« Not a Surprise | Main | Jodie Marsh »

February 22, 2007

Statistics Bleg

Just a question that's been running aoround my mind. As this EUObserver report puts it, we measure the poverty rate in the EU by measuring the population of a country against the median income of that country.

When we look at the US poverty rate (well, in one of the forms of the calculation, anyway) we look at the population of the entire place and mark it against the national median income.

But given the huge regional variations in income across the US, is this in fact comparing like with like? Shouldn't we be comparing the EU poverty rate by measuring the whole population against the EU median income?

Just wondered whether that had ever actually been done?

February 22, 2007 in Science | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference Statistics Bleg:


>But given the huge regional variations in income across the US, is this in fact comparing like with like? Shouldn't we be comparing the EU poverty rate by measuring the whole population against the EU median income?

I'd actually suggest the opposite is more valid, statistically. Compare the EU poverty rate with the US poverty rate measured within the various socioeconomic regions or proxies thereof (eg States). That is, the US rate is more inaccurate than the EU rate, more badly blurred/reduced in descriptive power by mingling separate populations, so we'd be better off improving the weaker than degrading the better.

Assuming that's a practicable option, of course...
Mate of mine's the Senior US Economist in NY for [megabank] and he ruefully admits to my criticisms of most analyses of Housing figures that he can give you screw-size by screw-size breakdowns of mid-west Fastener manufacturing volumes and sales, but he can not distinguish between bulge-bracket house sales and poverty-line house sales. The data's not collated by the Fed. Many a slip twixt cup and lip, to coin a whole new concept.

Tim adds: As far as I'm aware, the US census has pverty figures for each state: but they're against national median, not state median.

Posted by: Saltation | Feb 22, 2007 3:12:31 PM

It seems to me that "relative poverty" is either or both of: (i) a measure of one possible "justification" for envy; (ii) a measure of failure to achieve some communist nirvana.

If one is cushioned with provision, in extremis, of an adequacy (though not generosity), in absolute terms, of: food, roof, clothing, warmth, medical treatment and education of one's children, is that not enough?

Well, perhaps too, some time, space and limited financial help to recover from some personal disaster.

How far short of this actually are we in the UK?

Best regards

Posted by: Nigel Sedgwick | Feb 22, 2007 7:21:32 PM

Not to mention the regional variations in cost of living, without which I'd argue income variations aren't quite as useful.

Deliberately picking what I assume is an extreme example, but illustrates the point, compare Indianapolis, IN (pop. 780,000) to San Francisco, CA (pop. 740,000).

To maintain the standard of living that $25k/year gets you in Indy requires $62k/year in SF.

That’s of course only looking at basic costs, not total utility; I imagine many feel the weather, ocean, restaurants and nightlife of SF are superior to Indy. But if you’re looking at poverty, I don’t think those are as much of a concern.

Posted by: Bill | Feb 22, 2007 7:28:16 PM

I suppose as an addendum I should add that if income variations directly correlate to cost of living variations, then of course you’d only need the income variations. But I’m skeptical as I think many people are willing to take significant basic cost hits for the added utility of an area (culture, etc.).

Posted by: Bill | Feb 22, 2007 7:48:54 PM

Going on about "relative poverty", what about renaming (and slightly recalibrating) it as "relative wealth". Though that would be without changing the meaning from the spread of income within the group (or at least the spread on the lower side of the mean, relative to the mean).

Then it might not sound so bad (at least to many people).

But does not the fact that one can rename such a thing, and change how well it looks or sounds, without seriously changing what it means, show it in its true colours?

Best regards

Posted by: Nigel Sedgwick | Feb 22, 2007 8:01:09 PM

Last time I looked, the various nations that make up the eu are still sovereign countries (just about). So is the USA. Thus if one is comparing nation state to nation state the current practice would be valid.

And if one were comparing trading bloc to trading bloc, surely the best yardstick to measure the eu against would be NAFTA.

And as for breaking up a country into "suitable economic areas", isn't there a danger that it would be a very subjective exercise? I imagine it could be quite easy to compare say Louisiana without New Orleans to Greater London or the South East of England and come up with some very misleading answers.

Posted by: The Remittance Man | Feb 23, 2007 10:09:10 AM