N.B. This is a draft document. Please do not quote without the author’s permission
(email@example.com) Comments are very welcome.
Amartya Sen’s capability approach and poverty analysis
Rod Hick, PhD Candidate, London School of Economics
Paper presented at the Social Policy Association conference, Edinburgh, 1 July 2009.
The Capability Approach, developed and pioneered by the economist Amartya Sen, provides
a conceptual framework for analysing well-being and a strong critique of existing traditions
in welfare economics. The central tenet of the approach is that the appropriate „space‟ in
which to conceptualise and measure well-being is not in terms of primary goods or in utilities
(whether in the form either of happiness or preferences) but rather in terms of a persons
capabilities; that is, in the real freedoms that they have reason to value (1999: 74). While the
development and initial application of the approach occurred in the fields of welfare
economics and development, it is growing in popularity with academics from a wide range of
disciplines. However, is the approach useful to those of us who are interested in social policy
issues, and in particular to those interested in the conceptualisation and measurement of
poverty, social exclusion and related issues? How great a departure does it entail from
existing social policy traditions? Furthermore, do well-known problems with
operationalisation render it of limited use for applied policy analysis?
This paper comprises four sections. The first will present an outline of the capability
approach and its central concepts. The second will compare the approach to three existing
traditions of social policy analysis (i) the income poverty tradition, the (ii) deprivation
indicators tradition and (iii) the social exclusion approach. Third, the paper will consider the
most contentious of the concepts within the CA itself – that of capability – and suggest a
potential operationalisation that is both feasible and adequate for the requirements of social
policy research. Finally, the paper will consider the conceptual ground covered by the
capability approach and will briefly consider how this can be reconciled within social policy
Section 1: The Capability Approach
The central concepts: Functioning, Capabilities and Capability
The primary concepts of the capability approach are functionings and capabilities. Sen‟s
concept of functioning refers to the various things a person may succeed in „doing or being‟
(1999: 75); that is, a person‟s achievements in terms of (primarily) objective well-being,
while capabilities refer a person‟s real or substantive freedom to achieve such functionings 2
(1999: 73). Thus, functionings can be viewed as the various outcomes a person may achieve
(being healthy, participating in social activities, and so forth), while capabilities refer to the
real, as opposed to formal, opportunities to achieve these outcomes (the ability to be healthy,
the ability to participate in society, and so on). The various capabilities a person may possess
are components of their overall capability, which is conceived as a set which „reflects the
alternative combinations of functionings the person can achieve, and from which he or she
can choose one collection‟ (1993: 31). Thus, while rarely drawn as a clear distinction in the
literature, a person may be seen as possessing a range of capabilities (opportunities) which,
combined, comprise their overall capability (Gasper, 1997). The distinction between
functionings and capability is thus between „achievements on the one hand, and freedoms or
valuable options from which one can choose on the other‟ (Robeyns, 2005b: 95) and Sen
views the process of development as a process of expanding the real freedoms that people
enjoy (1999: 36). Despite the rather abstruse terminology, there is a clear link to a wider
literature on outcomes and opportunities. For this reason, and to make this link salient, I will
on occasion include „outcomes‟ and „real opportunities‟ in parentheses when discussing
functionings and capabilities respectively.
In order to highlight the distinction that Sen makes between the capability approach and
alternative approaches to well-being, it may be useful to draw further on the concepts of
commodities and the characteristics. Any commodity is viewed a „possessing‟ a range of
characteristics. For example, food may offer the characteristic of nutrients; a car the
characteristic of transport. Thus, „[S]ecuring amounts of these commodities gives the person
command over the corresponding characteristics‟ that the goods provide (Sen, 1987: 6).
However, even if we know the commodities possessed by individuals, this does not tell us
what individuals are able to do with these commodities; that is, the functionings that they are
able to achieve because the rate at which individuals are able to convert these characteristics
into functionings will vary due to differences in individual, social and environmental factors
(1999: 70-72). For example, a fixed amount of food (nutrients) will provide different levels of
functioning to a child compared to an adult. Furthermore, the functionings themselves may
also result in utility, such as the enjoyment that a person may derive from cycling a bike (Sen,
1983). Given that our interest in commodities is of instrumental importance in comparison
with the intrinsically important nature of functionings, it may be argued that to focus on
commodities rather than on what they can allow us to do or be is succumb to „commodity
fetishism‟ (Sen, 1987: 18) while utility is deemed to be an inadequate measure due to its
adaptive nature and limited relationship to objective well-being. Thus in the chain from
commodities – characteristics – functioning – utility, Sen argues that ethical evaluation
should be concerned with a person‟s capabilities and functionings, which with the latter
considered to be constitutive components of their well-being (1992: 39).
In conceptualising capability as a set of the alternative combinations of functioning a person a
person is able to achieve, Sen‟s concept is analogous to an opportunity set in the field of
social choice theory. The vectors themselves contain only valuable functionings and the
ability to exercise some choice is itself considered to have value. Thus, seen in this light, „the
value of a set can be reduced when the number of elements is reduced (it is also determined 3
by having opportunities that are not chosen) but it could not be enhanced by an increase of
trivial choices (see Sen, 1992: 63).‟ (Comim, 2008: 165). Drawing on the notion of positive
freedom, the concern is not merely with formal or legal freedoms but rests on the stress on
the real or substantive freedom of people to live a life that they value. Thus, Sen argues that
„[d]evelopment requires the removal of major sources of unfreedom: poverty as well as
tyranny, poor economic opportunities as well as systematic social deprivation, neglect of
public facilities as well as intolerance or overactivity of repressive states (Sen, 1999: 3).
