Measuring Human Development in a High-Income Country: A Conceptual Framework for Well-Being Indicators

Abstract

This paper is concerned with the construction of an appropriate conceptual framework for measuring human development with a focus on high-income countries. Too often, the measurement exercise is based on a purely empirical basis where indicators simply reflect data availability and “conventional wisdom”. This is likely to misguide policy-makers. We deal with two core points for the construction of a conceptual framework: (a) specification of the theoretical approach and (b) identification of the relevant categories of indicators. The paper endorses the capability approach which is the theoretical underpinning of human development. In line with this perspective, it offers a view of the relationships between key concepts such as human development, well-being, capabilities, and functionings. Based on this framework, it then tries to identify which typology of indicators is more suitable for measuring people's functionings. Building on a multidisciplinary literature, we classify indicators as input, output, outcome, and impact indicators, and conclude that outcome indicators are the best solution for measuring functionings. Finally, the paper provides examples of theoretically robust indicators and argues for a focus on more advanced functionings in high-income countries.

Keywords:

• human development
• well-being
• functionings
• social indicators
• high-income countries

JEL classifications:

• O150
• I300
• D630
• C430
• C180

1. Introduction

Most of the international debate on human development has drawn attention to low- and middle-income countries. However, in a historical moment in which governments employ mainstream austerity policies but there is a resurgence of alternative frameworks for studying present-day society, we believe that it is also essential to engage in this debate in the so-called “developed” societies. This paper is concerned with the construction of an appropriate conceptual framework for measuring human development with a focus on high-income countries. These human development indicators should provide a picture of people's life conditions in order to identify territorial disparities and tailor policies.

A rich international literature on human development, well-being, and quality of life indicators already exists. This includes the Canadian Index of Wellbeing (2011), the Happy Life Expectancy (Veenhoven, 1996), the Happy Planet Index (Abdallah, Michaelson, Shah, Stoll, & Marks, 2012), and the Better Life Index (Organisation for Economic Co-operation and Development [OECD], 2011). However, most of these international measures, including the official Human Development Index (HDI) of the United Nations Development Programme (UNDP), suffer from three interrelated shortcomings: (1) they do not have a sound theoretical foundation and often do not give a clear and rigorous definition of well-being and its dimensions; (2) indicators are assembled and utilized on a purely empirical basis and therefore simply reflect data availability and “conventional wisdom” (Galbraith, 1958); and (3) they fail to make the relevant distinction between “means” and “ends” of human development.1 In this paper, we try to address all these points.

What does identifying a conceptual framework mean? Following the guidelines of the Handbook on Constructing Composite Indicators (Nardo et al., 2008), elaborated by the OECD and the Joint Research Centre of the European Commission (EC-JRC), it means:

• clarifying the theoretical structure and defining the concepts (human development, well-being, etc.) accordingly,

• selecting the relevant dimensions within a multidimensional setting,

• selecting the relevant categories of indicators (Input? Output? Outcome? Impact?).

In this paper, we deal with the first and third point, while we do not elaborate on the problem of how to select human development or well-being dimensions.2 Therefore, in Section 2, we discuss the theoretical features of the capability approach and empirical issues related to its operationalization in terms of measurement. In Section 3, we review the scattered socio-economic literature—which also includes policy evaluation handbooks—that makes an explicit distinction between input, output and outcome indicators. With only this distinction in mind, we can identify indicators that more correctly measure—or can operate as proxies for—functionings. In Section 4, we combine these two literatures and discuss the indicators that, from a theoretical point of view, seem more suitable for measuring functionings in some relevant dimensions of human development. We will use examples of sound indicators for OECD countries. In this section, we also discuss how to deal with problems of data availability. In particular, we suggest that researchers dealing with these issues should be clear about the theoretical benchmark, why they rely on less than ideal solutions, and finally what are the assumptions—sometimes heroic ones—behind using less than ideal solutions rather than ideal ones. A direct consequence of this study is to verify whether the amount and quality of information available for high-income countries is adequate for making a valuable assessment of human development in the country as a whole and its territorial units (e.g., regions and provinces). Finally, Section 5 contains some concluding remarks.

