GDP and well-being

One irritating feature of the subjective well-being research is its preoccupation with GDP, or rather its preoccupation with what it thinks is our preoccupation with measuring production. It’s there again in the new economic foundation‘s recent National Accounts of Well-Being: Bringing real wealth onto the balance sheet.

The proposition that GDP cannot be used as an all-purpose proxy for societal well-being is a banality, not a critique. Nobody who believes that it is such a proxy is cited in this publication, and for good reason: nobody of any intellectual or political signficance (and quite possibly nobody at all) believes this to be the case. It is one of a vast number of statistical series collected and used in policy. Government policy is more likely to reduce than increase GDP through regulation and taxation steering resources away from their most productive uses.

The main point of National Accounts is to call for more regular collection of statistics on well-being, broadly defined. They’ve made a start with (in conjuction with Cambridge University) an interesting survey taking a multi-faceted look at well-being around Europe.

But it is not clear why an attack on GDP needs to be part of this argument. On their own account (p.40), it correlates reasonably well with observed differences in well-being between countries. And GDP growth is critical to keeping down one of the main causes of negative well-being, unemployment. This year is going to be a brutal reminder that while more money doesn’t necessarily make people much happier, involuntary loss of an income almost invariably makes them much less happy than they were before.

While I would like to see the nef survey results continue to be published, I very much doubt that they will be, as nef claims, useful to policymakers.

As the report notes, for life satisfaction there is significant stability over time – indeed, this is probably the most stable set of numbers in the history of social science. That self-reported well-being has fluctuated in a narrow range without clear trend for 60 years, despite massive changes in social circumstances over that same time period, should be a warning that what policymakers do really doesn’t have a huge impact on this measure.

Indeed, the familiar argument that higher GDP has failed to increase average happiness can be turned against every other trend and change: the welfare state has not made us happier, new technology has not made us happier, feminism has not made us happier (it hasn’t even made women happier), more education has not made us happier, the sexual revolution has not made us happier. Et cetera, et cetera.

Once various basics of prosperity and social order are in place, well-being levels are primarily due to personal characteristics and behaviours that are very hard for policymakers to get at such as personality, socialising, friendship, and love. No amount of policy effort is likely to make more than the most marginal difference to these things. Frustrating as it must be to believers in big government, some of the most important things in life we just have to do ourselves.

22 Responses to “GDP and well-being

  • 1
    Brendan
    January 28th, 2009 22:19

    Part of the problem with using GDP as proxy for well-being, that is often overlooked by macroeconomists, is that it includes government purchases at cost. If the government hires people to produce stuff that is worthless, that stuff is included in GDP just as much as if the government buys something valuable that people want. When calculating GDP, the national income accountants do not pass judgment on the social utility of government spending.

  • 2
    entropy
    January 28th, 2009 22:35

    Let’s face it, attacking GDP is really a proxy for a pre-emptive attack on any economists that might call them on the proposed solution to their non existent problem: some sort of government regulation to make people happier in the approved fashion.

    Brendan, public sector production of something worthless will still lower GDP as the return on the investment will be negative. Government, or anyone else for that matter, only contributes to GDP when the value added is greater than the investment. As a public sector employee myself I am having trouble thinking of a case where that would be so, except in my office of course 🙂

  • 3
    entropy
    January 28th, 2009 22:38

    I should clarify that the investment by the government in the worthless activity would not be available for investment in an activity that is not worthless and would in fact add value, consequently ensuring that the GDP is lower than it would be otherwise.

  • 4
    Robert Wiblin
    January 28th, 2009 22:47

    What did you make of this critique of the Easterlin paradox that generated some buzz last year:

    http://freakonomics.blogs.nytimes.com/2008/04/16/the-economics-of-happiness-part-1-reassessing-the-easterlin-paradox/

    http://www.nytimes.com/2008/04/16/business/16leonhardt.html?_r=2&oref=slogin

  • 5
    Dave Bath
    January 28th, 2009 23:17

    Andrew said: “In their own account (p.40), it correlates reasonably well with observed differences in well-being between countries.”
    Maybe. GDP per capita is probably better.

    One of the problems with “happiness” is that it is subjective… wealthier US citizens were less happy than many in poorer South American countries for example.

    So, what about objective metrics that DIRECTLY measure well-being? Who should define them? How about the indicators used by the UN, such as the HDI (Human Development Index), found (with lots of stats nearby to keep Andrew more than happy).

    The assumption is that literacy, lifespan, infant mortality DIRECTLY measure the possibilities of well-being if everyone had the same mindset.

