste system may create an acceptance of hierarchy and constrain the motivation for
mobility, but these beliefs are also simultaneously external constraints ...." (21-22)
Culture constitutes resources ("cultural capital") that can be used by some groups to keep others poor while
enriching themselves. A related concept, popularized by Robert Putnam, is "social capital," which "emphasizes
the social networks available to people to access and mobilize resources, and contributes to inequality because
elites are able to access internal and external social networks that are more powerful and wealthy. By contrast,
the poor have less influential networks that, while helping them cope with the vicissitudes of life, restrict their
chances for mobility." (23)
While the "culture of poverty" theory was criticized as being negative, static, and tending to "blame the victim"
-- a charge that Oscar Lewis would deny -- Putnam's "social capital" has been viewed as too sanguine about its
contributions to participatory democracy and as indiscriminate about the kinds of civic associations that might
make such contributions. However, where Lewis’ could once be faulted for arguing erroneously that the poor
were marginalized or were outside the cultural mainstream, as other authors had also done, such ideas have been
vindicated by more recent developments in Latin America. In this context,
" ... there is now increasing evidence that although classic marginality may have lacked empirical veracity in its
earliest iteration, changing economic conditions born of the structural adjustment and austerity of the 1980s,
together with neo-liberal restructuring of the 1990s, is today creating the very conditions and cultural
constructions conceived and predicted by Nun, Quijano and Lewis in the 1960s. Rising unemployment,
declining opportunities in even informal sector activities, a rise of private provisioning within a barter economy
(the
Theories of Poverty 11
trueque system in Argentina), social exclusion and new dimensions of marginalization, rising violence and
insecurity - these are all-too-frequent features of the contemporary urban scene." (Ward 2004: 185-186)
"Social exclusion” is thus another term associated with poverty. This phrase refers to denial of membership and
participation in community or social life to certain people because of their individual or group characteristics. It
has greater resonance in Europe and other countries with many foreigners. Refugees, migrants, and guestworkers,
for example, may be denied citizenship and its rights. This "has resulted in the charge that the EU is becoming
'Fortress Europe' (Mitchell & Russell, 1994)." Even some who enjoy citizen status, however, may be "denied full
participation in social, cultural and economic life....” The concept is broader than poverty ("rediscovered" in the
1960s) and wider than multiple deprivation ("discovered" in the 1970s). (Rees 1998: 20; Streeten 1998: 24).
Racial, ethnic, linguistic, religious, and other cultural characteristics have of course been well recognized as
enduring bases of undue exclusion, discrimination, inequality, and poverty. Current theories, however, warn
against mistaking culture for nature, i.e. regarding it as immutable. Though it is durable, culture is amenable to
variation and change, with changing circumstances and through public action that is both "culture-sensitive and
socially-aware." This strategy, though, may have its pitfalls, such as possible elite capture of a poverty-reduction
scheme undertaken through rather than around a traditional community power structure.
Whereas periods of prosperity have justified some optimism about the possibility of escape from poverty, current
conditions reflect a convergence of economic, social, and political forces in the geographic concentration and
isolation of the poor, particularly in urban areas.
Spatial Dimensions: Urban Poverty
Recap of Theoretical Approaches
Thus far, the theories of poverty outlined above may be grouped into two kinds of approaches: "individual” and
"structural” approaches. The first emphasizes "the characteristics, attitudes, or behavior of the poor as the roots
of poverty,” and is exemplified by theories of the culture of poverty and human capital theory. The second
approach stresses the amount and distribution of economic and social opportunities "‘external to, and coercive
over'" individuals and that result partly from political economy. (Cotter 2002: 535-536). Put another way, some
theories attribute poverty exclusively or primarily to certain characteristics of the poor as groups or as
9
individuals, while others blame mainly a or "the system” of factors and forces external to the poor or somehow
beyond their control.
As Cotter suggests, these two theoretical approaches may and are often combined as complementary rather than
contradictory approaches, such as those embodied in contemporary "multi-dimensional" notions of poverty.
However, debate may persist as to the relative primacy of certain basic sources over others. This is the case, for
example, between geographic and institutional factors. (Rodrik & Subramanian 2003, Sachs 2003).
Place and Poverty: Urbanization
Despite claims that the world has shrunk due to advances in modern technologies and to other globalizing forces,
some authors have reasserted the independent influence of place and space on development and poverty.
Geographic location, climate, and natural resources (or the lack thereof) remain important determinants of
agricultural and industrial productivity through their relationships with transport costs, "the disease burden," and
technology. Landlocked or hinterland countries or areas far away from major markets are unlikely to prosper due
to high costs of transportation, communication, and other means of access to resources outside. Thus, human
groups in such isolated conditions could be "fated" to intergenerational poverty. (Sachs 2003; Gallup & Sachs
1999; Krugman 1999; Diamond 1999).
The mounting urbanization* of the global population has underlined the importance of the rural and urban
patterns of poverty incidence. Most poverty remains rural, but has been urbanizing since the 1950s at rapid rates
-- mostly in countries whose cities are least prepared to receive fresh waves of migrants. (Ravallion 2003; Piel
1997). Although migration from rural areas has accounted for much of the urbanization process and the poverty
associated with it, and there may have been some amount of "ruralization” or hybridization in urban areas, the
differences as well as commonalities between the two contexts are nonetheless important in characterizing and
explaining the problem.
Urban Context of Poverty
The poverty consequences of migration may be summed up as follows:
1. The migrant encounters greater density, scale, diversity and complexity in the urban environment. "Urbanism"
imposes more intense, competitive pressures for survival with fewer of the comforts of community. The sources
of both affluence and poverty become more diverse and complex in the city.
2. Livelihood resources are transformed and seem scarcer. Instead of living "off the land," the migrant must
count more on location and access to jobs and markets. The urban economy is more of a money economy, with
food prices higher than in rural areas. (Streeten 1998: 17)
3. Decent shelter and land tenure are more expensive and harder to come by, with "informal" settlements often
the only option for poor in-migrants. Housing is a more complex commodity subject to more public regulations
and anti-poor practices like large-lot zoning and bank "red-lining." (Strassmann 1996: 212-213)
4. Needs as well as resources change in the urban setting. Luxuries like cars, TV, and telephones become
necessities, and "inventions" like bottled drinking water become "the mother of necessity." (Diamond 1999: Ch.
13).
5. For some, the city is an opportunity for mobility and integration in the economic and social mainstream. For
most poor migrants, it means concentration and isolation in a trap of marginal existence.
* The UN Centre for Human Settlements (HABITAT) states that "The greatest challenge will present itself in
Africa and Asia, where an explosive demographic change is expected.... By 2015, 153 of the world's 358 cities
with more than one million inhabitants will be in Asia. Of the 27 'megacities' with more than million inhabitants,
15 will be in Asia. There are even indications of forthcoming megacities with 20 or even 30 million inhabitants,
urban agglomerations of a size never known before in Šhuman history, most of which will be in developing
countries."
David Satterthwaite, while acknowledging that there are basic similarities between rural and urban
poverty, also points out their differences in the following illustration:
"For instance, the cause of poverty for a rural household that relies on a small landholding and that
suffers from a low crop yield is not the same as for an urban household in a squatter shack community
whose main income earner has lost a job due to recession or ill health or has suffered a drop in real
10
income. Programs aimed at reducing rural and urban poverty need to recognize these differences."
(Satterthwaite 2001: 1).
Satterthwaite adds, however, that there is also a "need to recognize the links between rural and urban
areas. A rural household’s response to poor crop yields may be to send one of its members to an urban
areas to seek work; an urban household may respond to declining income by sending their young
children to rural areas."
Cumulative Deprivations
A recent World Bank Sourcebook for Poverty Reduction Strategies (2002) presents a similarly “stylized
comparison” of rural and urban areas and their respective challenges for the poor, but also stresses the linkages
and mobility between them. (World Bank Klugman 2002: 62-63). In its chapter on urban poverty, the authors
present a diagram (reproduced below) to illustrate the point that “Urban poverty is often characterized by
cumulative deprivations – that is, one dimension of poverty is often the cause of or contributor to another
dimension.”
SOURCE: Baharoglu & Kessides 2002: Figure 16.1, p. 127.
Cumulative Impacts of Urban Poverty
Lack of access to credit for business or house
Lack of employment; inability to have a regular job, lack of regular incomand social security, poor nutrition
Poor health, poor education
Sense of insecurity, isolation, and disempowerment
Inability to afford adequate housing
Tenure insecurity, evictions, loss of small savings invested in housing
Unhygienic living conditions, low–quality public services
Satterwaite has argued that the scale and depth of absolute poverty in developing countries “have long
been underestimated for two reasons. The first is that estimates are based only on income levels and
take no account of other deprivations such as very poor housing conditions and lack of basic services.
The second is that the income–based poverty lines used to make these estimates are set too low in
relation to the cost of basic needs in most urban centers…” (Satterwaite 2003: 75). This author likewise
challenges the notion that the urban poor contribute to environmental degradation, and, on the contrary,
finds “strong evidence that urban environmental hazards are major contributors to urban poverty.” (73).
To some observers, the "pauperization" of earlier centuries is being repeated on a grander scale with modern
technologies helping multiply farm productivity for capitalists at the expense of most peasants, with the
productivity ratio between them rising from 10:1 to nearly 2000:1. (Amin 2003: 1) Driven from stagnant rural
areas, they end up in the ghettoes and informal settlements of urban areas where most remain bottled up and
suffer from "spatial mismatches" between the jobless and job opportunities -- metropolitan in the West but on an
international scale in poor countries. While some urban poor manage to move out and up, their very departure
becomes another deprivation (i.e. of community leadership and social capital) for those left behind. (Cf.
Browning & Cagney: 2003).
11
Concentration and Segregation
The concentration and segregation of the urban poor are themselves a defining process by making people more
acutely aware of poverty and inequality. Massey provides an interesting historical account of this process. In
premodern times, deprivation existed at low densities, with the masses of the poor hardly in contact with the
small elite except in the cities. Here "preindustrial technologies permitted neither the separation of work from
residence nor the segregation of the elite from the masses." With the 19th century industrial revolution and
urbanization, however, affluence and poverty became geographically concentrated for the first time. "Within
cities new transportation and communication technologies allowed the affluent to distance themselves spatially
as well as socially from the poor, causing a rise in the levels of class segregation and a new concentration of
affluence and poverty." (Massey 1996: 396).
Massey notes that this process eased shortly after World War II when the developed countries experienced an
economic boom that a growing middle class shared with the rich. But after 1970, "the promise of mass social
mobility evaporated and inequality returned with a vengeance, ushering in a new era in which the privileges of
the rich and the disadvantages of the poor were compounded increasingly through geographic means." Massey
predicts that the segregation will continue in the 21st century, as most of the poor will live in urban areas and the
rich will seclude themselves in enclaves of affluence. (Massey 1996: 396). The "edge cities" and gated
communities in the U.S. have already highlighted this process, though they are not a novelty in the metropolises
of poor countries.
Some changes in the sources of urban population growth and poverty have been suggested by recent studies. One
is that while migration accounted significantly for urbanization in LDCs in earlier periods, in recent years it has
been due more to urban expansion into rural areas as well as natural increase and migration. Moreover, the poor
have not necessarily been the most active migrants, and the growth in the numbers of the urban poor has been
due to second–generation natural increase as well as in–migration. (Baharoglu & Kessides 2004: 127).
According to another
Theories of Poverty 15
recent report, “… many slum dwellers are not new immigrants who recently arrived from rural areas in search of
better livelihoods. Today, many of the 100,000 pavement dwellers in Mumbai, for instance, are
second–generation residents …” (Buckley & Kalarickal 2004: 3).
Institutions: Government Policy as Cause
Institutional explanations of poverty, such as those traceable to Veblen, have abounded, especially with the
advent of "new institutionalism” in the social sciences. Some authors argue that institutional rather than
geographic and other factors explain the relative wealth and poverty of nations. Their statistical studies, they
claim, show that "the quality of institutions,” by which they mean the "rules of the game in society" particularly
those pertaining to property rights and the rule of law, "overrides everything else.... geography has, at best, weak
direct effects on incomes, although it has a strong indirect effect through institutions by influencing their quality.
Political theories about the state, including the welfare state, differ on the question whether it matters at all in
influencing social well-being. (Myles & Quadagno 2002). The weight of current scholarly opinion, though,
seems to be on the affirmative side: Government and its policies could create wealth and poverty, and reduce or
perpetuate inequality. Both have come in for persistent criticism as culprits, however, especially from the
standpoint of neoclassical economists who regard almost any state intervention in the market as "distortionary"
and inefficient. (Section 3 of this Educational Package will deal with the positive side of public policy on
poverty).
“Bad” Economic Policies
Birdsall provides a litany of "bad economic policies" that help to perpetuate poverty in countries like those of
Latin America:
•Populist spending programs, ostensibly aimed at benefiting the poor but instead bringing inflation and
raising interest rates that hurt the poor as well as other groups
•Protectionism and price controls on food and other products most consumed by the poor
12
•Minimum wage laws and special worker entitlements that deter the creation of more job opportunities
•Special credit, foreign exchange, and regulatory privileges that favor a wealthy minority
•Underpricing of public services thus leading to their wasteful use.
From another perspective, critics of neoclassical or neoliberal reforms usually score the family of economic
policies (trade liberalization, privatization, deregulation, decentralization) usually prescribed by international aid
agencies like the IMF that usually amount to austerity measures against public spending, especially of funds
meant for social services and the poor. Policies and programs that increase foreign and public indebtedness,
remove subsidies for the poor (but lavish "incentives" on foreign and domestic investors), and impose draconian
measures on urban squatters and slum-dwellers without adequate alternatives and "safety nets" also help keep
the poor impoverished.
Some public policies discriminate against rural migrants to cities, like the hukou system which denied
basic rights to city land, housing, jobs and services to China's "floating population" of 150 million until
2001 (Mackenzie 2002: 305-306). Metropolitan development plans and regulations that resettle
informal settlers away from their jobs or otherwise encourage urban sprawl exacerbate the marginal
conditions of the urban poor by multiplying dislocation and "mismatch" problems.
Technology and Technocracy
Technology policies and programs are also figuring prominently in worsening or alleviating socio-economic
divides. Too much of the communication flows fueling globalization "tend to be a placeless preserve of large
corporations and big government." (Thrift 1990: 34). Transportation schemes that favor private cars against mass
transit keep the urban poor bound to the ground. Expensive technological solutions, however, become problems
themselves. This is a basic dilemma demonstrated by the "digital divide" which is proving less tractable than was
hoped for despite measures to address the ICT needs of the poor. (Cf. Prasch 2003; Elliesen 2003)
More generally, aggregative and indirect policies that give primacy to economic growth are also likely to
aggravate urban and rural poverty. Greater sensitivity to the spatial and social incidence of poverty, though, has
been shown by more recent development techniques and strategies that include poverty mapping and geographic
as well as social targeting. (Bigman & Fofack 2000. See Section 3 of this Package). Bureaucratic and political
incompetence and corruption, however, limit the efficacy of technocratic approaches and drain limited public
resources that could redound to the poor.
Technocratic insulation from mass-based politics may be one way out, but technoracy presents its own dark side
of likely indifference to the plight of the poor. (Deolalikar & Pernia 2003: 7; UNRISD 2004:1). These are
problems primarily of poor countries, while richer countries have faced different challenges. The U.S. has begun
to substitute "workfare” for welfare benefits due to adverse public opinion of beneficiaries, dwindling pension
funds, aging populations, and other problems of affluence. There have been different types of welfare states,
however, with some having more "niggardly" social policies than others. (See Myles & Quadagno 2002).
Problems of governance, especially in developing countries, have been attributed to lack of information,
transparency, and accountability in administrative and political institutions. The poor also do not have adequate
mechanisms and resources for effective representation, participation and empowerment vis a vis government
agencies. Good governance proposals include giving the poor greater "voice", the intercession of civil society
organizations, and opportunities for "exit," but such recourses have also proved problematic. (On the limits of
community participation, see Mansuri & Rao 2004). Normative changes, e.g. in the political culture of both the
ruled and the rulers, regarding the rights of the poor to have a much bigger "place at the table,” are also
prerequisite conditions for the poor to organize and mobilize. (Cf. Ward 2004: 186).
Concluding Note
Poverty is a multidimensional social problem that hinges on but goes well beyond not having enough income
and the means to meet basic needs. While poverty may seem simple and obvious in its symptoms, especially to
its victims, it has complex and deeper material and cultural causes. (Which does not mean that “root” causes
should be addressed to the neglect of the symptoms, which also deserve remedy of a direct and immediate kind).
Historical circumstances and developmental stages have determined which causes predominate in particular
places and
13
Theories of Poverty 17
periods. Some basic factors still animate contemporary public debates, for example, the Malthusian fear of
overpopulation versus the attribution of poverty to the lagging rate of economic growth or to maldistribution of
prosperity and power.
But the problem has grown both more complex and larger with the world. Urban poverty suggests that while
geography mattered more in much earlier epochs, institutions (political, economic, social) have a much wider
remit and role to play today, to mediate the various actors and forces involved and to cooperate in any serious
effort to combat poverty in an urbanizing world. Thus, international aid agencies have just discovered and
acknowledged what some disciplines have probably long known, that governments can play a positive as well as
negative role’ “that institutions and governance are key determinants of sustained growth and poverty
reduction…. the 1990s awakened interest in institutions and governance….” (Wolfenson & Bourguignon 2004:
5).
Yet space retains its place even in this fast–shrinking and urbanizing globe. As many have observed, it is risky to
generalize about different cities to suit one–size–fits–all policies, even if they may fall into certain definable
categories. In order to be sensitive to the nuances presented by space, development and poverty reduction
policies must be context–specific and spatially as well as socially targeted, whether the poor migrate between
rural and urban areas or remain trapped in both. Institutions must still grapple with the geography of poverty and
development.
14
References
Amin, Samir (Oct. 2003): "World Poverty, Pauperization, & Capital Accumulation," Monthly Review 55/5, pp.
1-9.
Baharoglu, Deniz and Christine Kessides (2002): “Urban Poverty,” A Sourcebook for Poverty Reduction
Stategies, ed. By Jeni Klugman. Washington D.C.: The World Bank, vol. 2, ch. 16, pp. 124-159.
Bigman, David, and Hippolyte Fofack (2000): "Geographical Targeting for Poverty Alleviation: An Introduction
to the Special Issue," The World Bank Economic Review 14/1, pp. 129-145.
Birdsall, Nancy (Summer 1998): "Life is unfair: Inequality in the world," Foreign Policy 111, pp. 75-80.
Browning, Christopher R., and Kathleen A. Cagney (Dec. 2003): "Moving Beyond Poverty: Neighborhood
Structure, Social Processes, and Health," Journal of Health and Social Behavior 44/4, pp. 552-571.
Buckley, Robert M. and Jerry Kalarickal (October 2004): “Shelter Strategies for the Urban Poor: Idiosyncratic
and Successful, but Hardly Mysterious,” World Bank Policy Research Working Paper 3427, p. 35.
Budd, Edward C., ed. (1967): Inequality and Poverty. New York: W.W. Norton & Co.
Clark, Charles M.A. (Jun. 2002): "Wealth and poverty: On the social creation of scarcity," Journal of Economic
Issues 36:2 , pp. 415-420.
Cotter, David A. (Dec. 2002): "Poor people in poor places: Local opportunity structures and household poverty,"
Rural Sociology, 67/4, pp. 534-555.
Deolalikar, Anil B., and Ernesto M. Pernia, eds. (2003): "Introduction," Poverty, Growth, and Institutions in
Developing Asia. Hampshire, UK: Palgrave, for the Asian Development Bank, pp. 1-9. (See also Alex B.
Brillantes, Jr. "Public Sector Reform and Poverty Reduction," Ch. 4 in this volume.)
Diamond, Jared (1999): Guns, Germs, and Steel: The Fates of Human Societies. New York: W.W. Norton &
Company.
Elliesen, Tillmann (Jul. 2003): "The real 'digital divide' is inside the poor countries," [Interview with Mike
Jensen], D & C Development and Cooperation, 30/7, pp. 302-303.
Foster, James E. (May 1998): "Absolute versus relative poverty," The American Economic Review 88/2, pp.
335-341.
Gallup, John Luke, and Jeffrey D. Sachs with Andrew D. Mellinger (1998): "Geography and Economic
Development," Annual World Bank Conference on Development Economics [ABCDE], ed. by Boris Pleskovic
& Joseph E. Stiglitz. Washington, D.C.: The World Bank, 1991, pp. 127-188 (including comments by J. Vernon
Henderson and Anthony J. Venables, and proceedings summary).
Goldsmith, Teddy (Jul./Aug. 2001): "Poverty--the child of progress," The Ecologist 31/6, pp. 44-46.
Gordon, David M. (1972): Theories of Poverty and Underemployment. Lexington, MA: D.C. Heath and
Company.
Harrington, Michael (1962): "The Invisible Poor," The Other America. The Macmillan Co., Reprinted in Budd,
ed. Inequality and Poverty, pp. 140-151.
Krugman, Paul (1998): "The Role of Geography in Development," ABCDE, pp. 89-123 (including comments by
Paul A. David and James E. Rauch, and proceedings summary).
Lampman, Robert J. "Approaches to the Reduction of Poverty: I," Inequality and Poverty, ed. by Budd, 174-182.
15
Lewis, Oscar (Jan/Feb 1998): “The Culture of Poverty,” Society 35/2, pp. 7-9.
Mackenzie, Peter W. (Fall 2002): "Strangers in the City: The Hukou and Urban Citizenship in China," Journal of
International Affairs 56/1, pp. 305-319.
Mansuri, Ghazala, and Vijayendra Rao (Feb. 2004): "Community-Based and -Driven Development: A Critical
Review," World Bank Policy Research Working Paper 3209, http://econ.worldbank.org, p. 77.
Massey, Douglas S. (Nov 1996): "The Age of Extremes: Concentrated Affluence and Poverty in the Twenty-First
Century," Demography 33/4, pp. 396-412. (See comments by Sheldon Danziger and others in this same issue.)
Mills, Edwin S., and Ernesto M. Pernia (1994): "Introduction and Overview," Urban Poverty in Asia: A Survey
of Critical Issues, ed, by Ernesto M. Pernia. Hong Kong: Oxford University Press, pp. 3-51.
Mokyr, Joel (Dec. 1998): "Secrets of Success," Reason 30/7, pp. 70-74.
Myles, John, and Jill Quadagno (March 2002): “Political Theories of the Welfare State,” The Social Service
review 76/1, pp. 34-57.
Myrdal, Gunnar (1971): Asian Drama: An Inquiry Into the Poverty of Nations. Abridged by Seth King. New
York: Pantheon Books.
Pernia, Ernesto M., ed. (1994): Urban Poverty in Asia: A Survey of Critical Issues. Hong Kong: Oxford Univ.
Press.
Piel, Gerard (Jan./Feb. 1997): "The Urbanization of Poverty Worldwide," Challenge 40/1, pp. 58-71.
Pieterse, Jan Nederveen (2004): Globalization & Culture, Global Melange. Lanham: Rowman & Littlefield
Publishers, Inc.
Polanyi, Karl (1944, 1957): The Great Transformation: The Political and Economic Origins of Our Time. Boston:
Beacon Press.
Prasch, Robert E. (Jun. 2003): "Technical change, competition, and the poor," Journal of Economic Issues 37/2.
Rao, Vijayendra, and Michael Walton (2004): "Culture and Public Action: An Introduction," Culture and Public
Action, ed. by Rao and Walton. Stanford Univ. Press, p. 66.
Ravallion, Martin (Aug. 2002): "On the urbanization of poverty," Journal of Development Economics 68/2, Abs.
Rein, Martin, and Christopher Winship (Jul./Aug. 1999): "The dangers of 'strong' causal reasoning in social
policy," Society 36/5, p. 20.
Rees, Teresa (Feb. 1998): "Social Exclusion and Equal Opportunities," International Planning Studies 3/1, pp.
15-34.
Rodrik, Dani, and Arvind Subramanian (Jun. 2003): "The Primacy of Institutions," Finance & Development 40/2,
pp. 31-34.
Ross, Eric B. (Dec. 1998): "Malthusianism, counterrevolution, and the Green Revolution," Organization &
Environment 11/4, pp. 446-450.
Sachs, Jeffrey D. (Jun. 2003): "Institutions Matter, but Not for Everything," Finance & Development 40/2, pp.
38-41.
16
Satterthwaite, David (Sept. 5, 2001): "Rural and Urban Poverty: Understanding the Differences," Economic
Perspectives [e-journal of the U.S. State Department, http://usinfo.state.gov/journals/journals/htm] pp. 1-6.
Sen, Amartya (1999): Development as Freedom. Oxford: Oxford Univ. Press.
Srinivasan, T.N. (May 1994): "Human Development: Means and Ends," American Economic Review 84/2, pp.
238-243.
Strassmann, W. Paul (Mar. 1996): "Limits to Market Empowerment for Housing in Developing Countries: The
Case of Land," Journal of Economic Issues 30/1, pp. 211-222.
Streeten, Paul (May 1994): "Human Development: Means and Ends," American Economic Review 84/2, pp.
232-237.
Streeten, Paul (Fall 1998): "Beyond the Six Veils: Conceptualizing and Measuring Poverty," Journal of
International Affairs, 52/1, pp. 1-31.
Therien, Jean-Philippe (Aug. 1999): "Beyond the North-South divide: The two tales of world poverty,” Third
World Quarterly 20/4, pp. 723-742.
Thrift, Nigel (Feb. 16, 1990): "Exploring the Space of Flows,” New Statesman 3/88, pp. 33-34. Review of
Manuel Castells, The Informational City, among others.
Todaro, Michael P. (2001): "Growth, Poverty, and Income Distribution,” Economic Development. 7th ed.
Addison Wesley Longman. Ch. 5, pp. 151-194.
U.N. Centre for Human Settlements (HABITAT) (2001): Cities in a Globalizing World. London: Earthscan
Publications Ltd.
UNDP [U.N. Development Programme] (1997): Human Development Report 1997. Washington, D.C.: Oxford
Univ. Press for UNDP.
UNRISD [U.N. Research Institute for Social Development] (Aug. 1, 2004): "Technocratic Policy Making and
Democratic Accountability," UNRISD Research and Policy Brief 3, p. 4.
Ward, Peter M. (2004): "Introduction," "From the Marginality of the 1960s to the 'New Poverty' of Today,” Latin
American Research Review 39/1, pp. 183-187.
Wolfensohn, James D. and Francois Bourguignon (October 2004): “Development and Poverty Reduction:
Looking Back, Looking Ahead,” paper prepared for the 2004 Annual Meetings of the World Bank and IMF, The
World Bank.