2. The Capability Approach and Measurement Challenges

There are different concepts and terms in the human development and capability literature: human development, capability, functioning, agency, well-being, freedom, and others. Therefore, before discussing measurement, we must examine what should be measured.

First, it is useful to distinguish between “human development” and “capability approach” and clarify their relation. UNDP (2008) has summarized the core idea of the founder of the Human Development Report (HDR), Mahbub ul Haq, in this way:

Human Development is a development paradigm […]. It is about creating an environment in which people can develop their full potential and lead productive, creative lives in accord with their needs and interests. […] Development is thus about expanding the choices people have to lead lives that they value. […]

Fundamental to enlarging these choices is building human capabilities—the range of things that people can do or be in life.

From this brief definition and other works (Fukuda-Parr, 2003a, 2003b), it is clear that human development is a “paradigm” and is about building human “capabilities”. In fact, this paradigm

is based on what economist and Nobel laureate Amartya Sen calls the capabilities and functionings approach. Not only it is important to achieve more ‘functionings’, but it is essential for people to have the ‘capabilities’ or the freedom to achieve these. (UNDP, 2007, p. 1)

Sen (1989) himself titled one of his papers “Development as Capability Expansion.”

On the other hand, according to Robeyns (2011, p. 1):

The capability approach is a theoretical framework that entails two core normative claims: first, the claim that the freedom to achieve well-being is of primary moral importance, and second, that freedom to achieve well-being is to be understood in terms of people's capabilities, that is, their real opportunities to do and be what they have reason to value.

Therefore, while human development is a “paradigm”, the capability approach is its underlying “theoretical framework”; in other terms, the capability approach provides the theoretical foundations to the human development paradigm. Since the latter is concerned with building and expanding capabilities, measuring human development or measuring capabilities is substantially equivalent.

Before discussing problems concerning the measurement of capabilities, we must consider their building blocks, namely the functionings.3 In fact, a person's capability “represents the freedom that a person has in terms of the choice of functionings, given his personal features […]. It reflects the various combinations of functionings (‘beings’) he can achieve” (Sen, 1985, pp. 10–11). Consequently, unless we make orthodox hypotheses of individual maximizing behavior (Sen, 1999), evaluating achieved functionings is substantially different from evaluating capabilities, since the latter includes both achieved functionings and other accessible valuable functionings that have not been chosen by the person, which represent the opportunity aspect of his or her freedom.

Sen (1999, p. 75) claims that:

The evaluative focus of this ‘capability approach’ can be either on the realized functionings (what a person is actually able to do) or on the capability set of alternatives she has (her real opportunities). The two give different types of information—the former about the things a person does and the latter about the things a person is substantively free to do.

Therefore, if we go back to our initial question about “what we should measure,” there are at least two possible answers: we could measure achieved (or realized) functionings or the capability set of alternatives. Since a person's well-being can be seen as an evaluation of his or her achieved functionings (Sen, 1985), choosing the first focus means measuring people's well-being. A relevant implication is that well-being should be considered a constitutive aspect of human development. If we choose the second focus, we should measure the various functionings that are valuable and feasible for a person (regardless of whether they are chosen or not) or, in other words, his or her substantial freedoms. Thus, while the focus on achieved functionings lead us toward the measurement of well-being, the focus on the broader capability set leads us towards the measurement of freedom.4 It is important to bear in mind that, according to Sen, both evaluative focuses belong to the capability approach. Consequently, measuring human development includes both well-being and freedom.

The choice between the two focuses essentially depends on two circumstances: the objectives of the evaluation and which information can be used or collected. In some cases, the focus on achieved functionings can be sufficient, at least at a first stage: for example, when analyzing famines or extreme poverty, if there are tight time constraints, and when we need rapid assessments. In other cases, the focus on achieved functionings is chosen because of available information. This is the case of the human development indicators elaborated by UNDP: given the objective of having systematic and regular intercountry comparisons, the only comparable information available concerns a few achieved functionings. In fact, the HDI is a well-being index.

