Wellbeing Measurements

The experience of wellbeing varies with time

The salience of non-commodities casts doubt on the welfarist assumption that all well-being can be priced. Together with the rise of the service sector, the emergence of extended accounts may be regarded as acknowledging the secular shift of welfare away from its prior association with material production. Extended accounting relies heavily on the allocation of time. There is also a strand of research which takes time-use as the measure of welfare. This lends itself both to micro and to macro applications.

The concept of sustainable consumption goes back to Hicks, who defined it as the maximum value of consumption which would leave the individual afterwards as well-off as before. In their study of 1972, Nordhaus and Tobin attempted to estimate 'sustainable' welfare, which they took as consumption plus net investment.18 They acknowledged the need to take account of the depletion of non-renewable natural resources. This had been anticipated by Kapp in 1950, and several estimates were produced during the 1970s. Weitzman provided theoretical underpinning: national product net of asset depletion also described the discounted sustainable productive potential of the economy. A more normative and radical approach was pioneered by Zolotas in 1981. He incorporated pollution and natural resource depletion into a set of extended accounts for the period 1950-1975, and also imputed shadow costs to some social detriments. His index of the economic aspects of welfare [EAW] rose progressively more slowly than GNP, and he envisaged a time, a generation hence, when an increment of GNP would produce no welfare at all. He argued that this was a systemic feature: 'beyond a certain point, economic growth may cease to promote social welfare. In fact, it would appear that, when an industrial society reaches an advanced state of affluence, the rate of increase in social welfare drops below the rate of economic growth, and tends ultimately to become negative.'

The national accounts are silent about distribution, reflecting perhaps the utilitarian bias of their origins. Concern over inequality has motivated a good deal of the effort to devise measures of welfare. Atkinson's influential index provided a measure for evaluating the effect of income inequality on welfare, which could be adjusted to the amount of inequality tolerated (or desired). This has been applied to extended accounts by Beckerman, Crafts, and Jackson et al.

After a hiatus in the 1980s, extended accounting took a more radically critical turn. Daly and Cobb continued to develop the Zolotas model: they incorporated inequality (based on Gini coefficients) into a new measure, the Index of Sustainable Economic Welfare (ISEW). This had the effect of depressing the index: inequality has worsened since the 1970s, and was one of the main offsetting effects of economic growth. The principle of 'sustainability' in the rubric referred primarily to the depletion of non-renewable resources. While GNP continued to grow, the British ISEW declined overall by about 50 percent between the 1975 and 1990. Later versions show smaller declines. The innovation has caught on, and ISEW measures are available for Australia, Austria, Chile, Germany, Italy, the Netherlands, Sweden, the UK and the USA. In all of these countries except Italy, ISEW shows growth until the 1970s, with stagnation or decline afterwards.

If we are to think of society as a unitary actor, then according to the ISEW, the growth in economic activity since the mid 1970s has been producing not an increase, but rather a reduction in aggregate welfare. Since the 1970s, from this perspective, the pursuit of further growth has been irrational. It is only myopia and habit which allow it to continue in the face of negative welfare returns. The problems of aggregation should be borne in mind, and the imputation of environmental depletion is open to criticism.

Welfare measures make cognitive demands, which define the user community. The choice of what to measure is essentially normative, i.e. it requires some level of social assent. For wide acceptance, cognitive demands need to be kept to a minimum. It is 'headline' measures that count. Politicians and activists operate in conditions of attention constraint. Hence successful indicators are typically in the form of simple cardinal or interval summary measures, such as the growth rate, inflation, or central bank base rate. Binary indicators, such as the 'poverty line' and the unemployment rate, also have a deceptive simplicity. At the other extreme, academic experts have ample cognitive resources, and compete in analytical subtlety.

