handbook of risk theory volume 176 || the capability approach in risk analysis
TRANSCRIPT
39 The Capability Approach inRisk Analysis
S. Roeser, R
DOI 10.100
Colleen Murphy . Paolo GardoniTexas A&M University, College Station, TX, USA
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 980
History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 981
Defining Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 981
The Capability Approach and Gaging Capabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 983
Current Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 984
Use of the Capability Approach in Risk Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 984
Risk Determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 984
Risk Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 989
Risk Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 991
Use of Risk Theory in the Capability Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 993
Further Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 994
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 995
. Hillerbrand, P. Sandin, M. Peterson (eds.),Handbook of Risk Theory,
7/978-94-007-1433-5_39, # Springer Science+Business Media B.V. 2012
980 39 The Capability Approach in Risk Analysis
Abstract: The standard meaning of risk adopted by risk analysts from a broad range of fields
is that risk is the probability that a certain set of consequences will occur given a hazardous
scenario. Risk analysis is the process of determining the probability of occurrence and
consequences as well as evaluating the determined risks. Capability refers to the genuine
opportunity that an individual has to do and become things of value, such as being
educated and maintaining bodily integrity. Such doings and beings are called functionings.
A capability approach provides a distinct evaluative space for conceptualizing and judging
states of affairs on the basis of how the capabilities of individuals are affected. This chapter
examines the reciprocal contributions of a capability approach and risk theory. A capability
approach has been used to enrich risk theory in three ways. First, it has been argued that
the consequences component of risk should be conceptualized and assessed in terms of the
impact of a hazardous scenario on capabilities, instead of either resources or utility. Second,
instead of evaluating risks on the basis of either public or expert judgment, risks should be
judged acceptable or tolerable on the basis of whether certain threshold levels of capabilities
are maintained. Third, a capability approach provides a distinct alternative to a prominent
way of managing risks, cost-benefit analysis. Instead of comparing alternative policies on the
basis of their relative advantages and disadvantages, policies should be evaluated on the basis
of whether they address unacceptable or intolerable risks and on the basis of their likely
affectability. Likely affectability considers the dollar per unit change in the expected impact of
a hazardous scenario on capabilities. Risk theory has enriched the capability approach in two
ways. First, considering risk has highlighted an important dimension of capability, security,
not previously recognized. Second, attempts to operationalize a capability approach to risk
have led to the development of a novel way to assess capabilities, and not simply actual
functioning achievements.
Introduction
In this chapter we discuss a capability approach in risk analysis and discuss how risk analysis
has enriched theorizing about capabilities. The standard meaning of risk adopted by risk
analysts from a broad range of fields is that risk is the probability that a certain set of
consequences will occur given a hazardous scenario (Kaplan and Garrick 1981; Hansson
2007b). This definition is more specific than the broader definition of risk found in the
philosophical literature according to which risk is ‘‘a possible scenario in which adverse events
take place’’ (Hansson 2007b). Risk analysis is the process of determining the probability of
occurrence and consequences as well as evaluating the specified risks. There are three different
components of risk analysis: risk determination, risk evaluation, and risk management.
Capability refers to the genuine opportunity that an individual has to do and become things
of value, such as being educated and maintaining bodily integrity. The capability approach
provides a distinct evaluative space for conceptualizing and judging states of affairs on the basis
of how the capabilities of individuals are affected.
There are four sections in this chapter. The first defines capability and a capability approach
and discusses how capabilities are assessed. The second provides an overview of how the
capability approach has been applied to risk analysis as well as how ideas from risk analysis
have been used to deepen our understanding of capabilities. The third outlines two main areas
for further research in a capability approach to risk.
The Capability Approach in Risk Analysis 39 981
History
Defining Capability
There are many valuable states and activities, such as being adequately nourished, being
educated, and participating in the political life of a community, which a given individual
may or may not be in a position to achieve. In the capability approach pioneered by Amartya
Sen (1989, 1992, 1993, 1999a, b) and Martha Nussbaum (2000a, b, 2001) such states and
activities are called functionings. An individual enjoys a particular capability if he or she has
a genuine opportunity to achieve a specific valuable functioning. In practice, capabilities are
interdependent. An individual might have a genuine opportunity to have a rewarding career
and a large family, but not have an opportunity to achieve both at the same time.
An individual’s general capability captures this interconnectedness. An individual’s general
capability is a function of the various combinations or vectors of functionings that he or she
has a genuine opportunity to achieve (Sen 1992).
Two general factors influence the capability of an individual: what he or she has and what
he or she can do with what he or she has (Wolff and de-Shalit 2007). The resources of an
individual provide information about what he or she has. Internal resources refer to the assets
an individual possesses, including skills, talents, and psychological well-being. External
resources encompass the other means available to him or her, including income, wealth, and
the support of family. For example, whether an individual has a genuine opportunity to be
educated depends on her mental resources and talents. However, it may also depend on
income, insofar as there is tuition or costs for lodging, meals, and books associated with
schooling; morever, schooling often reduces an individual’s ability to earn income in the
short term. Family support for education is also critical; families must be willing, for example,
to allow a child to attend school instead of working inside or outside of the home. The
resources of an individual are not the only factor that influences what genuine opportunities
are available to him or her. There are variations in what individuals can achieve with a given set
of resources (Sen 1993, 1999b). Equally salient for determining the genuine opportunities of an
individual is the social and material structure within which he or she acts. Customs and
traditions, laws, the physical infrastructure of a community, and language are all examples of
some of the dimensions of this social and material structure. It is this structure that influences
what an individual can do or become with the resources at his or her disposal. An individual
may have a set of talents and skills. However, whether an individual can become educated given
his or her talents and skills depends on the condition of the physical infrastructure (e.g.,
whether there are roads or paths through which he or she can access a school), social norms,
and laws (e.g., whether there are formal or informal restrictions on him or her attending
school). Such factors impact whether and how resources contribute to achieved functionings
(Robeyns 2005, p. 99).
An individual’s capability is morally significant because of its connection with individual
well-being. According to Sen (2009), well-being can refer to either the ‘‘promotion of an
individual’s well-being’’ or the promotion of an individual’s agency goals. Agency goals are
a broader category of goals, encompassing everything an individual has reason to pursue.
Some of the goals of an individual may include his or her ownwell-being, and so an increase in
well-being will be an increase in agency achievement. However, some agency goals may not be
connected to an individual’s well-being; indeed the achievement of some agency goals can
982 39 The Capability Approach in Risk Analysis
be at the expense of individual well-being. Using the freedom/achievement distinction,
an individual’s well-being might relate to well-being achievement or well-being freedom.
Keeping these distinctions in mind, Sen claims that capability, broadly understood as the
dimension of freedom concerned with substantive opportunities, can include both well-being
freedom and agency freedom. He argues that more capability often enhances well-being, when
well-being is defined in terms of agency. However, because that same agency freedom can be in
tension with, or indeed go against, individual well-being freedom depending on the goals
a particular agent has, capability may not always enhance well-being understood as well-being
freedom. Capabilities provide information about the opportunity individuals have (or lack)
to exercise their agency in shaping their lives.