Considering that almost all statistical systems, with very few exceptions, are designed to collect information only on the socio-economic conditions of people and not on what they are “substantively free to do”, standard available statistics include scarce or no information regarding capabilities.5 Thus, capabilities cannot normally be measured using traditional available data. Until statistical institutes systematically collect information on capabilities, the only way to measure capabilities is by designing and carrying out ad hoc surveys so that the necessary information can be collected.

However, although the evaluative focus on the capability set leads to a full-fledged application of the capability approach, “the problem of valuation is not, however, one of an all-or-nothing kind” (Sen, 1999, p. 78). Looking forward to innovations in statistical systems that cover information on capabilities, we can intelligently utilize the available statistics to measure achieved functionings. There are often plenty of social, demographic and economic statistics available, especially in high-income countries, that can be utilized to measure functionings, i.e. well-being.

In the following sections, we will discuss the main conceptual and methodological issues regarding generally available socio-economic indicators that could be utilized in the measurement of well-being as achieved functionings, with special reference to high-income countries. Paradoxically, although there are more socio-economic data and quality of life indicators available for high-income countries than for other countries, there are only few human development measures available for the former. Apart from the HDR for the USA (Burd-Sharps, Lewis, & Martins, 2008) and the standard human development indicators available in the global HDRs produced by the UNDP, there are no national HDRs for high-income countries that could include sub-national indicators or country-specific measures. Although this gap is largely due to the right priority assigned by the international community to countries with low and middle human development levels—reflected in the institutional mandate of UNDP, it is puzzling why national and regional (e.g., European Union) institutions, both governmental and research, of high-income countries have not decided to investigate their human development.

3. Categorization of Socio-Economic Indicators

In this section, we focus on the scattered literature dealing with the “categorization” of indicators, i.e., with specific features that indicators should have in order to represent a phenomenon. Here, we do not refer exclusively to indicators of well-being, quality of life or human development, but to a broader range of socio-economic indicators.

For a long time, the standard economic literature has talked about the production function, that is, the process through which a commodity or national domestic product is obtained. Given a set of inputs that range potentially from physical, natural, and human capital to labor and institutions, we can obtain a given amount of output. Economic analysis, mainly based on neoclassical principles, is interested in understanding which inputs to use and how to combine them in the “optimal” way, i.e., to obtain the highest possible amount of output. The output/input ratio, thus, is the core of the efficiency analysis.

The distinction between inputs and outputs is also at the center of program and policy evaluation. In the logical framework—a well-known tool to plan an intervention and also allow a control by external financing agencies—we find space left to input and outputs with related “indicators”. In the input cells we often find items, such as resources allocated to the building of a school or a hospital, while output indicators are more concerned with the number of schools and hospitals built, and the number of pupils attending school. Inputs allow outputs to be produced which, based on a series of assumptions stated in the phase of project design, should help to reach the intended goals. Even external financing organizations—whether they are the European Union or national ministries—rarely challenge the initial ratio of the project: thus, they often limit their work to “controlling and measuring the inputs and immediate outputs of a program—how much money is spent, how many textbooks are distributed—rather than on assessing whether programs have achieved their intended goals” (Gertler, Martinez, Premand, Rawlings, & Vermeersch, 2011, p. 3).

The literature on the measurement of education, health, well-being, and related phenomena has also made extensive use of the classification of indicators in input and output indicators, although these definitions do not always converge (Adelman & Morris, 1972; Booysen, 2002; Dasgupta, 1990; McGranahan, 1972; Nardo et al., 2008). For the measurement of well-being, Dasgupta (1990) uses this categorization: following the basic needs approach as a conceptual framework, he argues that the consumption of goods and services, such as health facilities, medicine and education, should be conceived as inputs for the “production” of final outputs, namely life expectancy, literacy, and knowledge.