World Bank (2000): Attacking Poverty: World Development Report 2000/2001. Washington, D.C.: Oxford Univ.
Press, for the World Bank.
17
MODULE 1.2: MEASUREMENT AND POVERTY MAPPING
Measurement of Poverty
Issues in Measuring Poverty
Poverty measures are usually based on the income (or consumption) of the household in an area. But sometimes
other indicators of social service availability or social outcome are used singly or in a composite index. Some
prefer composite poverty measures because they convey the essence of poverty no as purely economic, but also
include social dimensions. Household income or expenditure does not capture aspects of welfare such as health,
literacy or access to public services. The availability of clean drinking water or public health centers, for
example, matter to an individual’s standard of living but is not reflected in a measure of consumption or income.
Households with access to free public services are better off than those without, even though their income or
expenditures may be the same.
Current Practice
Common practice starts by identifying a single monetary indicator of household welfare. Let the indicator value
for the ith household be denoted as i y . This indicator tends to be either total expenditure on consumption of
total income over some period. Next, a set of poverty lines, denoted as i z , is defined. These estimate the cost to
the household of the level of welfare needed to escape poverty, that is, it is agreed, at least, implicitly, that lower
values of / i i y z mean that a typical member of the household is absolutely poorer. Practice varies in terms of
the information used in setting the z’s. “Best practice” is to adjust for differences in the prices faced (over time or
space, in as much detail as data permit) and household demographics. Alternatively, one can introduce deflators
at the first stage of defining the household welfare indicator and have only one poverty line. Another method is
to set the z’s as a constant proportion of the mean for some subgroup to which i belongs.
Finally an aggregate poverty measure is identified, which summarize the information contained. In the measured
y’s and z’s. The most common measure is the headcount index, given by the proportion of the population for
whom / 1 i i y z . This measure, however, possesses undesirable properties, such as the fact that when a poor
person becomes poorer the headcount index of poverty will not increase. A large literature has proposed and
studied enumerable alternative measure, though as yet headcount index continues to be popular.
FGT Poverty Measures
The Foster-Greer-Thorbecke (FGT) family of poverty measures is highly regarded because it meets all the
axioms desirable in income-based poverty measures and contains a parameter, , that can be set according to the
society’s sensitivity to the income distribution among the poor. Let The FGT poverty measure is expressed as
1
1
( ; )
q
i
i
g
P y z
n z



_ _ _ _
_ _
_
where 1 2 , , , n y y y y _ is the vector of household incomes in increasing order, z is the predetermined
poverty line, i i g z y is the income shortfall of the ith household, ( ; ) q q y z is the number of poor
households (having income no greater than z ) and ( ) n n y is the total number of households.
When = 0, the FGT measure collapses to the headcount index, or the percentage of the population that
is below the poverty line. That is,
0
q
P H
n

This measure, while useful for general poverty comparisons, is insensitive to differences in the depth of poverty.
When = 1, the FGT measure gives the poverty gap, a measure of the average depth of poverty. 1 P is
18
expressed as
1
1
1 1 q
i i
i
g g
P
n z n qz 

_ _
When = 2, the FGT 2 P measure weights heavily income inequality among the poor. An additional peso that
reaches the poor will matter more than one reaching the only slightly poor. 2 P is expressed as
2
2
1
1 q
i
i
g
P
n z 
_ _ _ _
_ _
_
The use of FGT class of measures require the definition of a poverty line and is calculated on the basis of
disaggregated data (either household level, or aggregated for a few groups such as quintiles). The FGT measure
can be decomposed for population subgroups. If the population is divided into m collections of households
1, , j m _ with ordered income vector ( ) j y and population size j n , then
( )
1
( ; ) ( ; )
m
j j
j
n
P y z P y z
n 

_
Pis additively decomposable with population share weights. The decomposition allows a quantitative, as well
as qualitative, assessment of the effect of changes in subgroup poverty on total poverty. In fact, increased
poverty in a subgroup will increase total poverty at a rate given by the population share / j n n . That is, the
larger the population share, the greater the impact. The quantity ( ) ( / ) ( ; ) j
j j T n n P y z may be interpreted as
the total contribution of a subgroup to overall poverty while 100 / ( ; ) j T P y z is the percentage contribution of
subgroup j .
References
Foster, J., Greer, J. and Thorbecke, E. (1984): A Class of Decomposable Poverty Measures, Econometrica, 52(3),
pp. 761-766.
Ravallion, Martin (1996): Issues in Measuring and Modeling Poverty, World Bank Policy Research Working
Paper.
19
Geographical Targeting of Poverty Alleviation Programs
Practical Considerations in Designing a Poverty Alleviation Program
Geographical targeting of welfare programs is common in developing countries and is often used in conjunction
with additional targeting criteria to narrow the beneficiary population and thus reduce costs. The challenge for
policymakers is to use the available resources to provide the greatest possible assistance to those who need it
most. In the absence of reliable information on personal income, the first best solution of identifying the poor
and directing all benefits only to them is not feasible. Even in industrial countries that have the necessary data, it
is not possible to ascertain whether targeted programs do indeed reach all of the poor and do not leak to the
nonpoor.
In the past, since most developing countries did not have reliable information on individual income, many chose
programs with universal coverage. However, governments had to drastically reduce or even terminate these
programs due to growing budget constraints. Some countries replaced universal coverage with means testing.
However, in the absence of reliable information on household income, means testing led to massive leakage.
The absence of reliable information for identifying the poor, on the one hand, and the mounting constraints on
public resources, on the other, made targeting by means of indirect indicators the only viable alternative for most
developing countries. The indicators used to determine eligibility could include the household’s size, the number
of children in the household, the size of the household’s landholdings or other assets, and the region in which the
household was located.
Methods of Geographical Targeting
The optimum solution in welfare programs, from a theoretical point of view, is to identify the target population
and design the most effective program for this group. In most cases, however, it is not possible to identify the
target population since this requires information that is not observable and thus difficult to verify. In poverty
alleviation programs, the target population is the group of households with incomes below a certain minimum
level necessary to provide basic needs. Household income is often difficult to observe, however, and efforts to
assess its value and thus identify the target group may involve prohibitive costs. These costs consist not only of
direct administrative expenses for collecting the necessary information on income, but also of indirect costs due
to incentives that the program may give individuals either to modify their behavior or to falsify information on
their income in order to qualify for the program’s benefits. Poverty alleviation programs such as income transfers
or food subsidies to the poor, for example, may provide incentives to work less, cut earnings, or underreport
income in order to quality. Even in countries that have an accurate income reporting system, frequent means
testing is necessary to verify that only households that meet the criteria remain on the eligibility lists.
The difficulties and expenses involved in identifying eligible households leave two options: either to implement
universal programs that cover the entire population, or to use observable indicators that are highly correlated
with the relevant unobserved variables, such as income, in order to determine eligibility. Universal programs are
too expensive for most developing countries, and even many industrial countries find the rising welfare costs
daunting. The only viable option, therefore, is to use some form of targeting. This, however, requires a careful
choice of the targeting criteria, the observable indicators that will determine eligibility, and the programs that be
fit the specific conditions of the country.
Aside from geographic targeting, there are other targeting options available to developing countries. These
options fall into three categories, namely, self-targeted programs, programs targeted on the basis of household
characteristics and programs targeted on the household’s place of residence.
The Costs and Effectiveness of Targeted Programs
The effectiveness of a targeted program depends on the share of the target population – that is, the percentage of
the total poor population – that is actually covered by the program. This, in turn, depends on the accuracy with
which the observable indicators predict the unobserved variables that are the basis for determining eligibility. For
poverty alleviation programs, the desired indicators are those that are highly correlated with income. The
effectiveness of the set of indicators 1, , ( ) m _ depends on the probability expressed as
20
* Prob ( | for 1, , ) for all households i ik k y z k K _
where ik is the household’s level of the kth indicator (for example, the number of children) and *
k is the
critical value of the indicator that determines eligibility, i y is the unobservable income the ith household, and z
is the poverty-line income, below which the household is considered poor and therefore eligible for the program.
In practice, a data set containing complete information on both income and the indicators that are highly
correlated with income is used and the reduced form of the analysis is
1 1, , i q iq m m
q
P x _ _
where i P
is either the level of consumption expenditures or a dummy variable which indicates whether the
household is considered poor or not. We can distinguish between two types of indicators: 1 ( , , ) i iq x x _ are
household or personal indicators, and 1 ( , , ) m _ are dummies that identify the place of residence. When the
dependent variable is itself a dummy variable, the econometric analysis is transformed into a logit analysis, and
the equation then estimates the probability of a person being poor, assuming his personal indicators are within a
given range and his place of residence is in a given area.
The overall costs of the program have the following there components: (1) the direct (administrative) costs of
obtaining the information on these indicators; (2) the indirect costs dues to incentives that eligibility for the
program may provide; and (3) the costs of providing the benefits to the population covered by the program.
References
Baker, Judy and Margaret Grosh (1994): Poverty Reduction Through Geographic Targeting: How Well Does It
Work?, World Development, Vol. 22, No. 7, pp. 983-995.
Bigman, David and Uwe Deichmann (2000): Geographical Targeting: A Review of Different Methods and
Approaches, The World Bank Economic Review.
21
Background on Poverty Mapping
Higher-resolution maps are useful to decision-makers and researchers in part because they powerfully illustrate
the spatial heterogeneity of poverty. They are of special interest to environmental scientist and other researchers
working with spatial information on land cover change, ecosystem goods and services, infrastructure
development, and market integration, and similar topics with locational aspects.
There is as yet no standard methodology for producing high-resolution poverty maps. Various methods have
been used and refinements of techniques continue to be developed. Data needs differ depending on the analytical
methods chosen, and various methods have different implications for the timeframe and costs involved in
conducting the analysis. Moreover, some methods require a higher level of statistical and econometric expertise
than do others.
The choice of methods and data sources for poverty mapping should be determined according to the purpose for
which the resulting map will be used, which often dictates the appropriate level of precision and resolution. In
developing countries, it is also important to take into account the prevailing level of technical and human
capacity development.
Preparation of a poverty map may be driven by demand (e.g. need for information and analysis for program
design and/or implementation) or by supply (e.g. researcher interest in testing or refining a methodology). Ideally,
a poverty mapping exercise will emerge from and be shaped by the process of policy dialogue between map
producers and users. Through policy dialogue, map producers and users can work together to explore the specific
purposed of a proposed poverty mapping effort. Technical experts can help increase decision-makers’ awareness
of the potential users of poverty mapping as well as the inherent limitations of these techniques. Such
discussions can help illuminate important issues, not only with respect to choice of method and data source, but
also spark ideas concerning collaboration between various researchers and institutions, capacity development,
dissemination of resulting data products, and long-term sustainability of the mapping effort.
Generic Steps in Poverty Mapping
These eight generic steps involved in a poverty mapping effort highlight key decision points faced by researchers
and map producers. Not every poverty-mapping exercise will include all eight steps or follow them sequentially.
1. Define the purpose and expected use of mapping
In an ideal world, all poverty mapping would start here. Maps may be needed to show that certain
regions are disadvantaged, to rapidly assess options interventions, to target public investment to areas of
greatest need, or to investigate specific causes of poverty. The purpose and intended use of poverty maps
determine the scope and the required precision of the mapping exercise and should shape
methodological choices.
2. Select measure(s) of poverty and human well-being
Choosing an indicator or indicators of poverty is a pivotal step in map production. Poverty is a
multi-dimensional phenomenon, including economic, social, and other aspects of human well-being. The
selected indicator may be a monetary or non-monetary variable – for example, the proportion of
households below a certain income level or the proportion of households without access to sanitation.
Researchers sometimes distinguish between status and outcome variables – e.g. access to safe drinking
water (status) versus incidence of waterborne diseases (outcome) – but because indicators of poverty are
interdependent, the distinction between status and outcome measures is not always clear. A poverty
indicator may measure a single important dimension of human well-being, such as household
expenditure compared to a minimum necessary level or poverty line. Alternatively, the indicator may be
multidimensional, for instance, a composite index that depicts in basic deficits in basic human needs,
such as education, health care, and sanitation. Each type of poverty indicator has its own strengths and
weaknesses, and the choice of indicator will certainly influence who is classified as “poor”.
22
3. Select input data
Data used to construct a poverty map typically are drawn from population or agricultural censuses,
household surveys, or spatial (GIS) databases in which values are fixed to specific locations on a grid.
Increasingly, poverty mapping relies on data from many sources. Data used in poverty mapping may
vary in coverage, collection method, and level of resolution, all of which may have methodological
implications.
4. Select method of estimating or calculating poverty indicator
Researchers may choose to estimate a single variable, such as per capita household expenditures
compared to a specific standard of living (i.e. poverty line). Alternatively, they could use a composite
index, which may be calculated by simple regression (i.e. equal weighting) of a few variables or by
multivariate analysis, such as principal components or factor analysis.
5. Select a method to calculate, estimate, or display poverty indicator for geographic area
Depending on the chosen poverty indicator, input data and method of estimation/ calculation, researchers
will have different options for calculating or estimating the poverty indicator across a geographic area.
For instance, if map producers are using census-level data made available at the household level, then
simple aggregation of the data of the selected geographic unit may suffice. However, researchers often
need techniques that are more sophisticated. Poverty maps often combine census data (featuring
complete country coverage) with household survey data (encompassing a representative sample of the
selected population). This is accomplished by means of advanced statistical methods based on
econometric techniques, sometimes referred to as small area estimation. Combining data from these two
sources enables a poverty mapping study to benefit from both the complete spatial coverage of the
census and from a relevant poverty indicator in the household survey. Such statistical techniques help
overcome the survey’s insufficient sample size, which could not be aggregated to small administrative
units, and the census’ lack of an appropriate poverty measure.
6. Decide on the number of units for final map (resolution) to present poverty data
For many poverty-mapping methods, this step is often combined with the previous one. In the case of
small area estimation relying on household-unit data, researchers cannot map an individual household;
they must aggregate household-level data to larger units to reduce the statistical error in their prediction
model. Sensitivity tests conducted by researchers suggest that a minimum of 5,000 households is needed
to reduce statistical errors to an acceptable level. The number of households required may be
significantly higher in other cases, especially if the statistical model is not as strong in its predictive
power.
7. Produce and distribute maps
Mapping software is used to produce a spatial representation of the geographic distribution of calculated/
estimated poverty indicators. Maps and supporting analyses are distributed to the targeted
decision-makers. Increasingly, map producers are supplementing hardcopy maps with other products,
such as interactive decision-support tools and/ or datasets on compact discs, aimed s various audiences
(technical, general. or mixed).
8. Monitor usage and feedback
Poverty maps used for various purposes, ranging from identifying and understanding the causes of
poverty, to assisting in program development and policy formulation, to guiding allocation of
anti-poverty investments and expenditures. Map producers should monitor and evaluate the various
ways in which their maps are being used by decision-makers and/or researchers, and users should
23
provide feedback on the impact and limitations of poverty maps.
Poverty mapping and small area estimation
Poverty maps based on small area estimation method rely on sophisticated econometric techniques and a set of
identical variable (e.g. household characteristics and educational background) in both a census and a surveyed
representative sample of the overall population. By combining census and household survey data, researchers
benefit from the strengths of each instrument – a census’ complete coverage of a country and a survey’s more
detailed information. The survey provides the specific poverty indicator and the parameters, based on regression
models, to predict the poverty measure for the census.
Typically, the poverty indicator is an expenditure-based indicator of welfare, such as the proportion of household
that falls below a certain expenditure level (i.e. poverty line). In recent years, researchers have relied in two
principal methods for their small-area based poverty maps. The first requires access to detailed
household-unit-level data from a census. If such household-unit data are unavailable, unreliable, or incomplete –
as is frequently the case in many developing countries – researchers have applied average values for a given
indicator at the community level.
Small area estimation-based household-level survey data generally are more accurate and reliable than those
based on community-based averages. Indeed, the small area estimation technique using household-unit data is
the only poverty mapping method that generates an estimate of statistical error. However, the technical and data
requirements of this technique are relatively rigorous, and the approach works best in countries with regular and
comprehensive national censuses and household survey. Community-level averages are more readily available,
but using the small area estimation technique with such data generates an uncertain error, and the datasets used
may not provide a good proxy for the poverty indicator that the researcher seeks to measure.
Other poverty mapping methods
Although the “newest” type of poverty maps are based on small area estimation techniques, other methods have
a longer history of application and important lessons have been learned in the course of their use. Many such
methods feature the use of composite indexes, including the Human Development Index (HDI) originated by the
United Nations Development Programme (UNDP), as well as various basic needs measures. The latter,
sometimes referred to as “unsatisfied basic needs” indexes, have been used primarily in Latin America.
One advantage of composite indexes is that they are intuitive and easy for general audience to understand.
Moreover, this approach requires less advanced statistical expertise than small area estimation. Composite
indicators are stronger on the social dimensions of poverty and, on first impression, they appear to better capture
the multidimensional nature of human well-being. The most serious criticism of composite indexes is that their
weighting of variables can be arbitrary and theoretically unsound. Even a small change in the weighting scheme
could easily lead to a change in the proportion of households classified as poor and overturn the ranking of
geographic areas identified as poor.
Caveats
Although poverty mapping can be a powerful tool for analyzing poverty and communicating the results to
technical and non-technical audiences, experts hasten to point out the limitations of these techniques. Poverty
maps are not a panacea for understanding or solving poverty problems; they are only one tool among many for
investigating the complex phenomenon of poverty. They should be used in conjunction with other information
and analysis that provide context and ground-truthing within communities.
Poverty maps can be used to explore the spatial aspects of various components of human poverty. However,
indirect estimation of poverty, as opposed to direct observation in the field, introduces some degree of
uncertainty. Careful additional analyses are needed before conclusions are drawn on any meaningful correlation,
much less causal relationships between these variables. In addition, it is important that poverty mapping is
always seen in the overall context of a country’s decision-making processes. Technical tools like poverty maps
run the risk of being abandoned once initial donor support or funding has waned. To ensure a path of sustained
use and support for poverty map, fundamental questions needs to be addressed, such as how to retain skilled
analysts in the public sector, overcome limited or lacking demand and funding from policymakers, and convince
decision-makers that continued investment in poverty maps is worthwhile in an environment that does not follow
24
a purely technical approach to decision-making.
References
Background on Poverty Mapping, Experiences with the Development and Use of Poverty Maps
25
Small Area Estimation Technique
Small area estimation has received a lot of attention in recent years due to growing demand for reliable small
area estimators. Traditional area-specific direct estimators do not provide adequate precision because sample
sizes in small areas are seldom large enough.
Sample surveys are used to provide estimates not only for the total population but also for a variety of
subpopulations (domains). “Direct” estimators, based only on the domain-specific sample data, are typically
used to estimate parameters for large domains. But sample sizes in small domains, particularly small geographic
areas, are rarely large enough to provide direct estimates for specific small domains.
For example, the Family Income and Expenditure Survey (FIES) of the National Statistical Office is designed to
provide direct estimates with acceptable precision at the provincial level. But, to have large enough sample to
support reliable direct estimates, for say, all cities and municipalities, and for all subareas, particularly barangays,
is never possible in practice. In such an example, cities/ municipalities and barangays may be regarded as “small
areas: because the area-specific sample sizes are small (or even zero). In making estimates for such small areas,
it is necessary to “borrow strength” from related areas to form “indirect” estimators that increase the effective
sample size and thus increase the precision. Such indirect estimators area based on either implicit or explicit
models that provide a link to related small areas through supplementary data such as recent census counts and
current administrative records. Indirect estimators based on implicit models include synthetic and composite
estimators, while those based on explicit models incorporating area-specific effects include empirical Bayes
(EB), empirical best linear unbiased prediction (EBLUP) and hierarchical Bayes (HB) estimators.
Small Area Models
It is now generally accepted that when indirect estimators are to be used they should be based on explicit model
that related the small areas of interest through supplementary data such as last census data and current
administrative data. An advantage of the model approach is that it permits validation of models from the sample
data. Small area models may be broadly classified into two type: area level and unit level.
Area Level Models
Area-specific auxiliary data, i x , are assumed to be available for the sampled areas ( 1, , ) i m _ as well as the
non-sampled areas. A basic area level model assumes that the population small area mean i Y or some suitable
function ( ) i i g Y , such as log( ) i i Y , is related to i x through a linear model with random area effects
i :
, 1, , i i i i m x _ _
where _ is the p-vector regression parameters and the ’s are uncorrelated with mean zero and variance 2
.
Normality of the i is also often assumed. The above models also holds for non-sampled areas. It is also
possible to partition the areas into groups and assume separate models of the same form across groups.
We assume that direct estimators ˆ
i Y of i Y are available whenever the area sample size 1 i n . It is also
customary to assume that
ˆ
i i i e 
where ˆ ˆ ( ) i i g Y and the sampling errors i e are independent (0, ) i N with known i . Combining this
sampling model with the “linking” model, we get the well-known area level mixed model of Fay and Herriot
ˆ
i i i i e x _
Note that under the above formulation, both design-based random variables i e and model-based random
variable i are involved. In practice, sampling variances i are seldom known, but smoothing of estimated
26
variances ˆi is often done to get stable estimates i which are then treated as the true i . An advantage of
the area-level model is that the survey weights are accounted for through the direct estimators ˆ
i .
Unit Level Models
A basic unit level population model assumes that the unit y-values ij y , associated with the units j in the areas
i , are related to the auxiliary variables ij x through a one-way nested error regression model
, 1, , , 1, , ij ij i ij i y e j N i m x _ _ _
where 2 (0, ) i IID N _ are independent of 2 (0, ) ij e e IID N _ and i N is the number of population units
in the i-th area. The parameters of interest are the total i Y or the means i Y . The above model is appropriate
for continuous variables y. To handle count or categorical (e.g. binary) y variables, generalized linear mixed
models with random small area effects, i , are often used.
References
Battese, G.E., R.M. Harter and W.A. Fuller (1988): An Error Component Model for Prediction of Country Crop
Areas Using Survey and Satellite Data, Journal of the American Statistical Association, Vol. 83, pp. 28-36.
Fay, R.E. and R.A. Herriot (1979): Estimates of Income for Small Place: An Application of James-Stein
Procedures to Census Data, Journal of the American Statistical Association, Vol. 74, pp. 269-277.
Rao, J.N.K. (1999): Some Recent Advances in Model-Based Small Area Estimation, Survey Methodology, Vo.
25, No. 2, pp. 175-186.
27
Spatial Microsimulation Approach
Background
Microsimulation models in economic applications may be divided into static, dynamic and longitudinal types.
All these model work with cross-section data. A static microsimulation model treats a fixed number of
microunits and the attributes of this same set of microunits are reweighted to account for changes in
demographic structure over time. A dynamic microsimulation model, on the other hand, ages each microunit
individually by an empirical survival probability. The main difference between a dynamic microsimulation and a
static one is the aging procedure. A static aging procedure is relatively well-suited short- and medium-range
forecasts, provided it can be assumed that the characteristics of the population under examination do not change
rapidly. Athird type of microsimulation model found in economic literature are dynamic longitudinal models
which create synthetic microunits and forecasts a microunit’s whole lifecycle from birth to death. Thus, a
dynamic microsimulation model does not forecast the characteristics of real sample units but the assigned
characteristics of synthetic microunits. Presently, a major dynamic microsimulation model in the U.S. is
CORSIM developed at Cornell University.
Spatial microsimulation and regional analysis
Clarke and Holm (1987) provides a through presentation on how microsimulation methods can be applied in
regional science and planning analysis. Clarke (1996) points out that there are two major works involved in
applying microsimulation methods in spatial analysis. The first involves the creation of a microdata set using
conditional probabilities and contingency tables. A method called iterative proportional fitting is also used to
create probabilities using data sets which have different spatial scale. The next step involves the creation of a
sample of individuals or households based on the set of probabilities. Moreover, Clarke (1996) identifies three
factors why application of microsimulation method to regional science have not received wide attention in the
past. The first pertains to the immense computational problems and costs required in actual simulation work. The
computational problems are also substantially increased when the spatial dimension is considered. Presently,
however, computer technologies in both hardware and software offers considerable opportunities and flexibility
for large-scale simulations. Another stumbling block has been the lack of microdata sets to calibrate or test
results of simulations. This is certainly changing as more and better data sets are becoming available. A third
reason relates to the lack of sustained efforts by those involved in spatial microsimulation.
Spatial microsimulation process
Clarke (1996) illustrates how microsimulation can be employed for the creation of a micro-level population with
the population characteristics: age, sex, marital status and household tenure as shown in the figure below.
Supposing that age, sex, and marital status of the household head is available from the census, it is then possible
to estimate probabilities of household tenure. The first synthetic household has the following characteristics:
male household head, aged 27, married. The estimated probability that a household of this type would be
owner-occupied is 70. The next step in the procedure is to generate a random number to see if the synthetic
household gets allocated to the owner-occupier category. The random number in this example is 0.542 which
falls within the 0.001 to 0.700 range needed to quality as owner-occupied. The same procedure is then carried
out sequentially for the tenure allocation of all synthetic households. It should be noted that that difficult task in
microsimulation is to specify which variables are independent upon others and to determine the ordering of
probabilities.
28
Steps 1st 2nd Last
Head of household (hh)
1. Age, sex, and
marital status (M)
of hh
2. Probability of hh
of give age, sex,
and M being an
owner-occupier
3. Random number
(computer generated)
4. Tenure assigned
to hh on basis of
random sampling
5. Next hh (keep
repeating until a
tenure type has
been allocated to
every hh)
Age: 27
Sex: male
M: married
0.7
0.542
owner-occupied
Age: 32
Sex: male
M: married
0.7
0.823
rented
Age: 87
Sex: female
M: divorced
0.54
0.794
rented
Example of spatial microsimulation process (Clarke, 1996)
29
It is noted that a large number of microsimulation models in existence are inherently aspatial. This means that
existing microsimulation models does not incorporate sufficient geographic detail so as to allow richer analysis
at fine spatial levels. Microsimulation models would potentially find relevant applications to policy simulations
at neighborhood levels, and even voting and school districts.
At present, there are two prominent large-scale spatial microsimulation models in existence, namely, SimLeeds
and the SVERIGE Spatial Microsimulation Model. SimLeeds is spatial microsimulation model of Leeds
developed at the University of Leeds. SVERIGE or System for Visualizing Economic and Regional Influences in
Governing the Environment is a spatial dynamic microsimulation model constructed by the Spatial Modelling
Centre in Kiruna, Sweden. It is the first national level spatial model capable of analyzing the spatial
consequences of public policies. The treatment of space is achieved by incorporating regional attributes in
modelling the various socioeconomic modules of the model and constructing from scratch a model for modeling
internal migration in Sweden.