    There is a good interactive comparison tool that compares HDI to per cap GDP “here. Unfortunately, you cannot compare two countries of your choice – they pick a nearby country

    While cuba isn’t on the list, I did some calculations a couple of years ago, using the Living Planet Report from the WWF and the CIA World Factbook, doing quick multiply-throughs of GDP per cap at PPP and HDI. Roughly, Castro was 10 times more efficient at turning a dollar of PPP GDP per cap into HDI. Indeed, Castro was the ONLY country in 2006 to reach UN development targets for HDI and sustainability, and infant mortality/vaccination rates were better than the US, but not as good as Canada.

    Meanwhile, NZ is about 20% more efficient than we are at turning PPP GDP per cap into well-being outcomes.

    I’ll admit it ain’t perfect, but if you want to measure the effectiveness of a system, I think looking at objective human outcomes that best reflect well-being is the way to go. Using GDP without factoring in HDI-style metrics of human welfare is like guessing your distance travelled in a car from fuel consumption, or using shoe size as a proxy for height.

    And anyway, you can always get (unsustainable) luxury by borrowing, so change_in(GDP + CurrentAccountBalance)/per cap is I think better if you only want to use indicators from the back pages of The Economist.

    BTW: What would the GDP be for a country we NOBODY was in formal employment but everybody had an adequately sized veggie patch, a cow and some chooks?

  • 6
    M J Warby
    January 29th, 2009 02:00

    Entropy,
    GDP = private consumption + government consumption + investment + exports – imports.

    I do not think recurrent expenditure would cause negative return on investment, given it is not actually investment. So Brendan’s point would hold for government consumption expenditure.

    Your point on motivation is, however, very plausible.

  • 7
    Andrew Norton
    January 29th, 2009 07:39

    “What would the GDP be for a country we NOBODY was in formal employment but everybody had an adequately sized veggie patch, a cow and some chooks?”

    I think this is one reason why many poor countries (while undoubtedly very poor) are not as poor as the $1 or $2 a day figures suggest. They have home production which is not counted in GDP.

    And some of the increase in GDP in the West over the last few decades over-states the real increase in production, as goods and services are switched from the household to the market sector as women go out to work and then pay other people to do child minding, cleaning, food preparation etc.

  • 8
    conrad
    January 29th, 2009 08:07

    “While cuba isn’t on the list”
    .
    I don’t know why people feel obliged to use Cuba as an example — surely given the number of Cubans that live in the US (let alone the ones that want to live in the US) it offers evidence against all of these measures being meaningful. If you really love left wing politics, then I’m sure places like Sweden are better examples than Cuba. Of course if the measures of importance are literacy, infant mortality and life-span, then Singapore, Japan, and Hong Kong are ahead of Sweden (as is Australia), so there goes your socialist hypothesis.

  • 9
    Andrew Norton
    January 29th, 2009 09:33

    “One of the problems with “happiness” is that it is subjective…”

    No, that’s the whole point of it – it is something you feel, and this research clearly shows that what we think should be ‘objectively’ good, does not reliably translate into subjective feelings. This makes things very difficult for governments, whose main levers of laws and dollars can only easily get at ‘objective’ factors.

  • 10
    Russell
    January 29th, 2009 12:11

    I guess it’s the heat making you so touchy, but the reason some of us are annoyed about the GDP thing is that people like you start off sentences with “And GDP growth is critical to ….”

    The Robert Kennedy quote shows that GDP can be an irrelevant measure … what good will GDP growth be as the planet fries from global warming?

  • 11
    Andrew Norton
    January 29th, 2009 12:36

    Russell – My point precisely. GDP is irrelevant to measuring or assessing many things, which (unsurprisingly) is why nobody uses it for those things. I do not say ‘GDP is low today’ when I mean ‘the weather is appalling today’. However, it is equally silly not to use GDP where it is relevant, as it clearly is to employment.

  • 12
    Russell
    January 29th, 2009 12:48

    But wouldn’t it be better to talk precisely about employment, if you’re talking about employment: full-time/part-time, casual/permanent, skilled/unskilled etc. Bringing in GDP just complicates the point with a lot of more complex factors.

  • 13
    derrida derider
    January 29th, 2009 12:51

    Yes, when the question is put explicitly all economists and policymakers will agree that GDP is not a measure of human welfare. That doesn’t stop most of them from having a habit of mind that implicitly assumes it is.

    GDP leaves out all non-market production. GDP ignores sustainability of production. Aggregate GDP ignores how GDP is distributed among people. None of these are problems if we use GDP for its intended purpose – valuing aggregate market production in a given period – but when we use it more widely problems arise.

    So we get policy that replaces non-market with market production rather than boosting total production. That does not value finite natural resources properly for policymaking. And that ignores who wins and who loses by the policy.