In a recent relevant work prepared by the OECD and the EC-JRC, Nardo et al. (2008) indicate a series of steps to follow for the construction of composite indicators. The first one concerns the identification of a “theoretical framework”: depending on the conceptual apparatus a researcher follows, he/she may provide slightly different definitions of the concept being measured (e.g., well-being or quality of life) and select different dimensions. The following element of this step consists in choosing whether to use input, output, or “process” indicators. As argued in this handbook, “too often composite indicators include both input and output measures” (Nardo et al., 2008, p. 22). Therefore, the measurement part has to follow the theoretical framework strictly: this implies being aware that all the indicators used to measure the different dimensions have to be of the same type (input or output). For example, an index of innovation performance could incorporate a number of new products and services (output) and not research and development expenditures.

In the specific debate on development indicators, the dichotomy input versus output indicators often coincides with the dichotomy means versus ends of development (Booysen, 2002). Some authors argue that composite development indicators should point to either means or ends (e.g., Adelman & Morris, 1972), while others such as McGranahan (1972) claim that these variables should all focus on intended ends (or outputs). Besides some differences—which will be partly addressed in the following section dealing with well-being and human development indicators—the thought of all the above authors is in line with the recommendations of OECD-JRC to combine only indicators of the same typology.

There are serious problems in putting indicators either in the input or output box. That is why, more recently, scholars and organizations/institutions have tried to move beyond this strict dichotomization. Similarly, it is limiting to think about means and ends: we should at least refer to intermediate and final ends since we refer to semi-finished and final products in economic language. Therefore, nowadays there is some convergence in classifying indicators as input, output, outcome, and impact indicators (Gertler et al., 2011; Mandl, Dierx, & Ilzkovitz, 2008; Save the Children UK 2008; UNDP, 2010; Vos, 1996). This distinction is important for program monitoring and evaluation activities: “monitoring is a continuous process” that usually “tracks inputs, activities, and outputs,” while evaluation is a “periodic, objective assessment” of the outcome and the impact of the program, i.e., its long-run effects and the effects on the whole society (Gertler et al., 2011, p. 7).

Mandl et al. (2008) provide a very interesting graphical illustration of the linkages between input, output, and outcome indicators. In their words, “The monetary and non-monetary resources deployed (i.e. the input) produce an output. For example, education spending (input) affects educational attainment rates (output). The input-output ratio is the most basic measure of efficiency.” In contrast, “Effectiveness relates the input or the output to the final objectives to be achieved, i.e. the outcome.” Finally, impact indicators refer to the potential (positive and negative) effects on other sectors.

Even with this extended categorization, we can encounter problems in allocating some indicators to one box or another. Thus, we endorse the argument advocated by Vos (1996, p. 4) according to which, “it is best to think of a chain of indicators” highlighting causal relationships between basic input indicators, output, outcome, and impact indicators. Given these problems, we attempt to classify education indicators as follows:

• Input indicators: Public and private expenditures in education, school resources (monetary input), number of teachers and students (their ratio can be itself conceived as an input indicator of the quality of education), class size and instruction, teaching material, quality, and adequacy of curriculum.

• Output indicators: Enrolment rate, attendance rate, dropout rate, and repetition rate (sometimes also defined as measures of “access” to education).

• Outcome indicators: Completion rates (in between output and outcome indicator), literacy rates, expected number of completed years of schooling, and standardized test measures of student and adult achievement in terms of literacy and numeracy.

• Impact indicators: Earnings/wages, employment/unemployment rate, nutritional indicators, and health indicators.

Now, if the aim of a measurement exercise is to find an indicator of the supply of an education system, then we may think of focusing on input indicators; if the aim is to verify the immediate result of the effort in the education system, we can rely on output indicators or on a input/output ratio; finally, if the final goal is to verify the changes in the education performance, then we need to utilize outcome indicators or, indirectly, impact indicators.

