14 September 2009

Beyond GDP?

Joseph Stiglitz wants to develop new metrics to assess well being, and cites the limitations of conventional metrics of GDP (pictured above in comparison to infant mortality, from Gapminder.org). He writes in the FT today:
A political leader attempting to promote the well-being of his citizens is pulled in different directions: he will be graded on economic performance but there are many other dimensions to the quality of life, including the state of the environment. While there is no single indicator that can capture something as complex as our society, the metrics commonly used, such as gross domestic product, suggest a trade-off: one can improve the environment only by sacrificing growth. But if we had a comprehensive measure of well-being, perhaps we would see this as a false choice. Such a metric might indicate an increase in wellbeing as the environment improved, even if conventionally measured output went down.

This was one of several motivations for Nicolas Sarkozy, president of France, when he established the International Commission on the Measurement of Economic Performance and Social Progress, which I chaired and for which Amartya Sen served as adviser and Professor Jean-Paul Fitoussi of the Institut d’Etudes Politiques served as co-ordinator, and whose final report is issued on Monday.

National income statistics such as GDP and gross national product were originally intended as a measure of market economic activity, including the public sector. But they have increasingly been thought of as measures of societal well-being, which they are not. Of course, good statisticians have warned against this error. Much economic activity occurs within the home – and this can contribute to individual well-being as much as, or more than, market production.

There are concerns, too, that a focus on the material aspects of GDP may be especially inappropriate as the world faces the crisis of global warming. Should we “punish” a country – in terms of our measure of performance – if it decides to take some of the fruits of the increase in productivity from the advancement of knowledge in the form of leisure, rather than just consuming more and more goods?

What we measure affects what we do. If we have the wrong metrics, we will strive for the wrong things. In the quest to increase GDP, we may end up with a society in which most citizens have become worse off.
The French commission that Prof. Siglitz is participating on can be found here, along with their interim report issued last June.

While I agree with many of the critiques of GDP asa measure of well being, at the same time it does say something about well being (see the data at gapminder.org). I am not optimistic that new metrics will ever replace GDP, however effective policy making no doubt depends upon a more comprehensive view that GDP alone affords.

From this perspective, I think that Stiglitz is misguided when he writes of GDP that "Flawed statistics may also lead us to make incorrect inferences." This is undoubtedly a true statement, however, it is not that GDP is a "flawed statistic" -- it actually measures quite well what it is supposed to measure. The issue has to do with the use of GDP in policy making as an indicator. This is a little like using student standardized test scores as the sole metric of a graduate student applicant's prospects for success in our graduate program. Sure it works in many cases, but students are more than just a single number, and maybe GPA, letters, interviews and statement of application should be considered as well. Can all this be boiled down to a new and improved index? Perhaps, but I doubt it.

Decision making will continue to require consideration of imperfect and incomplete statistical indicators. Efforts to boil down complex issues like well-being to a single indicator are likely to fail for the same reasons that GDP fails as a single indicator. For better or worse, we will have to continue to rely on multiple indicators about contested and uncertain values.


  1. And what exactly is the point of this? To what purpose will this new metric be used? There is no international GDP contest, and no World GDP Cup. Do people risk their lives walking through the desert from Mexico to the United States because they compared the GDP of Mexico and the United States? I can only imagine that the only reason for this foolishness is to bump France up the GDP list a few places.

  2. The gdp fixation is worse than even Stiglitz says. There were plenty people warning about a coming de-leveraging crash but the reply from the failed rock star economists was always that the gdp numbers were healthy so there was no reason to panic. They even manged to persuade themselves that all that debt was a good thing - a sign of confidence. They didn't quite grasp that the gdp was artificially inflated by that selfsame debt. We were just pulling forward money from the future to inflate the numbers in the present and it had to stop at some point.

    Similarly it doesn't genius to realize that creating money out of thin air to encourage spending is fueling short-term growth but long-term misery unless that money is actually invested in something that will make more money than you are spending. Military operations, volatile house prices and anything else not actually adding real value shouldn't count.

  3. Reminds me of bogus measures of "livability", etc. which seek to compare cities on the basis of an index composed of a variety of statistics ranging from the arts to pollution. Anytime you actually look at the weighting of the various factors, it becomes obvious why the authors' preferences diverge markedly from the way real people actually vote with their feet.

  4. Hi,

    GDP (per capita) also correlates well with HDI (Human Development Index, which includes life expectancy, education, literacy, etc., as well as GDP:



  5. Stiglitz could have been more clear, but he seems to believe that the flaw with GDP is not that it doesn't accurately measure what it's supposed to measure, but that what it measures is a very limited glimpse of economic activity. From the FT article:

    "Flawed statistics may also lead us to make incorrect inferences. In the years preceding the crisis, many in Europe, focusing on America’s higher rates of GDP growth, were drawn to the US model. Had they focused on metrics such as median income – providing a better picture of what is happening to most Americans – or made corrections for the increased indebtedness of households and the country as a whole, their enthusiasm might have been more muted.

    "No good accountant would ignore the depreciation of a company’s capital, but the standard GDP measure not only does that but also takes no account of resource depletion and environmental degradation. Our increased awareness of the scarcity value of environmental resources makes this lacuna especially troubling."

    His argument is that measures of economic growth ought to account for depreciation of capital and resources, as well as how broadly the growth of an economy is spread across its people. GDP doesn't do that.

  6. David,

    I think it obvious that changes in wealth are a better metric than GDP. Of course, Keynes made this very mistake of focusing on the flow rather than the stock. And that despite the "broken window fallacy" having been explained many decades previous.

    Too bad those preaching the myth of "green jobs" are still making these mistakes today.

  7. Is GDP in the above chart calculated in dollars from exchange rates or purchasing power parity?

    It matters a bit, I believe.

    Stan-the glaring light of Bastiat is anathema to the Carbon Vampires.