03 August 2012

Does More Income Mean More Well Being?

Jan Delhey and Christian Kroll of Jacobs University Bremen have a new paper out (here in PDF) with the Social Science Research Center Berlin (Wissenschaftszentrum Berlin für Sozialforschung) in which they ask and answer an interesting question: Do recently proposed metrics to replace GDP actually do better than GDP in predicting peoples reported levels of well being?

Their answer, with one exception, is no.

Here is what they find when they compare income (GNI in PPP) and "six different quality of life" (QOL) metrics proposed to fix/complement/replace GDP with self-reported "subjective well being" averaged for 34 OECD nations:
Do the new QOL measures perform better than the GNI? The surprising answer is no, not across the board (see Table 1). Only one single measure outperforms the GNI, namely the Better Life Index (exclusive of subjective well-being). . . average life contentment is most closely associated with the objective well-being indexes, regardless of which one we consider...

[M]ost new measures fail to deliver what they promise, if we use the average subjective well-being as expressed in representative surveys as the yardstick. Only the OECD’s Better Life Index (to repeat: exclusive of [subjective well being], of course) has an advantage over the GNI in this respect. In contrast, all the other QOL measures employed do not perform better in predicting subjective wellbeing than the GNI, and one measure—the Index of Social Progress—performs worse.
With respect to GDP, the authors find a strong relationship that does not "level off," contrary to conventional wisdom and other findings. They explain:
How well does the GDP predict the national SWB? As Figure 3 [immediately above]shows, the correlation between GNI per capita (in purchasing power parities, to adjust for different price levels) and average SWB is very strong (.58) and highly significant. By and large, people enjoy life more in richer countries. There is also no sign of a leveling of this association among the richest nations, as it is often claimed (Inglehart and Klingemann 2000). If one uses a logged scale of GNI per capita, as advocated by many economists, the regression line is even slightly concave, again suggesting that there is no decreasing marginal utility of national income.
 What do these findings mean? The authors explain:
One straightforward lesson is that economic activities and the affluence they create actually do make life worthwhile for a huge majority of people—even among the OECD countries, of which many are affluent. This suggests that what Samuelson and Nordhaus observed 60 years ago is still valid today: “People do not live by bread alone. Nor does society live by GNP alone. But on our way to that utopian state of affluence where concern about material well-being will disappear, we do need a summary measure of aggregate economic performance”...
What about the one index that closely outperformed GDP?
The measure that indeed does make a better job than the GNI of predicting national subjective well-being is the Better Life Index. This holds for all the OECD countries, and in particular for the subset of the richest/core countries. Obviously, its life domains and selection of indicators capture the life facets that OECD citizens truly value, over and above the achieved level of economic advancement. More research is needed here to find out exactly what these extra components are that give the Better Life Index added value. Given the fact that previous research has revealed that social capital (especially when compared with the GDP) matters more for the SWB of rich countries than for that of poorer ones (Kroll 2008), our informed guess is that issues of social capital, social cohesion, and greater equality make the difference here.
In short:
To summarize our argument, for an economic indicator never intended to assess national well-being, the GDP is surprisingly successful in predicting a population’s subjective well-being. At the same time, the theoretical claim of the social indicators movement about the multi-dimensionality of human concerns is a valid criticism, and conceptually it should be possible to come up with performance measures that embrace this multi-dimensionality better than an economic performance measure alone.
Writing in the FT today, Samuel Brittan explains a risk that accompanies efforts to develop alternative metrics of wellbeing:
Economists have known for a very long time that there are many components of utility not taken into full account in measures of GDP per head. It excludes leisure, the value of work undertaken in the home, environmental harm and benefits. Some have tried to construct more comprehensive human development indices, also covering matters such as literacy, access to clean water and life expectancy. My view has always been that these matters are best considered separately rather than combined in an overall measure that inevitably reflects the personal values of those who draw it up.
 The findings of Delhey and Kroll make it difficult to argue with Brittan's conclusion:
[A]t the end of the day, the government can best contribute to happiness by the indirect route of concentrating on its traditional responsibilities for public order, national security, creating the conditions for prosperity and correcting gross disparities in income and wealth.
Calls to replace GDP are common these days. Any such metric should meet the basic empirical test of doing better than GDP in its relationship with outcomes that people care about. Most proposed metrics fail this basic test.