'Sustainability' has become a normative public policy objective in the UK. This perspective of diminishing returns to economic growth is captured in research programme for the study of normative social indicators. In the 1960s and early 1970s the idea that access to certain goods constitutes a precondition of welfare. re-emerged as the 'social indicators' movement. This was also inspired by the idea that real welfare was not captured by the System National Account indicators. Typically the goods in question consisted of nutrition, housing, education, health and life expectations, environmental quality, crime, and poverty levels. They might also include such objectives as the freedoms of movement, expression, and political organisation. Implicit in social indicators is some notion of adequacy: there is too little of some things, such as nutrition, housing or education; or too much of others, such as poverty, inequality or crime. Social indicators are rarely scaled in the metric of money, or set within an accounting framework.

By the early 1970s, several leading countries and international bodies had published one-off or serial collections of social indicators. This enterprise has not abated. Social indicators relied implicitly on a social-democratic consensus, with an egalitarian bias and a quest for social inclusion. But there was a lag between impulse and execution, and by the time social indicators were delivered, the impetus of social democracy was spent. Priorities for social expenditure had already been set in the 'golden age' period of expansion and the 1970s were a period of fiscal retrenchment. Social consensus swung away from equality and towards competition, from the left towards the right. The absence of a coherent accounting framework was another disadvantage.

In developing countries, deprivation was not relative but absolute. In the 1970s a 'basic needs' movement identified a bundle of goods that might claim priority over economic growth. Morris argued that if encompassing was beyond reach, there was a virtue in parsimony. He introduced an unweighted ‘Physical Quality of Life Index' [PQLI], made up of infant mortality, literacy and life expectation at age one, as a single measure of welfare.

What followed shows how social indicators not only depend on norms, but can also help to create them. The focus on basic needs came into conflict in the 1980s with the World Bank/IMF 'structural adjustment' policy, and with the increasing market orientation within development economics. The results of these programmes have been mixed, but the impression was that costs fell often disproportionately on the poor. In the late 1980s, dissatisfaction with the 'structural adjustment' programme inspired the creation of a new social indicator, the Human Development Index [HDI].40 This is made up of income per head, life expectation at birth, and an education indicator, expressed in a single figure between 0 and 1. It has gained wide acceptance, and may have played some role in the partial retreat of the World Bank, and its acknowledgment of poverty as a policy objective.

It is interesting to compare HDI and GNP with Sen's 'capabilities' approach, which has attracted a great deal of discussion. Sen moved from an axiomatic 'welfarist' position to the view that income alone does not satisfactorily capture welfare. In keeping with Liberal values, he has not privileged any particular good. Even under indigence it was necessary to respect individual priorities.44 Well-being constitutes having the 'capabilities' to achieve valuable 'functionings'. Both of these categories extend beyond the purely economic. Sen has not embodied his approach in any metrics (it has influenced the Human Development Index), so for all of its normative cogency and conceptual sophistication, his work has not yet achieved the policy impact of the HDI. 'Alternative' approaches, especially Sen's capability/functioning approach, and the various 'sustainability' measures are congruent to some extent with non-utilitarian ethical frameworks. The most extensive social-indicator test so far, has confirmed that the relation of well-being and income per head is weak, both in cross section and over time.

Association between health outcomes and economic status has long been observed internally within countries.51 It is not entirely clear how much ill-being arises simply from material deprivation, and how much from the psychic costs of exclusion. What is lacking are standard, simple, social indicators of the consequences of inequality for affluent societies, an index of deprivation and detriment similar perhaps to the PQLI or HID. Such indices are currently under construction by Michael Noble and his team at Oxford.

Indicators are all oriented towards development. They are essentially catch-up indices, calibrated to current best practice. They fail to address the original impulse of the social indicators movement, which was finding a way of measuring welfare in affluent societies; not only the welfare of the poor in those societies, but also of those who are working, healthy, and reasonably well-off. How does economic growth affect such people, and is it worth the cost?