Conceptualizing and assessing individual well-being in terms of capabilities differs impor-
tantly from two common metrics for assessing well-being: resources and utility. Consider
resources, or evaluating well-being on the basis of the market commodities (e.g., income,
resources) or primary goods (e.g., wealth, income, social bases of self-respect) an individual
enjoys. Resources are one necessary condition for the well-being of individuals. As we saw
earlier, resources influence the capabilities of individuals. An increase in resources can also be
the consequence of certain functioning achievements. However, one problem with the
resource framework is that it mistakes means for ends. Resources are taken to be intrinsically
good, instead of instruments for the achievement of what is intrinsically good, namely,
capabilities. Typically, we care about income not for its own sake but because of what it
allows us to do and become. A second problem with the resource-based framework is that
it ignores interpersonal variability, or the different rates at which diverse individuals can
use resources to achieve a particular functioning. As David Crocker (2008, p. 114) writes,
‘‘Due to variations among individuals, the same commodity either may help some and harm
others or may promote the well-being of some a lot and of others only a little. Although food
intake normally will enhance human functioning, it will kill the person choking on a fish bone.
To function well, Milo the wrestler needs on the one hand, more food than the infant and
the disabled and, on the other hand, less food than a wrestler of a similar size but stricken
with parasites.’’
Another method for assessing individual well-being is in terms of utility. Welfare econom-
ics is the discipline devoted to the assessment of the goodness of states of affairs and policies in
terms of their impact on well-being. Well-being/advantage is usually defined in terms of utility.
Utility is defined as happiness, where happiness is often understood as desire fulfillment (Sen
2009). The utility approach avoids one weakness with the resource framework: it does not
fetishize goods but makes individuals the fundamental unit of concern (Crocker 2008).
However it has important limitations. First, happiness and desire fulfillment are necessary
but not sufficient for well-being. Individuals who are persistently deprived may adapt to their
circumstances to make life tolerable, learning to ‘‘take pleasure in small mercies’’ and refusing
to desire or hope for change in their circumstances. If we assess the well-being/advantage of
such individuals on the basis of their happiness alone, then we would fail to get an accurate
picture of their actual disadvantage. Conversely, ‘‘people may have well-being and even
opulence (be ‘well off ’) and yet be unhappy and frustrated; their unfulfilled desires may be
for rare Rioja wine and top-of-the-line Mercedes’’ (Crocker 2008). As these examples illustrate,
the informational basis of well-being/advantage in welfare economics is incomplete. It should
be broadened to include factors such as substantive opportunities, negative freedoms, and
The Capability Approach in Risk Analysis 39 983
human rights. Omitting this information prevents us from making important distinctions in
our judgments of the relative advantage of individuals who enjoy the same level of happiness,
but differ dramatically along these other dimensions. Omitting this information also leads to
distorted assessments.
The Capability Approach and Gaging Capabilities
The capability approach offers an evaluative space in terms of which judgments about states
of affairs should be made. For example, in a capability approach to development, ‘‘[T]he
purpose of development is to improve human lives by expanding the range of things that
a person can be and do, such as to be healthy and well nourished, to be knowledgeable, and to
participate in community life’’ (Fukuda-Parr and Kumar 2003). Societies are assessed as more
or less developed from this perspective, then, not on the basis of their wealth but rather on the
basis of the degree to which they ensure the ability of individuals to be healthy, knowledgeable,
and a participant in their community. The capability approach is used in a wide variety of
fields including social choice theory, welfare economics, development ethics and economics,
moral and political theory, disaster studies, and education (Robeyns 2006; Gardoni and
Murphy 2010).
To be able to use the approach to evaluate different states of affairs, it must be possible to
assess the capabilities of individuals. The earliest attempt to operationalize the capability
approach was in development economics. The United Nations and development agencies
now assess development on the basis of the functionings achievement of members of
a country. The Human Development Report (HDR), published annually since 1990, uses the
Human Development Index (HDI) to assess development. The HDI continues to provide
a prominent model for assessing functionings achievements in other fields. In the HDI, the
functionings used for assessing development include the ability to live a long and healthy life,
the opportunity to be knowledgeable, and the ability to have a decent standard of living.
Capabilities are not directly observable or quantifiable. In practice, indicators are used to try to
indirectly gage either actual functionings achievements and/or capability, the freedom to
achieve functionings (Raworth and Stewart 2003; United Nations Development Program
2007). Indicators are used to quantify each functioning. To illustrate, one indicator for the
ability to live a long and healthy life is life expectancy at birth. For the opportunity to be
knowledgeable the adult literacy rate is used. Gross Domestic Product (GDP) per capita is an
indicator selected for the ability to have a decent standard of living. The data collected for each
indicator is compared to a scale of minimum and maximum values (goalposts). In the case of
life expectancy, the average life expectancy rate for individuals in a given country is compared
to the minimum value of 25 years and the maximum value of 85 years (United Nations
Development Program 2000). This comparison, also called normalization or scaling, gives
some context and meaning for the specific life expectancy rate of a given country, providing
a picture of what the level of functioning achievement for that indicator is (Jahan 2003). The
goalposts represent reasonable minimum and maximum values based on an analysis of
historical data (Raworth and Stewart 2003). Finally the normalized or scaled values of the
individual indicators are combined in an unweighted average. No weights are typically used
because each functioning is taken to be equally important.
984 39 The Capability Approach in Risk Analysis
Current Research
In this section we first consider how the capability approach has been applied to risk analysis.
We then discuss two important ways in which the concept of capability has been enriched by
risk studies.
Use of the Capability Approach in Risk Theory
The study of risk is highly interdisciplinary. A diverse range of risks are the subject of
examination, including those associated to environmental, technological, man-made, and
natural hazards. This section describes the contributions that the Capability Approach has
made to risk analysis. We consider in particular the contributions to the three components of
risk analysis: risk determination, risk evaluation, and risk management.
Risk Determination
Risk determination refers to the quantification of the the probability of occurrence of a
particular hazard and its associated consequences. Our focus in this section is not on the
probability component of risk, for which there are a wide variety of methods of quantification
available. For a description of such approaches see Pate-Cornell (1996), Haimes (2004), and
Cullen and Small (2004). Rather, our interest is in how the consequence components of risk
are often characterized and assessed. The first contribution of the capability approach to
risk theory is in our understanding of how the consequences from hazardous scenarios
should be conceptualized and assessed. Below we first describe two common approaches to
conceptualizing and quantifying the consequence components of risk and their limitations.
We then discuss how the capability approach has been proposed as an alternative, and how
this approach avoids the limitations with existent approaches.