Spatial microsimulation of informal households in Metro Manila
Tiglao (2002) embodies the first application of spatial microsimulation approach to for informal households in
Metro Manila and elsewhere. At present, InformalSim covers the City of Manila only. However, the model can
be easily extended to cover other cities and municipalities in Metro Manila. Manila consists of 54 traffic analysis
zones, 900 barangays and around 1.65 million persons in 1990.
The figure below shows the spatial microsimulation process of InformalSim. The object of the microsimulation
is to estimate characteristics of households in the microdata that would allow identification of informal
households. First, a baseline population consisting of all households in the 1990 CPH is initialized. Then, the
economic activity of household head is estimated using conditional probabilities from the 1996 MMUTIS data.
Assignments of economic activity are done using Monte Carlo sampling based on the characteristics of the
household head, namely, sex, age, and location. Next, occupation and employment sector probabilities are
computed and assignments are done using Monte Carlo sampling. The occupation and employment sector
probabilities are estimated using multinomial logit models which are calibrated using the 1997 FIES data set.
The next stage involves the estimation of household incomes based on the characteristics of the household head.
To achieve this, the employment status of the household head is first determined. The employment status of the
household is estimated using a probit model, that is, a binary choice model of being a wage earner (i.e. formal sector)
or self-employed (i.e. informal sector). Then, conditional on employment status, the household income is computed
using a regression model with correction for selectivity in the lines of Lee (1978).
The next step involves the estimation of permanent income of the household. Permanent income is needed to estimate
the value. Housing values is estimated using in two steps. First, housing tenure choice is estimated using a probit
model of whether the household is under formal or informal housing. Formal housing consists of owner-occupiers and
renters. On the other hand, informal housing are attributed to households who own the house but rents (with or
without consent of owner) the land. Then, housing value is computed using a regression model conditional on the
tenure status with the appropriate correction for selectivity bias in the lines of Lee and Trost (1978). Simulated values
can then be visualized using GIS and the output can be analyzed in a ‘complete-data’ setting.
The next figure presents the object representation of household microdata. The object-oriented approach to
spatial microsimulation modeling is proposed by Ballas et al. (1999). Object-oriented programming offers a very
flexible platform for estimation and handling of very large data sets. InformalSim is implemented in Java.
30
Initialize base households
using 1990 CPH data
(age, sex, marital status,
and education of household
head, household size)
Assign occupation of
household head based
on Monte Carlo sampling
Assign employment sector
of household head based
on Monte Carlo sampling
Compute occupation
probabilities from
Occupation Choice Model
Compute employment
probabilities from
Employment Choice Model
Estimate household income
based on characteristics
of household head
Compute employment
status probabilities and
assign employment status
by Monte Carlo sampling
Compute bias-adjusted
household income function
based on employment status
Compute economic activity
rate of household head
Estimate permanent
Income of household
Compute housing tenure
status probabilities and
assign housing tenure status
by Monte Carlo sampling
Compute bias-adjusted
housing value function
based on tenure status
Estimate housing tenure
and housing value
31
Spatial microsimulation of informal households
MEMBER
Variables
Province-ID
District-ID
Barangay-ID
Household-ID
Member-ID
Relation to hh head
Age
Sex
Marital status
Education
(Occupation)
(Employment sector)
(Income)
Methods
GetEconomictActivity
GetOccupation
GetEmploymentSector
GetIncome
...
HOUSEHOLD
Variables
Province-ID
District-ID
Barangay-ID
Household-ID
Household size
Age of hh head
Sex of hh head
Marital status of hh head
Education of hh head
(Economic activity of hh head)
(Occupation of hh head)
(Employment sector of hh head)
(Employment status of hh head)
Members [Vector]
Building type
Roof type
Wall type
State of repair
Year built
(Household income)
(Housing status)
(Housing value)
Methods
GetEconomicActivityofHead
GetOccupationofHead
GetEmploymentSectorofHead
GetEmploymentStatusofHead
GetHouseholdIncome
GetHousingStatus
GetHousingValue
...
Baseline
Characteristics
Unobserved
Characteristics
Computational
Objects/ Models
Object representation of household microdata
There are two major objects in InformalSim, namely, the member object and the household object. These two
objects contain variables and methods. Variables correspond to the actual characteristics of the respective objects.
Variables are of two types-baseline (i.e. observed) and unobserved. Methods contain computational codes or
models that operate on the variables. Each household object contains a vector (or collection) of member objects
as would be true in the physical sense. This representation is completely convenient as the characteristics of the
household are entirely dependent on the members that comprise it. Moreover, the approach allows limitless
flexibility as future implementations may be conveniently incorporated into the structure.
The current implementation of InformalSim consists of 10 modules. The modules are executed sequentially as
outlined in the microsimulation process shown in Figure 3. The modules are:
1) Economic activity
2) Occupational choice
3) Employment sector choice
4) Employment status
5) Household income
6) Permanent income
7) Housing tenure status
8) Housing value
9) Inequality measures
32
10) Mapping and visualization
33
Mean household incomes Informal Employment(% of households)
34
Informal housing tenure Mean housing values (% of households)
Gini coefficient
The above series of map outputs depicts the results of the spatial microsimulation exercise. If spatial
microsimulation approach is to be effectively applied to policy simulation then calibration and validation is vital
in the modeling process. Tiglao (2002) proposes a process for validating outputs of spatial microsimulation. The
process has been implemented. The validation involves a rather sophisticated small area estimation model of
mean household incomes using spatial covariates. The small area model provides reliable estimates of mean
household income that can be used as benchmark values. The small area estimation utilizes methods using
generalized linear mixed model (GLMM) in unbiased estimates of parameters of interest using covariates in
either univariate or multivariate setting. Estimation is performed under various approaches including Best Linear
Unbiased Prediction (BLUP) or Bayesian methods (e.g. Empirical and Hierarchical Bayes). Tiglao (2002)
pursues Empirical BLUP in the lines of Battese et al. (1988) in deriving estimates of mean household incomes at
35
the traffic zone level using car-ownership and dwelling unit sizes as covariates in explaining incomes.
References
Ballas, D. and G. Clarke (2000): GIS and Microsimulation for Local Labour Market Analysis, Computers,
Environment and Urban Systems, 24, pp. 305-330.
Ballas, D., G. Clarke and I. Turton (July 1999): Exploring Microsimulation Methodologies for the Estimation of
Household Attributes, Paper presented during the 4th International Conference on GeoComputation,
Fredericksburg, Virginia, pp. 25-28.
Battese, G. E., R. M. Harter and W. A. Fuller (1988): An Error-Components Model for Prediction of County
Crop Areas Using Survey and Satellite Data, Journal of the American Statistical Association, 83(401), pp. 28-36.
Clarke, G. (1996): Microsimulation for Urban and Regional Policy Analysis, Pion Ltd., London.
Clarke, G. and E. Holm (1987): Microsimulation Methods in Spatial Analysis and Planning, Geografiska
Annaler B, 69(2), pp. 145-164.
Lee, L. F. and R. P. Trost (1978): Estimation of Some Limited Dependent Variable Models with Application to
Housing Demand, Journal of Econometrics, 8, pp. 357-383.
Tiglao, N. C. (2002): Small Area Estimation and Spatial Microsimulation of Household Characteristics in
Developing Counties with Focus on Informal Settlements in Metro Manila, Unpublished Ph.D. Dissertation,
University of Tokyo.
36
MODULE 1.3: PUBLIC ACTION FOR URBAN POVERTY
REDUCTION: POLICY OPTIONS AND ANTIONA AND
NATIONAL-LOCAL GOEVERNEMNT ROLES
Urban Poverty Reduction: Vulnerability and Asset Ownership,
Implications for Public Action
Objectives of the Module:
This module seeks to provide the students the opportunity to: a) learn how public action meant to address urban
poverty must stem from a full appreciation of the complex nature of urban poverty, (i.e., interrelated concepts of
cumulative deprivations, vulnerability and asset), and b) examine how the passage of the Urban Development
and Housing Act of 1992 represents one such form of public action and c) analyze this specific policy in terms of
the roles of the national and local government units in implementing this law.
Required Readings
There are two required readings for Lesson I. The first material is Arsenio Balisacan’s account of the nature,
causes and policy measures taken to address urban poverty in the Philippines. This was published in 1994. The
second material draws from the work of Baharoglu and Kessides, which was published as part of a two volume
Sourcebook for Poverty Reduction Strategies edited by Jeni Klugman. The World Bank published this in 1992.
Arsenio M. Balisacan, “Urban Poverty in the Philippines: Nature, Causes and Policy Measures,” in Asian
Development Review Vol. 12 No.1 (1994); an expanded version of this work was published in Arsenio Balisacan.
Poverty, Urbanization and Development Policy: A Philippine Perspective (University of the Philippines Press).
1994.
Balisacan’s work provides an appropriate Philippine context within which one can better appreciate the
Baharoglu and Kessides chapter. It describes and analyzes urban poverty in the Philippines using data from the
early sixties up to the early nineties. This is done through a characterization of urban poverty as manifested in
the profiles of urban households, looking at the levels of employment and unemployment of household heads
and how poverty affects the household members, particularly the children. It also reviews employment and
earning differences of urban poor household, access to basic needs and services, health and nutrition and housing
and the environment.
This study traced the roots of urban poverty in the Philippines to three principal causes:
i) industrialization policies that unduly encouraged the concentration of infrastructure and social
services in major urban centers, particularly Metro Manila;
ii) trade and macroeconomic policies that severely penalized agriculture, labor-intensive exports
and small scale and medium scale manufacturing establishments; and
iii) public spending policies that accorded little attention to human capital formation for the poor”
Deniz Baharoglu and Christine Kissides, “Urban Poverty,” Jeni Klugman, ed. A Sourcebook for Poverty
Reduction Strategies Washington D.C.: The World Bank, 2002)
It is crucial for any one who seeks to develop strategies aimed at minimizing urban poverty to grasp the full
implications of three major foundations of this chapter. The first one is to list the five dimensions of poverty: a)
income/consumption, b) health, c) education, d) security and e) empowerment. More than this listing, it is
equally important to know that when one talks of urban poverty as characterized by “cumulative deprivations,”
this means that any of these five dimensions is “often the cause of or contributor to another dimension.” The
chart that depicts this nature of urban poverty is presented in Slide 18 of Module 1.
37
The second element of the nature of urban poverty which is stressed in this chapter is the how poverty is linked
to vulnerability. Vulnerability here is the risk or the probability of one falling into urban poverty. This
vulnerability is related to three characteristics of urban life: “1) commoditization ( reliance on the cash economy),
2) environmental hazard( stemming from the density and hazardous location of settlements and from exposure to
multiple pollutants) and social fragmentation ( lack of community and of inter-household mechanisms for social
security, compared to those in rural areas).”
The third concept is that of asset ownership. Baharoglu and Kessides describe this relationship as follows:
“Vulnerability is closely linked to asset ownership. The more assets people have, the less vulnerable they are; the
fewer assets held by households, the greater their insecurity. The types of assets fall under the headings of labor,
human capital- health, education, and skills; productive assets- often, the most important of these is housing;
household relations; and social capital.” (p. 124)
Slides 5, 6 and 7 of this lesson present how these three elements of urban poverty are linked. These slides,
however, leave out one column of Table 16.2, i.e. “policy areas to strengthen assets.” It should be pointed out
that the security dimension mentioned earlier is now presented in three forms: tenure security for housing,
financial security and personal security. Empowerment is made more concrete through its manifestation in
“social and political exclusion” that eventually leads to disempowerment. Each of the columns in these three
slides presents how the dimensions of urban poverty are linked to vulnerability and asset ownership. (p. 128).
Slide 8 lists the benefits that may result in the adoption of public action to address urban poverty. First,
reducing social and economic inequalities in urban areas can decrease inequality and also reduce social tensions
that can lead to social and political clashes. Second, it can prevent large-scale health and environmental
problems. Slum areas which do not have access to basic health services may experience health epidemics which
can be a health threat to the whole city. Inadequate water supply, sanitation, solid waste disposal and lack of
water drainage can also affect the over-all water supply level for cities. Third, slum areas are also most
vulnerable to disasters such as floods, earthquakes and industrial accidents. Addressing the problems of the
urban poor can minimize the adverse effects of environmental hazards and disasters. Fourth, improvements in
the living conditions of the urban poor can also support local economic development. As members of the city’s
labor poor, households residing in slum areas can increase their labor productivity if they have access to basic
social services. Programs that address these problems not only increase labor productivity, it also increases the
capacity of these households to register demand for consumer goods and services, which may boost economic
activities in the area. Finally, productive labor and health households will not only stimulate the local economy,
it can also trigger more national economic activities.
What policy options may strategists consider when they want to reduce urban poverty? Proceeding from the
different dimensions of poverty, Baharoglu and Kessides see the need for “policy and institutional reforms at the
national as well as at the city level in order to improve the conditions facing the poor. Programs that can directly
benefit the poor in the short to medium term may be scaled up. Policy and institutional reforms can promote the
longer term scope for poverty reduction by fostering the broad based economic growth of cities through the
development of efficient and well integrated markets for labor, land and housing, and finance, and through
effective public finance and responsive governance.” (p.134) Slide No 9 lists these five major categories of
policy interventions that are discussed in detail in the chapter.
Under the rubric of labor markets and employment, the chapter elaborates how policy interventions can cover
the following: a) support to small enterprises and micro-enterprises (including street vendors); b) increasing
access to job opportunities; c) supporting residual subsistence (urban agriculture); d) supporting home-based
income-generating activities and e) safety nets and social insurance.
On land, housing and urban services, the chapter sees the “vulnerability of the urban poor as exacerbated by the
inadequate provision of basic public services, as well as by the policy and regulatory frameworks that govern
land and housing supply and property rights.” ( p.139) It thus provide details on policy reforms in the areas of
38
a) tenure security and property rights, b) land and infrastructure development regulations; and c) planning
procedures, building codes and construction permits.
On financial markets, the chapter stresses how “lack of access to credit increases the vulnerability of the urban
poor by constraining their ability to improve their homes, which in some cases is also their workplace and the
venue of any new businesses. Credit underwriting is a major problem since the poor do not have property to use
as collateral and often lack regular incomes.” ( p. 142). Moreover, the poor can hardly accumulate their savings
and they usually have little access to formal savings program. Policy interventions to open access to financial
services to the poor may include: a) encouraging financial organizations to lend to micro-enterprises; b)
supporting local NGOs and banks in making credit available by providing seed funding and/or guarantees and c)
monitoring and regulating the performance of financial intermediaries who collect from the general public (p.
142).
Public finance, as distinguished from financial markets, cover efforts on: a) cost recovery, tariffs and subsidies,
and b) decentralization and intergovernmental relations. It categorically states that: “Poverty-oriented programs
should not aim to achieve full cost recovery from beneficiaries – that would defeat the underlying redistributive
objectives. Programs designed to address urban poverty must be clear about its whether its target beneficiary is
the community or individual households. Public or communal services such as storm drainage, urban roads and
footpaths are financed by general taxes. Users can pay for private goods and services such as land title and
electricity connection. On decentralization and intergovernmental relations, the chapter discusses the importance
of clear responsibilities and consistent within intergovernmental fiscal relations on service provision. Local
governments must be financially stable to ensure that the services they provide for the poor in their areas are not
inadequately funded.
Finally on urban governance and capacity building, the chapter describes and discussed the value of policy
actions for accountability and responsiveness to the public, anti-corruption policies and practices and the need to
develop local capacity to implement programs at the local level.
Slide No. 10 depicts the proposed national-local feedback processes for urban poverty reduction strategies. It
shows the need for all the sectors, i.e., private sector, government and civil society, to work together on these
strategies.
Baharaglo and Kessides explain the relationship, thus: “Urban poverty reduction strategies need to be grounded
and implemented at the local level. The scaling up of such strategies, however, requires both the national and
local government to act simultaneously, eliminating impediments at both levels. While the central government is
working on a nationwide scale, addressing policy matters and regulatory impediments and initiating new
programs, the local authorities should be designing strategies to make appropriate intervention and regulatory
changes in the city. .Local experiences should be fed back to the national government to influence its support of
cities and to assist the redesign of national programs.” (p.149).
Slides 11-13 present urban poverty reduction strategies that national government can take at four levels: policy
action, program innovations, regulatory framework and monitoring and coordination. Slides 14 to 16, on the
other hand enumerate what the local governments can do at five levels: policy issues, program innovations,
regulatory framework, monitoring and evaluation and financial issues.
Illustrative of how poverty reduction strategies can be structured, these levels of intervention will be used when
we analyze the adoption and implementation of the country’s Urban Development and Housing Act of 1992, an
example of a poverty reduction policy and the subject of the next lesson for this module.
39
Republic Act No. 7279: Formulation and Implementation of the Urban
Development and Housing Act of 1992
Objectives of the Session:
By the end of this session, the students should be able to a) explain the socio-politico context within which RA
7972 was formulated, b) the history of the formulation of the law, c) the content of the law and the its
implementing rules and regulations, and d) issues and concerns that were raised in the implementation of RA
7972, particularly the roles and responsibilities of the national and local governments.
The student’s familiarity with the law should be shown through the capacity to analyze the content of the law in
terms of: a) how it seeks to minimize the vulnerability of the urban poor and b) how it enhances the capacity of
the urban poor to increase their assets, c) how it delineates the roles and responsibilities of the national and local
government units in implementing the law.
The Socio–Politico Context after the end of Martial Law in 1986
The fall of Marcos through the 1986 people power revolution led to unprecedented policy initiatives through
which the country sought to institutionalize the engagement of non-government and people’s organization in all
levels of decision-making.
Slides 1 to 3 present pertinent provisions of the 1986 constitution and the Local Government Code of 1991.
Prompted by the key role that civil society organizations played in ousting former president Marcos through the
historic 1986 people power revolution, the 1986 Constitutional Convention saw the value of including separate
provisions for non-governmental and independent people’s organizations, as enunciated in Articles II and XIII
in slide 1.
The democratization process also led to the devolution of power from the national government to local
government units. While crucial decisions were then made in what was referred to as “imperial” Manila, the
passage of the Local Government Code (RA 7160) in 1991 drastically reduced this power when it devolved
substantial powers to local government units which became more autonomous.
Porio hailed the Code as having “provided for people’s participation in local governance, devolved central
powers and responsibilities to LGUs and increase their share of state revenues. Previously people’s participation
was largely confined to voting through elections. In the current code, citizen groups can initiate the formulation
of municipal ordinances and the recall of officials. Likewise, theyh can also participate in the planning and
monitoring of government projects. (Porio, 1997: 12). The Code also devolved to local governments the
responsibility to delivering basic social services, define land use, and promote tourism and ecological balance.
It likewise grants taxation powers and increase revenues both from local and national sources.
Reacting to the long years of martial rule under Marcos, the Filipinos, in 1986-1992, instituted political reforms
that would protect them from leaders who would want to perpetuate themselves in power like Marcos. Thus,
term of President Corazon Aquino witnessed innovations that served to protect the people’s rights against
possibilities of another strong man imposing a dictatorship on the country. Such political policy changes were
enshrined in the constitution, in the devolution of power to local government units and to the passage of laws
like RA 7279, otherwise known as the Urban Development and Housing Act of 1992.
The Passage of RA 7279
Slide No. 12 shows that it took all of 5 years before this law was passed. Over two years, a number of public
hearings were conducted. In 1991, the Committee on Urban Planning, Housing and Resettlement finally
recommended approval of a substitute SB 234. This law was signed into law by President Aquino on March 24,
1992 and took effect on March 29 of the same year.
40
How was the passage of this law received? Karaos acknowledged how the urban poor mobilized after the
1986 people power revolution. Popular mobilization, she asserted, become the urban poor’s visible
mechanism for demand-making and for influencing urban policies. She analyzed the passage of RA 7279 thus:
“One of the sector’s most important victories was the passage of the 1992 Urban Development and Housing
Act (UDHA). This law resulted from the active lobbying by NGOs, civic groups, the Church and urban poor
organizations, pressing local governments to observe a moratorium on squatter demolitions. With the law’s
enactment, local governments are ordered to meet legal requirements before a demolition, conduct a land
inventory and beneficiary registration, and identify sites for socialized housing.” (Karaos:69).
Republic Act 7279: The Urban Development and Housing Act
Slides 6-10 present the key features of this law. Slide 6 summarizes the objectives of the law and the principles
that it supports. Slide 7 explains how the formulation of a National Urban Development and Housing Framework
was provided for in this law. It likewise identifies the agency mandated to formulate this framework and how
this framework is to be prepared. Slides 8-10 highlight how the law supports decentralization, engages the
private sector and introduces the concept of the community mortgage program in the provision of social housing.
An outline of the 13 articles and the 49 provisions of the law give us a total picture of what this law covers:
Article 1. Title, Policy, Program and Definition of Terms
Sec. 1. Title
Sec. 2. Declaration of State Policy and Program Objectives
Sec. 3 Definition of terms.
Article II. Coverage and Exemptions
Sec. 4. Coverage
Sec. 5 Exemptions
Article III. National Urban Development and Housing Framework
Sec. 6 Framework for Rational Development
Article IV. Land Use, Inventory, Acquisition and Disposition
Sec. 7 Inventory of Lands
Sec. 8 Identification of Sites for Socialized Housing
Sec. 9 Priorities in the Acquisition of Land
Sec. 10 Modes of Land Acquisition
Sec. 11 Expropriations of Idle Lands
Sec. 12. Disposition of Lands for Socialized Housing
Sec. 13. Valuation of Lands for Socialized Housing
Sec. 14. Limitations on the Disposition of lands for Socialized
Housing
Article V. Socialized Housing
Sec. 15. Policy
Sec. 16 Eligibility Criteria for Socialized Housing Program
Beneficiaries
Sec. 17. Registration of Socialized Housing Beneficiaries
Sec. 18. Balanced Housing Development
Sec. 19. Incentive for the National Housing Authority
Sec. 20. Incentives for Private Sector Participating in Socialized
Housing
Sec. 21. Basic Services
Sec. 22. Livelihood Component
Sec. 23 Participation of Beneficiaries
Sec. 24. Consultation with Private Sector
Article VI. Areas for Priority Development, Zonal Improvement Program
Sites and Slum Improvement and Resettlement Program Sites
Sec. 25. Benefits
41
Article VII. Urban Renewal and Resettlement
Sec. 26. Urban Renewal and Resettlement
Sec. 27. Action against Professional Squatters and Squatting
Syndicates
Sec. 28. Eviction and Demolition
Sec. 29. Resettlement
Sec. 30. Prohibition against New Illegal Structures
Article VIII. Community Mortgage Program
Sec. 31. Definition
Sec. 32. Incentives
Sec. 33. Organization of Beneficiaries
Article IX. Related Strategies
Sec. 34. Promotion of Indigenous Housing Materials and
Technologies
Sec. 35. Transport System
Sec. 36. Ecological Balance
Sec. 37. Population Movements
Sec. 38. Urban –Rural Interdependence
Article X. Program Implementation
Sec. 39. Role of Local Government Units
Sec. 40. Role of Government Housing Agencies
Sec. 41. Annual Report
Article XI. Funding
Sec. 42. Funding
Sec. 43. Socialized Housing Tax
Article XII. Transitory Provisions
Sec. 44 Moratorium on Eviction and Demolition
Article XIII. Common Provisions
Sec. 45. Penalty Clause
Sec. 46. Appropriations
Sec. 47. Separability Clause
Sec. 48. Repealing Clause
Sec. 49. Effectivity Clause
Implementing Rules and Regulations for RA 7279
Slides 11 and 12 list the various implementing rules and regulations that are meant to govern the implementation
of specific provisions of this law. The Housing and Land Use Regulatory Board (HLURB) issued the guidelines
for the inventory and Identification of Lands and Sites for Socialized Housing. For its part, the Housing and
Urban Development Coordinating Council (HUDCC) provided guidelines for the acquisition, valuation,
disposition and utilization of lands for socialized housing. The Department of Finance (DOF) has guidelines on
equitable land valuation for socialized housing. Together, the Department of Interior and Local Government
(DILG) and the HUDCC require compliance with their implementing rules and regulations on the registration of
socialized housing beneficiaries.
The HLURB also issued guidelines on Balanced Housing, while the DOF provided incentives to
government-owned and controlled corporations and local government units and the private sector to participate
in the socialized housing and community program. DILG and HUDCC also have their guidelines on the
observance of proper and humane relocation and resettlement procedures.
The National Urban and Housing Development Framework (NUHDF)
RA 7279 calls for the formulation of a National Urban and Housing Development Framework by the HUDCC
42
and the HLURB. Slides 13, 14 and 15 present the vision and the basic principles and considerations that
governed the first NHUDF for 1993-1998. A second NHUDF was formulated for 1998-2004. Prepared in a
participatory process, the 1998-2004 NHUDF was reviewed by all levels of government and all sectors of civil
society in a series of twelve (12) regional consultations held throughout the country from September to October
1999. Guided basically by the same principles that underlie the first NUHDF, the second framework was
organized along six key themes: 1) Urban growth, integration and metropolitanization, 2) Urban land resource
management, 3) Urban environmental management, 4) Physical and social infrastructure, 5) Housing and
regulations, and 6) Urbanization and governance management.
Implementing RA 7279: Issues and Problems
Slides 16-21 discuss how RA 7279 was received by the urban poor and civil society, church groups and local
government units. Concerns over the inability of local government units and some national government offices
to fully implement the law are raised in these slides as these are captured in the literature and articles in
newspapers of national circulation.
Critics of the law complain of Executive Order 152. Issued on December 2002, Executive Order No 152
designates the Presidential Commission for the Urban Poor (PCUP) as the sole clearing house for the conduct of
demolition and eviction activities involving the homeless and underprivileged citizens and establishing for the
purpose a mechanism to ensure strict compliance with the requirements of just and humane demolition and
eviction under the Urban Development and Housing Act of 1992, and for other purposes.
43
MODULE 1.4: PROJECT MANAGEMENT1
Introduction
To enhance the quality of life of the inhabitants, public institutions have a responsibility to initiate and manage
development undertakings to address poverty (e.g., by eliminating social evils such as unemployment);and to
redress historical inequalities.
Development projects are the means by which the development programs are to be implemented; thus, the need
to explore the role of public managers in managing such projects.
We can consider the following:
•definition of projects, particularly those with a development focus;
•place of project management in administration;
•management skills required of the public manager in managing development projects; and
•overview of the various phases of the project management cycle, with specific reference to the role of
the public manager in each phase.
As to project management cycle, the following phases will be described, namely:
•project conception which includes project identification, formulation and preliminary design;
•project preparation whereby the project proposal is subjected to a series of feasibility analyses and then
submitted for appraisal;
•project monitoring and implementation; and
•project evaluation.
