1 Introduction
Despite the façade of precise estimates, we do
not really know what has been happening to
global poverty all things considered, particularly
when measured against rising costs for education
and healthcare, which are mostly not factored
into poverty measures, for technical reasons.
However, the mechanisms by which the
Millennium Development Goals (MDGs) might
have effected poverty reduction are not at all
clear. The recent literature on the impact of aid
on growth offers little useful insight given that it
largely ignores the major global structural
processes that condition the impact of aid flows,
such as those reflected by global balance of
payments asymmetries. Moreover, the emphasis
in the MDGs on absolute measures and the
implicit bias towards targeting quite possibly
undermine poverty reduction in many contexts,
particularly if poverty is primarily considered as
an outcome of dynamic processes of social
stratification and subordination.
Serious consideration of the erosion of decent
employment and wages and the increasing
segmentation of social security systems
throughout the developing world is particularly
needed if we are to truly embrace a pro-poor
agenda, i.e. not one that merely reduces absolute
poverty regardless of inequality, but one that
actually promotes equitable sharing without
double standards.
2 Measuring poverty reduction
At the outset, the impact of the MDGs on poverty
reduction, here defined as income poverty as per
Goal 1 of the MDGs, is very difficult to assess
because poverty reduction itself is very difficult to
measure. The international estimates provided by
the World Bank are highly contested, even
though they are the main data relied upon by the
leading international organisations spearheading
the promotion of the MDGs. This received much
attention through the debate between Reddy and
Pogge (2002a,b) and Ravallion (2002), which was
furthered by Wade (2004) among others. Much of
the contention relates to the way that the
purchasing-power-parity (PPP) poverty line had
been calculated and standardised. Together with
other considerations, Reddy and Pogge (2002a)
conclude that the World Bank’s estimates of the
magnitude, distribution or trend of global
income poverty are neither meaningful nor
reliable.
36
Towards Genuine Universalism
within Contemporary Development
Policy
Andrew M. Fischer
*
Abstract It is very difficult to know the impact of the MDGs on poverty reduction. On the one hand,
poverty measurements are ambiguous, arbitrary and contested, even in the best of cases such as China and
India. On the other hand, the mechanisms by which MDGs might have effected poverty reduction are not at
all clear, particularly in light of the major global structural processes that condition the impact of aid flows
and development more generally. Moreover, the emphasis in the MDGs on absolute measures and the
implicit bias towards targeting quite possibly undermine poverty reduction in many contexts. Hence, this
article argues that the MDGs should be replaced by a re-politicisation of the mainstream development
agenda, together with a genuine revival of emphasis on universalistic modes of social policy as viable means
of dealing simultaneously with poverty and inequality.
IDS Bulletin Volume 41 Number 1 January 2010 © 2010 The Author. Journal compilation © Institute of Development Studies
Published by Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA
This is just one example of the contention
underlying the façade of precision in the field of
poverty studies. Given the wide variety of fairly
arbitrary assumptions and choices that are
required in order to first choose a line and then
to apply this line to presumably accurate survey
data in ways that are broadly consistent,
accurate and comparable across time and
regions, it is no surprise nor exaggeration that
poverty estimates themselves are quite arbitrary,
even before considering World Bank PPP
adjustments. Saith (2005) eloquently synthesises
the perils, although practitioners of poverty
measurement have acknowledged the pitfalls for
decades. Indeed, in the opening pages of Poverty
Comparisons, Ravallion (1992: 2–3) admits that
poverty lines are arbitrary, albeit he argues that
even arbitrarily chosen lines nonetheless allow
for comparison and evaluation so long as they are
accurately adjusted over time.
However, the issue of adjustment leads to the
classic quandary that we simply do not know
whether poverty trends over time reflect actual
changes or else errors of adjustment. The
quandary is all the more problematic given that
the incomes of populations in poor countries
tend to be densely clustered around a typical
line, resulting in an extreme sensitivity of
poverty estimates to small adjustments to the
line, whether or not these are accurate. For
example, Székely et al. (2000) apply sensitivity
analysis to household survey data from 17 Latin
American countries in the 1990s. By varying the
poverty line parameters within reasonable
boundaries, they estimated poverty rates as lying
anywhere from 12.7 per cent to 65.8 per cent of
the total population. Moreover, the ranking of
countries with respect to poverty rates was also
highly sensitive to their exercise. Hussain (2003)
demonstrates a similar point with respect to
urban poverty in China regarding both
sensitivity and rank orderings across Chinese
provinces. Helwege and Birch (2007: 6) note
that, ironically, ‘the institutions that generate
poverty data are well aware of how
methodological choices affect poverty estimates.