Limitations with Common Frameworks to Risk Determination
Risk analysts and those who provide the inputs for a risk analysis, including engineers and
social scientists, have historically adopted a broad resource-based approach to the consequence
component of risk (Rowe 1980; Vose 2000; Bedford and Cooke 2001; Haimes 2004). In one
framework, the numbers of resources of various kinds that are lost, including structures,
money, time (in terms of delays in construction), and/or individuals are counted. Counting
the consequences of risks in this manner has a number of limitations. First, a concentration on
resources lost does not provide information about how the lives of individuals are affected. As
we noted in the discussion in the previous section, there are different rates at which a given
resource can generate an opportunity to achieve a valuable functioning. Thus, identifying the
number of structures damaged does not alone tell us what that loss means for the individuals in
the community affected. Losses of a given structure can have a greater or lesser impact
depending onwhether the structure is a hospital, an empty building, or an apartment complex,
for example. Second, defining consequences in terms of resources implicitly treats what are
important means to achieving well-being as ends in themselves. Third, risk analysts, engineers,
and social scientists increasingly recognize the need to consider consequences beyond those
traditionally taken into account in risk analysis. However, the resource framework does not
provide a principled way to demarcate what losses should count or are salient for purposes of
The Capability Approach in Risk Analysis 39 985
risk analysis. Indeed, consequences are often selected in practice on the basis of their ease
of quantification (Murphy and Gardoni 2006; Gardoni and Murphy 2009).
A second alternative framework evaluates consequences on the basis of the utility lost in
a given hazardous scenario. Utility is commonly determined by examining the preferences or
choices of individuals. An underlying assumption is that the satisfaction of preferences
increases welfare. Preferences or choices are assessed inmonetary terms. Thus, all consequences
of interest (e.g., the loss of a human life, harms to the environment) are converted into
a monetary figure. This figure is determined on the basis of individual ‘‘willingness to pay’’
(Sunstein 2005). That is, the cost associated with a risk is based on an examination of what
individuals would need to be paid to be willing to be exposed to a certain risks. These figures
are characteristically based on actual market activity. As Cass Sunstein (2005) describes it,
‘‘Suppose that people must be paid $600, on average, to eliminate risks of 1/10,000; suppose,
for example, that workers who face risks of that magnitude generally receive $600 in additional
wages each year. If so, the VSL [Value of a Statistical Life] would be said to be $6 million.’’
Contingent valuation surveys are used in cases where there is no market evidence on which to
make an assessment. For example, to assess the loss of a coral reef, individuals may be asked
what they are willing to pay to salvage coral reefs (Sunstein 2005; Hansson 2007a).
A number of objections have been raised to conceptualizing consequences in terms of
utility. Most fundamentally, assigning costs to certain consequences on the basis of market
choices and preferences does not accurately convey the costs of certain consequences,
either over- or under-estimating their significance. This is so for a number of reasons. First,
individuals may have misperceptions about or be indifferent toward risks they face. One
common case of misperception of risks is that rare, dramatic risks might be feared more
than ordinary risks. In the workplace, individuals may not be fully knowledgeable about
the occupational risks they face when accepting a position; when they do learn of certain
risks the costs to leave may be grave, including loss of seniority or pension if they seek
employment elsewhere (Anderson 1988). Environmental costs may not be a source of grave
concern, in part based on a failure to understand the significance of the environment to us.
As a result, the costs assigned may lead to a corresponding over- or under-valuation of
certain kinds of losses (Slovic 1987; May 2001; Murphy and Gardoni 2006).
Second, even when individuals have an accurate perception of the risks they face and are
not indifferent to their consequences, the willingness to pay amount may not capture the way
individuals value certain costs, such as the loss of their lives. Risks that are job-related may not
be voluntarily accepted, but rather chosen because they were the only form of employment
available (Wolff and de-Shalit 2007). As Elizabeth Anderson (1988) writes, ‘‘Workers with
families commonly do not regard their choices about risks to their lives as acts of consumers
out to maximize their personal utilities. Rather, they see their choices as attempts to discharge
their responsibilities to their families. Bound by their responsibilities to others, they do not feel
free to risk their lives at will for pay, as they sometimes acknowledge having done in their youth.
But neither do they feel free to jeopardize their families’ means of livelihood and future
prospects by simply quitting their hazardous jobs and risking long-term unemployment.’’
Thus, choices to accept certain wages may not reflect an individual’s judgment about
the value of his/her life, but rather the judgment about how it is necessary to fulfill certain
responsibilities.
An additional limitation is that the utilitarian framework does not focus attention on what
should be of fundamental interest, namely, how the genuine opportunities of individuals will
986 39 The Capability Approach in Risk Analysis
change if certain hazards are realized. Instead of valuing and prioritizing the promotion of the
agency of individuals, the utilitarian framework treats the losses of human lives as a market
commodity. Furthermore, information about the monetary costs of consequences does not
translate into information about how the freedom of individuals to do and become things of
value will be restricted. That is, what are the implications of certain forms of environmental
damage for the lives of individuals? Concentrating onwhat individuals may be willing to pay to
avoid certain losses does not provide information about the ways in which the genuine
opportunities open to individuals may change should certain hazards be realized.
A Capability Approach to Risk Determination
In a capability approach, the consequences of hazardous scenarios are conceptualized and
assessed in terms of changes in capabilities. Risk is then defined as the probability that capabilities
will be reduced (Murphy and Gardoni 2006). Conceptualizing the consequences component of
risk from a capability approach overcomes the limitations with both the resource and utilitar-
ian framework. It focuses the attention of risk analysts directly on how a hazardous scenario
will affect the welfare of individuals within a community, insofar as consequences are defined
in terms of the change in the genuine opportunities of individuals to achieve valuable states
and activities. It thus avoids the mistake of focusing on commodities, which are instrumental
means but not valuable ends in themselves. Furthermore, the capability approach can account
for differences in interpersonal conversion rates in using resources to achieve valuable func-
tionings; the capability approach focuses directly on the freedom to achieve valuable function-
ings itself. The capability approach also does not define the consequences of hazards in terms of
what individuals would be willing to pay. It thus avoids the commodification of costs like the
loss of a human life and problems with the lack of information about, concern for, or ability to
avoid certain risks plaguing the utilitarian analysis. What individuals would be willing to
pay to avoid certain consequences plays no role in determining what the consequences of
a hazardous scenario might be. Instead, the impact of a hazard on the genuine opportunities
of individuals is the focal point of concern. The capability approach can account for the
broader indirect impact that a hazard might have, because built into the framework is
the recognition of the fact that both resources and the social and material structure of
a community profoundly affect the opportunities open to individuals. Concern with the
freedom of individuals to achieve valuable functionings invites questions about whether and
in what ways the resources of individuals or the social and material structure of a community
have changed. Finally, concentrating on capabilities, rather than achieved functionings, reflects
a respect for the commitment of liberal governments to ensure that a range of options of ways
of living are open to individuals, instead of promoting a particular way of living itself.
