Three points are crucial, and I would urge member states to consider them seriously.

1. We should measure progress in a way that can guide policy. This requires a single over-arching measure of how we are doing. The alternative is the easy option of saying that lots of things matter. But this does not help us resolve the policy trade-offs.
2. The right single measure of progress must be the one that is self-evidently good. The only such measure is the happiness of the population  - and the equivalent absence of misery. This is the proper measure of the quality of life. We should of course combine with it the impact of current conditions on the happiness of future generations, and any changes in the length of life.
3. For practical purposes the most practicable measure of happiness is life-satisfaction. This has the big advantage for policy-makers that it depends on satisfaction with the different domains of life (family, work, community, health, income and so on), which in turn depend on specific public policy inputs. It thus provides an indicator for which we can develop a causal model that is useful to policy-makers.

1. A SINGLE OVER-ARCHING MEASURE
Our measure of progress should be the same as the measure we use to evaluate our current policy options. For the best policy is the one that leads to the greatest progress. But, when we make policy choices, it is not enough to know how a policy affects educational  achievement, physical health, environmental quality and so on. There has to be a metric for comparing these outcomes in order to say whether one combination of outcomes is better than another.
This is essential. It is not enough to simply list the impact of a policy on each human “capability”, as Sen argues.  There must be a way of aggregating these impacts, otherwise we shall rarely be able to rank one policy over another – and there are an awful lot of plausible policies.
To assign arbitrary weights to different types of outcome (as in the Human Development Index) is not satisfactory. It will inevitably reflect the judgements and perspective of the elite who devise the weights. The only justifiable system of weights is that which reflects the experience of the population regarding how the different outcomes affect their overall well-being.

2. HAPPINESS AS THE CRITERION
So what is that measure of well-being? I think there is one simple test of what is the ultimate good for humans – that we find it self-evidently good. So we can list all the goods we consider important: freedom, health, achievement, income, happiness and so on. We can then ask of each, Why is it good? If we ask, for example, why freedom is good, people will say slavery makes people miserable. So does ill-health, and so on. But if we ask, why does it matter if people are miserable or happy, no reason can be given.  It is self-evident.

This is the philosophical reason for the long-established tradition of thought which believes that the best societies are those in which there is the most happiness and the least misery. In the eighteenth and nineteenth centuries this was a commonplace view, and it helped to usher in a more humane social order. But in the early twentieth century it took a severe blow from the growth of philosophical scepticism about whether you could know how anyone else felt. This was compounded by the behaviourist revolution in psychology which believed that all you could study was behaviour and not feelings or motives.
Fortunately psychology has changed track in the last fifty years, and returned to the view that feelings have an objective reality. For example, researchers have correlated the happiness which an individual self-reports and the corresponding estimate made by a friend. The correlation is good – which (when you think of it) is an essential condition for effective friendship and social life. But more decisive in the intellectual debate has been the discovery by neuro-psychologists of the areas of the brain where happiness and misery are experienced. Here we have good correlations (both across time and across individuals) of self-reported happiness and electronic readings in the brain. This I believe should settle the argument about whether our feeling have sufficient objective reality to be taken very seriously.

I just gave a philosophical reason for taking them seriously. This is reinforced by the corresponding empirical finding that when people are asked what they most want in life the majority reply “To be happy”.
 Before going further, let me deal with the obvious question: “Are you advocating that everybody should pursue their own selfish happiness?” The answer of course is No, for two reasons. First, if everyone is selfish, all will lose from the selfishness of others. But second the selfish individual will himself lose (even if others are unselfish). For psychology shows that, as people come to care more about the happiness of others, they themselves become happier. 
 So happiness and misery are the key indicators of the quality of life. This is really fortunate because it relieves policy-makers of making judgements about how different dimensions of experience should be weighted – we accept the internal valuations which spontaneously occur within individuals.