Ultimately, GDP will all but certainly remain core to efforts to measure well being. That said, there are important dimensions that it misses. Bringing those dimensions into view will not, in my view, be accomplished by inventing a better single metric, but by realizing that no one metric captures all that matters and recognizing that understanding well being is complex, multi-dimensional and involves trade-offs that people do not agree about. Multiple metrics will thus aid in both clarity and focus, and help to avoid the risk of experts seeking to impose their values on the broader community via stealth accounting.


  1. "By this time the soma had begun to work. Eyes shone, cheeks were flushed, the inner light of universal benevolence broke out on every face in happy, friendly smiles" -- Brave New World, Aldous Huxley

    Subjective Well Being, you say? I find the concept of governments striving for such things in aggregate, or even needing to measure them, more than a little creepy.

  2. ==>> I find the concept of governments striving for such things in aggregate, or even needing to measure them, more than a little creepy. <<==

    Would that also apply to GDP? All single metrics, are obviously, insufficient to gain a comprehensive picture. The problem with GDP, IMO, is that as an aggregate, it fails to capture changes in income inequality.

  3. Well-being can be measured as inversely proportionate to an individual's propensity to escape reality through reality distorting activities and influences. It can typically be attributed to behaviors which are unproductive or pursued in excess. It is most likely suffered by individuals incapable of self-moderating behavior. This predisposition can be observed in individuals who are poor and rich alike. It is not uniquely correlated with material possessions or generally conceived wealth.

  4. -2- Joshua,

    At least GDP is, well, objective. I'm much more interested in GDP measurement than income inequality. What's the point of measuring income inequality?

  5. Is it fair to say that those who criticize GDP as a measure of well being all live in countries with high GDPs? Funny how that works.

    Please note that people do not risk their lives migrating from high GDP countries to low GDP countries.

  6. ==>> Well-being can be measured as inversely proportionate to an individual's propensity to escape reality through reality distorting activities and influences. <<==

    Interesting. Buddhist macro-economics. I like it.

  7. ==>> What's the point of measuring income inequality? <<==

    Come in to my parlor said the spider to the fly?

    I think that the significance of income inequality is an interesting discussion, but my guess is that if you ask that question in that fashion (as if it is clear that there is no point), there's no point in my trying to answer it.

    To continue the Buddhist theme:

    Is there a point to answering a question about something for which there is no point?

  8. Joshua:

    I am not proposing an economic system, but rather an objective diagnostic test to determine efficacy of a culture. Economics is a subset of this larger paradigm. What people will do with that knowledge is mainly of individual concern.

    My predisposition is to refrain from coercing individual behavioral adjustments, or inflicting punitive damage, unless there is a well-established, overriding concern to society. My outlook on life/reality demands that behaviors are classified in three classes, namely by their suitability for normalization, tolerance, or rejection. There are few behaviors which are suitable for the first class and similarly few that merit the last. Most behaviors fall into the second class.

    My desire would be to use this or a diagnostic test to provide feedback to individuals in order to enable and promote self-moderating behaviors. This capability is, after all, the prerequisite for liberty, and is of the utmost importance to the individual, and achieving an optimal balance is critical to the viability of a society.

  9. Interesting article but to say that these findings make it difficult to argue with Brittan's conclusion seems somewhat overstretched. What the role of government is in improving people's happiness is a very good question. What metrics we should use is another one but to jump from saying:

    1) GDP is well correlated to well-being, often better than other metrics + it may not be a good idea/possible to develop one meta-metric, many metrics would be better.
    2) the government should focus on its traditional roles if it wants to improve well-being.

    all 1) supports is that improving GDP should probably be an objective of a government interested in improving people's well-being, it certainly doesn't say anything about correcting gross disparities of wealth or security.

    And even more importantly, the fact that we shouldn't seek a meta-metric or that GDP is strongly correlated to well-being clearly doesn't entail, of itself, that a well-being promoting government should focus solely on GDP, the article points to social capital (the Spirit Level to equality). Yet is improving social capital really a 'traditional' role of government?