Operationalisation of the approach
It seems almost duplicitous to discuss the capability approach, let alone in some way
advocate it, without addressing issues that exist with regard to its operationalisation, which
has become a major issue (some might say the major issue) for those working with the
approach. Put simply, Sen has neither provided a list of capabilities and their respective
weightings nor suggested the method by which this might occur, save that the choices
involved should be explicit and should be subject to public participation and scrutiny. His
decision to leave the approach „deliberately incomplete‟ has been criticised by some who
argue that he needs to provide such a list of important capabilities or, at least, greater
guidance here (e.g. Nussbaum, 2000a,b). In addition to determining which capabilities we
will be interested in, some authors have argued that it is not at all clear how capabilities can
be measured directly (Comim, 2008; Krishnakumar, 2007), particularly due to the inclusion
of hypothetical examples or non-chosen alternatives that the approach deems constitutive of
an individual‟s well-being (Lelli, 2008). Basu has suggested that Sen‟s concepts are more
complex and difficult to measure that he acknowledges (1987: 71). As regards the selection
of capabilities, Sen has argued that „to have such a fixed list, emanating from pure theory, is
to deny the possibility of fruitful public participation on what should be included and why‟
(2006: 362) and that „pure theory cannot „freeze‟ a list of capabilities for all societies for all
time to come, irrespective of what the citizens come to understand and value‟ (2006: 363). He
has thus criticised the idea of the fixed list of central capabilities, or what he has described as
„a giant mausoleum to one fixed and final list of capabilities‟ (2006: 365).
Section 2: Relationship to social policy traditions
Despite the use of new terminology, we may legitimately question the extent to which the
approach differs from some existing traditions in Social Policy. This section will attempt to
briefly highlight some differences between three traditions of social policy analysis: the
income poverty approach, relative deprivation/deprivation indicators approach, and the social
Relative income poverty tradition
As part of his critique of using income as a measure of well-being, Sen has criticised standard
income poverty measures for focussing on an instrumental variable rather than ends that are
intrinsically important. This draws upon the notion of intrinsic ends and instrumental means 4
that is discussed by Aristotle in his Nichomachean Ethics: „[c]learly, then, things can be
called good in two senses: some as good in their own right, and others as a means to secure
these. Let us, then, separate the things that are good in themselves from those that are merely
useful‟ (Aristotle, 1953: 11-12). However, in addition to this distinction, Sen has criticised a
focus on income because the rate at which individuals will be able to convert a given amount
of income into intrinsically important outcomes (functionings) will vary between of the
different needs that individuals and groups have.
Furthermore, while there may be advantages in an income measure, including its relative
simplicity and summary nature, such a measure is inherently unidimensional and cannot
recognise the plural nature of well-being (and deprivation). Indeed, such a measure may be
particularly poor at highlighting situations where some dimensions of deprivation intensify
despite rising income or where little relationship is found between income and a particular
While traditional relative poverty lines (say, at 60% of median income) are genuinely relative
and may provide information about a chosen standard over time, they have been criticised on
the grounds that they do not relate to any concept of individual need (Gordon, et al., 2000),
but remain popular largely due to their simplicity (Ravillion, 1996) which may facilitate
public discussion about the issues of inequality and poverty in the public domain more easily
than more complex measures. The most common income poverty measure used in the social
policy field is undoubtedly the headcount ratio, H, and Sen has been particularly critical of
this measure, which he described as being „obviously a very crude index‟ and has criticised
for being „completely insensitive to the distribution of income among the poor‟ (1976: 219).
Indeed, there are many other widely appreciated problematic features of the headcount
measure including that while it measures poverty elimination, it fails to register poverty
alleviation in any way, and thus creates incentive effects for government to focus resources
on the poor who are least in need (in order to bring them that short distance over the poverty
line) (Zheng, 1997; Ravillion, 1996), though the simultaneous use of a number of poverty
lines may provide greater understanding of the income distribution than relying on one
poverty line in isolation.
Nonetheless, despite the variety of income income-based measures, the capability approach is
clearly distinct from them all in terms of its recognition of the multidimensionality of wellbeing, focus on intrinsic ends rather than instrumental means, consideration of non-income
influences on well-being and the effects of differing needs on the conversion rates from
income to functionings (outcomes).