There are however two further items which need to be incorporated before we have a complete measure of progress.
1. Life expectancy. Clearly people value life-years as well as the quality of each life year. On certain assumptions we can multiply the two together to get a measure of Quality Adjusted Life Years.  We should do this using overall quality of life and not the ‘health-related’ quality of life studied by health economists.
2. Impact on future generations. The happiness of future generations matters as much as our own. So it should not count as progress if what we now enjoy is at the expense of the future. This issue has been discussed at length in the Stiglitz report  and I agree with their general approach in terms of stocks of assets.

3.      PRACTICALITIES: LIFE-SATISFACTION
The measure
This brings me to the practical issues. We need a measure that has already been intensively studied and whose properties are well known. For example it would help if we already had a causal model for how it is determined, and it would be particularly helpful if the determinants corresponded to different spheres of life which policy-makers recognise.
Of all the measures of subjective well-being, life-satisfaction has been the most intensively studied.  In one extremely helpful approach life-satisfaction depends on satisfaction with the different domains of life (see Figure). And domain-satisfaction depends in turn on more detailed causes including public service delivery and satisfaction with the service. It thus provides information of real use to policy-makers.

Uses
So how would a policy-maker use regularly-collected data? There are many ways.
1. He would monitor trends in the aggregate life of the community and see which areas are going better and which worse.
2. He would look hard at the distribution  of well-being, and see who is in misery and who is flourishing.
From 1 and 2 he would get ideas for policy change. There are then two ways in which he could assess these possible changes in terms of their effect on overall life-satisfaction:
3. One would be to use his causal model to form an impression of what difference an action might make – by summing the effects down the different channels indicated in the diagram.
4. A more exact estimate would come from a controlled experiment.
In every case the final outcome of course would be the life-satisfaction of the population, and the policy-maker might wish to give more weight to reductions in misery, compared with increases in happiness. That would be an ethical choice but all other steps in the analysis would be based on positive social science. It seems to me highly likely that this type of quantitative analysis will become routine practice in many governments within the next 20-30 years.

I have been involved in cost-benefit analysis using the money metric for 40 years. But its weakness becomes increasingly apparent – it inevitably ignores income distribution (treating each dollar as equivalent) and it cannot handle many of the most important influences on human happiness which are social influences not amenable to willing-to-pay analysis. The concept of QALY used in health research is much better but, as practised, it involves arbitrary weighting of different dimensions of well-being. So, once again, we come to the over-arching criterion of life-satisfaction. I believe we shall get there but only if we are single-minded and rigorous about it.

CONCLUSION
 It is wonderful that the OECD has taken this initiative to find a better concept of progress. But it would be a tragedy if current scepticism about the GDP criterion led to no substantive changes of policy, due to disagreement about a coherent alternative. The money metric has a strong but limited logic. It will only be displaced by some other metric. If what we ultimately value is the quality of human experience, that metric should be centred on happiness and misery.

REFERENCES
Diener, E., Helliwell, J. and Kahneman, D (eds) (forthcoming) International Differences in Well-Being, New York: Oxford University Press
Lyubomirsky, S. (2008) The how of happiness: A scientific approach to getting the life you want. New York: Penguin Press

Sen, A. (2009)  The Idea of Justice, Harvard University Press

Stiglitz, J., Sen, A. and Fitoussi, J. (2009) “Report of the Commission on the Measurement of Economic Performance and Social Progress”

This paper has been written within the context of the 3rd OECD World Forum on "Statistics, Knowledge and Policy. Charting Progress, Building Visions, Improving Life", Korea, October 2009.

Lord Layard is Emeritus Professor of Economics at the London School for Economics (LSE).

Lord Layard directs the Wellbeing Programme at the Centre for Economic Performance at the LSE, and is the author of the influential book Happiness.

He is currently advising the Government on mental health policy, and is the author of The Depression Report - A New Deal for Depression and Anxiety Disorders, which was published in June 2006.