Development Project
Development projects will be the devices through which various institutions will be expected to implement
development policies. The focus of these projects will be on initiating changes in government spending priorities,
launching long-term development programmes, and delivering concrete benefits to disadvantaged communities.
However, before commencing with a detailed analysis of such projects, one must define a project and
characterize development projects.
There is no standard, accepted definition of a project, although there are commonalities in various definitions. A
“project” is defined by the Project Management Body of Knowledge (in Oosthuizen, 1994:42) as “any
undertaking with a defined starting point and specific objectives whose attainment marks the completion of the
project.” Most projects depend on finite or limited resources by which the objectives are to be achieved. The
Triple Constraint refers to projects having a three-dimensional objective, namely the simultaneous
accomplishment of the performance specification, the time schedule, and the budget. ( Rosenau, 1992:2)
A similar definition of a “project” is provided by Kerzner (1992:2) who considers a “project” to be “any series of
activities which has a specific objective to be completed within certain specifications;
has defined start and end dates; and consumes resources such as finance, personnel and equipment.
A project can be distinguished from continuous processes and service-rendering in the public sector in that it is a
non-routine, non-repetitive, once-off undertaking.” Burton and Michael (1992:3) elaborate by stating that
ongoing work does not constitute a project, yet reorganising the manner in which the work is done could be
considered a project. Thus, each project is unique since it is carried out only once, its implementation cannot be
1From: H J Nel , PROJECT MANAGEMENT FOR DEVELOPMENT: THE ROLE OF THE PUBLIC MANAGER,
http://www.up.ac.za/academic/soba/SAAPAM/administratio%20publica/vol8no2/nel.htm (with minor editing).
44
rehearsed, and it is temporary in duration. Added to this, a project originates because "something not done before
must be done" and is therefore characterized by the production of a unique product (Rosenau, 1992:6).
Atkins and Milne (1995:3-5) distinguish between “conventional” and “development” projects. It is asserted that
development projects extend the project activities, output and time frame beyond the scope of a conventional
project by: encouraging and assisting the beneficiary community to actively participate in the project and to take
ownership, in so far as possible, of the asset created; maximising the short-, medium- and long-term project
benefits to alleviate poverty in a sustainable and replicable manner; using the project as a vehicle for training and
building the capacity of the local community; enhancing employment opportunities through the use of
labour-intensive technologies; and minimising negative environmental impact and thereby enhancing
sustainability.
Baum and Tolbert (1985:333) provide a World Bank perspective and define a “development project” as a
“discrete package of investments and institutional actions designed to achieve a specific development objective
within a designated time frame.” Similarly, Shaghil and Mushtaque (1993:48) regard a “development project” as
a “technically predetermined set of interrelated activities involving the most effective use of resources to achieve
development objectives and provide goods and services to the benefit of targeted communities.”
The Reconstruction and Development Pretoria (RDP) White Paper (1994:15) makes provision for development
projects which conform to certain criteria, namely, high impact on communities served; economic and political
viability and sustainability; job creation; provision of basic needs; training and capacity development; some
existing capacity to start implementation; visibility; transparency; and affirmative action with respect to race and
gender.
Wallis (in Reddy, 1996:172) adds that two categories of projects can be identified in relation to the RDP, namely
presidential or "lead" projects; and projects initiated by local authorities yet aimed at RDP objectives.
According to Wallis (in Reddy, 1996:172) the first category of development projects serves a catalytic function,
enabling the programme to start and ensuring that certain identified needs are met as soon as possible. Such
projects are said to be catalysts in the sense that they are supposed to generate other projects and activities of a
similar type.
It is therefore apparent that a “project” is a “set of activities with defined start- and end-dates, specific objectives
to be achieved with certain constraints, and limited resources to be utilised.” In addition, it was pointed out that a
development project exhibits these characteristics although the scope thereof is broader than conventional
projects in various respects. For instance, it was asserted that development projects encourage capacity-building,
employment creation through the use of labour-intensive technologies, and sustainability or a concern for
minimising negative environmental impacts.
It was further indicated that a broad development programme can be concretized and implemented by
undertaking various development projects. This is reflected by the RDP and the relationship between a
development programme and project will be analysed in more detail in the section which follows.
Project Management in the Public Sector
A project can be regarded as a component of a programme. Shaghil and Mushtaque (1993: 44) point to an
important distinction between a project and a programme. A “programme” consists of a “group of similar, related
or allied projects.” This is supported by the United Nations (1971:2) which write that a “programme” is an
“organised social activity with a specific objective and comprises an interrelated group of projects.” A road
construction programme, for example, can be executed as a string of road construction projects linked in time,
space and function. Meiring and Parsons (1994, ch.5:21) describe a “programme” as a “set of specific action
steps which must be undertaken separately or simultaneously to attain policy objectives.” The provision of
services in the public sector commences with the formulation of a policy and continues with the establishment of
45
a programme which is to be implemented operationally by means of a number of interrelated projects. It can be
deduced that policy implementation comprises two phases, namely the planning and programming phase; and
the programme implementation phase. Project management can be regarded as a technique to be applied during
the programme implementation phase of policy implementation. A logical sequence becomes apparent whereby
goals and objectives are formulated and approved in policies, plans are made to ensure their achievement, and
programmes (consisting of projects) are initiated to implement policies and plans within a given time frame and
according to a specific sequence.
Van der Walt and du Toit (1997:314) support this by stating that, within the development framework and
philosophy of the RDP, public institutions have to undertake projects and programmes at local level to develop
“disadvantaged communities.” In this respect, local development programmes are divided into numerous
community development programmes, which, in turn, comprise various development projects. A development
programme is thus practically executed by means of development projects. This is graphically illustrated in
Figure 1.
It can be deduced that “project management” is a “specific technique which can be applied by public managers
to ensure the effective and efficient implementation of development policies.” To ensure that the community
concerned is served effectively and that the project will be completed within time, cost and quality constraints,
the public manager responsible for the project needs to demonstrate certain qualities and skills.
Management Skills for Project Management
The public manager must possess those skills relevant to the management of development projects, namely,
leadership, motivation, team-building, and the resolution of conflict.
Leadership Skills
“Leadership” is critical to high performance on projects and can be defined as the “ability of the public
manager to influence the behaviour of subordinate officials to support goal-attainment” (Harrison, 1992:253).
Leadership involves "getting followers to follow" and certain traits are significant in this regard, namely, a high
personal commitment to institutional goals; inspiring and maintaining commitment of subordinates to attain
these goals; and actions to give effect to these goals (Harrison, 1992:253).
Leaders have a responsibility to "strike a balance" between two types of orientation, namely, task orientation;
and employee orientation. Dessler (1986:353-354) “Task orientation” is characterised by behaviour which
stresses productivity and the technical aspects of the project and reflects an assumption that subordinates are
merely a means to an end. An employee-oriented leader views subordinates as human beings of intrinsic
importance and accepts their individuality and personal needs.
46
Public managers can improve their effectiveness by giving due consideration to both task needs as well as team
and individual needs of subordinates Reiss (1992:159). This is supported by Harrison (1992:257) who states that
essentially the public manager responsible for managing a project needs to have a high task orientation in that
the prime focus should be the effective attainment of project objectives through high performance. However, the
public manager will not necessarily achieve high performance if subordinates are alienated, demotivated, and
uncommitted to the project objectives; thus, the two dimensions of leadership are not mutually exclusive since
the employee-oriented dimension of leadership is supportive of the task-oriented dimension of a project.
Situational factors play a role in determining which leadership traits and styles are most effective. Generally, the
project setting favours a task-oriented leader, although there are certain situations which necessitate an
employee-oriented style; thus, it is stressed that the public manager should assess the situation and be
sufficiently flexible to adapt the leadership style to the particular conditions of the situation. Irrespective of the
leadership traits or styles exhibited by the public manager, management involves attaining objectives with and
through other people.;thus, in ensuring the effective and efficient attainment of project objectives, the public
manager needs to be able to motivate subordinate officials.
Motivational Skills
Reiss (1992:174) points out that motivated subordinates are productive and obtain satisfaction from achieving or
striving towards group objectives. Kerzner (1992:519) supports this by writing that team motivation has been
identified as having the strongest overall influence on project success, and is an important factor in all phases of
a project; thus, the public manager has to learn the skill of motivating subordinates individually and collectively.
A task of the public manager is to ascertain what motivates subordinates and then ensure that these motivations
are fulfilled in the project setting. Theories exist as to what motivates people, yet mention will only be made of
the suggestions of a few authors with respect to the essential elements of motivation. Approval, praise,
recognition, trust, job enrichment, and effective communications enhance motivation (Stallworthy and
Kharbanda (1983:87). There is a distinction between driving forces and restraining forces of motivation in the
project setting (Kerzner (1992:519-520). The former include factors, such as, good interpersonal relations, a
common goal, clear role definition, and project visibility. The latter relate to factors, such as, poor team
organisation, communication barriers, poor leadership, and uncertain rewards.
The public manager can motivate subordinates by clearly defining roles, responsibilities, and performance
expectations; assigning tasks which are challenging; giving honest appraisals of performance and giving credit
where it is due; providing a positive working environment; and involving subordinates in decisions relating to
the project (Anderson et. al., 1984:189; Kerzner, 1992:245).
The motivation of subordinates to perform well is a crucial aspect in determining the success of a development
project. To this end, the public manager needs to actively learn the skill of motivating subordinates by taking
cognisance of factors which motivate people and then ensuring that these factors are given due consideration
during the various phases of the project.
An important element of motivation is the feeling of belonging or being part of a team. It is, therefore, essential
that the public manager maximise the performance of subordinates by exercising team-building skills.
Team-Building Skills
Goodman and Love (1980:35) note that the subordinate officials involved in a particular project are not
necessarily from the same work unit, or even the same department. The public manager overseeing the
implementation of a development project thus needs to ensure that project success is achieved as a team effort.
47
This is supported by Child (1984:133) who asserts that it is natural for individuals to identify with their own
work unit or department, as opposed to other departments. However, if joint projects are called for, then project
teams drawing on officials of two or more departments need to be developed and maintained for the duration of
such projects.
The public manager requires team-building skills, especially when supervising team members drawn from
different work units or departments. Harrison (1992:269) defines a “team” as “any group of individuals involved
in a joint undertaking, such as a project, where interaction and interdependency between the individuals are
required for efficiency and effectiveness.”
Giesen (1996:27) summarises by providing a "winning" model for team-building. This model comprises a
number of steps to be undertaken by the public manager when attempting to establish a project team, namely,
define the purpose and goals of the team; establish team composition and roles;
clarify team rules and responsibilities; integrate individual personalities; manage team performance; and evaluate
team productivity.
The public manager responsible for a development project requires the skill of taking individuals, often from
different work units or departments, and building an effective project team. There are various techniques
whereby team-building can be achieved by the public manager all of which, if performed effectively, can lead to
positive outcomes such as the exchange of quality information among team members, improved decision-making
and problem-solving, as well as more effective systems of control and feedback regarding performance.
Ineffective team-building may result in interpersonal conflict among team members which will obviously be
detrimental to project performance .Conflict situations often arise in the project setting even where
team-building has been effective.
Conflict Resolution Skills
A significant problem in the achievement of teamwork and co-operation in the project setting is the existence of
conflict between individuals and groups Harrison (1992:276-277). Rosenau (1992:190) adds that projects are
fraught with conflicts in that they are temporary undertakings within a more permanent institution. Thus, public
managers responsible for overseeing the implementation of development projects require a high tolerance for
conflict, as well as specific skills to resolve conflict. A number of research studies indicate that public managers
can resolve conflict, irrespective of the source, by making use of various conflict resolution modes, namely,
withdrawal or retreating from an actual potential disagreement; smoothing or de-emphasising areas of difference
and emphasising areas of agreement; compromising or searching for solutions which are mutually satisfactory to
parties involved in the conflict situation; forcing or exerting a viewpoint at the potential expense of another
(characterised by a win/lose perspective of conflict resolution); and confrontation or facing the conflict directly
whereby parties involved in the conflict work through their disagreements by adopting a problem-solving
approach (Kerzner, 1992:419-420).
Harrison (1992:295-296) asserts, however, that the abovementioned modes of conflict resolution are mostly
applicable to the resolution of disagreements when they arise, and certain modes such as withdrawal may resolve
the difference yet increase the underlying source of the conflict. The public manager should seek to manage
conflict by preventing it before it occurs. This can be achieved by creating conditions conducive to teamwork
and collaboration, as well as recognising the first signs of the deterioration of relationships and taking action to
reverse the movement.
Conflict is inevitable in the project setting and that the public manager responsible for overseeing the
implementation of a development project requires conflict resolution skills. It is first necessary for the public
manager to diagnose the situation and identify potential sources of conflict, which may range from personality
differences to competition for scarce resources. Based on this diagnosis, the public manager can then adopt a
mode of conflict resolution deemed to be most suitable in addressing the conflict situation. However, it was
48
stressed that it is not sufficient to merely resolve differences, but to also deal effectively with the underlying
causes of conflict. In other words, the public manager must not only resolve conflict but also manage it by
identifying potential sources of disagreement and dealing with these before they result in disruptive
disagreements.
In addition to demonstrating the abovementioned skills, a public manager needs to follow a series of practical
steps to launch and effectively manage a development project.
The Role of the Project Manager in Project Management
Project management, as a component of programme implementation, is undertaken in various phases which are
interrelated and follow a logical progression. Numerous views exist regarding the phases constituting the project
management cycle. Attention will be devoted to the role of the public manager in
project conception, which comprises the identification and formulation of a project proposal;
project preparation whereby the project proposal is subjected to a series of feasibility analyses, based
upon which a project report will be compiled and submitted for project appraisal;
project implementation or the actual execution of the development project by the project team; and
project evaluation or the detailed analysis of the overall effectiveness of the development project in
attaining objectives contained in broader development programmes and policies.
Figure 2 below indicates these phases diagrammatically.
The first phase of the project management cycle is therefore project conception.
Project Conception
Harrison (1992:241) writes that the conception phase of the project cycle involves taking an idea and converting
it into a formal project proposal. Project conception thus refers to that phase whereby,
the project is defined in conceptual terms;
the objectives of the project are identified; and
the requirements to complete the project are roughly defined.
Oosthuizen (1994:43) asserts that sufficient quality time needs to be devoted to project conception in that
project success is often contingent upon the extent to which project objectives are compatible with the
49
constraints imposed by limited resources. The project conception phase will henceforth refer to certain steps,
each of which will be analysed in detail, namely, project identification and formulation. In this regard, it is
necessary to commence with clarity concerning the essence of project identification as the first stage of project
conception.
Project Identification
The first task of the entire project cycle is simply to identify the project area (According to Goodman and Love
(1980:49). Ideally, projects should be a response to a readily apparent community need or a deficiency in the
development of the local environment. Projects therefore begin as ideas which contemplate movement via
concrete actions towards new or improved situations to meet identified community needs.
Shaghil and Mushtaque (1993:93-94) concur with this conceptualisation of project identification by further
describing it as the process whereby there is a search for "... an idea with development content ..." which could
find concrete expression in the form of a feasible course of action to develop the local community. It is added
that the conception of a project idea does not take place in abstraction, since the basic idea is to identify the type
of activities which are likely to contribute towards the attainment of development objectives contained in
national and provincial policies.
Project managers at local government level will have to target projects which will have the greatest impact in
terms of government policies, priorities and guidelines.
According to Baum and Tolbert (1985:339) projects originate from a multiplicity of sources. In practice, project
ideas often result from the identification of unsatisfied community demands or needs and possible means to meet
them; problems or constraints in the development process caused by shortages of essential facilities and material
and human resources. unutilised or under-utilised material and human resources and opportunities for their
conversion to more productive purposes; and
the need to complement other investments (such as providing roads and sanitation to a housing project).
Once a development project has been identified, it is necessary to develop a statement in broad terms which
indicates project objectives and outputs, as well as estimates of the resources required by the project. This stage
of the project management cycle is referred to as project formulation and preliminary design.
Project Formulation and Preliminary Design
At the project identification stage, a project may only be an idea with rough estimates of its desirability in terms
of national needs, as well as its possible cost and likely benefits. During the formulation stage, the project has to
be spelt out in greater detail and in more specific terms, in order to enable the decision-making bodies to
appraise it. This phase should therefore lay the foundation and provide the blueprints for all the other phases of
project management (United Nations, 1971:76).
Conyers and Hills (1984:132) assert that various aspects of the proposed project demand attention during this
stage:
nature and extent of resources to be utilised;
nature and extent of the product/service to be provided;
target group of the project;
duration of the project; and
relationship between the resources used and the outcome of the project.
Lister (1995:52) asserts that previous research points to the importance of a thoroughly-investigated project
50
definition which is clearly communicated and agreed upon by the relevant stakeholders. Baum and Tolbert
(1985:340) support this by stating that emphasis needs to be placed upon the importance of explicitly defining
the objectives that the project is intended to achieve.
In this respect, all funding applications for development projects may be submitted in the form of a "business
plan" whereby the preliminary design for the project is clearly set out. Such a business plan is prepared in
accordance with a prescribed structure, that is,
details of the project manager and implementation agency;
description of the project;
project capital estimate (labour, materials, training etcetera);
time and funding plans;
cash flow;
provision for maintenance and recurrent expenditure;
commitment to include labour-intensive technology where feasible to do so; and
the conformity of the project with RDP guidelines relating to community participation, the environment,
employment policy, and capacity building (Project Pro, 1994:7).
It is stipulated that the business plan submitted by a public institution must deal in detail with the priority of the
project in the area, the beneficiaries, as well as the socio-economic conditions and the local situation (Hazlewood,
1995:23). Furthermore, the plan is also required to give a detailed description of
identified local RDP policy objectives;
the number of people who will benefit from the project;
the project design and the compatibility thereof with the project area; and
the exact and quantifiable outputs of the project.
The logical framework approach is used in project formulation and preliminary design in that it requires
systematic and comprehensive information relating to the objectives, design, activities, and proposed outputs of
the project. This approach is laudable since it provides the project manager with a guideline with regards to the
detailed information required by decision-makers and lending institutions when submitting a project proposal for
approval and funding. In addition to this approach, the “comprehensive approach” in respect of project
formulation and preliminary design provides a blue print for action and review in the phases of the project cycle
following appraisal and approval, namely, implementation and evaluation.
The process which is launched when a project is identified continues through the stages of preparation and
appraisal and leads ultimately to the decision whether to commit funds to its implementation. Refining project
objectives and the means of achieving them is an important part of project preparation which will be focused
upon in the next section as the second phase of the project management cycle.
Project Preparation
Project preparation provides the basis upon which the loan or grant for the project will be determined, and
around which agreements will be signed Hazlewood (1995:15). The preparation of a project usually covers the
full rationale, a thorough feasibility analysis, plan schedule, and cost estimate of the proposed project - complete
with supporting documents, tables, schedules and special studies.
Similarly, it is the opinion of that the preparation of a project for approval is a "painstaking process" requiring
the assembling of a ll relevant data, their careful assessment, and the examination of all possible alternative
approaches (Shaghil and Mushtaque,1993:73). Project preparation involves making decisions relating to
technology, scale, location, costs and benefits, degree of risk and uncertainty, financial viability, and other factors.
The alternatives are evaluated to help select that combination which is likely to prove the most appropriate from
the technical, financial, economic, management, and other perspectives.
51
A significant component of project preparation is a statement of work (SOW) which describes, in detail, the
activities to be undertaken by the public institution in achieving the project objectives identified during the
formulation phase. A project proposal can be drawn up whereby each aspect of the Triple Constraint is specified,
namely, the performance specification (for example, the quality of service to be provided); the time required to
complete the project; and cost estimates for each project activity (Rosenau, 1992: 32).
The line of demarcation between project identification and preparation is often not clear. Shaghil and Mushtaque
(1993:95) provide a guide in this regard by stating that a distinction can be made on the basis of the difference
between "hypothetical concepts" and "action programmes". At the identification stage, projects may be
characterised as “hypothetical concepts” whose viability has yet to be established. The stage of project
preparation, however, can be distinguished by its attempt to investigate the project contribution, if implemented,
towards achieving identified project objectives; and practical possibilities of implementing the idea.
Project preparation mainly involves subjecting the initial formulation of project ideas to an analysis of their
feasibility. Upon completion of various feasibility analyses, a detailed project proposal can be submitted to the
legislature or funding institution (such as a development bank) for appraisal, and possible approval. It is essential
to provide an overview of feasibility analysis and project appraisal as requirements during project preparation.
Feasibility Analysis
Feasibility analysis examines whether it is possible to implement a project, given the standards and criteria set
forth in the preliminary design. Well prepared feasibility analyses question every aspect of the preliminary
design within the context of the actual project environment and determine whether a project can be satisfactorily
carried out with the financial, technical, human, material, and institutional resources available. Thus, together
with project appraisal, feasibility analysis functions as the interface between conception and reality. In other
words, these components of project preparation link the planning set of project tasks, namely, project
identification, formulation and preliminary design with the action-oriented set of tasks, namely, project approval,
implementation and evaluation (Goodman and Love, 1980:82).
Feasibility reports may be regarded as tentative proposals which are subjected to scrutiny of two types, namely,
scrutiny by the implementing agency, in this case a public institution, of its proposals which undergo many
revisions until a finally accepted scheme evolves and is submitted for appraisal; and a "third party," in this case,
the legislature at provincial or national level or a funding institution, whose responsibility it is to assess the merit
or worthiness of the proposal objectively and determine whether the project should be funded and implemented.
(Shaghil and Mushtaque , 1993:111-112)
Baum and Tolbert (1985:347) support these conceptions of feasibility analysis and write that such an analysis
should form the central core of project preparation. They believe that the purpose of conducting feasibility
analyses is to provide decision makers with the basis for deciding whether to proceed with the project and for
selecting the most desirable option for the attainment of national development objectives. As its name implies, a
feasibility analysis is undertaken to establish the justification or feasibility of the project as a whole in all its
relevant dimensions (technical, financial, and so forth). Each of these dimensions is analysed, not only separately
but also in relation to all the others, through a process of successive approximations.
Other crucial functions of feasibility analysis are summarised by Fourie (1997:10). Firstly, by examining project
objectives and questioning all assumptions, feasibility analysis provides a framework to reformulate the
preliminary design into the most appropriate design. Secondly, such an analysis assists in guiding the
implementation of the project by pointing out potential "trouble spots" and by discussing the use of possible
contingency plans. Finally, a complete feasibility analysis includes criteria and baseline measures which can be
used to monitor and evaluate the project during and after its implementation by the public institution.
52
Testing the feasibility of a project should relate to factors such as the objectives in view; overall nature of the
measures contemplated in the attainment of these objectives; and mutual consistency thereof.
Once the development project has been analysed in terms of its feasibility, a report must be compiled and
submitted to the relevant legislature or funding institution for appraisal.
Project Appraisal
During the appraisal of the project, decision makers need to satisfy themselves that the project meets the
conditions which enable it to proceed effectively and efficiently. The chief concern in this respect is to determine
whether or not the project is the best means of achieving policy objectives.
Conyers and Hills (1984:130-131) elaborate on the meaning of project appraisal by stipulating that it basically
means comparing alternative courses of action formulated during the initial, planning phases of the project
management cycle, prior to exercising a choice among them. The term "appraisal" embraces an ex ante analysis
of a proposed project or, in other words, constitutes an analysis conducted before a project is selected, funded
and implemented. Du Mhango (1993) supports this by describing the appraisal of a project as an aid to assist in
decision-making as to whether to approve and/or support the implementation of a project.
Project appraisal needs to address two questions, namely: will the project as it is designed meet its own
objectives as well as the wider needs of its location and nation?; and how does the project compare with other
projects it may be competing with for funding? (Goodman and Love, 1980:101).
An appraisal seeks to establish what will happen if a particular proposal is implemented, where the anticipated
effects will occur, who will gain and lose, when those effects will occur, and the efficiency of the investment in
relation to the resources used and the benefits derived (Conyers and Hills, 1984:132). Thus, a detailed appraisal
provides an indication of the aforementioned factors and, in so doing, needs to be based on reliable and accurate
information, as well as the use of sound techniques by which this information is analysed.
Project appraisal involves testing various project alternatives with the aim of selecting an option which is
technically competent, as well as generally practicable and feasible. In essence, there are certain criteria which
may be used in this testing process, namely checking the likelihood of current and future problems being
adequately addressed through the proposed project; flexibility of the project design or its capacity for
modification as a response to uncertainties about future conditions feasibility of the proposed activities in
relation to economic, political, or other constraints; and conformity of the proposal to accepted standards and
principles applying to similar projects (Lichfield (1975:153);
.
Project appraisal comprises the final stage of project preparation and is undertaken by the legislature or a
funding institution with the purpose of assessing whether a proposed development project warrants approval
and/or financial assistance. In the event of a project proposal being selected, the public institution can proceed
with the final design of the development project, following which the project implementation phase of the
project management cycle will commence.
Project Implementation
Hazlewood (1995:15) defines “project implementation” as that “phase of the project management cycle which
represents the actual establishment and operation of a development project.” Goodman and Love (1980:135)
elaborate by providing an overview of the various tasks of the public manager, during project implementation.
These tasks include:
•programming the implementation of the project by breaking it down into its component tasks and
activities;
53
•directing the implementation of the project by assigning project tasks and activities to project team
members;
•allocating and utilising necessary resources such as personnel, finance, time and materials; and
•coordinating, monitoring and controlling the performance of the project team and the use of project
resources in a manner which will ensure the completion of all project activities in an orderly and optimal
way.
Botha (1995:38) indicates certain issues when implementing a development project. Firstly, the scheduling of
activities and the assignment of responsibilities require a clear explanation. It is added that the responsibilities of
key role players needs clarification to ensure that optimal use is made of available resources, including human
and financial resources. Secondly, the degree of synergy between the institutional arrangements and the
implementation plan needs to be examined by demonstrating how project implementation will be coordinated
and supervised. Thirdly, the monitoring and control techniques need to be selected. It is pointed out that due to
the importance of maintaining public accountability, the control measures selected should be relevant to the
targeted communities as well as transparent in their use.
Project implementation is characterised by key tasks of the public manager responsible for the project. These
tasks can be summarised as producing a work breakdown structure which assigns responsibilities to each project
team member; scheduling project activities and establishing time frames for attaining project milestones;
ensuring the efficient and effective utilisation of resources allocated to the project; and monitoring and
controlling progress in attaining project objectives.
Specific management techniques can be applied to effectively implement projects. Among these techniques,
network analysis and its variations (for example, Programme Evaluation and Review Technique or PERT and the
critical path method) have found widespread application in scheduling and monitoring project implementation.