An alternative tradition of „direct‟ measurement of poverty lies in the use of deprivation
indicators. These typically include ask respondents whether they are unable to afford a range
of consumer goods and activities that are either possessed by a majority of the population or
are deemed necessities (e.g. Callan et al., 1993) While there are broad similarities in the 5
deprivation items chosen by different authors and collected by major social surveys, there is
less agreement about how to use such information to determine a poverty line, with a variety
of approaches adopted (e.g. Townsend, 1979; Gordon et al., 2000; Saunders and Adelman,
2006; Borooah, 2007; Whelan and Maitre, 2007; Whelan, 2007). An example of two
deprivation questions from the EU Survey on Income and Living Conditions is as follows:
Deprivation questions from EU Survey on Income and Living Conditions
Source: Central Statistics Office website (2009)
In terms of the question wording, we can see from the questions above that the indicators
used go beyond merely asking whether such activities occur, but also to ask whether the
reason for their non-occurrence is because of a lack of resources. This formulation follows
the influential critique by David Piachaud of Peter Townsend‟s deprivation index, where he
noted that preferences should not be conflated with poverty and that voluntary nonconsumption or non-participation should be removed from our calculus (1981). It should be
noted that this critique bears a close resemblance to Sen‟s distinction between functionings
(outcomes) and capabilities (real opportunities). However, in most of our current social
surveys the questions go beyond asking whether important activities occur, but stop short of
focussing on all of the constraints that might prevent individuals from participating in them:
rather, they focus on one particular constraint: that of resouces. Other constraints such as
disability, ill-health, discrimination, geographical isolation, inter alia, are simply not
considered. Thus, in terms of activities, the result of the Piachaud critique has been to shift to
something that may be seen as „quasi-capabilities‟ rather than the broader concept of
capabilities as outlined by Sen.
The question of whether we are interested in constraints of other kinds is a theoretical matter,
and the first question that may arise is whether we would be interested in additional
constraints at all. A second, and perhaps more searching question is whether, if we deem such
constraints of interest, we would wish to include these in our definition of poverty, or
whether we would want to draw on another concept in order to account for these.
Clearly, there are similarities between the capability approach and the deprivation indicators
approach. However, some differences are worth noting. First, deprivation indicators typically
include questions both of possession of commodities and participation in activities (beings or
doings). These receive distinct treatment in the capability approach with only the latter
deemed to be of intrinsic interest and the former requiring additional information before
possession information can be converted into functioning information due to the varying
conversion factor between income and commodities (inputs) and functionings (outcomes)
Does the household have family or friends for a drink or a meal once a month?
(If no, it is because the household can not afford to or is there another reason?)
Do the household members have hobbies or leisure activities?
(If no, is it because the household can not afford to or is there another reason?)6
that people will face. I do not propose to explore this issue here. Second, the wider range of
constraints on functioning that the approach would allow for, and indeed how these would be
integrated requires further consideration of its conceptual implications, to which we shall
Social exclusion approach
The concept of social exclusion has received substantial attention in recent years, typically as
a concept distinct but allied to that of poverty. Indeed, its rise to political currency has been
remarkable and used by, inter alia, the European Union and the UK government, which set
up a dedicated Social Exclusion Unit in 1997. The concept (or at least the terminology) of
social exclusion was first used by René Lenoir in France in the 1970s to identify those who
were falling through France‟s insurance-based social security programmes. This included
„the mentally and physically handicapped, suicidal people, aged invalids, abused children,
drug addicts, delinquents, single parents, multi-problem households, marginal, asocial
persons, and other “social misfits”‟ (Silver, 1994: 532). While such categories may indeed
include many vulnerable individuals, such a conceptualisation was derived from Republican
ideology which stressed the importance of social cohesion, which was challenged by
exclusion from such groups from this form of social provision (Evans, 1998). Seen in this
way, exclusion is a result of a break down in solidarity, which must be rectified by
„integration‟ or „insertion‟ (Silver, 1994: 541-2). Thus, it should be noted that the original
understanding of the concept was not directly in terms of low individual well-being, but
rather on the importance of cohesion and solidarity.
The social exclusion concept remains highly contested with the question of exclusion from
what? generating a multiplicity of answers. For some, the concept implies exclusion from
welfare institutions (see Room, 1995b: 243), the response to which is citizenship rights. For
others, it allows a focus on inability to participate in society for a wider range of constraints
then is usually considered in poverty research: it includes „discrimination, chronic ill-health,
geographical location, or cultural identification, for example‟ (Burchardt et al., 2002a). For
Berghman (1995), social exclusion is not an outcome at all, but rather a process, the outcome
of which is deprivation, conventially defined. Similarly, Paugam sees the process of
exclusion as one of „social disqualification‟ where deprivations multiply and one‟s
relationship to social life and relations becomes increasingly tenuous (Silver, 1994) where he
resulting exclusion may be irreversible.
Thus, one of the challenges in locating the social exclusion approach with respect to the
capability approach is the extremely contested nature of the social exclusion concept (Silver,
1994; de Haan, 1998). Sen himself has argued that the approach can be seen as a subset of
capability deprivation, with the concept focussing on the „relational features in the
deprivation of capability and thus in the experience of poverty‟ (Sen, 2000: 6). Thus, the
question of exclusion from what? generates different answers depending on national tradition
or political thought (de Haan, 1998: 17) and within individual traditions has been developed
and broadened over time. There might be widespread agreement on the need to fight 7
exclusion, but „fighting exclusion means different things to different people‟ (Silver, 1994:
It is argued that a settled view of the concept is ultimately required if it is to be useful as an
academic concept and not a mere rhetorical device. If „social exclusion‟ means different
things to different people (Silver, 1994; Evans, 1998), then there is a challenge in producing
coherent academic debate and a danger that the concept is used for political purposes: for
example, to obscure or deflect attention away from poverty or to valorise (paid) labour. The
use of the discourse of social exclusion to deflect attention away from poverty has been noted
by many authors (Room, 1995a; Berghman, 1995).