As a final remark, it is important to clarify that this framework works in a static setting and analyzes well-being dimensions separately. If we look at these relationships dynamically, educational outcomes become means or inputs to health and nutritional outcomes which, in turn, may be inputs into the “production” of outcomes related to emotional well-being and so on. In a dynamic setting, the system becomes far more complex.

4. Measuring Functionings in Some Relevant Dimensions of Human Development

Given the premises outlined in the previous section, how can we measure well-being in a developed country? Following the OECD-JRC guidelines, the initial point consists in the identification of the theoretical framework: as argued in Section 2, we endorse the capability approach. Leaving aside the very widespread debate on well-being dimensions, which indicators should we use? As argued by Nardo et al. (2008), we need to focus on either input, output, outcome or, finally, impact indicators and avoid mixing them up.6 Let us start with the official UNDP indicator of human development—the HDI.

The HDI—both the old one (before 2010) and the new one—includes input indicators (gross domestic product (GDP) or gross national income (GNI)), output indicators (e.g., combined gross enrolment rate) as well as outcome indicators (e.g., life expectancy at birth). This incoherence, however, should be analyzed in view of this indicator's specific goal. The choice of having a simple composite indicator, which includes both economic and social indicators for a comparison between all the countries in the UN system, is at the center of this final choice. Few dimensions, with both economic and social variables, were needed to have an indicator able to compete with the traditional economic indicators of development and well-being; the cross-country comparison, then, imposed the selection of variables for which data were available, starting with national income.

Using the words of Veenhoven (2005, p. 61), who uses the simple input–output classification,

There are two ways to assess how well people live. One is to consider to what extent the country provides conditions deemed essential for a good life. In this approach the emphasis is on societal input … The other approach is to assess how well people thrive. In this approach the emphasis is on societal output.

As argued by Chibber and Laajaj (2007, cited in Kovacevic 2010, p. 6), in the case of the HDI “The focus point should be the desired outcome rather than the means to reach these outcomes.” More in general, the well-known Sen–Stiglitz–Fitoussi Commission concluded that

First is the emphasis on people, on what they value as important for their daily life, and on the environment in which they develop. Taking the individual as the fundamental unit of analysis does not imply neglecting communities and institutions, but requires evaluating them in virtue of what they bring to the quality of life of the people participating in them. This perspective also implies focusing on the ‘ends’ of various human activities, while recognizing that their achievement can matter both intrinsically and instrumentally (i.e. to achieve some other goal). (Stiglitz, Sen, & Fitoussi, 2009, p. 144)

Therefore, in order to build a well-being indicator theoretically rooted in the capability approach, we should use variables indicating how people fare with regard to the relevant elements of their life and not others focusing on the means to achieve people's well-being. At the same time, impact indicators of education, such as people's wages and employment conditions, are the result of educational levels and many other interacting factors and therefore they are not adequate for measuring the phenomenon.

In addition to being incoherent with the human development framework, the use of input indicators has two important weaknesses: (1) it assumes a straightforward relationship between the means and the outcomes (sometimes not observed), and thus it does not consider, among other things, the role of personal, environmental, social, and institutional conversion factors (Sen, 1985)7; (2) it does not leave each country the possibility of identifying its way to enhance people's well-being (Chibber & Laajaj, 2007). Given that what gets measured gets done, if a means to enhance well-being is included in the indicator, we are sending the message to policy-makers that their performance will be evaluated on the basis of that instrument. This is a problem because that instrument may be only one of the means to enhance people's well-being and because it violates the principle of ownership, i.e., the possibility that national and local policy-makers can choose the policies to implement.

In some cases, having outcome indicators is more complicated. As argued in Section 3, people's education should be ideally measured by their competences and abilities. While the amount of data of this type for developed countries has recently increased [e.g., the Programme for International Student Assessment (PISA) standardized test scores], this information can still be lacking for many areas. In these situations, we can use proxy measures (or output indicators) such as enrolment or attendance rate (Nardo et al., 2008). In these cases, it is important to highlight the limitations of this variable: “indicators of school enrolment inform about access to education, but they may provide a misleading picture of outcomes if schools do not provide effective instruction” (Stiglitz et al., 2009, p. 167).