Such techniques break a project down into detailed activities and establish interrelationships between various
activities. This permits the project manager to arrange the project in terms of manageable components;
effectively co-ordinate various project activities; and obtain time and cost schedules for the project.
Not only do such techniques enhance internal project coherence, but they also save implementation time by
isolating problems into their appropriate project components (Goodman & Love, 1980:15). These techniques are
useful only if action is taken to correct any deviation by means of ongoing review and monitoring while the
project is being implemented. A final evaluation phase is required in order to assess the overall worth of the
project.
Project Evaluation
The project cycle does not end when a project has been implemented. The final stage of this cycle is that of ex
post evaluation which is retrospective in focus and takes place after the project has been implemented and
completed. An important purpose of such an evaluation is to ascertain the reasons for the apparent success or
failure of a project, in order to identify the features that deserve replication in future projects and certain pitfalls
which should be avoided. Ex post evaluation also has an accountability facet in that it informs the public
manager and policy-makers how effectively and to what extent individual projects have achieved the desired
results (World Bank, 1993:381). In addition to goal-attainment, ex post evaluation should also study the impact
of the project on the target group, as well at its impact on various components of the environment. Furthermore,
an exhaustive evaluation of each phase of the project cycle is required in order to determine its contribution to
the project in respect of budget, timetable and other factors (Goodman & Love, 1980:17).
Evaluation studies can assist the public manager in assessing various factors, such as
•the extent to which the project produced intended effects (e.g., how many households enjoy minimum
health standard);
54
•the distribution of benefits between different groups (e.g., how many low-income households enjoy
minimum health standards in proportion to middle- and high-income households); and
•the cost-effectiveness of the project as compared with alternative delivery systems (e.g., what is the
average cost of vaccinating a child in the new health centre compared with the average cost of
vaccination in a hospital) (World Bank, 1993: Annexure).
•The public manager may experience certain constraints when evaluating a completed project. Such
constraints may include goals change/not clear; inadequate criteria for measuring success;
•unintended consequences/spillover effect; cost of evaluation; timing of evaluation (ie too late/too soon);
and results of evaluation too academic or technical (World Bank, 1993: Annexure).
Related to and often arising from the evaluation of a project is the need for follow-up. For a project to achieve its
objective fully, smaller or related projects may need to be implemented almost immediately. In other words, if
follow-up action means the difference between a project being fully operational or not, then it is a wise
investment to undertake these activities in a timely manner (World Bank, 1993: Annexure).
The final task of the public manager with respect to the evaluation of a project is the refinement of policy and
planning. Policy makers and public managers will need to refine policies, plans and procedures in the light of
each completed project. Experiences and lessons learned should form the basis on which such policies and plans
are reviewed. Refinement of policies and plans is required in order to keep pace with changing circumstances
and the information provided through project evaluation can make a useful contribution in this regard (World
Bank, 1993: Annexure).
Conclusion
A “project” can be defined as a “set of activities with defined start and end dates, specific objectives and
constraints such as time, cost and quality.” Furthermore, a project was distinguished from a service or a process
in that it is a non-routine, non-repetitive and once-off undertaking. Particular emphasis was placed on
development projects in the sense that such projects aim to enhance employment opportunities; empower the
beneficiary community through training and capacity-building; and maximise the sustainability of project
benefits.
As the place of project management in administration, development projects are the "building blocks" of
development. In other words, public institutions can ensure the effective attainment of development policy
objectives by implementing a series of interrelated and interdependent development projects. It was pointed out
that project management is a specific technique which can be fruitfully applied by public managers to ensure
effective and efficient policy implementation.
However, the effectiveness of a project in attaining development objectives, is largely contingent upon the
management skills of the public manager responsible for the implementation of such a project. In this respect,
the public manager needs to possess skills such as the ability to lead project team members in a manner which
supports goal-attainment; motivate subordinates individually and collectively and thereby enhance productivity
and work performance; create an environment in which the subordinates and public manager work together as a
team; and identify sources of conflict in the project setting and resolve these so as to minimise the potential
dysfunctional impact thereof on work performance.
The public manager has a role to play in respect of
•project conception by which the need for a project is identified and a project proposal is
formulated;
•project preparation or the conducting of a series of feasibility analyses to ensure that the project
proposal is sufficiently prepared for appraisal;
•project implementation which comprises the actual execution of tasks to achieve the objectives
of the project; and
55
•project evaluation or the overall assessment of the effectiveness and efficiency of the project in
meeting identified development needs in the
•References
References
Andersen, E.S., Grude, K.V., Haug, T. and Turner, J.R. 1987. Goal-Directed Project Management, London,
Kogan Page.
Atkins, H. and Milne, C. 1995. Role of Consultants in Development Projects, Development Bank of Southern
Africa.
Baum, W.C. and Tolbert, S.M. 1985. Investing in Development, World Bank Publication.
Botha, J.H. de V. 1995. "Basic Needs and Urban Reconstruction: An Integrated Approach to Services", IMIESA,
April.
Burton, C. and Michael, N. 1992. A Practical Guide to Project Management: How to Make it Work in your
Organisation, London, Kogan Page.
Child, J. 1984. Organization: A Guide to Problems and Practice, Second Edition, London, Paul Chapman
Publishing.
Conyers, D. and Hills, P. 1984. An Introduction to Development Planning in the Third World, Chichester, John
Wiley & Sons.
Dessler, G. 1986. Organization Theory: Integrating Structure and Behaviour, Second Edition, Englewood Cliffs,
Prentice Hall.
Du Mhango, G. 1993. "Financial and Socio-Economic Analysis", Paper presented at World Bank Conference,
Port Elizabeth.
Fourie, D. 1997. "Project Evaluation in Practice", Paper presented at SAIMAS Conference. Port Elizabeth.
Giesen, S. 1996. "Project Management", Fred Pryor Seminar, Port Elizabeth, 28 October.
Goodman, L.J. and Love, R.N. 1980. Project Planning and Management: An Integrated Approach, New York,
Pergamon Press.
Harrison, F.L. 1992. Advanced Project Management: A Structured Approach, Third Edition, Aldershot, Gower
Publishing.
Hazlewood, L.R. 1995. "Designing Fundable Development Projects for the RDP", Munisipale en Openbare
Dienste, March.
Kerzner, H. 1992.. Project Management: A Systems Approach to Planning, Scheduling and Controlling, Fourth
Edition, New York, Van Nostrand Reinholdt.
Lichfield, N. 1975. Evaluation in the Planning Process, Oxford, Pergamon Press.
Lister, G. 1995. "How to Manage Projects in the New South Africa", Project Pro, Vol. 5, No. 6, November.
56
Meiring, M.H. and Parsons, C.R. 1994. Fundamental Public Administration: A Guide for a Changing South
Africa, Port Elizabeth, University of Port Elizabeth.
Oosthuizen, P. 1994. The Silent Revolution: Project Management ... How to Make Business Work, Arcadia, PM
Publishers.
Reiss, G. 1992. Project Management Demystified: Today's Tools and Techniques, London, E & FN Spon.
Republic of South Africa. 1994. White Paper on Reconstruction and Development, Pretoria, Government Printer.
Rosenau, M.D. 1992. Successful Project Management: A Step-by-Step Approach with Practical Examples,
Second Edition, New York, Van Nostrand Reinhold.
Shaghil, M. and Mushtaque, M. 1993. Project Planning and Management, Jaipur, Rupa.
Stallworthy, E.A. and Kharbanda, O.P. 1983. Total Project Management: From Concept to Completion,
Aldershot, Gower Publishing.
United Nations. 1971. Administration of Development Programmes and Projects: Some Major Issues, New York,
United Nations.
Van der Walt, G. and du Toit, D.F.P. 1997. Managing for Excellence in the Public Sector, Kenwyn, Juta.
Wallis, M. 1996. "Local Government and Development Planning in South Africa", in Reddy, P.S. (ed) Readings
in Local Government Management and Development: A Southern African Perspective, Kenwyn, Juta.
World Bank. 1993. "World Bank Project Management for Local Government", World Bank Conference, Port
Elizabeth, 29 November - 10 December.
World Bank. 1996. "RDP Picks up Speed in the Eastern Cape", Local Government Digest, August.
World Bank. 1995. "R850 million for RDP Municipal Projects", Munisipale en Openbare Dienste, Vol. 15, No. 4,
November.
World Bank. 1994. "Managing the RDP", Project Pro, Vol. 4, No. 6, November.
H J Nel , PROJECT MANAGEMENT FOR DEVELOPMENT: THE ROLE OF THE
PUBLIC MANAGER,
http://www.up.ac.za/academic/soba/SAAPAM/administratio%20publica/vol8no2/nel.htm
57
Project Planning
Project planning has several sub phases. It starts with the identification of possible projects from established
sources of ideas (e.g., central planning agency, NGOs, international financial institutions, corporate planning
division of multinational corporations). A government may contract a pool of experts or specialists to determine
the priority needs of a locality. By engaging the community in a discussion of their most pressing concerns, the
team may zero in on the most suitable project proposal. The project idea then becomes formally formulated and
proceeds to the next level. This could be aided through the articulation of lobby groups, the expression of interest
by target communities concerned, or the sponsorship of a local politician or policymaker. Project planning has
several sub phases. It starts with the identification of possible projects from established sources of ideas (e.g.,
central planning agency, NGOs, international financial institutions, corporate planning division of multinational
corporations). A government may contract a pool of experts or specialists to determine the priority needs of a
locality. By engaging the community in a discussion of their most pressing concerns, the team may zero in on the
most suitable project proposal. The project idea then becomes formally formulated and proceeds to the next level.
This could be aided through the articulation of lobby groups, the expression of interest by target communities
concerned, or the sponsorship of a local politician or policymaker.
“Formulation” calls for “the identification of key factors vital to the project, such as objectives and focus.” At
this stage, it is important to clearly delineate and pinpoint the exact primary and secondary objectives or goals of
the proposed undertaking. Word should also be given on the breadth of the goals; what are the long-term,
medium-term and short-term goals. “Focus” refers to “specified courses of action that the project intend to take
to achieve the stated targets or objectives.”
Since there would be competing projects for scarce funding, the project must have a high-priority status. Priority
is a function of timing and location. One project, for example, may be of highest priority to a residential area but
not for a commercial district or a project could be of extreme necessity for one area at one point in time but not
in another.
Long-term viability is another concern. Project formulation should take into account the full coverage period of
the project and the factors that could possibly affect the project’s performance or success. One factor that could
seriously affect a project’s viability in the long-run is the availability of resources throughout the project lifespan.
Support by politicians, and, most especially, the intended community beneficiaries could also spell a big
difference. Viability is very important since most projects have long-term results and benefits, thus, their
sustainability and continuity in the long-run should be maintained and strengthened.
Since most projects are not created in isolation but as part of a bigger program, which in turn heeds a certain
public policy, its impact on other projects should also be taken into consideration. The interrelations and
interactions of each project should be carefully thought out. One project may, for example, serve as a catalyst for
an other or one project may not be feasible unless another project is implemented. The harmful effects of these
possible scenarios could be mitigated by proper project timing.
Another important factor in project selection is the effect of the project on the national position or prestige of a
government, both local and international; the higher the tendency of a project to promote the government’s
position, the greater the chances of the project being approved, sustained and completed. Conversely, if the
project has a great possibility of resulting in failure, which could have adverse effects on the prestige of a
national government, it would have a poor chance of being selected.
After identifying one project as the best among the competing project proposals, project formulation proceeds to
the analysis of necessary economic (i.e., funding), technical (i.e., consultancy, manpower, facilities, land,
equipment), managerial (i.e. project managers, technical personnel or experts/analysts) and other related support
systems (i.e. political, administrative). This could be done by using existing resources, assuring that such support
systems critical for the project would be made available in the progress of the project or, lastly, by creating or
developing such resource systems right at the start. Delays in the delivery or availability of these support systems
could greatly handicap or derail the project. Another concern relative to support systems is the presence of
support structures that may come from the community or the policymakers and policy implementors.
58
Lastly, a redefinition of the project shall be made in light of the results of the identification of key project factors
and analysis of the project’s support systems. Redefinition would concentrate on the alternatives being proposed
to address the identified inadequacies or lapses on the last two steps mentioned. Depending on the scale of the
revision, a reformulation of the project may be called for. The product of reformulation would be a project
statement which would then serve as a sound basis for the project’s preliminary design, which is supposed to
have the following elements: (1) a list of project objectives; (2) an overview of the project with given
alternatives; (3) the technical elements of the project; (4) the project’s management plan, including operating and
implementing strategy and, lastly, (5) a budget. A detailed multi-aspect feasibility analysis could then be started
at this juncture.
Other project management scholars posit present slight variations in project planning. Franco et al. (1997, 24-27)
outlined the steps in reaching a good project proposal as follows: participation analysis; problem and opportunity
assessment and project identification; project definition; screening and; finalization of the project concept.
It is incumbent upon the project planners and managers to come up with a detailed and cost-factored activity
plan and timetables. After finishing the laborious feasibility study, the proposal would then be in the hands of the
funding agency. If it passed the pre-evaluation and appraisal and the rounds of project proponent-funding
organization negotiations, it is then poised for implementation. (Franco et al, 1997, 27-30)
Anti–Poverty Program/Project Administration 3
References
Goodman, Louis J. and Ralph N. Love (Eds.), Project Planning and Management: An Integrated Approach.
Pergamon Press, 1980, pp. 50-59
Franco, Ernesto A. et al. 1997, Project Management for Social and Economic Development., Manila, Anvil
Publishing Inc., pp. 24-30
59
Project Implementation
Implementation consists of (1) breaking down the project into discrete component tasks, operations or activities,
(2) establishing an organization and staff to conduct the actual implementation procedures (3) directing the
execution by assigning tasks and activities to groups or teams within the project organization and procuring and
allocating the necessary resources and (4) coordinating, monitoring and control of the performance of
responsible groups, optimum use of project resources and the completion of the project through an orderly and
timely fashion (Goodman and Love. 1980: 135).
In developing countries as borrowers, they are responsible for coming up with the implementation plan and the
crafting of the project management organization, although lending or financing institutions (i.e. Asian
Development Bank, World Bank) sometimes do extend technical assistance at this phase. Much is expected from
the project manager and the entire organization under him. They are accountable for the expenditures in the form
of procurement or workers’ salaries and other disbursements in the course of the undertaking, as well as in
clarifying which funds would come from the ADB or WB and which ones would be identified as government
counterpart funds. The organization is also mandated to submit their progress reports, particularly the project’s
financial standing, to the funding agencies and be open for any inquiry in their work performance.
Budgeting and cost estimation is one essential step in project implementation. There is a wide range of budgeting
techniques or methods that a project implementer or manager can choose from, among which are the top-down,
bottom-up, budget request, planning-programming budgeting system (PPBS) and zero-based budgeting (ZBB).
Each and every one of them has certain peculiar advantages, as well as disadvantages, thus caution should be
made in choosing the most suitable for a specific type of project. For purposes of cost estimation, it would be of
extreme aid if there would be an available form identifying the level of resource need, at what specific time will
it be most demanded and its availability status. (Meredith and Mantel, 1995:315-316).
Additional factors that should be looked into would be inflation, differential changes in cost factors, waste,
personnel replacement costs and contingencies for unexpected eventualities (Meredith and Mantel, 1995: 316).
Task scheduling is another important facet of project implementation. It calls for the appropriate sequencing of
project activities or the maintenance of a logical flow of work. The more realistic the timetable and the more
honest it is when it comes to divulging serious time constraints, the more reliable it becomes. The success of
scheduling is also a function of the acceptability of the functional organization which would use it for assessing
the tempo, as well as the quality, of work of the whole undertaking (Meredith and Mantel, 1995: 371-372).
The concern of resource allocation, on the other hand, would be to determine the best trade-offs between
resources at hand, including time, throughout the course of the project implementation. Aside from allocation,
the project’s scarce or limited resources are also carefully loaded and leveled. defined resource loading as “the
process of calculating the total load from project tasks on each resource for each time period of the project’
duration” while resource leveling is “concerned with evening out the demands for various resources required in a
project by shifting tasks within their slack allowances” (Meredith and Mantel , 1995:423).
Project implementers can choose from a wide range of existing mechanisms in planning and scheduling their
resource allocation and monitoring their work’s progress in the process. They can opt for the traditional
techniques, such as the Project Breakdown Structure (PBS), Gantt charts and Line of Balance Method (LOB) to
the more sophisticated Critical Path Method (CPM), Performance Evaluation and Review Technique (PERT) and
other recent network scheduling. These supervisory techniques are important in that they ensure the smooth flow
of the project, detect deviations and make timely rectifications.
A manual or guideline for operations is suggested by Franco as an indispensable item at this stage. In this
cardinal document, powers and responsibilities of team members are clearly spelled out to avoid confusion and
60
wasteful functional overlaps. This is most true for complex projects with huge funding. Orienting all project
actors could also be a plus factor (Franco et al, 1997, p. 285). The use of computer programs or softwares
throughout the project implementation, and even in project monitoring and evaluation, may prove to be of big
help.
References
Goodman, Louis J. and Ralph N. Love, (Eds) 1980, Project Planning and Management: An Integrated Approach.
Pergamon Press, pp. 135-182
Franco, Ernesto A. et al., l997, Project Management for Social and Economic Development. Manila: Anvil
Publishing Inc., pp. 30-31; 285
Meredith, Jack R. & Samuel J. Mantel Jr., 1995, Project Management: A Managerial Approach 3rd ed.
61
Project Monitoring and Evaluation
In a normal project cycle, “monitoring and evaluation” is usually the last stage, notwithstanding the fact that
evaluation and monitoring mechanisms could also be specifically designed or built to check the progress of the
project from programming to implementation and handover to make timely corrections. Its basic purpose is to
ascertain if the desired objectives or goals laid out in the project formulation and planning were met. In case the
intervention failed to address the issue, monitoring and evaluation also look into the reasons behind such failure.
That this stage is the last in the project cycle also contributes to the low priority being given to it by
many. The project implementation phase often gets the bigger attention and recently project planning
began to challenge this traditional position. The unconscious, if not deliberate, neglect of project
evaluation may be sociopolitical in nature. Some politicians may not want the public to know how the
funds were used or where these funds went. In this scenario, induced amnesia has been the resort to
evade questions arising from accountability and transparency. In some instances, the neglect may be
due in part to the absence of a competent team or staff to conduct the evaluation procedures or the low
esteem given by the project planners to evaluation. Although not deliberate, this narrowness should not
be considered as a reasonable excuse.
The importance of evaluation and monitoring is highlighted, particularly in less developed countries,
by the scarcity of resources which offers no room for wastages and unnecessary expenditures.
The Organization for Economic Cooperation and Development (OECD) defined “project evaluation”
as “the attempt to assess the results of an activity, as a function of the results, of the means employed to
achieve them”. Here “effectiveness” pertains to “results,” while “efficiency” refers to “means”.
“Effectiveness” concerns meeting objectives, while “efficiency” refers to the “mobilization and
organization of resources necessary to carry out the project’s aims at the least cost.” The latter could be
expressed in terms of capital-output ratios or output per working hour. Different NGOs, POs, aid
agencies of First World governments and international lending or financing institutions (e.g,, World
Bank, Asian Development Bank) come up with their own concepts or orientations in project evaluation,
but all of them share the fundamentals.
Evaluation, by nature, takes a comprehensive and all-encompassing perspective. The success or failure
of the whole intervention, not just one or two aspects of it, is measured against its outlined objectives.
The results of project monitoring and evaluation (particularly post-project evaluation) provide
reference that is of great use for project planners or policymakers in the future, especially if they are to
undertake a similar action. In the case of evaluation during planning or implementation stage, the
lessons or observations would serve as the basis for appropriate and timely modifications.
There are a number of evaluation approaches that can be of use for project evaluators.
Koppel(1976,8-13) gave three of these: evaluating objectives, evaluating subjectives and evaluating
levels. Evaluating objectives is the most popular among the trio. But objectives are not restricted to
those which are written explicitly, as implicit objectives are also looked into. This evaluation approach
also gives due attention to inputs, formations, outputs and effects. Evaluation by subjective, on the
other hand, focuses on the process itself and on the questions “who benefited”, “who decided” and
“who paid”. Lastly, evaluating by levels make a distinction between level of activity (i.e. program,
projects, and activities) and identified objectives per level of activity.
Consistency is one project management guideline that could greatly influence project evaluation.
“Consistency” refers to “both form and substance of project documentation and procedures.” A
62
databank equipped with up-to-date records could speed up the evaluator’s work. Formalizing uniform
evaluation rules and procedures could in sustaining and enhancing projects.
There is a wide range of evaluation techniques that evaluators could choose from. One of the popularly
subscribed to is the social cost-benefit analysis technique. The terms “cost” and “benefit” are not
restricted to their usual meanings but are expanded to expressed “opportunity losses” (or “opportunity
costs”) and “prices of social goods”. After coming up with a social rate of discount that will serve as
the cut-off point, projects offering the highest rate of return would be deemed “acceptable,” while those
who fall below the benchmark would be considered “unacceptable.” Other techniques used in
evaluation would be the resort to control groups for data gathering, comparative analysis, baseline
measures and sampling. Quantitative financial techniques could come in the form of variance analysis.
There is a raging contention on which evaluation technique is more “superior”- the quantitative, which
is branded as “scientific” and “systematic” or the qualitative, which is argued as most apt for difficult
to quantify costs and benefits and is best represented by the case study.
The orientation and composition of the evaluators is vital for the fulfillment of project evaluation, as
they are mirrors of the evaluation itself. The common scenario for a typical government undertaking is
that the central or national planning agency is the one who will carry out the evaluation. In the private
sector, its counterpart would be the corporate planning department or division. This arrangement is
called “in-house” or “internal evaluation.” The opposite of this is the “external team”- an independent
private team of experts and consultants who bring with them their own evaluation frameworks and has
less structured.
The evaluation team, internal or external, is usually composed of individuals who come from
multidisciplinary backgrounds and has a certain degree of autonomy from the project implementers. At
a minimum, the team is usually composed of a technical expert and a project management expert.
Candidates may come from the academe or research institutions and the private sector.
There are different implications in using an internal team as against an external team. Internal teams
have familiarity and access to almost every bit of information, although some contest the objectivity of
its findings. External teams, on the other hand, are less familiar with the intricacies of the work, less
access to background data, and are utilizing different evaluation perspectives and tools that may not
necessarily be pertinent to the project. Nevertheless, external teams are perceived to furnish more
“objective” results. To get the best of both kinds of teams, some projects utilize an optimum mix of
internal and external staff to facilitate evaluation. For an inter-agency government project, the
evaluators may come from the representatives of each agency concerned.
The evaluation plan should contain the criteria to be used, evaluation technique, time frame for the
conduct of the evaluation, budget, and organization of staff requirements. Participation of the intended
beneficiaries in the actual evaluation, as well as in the other phases of the project cycle, is also a vital
element that could determine the success of the project.
In terms of measuring the impact of a project, it is best to design a longitudinal study since it may take,
on a minimum, five years to observe the real impacts of the project. This is very true for socio-civic
interventions whose “returns of investment” may take longer to become evident.
63
References
Project Planning and Management: An Integrated Approach. Ed. By Goodman, Louis J. and Ralph N. Love.
Pergamon Press, 1980, pp. 213-232.
Koppel, Bruce 1976. The Evaluation Factor: A Handbook to remind Evaluators of the Complexity of their Task,
Honolulu: East West Center, pp. 8-13.
64
Concepts and Techniques for Impact Evaluation
Baker, Judy L. (May 2000). Evaluating the Impact of Development Projects on Poverty A Handbook for
Practitioners (Chapters 1 to 3). The International Bank for Reconstruction and Development/THE WORLD
BANK, Washington, D.C. (with minor editing)
A “comprehensive evaluation” is an “evaluation that includes monitoring, process evaluation, cost-benefit
evaluation, and impact evaluation.” “Monitoring” will help to assess whether a program is being implemented
as was planned. A program monitoring system allows continuous feedback on the status of program
implementation. Process evaluation is concerned with how the program operates and focuses on problems in
service delivery. “Cost-benefit” or “cost-effectiveness evaluations” assess program costs (monetary or non
monetary), in particular their relation to alternative uses of the same resources and to the benefits being produced
by the program. “Impact evaluation” is intended to determine more broadly whether the program had the desired
effects on individuals, households, and institutions and whether those effects are attributable to the program
intervention. Impact evaluations can also explore unintended consequences,
whether positive or negative, on beneficiaries.
Some of the questions addressed in impact evaluation are: How did the project affect the beneficiaries? Were any
improvements a direct result of the project, or would they have improved anyway? Could program design be
modified to improve impact? Were the costs justified? Other factors or events besides the outcome of a project
may be correlated with the outcomes but are not caused by the project. To ensure methodological
rigor, an impact evaluation must estimate the counterfactual (what would have happened had the project never
taken place or what otherwise would have been true). This is accomplished through the use of comparison or
control groups (those who do not participate in a program or receive benefits), which are subsequently compared
with the
treatment group (individuals who do receive the intervention). Control groups are selected randomly from the
same population as the program participants, whereas the “comparison group” is the “group that does not receive
the program under investigation.” Both the comparison and control groups should be comparable with the
treatment group in every way, with the only difference between groups being program participation.
Determining the counterfactual can be accomplished using several methodologies which fall into two broad
categories, “experimental designs” (randomized), and “quasi-experimental” designs (nonrandomized). It is
difficult to net out the program impact from the counterfactual conditions which can be affected by history,
selection bias, and contamination. Qualitative and participatory methods can also be used to assess impact. These
techniques often provide critical insights into beneficiaries' perspectives, the value of programs to beneficiaries,
the processes that may have affected outcomes, and a deeper interpretation of results observed in quantitative
analysis. The strengths and weaknesses of each of these methods are discussed in more detail below. No
technique is perfect and the evaluator must make decisions about the tradeoffs for each method chosen. Early
and careful planning will, however, provide more methodological options in designing the evaluation.
Experimental Designs
“Experimental designs,” also known as “randomization,” are generally considered the most robust of the
evaluation methodologies. By randomly allocating the intervention among eligible beneficiaries, the assignment
process creates comparable treatment and control groups that are statistically equivalent to one another, given
appropriate sample sizes. This is a very powerful outcome because, in theory, the “control groups” generated
through random assignment serve as a perfect counterfactual, free from the selection bias issues. The main
benefit of this technique is the simplicity in interpreting results-the program impact on the outcome being
evaluated can be measured by the difference between the means of the samples of the treatment group and the
control group.
