Amartya Sen: Capability and Well‐Being
Abstract and Keywords
Gaertner explores the aspect of measurability within Sen's capability approach, remarking that it is extremely difficult to determine weighing schemes for set evaluation and a set of basic functionings that captures definitions of what is basic in different cultural contexts. Furthermore, the interdependencies of freedoms and the variance in capability sets over time add to the degree of complexity in the evaluative exercise.
During the last three or four years the West German economy has witnessed an annual growth rate of GNP of between 3 and 4 per cent, while the number of persons out of work remained at roughly 2 million during the same period of time. Some people question this figure, others seem to get used to it. What is noteworthy is the fact that in spite of a thriving economy the unemployment rate seems to have got stuck at a rather high level (be it 2 million or only 1.5 million). In my opinion, the latter phenomenon should also be viewed from a somewhat different angle, namely the growing length of time many of these people have actually been jobless. It is being reported by doctors and psychologists that individuals who are out of work over a long period are suffering from this situation psychologically — and not only in terms of the obvious real income losses. They get isolated within society and start losing the capacity to do and initiate certain things, a capacity which they formerly possessed. This change is twofold. In Sen's terminology, their vector of functionings is altering, and at the same time their capability, namely, their ability to achieve ‘various alternative combinations of beings and doings’ is shrinking. This fact does not manifest itself in the official statistics, and quite often it remains unobserved in discussions about unemployment when growth and inflation rates, productivity changes, and other economic indicators are quoted.
It is still quite often the case that developing countries are only compared in terms of GNP per head. In his Commodities and Capabilities Sen has shown that India and China are close together in terms of GNP per head but quite far apart in terms of basic capabilities of survival and education, such as ‘the ability to live long, the ability to avoid mortality during infancy and childhood, the ability to read and write, and the ability to benefit from sustained schooling’ (p. 76).
These basic capabilities are certainly of great importance in an analysis of poverty. They are no longer of much interest in a comparison of highly developed countries, for the simple reason that most of the truly elementary functionings are achieved to a very high degree in those economies. It seems to me, however, that another set of basic functionings is more and more coming to the fore in highly industrialized countries, functionings we have forgotten to think about since everyone had accomplished them formerly. Some of these basic functionings are in danger of becoming unachievable (or only partially achievable), others appear to be out of reach for the time being. Let me just name a few: to drink tap‐water, to take a swim in rivers or in the sea, to eat fish from the sea, to breathe clean air in a metropolitan area, to walk around at night without fear, to live without the danger of an accident in a nuclear power plant.
(p.63) All these functionings are very basic though, admittedly, their aspect of ‘basicness’ is different from that in Sen's comparison between India and China. I find still another category of functionings worth mentioning, which seems to be closely correlated with the standard of living. Here, too, I want to mention only a few items: to receive further education, to be regularly employed, to take a holiday, to participate in social life. Some of these functionings are to some extent related to the notion of opulence. They have, however, a lot to do with a person's ability to choose between different ways of living, in other words, with a person's capability set.
The selection of a class of valuable functionings is obviously context‐dependent. In some cases context should be interpreted in terms of the stage of development of the economies under investigation, in other cases it would be more appropriate to consider the cultural and historical development of a particular society. Sen writes that the choice of the class of functionings is intimately related to the selection of objects of value.
Is there something like a complete list of relevant functionings? In Sen's comparison between India and China, for example, which is very much in favour of China, should one perhaps include an item reflecting the historical fact that under Mao many people had to give up their proper occupation in order to work in agriculture or the coal‐mines—a fact which severely limited the capability set of the Chinese intellectual at that time?
The following remark refers to the aspect of measurability within Sen's capability approach. Elementary functionings such as life expectancy, infant mortality, or adult literacy rate are relatively easy to measure, even on a cardinal scale. But how about more complex functionings such as achieving self‐respect, taking part in social and or political life, being happy in one's job? It seems to me that for these items the measurability issue is quite difficult to resolve. Could it be solved indirectly, say, via commodities and prices? My first reaction is that such a proposal would introduce a profound inconsistency into Sen's analysis. In Commodities and Capabilities Sen argues convincingly against the opulence‐focused approach and now, via the measurability issue, the perspective of commodity command would re‐enter the scene. On the other hand, one should not forget that the availability of good‐quality clothing at low prices, for example, makes it easier ‘to appear in public without shame’, a functioning mentioned by Adam Smith. Larger apartments provide the possibility of entertaining friends, and more leisure (i.e. fewer working hours) makes it easier to take part in the life of the community. Expenditure on leisure activities could, to some extent, serve as a proxy for the degree of social integration of a particular person or household.
In connection with the identification of the objects of value, Sen mentions the dominance ranking between vectors of functionings or capabilities. How far does the dominance relationship take us? This clearly depends on what is being compared to what.
It may be of some interest to report on a simple calculation which I have (p.64) made on the basis of a vector of basis capabilities (namely, infant death rate, life expectancy, number of inhabitants per medical doctor, illiteracy rate, consumption of calories) and GNP per capita. I collected these data for about 130 countries and looked at the percentage of cases where a comparison between any two countries was possible via a simple vector dominance relationship (thereby avoiding the question of attaching appropriate weights).
I found out that for East European countries a simple vector dominance held in 16 per cent of the binary comparisons; for Western democratic countries (including Canada, Japan, and the USA), the percentage rate was 17; between elements of the first group and the second group, it was 21.4; among the first thirty countries in terms of GNP per head (‘the richest’), it was 26; among the last thirty countries in terms of GNP per head (‘the poorest’), it was 23; between elements of the richest group and the poorest group, vector dominance held in roughly 90 per cent of all binary comparisons. This shows that for some of the underlying issues, the simple dominance ranking cannot take us too far. Weighting therefore seems unavoidable, and this is clearly emphasized by Sen.