A recent comparison of countries found that SWB rose moderately but significantly with income, with a large variance, of which 37 percent was explained. But subsequent analysis of the same data indicated an inflection point at the 75th percentile: above that level, income did not provide any increment to subjective well-being.

Interestingly, a rise of income over time for particular individuals produced no improvement at all in well-being. Among the determinants of SWB, the quality of relationships, of leisure, and of work experience counted a considerably more for aggregate well-being than income and consumption measures. Materialism, a preoccupation with economic well-being, was negatively correlated with SWB, and especially so in those who believe that more money would make one happier.

The determinants of SWB have more recently been investigated on very large samples over time, in both the United States and in Europe. The findings are consistent with previous ones. They show little variation over time. Absolute income counts for little but relative income (i.e. position in the ladder of earnings) has significant influence on wellbeing, from the second quartile upwards. The positive cross-sectional correlation of income and SWB within countries has long suggested a link from static to dynamic approaches, and to the 'relative income hypothesis', which states that what counts is not absolute income, but relativities.70 From this point of view, if distribution is unchanged, then even a large rise in income will leave no impact on well-being. The reason for this, it is argued, is that as incomes increase, so do consumption norms.71 Consumers become habituated to new levels of consumption.

Longitudinal studies indicate that personality is a strong predictor of SWB. One review concludes that happiness is more a trait than a state. A study of separated identical twins suggests that neither social and economic status, educational attainment, family income, marital status or religious commitment could account for more than three percent of the variance in 'happiness'. About half the variance was associated with genetic variation. Subsequent re-testing suggested that about 80 percent of the stable element in well-being was heritable. Culture provides another element of stability. In affluent societies, culture appears to affect SWB more strongly than income. One study suggests that the hedonic ideal of distinctly individual welfare, utility or happiness might be an ethnocentric cultural construct that is peculiarly Nordic, Anglo-Saxon or Protestant. The long-term persistence of SWB scores in particular countries even brings to mind the idea of 'national character'.

The most consistent application of the relative income concept hypothesized that expectations are formed by comparisons with parents, and that demographic cycles mean that, different cohorts have different expectations.81 Satisfaction has moved pro-cyclically, with the young adults of the 'golden age' exceeding their own modest expectations, while the successor 'baby boom' having its higher expectations disappointed.

A different psychological approach is to investigate the hedonic dynamics of satisfaction, which explain why people might make choices which do not improve their welfare. This approach presents a more nuanced account of satisfaction than the crude measure of SWB. In welfare economics, it is assumed axiomatically that the consumer is well-informed, self-aware, consistent, and acquisitive. These assumptions are necessary if 'revealed preference' is to equal welfare. Other approaches do not have such confidence in the cognitive abilities of the consumer. That individual choice might not maximise welfare is an old but neglected theme in welfare discourse. Research in the hedonics of satisfaction continues. It stresses that satisfaction depends on habituation, anchoring, contrast and temporal effects; that the experience of satisfaction varies with time, and changes between decision, experience and retrospection.

The psychic payoff of rising absolute income is small, but gains in relative income are rewarding. There are high levels of satisfaction with stable attributes which cannot easily be changed, such as personality, gender, and nationality. Stability and habituation appears to promote well-being.

There is a view that ill-being does not belong on the same dimension as well-being. 93 'Prospect theory' argues that losses are more acutely experienced than gains. Unemployment and discrimination have a much more powerful effect on well-being than material gains. It may be easier to reach consensus about welfare bads than about welfare goods.

This does not mean that GNP goods have lost their relevance permanently. A society dependent on exponential growth for a stable experience of well-being might suffer badly if growth is withdrawn. Many societies have yet to arrive at that state of abundance, and can still anticipate large welfare returns to growth. And a shift away from GDP goods towards leisure or short-term gratification might eventually return us to the economy of pain.

AVNER OFFER, ECONOMIC WELFARE MEASUREMENTS AND HUMAN WELLBEING, Nuffield College working papers, Oxford University, MARCH 2000