65
While experimental designs are considered the optimum approach to estimating project impact, in practice there
are several problems. First, randomization may be unethical owing to the denial of benefits or services to
otherwise eligible members of the population for the purposes of the study. An extreme example would be the
denial of medical treatment that can turn out to be lifesaving to some members of a population. Second, it can be
politically difficult to provide an intervention to one group and not another. Third, the scope of a program may
mean that there are no non treatment groups such as with a project or policy change that is broad in scope
-examples include an adjustment loan or programs administered at a national level. Fourth, individuals in control
groups may change identifying characteristics during the experiment that could invalidate or contaminate the
results. If people move in and out of a project area, they may move in and out of the treatment or control group.
Alternatively, people denied of a program benefit may seek it through alternative sources, or those being
offered a program
may not take up the intervention. Fifth, it may be difficult to ensure that assignment is truly random. And, finally,
experimental designs can be expensive and time consuming in certain situations, particularly in the collection of
new data.
With careful planning, some problems can be addressed in the implementation of experimental designs, as in the
random selection of beneficiaries, to provide both a politically transparent allocation mechanism and the basis of
a sound evaluation design, as budget or information constraints often make it impossible to accurately identify
and reach the most eligible beneficiaries. A second way is bringing control groups into the program at a later
stage once the evaluation has been designed and initiated. The random selection determines when the eligible
beneficiary receives the program, not if they receive it. This was done in the evaluation of a nutrition program in
Colombia, which provided the additional benefit of addressing questions regarding the necessary time involved
for the program to become effective in reducing malnutrition (McKay 1978). Finally, randomization can be
applied within a subset of equally eligible beneficiaries, while reaching all of the most eligible and denying
benefits to the least eligible, as was done with education projects in the El Chaco region for the Bolivia social
fund evaluation (Pradhan, Rawlings, and Ridder 1998). However, if the latter suggestion is implemented, the
results produced from the evaluation will be applicable to the group from which the randomly generated sample
was selected.
Quasi-Experimental Designs
Quasi-experimental (nonrandom) methods can be used to carry out an evaluation when no treatment and
comparison groups can be selected through experimental design. These techniques generate comparison groups
that resemble the treatment group, at least in observed characteristics, through econometric methodologies,
which include matching methods, double difference methods, instrumental variables methods, and reflexive
comparisons (see Box 1.2). When these techniques are used, the treatment and comparison groups are usually
selected after the intervention by using nonrandom methods. Therefore, statistical controls must be applied to
address differences between the treatment and comparison groups and sophisticated matching techniques must
be used to construct a comparison group that is as similar as possible to the treatment group. A comparison group
is also chosen before the treatment, though the selection is not randomized. The main benefit of
quasi-experimental designs is that they can draw on existing data sources and are thus often quicker and cheaper
to implement, and they can be performed after a program has been implemented, given sufficient existing data.
The principal disadvantages of quasi experimental techniques are that (a) the reliability of the results is often
reduced as the methodology is less robust statistically; (b) the methods can be statistically complex; and (c) there
is a problem of selection bias. In generating a comparison group rather than randomly assigning one, many
factors can affect the reliability of results. Statistical complexity requires expertise in the design of the evaluation
and in analysis and interpretation of the results, which may not always be possible.
The third problem of bias relates to the extent to which a program is participated in by subgroups of a target
population, thus affecting the sample and the results. There are two types of bias: those due to differences in
observables or something in the data, and those due to differences in unobservables (not in the data), often called
selection bias. An observable bias may include the selection criteria through which an individual is targeted, such
66
as geographic location, school attendance, or participation in the labor market. Unobservables that may bias
program outcomes could include individual ability, willingness to work, family connections, and a subjective
(often politically driven) process of selecting individuals for a program. Both types of biases can yield inaccurate
results, including under- and overestimates of actual program impacts, negative impacts when actual program
impacts are positive (and vice versa), and statistically insignificant impacts when actual program impacts are
significant and vice versa. (See, for example, LaLonde 1986, Fraker and Maynard 1987, LaLonde and Maynard
1987, and Friedlander and Robins 1995.) It is possible to control for bias through statistical techniques such as
matching and instrumental variables, but it is very difficult to fully remove them which remains a major
challenge for researchers in the field of impact analysis. Among quasi-experimental design techniques,
matched-comparison techniques are generally considered a second-best alternative to experimental design. The
majority of the literature on evaluation methodology is centered around the use of this type of evaluation,
reflecting both the frequency of use of matched comparisons and the many challenges posed by having
less-than-ideal comparison groups. In recent years there have been substantial advances in propensity score
matching techniques (Rosenbaum and Rubin 1985; Jalan and Ravallion 1998).
Box 1.1 The Problem of Selection Bias
Selection bias relates to unobservables that may bias outcomes (for example, individual ability, preexisting conditions).
Randomized experiments solve the problem of selection bias by generating an experimental control group of people who
would have participated
in a program but who were randomly denied access to the program or treatment. The random assignment does not remove
selection bias but instead balances the bias between the participant and non participants samples. In quasi-experimental
designs, statistical models (for example, matching, double differences, instrumental variables) approach this by modeling
the selection processes to arrive at an unbiased estimate using nonexperimental data. The general idea is to compare
program participants and non participants holding selection processes constant. The validity of this model depends on
how well the model is specified.
A good example is the wages of women. The data represent women who choose to work. If this decision were
made, we could ignore the fact that not all wages are observed and use ordinary regression to estimate a wage
model. Yet the decision by women to work is not made randomly-women who would have low wages may be
unlikely to choose to work because their personal reservation wage is greater than the wage offered by
employers. Thus the sample of observed wages for women would be biased upward. This can be corrected for if
there are some variables that strongly affect the chances for observation (the reservation wage) but not the
outcome under study (the offer wage). Such a variable might be the number of children at home.
Source: Greene (1997).
This method is very appealing to evaluators with time constraints and working without the benefit of baseline
data given that it can be used with a single crosssection of data. This technique is, however, dependent on having
the right data because it relies on over sampling program beneficiaries during the fielding of a larger survey and
then "matching" them to a comparison group selected from the larger core sample of the overall effort, often a
national household survey. Given the growth in the applications of large surveys in developing countries, such as
the multipurpose Living Standards Measurement Studies, this evaluation method seems particularly promising. A
good example is the evaluation of a public works program, TRABAJAR, in Argentina (Jalan and Ravallion 1998,
Annex 1.1, and chapter 4).
67
Box 1.2 Summary of Quantitative Methods for Evaluating Program Impact
The main methods for impact evaluation are discussed below. Because no method is perfect, it is always desirable to
triangulate.
Experimental or Randomized Control Designs
* Randomization, in which the selection into the treatment and control groups is random within some well-defined set of
people. In this case there should be no difference (in expectation) between the two groups besides the fact that the treatment
group had access to the program. (There can still be differences due to sampling error; the larger the size of the treatment
and control samples the less the error.)
Nonexperimental or Quasi-Experimental Designs
* Matching methods or constructed controls, in which one tries to pick an ideal comparison that matches the treatment
group from a larger survey. The most widely used type of matching is propensity score matching, in which the comparison
group is matched to the treatment group on the basis of a set of observed characteristics or by using the "propensity score"
(predicted probability of participation given observed characteristics); the closer the propensity score, the better the match.
A good comparison group comes from the same economic environment and was administered the same questionnaire by
similarly trained interviewers as the treatment group.
* Double difference o r difference-in-differencemse thods, in which one compares a treatment and comparison group (first
difference) before and after a program (second difference). Comparators should be dropped when propensity scores are used
and if they have scores outside the range observed for the treatment group.
* Instrumental variables or statistical control methods, in which one uses one or more variables that matter to participation
but not to outcomes given participation. This identifies the exogenous variation in outcomes attributable to the program,
recognizing that its placement is not random but purposive. The "instrumental variables" are first used to predict program
participation; then one sees how the outcome indicator varies with the predicted values.
* Reflexivec omparisonsi,n which a baseline survey of participants is done before the intervention and a follow-up survey is
done after. The baseline provides the comparison group, and impact is measured by the change in outcome indicators before
and after the intervention.
Qualitative Methods
Qualitative techniques are also used for carrying out impact evaluation to determine impact by the reliance on
something other than the counterfactual to make a causal inference (Mohr 1995). The focus is on understanding
processes, behaviors, and conditions as they are perceived by the individuals or groups being studied (Valadez
and Bamberger 1994). For example, qualitative methods and, particularly, participant observation can provide
insight into the ways in which households and local communities perceive a project and how they are affected by
it. Because measuring the counterfactual is at the core of impact analysis techniques, qualitative designs have
generally been used in conjunction with other evaluation techniques. The qualitative approach uses relatively
open-ended methods during design, collection of data, and analysis. Qualitative data can also be quantified.
Among the methodologies used in qualitative impact assessments are the techniques developed for rapid rural
assessment, which rely on participants' knowledge of the conditions surrounding the project or program being
evaluated, or participatory evaluations in which stakeholders are involved in all stages of the
evaluation-determining the objectives of the study, identifying and selecting indicators to be used, and
participating in data collection and analysis. For a detailed discussion on participatory methods see World Bank
(1996), The World Bank Participation Sourcebook.
The benefits of qualitative assessments are that they are flexible, can be specifically tailored to the needs of the
evaluation using open-ended approaches, can be carried out quickly using rapid techniques, and can greatly
enhance the findings of an impact evaluation through providing a better understanding of stakeholders'
68
perceptions and priorities and the conditions and processes that may have affected program impact. Among the
main drawbacks are the subjectivity involved in data collection, the lack of a comparison group, and the lack of
statistical robustness, given mainly small sample sizes, all of which make it difficult to generalize to a larger,
representative population. The validity and reliability of qualitative data are highly dependent on the
methodological skill, sensitivity, and training of the evaluator. If field staff are not sensitive to specific social and
cultural norms and practices, and nonverbal messages, the data collected may be misinterpreted. And finally,
without a comparison group, it is impossible to determine the counterfactual and thus causality of project impact.
Integrating Quantitative and Qualitative Methods
There is a growing acceptance of the need for integrating the two approaches. Impact evaluations using
quantitative data from statistically representative samples are better suited to assessing causality by using
econometric methods or reaching generalizable conclusions. However, qualitative methods allow the in depth
study of selected issues, cases, or events and can provide critical insights into beneficiaries' perspectives, the
dynamics of a particular reform, or the reasons behind certain results observed in a quantitative analysis.
Integrating quantitative and qualitative evaluations can often be the best vehicle for meeting the project's
information needs. In combining the two approaches, qualitative methods can be used to inform the key impact
evaluation questions, survey the questionnaire or the stratification of the quantitative sample, and analyze the
social, economic, and political context within which a project takes place, whereas quantitative methods can be
used to inform qualitative data collection strategies, to design the sample to inform the extent to which the
results observed in the qualitative work can be generalized to a larger population by using a statistically
representative sample, and, statistical analysis can be used to control for household characteristics and the
socio-economic conditions of different study areas, thereby eliminating alternative explanations of the observed
outcomes. There are several benefits of using integrated approaches in research (Bamberger, 2000) that also
apply to impact evaluations. Among them:
•Consistency checks can be built in through the use of triangulation procedures that permit two or more
independent estimates to be made for key variables (such as income, opinions about projects, reasons for
using or not using public services, and specific impact of a project).
•Different perspectives can be obtained. For example, although researchers may consider income or
consumption to be the key indicators of household welfare, case studies may reveal that women are
more concerned about vulnerability (defined as the lack of access to social support systems in times of
crises), powerlessness, or exposure to violence.
•Analysis can be conducted on different levels. Survey methods can provide good estimates of individual,
household, and community level welfare, but they are much less effective for analyzing social processes
(social conflict, reasons for using or not using services, and so on) or for institutional analysis (how
effectively health, education, credit, and other services operate and how they are perceived by the
community). There are many qualitative methods designed to analyze issues such as social process,
institutional behavior, social structure, and conflict.
•Opportunities can be provided for feedback to help interpret findings. Survey reports frequently include
references to apparent inconsistencies in findings or to interesting differences between communities or
groups that cannot be explained by the data. In most quantitative research, once the data collection phase
is completed it is not possible to return to the field to check on such questions. The greater flexibility of
qualitative research means that it is often possible to return to the field to gather additional data. Survey
researchers also use qualitative methods to check on outliers-responses that diverge from the general
patterns. In many cases the data analyst has to make an arbitrary decision as to whether a household or
community that reports conditions that are significantly above or below the norm should be excluded (on
the assumption that it reflects a reporting error) or the figures adjusted.
Qualitative methods permit a rapid follow-up in the field to check on these cases. In practice, the integration of
quantitative and qualitative methods should be carried out during each step of the impact evaluation. For
69
illustration, the Nicaragua School Autonomy Reform Case provides a good example of integrated methods.
Quantitative methods following a quasi-experimental design were used to determine the relationship between
decentralized management and learning and to generalize results for different types of schools. In addition,
qualitative techniques, including a series of key informant interviews and focus group discussions with different
school based staff and parents, were utilized to analyze the context in which the reform was introduced, examine
the decision making dynamics in each school, and assess the perspectives of different school community actors
on the autonomy process (see Annex 1.11).
Other Approaches to Impact Evaluation
Two other topics are particularly relevant to the discussion of evaluating the poverty impact of projects: (a)
approaches to measuring the impact of structural adjustment programs, and (b) theory-based evaluations.
Evaluating Structural Adjustment Programs. There has been substantial debate on the impact of structural
adjustment programs on the poor. Much of the evidence used to support this debate is, however, based on
deficient assumptions and methods. As with other projects, the policy changes under structural adjustment
projects must be (a) compared with relevant counterfactuals that would respond to the same macroeconomic
constraints, and (b) analyzed in the context of the local economic structure and based on empirical information
from household surveys. This, however, is very difficult for three reasons. First, policy changes may have
economy-wide impact, making it impossible to find comparison groups that are unaffected. Second, because of
exogenous factors, lags, feedbacks, and substitutions, any changes in the well-being of the poor must be
interpreted with extreme caution. And, third, it is difficult to predict what would have happened if adjustment
had not taken place, what alternative policies a government might have pursued and what the resulting impact
would have been on the poor. In the literature, several approaches have been used, each with its own
shortcomings. The techniques are in many cases similar to those described in Box 1.2, though, as shown in Box
1.3, estimating the counterfactual requires vast assumptions that may substantially affect the validity of the
results. This is most viably handled by isolating specific policy changes that would affect the population, such as
exchange rate policies, trade policies, reductions in public expenditures, and reductions in public sector
employment. Yet even with this approach it can be difficult to isolate the impact of specific policies. For
examples, see Killick (1995), Poppele, Summarto, and Pritchett (1999), Bourguignon, de Melo, and Suwa (1991),
and Sahn, Dorosh, and Younger (1996).
Box 1.3 Summary of Methods Used to Evaluate Adjustment Policies
Approaches with No Counterfactual
* Qualitative studies that assess conditions of the population (often identifying vulnerable subgroups) before, during, and
after adjustment policies are implemented through focus groups, interviews, and other qualitative techniques.
* "Before and After," which compares the performance of key variables during and after a program with those prior to the
program. The approach uses statistical methods to evaluate whether there is a significant change in some essential variables
over time. This approach often gives biased results because it assumes that had it not been for the program, the performance
indicators would have taken their pre-crisis-period values.
Approaches that Generate a Counterfactual Using Multiple Assumptions
* Computable general equilibrium models (CGEs) that attempt to contrast outcomes in treatment and comparison groups
through simulations. These models seek to trace the operation of the real economy and are generally based on detailed social
accounting matrices collected from data on national accounts, household expenditure surveys, and other survey data. CGE
models do produce outcomes for the counterfactual, though the strength of the model is entirely dependent on the validity of
the assumptions. This can be problematic as databases are often incomplete and many of the parameters have not been
estimated by formal econometric methods. CGE models are also very time consuming, cumbersome, and expensive to
generate.
* With and without comparisons, which compare the behavior in key variables in a sample of program countries with their
70
behavior in nonprogram countries (a comparison group). This is an approach to the counterfactual question, using the
experiences of the comparison group as a proxy for what would otherwise have happened in the program countries. It is,
however, quite difficult to achieve a true comparison group. The method assumes that only the adoption of an adjustment
program distinguishes a program country from the comparison group and that the external environment affects both groups
the same.
* Statistical controls consisting of regressions that control for the differences in initial conditions and policies undertaken in
program and nonprogram countries. The approach identifies the differences between program and nonprogram countries in
the preprogram period and then controls these differences statistically to identify the isolated impacts of the programs in the
postreform performance.
Theory-Based Evaluation. The premise of theory-based evaluations is that programs and projects are based on
explicit or implicit theory about how and why a program will work. The evaluation would then be based on
assessing each theory and assumptions about a program during implementation rather than at a midpoint or after
the project has been completed. In designing the evaluation, the underlying theory is presented as many
microsteps, with the methods then constructed for data collection and analysis to track the unfolding of
assumptions. If events do not work out as expected, the evaluation can say with a certain confidence where, why,
and how the breakdown occurred. The approach puts emphasis on the responses of people to program activities.
Theories direct the evaluator's attention to likely types of near-term and longer-term effects. Among the
advantages are, first, that the evaluation provides early indications of program effectiveness during project
implementation. If there are breakdowns during implementation, it is possible to fix them along the way. Second,
the approach helps to explain how and why effects occurred. If events work out as expected, the evaluation can
say with a certain confidence how the effects were generated. By following the sequence of stages, it is possible
to track the microsteps that led from program inputs through to outcomes. The shortcomings of the approach are
similar to many of the other methodologies. In particular, (a) identifying assumptions and theories can be
inherently complex; (b) evaluators may have problems in measuring each step unless the right instruments and
data are available, (c) problems may be encountered in testing the effort because theory statements may be too
general and loosely constructed to allow for clear-cut testing, and (d) there may be problems of interpretation
that make it difficult to generalize from results (see Weiss 1998). An example of theory-based technique is being
piloted by the Operations and Evaluation Department of the World Bank to evaluate the impact of social
investment funds on community-level decision making processes, traditional power structures and relationships,
and community capacity, trust, and well-being. This will be based on the theory that priority groups can
effectively implement a project and operate and maintain the investment created by the project. A set of main
assumptions and sub assumptions has been set out and will be tested using existing household survey data, as
well as a specially designed survey instrument for a smaller sample, and focus groups and other PRA techniques.
The information from each of these data sources will be triangulated in the analysis.
Cost-Benefit or Cost-Effectiveness Analysis
Cost-benefit or cost-effectiveness analysis enables policymakers to measure program efficiency by comparing
alternative interventions based on the cost of producing a given output. It can greatly enhance the policy
implications of the impact evaluation and therefore should also be included in the design of any impact
evaluation. (For a discussion of cost-benefit and cost-effectiveness analysis, see Handbook on Economic
Analysis of Investment Operations, World Bank 1996.)
Cost-benefit analysis attempts to measure the economic efficiency of program costs versus program benefits, in
monetary terms. For many projects, especially in the social sectors, it is not possible to measure all the benefits
in monetary terms. For example, the benefits of a program to provide school inputs (textbooks, classroom
furniture, preschool programs) would be increased learning. Instead of measuring monetary outcomes, learning
achievement scores could be used to quantify the benefits. This would require cost-effectiveness analysis. The
concepts for both types of analysis are the same.
The main steps of cost-benefit and cost-effectiveness analysis are to identify all project costs and benefits and
71
then compute a cost-to-effectiveness ratio. In calculating costs, the value of the intervention itself should be
included, as well as all other costs, such as administration, delivery, investment costs (discounted to the net
present value), the monetary value of freely provided goods or services, social costs such as environmental
deterioration, and health hazards. Benefits can be monetary, such as gain in income, or the number of units
delivered, test scores, or health improvements. When benefits cannot be quantified, it is possible to use
subjective indicators, such as ranking or weighting systems. This approach can be tricky in interpreting
subjective scores. Once the costs and benefits have been determined, the cost-effectiveness ratio (R) is then R =
cost/unit (or benefit). This ratio can then be compared across interventions to measure efficiency. In theory, this
technique is quite straightforward. In practice, there are many caveats in identifying and quantifying the costs
and benefits. It is important to ensure that appropriate indicators are selected, that the methodologies and
economic assumptions used are consistent across ratios, and that the ratios are indeed comparable. Measuring
cost-effectiveness can be best carried out when included in the evaluation design from the earliest stages. This
allows for the collection of the necessary cost and benefit information and ensuring consistency.
Choosing a Methodology
Given the variation in project types, evaluation questions, data availability, cost, time constraints, and country
circumstances, each impact evaluation study will be different and will require some combination of appropriate
methodologies, both quantitative and qualitative. The evaluator must carefully explore the methodological
options in designing the study, with the aim of producing the most robust results possible. Among quantitative
methods, experimental designs are considered the optimal approach and matched comparisons a second-best
alternative. Other techniques, however, can also produce reliable results, particularly with a good evaluation
design and high-quality data. The evidence from the "best-practice" evaluations reviewed for this handbook
highlights that the choice of impact evaluation methodologies is not mutually exclusive. Indeed, stronger
evaluations often combine methods to ensure robustness and to provide for contingencies in implementation.
Joining a "with and without" approach with a "before and after" approach that uses baseline and follow-up data
is one combination strongly recommended from a methodological perspective (Subbarao and others 1999).
Having baseline data available will allow evaluators to verify the integrity of treatment and comparison groups,
assess targeting, and prepare for a robust impact evaluation. This is true even for randomized control designs.
Although randomization ensures equivalent treatment and comparison groups at the time of randomization, this
feature should not influence evaluators into thinking that they do not need baseline data. Indeed, baseline data
may be crucial to reconstructing why certain events took place and controlling for these events in the impact
assessment.
Incorporating cost-benefit or cost-effectiveness analysis is also strongly recommended. This methodology can
enable policymakers to compare alternative interventions on the basis of the cost of producing a given output.
This is particularly important in the developing-country context in which resources are extremely limited. Finally,
combining quantitative and qualitative methods is the ideal because it will provide the quantifiable impact of a
project as well as an explanation of the processes and interventions that yielded these outcomes. Although each
impact evaluation will have unique characteristics requiring different methodological approaches, a few general
qualities of a best-practice impact evaluation include:
•An estimate of the counterfactual has been made by (a) using random assignment to create a control group
(experimental design), and (b) appropriately and carefully using other methods such as matching to create a
comparison group (quasi-experimental design).
•To control for pre- and postprogram differences in participants, and to establish program impacts, there are
relevant data collected at baseline and follow-up (including sufficient time frame to allow for program impacts).
•The treatment and comparison groups are of sufficient sizes to establish statistical inferences with minimal
attrition.
•Cost-benefit or cost-effectiveness analysis is included to measure project efficiency.
•Qualitative techniques are incorporated to allow for the triangulation of findings.
72
Key Steps in Designing and Implementing Impact Evaluations*
Undertaking an impact evaluation study can be challenging and costly, with implementation issues arising at
every step of the way. These challenges highlight the importance of a well-designed study, a committed and
highly qualified team, and good communication between the evaluation team members. By incorporating the
evaluation early into the design of a project, it will be possible to obtain results in a timely way so that the
findings can be used for mid project adjustments of specific components.
Regardless of the size, program type, or methodology used for the evaluation, there are several key steps to be
carried out as outlined below (Box 2.1). A discussion of these steps as well as a discussion of the many issues
that may arise in implementation is important. The sequencing of these steps is critical, particularly in ensuring
the collection of necessary data before the project begins implementation. Early planning provides the
opportunity to randomize, to construct ex ante matched comparisons, to collect baseline data, and to identify
upcoming surveys that could be used in a propensity score matching approach. All of the design work and initial
data collection should be done during project identification and preparation. Ideally, some results will be
available during the course of project implementation so they can feed into improving the project design if
necessary.
Determining Whether or Not to Carry Out an Evaluation
A first determination is whether or not an impact evaluation is required. As discussed above, impact evaluations
differ from other evaluations in that they are focused on assessing causality. Given the complexity and cost in
carrying out impact evaluation, the costs and benefits should be assessed, and consideration should be given to
whether another approach would be more appropriate, such as monitoring of key performance indicators or a
process evaluation.
(These approaches should not be seen as substitutes for impact evaluations; indeed they often form critical
complements to impact evaluations.)
Box 2.1 Main Steps in Designing and Implementing Impact Evaluations
During Project Identification and Preparation
1. Determining whether or not to carry out an evaluation
2. Clarifying objectives of the evaluation
3. Exploring data availability
4. Designing the evaluation
5. Forming the evaluation team
6. If data will be collected:
(a) Sample design and selection
(b) Data collection instrument development
(c) Staffing and training fieldwork personnel
(d) Pilot testing
(e) Data collection
(f) Data management and access
During Project Implementation
7. Ongoing data collection
8. Analyzing the data
9. Writing up the findings and discussing them with policymakers and other stakeholders
10. Incorporating the findings in project design
And perhaps the most important inputs to the decision of whether or not to carry out an evaluation are strong
political and financial support. The additional effort and resources required for conducting impact evaluations
are best mobilized when the project is innovative, is replicable, involves substantial resource allocations, and has
well-defined interventions.
73
Impact evaluations should also be prioritized if the project in question is launching a new approach such as a
pilot program that will later be under consideration for expansion based on the results of the evaluation, or the
new World Bank Learning and Innovation Loans. This rationale made the Nicaraguan school autonomy reform a
good candidate for an impact evaluation. The evaluation study accompanied the government's testing of a new
decentralized school management model from its pilot stage in the mid-1990s through its expansion to almost all
secondary schools and about half of all primary schools today. The evaluation was managed by a closely
coordinated international team including local staff from the Ministry of Education's research and evaluation unit
and the World Bank's Primary Education Project coordination office in Managua. Their involvement ensured that
the evaluation informed key policy decisions regarding the modification and expansion of the pilot. Another
important consideration is to ensure that the program that is to be evaluated is sufficiently developed to be
subject to an impact evaluation. Pilot projects and nascent reforms are often prone to revisions regarding their
content as well as how, when, and by whom they will be implemented. These changes can undermine the
coherence of the evaluation effort, particularly experimental designs and other types of prospective evaluations
that rely on baseline and follow-up data of clearly established treatment and control groups. Where the policies
to be evaluated are still being defined, it may be advisable to avoid using an impact evaluation in order to allow
for flexibility in the project. Gaining support from policymakers and financiers for an impact evaluation can be
challenging but is a prerequisite for proceeding. They must be convinced that the evaluation is a useful exercise
addressing questions that will be relevant to decisions concerning the evaluated program's refinement, expansion,
or curtailment. They must also be convinced of the legitimacy of the evaluation design and therefore the results,
particularly when the results are not as positive as anticipated. Financing for an impact evaluation remains a
difficult issue for program managers and client counterparts alike. The financing issue is compounded by the fact
that data on evaluation costs are usually difficult to obtain. The stickiest issue arises from the public good value
of the evaluation: if the results of the evaluation are going to be used to inform policies applied outside of the
national boundaries within which the evaluation is conducted, why should an individual country bear the cost of
the evaluation? Among the case studies that had information on sources of funding, the information shows that
countries often assume the majority, but not the entirety, of the evaluation costs.