Room (1995b: 233-4) has suggested that the shift from (income) poverty to social exclusion
entails a broadening of focus on three fronts. Firstly, it implies a shift from focussing on
income/expenditure to multi-dimensional disadvantage. Second, it implies a shift from static
outcomes to dynamic processes. Finally, there is a shift in focus from the household or the
individual to the local community in terms of a spatial dimension. While these may indeed be
features of a social exclusion approach, the extent to which they are, in fact, unique or real
differences from a poverty perspective can be questioned (Burchardt et al., 2002a) and the
indeed the distinction between the social exclusion and poverty approaches is only substantial
is one looks at the indirect measurement of (income) poverty (e.g. Room, 1995a).
In terms of what is agreed, there is widespread agreement that the concept is multidimensional, with a focus on exclusion in economic, political, cultural and social dimensions
(Berghman, 1995; de Haan, 1998). It should be noted that rather than the exclusion
perspective being inevitably broader or more comprehensive (Berghman, 1995) than the
poverty perspective, if poverty is seen as the direct measurement of deprivation, social
exclusion may, in fact, be narrower, if it is limited to exclusions that are „relational‟ (e.g. Sen,
2000). In terms of the causes of exclusion, there is a widespread agreement that the approach
includes a broader range of constraints than simply „lack of resources‟ as is typical in poverty
research (Evans, 1998; de Haan, 1998; Burchardt et al., 2002a).
However, I argue, however, that what remains contested is the very core of the concept –
what it is we want it to do. This, it is argued, is a more deep-seated problem. It is the very
purpose and perspective of the approach that is at issue. There remain two distinct, albeit
related, questions in my opinion which remain unanswered in the literature.
First, it is not clear whether social exclusion is an outcome or a process. While the idea that
social exclusion can be seen in dynamic terms is often professed as an advantage, there is a
distinction between adopting a dynamic perspective and an approach focussing on a process
and it is useful to highlight this distinction. Room (1995b: 237) notes that „[i]t is not enough
to count the numbers and describe the characteristics of the socially excluded; it is also
necessary to understand and monitor the process of social exclusion and to identify the
factors that can trigger entry or exit from situations of exclusion‟. Here exclusion is presented
as an outcome, and we are interested in a dynamic perspective in order to understand the 8
process of entry and exit to exclusion. It should be noted that thus is not entirely distinct from
research into the causes of poverty. However, elsewhere, he notes „[s]ocial exclusion is the
process of becoming detached from the organisations and communities of which the society
is composed and from the rights and obligations that they embody‟ (Room, 1995b: 243).
However, these two scenarios are distinct. In the first we identify a group (the excluded) and,
ex post, observe the risk factors that predict their exclusion. In the second, it is the risk factors
themselves that are of interest, and we do not know (or even perhaps care) whether these did,
in fact, result in the outcome in question for each individual. If exclusion is an outcome then
a focus on the processes that leads to it is in no way unique to social exclusion research –
such a focus can be, and indeed is, frequently employed in poverty research. However, if
social exclusion is not an outcome, but rather is a process leading to deprivation (as
Berghman, 1995 suggests), then we will need to be clear about this and distinguish risk
factors from outcomes. Are we interested only in those for whom the process of „social
exclusion‟ leads to „deprivation‟? Or are we interested in all those who are suffering the
process of social exclusion? These are distinct options and require greater consideration and
The second (related) question is why is social exclusion is important and, if social exclusion
is to be seen as a bad thing, who it is bad for. There seems to be broad agreement that social
exclusion is a bad thing; however, there is insufficient agreement, I argue, about who social
exclusion is bad for. Is it bad for the individual, preventing her to live as she would like, and
thus related to her well-being? Or it is bad for society or the social fabric? This is, in my
opinion, an important distinction. For example, Silver (1994: 534) notes that „[i]n terms of
Durkheimian rhetoric, exclusion threatens society as a whole with the loss of collective
values and the destruction of the social fabric‟. Similar questions are pertinent when Room
notes that exclusion may be seen in terms of outcomes such as inadequate social
participation, lack of social integration and lack of power (Room, 1995b: 243). Here the
question of voluntary social exclusion may be raised (see e.g. Barry, 2002); one which has
important implications for how we understand the concept.
For all the advantages of the concept of social exclusion in terms of the potential additional
terrain that it may cover, the multiplicity of understandings not just of the dimensions
considered, but also of its normative underpinning, must be viewed a weakness of the
concept, as currently understood.
Section 3: The capability set and individual capabilities
The essential distinction between the capability concept and the functioning concept, and
indeed the rationale for the former, is in recognising the normative importance of
opportunities and the requirement not to conflate choice with constraint. This is, in part, to
acknowledge the role that choice plays in mediating the relationship between opportunity and
outcome – two individuals who share the same real opportunities at any point in time may not
display the same outcomes. Indeed, in many ways this is not new: a central plank of 9
Piachaud‟s (1981) critique of Townsend was the admonishment that poverty should not be
conflated with preferences (more on this later).