To assess the impact of a hazard, it is necessary to operationalize the capability approach to
risk. Initial efforts to operationalize the capability approach built on the framework are provided
by the HDI for development, which assesses achieved functionings (Gardoni and Murphy
2009). For the purposes of quantifying the consequences of risk, a Hazard Impact Index HII
was developed (Gardoni andMurphy 2009). TheHII is constructed in four steps. The first step
involves selecting the capabilities relevant for the kind of risk in question; that is, it is necessary
to identify which capabilities will provide an accurate picture of the impact of a hazard. Gardoni
and Murphy (2009) have argued for a set of criteria that any argument for the capabilities to be
used in risk analysis must satisfy. Briefly, a given capability is relevant for risk analysis insofar as it
reflects the underlying values and concerns with risk. To ensure that a capability risk analysis is
The Capability Approach in Risk Analysis 39 987
both comprehensive and, at the same time, practically possible to perform, theminimumnumber
of capabilities should be selected and the capabilities chosen should provide information not
available from given capabilities (e.g., because there is a correlation between the enjoyment of two
capabilities). Gardoni andMurphy also proposed a preliminary list of capabilities to be used in risk
analysis conducted with respect to natural hazards.
Indicators must then be chosen for the selected capabilities. In the context of risk,
appropriate indicators must be shown to track the impact on the capability of interest of
a hazardous scenario (Gardoni and Murphy 2010). Such indicators will allow risk analysts to
assess the impact of a hazardous scenario on capabilities, and so assess the consequence
component of risk. Following the model provided by the HDI indicators were taken to track
the change in the level of achievement of each associated functioning. Each indicator must be
converted into an index that ranges from 0 (minimum achievement) and 1 (maximum
achievement) through a process of scaling analogous to that used in the HDI. Finally, all
indices are combined through an averaging process to create an aggregate measure of achieve-
ment. It is also noted that in the context of risk analysis the values of the indicators cannot be
simply measured but need to be predicted since the subject of risk analysis is future events. The
HII gives a snapshot of the predicted well-being after a hazard. A comparison between the HII
and the actual well-being before a hazard (also measured using a capability approach) provides
a measure of the impact of the hazard.> Figure 39.1 illustrates this formulation of the HII. F1, F2,. . ., Fk on the left column are
the functionings under consideration. Each functioning can be achieve at different levels 1
through q. For example, functioning Fkmight be achieved at levels Fk1, Fk2,. . ., Fkq. As shown in> Fig. 39.1, an individual P(i) might achieve F1 at level F12, F2 at level F2q, and so on, up to Fk at
level Fk1. Each level of achieved functioning for individual P(i) is then converted into
a corresponding index, for a total of k indices: I1(i), I2
(i),. . . Ik(i). The same process is repeated
for all the considered individuals P(1),. . ., P(n). An average of the indices over all individuals is
then computed (Avg.[Ik(i)], i = 1,. . ., n) for each functioning. Finally the HII is computed by
combining all the averages.
One limitation with this construction of the HII is that it assesses the level of achievement
of functionings, not capabilities. For assessing the least advantaged in society and for assessing
the impact of a natural hazard (e.g., flood, earthquake, or hurricane) in the emergency phase
Converting achieved functionings into indices
Fk Fk2 Fkq
F2 F21 F22
F1 F11
P(i)P(i) P(n)
HIIP(1)
I(1)1 I(i)1 I(n)
1
I(1)2 I(i)2 I(n)
2
I(1)k I(i)k I(n)
k Avg. [I(i)k]
Avg. [I(i)2]
Avg. [I(i)1] i = 1,...,n
i = 1,...,n
i = 1,...,n
F1q
Fk1
F12
F2q
. Fig. 39.1
Illustration of the current formulation of theHII andHDI (Published previously as Fig. 1, inMurphy
and Gardoni 2010, with kind permission from Taylor & Francis Ltd., http://www.informaworld.
com)
988 39 The Capability Approach in Risk Analysis
(the time that immediately follows a medium or large disaster) information about achieved
functionings may be sufficient for assessing capabilities. This is because we can typically assume
that the functionings frequently used in these contexts, which are basic functionings, will be
automatically chosen if individuals have the genuine opportunity to do so. Examples of basic
capabilities in Murphy and Gardoni’s (2010) sense include the capability to have adequate
shelter, avoid injuries, and be adequately nourished. Murphy and Gardoni’s use of the notion
‘basic capabilities’ departs from both Nussbaum and Sen. Nussbaum defines basic capabilities
as ‘‘the innate equipment of individuals that is the necessary basis for developing the more
advanced capabilities, and a ground of moral concern.’’ (2000b, p. 84). Sen defines basic
capabilities as capturing ‘‘the ability to satisfy certain crucially important functionings up to
certain minimally adequate levels.’’ (1993, p. 40). By contrast, we delimit basic capabilities in
terms of those functionings that, with rare exceptions, individuals will choose to achieve if they
have a genuine opportunity to do so. In the case of basic capabilities, the levels of achieved
functionings thus provide an accurate picture of the capabilities. For example, we can reasonably
assume that if some individuals in the emergency phase in the aftermath of a hazard are not
sheltered this is because they lack the capability to be so. However, when we gage the medium-
and long-term impacts of a hazard on capabilities, an assessment of functionings achievement is
typically an inaccurate proxy. This is because the medium- and long-term impacts of large
hazards, or the impact of smaller hazards, affect non-basic functionings (or basic capabilities in
amore subtlemanner) that we cannot safely assumewill be chosen given an opportunity to do so.
In addition, as discussed earlier, in practice only vectors of functionings are open to an
individual and choosing to achieve one functioning level might preclude the opportunity to
chose a certain level of a different functioning. That is, only certain combinations of function-
ings are open to an individual to achieve. However, the original formulation treated each
functioning in isolation and did not account for how different functionings might interact in
fact. (Murphy and Gardoni 2010). Furthermore, the original formulation assessed the average
impact on levels of functionings achievement across a population. However, an average impact
can mask important variations among a population. The same average may reflect a case
in which all individuals had roughly the same impact, or a case in which some individuals
were not impacted but a portion of the population was severely impacted. The variation in
impact may correlate with certain subgroups within a population, though it need not do so.
Finally, in evaluating the capability of an individual we need to consider two dimensions
captured by this approach: the quality of the possible vectors, and the extent of the freedom
that an individual has in choosing different vectors. In evaluating capability, we need to
capture both the quality of the options open to an individual as well as the quantity, or
range of vectors, he or she has a genuine opportunity to achieve. The quality of the possible
vectors provides information about the levels and kinds of functionings that can be achieved,
and, more generally, of the available opportunities. The range of vectors provides information
about the scope of freedom.