Clarifying Evaluation Objectives
Once it has been determined that an impact evaluation is appropriate and justified, establishing clear objectives
and agreement on the core issues that will be the focus of the evaluation up front will contribute greatly to its
success. Clear objectives are essential to identifying information needs, setting output and impact indicators, and
constructing a solid evaluation strategy to provide answers to the questions posed. The use of a logical (log)
framework approach provides a good and commonly used tool for identifying the goals of the project and the
information needs around which the evaluation can be constructed.
The log frame is based on a simple four-by-four matrix that matches information on project objectives with how
performance will be tracked using milestones and work schedules, what impact project outputs will have on a
beneficiary institution or system and how that will be measured, and how inputs are used to deliver outputs. It is
assumed that the project's intended impact is a function of the project's outputs as well as a series of other factors.
The outputs, in turn, are a function of the project's inputs and factors outside the project. Quantifiable measures
should then be identified for each link in the project cycle. This approach does not preclude the evaluator from
also looking at the unintended impacts of a project but serves to keep the objectives of the evaluation clear and
focused. Qualitative techniques are also useful in eliciting participation in clarifying the objectives of the
evaluation and resulting impact indicators. Although a statement of the objective would seem on the face of it to
be one of the easiest parts of the evaluation process, it can be extremely difficult. For example, statements that
are too broad do not lend themselves to evaluation. The objective statement in the Mexico PROBECAT
evaluation that the evaluation is about "the effect of the PROBECAT training program on labor market
outcomes" would be more precise if it were narrowed down to the effect of PROBECAT on hours worked,
hourly earnings, monthly salary, and time to first job placement for different types of workers. The Mexico
PROGRESA evaluation provides a good example of creating a clear outline and delineating multiple objectives
from the start with a separate discussion of each component-with objectives detailed in subcategories (Annex
74
1.10). This was particularly important because the intervention was quite complex, with the evaluation having to
address not only the program impact but also aspects of program operations targeting and timing.
Reviewing other evaluation components such as cost-effectiveness or process evaluations may also be important
objectives of a study and can complement the impact evaluation. Cost-effectiveness may be of particular concern
for policymakers whose decision it will be to curtail, expand, or reform the intervention being evaluated. On
issues related to service delivery, a process evaluation may be relevant to assess the procedures, dynamics,
norms, and constraints under which a particular program is carried out.
Exploring Data Availability
Many types of data, like cross-sectional or panel surveys to qualitative open-ended interviews, can be used to
carry out impact evaluation studies. Ideally this information is available at the individual level to ensure that true
impact can be assessed. Household level information can conceal intra household resource allocation, which
affects women and children because they often have more limited access to household productive resources. In
many cases, the impact evaluation will take advantage of some kind of existing data or piggyback on an ongoing
survey, which can save considerably on costs. With this approach, problems may arise in the timing of the data
collection effort and with the flexibility of the questionnaire design. Box 2.2 highlights some key points to
remember in exploring the use of existing data resources for the impact evaluation. With some creativity, it may
be possible to maximize existing information resources.
At the most basic level, data on the universe of the population of interest will be required as a basis from which
to determine sample sizes, construct the sampling frame, and select the sample. Other types of data that may be
available in a given country and can be used for different impact
evaluations include (see Valadez and Bamberger 1994): household income and expenditure surveys; Living
Standards Measurement Studies (LSMSs); labor market surveys; records of cooperatives, credit unions, and
other financial institutions; school records on attendance, repetition, and examination performance; public health
records on infant mortality,
Box 2.2 Key Points for Identifying Data Resources for Impact Evaluation
* Know the program well. It is risky to embark on an evaluation without knowing a lot about the administrative and
institutional details of the program; that information typically comes from the
program administration.
* Collect information on the relevant "stylized facts" about the setting. The relevant facts might include the poverty map,
the way the labor market works, the major ethnic divisions, and other relevant public programs.
* Be eclectic about data. Sources can embrace both informal, unstructured interviews with participants in the program and
quantitative data from representative samples. However, it is extremely difficult to ask counterfactual questions in
interviews or focus groups; try asking someone who is currently participating in a public program: "What would you be
doing now if this program did not exist?" Talking to program participants can be valuable, but it is unlikely to provide a
credible evaluation on its own.
* Ensure that there is data on the outcome indicators and relevant explanatory variables. The latter need to deal with
heterogeneity in outcomes conditional on program participation. Outcomes can differ depending, for example, on whether
one is educated. It may not be possible to see the impact of the program unless one controls for that heterogeneity.
* Depending on the methods used, data might also be needed on variables that influence participation but do not influence
outcomes given participation. These instrumental variables can be valuable in sorting out the likely causal effects of
nonrandom programs (box 1.2).
* The data on outcomes and other relevant explanatory variables can be either quantitative or qualitative. But it has to be
possible to organize the information in some sort of systematic data structure. A simple and common example is that one has
values of various variables including one or more outcome indicators for various observation units (individuals, households,
firms, communities).
75
* The variables one has data on and the observation units one uses are often chosen as part of the evaluation method. These
choices should be anchored to the prior knowledge about the program (its objectives, of course, but also how it is run) and
the setting in which it is introduced.
* The specific source of the data on outcomes and their determinants, including program participation, typically comes from
survey data of some sort. The observation unit could be the household, firm, or geographic area, depending on the type of
program one is studying.
* Survey data can often be supplemented with useful other data on the program (such as from the project monitoring
database) or setting (such as from geographic databases). Incidence of different infectious diseases, number of women
seeking advice on contraception, or condom consumption; specialized surveys conducted by universities, nongovernmental
organizations (NGOs), and consulting groups; monitoring data from program administrators; and
project case studies.
Using Existing Survey Data. Many surveys may also be in the planning stages or are ongoing. If a survey
measuring the required indicators is planned, the evaluation may be able to over sample the population of
interest during the course of the general survey (for example, to use for
the propensity score matching approach) as was done for the Nicaraguan Social Investment Fund evaluation and
the Argentine TRABAJAR workfare program evaluation (Jalan and Ravallion 1998).
Conversely, if a survey is planned that will cover the population of interest, the evaluation may be able to
introduce a question or series of questions as part of the survey or add a qualitative survey to supplement the
quantitative information. The evaluation assessed the impact of the program on the nutritional status and food
security of poor households. Quantitative data included specific questions on household income and expenditure
and skills level, whereas qualitative data focused on women's empowerment-status and decisionmaking in the
household, social networks, self-confidence, and so forth.
Designing the Evaluation
Once the objectives and data resources are clear, it is possible to begin the design phase of the impact evaluation
study. The choice of methodologies will depend on the evaluation question, timing, budget constraints, and
implementation capacity. The pros and cons of the different design types should be balanced to determine which
methodologies are most appropriate and how quantitative and qualitative techniques can be integrated to
complement each other. Even after the evaluation design has been determined and built into the project,
evaluators should be prepared to be flexible and make modifications to the design as the project is implemented.
In addition, provisions should be made for tracking the project interventions if the evaluation includes baseline
and follow-up data so that the evaluation effort is parallel with the actual pace of the project.
In defining the design, it is also important to determine how the impact evaluation will fit into the broader
monitoring and evaluation strategy applied to a project. All projects must be monitored so that administrators,
lenders, and policymakers can keep track of the project as it unfolds. The evaluation effort, as argued above,
must be tailored to the information requirements of the project.
Evaluation Question. The evaluation questions being asked are very much linked to the design of the evaluation
in terms of the type of data collected, unit of analysis, methodologies used, and timing of the various stages. For
example, in assessing the impact of textbooks on learning outcomes, it would be necessary to tailor the
evaluation to measuring impact on students, classrooms, and teachers during a given school year. This would be
very different than measuring the impact of services provided through social fund investments, which would
require data on community facilities and households. The case studies in Annex I provide the other examples of
how the evaluation question can affect the evaluation design.
In clarifying the evaluation questions, it is also important to consider the gender implications of project impact.
At the outset this may not always be obvious, however; in project implementation there may be secondary
76
effects on the household, which would not necessarily be captured without specific data collection and analysis
efforts.
Timing and Budget Concerns. The most critical timing issue is whether it is possible to begin the evaluation
design before the project is implemented and when the results will be needed. It is also useful to identify up front
at which points during the project cycle information from the evaluation effort will be needed so that data
collection and analysis activities can be linked. Having results in a timely manner can be crucial to policy
decisions-for example, during a project review, around an election period, or when decisions regarding project
continuation are being made. Some methods require more time to implement than others. Random assignment
and before-and-after methods (for example, reflexive comparisons) take longer to implement than ex-post
matched-comparison approaches. When using before-and-after approaches that utilize baseline and follow-up
assessments, time must be allowed for the last member of the treatment group to receive the intervention, and
then usually more time is allowed for post program effects to materialize and be observed. Grossman (1994)
suggests that 12 to 18 months after sample enrollment in the intervention is a typical period to allow before
examining impacts. In World Bank projects with baselines, waiting for both the intervention to take place and the
outcomes to materialize can take years. For example, in the evaluation of the Bolivian Social Investment Fund,
which relied on baseline data collected in 1993, follow-up data was not collected until 1998 because of the time
needed for the interventions (water and sanitation projects, health clinics, and schools) to be carried out and for
effects on the beneficiary population's health and education outcomes to take place. A similar period of time has
been required for the evaluation of a primary education project in Pakistan that used an experimental design with
baseline and follow-up surveys to assess the impact of community schools on student outcomes, including
academic achievement. The timing requirements of the evaluation cannot drive the project being evaluated. By
their very nature, evaluations are subject to the time
frame established by the rest of the project. Evaluations must wait on projects that are slow to disburse and
generate interventions. And even if projects move forward at the established pace, some interventions take
longer to carry out, such as infrastructure projects. The time frame for the evaluation is also sensitive to the
indicators selected because many, such as changes in fertility rates or educational achievement, take longer to
manifest themselves in the beneficiary population.
Implementation Capacity. A final consideration in the scale and complexity of the evaluation design is the
implementation capacity of the evaluation team. Implementation issues can be very challenging, particularly in
developing countries where there is little experience with applied research and program evaluations. The
composition of the evaluation team is very important, as well as team members' experience with different types
of methodologies and their capacity relative to other activities being carried out by the evaluation unit. This is
particularly relevant when working with public sector agencies with multiple responsibilities and limited staff.
Awareness of the unit's workload is important in order to assess not only how it will affect the quality of
evaluation being conducted but also the opportunity cost of the evaluation with respect to other efforts for which
the unit is responsible. There are several examples of evaluation efforts that were derailed when key staff were
called onto other projects and thus were not able to implement the collection of data on schedule at the critical
point in time (such as a point during the school year or during agricultural season). Such situations can be
avoided through coordination with managers in the unit responsible for the evaluation to ensure that a balance is
achieved with respect to the timing of various activities, as well as the distribution of staff and resources across
these activities. Alternatively, it can be preferable to contract a private firm to carry out the evaluation (discussed
below).
Formation of the Evaluation Team
A range of skills is needed in evaluation work. The quality and eventual utility of the impact evaluation can be
greatly enhanced with coordination between team members and policymakers from the outset. It is therefore
important to identify team members as early as possible, agree upon roles and responsibilities, and establish
mechanisms for communication during key points of the evaluation. Among the core team is the evaluation
manager, analysts (both economist and other social scientists), and, for evaluation designs involving new data
collection, a sampling expert, survey designer, fieldwork manager and fieldwork team, and data managers and
77
processors (for a comprehensive guide to designing and implementing surveys, see Grosh and Mufnoz 1996).
Depending on the size, scope, and design of the study, some of these responsibilities will be shared or other
staffing needs may be added to this core team. In cases in which policy analysts may not have had experience
integrating quantitative and qualitative approaches, it may be necessary to spend additional time at the initial
team building stage to sensitize team members and ensure full collaboration. In building up the evaluation team,
there are also some important decisions that the evaluation manager must make about local capacity and the
appropriate institutional arrangements to ensure impartiality and quality in the evaluation results.
First is whether there is local capacity to implement the evaluation, or parts of it, and what kind of supervision
and outside assistance will be needed. Evaluation capacity varies greatly from country to country, and although
international contracts that allow for firms in one country to carry out evaluations in another country are
becoming more common (one example is the Progresa evaluation being carried out by the International Food and
Policy Research Institute), the general practice for World Bank-supported projects seems to be to implement the
evaluation using local staff while providing a great deal of international supervision. Therefore, it is necessary to
critically assess local capacity and determine who will be responsible for what aspects of the evaluation effort.
Regardless of the final composition of the team, it is important to designate an evaluation manager who will be
able to work effectively with the data producers as well as the analysts and policymakers using the data and the
results of the evaluation. If this person is not based locally, it is recommended that a local manager be designated
to coordinate the evaluation effort in conjunction with the international manager.
Second is whether to work with a private firm or public agency. Private firms can be more dependable with
respect to providing results on a timely basis, but capacity building in the public sector is lost and often private
firms are understandably less amenable to incorporating elements into the evaluation that will make the effort
costlier. Whichever counterpart or combination of counterparts is finally crafted, a sound review of potential
collaborators' past evaluation activities is essential to making an informed choice.
And third is what degree of institutional separation to put in place between the evaluation providers and the
evaluation users. There is much to be gained from the objectivity provided by having the evaluation carried out
independently of the institution responsible for the project being evaluated. However, evaluations can often have
multiple goals, including building evaluation capacity within government agencies and sensitizing program
operators to the realities of their projects once these are carried out in the field. At a minimum, the evaluation
users, who can range from policymakers in government agencies in client countries to NGO organizations,
bilateral donors, and international development institutions, must remain sufficiently involved in the evaluation
to ensure that the evaluation process is recognized as being legitimate and that the results produced are relevant
to their information needs. Otherwise, the evaluation results are less likely to be used to inform policy. In the
final analysis, the evaluation manager and his or her clients must achieve the right balance between involving the
users of evaluations and maintaining the objectivity and legitimacy of the results.
Data Development
Having adequate and reliable data is a necessary input to evaluating project impact. High-quality data are
essential to the validity of the evaluation results. As discussed above, assessing what data exist is a first
important step before launching any new data collection efforts. Table 2.1 links the basic evaluation
methodologies with data requirements. Most of these methodologies can incorporate qualitative and
participatory techniques in the design of the survey instrument, in the identification of indicators, and in input to
the identification of controls, variables used for matching, or in instrumental variables.
78
For evaluations that will generate their own data, there are the critical steps of designing the data collection
79
instruments, sampling, fieldwork, data management, and data access. This section does not outline the
step-by-step process of how to undertake a survey but rather provides a brief discussion of these steps. Some of
the discussion in this section is more relevant to evaluations based on the collection and analysis of larger-scale
sample surveys using quantitative data than for evaluations using qualitative data and small sample sizes.
Deciding What to Measure. The main output and impact indicators should be established in planning the
evaluation, possibly as part of a logical framework approach. To ensure that the evaluation is able to assess
outcomes during a period of time relevant to decision makers' needs, a hierarchy of indicators might be
established, ranging from short term impact indicators such as school attendance to longer-term indicators such
as student achievement. This ensures that even if final impacts are not picked up initially, program outputs can
be assessed. In addition, the evaluator should plan on measuring the delivery of intervention as well as taking
account of exogenous factors that may have an effect on the outcome of interest.
Evaluation managers can also plan to conduct the evaluation across several time periods, allowing for more
immediate impacts to be picked up earlier while still tracking final outcome measures. This was done in the
Nicaragua School Reform evaluation, in which the shorter-term impact of the reform on parental participation
and student and teacher attendance was established and the longer-term impacts on student achievement are still
being assessed. Information on the characteristics of the beneficiary population not strictly related to the impact
evaluation but of interest in the analysis might also be considered, such as their level of poverty or their opinion
of the program. In addition, the evaluator may also want to include cost measures in order to do some
cost-effectiveness analysis or other complementary assessments not strictly related to the impact evaluation. The
type of evaluation design selected for the impact evaluation will also carry data requirements. These will be
specific to the methodology, population of interest, impact measures, and other elements of the evaluation. For
example, if an instrumental variable approach (one of the types of matched-comparison strategies) is to be used,
the variable(s) that will serve as the instrument to separate program participation from the outcome measures
must be identified and included in the data collection. This was done for the Bolivian Social Investment Fund
impact evaluation, in which knowledge of the social fund and the presence of NGOs were used as instrumental
variables in assessing the impact of social fund interventions.
It can be useful to develop a matrix for the evaluation, listing the question of interest, the outcome indicators that
will be used to assess the results, the variable, and the source of data for the variable. This matrix can then be
used to review questionnaires and plan the analytical work as was done in the evaluation of the Nicaragua
Emergency Social Investment Fund.
Developing Data Collection Instruments and Approaches.
Developing appropriate data collection instruments to generate the required data to answer the evaluation
questions can be tricky. This will require having the analysts involved in the development of the questions, in the
pilot test, and in the review of the data from the pilot test. Involving both the field manager and the data manager
during the development of the instruments, as well as local staff-preferably analysts who can provide knowledge
of the country and the program-can be critical to the quality of information collected (Grosh and Munioz:1996.)
It is important to ensure that the data collected can be disaggregated by gender to explore the differential
impact of specific programs and policies.
Quantitative evaluations usually collect and record information either in a numeric form or as precoded
categories. With qualitative evaluations, information is generally presented as descriptive text with little or no
categorization. The information may include an individual's responses to open-ended interview questions, notes
taken during focus groups, or the evaluator's observations of events. Some qualitative studies use the precoded
classification of data as well (Bamberger, 2000). The range of data collection instruments and their strengths and
weaknesses are summarized in table 2.2.
The responses to survey questionnaires can be very sensitive to design; thus it is important to ensure that the
structure and format are appropriate, preferably undertaken by experienced staff. For example, the utility of
80
quantitative data has often been severely handicapped for simple mechanical reasons, such as the inability to link
data from one source to
another. This was the case in a national education assessment in one country where student background data
could not be linked to test score results, which made it impossible to assess the influence of student
characteristics on performance or to classify the tests scores by students' age,
gender, socioeconomic status, or educational history.
For both qualitative and quantitative data collection, even experienced staff must be trained to collect the data
specific to the evaluation, and all data collection should be guided by a set of manuals that can be used as
orientation during training and as a reference during the fieldwork. Depending on the complexity of the data
collection task, the case examples show that training can range from three days to several weeks. Pilot testing is
an essential step because it will reveal whether the instrument can reliably produce the required data and how the
data collection procedures can be put into operation. The pilot test should mimic the actual fieldwork as closely
as possible. For this reason, it is useful to have data entry programs ready at the time of the pilot to test their
functionality as well as to pilot test across the different populations and geographical areas to be included in the
actual fieldwork.
Sampling. “Sampling” is an art best practiced by an experienced sampling specialist. The design need not be
complicated, but it should be informed by the sampling specialist's expertise in the determination of appropriate
sampling frames, sizes, and selection strategies. The sampling specialist should be incorporated in the evaluation
process from the earliest stages to review the available information needed to select the sample and determine
whether any enumeration work will be needed, which can be time consuming. As with other parts of the
evaluation work, coordination between the sampling specialist and the evaluation team is important. This becomes
particularly critical in conducting matched comparisons because the sampling design becomes the basis for the "match"
that is at the core of the evaluation design and construction of the counterfactual.
81
82
In these cases, the sampling specialist must work closely with the evaluation team to develop the criteria that
will be applied to match the treatment and comparison groups.
There are many tradeoffs between costs and accuracy in sampling that should be made clear as the sampling
framework is being developed. For example, conducting a sample in two or three stages will reduce the costs of
both the sampling and the fieldwork, but the sampling errors and therefore the precision of the estimates will be
increased. Once the outcome variables and population(s) of interest have been determined by the evaluation team,
a first step for the sampling specialist would be to determine the power calculations (see Valadez and Bamberger
1994, pp. 382-84, for a discussion of the power calculation process). Since the power calculation can be
performed using only one outcome measure, and evaluations often consider several, some strategic decisions
will need to be made regarding which outcome indicator to use when designing the sample. After developing the
sampling strategy and framework, the sampling specialist should also be involved in selecting the sample for the
fieldwork and the pilot test to ensure that the pilot is not conducted in an area that will be included in the sample
for the fieldwork. Often initial fieldwork will be required as part of the sample selection procedure. For
example, an enumeration process will be required if there are no up-to-date maps of units required for the sample
(households, schools, and so forth) or if a certain population of interest, such as malnourished children, needs to
be pre-identified so that it can be selected for the purpose of the evaluation.
Once the fieldwork is concluded, the sampling specialist should provide assistance on determining sampling
weights to compute the expansion factors and correct for sampling errors and nonresponse. (Grosh and Munioz
1996 provide a detailed discussion of sampling procedures as part of household survey work. Kish 1965 is
considered one of the standard textbooks in the sampling field.) And finally, the sampling specialist should
produce a sampling document detailing the sampling strategy, including (a) from the sampling design stage, the
power calculations using the impact variables, the determination of sampling errors and sizes, the use of
stratification to analyze populations of interest; (b) from the sample selection stage, an outline of the sampling
stages and selection procedures; (c) from the fieldwork stage to prepare for analysis, the relationship between the
size of the sample and the population from which it was selected, nonresponse rates, and other information used
to inform sampling weights; and any additional information that the analyst would need to inform the use of the
evaluation data. This document can be used to maintain the evaluation project records and should be included
with the data whenever it is distributed to help guide the analysts in using the evaluation data.
Questionnaires. The design of the questionnaire is important to the validity of the information collected. There
are four general types of information required for an impact evaluation (Valadez and Bamberger 1994). These
include
•Classification of nominal data with respondents classified according to whether they are project
participants or belong to the comparison group;
•Exposure to treatment variables recording not only the services and benefits received but also the
frequency, amount, and quality-assessing quality can be quite difficult;
•Outcome variables to measure the effects of a project, including immediate products, sustained outputs
or the continued delivery of services over a long period, and project impacts such as improved income
and employment; and
•Intervening variables that affect participation in a project or the type of impact produced, such as
individual, household, or community characteristics-these variables can be important for exploring
biases.
The way in which the question is asked, as well as the ordering of the questions, is also quite important in
generating reliable information. A relevant example is the measurement of welfare, which would be required for
measuring the direct impact of a project on poverty reduction. Asking individuals about their income level would
not necessarily yield accurate results on their level of economic well- being. As discussed in the literature on
welfare measurement, questions on expenditures, household composition, assets, gifts and remittances, and the
imputed value of homegrown food and owner-occupied housing are generally used to capture the true value of
household and individual welfare. The time recall used for expenditure items, or the order in which these
83
questions are asked, can significantly affect the validity of the information collected. Among the elements noted
for a good questionnaire are keeping it short and focused on important questions, ensuring that the instructions
and questions are clear, limiting the questions to those needed for the evaluation, including a "no opinion" option
for closed questions to ensure reliable data, and using sound procedures to administer the questionnaire, which
may indeed be different for quantitative and qualitative surveys.
Fieldwork Issues. Working with local staff who have extensive experience in collecting data similar to that
needed for the evaluation can greatly facilitate fieldwork operations. Not only can these staff provide the
required knowledge of the geographical territory to be covered, but their knowledge can also be critical to
developing the norms used in locating and approaching informants. Field staff whose expertise is in an area other
than the one required for the evaluation effort can present problems, as was the case in an education evaluation
in Nicaragua that used a firm specializing in public opinion polling to conduct a school and household survey.
The expertise that had allowed this firm to gain an excellent reputation based on its accurate prediction of
improbable election results was not useful for knowing how to approach school children or merge quantitative
data sets. This lack of expertise created substantial survey implementation problems that required weeks of
corrective action by a joint team from the Ministry of Education and the World Bank. The type of staff needed to
collect data in the field will vary according to the objectives and focus of the evaluation. For example, a
quantitative impact evaluation of a nutrition program might require the inclusion of an anthropometrist to collect
height-for-weight measures as part of a survey team, whereas the impact evaluation of an educational reform
would most likely include staff specializing in the application of achievement
tests to measure the impact of the reform on academic achievement. Most quantitative surveys will require at
least a survey manager, data manager, field manager, field supervisors, interviewers, data entry operators, and
drivers. Depending on the qualitative approach used, field staff may be similar with the exception of data entry
operators. The skills of the interviewers, however, would be quite different, with qualitative interviewers
requiring specialized training, particularly for focus groups, direct observation, and so forth. Three other
concerns are useful to remember when planning survey operations. First, it is important to take into
consideration temporal events that can affect the operational success of the fieldwork and the external validity of
the data collected, such as the school year calendar, holidays, rainy seasons, harvest times, or migration patterns.
Second, it is crucial to pilot test data collection instruments, even if they are adaptations of instruments that have
been used previously, both to test the quality of the instrument with respect to producing the required data and to
familiarize fieldwork staff with the dynamics of the data collection process. Pilot tests can also serve as a
proving ground for the selection of a core team of field staff to carry out the actual survey. Many experienced
data collectors will begin with 10 to 20 percent more staff in the pilot test than will be used in the actual
fieldwork and then select the best performers from the pilot to form the actual data collection teams. Finally,
communications are essential to field operations. For example, if local
conditions permit their use, fieldwork can be enhanced by providing supervisors with cellular phones so that
they can be in touch with the survey manager, field manager, and other staff to answer questions and keep them
informed of progress.
Data Management and Access. The objectives of a good data management system should be to ensure the
timeliness and quality of the evaluation data. Timeliness will depend on having as much integration as possible
between data collection and processing so that errors can be verified
and corrected prior to the conclusion of fieldwork. The quality of the data can be ensured by applying
consistency checks to test the internal validity of the data collected both during and after the data are entered and
by making sure that proper documentation is available to the analysts who will be using the data. Documentation
should consist of two types of information: (a) information needed to interpret the data, including codebooks,
data dictionaries, guides to constructed variables, and any needed translations; and (b) information needed to
conduct the analysis, which is often included in a basic information document that contains a description of the
focus and objective of the evaluation, details on the evaluation methodology, summaries or copies of the data
collection instruments, information on the sample, a discussion of the fieldwork, and guidelines for using the
data.