However, while the concept of functioning has received widespread acceptance from the
point of view of academics working with the capability approach, the capability concept is
more contentious, with real questions about the extent to which is can be operationalised
(Krishnakumar, 2007; Comim, 2008). In terms of measurement, the capability concept poses
particular problems, because conceiving of a person‟s capability in terms of a set implies
identifying and valuing not only the bundle of functionings that a person chooses, but also the
other potential bundles that they did not. However, I wish to argue that Sen offers two
distinctive conceptions his capability concept: the first relates to individual capabilities, with
the latter relating to a person‟s aggregate capability, which is conceived of in terms of a set.
There are a number of advantages of considering the conceptualisation of aggregate
capability in terms of a set. First, it allows it not only to see the final functionings (outcomes)
that they choose, but also the full range and extent of opportunities from which they are
chosen. Sen has provided the related concept of „refined functionings‟ to describe outcomes
that take account of the available alternatives.
The second advantage of conceiving of capability as a set is that it allows us to understand
the trade-offs that occur achieving certain functionings. For example, this may allow us to
understand the extent to which functioning in one area (say, being in full-time employment)
cannot be achieved at the same time as others (being the primary carer for a child) and that
these choices are made explicit. However, this presents a rather counter-intuitive situation –
both options of paid employment and caring may be in a person‟s capability set, despite being
mutually exclusive options. Thus, while the conceiving of capability in terms of a set makes
such trade-offs explicit, we require more information than simply observing that an outcome
is included in a capability set before we can judge the extent to which it is truly feasible.
This shift from functioning to capability is often discussed as if it is one movement.
However, I argue that this is best seen as consisting of two. The first shifts from functionings
to capabilities; that is, from individual outcomes to individual opportunities. In order to
exclude situations where particular outcomes occur as a result of choice, we look not at
whether a person has achieved a particular outcome but rather at whether they had the real
freedom to do so. The second shift is from individual capabilities to a persons overall
capability, or from individual opportunities to a persons aggregate opportunity. Thus, it is
argued by the author that one can accept that capability represents the real freedom of a
person to do or be something that they have reason to value without necessarily developing
the concept as analogous to an opportunity set (as Sen does). It may be useful to consider
from a poverty perspective whether either or both of the aforementioned shifts are justified,
with some consideration given to the burden of measurement that they imply.
The shift from outcomes to opportunities is related to the Piachaud critique (1981) which
criticised Townsend for conflating preferences with poverty. The importance of 10
distinguishing constraints and choice is an issue that Sen has also raised on many occasions.
Given that the Piachaud critique has largely been accepted and indeed incorporated by
subsequent surveys, it would seem that the principle of such a shift from (individual)
functionings to (individual) capabilities is not entirely controversial. However, it must also be
recognised that a such a shift from functionings to capabilities is not only limited to
constraints involving a lack of resources, but all those that influence a person‟s real or
substantive freedom to achieve certain outcomes. This is an important distinction, to which
we will return.
In considering the shift from individual capabilities (outcomes) to a person‟s overall
capability, we must question the value of unchosen opportunities and trade-offs in making
our judgement. In terms of the former, it should first be recognised that conceiving of a
person‟s capability as a set is to look at their outcomes and opportunities in a positive sense,
by which I mean, to look at all of the things that they do achieve and all of those that they
could have achieved.
First of all, in understanding poverty, it might be argued that we are not interested in high
levels of functioning in particular dimensions or, indeed, in all dimensions (such as travelling
around the world, say). This is akin to the „focus axiom‟ in income poverty measurement in
which we focus ignore the living standards of the rich. Thus, in direct measurement, we may
be interested not all of things a person does do but, in a negative sense, in the needs that they
are unable to meet. Such an approach is in no way unique (see e.g. Doyal and Gough, 1991).
Thus, the extensive information that will be contained in the capability set may be considered
redundant for our purposes.
The second consideration is the extent to which understanding trade-offs amongst important
variables is deemed to occur. An example of this might be where a parent may choose to limit
social participation in order to provide greater opportunities for their children. Such trade-offs
are indeed of interest. However, it is by no means clear how they influence the valuation of a
person‟s aggregate capability, so this advantage comes with non-trivial measurement
Given this difficulty, and the lack of real advantage of in understanding the breadth of the
capability set in terms of poverty research, it may be argued that the second shift from
individual capabilities to a person‟s aggregate capability does not justify the burden of
measurement that it implies. If understanding trade-offs amongst necessities are deemed to be
an important advantage, it may be worthwhile considering them (either implicitly or
explicitly) elsewhere rather than including them in the concept of the capability set.
Towards measuring capabilities
One way that we may decide to gather information about a person‟s capabilities is via an
extension of the deprivation indicators that are typically collected in many social surveys.11
However, there remains considerable work to be done before such a process could be said to
In terms of the „object of interest‟ – that is, in the deprivations we are interested in – we may
consider a broader range of deprivations than simply „material deprivation‟ or „social
exclusion‟. This is not to suggest that anything can be included here – but a focus on
capability deprivation would allow consideration of wider dimensions because it would not
require anchoring these around notions of resources (in the case of many poverty indicators)
or relational features of social life (in the case of social exclusion).