For this reason,Murphy andGardoni (2010) proposed amethod for determining vectors of
capabilities of an individual. The fundamental idea is that functioning achievements of other
people says something about the choices that are genuinely open to an individual in the same
group. Knowing that other individuals have chosen to realize a certain opportunity is an
indication that such opportunity actually exists and allows us to distinguish between
a situation where an individual chooses not to actually take advantage of an opportunity
versus a situation in which such an opportunity does not exist. This method also accounts for
Fkq
P(i)
C(i)HII (1) HII (i) HII (n)
Converting achieved functionings into indices
I(1)k I(i)k I(n)
k
I(1)2 I(i)2 I(n)
2
I(1)1 I(i)1 I(n)
1
P(1) P(i) P(n)
Avg. [HII (i)]
Basic functionings
Non-basic functionings
St. Dev. [HII (i)]
i = 1,...,n
i = 1,...,n
F11F12
V(n)V(i)V(1)
F2iF2q
Fk1
F1q
F21
Fk1
F1
F2
Fk
. Fig. 39.2
Illustration of the new formulation of the HII (Published previously as Fig. 2, in Murphy and
Gardoni 2010, with kind permission from Taylor & Francis Ltd., http://www.informaworld.com)
The Capability Approach in Risk Analysis 39 989
vectors, rather than isolated functionings, by computing the impact of a hazard at the
individual level and then determining the achieved functionings across individuals.> Figure 39.2 illustrates the formulation proposed by Murphy and Gardoni (2010). An
individual P(i) can only choose among a set of vectors of achieved functionings, V(1),. . ., V(n).
For example, P(i) can chose V(i) which means choosing F12, F2q,. . ., Fk1. After transforming the
achieved functionings into indices, we need to compute the HII(i) (or the HDI(i) if this
approached is applied to human development) for individual P(i) and only then determine
the statistics (average and standard deviation) of the HII(i) across all individuals, indicated in> Fig. 39.2 as Avg.[HII(i)] and St. Dev.[HII(i)], respectively. In this formulation, Avg.[HII(i)]
tells us about the quality of the possible options, while the St. Dev.[HII(i)] captures the extent of
freedom.
Risk Evaluation
As its name implies, risk evaluation is the process of assessing the information provided
through a risk determination. Of particular interest is whether a risk is acceptable or not.
Below we survey two common methods for judging the acceptability of risks and their
corresponding limitations. A third approach, cost-benefit analysis, is considered in the next
section on risk management. We then present the capability approach to acceptable risk.
Limitations with Common Frameworks to Risk Evaluation
One approach to determining the acceptability of a risk is to consider whether the public finds
a given risk acceptable (Hunter and Fewtrell 2001). Underlying this approach is a commitment
to democratic governance, where democracy is taken to imply a form of majority rule.
In a democracy, policies and decisions are justified insofar as they express the will of the
people. However, there are two important limitations with using public judgment as the basis
for determining the acceptability of risks. First, as noted earlier, the public often has limited,
and sometimes inaccurate, knowledge about risks; thus judgments will not be based on correct
assessments. Furthermore, the public may be indifferent about risks they face. All of these
factors undermine the confidence we can have in the assessments of the acceptability of a given
990 39 The Capability Approach in Risk Analysis
risk the public may make. Second, the public judgment standard for acceptable risk mistakenly
equates democracy with majority rule. However, it is widely acknowledged by liberal demo-
cratic theorists that individual rights are at the core of democracy and constrain what the
majority may want to do. Furthermore, democratic processes should not take preferences or
views of citizens as a given, but rather aim to help develop an informed citizenry. Thus,
democracy does not require that the initial, often uninformed, preferences of citizens be the
basis for evaluating risks a community faces.
An alternative approach is to leave judgments about the acceptability of risks to experts.
Professional risk analysts have the most complete information about the character of risks and
the knowledge to understand how to interpret the results from a risk analysis. Often risk
analysts use as the basis for their judgments their past experience, the standards of their
profession, or the preferences and desires of their client (Fischhoff et al. 1981). However, in
practice experts may be biased in their evaluation of the acceptability of risks, failing to take
into consideration aspects of risk that matter to the general public or how risks impact society
as a whole, as opposed to their client in particular. Finally, the criteria experts use to judge a risk
as acceptable are frequently implicit, making it difficult to scrutinize or evaluate the justifi-
ability of a particular decision.
A Capability Approach to Risk Evaluation
In addition to providing an alternative characterization of the consequences of a hazardous
scenario, the capability approach provides resources for a novel way of conceptualizing the
criteria in terms of which a risk should be judged acceptable. Motivating this approach is the
key claim that justice requires that individuals enjoy a certain threshold level of capability
(Nussbaum 2000b). Judgments about the acceptability of a risk, then, should take into account
whether or not individuals will continue to enjoy threshold levels of capabilities in the
aftermath of a hazard (Murphy and Gardoni 2008).
There are two thresholds that are relevant in this context. The first is the acceptable
threshold. This threshold specifies what minimum level of capabilities it is acceptable for
individuals to enjoy over any period of time; this is the level below which capabilities should
not fall ideally. In the aftermath of a hazard, especially the immediate aftermath, it may not be
practically possible for all individuals to enjoy the acceptable level of capabilities. A framework
for risk evaluation should recognize this possibility, while at the same time setting limits on
what is permissible even in the immediate aftermath of a hazard. This is where the second
threshold comes in. The tolerability threshold is lower than the acceptable threshold. It
specifies what level of capabilities is tolerable in the immediate aftermath of a hazard, provided
this lower level is temporary, reversible, and does not fall below a lower tolerable limit. The
tolerability threshold reflects the moral necessity of avoiding permanent damage to the well-
being of individuals (Murphy and Gardoni 2008).
To account for the role of public deliberation, Nussbaum (2000b) and Sen (2009) rightly
note, is an important component of the capability approach. In a capability approach to
acceptable risk (Murphy and Gardoni 2008), the precise specification of the acceptable and
tolerable threshold can be the product of internal democratic processes. The goal of such public
processes is to articulate standards that establish realistically ambitious levels of acceptable and
tolerable capabilities, given a society’s actual conditions. Because we are dealing with risks, in
which the precise consequences of a hazard are unknown, the specification of the thresholds of
acceptable risk must also take into account probability. That is, in addition to specifying the
The Capability Approach in Risk Analysis 39 991
level of capabilities judged acceptable, the acceptable threshold must also specify the proba-
bility that capabilities will not fall below the acceptable threshold whichmust beminimally met
for a risk to be judged acceptable. The same consideration is true of the tolerability threshold.
In risk evaluation, then, we assess the probability that the level of capabilities across
a population is below the thresholds of acceptable and tolerable risk. To take into consideration
questions of justice in the distribution of risks, a process of disaggregation may be used.
This involves determining the probability distribution (as a measure of the likelihood) of the
capabilities of subgroups within a population. The probability that the level of capabilities for
various subgroups is below the thresholds of acceptable and tolerable risk may then be assessed.
The capability approach to acceptable risk avoids the limitations with relying either only on
public judgment or on expert judgment. It finds a central role for the public to play in shaping
the criteria of acceptable risk in setting the thresholds of acceptability. At the same time,
judgments about the acceptability of particular risks can be done by experts, whose evaluation
will be shaped by the standards the public sets.
