What is the “good life”? Can we measure it?


This page in:

Material consumption is the basis of traditional welfare measurement in economic practice. For example

-          In standard economic models, an individual’s investment choices are often determined with respect to the discounted streams of expected lifetime consumption under different scenarios.

-          Mainstream poverty analysis takes consumption as the operational measure of the “welfare space” and stipulates a consumption threshold below which an individual or household is deemed poor.

While the consumption approach to welfare has much merit, I think every researcher recognizes certain limitations. There exists a long tradition of discontent with consumption as the paradigmatic welfare measure, perhaps most eloquently expressed by Amartya Sen and Martha Nussbaum in numerous works that develop the “capabilities” approach to welfare conceptualization.

Recognizing the theoretical limitation of consumption (or income) based welfare measures, and proposing alternatives, is only part of the challenge for welfare economists and policy analysts. The equal challenge is how the researcher successfully operationalizes any alternative conception of welfare.

Numerous contrasting approaches to alternative welfare measures are grouped under the general heading of subjective well-being (SWB). As one example perhaps closest related to traditional economic welfare measures, subjective beliefs of consumption adequacy seek an individual’s perceptions of sufficiency either in total consumption or constituent components such as food or housing. Using this approach Pradhan and Ravallion have constructed subjective poverty lines and poverty measures and contrasted them with the traditional “objective” poverty measures.

A more prevalent SWB notion centers around the concept of “happiness”, which may seem a natural choice for economic analysis as notions of “happiness” are congruent with the utilitarianism of Jeremy Bentham, one founding figure of modern economics. The paradigmatic happiness question takes the form: “All things considered, how would you say things are these days – would you say that you are very happy, pretty happy, or not too happy?”

Richard Easterlin analyzed variants of this question to find that individual happiness does not increase as the average income in society rises, perhaps due to hedonic adaptation. This is the famous Easterlin paradox that recognizes, while at one moment in time happiness is positively related to income, happiness does not increase as a country’s economy grows over time.

More recently, the happiness approach to SWB has been superseded by “life satisfaction”. The life satisfaction question, as exemplified in the World Values Survey variant asks: “All things considered, how satisfied are you with your life as a whole these days?” The respondent is then asked to choose from a scale of 1 to 10.

The contrast between “happiness” and “life satisfaction” is that the former is an emotion (an affect) while the latter is a reflective assessment.  In principle these two measures can reflect entirely differently domains of welfare – it is certainly possible at a given point in time to be happy yet unsatisfied with life, or vice-versa. Nevertheless life satisfaction and happiness measures do tend to co-vary. For example soon after the end of apartheid in South Africa in 1994 both happiness and life satisfaction were far higher among the black population than in 1988 during the apartheid era. However by 1995 both measures had fallen back to their 1988 levels.

The Easterlin paradox has also been observed with life-satisfaction measures. Certainly the observation that economic growth is not related to changes in life satisfaction or happiness presents challenges to proponents of economic development if one takes these measures to meaningfully represent fundamental notions of welfare. (Recently much scholarly debate has centered on whether the Easterlin paradox does indeed exist in the world. See Easterlin and others for a defense of the original empirical findings and Sacks, Stevenson, and Wolfers for a challenge to the paradox.)

One obvious objection to all of these SWB measures involves inter-subjective comparison. Of course the subjective understanding of the SWB question, both with respect to the notion of satisfaction and the application of an unanchored scale, is difficult to standardize across individuals and, especially, populations with different cultural orientations.

However my discomfort with “life satisfaction” measures and their ilk is due to a somewhat different feature– they are broad summary measures encompassing potentially contradictory aspects of lived existence. How might you answer “how satisfied am I with life as a whole” if you are quite happy with your professional life but dissatisfied with your family life? I suppose you will weigh these two domains against each other. The relative weights people place on these domains surely varies across individuals and across the life-cycle within an individual. Given these difficulties, does it make sense for public policy to posit improvements in “life satisfaction” as a goal?

Instead, let me suggest a thought experiment –quietly reflect for a full 10 minutes on your conception of a “good-life” and the fundamental properties that a “good life” need contain. If I meditate on what I fundamentally value in a “good life”, I arrive at three conceptually distinct domains that I now list as:

·         Autonomy, in the sense of the ability to influence the course of my life both in the short run (how will I allocate my time this week, what will I eat for dinner) and the longer run (where will I live, how will I earn my livelihood).

·         Relatedness, in the sense of the maintenance and growth of sustaining interpersonal relations and the feeling of emotional connectedness to friends, family, and community.

·         Security, in the sense of freedom from (or insurance against) risks to physical and emotional integrity such as famine, ill-health, environmental catastrophe, divorce, violence, etc.

In regards to my list, the positive relation between economic development and welfare may be fairly clear with respect to many aspects of the security domain – much of human advancement can be seen as directed towards improving our fundamental security in the physical world (and, yes, other aspects of my security domain can be worsened by economic development). However there is no clear relation between development and either autonomy or relatedness. Autonomy is a feeling of control over life, which we hope is related to development, but not necessarily so. Relatedness can either be beneficially or detrimentally affected by development processes, at least in the short run. In the longer run, forms of human relatedness may adapt to any economic situation.

These three domains are simply the musings of one applied micro-economist, but I have a feeling that if you undertake a similar exercise you will also delineate conceptually distinct domains. Indeed fairly recently I have come upon self-determination theory, one school of thought in psychology, and their stipulated core psychological needs of competence, relatedness, and autonomy. (Hmm, a little familiar.)

While self-determination theory argues that these core psychological needs are universal (yet can be differentially expressed across cultures) the challenge remains as to how these constructs might be meaningfully measured and analyzed. The founders of self-determination theory do propose some survey examples but no example seeks to aggregate the three core needs into one measure. Policy choices may differentially affect these core psychological needs and, if so, combining the three needs into a summary measure will obscure rather than inform accurate measurement of SWB.

My gut feeling is that any unidimensional measure of SWB will be inadequate to summarily capture the overlapping domains of lived experience both that we collectively value and that can be affected by policy choices. Nevertheless the day where we achieve consensus over a suitably flexible and operationally tractable conception of the “good life” is surely far off. Until then, let’s proceed with caution when using the existing SWB measures of happiness and life satisfaction.

Additional note: The Oxford Poverty and Human Development Initiative advocates the broadening of welfare measurement and devotes a lot of effort to operationalizing various measures. They have a very informative website.


Jed Friedman

Senior Economist, Development Research Group, World Bank

Join the Conversation

August 03, 2011

Martin Seligman on positive psychology has a nice typology, this blog agglomerated the "happiness" talks.


August 03, 2011

This is an interesting angle to look at:


Thomas de Hoop
August 04, 2011

I would agree with your discomfort with life-satisfaction measures when people would argue that life-satisfaction should be the sole purpose of public policy. However, I would say that life-satisfaction measures can certainly help policy makers in the identification of the possible trade-offs between autonomy, relatedness, and security.

Me and my colleagues wrote a paper that is concerned with the trade-off between autonomy and relatedness (at least that's how I would call it in the terminology you use) when autonomy is associated with the violation of gender norms.

You can find it here:


Jed Friedman
August 04, 2011

for the very interesting links, thoughts, and papers!

Arianna Legovini
August 08, 2011