One of the central debates within the capability literature is which capabilities we are
interested in – and in particularly whether any capability list is to be determined by the
individuals themselves (e.g. Alkire, 2002) or by others – for example, by an analyst or some
form of social participation. Sen‟s contention that we should focus on capabilities that we
have „reason to value‟ has only served to confuse in this area, and while this is a major issue
with the approach as a whole (and is addressed by Goerne‟s paper in this symposium), it is
argued here that this debate must take a somewhat different direction when we are focussing
specifically on poverty analysis.
This is because the variables used to distinguish the poor from the non-poor (or even, in a
broader sense, the disadvantaged from the non-disadvantage) cannot be determined by the
individual themselves (private jets and holidays in Mauritius for me; £10 to spend on a
Saturday night for you). There are certain well-established methods of identifying „socially
perceived necessities‟ that we will need to consider in determining these capabilities rather
than such decisions being left either to the individual on the one hand, or the analyst on the
other making this decision. This is an important issue, but is not addressed further here.
However, it should be noted that that a focus on wider capability deprivations would include
the possibility of consideration of indicators relating to social need that were not particularly
related to resources or relational deprivation. Thus, the breath of the concept of social need
may be preserved.
In terms of the constraints of interest, the focus on real freedom implies a wider range of
constraints and not focussing solely on „lack of resources‟. For some, the „lack of resources‟
condition is an important aspect of the poverty concept and wider constraints should not be
considered. The argument here is not to take issue with this (particularly given that this is
something of an accepted part of the poverty orthodoxy) but rather to argue for the
importance of other constraints in our moral analysis. As Sen notes, „If our paramount
interest is in the lives that people can lead – the freedom they have to lead minimally decent
lives – then it cannot but be a mistake to concentrate exclusively only on one or other of the
means to such freedom. We must look at impoverished lives, and not just at depleted wallets
(Sen, 2000: 3).12
I argue that what is required is a reconsideration of the relationship between poverty and
well-being, thus clarifying the terrain of analysis before subsequently dividing this terrain
amongst the various concepts.
Section 4: The poverty concept and the capability approach
By now it should be obvious that suggested method of measuring capabilities has a clear
similarity with the existing direct measurement of poverty and use of deprivation indicators.
However, it is also a broader notion both in terms of the „object of interest‟ and the
„conditions of interest‟. There will undoubtedly be questions about whether this is „really‟
poverty at all, and it is an answer to this question that I begin to sketch an answer to here.
In Poverty and the United Kingdom, Townsend‟s deprivation indictators, as I understand
them, were intended to examine the full range of conditions of living that were „customary, or
are at least widely encouraged or approved, in the societies in which they belong‟ (1979: 31).
Indeed, this breadth was stressed by Townsend himself – in discussing the use of deprivation
indicators to measure poverty, he noted that „[i]n principle, such a list might be developed, as
I have suggested, from an exhaustive analysis of the amenities available to, and the customs
or modes of living of, a majority of the population‟ (1979: 251). While such an exhaustive list
may always prove elusive, he notes that in terms of his own survey, „we sought only to
ensure that all the major areas of personal, household and social life were represented in out
questionnaire‟ (1979: 251). The breadth of the sixty indicators measured is indeed
impressive: six items of dietary deprivation, four of clothing, four relating to fuel and light,
nine relating to household facilities, four to household conditions and amenities, twelve to
conditions at work, five to health, one to education, five were environmental, four related to
the family, two to recreational activities and four to social activities.
We are not told a great deal about the process of reducing these sixty into a more manageable
index of twelve items other than it sought to focus on „those indicators which apply to the
whole population‟ (1979: 251) However, this process is of some interest, because the
resulting index is in no way representative of the original breadth: four items of dietary
deprivation, two each relating to the family, recreation and social deprivation (which, in fact,
are remarkably similar and, with the exception of one item, relate to social activities with
those outside the household), one item of household facilities (no refridgerator) and one of
housing conditions (no sole use of WC, sink or wash basin, bath or shower and cooker). Not a
single item from the categories of conditions at work, health, education, environmental
deprivation, fuel and light and clothing are included in the summary deprivation index.
My intention with the aforementioned discussion is not to criticise the selection (parsimony is
always required) but rather to demonstrate that poverty, and the original direct measurement
of poverty in particular, did begin from a broad notion of that concept and situate it within
wider considerations of well-being and need.13
Furthermore, in terms of the conditions in which we are interested in deprivation, Piachaud‟s
influential critique of Townsend (1981) led to a shift in the terms in which we understand
deprivation, following his observation that doing without some of the items of the deprivation
index might be „as much to do with tastes as with poverty‟ (1981: 420). Indeed, in attempting
to draw a distinction between deprivation by choice and by constraint, he was making a
similar distinction between the functioning and capability concepts provided by Sen.