  • 14
    Rajat Sood
    January 29th, 2009 14:31

    Many of the comments above are besides the point. The barrier to policy-makers implementing more aggressive climate change/’sustainability’ and redistributional policies is not an obsession with GDP, it is an obsession with votes. Voters care about jobs and incomes, for better or worse. Do you think Rudd would change his economic stimulus policies if he were told or reminded that GDP is not a perfect measure of well-being?

  • 15
    Brendan
    January 29th, 2009 18:22

    Entropy, suppose the government borrows some money and:

    a) uses the money to give me a lump-sum payment (such as a tax rebate) and I choose to spend my free time sitting at home reading Andrew Norton’s blog

    b) uses the money to hire me to sit at home and read Andrew Norton’s blog.

    Now, (a) and (b) are identical in terms of final allocations and social welfare. I am doing the same thing and the money flows are the same, but the figures in the national accounts are different (transfer payments are not included in government spending).

    In (b), I am employed producing a government service and so in the macro statistics this will show more hours worked, and a higher GDP.

  • 16
    2dogs
    January 29th, 2009 18:51

    Frey & Stutzer 2002 have an excellent breakdown of the components of human happiness, using a non-prescriptive “life satisfaction” measurement.

    I am very suspicious of the measurements being used by National Accounts; their “indicators” might measure the authors’ happiness, but I am not sure they measure mine. There appears to be an agenda in the selection of them.

  • 17
    Andrew Norton
    January 29th, 2009 19:45

    Robert – The Stevenson and Wolfers research was certainly interesting, especially in their work checking the Japanese data and discovering a question change which undermined the thesis that Japanese prosperity had not increased Japanese subjective well-being.

    But the more important point in my mind is that while they may have found stronger statistical relationships between GDP and happiness than previous researchers, it remains the case that today’s numbers don’t look so different from those recorded in the 1940s. For example, a 1948 Morgan poll in Australia found 36% of respondents saying they were very happy and 60% saying they were quite happy. The 2007 Australian Survey of Social Attitudes found 32% very happy and 58% fairly happy. It’s possible that our aspirations for happiness are higher now, but these numbers are pretty similar.

    I think there is a pretty powerful explanation which is a version of the adaptation thesis. This is that most of us have a personality-driven happiness ‘setpoint’; and while events may push us up or down we tend to adapt back to the setpoint over time. The best research on this, a German panel study, found that this was true of about 80% of people.

  • 18
    Brendan
    January 29th, 2009 20:56

    Some recently updated papers from a conference on the measuring of subjective well-being.

  • 19
    Andrew Norton
    January 29th, 2009 21:03

    Brendan – I randomly clicked on the Kahneman paper, and found this quote:

    “The problem is not so much with the National Accounts themselves as with the fact that policy makers and the public often lose sight of their limitations, or misinterpret national income as the sole object of policy and primary measure of well-being.4”

    A footnote! But all it actually contains is a quote from Obama agreeing with the Kennedy quote about the limits of GDP. So it is evidence against rather than for the proposition being made.

    These guys can’t help themselves. Burning this straw man is such a ritual of the literature that it just has to be done.

  • 20
    Jeremy Bray
    January 30th, 2009 11:48

    ” … while more money doesn’t necessarily make people much happier, involuntary loss of an income almost invariably makes them much less happy than they were before.”

    This is, I believe, the beginning and the end of the argument, and a good part of the reason why our societies pay so much attention to the measurement of economic progress.

  • 21
    TerjeP (say tay-a)
    January 31st, 2009 09:18

    I recently saw a picture of kids in the ruins of Gaza playing on a make shift see-saw beside the ruins of their home and smiling from ear to ear. As such I think a see-saw outside every home is the secret to happiness. 😉

    GDP isn’t a balance sheet. It is a measure of flow. The federal budget is also a flow rather than a balance sheet. So often the success of government is measured by the amount they spend rather than what we have. Or at least that seems to be the mindset when they brag endlessly about how many billions of our dollars are being spent on a heathcare system that still does not work. Why people see merit in a more expensive government that still doesn’t deliver is a bit beyond me.

  • 22
    Robert Wiblin
    February 2nd, 2009 01:51

    I guess then we should be making a list of things that cause people to not average around their happiness setpoint, or that causes their setpoint to change substantially. Then we should try to affect these things and ignore everything else.

    What sort of things would appear on the list? Do you have data on the effect on relationships (outside marriages, which always seems to be top on the list of this happiness research)?

    I would guess as a start: debilitating disease and mental illness, serious physical pain, bad marriages and disfunctional families, bullying at school and workplace, chronic loneliness, drug abuse, no sun.