    Besides, this ignores the broader question about what the role of government should be. Obviously, that is a thorny issue but it seems a little odd to not address it more explicitly as the very idea that the government's role is to increase people's well-being is a pretty controversial one..

  10. If their evidence for "With respect to GDP, the authors find a strong relationship that does not "level off," contrary to conventional wisdom and other findings" is just based on that one regression they show, that's a poor piece of evidence and demonstrates an over-reliance on statistical methods without an appreciation for how they work or what they mean. The regression is highly influenced by three outliers (USA, Luxembourg and one other), without which the regression would appear quite different. I would also hesitate to call an r-squared value of .58 "very strong." It is suggestive, definitely not weak, but also definitely not "very strong."
    As well, I wonder what set of countries they are sampling, since that is clearly not all of them.
    I am unconvinced.

    Of course, maybe if I read the article it they would have a more convincing piece of evidence. But Pielke clearly couldn't identify one.

  11. Is it fair to say that those who criticize GDP as a measure of well being all live in countries with high GDPs? Funny how that works.

    or as Ogden Nash put it:

    The only incurable troubles of the rich are the troubles that money can't cure,
    Which is a kind of trouble that is even more troublesome if you are poor.
    Certainly there are lots of things in life that money won't buy, but it's very funny --
    Have you ever tried to buy them without money?

  12. The point of measuring income inequality is so you can see what's coming down the road. Income inequality correlates with a lot of negative social outcomes in weathly democracies (http://www.youtube.com/watch?v=cZ7LzE3u7Bw).

  13. The problem with GDP, IMO, is that as an aggregate, it fails to capture changes in income inequality.

    In a democracy the amount of income inequality is chosen by the people. In the end the US doesn't have more than Sweden because the government wants it that way, but because the people do.

    When a government aims to increase income equality against the will of the people, the result is decreased wealth and eventually decreased happiness. It's been tried, both by autocratic soviets and bits of democratic Europe, and it doesn't work very well.

    Governments should strive to increase wealth and democracy. The people should then chose what to do with their wealth and freedom. If they chose to redistribute it, then so be it. But aiming for that takes the eye off the actual prize for governments, which is a healthy economy in the first place.

  14. Strikes Eli that median family income would do a better job than GDP per person.

  15. Median family income is a terrible statistic to use.

    Apart from "family" being almost impossible to define, is leaves out huges chunks of the economy that are not personal income.

    I live in a country that has comprehensive medical and accident cover. That vastly improves my real wealth, without ever coming near my income.

    Medians also don't cover the bottom 25% who are the "at risk". It's not a lot better than mean in that regard.

  16. I don't find the relationship between GDP and subjective well being to be particularly convincing, regardless of the R^2 value. With only three high-GDP countries (Norway, Luxembourg, U.S.), it's impossible to know if the trend is a fluke or not. What would happen if we included the non-OECD, high-GDP countries like Qatar, Singapore, Brunei, and UAE?

    I also doubt that subjective well-being is particularly indicative of anything. Does anyone think it's possible for Mexicans to be better off (happier?) than people in Belgium, Finland, Austria, France, U.S., and most other countries? Silly.

    I would expect that the best (objective) measure of well being might be net immigration. If a lot more are trying to get in than out, it's probably a pretty good place to be.

  17. Well, actually median family income does a better job than per capita GDP just using a quick and dirty and btw, if you read the paper, they are not using r^2 but r, and median family income does a hell of a lot better than GDP per capita at capturing the lowest 25%'s plight..

  18. I think that another factor missing from this debate is that (although it sounds cynical) economic growth is not only about "creating good" but "preventing bads." If we look back at human history it is punctuated by a lot of horror and misery with the occasional great work of art.

    Human societies are chaotic systems and in economic growth we've found an instrument that mostly seems to reduce the horrible things humans are capable of. So while you can argue GDP is not the best measure of the good you also have to take into account how good it is at preventing the bad (and don't even get me started on the "ugly")

    Conversely, it's interesting to think about whether an alternative measure or policy goal (like subjective happiness) would be as effective at prevent such things.