In Table 1, we attempt to classify indicators in outcome and output indicators for different well-being dimensions and specific capabilities/functionings. For any well-being dimension, in fact, we have more capabilities/functionings: in the health dimension, for example, we have the capability “being able to prevent avoidable diseases” or more advanced capabilities such as “being in a good health status”. Similarly, in the education dimension, “being literate” can be considered as a basic capability, while “having basic education” or “having higher skills/knowledge” is a more advanced one. Though originally proposed for all countries, given its mandate, the UNDP has applied the human development approach almost entirely to low- and middle-income countries. As a consequence, the HDI is also an indicator based on a conceptualization of human development as the expansion of very basic capabilities. As Anand and Sen (1993) argue, if we wish to apply the same concept to high-income countries, where most of the elementary capabilities are activated for almost the entire population, we need to focus on more advanced capabilities. Indeed, literacy rates or morbidity rates for infectious diseases are not sufficient to show disparities in education and health among developed countries and regions.

Table 1 Examples of Output/Outcome Indicators for Different Well-Being Dimensions and Functionings in a High-Income Country

Well-being dimension	Capability/functioning	Output indicator	Outcome indicator
Health	Being in a good health status	Percentage of population with access to different health services when needed	(1) Healthy life expectancy at birth(2) Disability-adjusted life expectancy at birth(3) Life expectancy at birth(4) Median health conditions (subjective measure)
	Preventing avoidable diseases	Percentage of children immunized against infectious diseases	(1) Morbidity rates for specific diseases (e.g., tuberculosis, malaria, measles, and HIV-Aids)^a
Education	Having basic education	(1) Enrolment rate in primary education(2) Attendance rate in primary education(3) Enrolment rate in lower secondary education(4) Attendance rate in lower secondary education	(1) Completion rate in primary education(2) Completion rate in lower secondary education(3) Average results in PISA standardized test scores(4) Average results in INVALSI test scores
	Having higher knowledge/skills	(1) Enrolment rate in upper secondary education(2) Attendance rate in upper secondary education(3) Enrolment rate in tertiary education	(1) Completion rate in upper secondary education(2) Percentage of population with a university diploma (BA)(3) Percentage of population with a university degree(4) Percentage of population with a Ph.D.
	Being literate	(1) Enrolment rate in primary education(2) Attendance rate in primary education(3) Percentage of adult population that has attended literacy programs	(1) Average results in PISA standardized test scores(2) Average results in INVALSI test scores(3) Percentage of adult population able to read and write
Shelter/safety	Living in a decent shelter	Number of dispossess per 1000 citizens/buildings^a	(1) Percentage of population living in a decent house (pavement, roof, bathroom, access to potable water, electricity, etc.)
Employment	Holding a decent job	(1) Employment rate(2) Unemployment rate^a(3) Not in Education, Employment or Training (% of 15–64 y.o. population)^a	(1) % of 15–64 y.o. population with a permanent job(2) % of 15–64 y.o. population with a precarious job^a(3) % of 15–64 y.o. population with a full-time job(4) % of 15–64 y.o. population with a part-time job(5) % of 15–64 y.o. population with a “protected” contract
Participation	Participating in political/civic life	(1) Percentage of the population registered in an association	(1) Percentage of the population >14 y.o. carrying out free-of-charge activity for associations/unions(2) Percentage of the population >14 y.o. participating in meetings of associations(3) Percentage of the adult population voting(4) Percentage of the population >14 y.o. getting informed about politics
Nutrition^b	Being adequately nourished	Percentage of the population with access to the minimum amount of calories required	(1) Malnutrition rate(2) Obesity rate(3) Average body mass index

a These variables indicate deprivations. To transform them into achievements, the final indicators must be calculated as “1(or 100) —deprivation indicator”.

b For an in-depth analysis of how to analyze food and nutrition security through the capability approach, see Burchi and De Muro (2012).