They simply have not established standardised
approaches to measuring poverty’.
Moreover, poverty trends can differ depending on
where or how a line is drawn. For instance,
poverty was decreasing in China from 1998 to
2000 according to the unreasonably-low official
poverty line (i.e. the one often cited in World
Bank publications in the early 2000s), but it was
rising according to the more reasonable absolute
poverty line calculated by the Chinese National
Bureau of Statistics (Hussain 2001, cited in
Fischer 2005: 96–9). Similarly, based on their
efforts to correct inconsistencies in Indian survey
data from 1999–2000, Himanshu and Sen (2004)
conclude that there had been little poverty
reduction in India in the 1990s, contrary to the
dominant consensus. Deaton and Kozel (2005:
117) contend that such claims are ‘frankly
political’ and that there is good evidence that
poverty fell. However, further findings presented
by Himanshu (2007), based on new data from
2004–05, confirmed his earlier results for the
1990s, alongside other studies that repeatedly
found that calorie deprivation had actually
increased in India in the 1990s (e.g. see
Meenakshi and Vishwanathan 2003). Likewise,
in their assessment of alternative poverty
estimates from the World Bank and the United
Nations for Latin America, Helwege and Birch
(2007) advise caution in interpreting trends from
any of these data.
Our ability to track poverty trends over time is
critically based on our presumption that we can
accurately measure all of the changing cost
factors faced by poor households together with
their changing patterns of livelihood and
consumption, in contexts of often rapid
structural change. Notably, the World Bank
recently revised upwards its estimates for global
PPP poverty rates, including an upward
adjustment of about 40 per cent for China. This
was based on new and improved cost of living
data from 2005, which revealed a substantially
higher cost of living for the poor in developing
countries than was previously estimated from
older data. However, Chen and Ravallion (2008)
stress that, although poorer than previously
thought, the world was no less successful in
poverty reduction. They reach this conclusion
simply by deflating the new PPP poverty lines by
the official consumer price indices of each
country back to 1981 (Chen and Ravallion 2008:
14–15). In other words, the fact that the
resultant trends were the same as before is
merely an artefact of their assumption that the
source of error was the same in 1981 as it was in
2005. In so doing, they completely sidestep the
question of whether the poor faced greater cost
of living increases than suggested by the general
IDS Bulletin Volume 41 Number 1 January 2010
37
consumer prices indices, which is quite possible
given the notable increases in inequality in many
countries over the same period.
1
Thus, while
Chen and Ravallion provide a politically
convenient narrative, many intractable problems
remain lost in past surveys.
Arguably, the fundamental Achilles heel of the
income poverty approach is that education and
health costs are mostly not included in the
calculation of poverty lines. The exclusion is for
technical reasons, given that they constitute
large and highly irregular expenditure items
across households and across time. However, they
are included in the expenditures of surveyed
households. This renders the comparison of
poverty rates very difficult across households, let
alone across regions with different provisioning
systems or else across time when the costing and
supply of education or healthcare changes.
Increasing costs of education or healthcare, or
else increasing school enrolments in schools that
charge fees, would be invisible to most
conventional poverty measures even though they
effectively raise the poverty line for a large
proportion of households, leading to an
important source of underestimation of poverty
rates in such contexts. This weakness is
recognised in some of the literature, but it is also
generally sidestepped in the same literature (e.g.
see Ravallion 1992: 12, 28).
The innovative work by van Doorslaer et al. (2005)
is one exception in this regard. Merely by
deducting catastrophic out-of-pocket payments
for healthcare from the expenditures of
households surveyed in 11 low- and middle-
income countries in Asia (most surveys taken
around 2000), they show that poverty rates across
Asia increased from 19.3 per cent to 22 per cent,
or an increase of 78 million people. Notably, this
does not take into account the repressed
expenditure of poor people who would have
otherwise spent more on necessary healthcare
were it not for lack of funds, which is a problem
that the authors consider but cannot measure.
As insightful as this work is, it is nonetheless
based on insights from single surveys taken at
one particular point in time. The results offer no
indication on how these considerations might
alter our perception of trends over time,
particularly in cases where education and
healthcare costs have increased substantially. For
instance, falling income poverty rates in China
since the beginning of the reform period do not
factor into consideration the parallel shift from
very cheap to very expensive healthcare. More
specifically, while general consumer prices in
China remained more or less unchanged between
1997 and 2004, tuition fees and healthcare
services were extremely inflationary, especially in
some of the poorest western provinces such as
Qinghai, where prices for healthcare services
more than quadrupled (see Fischer 2005: 48). By
ignoring such dramatically changing price
structures (as do Chen and Ravallion 2008), we
simply do not know to what degree the
appearance of improving poverty rates merely
represents increasing relative prices for these
essential services not included in the poverty line.