Risk Management
Communities and individuals want to knowwhat risks they face, and tomake judgments about
whether such risks are acceptable, in part so that they can effectively manage risks. The
information provided by risk determination and risk evaluation provides a foundation for
ensuring that policies about risk are formulated in a well-informedmanner. Such policies must
take into account the limited resources available to governments and agencies as well as the
other competing priorities that governments must address. Risk management strategies pro-
vide a framework for aiding policy makers, who are generally not experts in risk analysis, in
deciding which risks should be prioritized formitigation action andwhichmitigation strategies
should be taken. Below we first discuss one prominent risk management framework, cost-
benefit analysis and its limitations.We then consider a capability approach to riskmanagement.
Limitations with Common Frameworks to Risk Management
A common method of evaluating and managing risks is cost-benefit analysis (CBA). CBA is
a broadly utilitarian approach in which two or more options for public policy decision making
are compared on the basis of their respective advantages and disadvantages. The advantages
and disadvantages are defined in terms of the consequences. All consequences (which may
include economic costs, risks of death, environmental harm) are assigned a numerical, typically
monetary, value. The monetary value is assessed based on the preferences of individuals, where
such preferences are measured based on the amount of money an individual would be willing
to pay to avoid certain risks or the amount of money he or she would agree to be paid to be
exposed to certain risks. Market information or valuation surveys provide the basis for
monetary figures. After quantifying the relative advantages and disadvantages, advantages
are weighed against the disadvantages numerically. The option with the greatest net benefit
(taking into account benefits minus costs) is then selected (Sunstein 2005; Hansson 2007a).
There are two strengths of CBA, which explains in part why it is favored by policy makers.
First, it provides a straightforward decision-making strategy for determining which risks to
address and in what way to address them. Second, it takes as a priority the efficient allocation of
scarce resources.
992 39 The Capability Approach in Risk Analysis
However, CBA has also been the subject of a number of criticisms, which parallel those
raised to a utility approach to risk determination. To briefly review, first, as noted earlier, many
critics of cost-benefit analysis object to allowing the market to establish a distribution of risks
on the basis of what individuals are willing to pay or be paid. For critics, this does not recognize
and cannot guarantee the obligations that both employers and communities have to guard
against exposing employees or individuals to certain kinds of risks. Second, relying on market
mechanisms to quantify risks is problematic because it assumes that individuals have full
knowledge about the risks they face when they determine what they would be willing to be paid
to be exposed to certain risks and that they base their decision on their assessment of how they
value what will be put at risk, such as their lives. However, neither assumption is warranted in
practice. As noted above, individuals are often not fully knowledgeable about the risks they
face. Furthermore, choices about risk exposure are often a function of factors, such as limited
income or employment opportunities, that do not reflect how individuals value certain goods.
Third, CBA fails to sufficiently recognize that the protection of life is not simply a question to
which a certain cost should be assigned, but reflects a principle or ideal to which societies
should be committed. Relatedly, it is problematic to assess a monetary value to the loss of
human life or a monetary value to other noneconomic goods. Fourth, there are important
issues of fairness and equality in the distribution of risks which the CBA cannot take into
account, given its emphasis on the aggregate costs and benefits (Asveld and Roeser 2009, esp.
the chapters by Hansson, Cranor, and MacLean).
A Capability Approach to Risk Management
In a Capability Approach to risk management, the overarching aim is to protect and promote
individuals’ capabilities (Murphy and Gardoni 2007). Given this aim, the first step in evalu-
ating which risks to address is to determine whether a given risk is acceptable or tolerable.
Priority should be given to risks judged intolerable, followed by a consideration of risks judged
unacceptable.
When evaluating a range of policy options to address and mitigate unacceptable or
intolerable risks, policy makers should determine which policies are viable options. Viable
options should be reasonably anticipated to bring the predicted level of capabilities above
the tolerable or acceptable thresholds, respectively. This can be achieved by reducing the
probability of occurrence of a given hazard and/or its expected impact. Because resources
are limited and risk management is not the only objective to pursue, choices among
viable options need to bemade. In a Capability Approach, we select from among viable options
on the basis of the likely affectability of relative policies. That is, from among policies
considered, we can select a policy to pursue based on the expected dollar per unit change
in the impact of a hazard. Monetary quantification can be used as a measure of the cost of
a given policy; its effectiveness is then conceptualized in terms of the increase or decrease in
capability levels.
The capability approach to risk management avoids some of the main criticisms leveled
against CBA. In a capability approach the thresholds that set the criteria for acceptable and
tolerable risk reflect the judgment that justice requires that individuals not be exposed to
certain kinds of risks and that society has an obligation to ensure that individuals are protected
from such risks. Because the quantification of the impact of risks and of mitigation strategies is
defined and assessed in terms of capabilities, the Capability Approach does not involve
quantifying the value of a human life or commodifying diverse kinds of goods. At the same
The Capability Approach in Risk Analysis 39 993
time, assessing policies on the basis of their likely affectability ensures that consideration is
given to the fact that resources are scarce and that it is important to ensure that public resources
are used in an efficient manner.
Use of Risk Theory in the Capability Approach
The concept of risk has been used to enrich our understanding of capability. In their work on
assessing advantage and disadvantage from a capability approach, philosophers JonathanWolff
and Avner de-Shalit (2007) have argued that it is important to consider the security of any
functioning achievement when assessing the capability of an individual. An individual has
a genuine opportunity to achieve a given functioning only if that achievement is one that can
be sustained; it is not an achievement that is at undue risk. A functioning achievement is at
undue risk, in their view, insofar as its achievement is temporarily and involuntarily put in
jeopardy by the pursuit of other functionings achievement. In other words, a functioning
achievement is insecure when an individual is ‘‘forced to take risks that in one way or another
are bigger than others are being exposed to or take’’ (Wolff and de-Shalit 2007). To illustrate,
the achievement of being sheltered may be insecure and at risk insofar as an individual has an
uncertain income stream stemming from irregular employment that results in unstable ability
to pay rent. Alternately, an individual may be forced to constantly put her bodily integrity
at risk, insofar as the only available sources of employment are dangerous occupations
(Wolff and de-Shalit 2007).
The inclusion of the idea of secure functionings achievement is becoming widely accepted
and adopted in capability analyses. Capability is now understood to represent the genuine
opportunity an individual has to achieve a particular functioning in a secure manner. Thus,
concern with the functionings achievements being put at risk in an involuntary or dispropor-
tionate manner is a central issue. Wolff and de-Shalit suggest statistical information
can provide some insight into the security of functionings achievement. In their view, we
can assess the security of functionings achievement by looking at trends over time for specified
subgroups. In their words, (2007, pp. 116–117),
" The fact that I have a job today says nothing about whether it is a day’s casual work or a sinecure
for life. However, taking a wider view and looking at the individual’s social circumstances or
context immediately provides more information. Indeed, statistics will provide much of what we
need. Imagine that among certain groups – perhaps the young, recent immigrants, or the low
paid – there is a high degree of mobility in employment or housing, with those moving jobs or
homes also experiencing periods of unemployment or homelessness. This, then, gives a prima
facie reason to believe these functionings are not achieved securely by people within these
groups. . . In general, then, although individual functioning is not an indicator for the degree of
security, statistics often can be.