It is recommended that the data produced by evaluations be made openly available given the public good value
84
of evaluations and the possible need to do additional follow-up work to assess long-term impacts by a team other
than the one that carried out the original evaluation work. To facilitate the data-sharing process, at the outset of
the evaluation an open data access policy should be agreed upon and signed, establishing norms and
responsibilities for data distribution. An open data access policy puts an added burden on good data
documentation and protect ing the confidentiality of the informants. If panel data are collected from the same
informants over time by different agencies, the informants will have to be identified to conduct the follow-up
work. This requirement should be balanced against the confidentiality norms that generally accompany any
social sector research. One possible solution is to make the anonymous unit record data available to all interested
analysts but ask researchers interested in conducting follow-up work to contact the agency in charge of the data
in order to obtain the listing of the units in the sample, thereby giving the agency an opportunity to ensure
quality control in future work through contact with the researchers seeking to carry it out.
Analysis, Reporting, and Dissemination
The analysis of the evaluation data requires collaboration between the analysts, data producers, and
policymakers to clarify questions and ensure timely quality results. Problems with the cleaning and interpretation
of data will arise during analysis and require input from various team members. There are techniques and
challenges of carrying out quantitative analysis based on statistical methods. There are also many techniques for
analyzing qualitative data (see Miles and Huberman 1994). Two commonly used methods for impact evaluation
are mentioned- “content analysis” and “case analysis” (Taschereau 1998). “Content analysis” is used to analyze
data drawn from interviews, observations, and documents. In reviewing the data, the evaluator develops a
classification system for the data, organizing information based on (a) the evaluation questions for which the
information was collected; (b) how the material will be used; and (c) the need for cross-referencing the
information.
The coding of data can be quite complex and may require many assumptions. Once a classification system has
been set up, the analysis phase begins. This involves looking for patterns in the data and moving beyond
description toward developing an understanding of program processes, outcomes, and impacts. This is best
carried out with the involvement of team members. New ethnographic and linguistic computer programs are also
now available, designed to support the analysis of qualitative data.
“Case analysis” is based on case studies designed for in-depth study of a particular group or individual. The high
level of detail can provide rich information for evaluating project impact. The processes of collecting and
analyzing the data are carried out simultaneously as evaluators make observations as they are collecting
information. They can then develop and test explanations and link critical pieces of information.
First, analysis commonly takes longer than anticipated, particularly if the data are not as clean or accessible at
the beginning of the analysis, if the analysts are not experienced with the type of evaluation work, or if there is
an emphasis on capacity building through collaborative work. In the review of the case studies considered for
this article, the most rapid analysis took approximately one year after producing the data and the longer analysis
took close to two years. The case in chapter 3 illustrates some of the many steps involved in analysis and why it
can take longer than anticipated.
Second, the evaluation manager should plan to produce several products as outputs from the analytical work,
keeping in mind two elements. The first is to ensure the timing of outputs around key events when decisions
regarding the future of the project will be made, such as mid-term reviews, elections, or closings of a pilot phase.
The second is the audience for the results. Products should be differentiated according to the audience for which
they are crafted, including government policymakers, program managers, donors, the general public, journalists,
and academics.
Third, the products will have the most policy relevance if they include clear and practical recommendations
stemming from the impact analysis. These can be broken into short- and long-term priorities, and when possible,
should include budgetary implications. Decisionmakers will be prone to look for the "bottom line."
85
Finally, the reports should be planned as part of a broader dissemination strategy, which can include
presentations for various audiences, press releases, feedback to informants, and making information available on
the Web. Such a dissemination strategy should be included in the initial stages of the planning process to ensure
that it is included in the budget and that the results reach the intended audience.
Bibliography
The word processed describes informally reproduced works that may not be commonly available through
libraries.
Abadie, A., J. Angrist, and G. Imbens. 1998. "Instrumental Variables Estimation of Quartile Treatment Effects."
National Bureau of Economic Research Working Paper Series, No. 229:1-28.
Atkinson, Anthony. 1987. "On the Measurement of Poverty." Econometrica 55: 749-64.
Bamberger, Michael.2 000. Integrating uantitative and Qualitative Methods in Development Research
Washington, D.C.: World Bank.
Barnes, Carolyn. 1966. "Assets and the Impact of Microenterprise Finance Programs." USAID AIMS Project
Brief 6. Washington, D.C.: USAID.
Barnes, Carolyn, and Erica Keogh. 1999. "An Assessment of the Impact of Zambuko's Microenterprise Program
in Zimbabwe: Baseline Findings." USAID AIMS Project Brief 23. Washington, D.C.: USAID.
Benus, Jacob, Neelima Grover, and Recep Varcin. 1998. Turkey: Impact of Active Labor Programs. Bethesda,
Md.: Abt Associates.
Benus, Jacob, Neelima Grover, Jiri Berkovsky, and Jan Rehak. 1998. Czech Republic: Impact of Active Labor
Programs. Bethesda, Md.: Abt Associates.
Besharov, Douglas J., Peter Germanis, and Peter H. Rossi. 1997. Evaluating Welfare Reform: A Guide for
Scholars and Practitioners. College Park, Md.: University of Maryland.
Bindlish, V., R. Evenson, and Mathurin Gbetibouo. 1993. Evaluation of T & V-Based Extension in Burkina Faso.
Washington, D.C.: World Bank.
Bloom, Howard S., Larry L. Orr, George Cave, Stephen H. Bell, Fred Doolittle, and Winston Lin. 1994. The
National JTPA Study: Overview,Impacts, Benefits, and Costs of Title II-A. Bethesda, Md.: Abt Associates.
Blundell, Richard W., and R. J. Smith. 1993. "Simultaneous Microeconometric Models with Censoring or
Qualitative Dependent Variables." In G. S. Maddala, C. R. Rao, and H. D. Vinod, eds., Handbook of Statistics.
Vol. 11. Amsterdam: North Holland.
Bourguignon, Francois, Jaime de Melo, and Akiko Suwa. 1991. "Distributional Effects of Adjustment Policies:
Simulations for Archetype Economies in Africa and Latin America." World Bank Economic Review 5 (2):
339-66.
Burtless, Gary. 1995. "The Case for Randomized Field Trials in Economic and Policy Research." Journal of
Economic Perspectives 9 (Spring): 63-84.
Card, David, and Philip K. Robins. 1996. "Do Financial Incentives Encourage Welfare Recipients to Work?
Evidence from a Randomized Evaluation of the Self-Sufficiency Project." National Bureau of Economic
86
Research Paper 5701, August. Cambridge, Mass.: NBER.
Carvalho, Soniya, and Howard White. 1994. "Indicators for Monitoring Poverty Reduction." World Bank
Discussion Papers 254. Washington, D.C.: World Bank.
Chen, Martha Alter, and Donald Snodgrass. 1999. "An Assessment of the Impact of Sewa Bank in India:
Baseline Findings." USAID AIMS Project, Processed. Washington, D.C.: USAID.
Cohen, Monique, and Gary Gaile. 1997. "Highlights and Recommendations of the Virtual Meeting of the CGAP
Working Group on Impact Assessment Methodologies." USAID AIMS Project. Washington, D.C.: Management
Systems International. 1998. "Highlights and Recommendations of the Second Virtual Meeting of the CGAP
Working Group on Impact Assessment Methodologies." USAID AIMS Project. Washington, D.C.: Management
Systems International.
Dar, Amit, and Indermit S. Gill. 1998. "Evaluating Retraining Programs in OECD Countries: Lessons Learned."
The World Bank Research Observer 13 (February): 79-101.
Dar, Amit, and Zafiris Tzannatos. 1999. "Active Labor Market Programs: A Review of Evidence from
Evaluations." World Bank Social Protection Discussion Paper 9901. Washington, D.C.: World Bank.
Dehejia, Rajeev H,. and Sadek Wahba. 1998. "Causal Effects in Non-Experimental Studies: Re-Evaluating the
Evaluation of Training Programs." NBER Working Paper Series 6586. Cambridge, Mass.: NBER
http://www.nber.org/papers/w6586.
Dennis, Michael L., and Robert F. Boruch. 1989. "Randomized Experiments for Planning and Testing Projects in
Developing Countries: Threshold Conditions." Evaluation Review 13 (June): 292-309.
Diagne, Aliou, and Manfred Zeller. 1998. "Determinants of Access to Credit and Its Impact on Income and Food
Security in Malawi." Manuscript submitted to the International Food Policy Research Institute (IFPRI)
Publication Review Committee for consideration as an IFPRI Research Report.
Diop, F., A. Yazbeck, and R. Bitran. 1995. "The Impact of Alternative Cost Recovery Schemes on Access and
Equity in Niger." Health Policy and Planning 10 (3): 223-40.
Donecker, Jane, and Michael Green. 1998. Impact Assessment in Multilateral Development Institutions. London:
Department for International Development.
Dunn, Elizabeth. 1999. "Microfinance Clients in Lima, Peru: Baseline Report for AIMS Core Impact
Assessment." USAID AIMS Project, Processed. Washington, D.C.: USAID.
Edgecomb, Elaine L., and Carter Garber. 1998. "Practitioner-Led Impact Assessment: A Test in Honduras."
USAID AIMS. Washington, D.C.: USAID.
Ezemenari, Kene, Anders Ruqvist, and K. Subbarao. 1999. "Impact Evaluation: A Note on Concepts and
Methods." World Bank Poverty Reduction and Economic Management Network, Processed. Washington, D.C.:
World Bank.
Foster, James, J. Greer, and Erik Thorbecke. 1984. "A Class of Decomposable Poverty Measures."
Econometrica5 2: 761-65.
Friedlander, Daniel, and Gayle Hamilton. 1993. The Saturation Work Initiative Model in San Diego: A Five-Year
Follow-Up Study. New York: Manpower Demonstration Research Corporation.
87
EVALUATING THE IMPACT OF DEVELOPMENT PROJECTS ON POVERTY
Friedlander, Daniel, and Philip K. Robins. 1995. "Evaluating Program Evaluations: New Evidence on
Commonly Used Non experimental Methods." American Economic Review 85 (September): 923-37.
Friedlander, Daniel, David H. Greenberg, and Philip K. Robins. 1997. "Evaluating Government Training
Programs for the Economically Disadvantaged." Journal of Economic Literature 35 (December): 1809-55.
Fuller, Bruce, and Magdalena Rivarola. 1998. Nicaragua's Experiment to Decentralize Schools: Views of Parents,
Teachers and Directors. Working Paper Series on Impact Evaluation of Education Reforms, paper no. 5.
Washington, D.C.: World Bank.
Gaile, Gary L., and Jenifer Foster. 1996. "Review of Methodological Approaches to the Study of the Impact of
MicroEnterprise Credit Programs." USAID AIMS Brief 2. Washington, D.C.: USAID
Glewwe, Paul, Michael Kremer, and Sylvie Moulin. 1998. "Textbooks and Test Scores: Evidence from a
Prospective Evaluation in Kenya." Development Research Group, World Bank, October 28.
Goodman, Margaret, Samuel Morley, Gabriel Siri, and Elaine Zuckerman. 1997. Social Investment Funds in
Latin America: Past Performance and Future Role. Washington D.C.: Inter American Development Bank.
March.
Government of Denmark. 1995. Methods for Evaluation of Poverty Oriented Aid Interventions. Copenhagen:
Ministry of Foreign Affairs.
Greene, W. H. 1997. Econometric Analysis. Hemel Hempstead, New Jersey: Prentice Hall Press.
Greenberg, David, and Mark Shroder. 1997. The Digest of Social Experiments, 2nd ed. Washington, D.C.: The
Urban Institute Press.
Grosh, Margaret E., and Juan Mufioz. 1996. "A Manual for Planning and Implementing the Living Standards
Measurement Study Survey." LSMS Working Paper #126. World Bank, Washington, D.C.
Grossman, Jean Baldwin. 1994. "Evaluating Social Policies: Principles and U.S. Experience." The World Bank
Research Observer 9 (July): 159-80.
Habicht, J. P., C. G. Victoria, and J. P. Vaughan. 1999. "Evaluation Designs for Adequacy, Plausibility and
Probability of Public Health Programme Performance and Impact." International Journal of Epidemiology 28:
10-18.
Harrell, Adele, Martha Burt, Harry Hatry, Shelli Rossman, Jeffrey Roth, and William Sabol. 1996. Evaluation
Strategies for Human Service Programs: A Guide for Policymakers and Providers. Washington, D.C.: Urban
Institute.
Heckman, James. 1997. "Instrumental Variables. A Study of Implicit Behavioral Assumptions Used in Making
Program Evaluations." Journal of Human Resources 32 (3): 441-61.
Heckman, James, R. Lalonde, and J. Smith. 1999. "The Economics and Econometrics of Active Labor Market
Programs." In 0. Ashenfelter and D. Card, eds., Handbook of Labor Economics, vol. III. New York: Elsevier
Science Publishing Co.
Heckman, James, and Richard Robb. 1985. "Alternative Methods of Evaluating the Impact of Interventions: An
Overview." Journal of Econometrics 30: 239-67.
88
Heckman, James J., and Jeffrey A. Smith. 1995. "Assessing the Case for Social Experiments." Journal of
Economic Perspectives 9 (2): 85-110.
Heckman, J., H. Ichimura, J. Smith, and P. Todd. 1998. "Characterizing Selection Bias Using Experimental
Data." Econometrica 66: 1017-99.
Hicks, Norman. 1998. Measuring the Poverty Impact of Projects in LAC. World Bank Latin America and the
Caribbean Region, Processed. Washington, D.C.: World Bank.
Holder, Harold D., Andrew J. Treno, Robert R Saltz, and Joel W. Grube. 1997. "Summing Up:
Recommendations and Experiences for Evaluation of Community-Level Prevention Programs." Evaluation
Review 21 (April): 268-78.
Hollister, Robinson G., and Jennifer Hill. 1995. Problems in the Evaluation of Community-Wide Initiatives. New
York: Russell Sage Foundation.
Hulme, David. 1997. "Impact Assessment Methodologies for Microfinance: Theory, Experience and Better
Practice." Manchester, U.K.: Institute for Development Policy and Management, University of Manchester.
International Food Policy Research Institute. 1998. Programa Nacional deEducaci6n, Salud, y Alimentaci6n
(PROGRESA): A Proposal for Evaluation (with technical appendix). Washington, D.C.: IFPRI.
Jalan, Jyotsna, and Martin Ravallion. 1998a. "Income Gains from Workfare: Estimates for Argentina's
TRABAJAR Program Using Matching Methods." Washington D.C.: Development Research Group, World
Bank. .1998b. "Transfer Benefits from Workfare: A Matching Estimate for Argentina." Washington, D.C.:
Development Research Group, World Bank.
Jamison, Dean T., Barbara Serle, Klaus Galda, and Stephen P. Heyneman. 1981. "Improving Elementary
Mathematics Education in Nicaragua: An Experimental Study of the Impact of Textbooks and Radio on
Achievement." Journal of Educational Psychology 73 (4): 556-67.
Karoly, Lynn A., Peter W. Greenwood, Susan S. Everingham, Jill Houb6, M. Rebecca Kilburn, C. Peter Rydell,
Matthew Sanders, and James Chiesa. 1998. Investing in Our Children: What We Know and Don't Know about
the Costs and Benefits of Early Childhood Interventions. Santa Monica, Calif.: Rand.
Kemple, James J., Fred Doolittle, and John W. Wallace. 1993. The National JTPA Study: Site Characteristics and
Participation Patterns. New York: Manpower Demonstration Research Corporation.
Khandker, Shahidur R. 1998. Fighting Poverty with Microcredit: Experience in Bangladesh. New York: Oxford
University Press for the World Bank.
Killick, Tony. 1995. IMF Programmes in Developing Countries, Design and Impact. London: England.
King, Elizabeth, and Berk Ozler. 1998. What's Decentralization Got to Do with Learning? The Case of
Nicaragua's School Autonomy Reform. Working Paper Series on Impact Evaluation of Education Reforms, paper
no. 9. Washington, D.C.: World Bank.
King, Elizabeth, Berk Ozler, and Laura Rawlings. 1999. Nicaragua's School Autonomy Reform: Fact or Fiction?
Washington, D.C.: World Bank.
Kish, Leslie. 1965. Survey Sampling. New York: John Wiley and Sons.
Levinson, James F., Beatrice Lorge Rogers, Kristin M. Hicks, Thomas Schaetzel, Lisa Troy, and Collette Young.
89
1998. Monitoring and Evaluation:A Guidebookfo r Nutrition ProjectM anagers in Developing Countries.
Medford, Mass.: International Food and Nutrition Center, Tufts University School of Nutrition Science and
Policy, March.
Manski, Charles, and Irwin Garfinkel, eds. 1992. Evaluating Welfare and Training Programs. Cambridge, Mass.:
Harvard University Press.
Manski, Charles, and Steven Lerman. 1977. "The Estimation of Choice Probabilities from Choice-Based
Samples." Econometrica 45: 1977-88.
McKay, H., A. McKay, L. Siniestra, H. Gomez, and P. Lloreda. 1978. "Improving Cognitive Ability in
Chronically Deprived Children" Science 2000(21): 270-78.
Meyer, Bruce D. 1995. "Natural and Quasi-Experiments in Economics." Journal of Business and Economic
Statistics, 13 (April): 151-161.
Miles, Matthew B., and A. Michael Huberman. 1994. Qualitative Data Analysis. London: Sage Publications.
MkNelly, Barbara, and Karen Lippold. 1998. "Practitioner-Led Impact Assessment: A Test in Mali." USAID
AIMS Brief 21. Washington, D.C.: USAID
MkNelly, Barbara, and Christopher Dunford (in collaboration with the Program in International Nutrition,
University of California, Davis). 1998. "Impact of Credit with Education on Mothers' and Their Young
Children's Nutrition: Lower Pra Rural Bank Credit with Education Program in Ghana." Freedom from Hunger
Research Paper No. 4, March.
Moffitt, Robert. 1991. "Program Evaluation with Non experimental Data." Evaluation Review 15 (3): 291-314.
Mohr, Lawrence B. 1995. Impact Analysis for Program Evaluation, 2nd ed. Thousand Oaks, Calif.: Sage
Publications. 1999. "The Qualitative Method of Impact Analysis." Processed.
Morduch, Jonathan. 1998a. "The Microfinance Schism." Harvard Institute for International Development,
Development Discussion Paper 626, February. 1998b. "Does Microfinance Really Help the Poor? New Evidence
from Flagship Programs in Bangladesh." Harvard University Institute for International Development, Processed,
June. . 1999. "The Microfinance Promise." Journal of Economic Literature 37(4): 1569-1614.
Newman, John, Menno Pradhan, Laura Rawlings, Ramiro Coa, and Jose Luis Evia. 2000. "An Impact
Evaluation of Education, Health, and Water Supply Investments of the Bolivia Social Investment Fund."
Processed.
Newman, John, Laura Rawlings, and Paul Gertler. 1994. "Using Randomized Control Designs in Evaluating
Social Sector Programs in Developing Countries." The World Bank Research Observer 9 (July): 181-202.
Operations Evaluation Department, World Bank. 1994. "An Overview of Monitoring and Evaluation in the
World Bank." June 30.
Poppele, J. S. Summarto, and L. Pritchett. 1999. "Social Impacts of the Indonesia Crisis: New Data and Policy
Implications." Social Monitoring Early Response Unit, World Bank, Washington, D.C. Processed.
Pradhan, Menno, Laura Rawlings, and Geert Ridder. 1998. "The Bolivian Social Investment Fund: An Analysis
of Baseline Data for Impact Evaluation." World Bank Economic Review 12 (3): 457-82.
Ravallion, Martin. 1994. Poverty Comparisons, Fundamentals in Pure and Applied Economics, vol. 56.
90
Harwood Academic Publishers. 1999. Monitoring Targeting Performance When Decentralized Allocations to the
Poor Are Unobserved. World Bank, Washington, D.C. Processed.
Ravallion, Martin, and Quentin Wodon. 1998. Evaluating a Targeted Social Program When Placement Is
Decentralized World Bank Policy Research Working Paper 1945. Washington, D.C.: World Bank.
Ravallion, Martin, Dominique van de Walle, and Madhur Gautam. 1995. "Testing a Social Safety Net." Journal
of Public Economics 57: 175-99.
Rawlings, Laura. 2000. "Assessing Educational Management and Quality in Nicaragua." In Bamberger:
Integrating Quantitative and Qualitative Methods in Development Research. Washington, D.C.: World Bank.
Rebien, Claus C. 1997. "Development Assistance Evaluation and the Foundations of Program Evaluation."
Evaluation Review 21 (August): 438-60.
Revenga, Ana, Michelle Riboud, and Hong Tan. 1994. "The Impact of Mexico's Retraining Program on
Employment and Wages." World Bank Economic R eview 8 (2): 247-77.
Rivers, Douglas, and Quang H. Vuong. 1988. "Limited Information Estimators and Exogeneity Tests for
Simultaneous Probit Models." Journal o f Econometrics3 9: 347-66.
Robinson, Sherman. 1991. "Macroeconomics, Financial Variables, and Computable General Equilibrium
Models." World Development,19(11): 1509-25.
Rosenbaum, P., and D. Rubin. 1983. "The Central Role of the Propensity Score in Observational Studies for
Causal Effects."
Biometrika 70: 41-55. 1985. "Constructing a Control Group Using Multivariate Matched Sampling Methods that
Incorporate the Propensity Score." American Statistician 39: 35-39.
Rossman, Gretchen B., and Bruce L. Wilson. 1985. "Numbers and Words: Combining Quantitative and
Qualitative Methods in a Single Large-Scale Evaluation Study. Evaluation Review 9 (3): 627-43.
Sahn, David, Paul Dorosh, and Stephen Younger. 1996. "Exchange Rate, Fiscal and Agricultural Policies in
Africa: Does Adjustment Hurt the Poor?" World Development, 24(4): 719-47.
Sebstad, Jennifer, and Gregory Chen. 1996. "Overview of Studies on the Impact of MicroEnterprise Credit."
USAID AIMS, June. Washington, D.C.: USAID
Sharma, Manohar, and Gertrud Schrieder. 1998. "Impact of Finance on Food Security and Poverty Alleviation-A
Review and Synthesis of Empirical Evidence." Paper presented to Workshop on Innovations in Micro-Finance
for the Rural Poor: Exchange of Knowledge and Implications for Policy, Ghana, November.
Smith, William. 1998. Group Based Assessment of Change: Method and Results. Hanoi, Vietnam: ActionAid.
Subbarao, K., Kene Ezemenari, John Randa, and Gloria Rubio. 1999. Impact Evaluation in FY98 Bank Projects:
A Review. World Bank Poverty Reduction and Economic Management Network, Processed, January.
Tan, J.P, J. Lane, and G. Lassibille. 1999. "Schooling Outcomes in Philippine Elementary Schools: Evaluation of
the Impact of Four Experiments." World Bank Economic Review, September.
Taschereau, Suzanne. 1998. Evaluating the Impact of Training and Institutional Development Programs, a
Collaborative Approach. Economic Development Institute of the World Bank, January. USAID. 1984. "Study of
91
Low Cost Housing." New Delhi: USAID. June 1995. "Assessing the Impacts of MicroEnterprise Interventions:
A Framework for Analysis." MicroEnterprise Development Brief, June.
Valadez, J., and M. Bamberger, ed. 1994. "Monitoring and Evaluating Social Programs in Developing
Countries." Economic Development Institute of the World Bank Series, World Bank, Washington, D.C.
van de Walle, Dominique. 1999. "Assessing the Poverty Impact of Rural Road Projects." World Bank. Processed.
Walker, Ian, Rafael del Cid, Fidel Ordofiez, and Florencia Rodriguez. 1999. "Evaluaci6n Ex-Post del Fondo
Hondurefio de Inversi6n Social (FHIS 2)." ESA Consultores, Tegucigalpa, Honduras.
Weiss, Carol H. 1998. Evaluation. Upper Saddle River, New Jersey: Prentice Hall.
Wilson, Sandra Jo. 1998. "A Multi-Method Proposal to Evaluate the Impact of CGIAR-Instigated Activities on
Poverty Reduction." Paper presented to the Annual Conference of the American Evaluation Association,
November.
World Bank, Poverty and Human Resources Division, Policy Research Department. 1994a. Impact Evaluation of
Education Projects Involving Decentralization and Privatization. Working Paper. Washington, D.C.: World Bank.
Learning and Leadership Center and Operations Policy Department. 1996. Handbook on Economic Analysis of
Investment Operations. Washington, D.C.: World Bank.
* , Operations and Evaluation Department. 1998a. 1998 Annual Review of Development Effectiveness.
Washington, D.C.: World Bank. , Poverty Reduction and Economic Management, Latin America and Caribbean
Region. 1998b. Measuring the Poverty Impact of Projects in LAC. Washington, D.C.: World Bank. Operations
and Evaluation Department. 1999. World Bank Agricultural Extension Projects in Kenya: An Impact Evaluation.
Report No. 19523. Washington, D.C.: World Bank. Poverty Reduction and Economic Management Unit, 1999.
Strategy for Impact Evaluation. Washington, D.C.: World Bank.
Wouters, A. 1995. "Improving Quality through Cost Recovery in Niger." Health Policy and Planning 10 (3):
257-70.
Zeller, Manfred, Akhter Ahmed, Suresh Babu, Sumiter Broca, Aliou Diagne, and Manohar Sharma. 1996.
"Security of the Poor: Methodologies for a Multi country Research Project." IFPRI Discussion Paper 11.
The findings, interpretations, and conclusions expressed in this paper are entirely those
of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations, or
to members of its Board of Executive Directors or the countries they represent. The World Bank does not
guarantee the accuracy of the data included in this publication and accepts no responsibility for any consequence
of their use. The material in this publication is copyrighted. The World Bank encourages dissemination of its
work and will normally grant permission to reproduce portions of the work promptly.
Permission to photocopy items for internal or personal use, for the internal or personal
use of specific clients, or for educational classroom use is granted by the World Bank, provided that the
appropriate fee is paid directly to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA
01923, USA.; telephone 978-750-8400, fax 978-750-4470. Please contact the Copyright Clearance Center before
photocopying items.
For permission to reprint individual articles or chapters, please fax a request with complete information to the
Republication Department, Copyright Clearance Center, fax 978-750-4470.
All other queries on rights and licenses should be addressed to the Office of the Publisher, World Bank, at the
92
address above or faxed to 202-522-2422.
ISBN 0-8213-4697-0
Library of Congress Cataloging-in-Publication Data
Baker, Judy L., 1960-
Evaluating the impact of development projects on poverty: a handbook for practitioners / Judy L. Baker
p. cm. - (Directions in development)
Includes bibliographical references.
ISBN 0-8213-4697-0
1. Economic development projects-Evaluation-Handbooks, manuals, etc.
2. Poor-Developing countries. I. Title II. Directions in development (Washington, D.C.)
HD75.9 .B35 2000
338.9'0068'4-dc2l 00-028325