It is in this sense that the exclusive focus on
conventional absolute income poverty measures
in Goal 1 of the MDGs can be said to be biased
against universalistic modes of social policy. A
movement towards free education or healthcare
financed through progressive taxation would not
necessarily appear as decreasing income poverty
even though it would lower the effective poverty
line for households previously paying fees.
Similarly, it is difficult to calibrate poverty rates
in any meaningful way across countries with very
different provisioning systems, such as between
Cuba with its free healthcare and education, and
Vietnam, which had the greatest reliance on out-
of-pocket payments among the Asian cases
studied by van Doorslaer et al. (2005). Indeed,
much of the appearance of rising incomes in the
latter case might actually signify deteriorating
social wealth.
3 Global structural processes and aid
The other side of the question on the poverty
impact of the MDG paradigm is that the
mechanisms by which MDGs might have effected
poverty reduction are simply not clear. Notably,
the major episodes that account for a large part
of the commonly cited absolute income poverty
reduction over the last several decades had little
to do with MDGs, for example China in the
1980s or the financially driven global economic
boom of 2002–07. The mechanisms that could
continue to drive poverty reduction now that this
boom has bust are even less clear and the MDGs
provide little guidance. Hope is placed on the aid
system. However, the recent literature on the
impact of aid on growth and poverty offers little
Fischer Towards Genuine Universalism within Contemporary Development Policy
38
useful insight, given that attention is rarely given
to the major global structural processes that
condition the impact of aid flows and
development more generally.
On this last point, as argued in Fischer (2009),
there is no sense in discussing whether aid is
good or bad for development, or whether more or
less aid is required for development, outside of a
much broader understanding of the key factors
underlying cases of successful development up to
the present, namely industrialisation and large
sunk investments in urban and other
infrastructure. These in turn usually necessitate
that developing countries run trade deficits
rather than surpluses. In an earlier
developmental epoch, aid was understood as
enabling poor countries to cover such trade
deficits, as best represented by the crucial role of
aid in the balance of payments of South Korea
well into the 1970s. The recent exceptionality of
China might seem to counter this rule, although
this is debatable. In any case, its replication or
sustainability is questionable if only because the
world economy has probably reached its limits of
disequilibria.
The role of aid must be considered in this light
because aid will have very different
macroeconomic implications depending on
whether a receiving country is in current account
deficit or surplus. Moreover, even if a country is
in deficit, aid would presumably have very
different effects on growth depending on
whether the deficit is due to productive
investment and accumulation, or simply due to
terms of trade and other contractionary shocks,
or to austerity and structural adjustment
programmes (SAPs). Indeed, many of the
arguments on aid and growth are rendered
trivial by their lack of consideration of these
broader structural considerations.
Similarly, it is meaningless to argue that large
increases in aid would produce better
development (e.g. Sachs 2005) without examining
the international mechanisms set in place to
deliver aid, which for much of the last three
decades have encouraged a haemorrhaging of
human, physical and financial resources from poor
to rich countries. The same can be said of
arguments that aid effectiveness could be
improved by improving incentive mechanisms
within poor countries (e.g. Easterly 2006),
irrespective of the globalised incentive
mechanisms that play a strong role in disorienting
local incentives. The frivolity of these arguments
is best highlighted by the massive US current
account deficit, many times the size of the annual
aid budget of the Organisation for Economic Co-
operation and Development (OECD) countries
and largely financed by the developing world.
More pertinent is the fact that, in the midst of a
crushing depression in the 1980s and into much
of the 1990s, Africa was experiencing net
outflows of capital that far exceeded any inflows
of aid. Collier (2007: 91–3) discusses this point to
some extent but does not place it within the
context of the systemic shifts in global economic
structures and ideologies that produced these
outcomes, with the result that he conflates many
symptoms as causes. This is expected given that
he broadly positions himself within the
hegemonic paradigms and institutions of this
period, as do Sachs and Easterly. Indeed, Collier
(2007) mostly avoids mention of the 1982 debt
crisis and seems to imply that SAPs themselves
constituted good economic policy.
Similar obfuscations are committed throughout
the cross-country regression literature on aid.