There is a second contribution that risk theory makes to the capability approach.
The method developed to assess capabilities in the context of risk analysis (Murphy and
Gardoni 2010) and illustrated in > Fig. 39.2 can also be applied to assess capabilities in its
original context of development economics and in other applications. As noted above, initial
efforts to operationalize the capability approach built on the model provided by the HDI for
development, which concentrated on assessing achieved functionings. Furthermore, the
994 39 The Capability Approach in Risk Analysis
original formulation treated each functioning achievement in isolation, while in practice an
individual P(i) can only choose among a set of vectors of achieved functionings.
Further Research
There are two main areas of further research for a capability approach to risk analysis that we
concentrate on in this section. The first centers on the issues to address to operationalize the
capability approach to risk. The second centers on the place of a concern with capabilities in an
overall perspective on risk evaluation and management.
The capability approach developed by Sen is deliberately underspecified. Sen (1992, 1993)
refrains from identifying the capabilities that should be used in any given evaluative exercise for
two reasons. Thefirst is that the capabilities that should be consideredwill always be to some extent
evaluative exercise specific; that is, the capabilities of interest when considering development are
not necessarily the capabilities of interest when assessing risk. The second reflects Sen’s commit-
ment to identifying valuable capabilities via democratic procedures, and in particular democratic
deliberation. These two commitments of Sen’s give rise to a series of theoretical questions for
a capability approach to risk. First, there is the issue of how evaluative exercise specific the selection
of capabilities should be. For example, should the capabilities selected vary according to the kind of
risk being assessed? The work done to date developing a capability approach to risk has been
conducted in the context of the risks posed by natural hazards. Are the capabilities relevant for
technological risks, for example, the same as those relevant for natural hazards? On what basis do
we answer this question? Second, if Sen is correct about the role of public deliberation, in the
context of risk there is a need to incorporate a place for public deliberation in the selection or
evaluation of capabilities and a method for such deliberation to become possible.
Making the capability approach operationalizable also requires further research into
a cluster of questions surrounding the quantification of consequences. Indicators must fulfill
at least two criteria (Gardoni and Murphy 2010). They must be representative of a given
capability and be intuitively plausible. The latter criterion is required to facilitate the adoption
and successful application of a capability approach in public policy discussion and decision
making about risk. Indicators in the context of risk are designed to track changes in capabil-
ities, so that we can understand the impact of a given hazardous scenario on a community.
One question concerns which indicators will be both representative of a capability and provide
information about the change in capabilities due to a hazard.
Another issue is that, unlike many current applications of the capability approach, indica-
tors in the context of risk must be predicted. With risks we are always dealing with future
possible states, not analyzing existing states as is the case with many current applications of the
capability approach. The basis on which an accurate and credible prediction of changes in
capabilities can be made must be determined and, in particular, whether current methods for
quantifying risks can be used for this purpose or whether novel methods for predictionmust be
developed. One method (Gardoni andMurphy 2010; Murphy and Gardoni 2010) is to forecast
the impact of future hazards by first assessing the impact of past disasters using available data.
The impacts of past disasters can then be used to estimate the likelihood of each potential
outcome ofHII for a given hazard type,H, of magnitude,M. Mathematically, such likelihood is
expressed as the conditional Probability Density Function (PDF), P HII H ;Mjð Þ, of HII forgivenH andM. Using the Total Probability Rule (Ang and Tang 2007), we can then estimate the
The Capability Approach in Risk Analysis 39 995
probability of future societal impacts accounting for their likelihood of occurrence integrating
out H and M, as
P HIIð Þ ¼Z
P HII H ;Mjð ÞP H ;Mð ÞdHdM
where P H ;Mð Þ is the joint PDF of H ;Mð Þ, which can be estimated for typical natural hazards
based on for example meteorological or seismological considerations for hurricanes and
earthquakes, respectively.
Further research is also required to determine whether, and to what extent, indicators for
selected capabilities must be either time-dependent or risk-dependent. With respect to time-
dependence, the issue is whether in trying to understand the long-term versus immediate
impact of a hazard the indicators must change.
The second general topic for further research concerns the place of capabilities within
a comprehensive framework of risk evaluation and risk management. One specific question for
further research is the place of a consideration of capabilities among the additional, diverse
moral considerations that a judgment of the acceptability of a risk must take into account.
The capability approach to risk developed to date provides an account of how the capabilities
dimension of risk should be evaluated, which reflects the fact that capabilities are constitutive
dimensions of individual well-being and that justice requires a minimum level of capabilities
be guaranteed for all individuals. However, the consequences of a hazardous scenario are not
the only factor that is relevant for risk evaluation purposes. It is well-established that the public
is concerned about and distinguishes among risks on the basis of the source of a risk, that is,
how a risk was brought about or created (Wolff 2006; Murphy and Gardoni 2011). The public
distinguishes among risks on the basis of whether they were created through negligence,
recklessness, and non-culpable behavior. Risks are also evaluated differently depending on
whether they were voluntarily taken or not. The distribution of benefits associated with taking
risks and the potential harm that may be realized should a hazardous scenario occur is morally
salient (Fischhoff et al. 1981; Slovic 2000; Asveld and Roeser 2009). One question for further
research then is: when evaluating a risk, onwhat basis should we weigh or take into consideration
the expected impact of a given risk on capabilities relative to the other moral factors that matter?
A second question centers on the broader implications of a capability approach for issues in risk
management. That is, if we conceptualize risks from a capability perspective, how should that
inform the principles that should guide the creation and regulation of risk? Consider techno-
logical risks. Should, for example, new technologies be permitted only insofar as it can be shown
that such technologies will enhance human capabilities? Should the design of technologies be
guided by a concern for capabilities? There are a number of authors working on extending the
capability approach to areas like design ethics (Oosterlaken 2009), information technology
(Johnstone 2007; Wresch 2007; Zheng 2007; Coeckelbergh 2010b) and AI technology in health
care (Coeckelbergh 2010a). Another question is whether it is necessary to adopt a capability
approach to design or to technology insofar as one adopts a capability approach to risk.
Conclusion
This chapter has provided an overview of the reciprocal contributions of a capability approach
and risk theory. A capability approach can enrich risk theory in three ways: (1) The conse-
quences component of risk can be conceptualized and assessed in terms of the impact of
996 39 The Capability Approach in Risk Analysis
a hazardous scenario on capabilities; (2) Judgments of the acceptability of risks can be made on
the basis of whether certain threshold levels of capabilities aremaintained; (3) The primary aim
of risk management can be to protect and promote individuals’ capabilities, by prioritizing
unacceptable or intolerable risks and by selecting policy options on the basis of expected dollar
per unit change in the impact on capabilities. On the other hand, risk theory has enriched the
capability approach by (1) drawing attention to an important dimension of capability, namely
security; and (2) providing resources for developing a novel way to assess capabilities instead of
functioning. Further work is needed to make the capability approach to risk theory operatio-
nalizable and to clarify the place of a concern for capabilities in an overall perspective on risk
evaluation and management.