There are many schemes for choosing relative weights—strictly speaking, there is an infinite number of weighting schemes. As long as there is no agreed‐on metatheory, any set of weights may be found arbitrary, except perhaps for an equal weighting system, which would satisfy an anonymity or neutrality condition. But neutrality itself is debatable. Maybe it is too much to ask for a metatheory since there are good reasons to argue that the construction of relative weights should depend on the proper context. Also, one should perhaps acquiesce in defining a certain range within which the relative weights could vary, a point mentioned by Sen. In an analysis and comparison of well‐being among very poor countries, Dasgupta (1989) has established a ranking of these countries in terms of the classical Borda method. It has often been argued that this method, too, is arbitrary, but Dasgupta's findings are quite illuminating, particularly his results on the degree of correlation between elements of a vector of positive rights and elements of a vector of negative rights.
Let us return to capabilities. Sen argues that a capability set should not be evaluated according to the actual achievements of a person (‘well‐being achievement’) but according to the set of real opportunities (‘well‐being freedom’). How can one define the set of real opportunities? This question is not an easy one. Is it true that well‐being freedom increases whenever a particular person's range of choice increases? That depends. It depend on the particular items which have become additionally available. Do new products, for example, increase well‐being freedom? This question was raised by Williams (1987) in response to Sen's ‘standard of living’ lectures (1987). The answer, I think, should be: not necessarily. A new washing powder normally does not create more freedom, since it does not represent a value‐object. It provides some functionings which are more or less irrelevant. Things may, however, be (p.65) different if the negative effects on the environment induced by the new detergent are less serious than those of the old product. It seems to me that for this kind of question the Gorman–Lancaster approach1 of looking at characteristics could be of some help. An evaluation in terms of goods characteristics is probably a lot easier to perform than an evaluation in terms of extensions of a person's freedom. However, not every new characteristic which is created is valuable, nor is any increase in some particular characteristic.
The above issue is closely related to the question of set evaluation. Can the value of the capability set be equated to the value of one of its elements, the chosen element, for example? Sen calls this rather truncated procedure ‘elementary evaluation’. Its merits are obvious, and if the selected element is chosen through maximization according to some intelligible criterion, then elementary evaluation may be satisfactory as a first approximation. The situation would, however, be quite different if the chosen element, that is, the achieved n‐tuple of functionings, was picked at random or arbitrarily.
The foregoing argumentation shows how difficult it is to assess well‐being freedom. A larger set of n‐tuples of functionings is not necessarily tantamount to a preferable set (sometimes a smaller set could even be a better set when information gathering and processing become too costly). The evaluative problem may be less complicated when the rights aspect in well‐being freedom is considered. It seems to me that there is a direct relationship between well‐being freedom and a particular society's bill of rights or, formulated more cautiously, a society's list of guaranteed—that is, actually protected—fundamental human rights. Freedom of thought means freedom of expression, freedom of the press is tantamount to access to a huge amount of information, freedom to choose one's place of work increases one's flexibility, and so forth. In other words, civil and political liberties increase an individual's capability set and therefore his or her well‐being freedom.
The situation again becomes more complicated when interdependencies are brought into the picture. My range of freedom, for example, is limited by your range of freedom, and vice versa. Also, capability sets do not seem to be given once and for all, that is, to be absolute and invariant. On the contrary, they vary over time due to cultural, economic, political, sociological, and technological changes (we have observed above that the choice of valuable functionings is context‐dependent). All this adds to the degree of complexity of the evaluative exercise.
There have, of course, been attempts to deal with some of the points mentioned above. The issue of rights and liberty, for example, has been extensively discussed by many philosophers as well as economists. In economics, Hayek and his liberal school are well known for their investigation into the relationship between economic actions and systems of rights. In social choice theory, (p.66) many scholars have considered the exercise of individual rights within collective choice procedures. Concerning the enlargement of opportunity sets, Koopmans (1964) formulated axioms which, when followed by an individual, make him or her prefer the augmented set to the original one.
In my opinion, much more work has to be done on these and related issues. The various questions which I have raised in my discussion should not give the impression that I am sceptical about the merit of Sen's approach. On the contrary, I find the focus on functionings and capabilities within an analysis of well‐being extremely significant and fruitful. Just to underline my conviction, I should like to refer again to the list of basic functionings mentioned at the beginning of my discussion, which apparently play an increasingly important role in highly industrialized countries—functionings, however, that are not adequately reflected in the official statistics many economists are used to working with. It would be misleading to think that Sen's approach to wellbeing and capability is primarily relevant to the analysis of poor nations, although his empirical studies in terms of the most elementary functionings have revealed an appalling degree of injustice in some of those countries. And the disclosure of this fact is, of course, highly significant, mostly for those who are suffering.
Dasgupta, P. (1989). ‘Well‐being: Foundations, and the Extent of its Realization in Poor Countries’. Discussion Paper, Stanford University, Stanford, Calif.
Gorman, M. W. (1956). ‘The Demand for Related Goods’, Journal Paper J3/29, Iowa Experimental Station, Ames, Iowa.
Koopmans, T. C. (1964). ‘On Flexibility of Future Preference’, in M. W. Shelly, II and G. L. Bryan (eds.), Human Judgments and Optimality. New York: John Wiley.
Lancaster, K. J. (1966). ‘A New Approach to Consumer Theory’, Journal of Political Economy, 74, 132–57.
Sen, A. (1985) Commodities and Capabilities. Amsterdam: North‐Holland.
—— (1987). The Standard of Living (The Tanner Lectures) Cambridge: Cambridge University Press.
Williams, B. (1987). ‘The Standard of Living: Interests and Capabilities’, in Sen (1987).