However, the result of Piachaud‟s critique was not just to remove preferences by focussing
where this occurred due to constraint, but rather to focus on one constraint in particular: that
of a lack of resources. While this makes sense in terms of the importance of resources for the
concept of poverty, for activities at least – it created a false dichotomy between those who did
not engage in certain activities because they could not afford them and those who did not
engage in said activities because they did not want to. The other potential reasons for nonparticipation that we have noted (disability, discrimination, etc) were subsequently
overlooked. Nonetheless, the dichotomy was seen to fulfil its purpose in terms of the
definition of poverty adopted. In terms of social exclusion, where additional constraints
beyond a „lack of resources‟ might be considered, this was at times recognised. The authors
of the Poverty and Social Exclusion in Britain report noting:
„It is important to note that factors other than price may also result in effective
exclusion from services and activities. Those with limiting long-standing illness or
disability were asked about difficulties in accessing services‟. (Gordon et al., 2000: 60)
Indeed, the results of Poverty and Social Exclusion survey itself demonstrated these
exclusions. Eighteen (18) percent of respondents reported „lack of time due to childcare
responsibilities‟ as preventing them from common social activities. Similarly, fourteen (14)
percent were prevented by being too old, ill sick or disabled, six (6) percent by having no-one
to go out with and five (5) percent due to having no vehicle or because of poor public
transport (Gordon et al., 2000: 62).
However, the importance of „resources‟ to the poverty concept influenced not just the
„condition of interest‟ (lack of resources) but also the „object of interest‟ (the indicators
themselves). The authors of the Breadline Britain survey argued that not only was poverty to
be understood as a condition of an „enforced lack of socially perceived necessities‟ (Mack
and Lansley, 1985: 44), but this consideration would influence not just the constraint being
considered but also the object of analysis. Given the addition of this condition to the poverty
concept, they noted that:
„The critical role of lack of resources to the concept of poverty also has wider implications,
because it determines which aspects of our way of life should be included in a minimum
standard of living aimed at measuring poverty. We decided that only those aspects of life
facilitated by access to money should be tested in the Breadline Britain survey‟. (Mack and
Lansley, 1985: 44)14
Thus, the decision to apply the „lack of resources‟ constraint influenced not just the constraint
side of the equation, but also the objects of interest in terms of the deprivation indicators
themselves. My purpose here is not to question the way in which the poverty concept has
been developed and refined, but rather to question what it means for our broader analysis.
I would like to propose that if activities are considered to be necessities, a lack of resources is
not the sole constraint that we may be interested in. This is not to question the central role
than lack of resources plays in our conception of poverty, but rather that other constraints
deserve recognition and that our field of study should not be entirely defined (and ultimately
truncated) by the poverty concept. While the Piachaud critique sought to remove preferences
from the equation, it is in my view simply incongruous to suggest that something is a serious
moral concern if one is doing without it because she is constrained by her lack of resources
but of little or no concern at all if as a result of other constraints. Thus, while a lack of
resources may well be the most important impediment to participating in social activities, it is
by no means the only one, and it is argued that these additional constraints deserve (more!)
Furthermore, I would suggest that we should not limit our understanding of „socially
perceived necessities‟ to deprivations that are particularly amenable to changes in income.
These may be of particular interest to the poverty concept itself, but may also have the effect
of limiting our understanding of need and wider deprivations.
The additional conceptual space
Clearly, then, a capability approach to poverty analysis requires some additional conceptual
space not provided for by the poverty concept, traditionally defined. It might be asked
whether the capability approach measures something altogether broader in nature then
poverty, per se, and we may thus question whether it has anything to offer poverty analysis
(Lister, 2004). I do not disagree that the capability approach does indeed take a broader focus
than much existing poverty analysis. However, I do not believe that this requires us to jettison
our existing concepts in favour of some radically new concept of poverty.
What I believe a capability approach to poverty analysis would entail is identifying a terrain
of analysis that is broader than the boundaries of the poverty concept. While this may seem a
somewhat odd distinction, it is one that should be familiar to us given the addition of the
allied concept of social exclusion to that of poverty in many recent studies. On the
assumption that there is, in fact, a distinction between these concepts, this implies the notion
that the boundary of one of the concept is not equivalent to what I describe as the „terrain of
analysis‟. Thus, there is nothing terribly unique about the idea of suggestion consideration of
additional terrain, per se.15
In some ways, this distinction relates to the origin and purpose of the respective approaches.
The initial exposition of the capability approach was in response to the „equality of what?‟
debate, where the central question was which informational space we should be focussing on
in our moral analysis and our considerations of equality. Such a focus on an „informational
space‟ perhaps renders it more suitable to questions of „terrain‟ (what should we be looking
at?) rather than understanding one concept (what is poverty?). Sen‟s conceptual writings are
used to develop a conception of well-being (and its inverse) – or at least a conception of wellbeing that would be suitable for social policy analysis – and the moral „terrain‟ or „space‟
under consideration is deduced from this. There may be a subsequent step which includes
diving the conceptual territory of analysis up between the relevant concepts – as Sen himself
suggests when discussed the potential of viewing the social exclusion approach as a subset of
the capability approach.
Conversely, the approach that poverty analysis has taken has often been rather different. This
has been something of a search for the meaning of that concept (the question of „what is
poverty?‟) and then attempt to operationalise and quantify whatever definition we agree on.
In this approach, an initial decision is made that poverty is an important social problem
worthy of analysis in order to understand its causes and to make recommendations for
policies to eliminate same. The approach is largely an investigation of what poverty is (the
implication being that we know that we are interested in it, regardless of what it is) and the
concept is redefined as knowledge in the field accumulates. My contention is not that such
developments are illegitimate (far from it), but rather that our analysis need not necessarily
follow the boundaries of one concept in any uniform way. Thus, the two approaches that I
have sketched here are in no way conflicting and can, indeed, be reconciled. The question of
how this might occur is beyond the scope of this paper, but is, I believe, and important area
for further work.