Acknowledgment
This research was supported primarily by the Science, Technology, and Society Program of the
National Science Foundation Grant (STS 0926025). Opinions and findings presented are those
of the authors and do not necessarily reflect the views of the sponsor.
References
Anderson E (1988) Values, risks, and market norms.
Philos Public Aff 17(1):54–65
Ang AH-S, Tang WH (2007) Probability concepts in
engineering: emphasis on applications to civil and
environmental engineering. Wiley, New York
Asveld L, Roeser S (2009) The ethics of technological risk.
Earthscan, London
Bedford T, Cooke R (2001) Probabilistic risk analysis:
foundations and methods. Cambridge University
Press, Cambridge
Coeckelbergh M (2010a) Health care, capabilities, and AI
assistive technologies. Ethical Theory Moral Pract
13:181–190
CoeckelberghM (2010b) Human development or human
enhancement? A methodological reflection on
capabilities and the evaluation of information tech-
nologies. Ethics Inf Technol. doi:10.1007/s10676-
010-9231-9
Crocker D (2008) The ethics of human development.
Cambridge University Press, New York
Cullen A, Small MJ (2004) Uncertain risk management
choices in risk analysis and society. In: McDaniels T,
Small MJ (eds) Risk analysis and society: an inter-
disciplinary characterization of the field. Cambridge
University Press, Cambridge, pp 163–212
Fischhoff B, Lichtenstein S, Slovic P, Derby SL, Keeney RL
(1981) Acceptable risk. Cambridge University Press,
Cambridge
Fukuda-Parr S, Kumar AKS (eds) (2003) Readings in
humandevelopment.OxfordUniversity Press,Oxford
Gardoni P, Murphy C (2009) A capabilities-based
approach tomeasuring the societal impacts of natural
and man-made hazards. Nat Hazard Rev 10(2):23–37
Gardoni P, Murphy C (2010) Gauging the societal
impacts of natural disasters using a capabilities
based approach. Disasters J Disaster Stud Policy
Manag 34(3):619–636
Haimes YY (2004) Risk modeling, assessment, and man-
agement, Systems Engineering and Management.
Wiley, Hoboken
Hansson SO (2007a) Philosophical problems in cost-
benefit analysis. Econ Philos 23:163–183
Hansson SO (2007b) Risk. Stanford encyclopedia of philos
ophy. http://plato.stanford.edu/entries/risk/. Accessed
23 March 2010
Hunter PR, Fewtrell L (2001) Acceptable risk. In: Fewtrell
L, Bartram J (eds) Water quality: guidelines, stan-
dards, and health. IWA, London, pp 207–227
Jahan S (2003) Evolution of the human development
index. In: Fukuda-Parr S, Shiva Kumar AK (eds)
Readings in human Development. Oxford Univer-
sity Press, Oxford, pp 128–139
Johnstone J (2007) Technology as empowerment: a
capability approach to computer ethics. Ethics Inf
Technol 9:73–87
Kaplan S, Garrick BJ (1981) On the quantitative defini-
tion of risk. Risk Anal 1:11–27
May P (2001) Organizational and societal consequences
for performance-based earthquake engineering.
PEER 2001/04, Pacific Earthquake Engineering
The Capability Approach in Risk Analysis 39 997
Research Center, College of Engineering, University
of California–Berkeley, Berkeley
Murphy C, Gardoni P (2006) The role of society in
engineering risk analysis: a capabilities based
approach. Risk Anal 26(4):1085–1095
Murphy C, Gardoni P (2007) Determining public policy
and resource allocation priorities for mitigating nat-
ural hazards: a capabilities-based approach. Sci Eng
Ethics 13(4):489–504
Murphy C, Gardoni P (2008) The acceptability and the
tolerability of risks: a capabilities-based approach.
Sci Eng Ethics 14(1):77–92
Murphy C, Gardoni P (2010) Assessing capability instead
of achieved functionings in risk analysis. J Risk Res
13(2):137–147
Murphy C, Gardoni P (2011) Evaluating the source of the
risks associated with natural events. Res Publica
17(2):125–140
Nussbaum M (2000a) Aristotle, politics, and human
capabilities: a response to Antony, Arneson,
Charlesworth, and Mulgan. Ethics 111(1):102–140
Nussbaum M (2000b) Woman and human development:
the capabilities approach. Cambridge University
Press, Cambridge
Nussbaum M (2001) Adaptive preferences and women’s
options. Econ Philos 17:67–88
Oosterlaken I (2009) Design for development: a capabil-
ity approach. Des Issues 25(4):91–102
Pate-Cornell ME (1996) Uncertainties in risk analysis: six
levels of treatment. Reliability Eng Sys Safety
54:95–111
Raworth K, Stewart D (2003) Critiques of the human
development index: a review. In: Fukuda-Parr S,
Shiva Kumar AK (eds) Readings in human
development. Oxford University Press, Oxford,
pp 140–152
Robeyns I (2005) The capability approach: a theoretical
survey. J Hum Dev 6(1):93–114
Robeyns I (2006) The capability approach in practice.
J Polit Philos 14(3):351–376
Rowe WD (1980) Risk assessment: theoretical
approaches and methodological problems. In:
Conrad J (ed) Society, technology, and risk assess-
ment. Academic, New York, pp 3–29
Sen A (1989) Development as capabilities expansion.
J Dev Plan 19:41–58
Sen A (1992) Inequality reexamined. Harvard University
Press, Cambridge
Sen A (1993) Capability and well-being. In: Nussbaum
M, Sen A (eds) The quality of life. Clarendon,
Oxford, pp 30–53
Sen A (1999a) Commodities and capabilities. Oxford
University Press, Oxford
Sen A (1999b) Development as freedom. Anchor Books,
New York
Sen A (2009) The idea of justice. Belknap Press of Har-
vard University Press, Cambridge
Slovic P (1987) Perception of risk. Science 236:280–285
Slovic P (2000) The perception of risk. Earthscan,
London
Sunstein C (2005) Cost-benefit analysis and the environ-
ment. Ethics 115:251–285
UNDP (United Nations Development Program)
(2000) Human development report 2000. Oxford
University Press, New York
UNDP (United Nations Development Program)
(2007) Human development report 2007. Oxford
University Press, New York
Vose D (2000) Risk analysis: a quantitative guide. Wiley,
New York
Wolff J (2006) Risk, fear, blame, shame and the regulation
of public safety. Econ Philos 22:409–427
Wolff J, de-Shalit A (2007) Disadvantage. Oxford Univer-
sity Press, Oxford
Wresch W (2007) 500 million missing web sites:
Amartya Sen’s capability approach and measures
of technological deprivation in developing coun-
tries. In: Rooksby E, Weckert J (eds) Information
technology and social justice. Information Science,
Hershey
Zheng Y (2007) Exploring the value of the capability
approach for E-development. Paper presented at the
9th international conference on social implications of
computers in developing countries, Sao Paulo