Table 1 is based on data available for most of the OECD countries given the enormous effort to uniformize statistics in these countries made in the last two decades. For example, in the health dimension, a suitable indicator would be the healthy life expectancy at birth, as suggested by the World Health Organization and the European Union.8 This indicator is better than life expectancy because the latter also depends heavily on other factors not related to health (car accidents, medical and technological progress, etc.). Life expectancy tells us how long people live and not whether they live in good health conditions. An example of output indicator is the percentage of population having access to different health services when needed: like the enrolment rate in education, this indicator provides information on the use of services, but the final outcome depends on many other factors, including the quality of the service.

Another stream in the literature makes a distinction between stock and flow variables. The new HDI, for example, includes stocks (mean years of schooling), flows (GNI), and “two expectations of stock variables conditional on current flows (life expectancy and expected years of schooling)” (Klugman, Rodríguez, & Choi, 2011, p. 21). Most of the literature reaches the conclusion that mixing them in one indicator is not necessarily a problem as long as they are justified from a theoretical point of view (Kovacevic, 2010). However, the HDI has been criticized because the presence of stock (and expected stock) variables does not make the composite index very responsive to external shocks or to the implementation of important policies in sectors like education and health. According to this view, the indicator is only a measure of “the outcomes of past efforts rather than the effects of present or recent policy changes” (Kovacevic, 2010, p. 6).

In order to address this point, we need to specify the final goal of the measurement exercise. Until now, we have implicitly considered the objective of a well-being indicator to be offering a static picture of people's life conditions. Our effort was to describe which outcome indicators could best serve this purpose. If this is the objective, is it really a problem if the indicator measures “the outcomes of past efforts”? Our answer is “no”: even the best policies in key human development sectors need time to have substantial impacts on the life of people. In contrast, if the objective of the measurement exercise is to find an indicator to use to make a prompt evaluation of the immediate impact of a policy or an external shock, we need to revise the traditional indicator. In particular, it should include only flow variables, which in many cases are equivalent to output indicators. For example, what is the short-run impact of the current economic crisis on health in Europe? Looking back at Table 1, we could not use outcome indicators such as the life expectancy at birth because we simply could not find any impact. To this end, using an output indicator like the proportion of people having access to health services when needed seems to be a better choice.

5. Conclusions

Many international indicators of well-being, quality of life, and related phenomena suffer from several problems ranging from the lack of a sound theoretical foundation and specification of the final objective, to the combination of input and output indicators without adequate justification. For this reason, we attempted to build a conceptual framework for measuring human development in high-income countries. Measuring and assessing human development in industrialized countries—an exercise rarely done—is of extreme importance for at least three good reasons: (1) to identify territorial disparities and tailor economic and social policies; (2) for the “dethronement of G.N.P.” (Morse, 1971); and (3) because description is prescription (Sen, 1980).

In this paper, we deal with two of the three core points for a construction of a conceptual framework, namely: (a) the clarification of the theoretical approach, which is needed to define the concept of human development and its pillars; and (b) the identification of the relevant categories of indicators. We do not engage in the discussion on how to select the relevant human development dimensions.

In Section 2, we provide a definition of the core concepts behind the capability approach and elaborate on the relationship between the capability approach and the human development paradigm, with the former being the theoretical foundation of the latter. We also stress the problems encountered when measuring human development. The most important consists in the choice between functionings and capabilities: while capabilities are the most adequate evaluative space, in order to measure human development in OECD countries and their territorial units, we need to rely on functionings. As argued by Sen, this is still coherent with the theoretical framework of the capability approach. Moreover, it allows us to use the massive—and often poorly used—statistical information on functionings available for high-income countries.

