Thursday, July 13, 2006

Divide and Clunker

Will Wilkinson gives a well-deserved thrashing to a truly stupid "study" by the New Economics Foundation purporting to show that Vanuatu is the "happiest place on earth."

Here's the short version: They took a measure of each nation's life satisfaction (as measured by happiness surveys and life expectancy) and divided it by a measure of that nation's resource usage. They called the resulting quotient the nation's happiness index. WTF?

An analogy. Suppose you want to rank a group of students from smartest to stupidest. So you need an index of smartness. So you take each student's IQ and divide it by... oh, what the hell, divide by parents' income. Totally arbitrary, right? At best, if income is just unrelated to intelligence, dividing by income adds irrelevant noise to your measurement. And if income is actually correlated with intelligence (say, because smarter people are both more likely to have high incomes and more likely to have smart kids), then you'd be making the smart students look stupider and the stupid students look smarter.

That, in essence, is what the New Economics Foundation did here. There is certainly a positive (though not perfect) correlation between life satisfaction and consumption of resources, so dividing by resource usage is a good way to produce results that bear an inverse relation to, um, the truth.


Joel Bernstein said...

I think the *idea* was to get an idea of which country had the highest amount of happiness per unit of resources used, because those people would be more "efficient" with their resources in generating happiness. Kind of like how I care more about the miles per gallon my car can get, more than the total number of miles it can drive on one tank.

Of course, it's still a load of hooey, but I can at least trace their faulty logic back to the source.

JB said...

You can find the full report here:

Some highlights from the top 20 "happiest" countries:

2. Colombia
6. Cuba
7. Honduras
8. Guatemala
12. Vietnam
15. Sri Lanka

Blar said...

I think that Will's criticism applies to the news write-up rather than the NEF publication itself. I skimmed the paper (at JB's link) and didn't see them ever refer to Vanuatu as the happiest place on Earth, or their Happy Planet Index (HPI) as a measure of a country's happiness. In fact, they say the opposite: "It is important to recognise from the outset that the HPI is not an indicator of the happiest country on the planet, or the best place to live" (p. 8). In other words, they probably aren't pleased by the news article either. Here's a central part of their justification of the HPI, also from p. 8. Compared to previous indices,

>>[t]he HPI takes a radically different approach to defining progress. With well-being as the ultimate end, and planetary resource consumption as the fundamental input, we can restate the goal of development as delivering high levels of well-being within the constraints of equitable and responsible resource consumption. The HPI reflects the extent to which countries succeed in achieving this goal.<<

They then define the HPI as (Life Satisfaction x Life Expectancy)/(Ecological Footprint), and go on to say:

>>The HPI is a measure of the ecological efficiency of delivering human well-being. It reflects the average years of happy life produced by a given society, nation or group of nations, per unit of planetary resources consumed. Put another way, it represents the efficiency with which countries convert the earth's finite resources into well-being experienced by their citizens.<<

So, unlike dividing IQ by parents' income, their index does have a justification behind it. It is an ideologically-laden justification, but they are open about that, and they defend it and make an argument for why it has advantages over other indicies.

Will Wilkinson said...

Re Blar's comment. Suppose, as I think is pretty clearly true, that well-being doesn't increase linearly with "ecological footprint." To grossly oversimplify for illustration, suppose "well-being" can be put on a scale from 1 to 10, and the default for people who aren't starving or at war is 6. And then suppose that an increase of an increment of well-being requires a doubling of "ecological footprint" from the last increment. To get to 9, the ecological footprint will have to be eight times greater than at the default. It will then fall out that a society at the default of peaceful sufficiency will be more "ecologically efficient" in producing well-being. So we get a place like Vanuatu at the top of the list.

The question then is, Why care at all about ecological efficiency, if more of it comes at a cost of lower well-being? The point of the NEF study, I take it, is that the level of "footprint" required for higher levels of well-being is "unsustainable," so will, over the long run, can't last. This is, I think, pretty clearly false. The "biocapacity" stuff is specious for precisely the reasons Julian Simon laid out so lucidly. If somebody came up with cold fusion tomorrow, the whole calculus would change. Etc.

Glen Whitman said...

Will -- exactly right; I make essentially the same point in the follow-up post (see links below). If there are diminishing marginal returns from resource usage, then average return per unit necessarily declines as the total return goes up.

Will Wilkinson said...

Glen, Sweet. Well said.