Data from two entirely different epochs of
development are usually merged together
without differentiating the radically different
structured settings in which aid might or might
not work. Crucial in this regard is the difference
between the epoch of developmentalism up to
the 1970s, when most Southern countries
experienced decent economic performance,
versus the current period that started in the
early 1980s, after which Southern performance
has been much worse outside of East and South
Asia and aid flows have become a trickle in
comparison to successive waves of capital flows
from poor to rich countries. Aid has since been
futile in producing any significant degree of net
global redistribution.
In the absence of these broader considerations,
the MDG agenda has tended to encourage an
over-inflated sense of self-importance among
international donors about the role that aid can
play in driving development. This exaggeration
is all the more evident when we consider that
probably the majority of aid flows return to their
donors, if ever they even leave the shores of their
donor, such as in the case of technical assistance,
IDS Bulletin Volume 41 Number 1 January 2010
39
consultancies and research funding. ActionAid
(2005) has produced some interesting research
on the uses and abuses of such ‘phantom aid’.
Academics should take these issues much more
seriously, although perhaps their avoidance
derives from fear of biting the hands that feed.
This being said, there are particular moments
when the aid industry does play a very pivotal
role. These are in times of economic crisis, when
developing countries start to face balance of
payments problems and potential financial crises.
The aid industry is then suddenly possessed with
much power, particularly gatekeepers such as the
World Bank and the International Monetary
Fund (IMF). If guided by an orthodox monetarist
agenda, this power can be wielded in punitive
destructive ways. This is precisely what happened
during the 1982 debt crisis. While the crisis was
fundamentally rooted in systemic disequilibria
stemming from the breakdown of the Bretton
Woods system, the ideological recasting of this
crisis as one of irresponsible borrowing by inept
and corrupt Southern governments and their host
of rent-seeking elites fashioned the justification
for policies that shifted most of the burden of
adjustment from Northern financial sectors to
Southern publics. The recasting also helped to
consolidate the rise of what is now called the
Washington Consensus. Similar dynamics
unfolded with the East Asian financial crisis in
1997–8 and the consolidation of the so-called
Post-Washington Consensus. Similar dangers
exist today.
4 Towards re-politicisation and genuine
universalism
In terms of what should replace the MDGs after
2015, it is important to recall three main
criticisms of the MDGs: that they do not pay
attention to employment, or inequality, and that
they de-politicise development debates.
2
Of these
three, de-politicisation arguably underlies the
weakness of the MDGs with respect to
employment and inequality.
The reasons for this might seem semantic, but
they are not. They are rooted in the fact that
policy choices are very political, even though
these choices are often couched in seductively
technocratic terms. De-politicisation then serves
to veil underlying agendas and allows
paradigmatic shifts in theory and practice to be
hidden behind principles of charity and altruism.
More specifically, the emphasis in the MDGs on
absolute measures and the implicit bias towards
targeting predisposes the MDG agenda to be co-
opted by an orthodox approach to development
policy.
In this regard, there is much truth to the claim
that the MDG agenda has been embedded within
the Washington Consensus (including its various
‘post’ reiterations). It is important to recall that
addressing poverty is not particularly antithetical
to this consensus. Indeed, the World Bank
dedicated its World Development Report specifically
to the theme of poverty for the first time in 1990.
Lipton et al. (1992) famously called this the ‘New
Poverty Agenda’. Accordingly, poverty is quite
comfortably explained by way of market
imperfections. Poverty persists precisely because
markets do not function efficiently, causing
market failures (i.e. involuntary unemployment
or incomes below potential). In times of
deregulatory zeal, it is usually argued that these
imperfections are largely caused by witless or
devious government interventions resulting in
countless economic distortions in need of
structurally adjusting. Until such adjustments
are completed, targeted safety nets should be
provided for those who fall through the cracks,
provided they are deemed as deserving.
Targeting then necessitates accurate poverty
measurement. The Post-Washington Consensus
basically adds the weakness of market-
supporting institutions (i.e. insecure property
rights, poor enforcement of contracts,
clientelism, corruption, etc.) to the list of reasons
why markets fail without challenging the
underlying logic. In other words, talking about
poverty does not necessarily signal a shift away
from this paradigm. It is often symptomatic of a
more conservative agenda that emphasises
charity and paternalism, versus more progressive
(i.e. redistributive) agendas that emphasise
equality and employment.
3
Similar tendencies exist in the more recent
iterations of the poverty agenda. For example,
multidimensional conceptions of poverty (which
are not at all new, although they have come to be
treated as such) can be easily co-opted into a
supply-side policy approach, more or less
Washington Consensus in nature, i.e. avoid
demand management and focus instead on
allowing supply to create its own demand,
particularly through the operation of efficient
Fischer Towards Genuine Universalism within Contemporary Development Policy
40