In Section 3, we review the literature on categorization of socio-economic indicators. While, in the past, indicators were divided into simple input–output indicators, more recently, we find a broader classification in input, output, outcome, and impact indicators in both documents on well-being indicators as well as impact evaluations. In the education domain, for example, we move from very basic input indicators, such as the public expenditure on education, to output indicators such as enrolment rates, and outcome indicators such as the results in the PISA standardized test scores for EU countries. Impact indicators, on the other hand, refer to externalities generated, for example, by public expenditures in education in non-education domains such as employment and health. As a conclusion of this analysis, in the literature we found substantial agreement on using only one category of indicators to measure socio-economic phenomena.

Using the capability approach as theoretical framework and considering the aforementioned classification of indicators, in Section 4 we elaborate on the typology of indicator that is more suitable for measuring functionings. We conclude that outcome indicators are the best solution since they reflect the real ends of human development. When data on outcomes are missing, a reliable assessment can still be made using output indicators; however, we suggest that researchers specify why they use second-best indicators and what are the limitations and assumptions made when using second-best rather than ideal indicators.

We also present examples of outcome and output indicators for six dimensions (health, education, shelter/safety, employment, participation, and nutrition) and for different functionings within each dimension. The latter is important because one of the crucial revisions to the human development indicators required when we concentrate on developed countries consists in the focus on more advanced functionings within each dimension: two examples are “being in a good health status” and “having higher knowledge” rather than more basic functionings such as “preventing avoidable diseases” and “being literate.”

Finally, we hope that this paper will contribute to a more rigorous approach to the measurement of human development, quality of life, and well-being and move away from “measuring without theory” (Koopman, 1947). Indeed, a lack of rigor in this exercise may well lead to the adoption of inappropriate policies.

Acknowledgements

The authors would like to thank the participants in the PRIN Workshop and in the EAPE conference, held in October 2012 in Modena and Krakow, respectively.

Additional information

Funding

This paper is an outcome of the National Research Project (PRIN) “Measuring Human Development and Capabilities in Italy: methodological and empirical issues” co-funded by the Italian Ministry of Education, University and Research (MIUR).

Notes

¹ For a further analysis of the limits of current well-being indicators in Italy, see Burchi and Gnesi (in press).

² For examples of methods used to select relevant dimensions, see Alkire (2008), Burchi et al. (2014), Vizard and Speed (in press).

³ A functioning is an achievement of a person: what he or she manages to do or to be. It reflects, as it were, a part of the ‘state’ of that person. (Sen, 1985, p. 10)

⁴ Nevertheless, some special achieved functionings can be considered equivalent to instrumental freedoms. A good example is “being educated,” and more importantly “being literate,” that expands our meta-capability to value and to choose. Consequently, measuring those special functionings does not give us just an evaluation of well-being, but also a partial evaluation of some relevant aspects of freedom.

⁵ Indeed, there is an increasing amount of available statistics related to various aspects of subjective (self-reported) well-being, such as happiness or life satisfaction. This trend reflects the growing attention, both at scientific and political levels, that such approach receives (although it is by no means a novel one). Some scholars (for instance Anand et al., 2009) think that information on subjective well-being can be utilized to measure capabilities. This hypothesis would deserve a critical scrutiny in the light of the issue of “mental conditioning and adaptive attitudes” discussed by Sen (1999, pp. 62–63).

⁶ Similarly, Klasen (2012) criticizes the framework at the basis of the United Nations Millennium Development Goals because some indicators measure well-being outcomes, while others outputs or inputs.

⁷ The conversion factors play a central role in the capability approach. They are fundamental to make a comprehensive assessment of human development or well-being, and to understand why resources may not be highly correlated with functionings. In this paper, we do not investigate in-depth the role of conversion factors as they are less relevant for the measurement exercise. By using outcome indicators, we implicitly account for the presence of conversion factors.

⁸ A detailed description of this indicator can be found at http://www.who.int/healthinfo/statistics/indhale/en/ (consulted: 5 June 2013).