It is argued that there are possibilities for a capability approach to poverty analysis within the
field of social policy. These are in identifying a „terrain of analysis‟ and in linking the
concept of poverty to wider notions of social need and well-being. A capability approach to
poverty analysis need not be seen as an attempt to tear up existing concepts or to deny the
importance of poverty, but rather as an attempt to locate existing concepts within a broader
framework of moral analysis within social policy. 16
Alkire, S. (2002). Valuing freedoms: Sen's capability approach and poverty reduction.
Oxford, Oxford University Press.
Aristotle (1953). The Nicomachean Ethics. London, Penguin
Barry, B. (2002). Social exclusion, social isolation and the distribution of income.
Understanding social exclusion. J. Hills, J. Le Grand and D. Piachaud. Oxford, Oxford
Basu, K. (1987). "Achievements, capabilities and the concept of well-being: a review of
Commodities and Capabilities by Amartya Sen." Social Choice and Welfare 4(1): pp. 69-76.
Berghman, J. (1995). Social exclusion in Europe: Policy context and analytical framework.
Beyond the threshold: The measurement and analysis of social exclusion. G. Room. Bristol,
Borooah, V. (2007). "Measuring economic inequality: deprivation, economising and
possessing." Social Policy and Society 6(1): pp. 99-109.
Burchardt, T., J. Le Grand, et al. (2002a). Introduction. Understanding social exclusion. J.
Hills, J. Le Grand and D. Piachaud. Oxford, Oxford University Press.
Callan, T., B. Nolan, et al. (1993). "Resources, deprivation and the measurement of poverty."
Journal of Social Policy 22(2): pp. 141-172.
Comim, F. (2008). Measuring capabilities. The capability approach: concepts, measures and
applications. F. Comim, M. Qizilbash and S. Alkire. Cambridge, Cambridge University
Doyal, L. and I. Gough (1991). A theory of human need. Basingstoke, Macmillan Education
Gasper, D. (2007). "What is the capability approach? Its core, rationale, partners and
dangers." The Journal of Socio-Economics 36: pp. 335-359.
Gordon, D., L. Adelman, et al. (2000). Poverty and social exclusion in Britain. York, Joseph
Krishnakumar, J. (2007). "Going beyond functionings to capabilities: An econometric model
to explain and estimate capabilities." Journal of Human Development 8(1): pp. 39-63.
Lelli, S. (2008). Operationalising Sen's capability approach: the influence of the selected
technique. The capability approach: concepts, measures and applications. F. Comim, M.
Qizilbash and S. Alkire. Cambridge, Cambridge University Press.
Mack, J. and S. Lansley (1985). Poor Britain. London, George Allen & Unwin.17
Nussbaum, M. C. (2000a). Women and human development: The capabilities approach.
Cambridge, Cambridge University Press.
Nussbaum, M. C. (2000b). "Women's capabilities and social justice." Journal of Human
Development 1(2): pp. 219 - 247.
Piachaud, D. (1981). "Peter Townsend and the Holy Grail." New Society 10 September
1981: pp. 418-420.
Ravillion, M. (1996). "Issues in measuring and modelling poverty." The economic journal
Reyles, D. Z. (2007). The ability to go without shame: A proposal for internationally
comparable indicators of shame and humiliation. OPHI Working Paper Series # 3. Oxford,
Oxford Poverty and Human Development Initiative: 46.
Room, G. (1995a). Poverty and social exclusion: The new European agenda for policy and
research. Beyond the threshold: The measurement and analysis of social exclusion. G. Room.
Bristol, Policy Press.
Room, G. (1995b). Conclusions. Beyond the threshold: The measurement and analysis of
social exclusion. G. Room. Bristol, Policy Press.
Saunders, P. and L. Adelman (2006). "Income poverty, deprivation and exclusion: a
comparative study of Australia and Britain." Journal of Social Policy 35(4): pp. 559-584.
Sen, A. (1976). "Poverty: an ordinal approach to measurement." Econometrica 44(2): 219 -
Sen, A. (1983). "Poor, relatively speaking." Oxford Economic Papers 35(2): pp. 153 - 169.
Sen, A. (1987). Commodities and capabilities. New Delhi, Oxford University Press.
Sen, A. (1992). Inequality reexamined. Cambridge, Massachusetts, Harvard University Press.
Sen, A. (1993). Capability and well-being. The quality of life. M. C. Nussbaum and A. Sen.
Oxford, Oxford University Press.
Sen, A. (1999). Development as freedom. Oxford, Oxford University Press.
Sen, A. (2000). Social exclusion: concept, application and scrutiny. Social Development
Papers. Manilla Asian Development Bank: 54.
Townsend, P. (1979). Poverty in the United Kingdom: A survey of household resources and
standards of living. Middlesex, Penguin.
Whelan, C. T. and B. Maitre (2007). "Levels and patterns of material deprivation in Ireland:
After the 'Celtic Tiger'." European Sociological Review 23(2): pp. 139-154.18
Zheng, B. (1997). "Aggregate poverty measures." Journal of Economic Surveys 11(2): pp.