Friday, June 18, 2010

So stats aren't his strong point

If I showed someone a graph which had a direct correlation between two sets of figures one can say that one has an effect on the other. Economists calculate this using the formula r squared, or the correlation coefficient.

r squared is between 0 and 1: at 0 there is no correlation. 1 is where the two are entirely dependent on each other. I used to have a great time during my degree collecting all kinds of data and calculating if they had anything to do with each other. And it also enabled me to occasionally throw into conversation the word heteroskedasticity.

This EU summit has, alas, proven that our new Prime Minister isn't up there with even basic stats, let alone yummy econometrics.

Mr Cameron said a larger EU had diluted the tendency for further bureaucratic integration. He said: “Now Europe is so much wider and broader, with the countries of eastern and central Europe as members, that will help push us in a more intergovernmental direction, which I support.
“The wider and broader Europe takes some of the pressure off further integration. But you also have to be on your guard.”

This was the excuse used for Greek entry to the EU in the 1980s and actually was an A Level economics question (where you don't tend to use any stats apart from the terms of trade).

It's quite clearly a nonsense which doesn't require any complex analysis but a basic look at the different treaties over the past 50 years.

When the A8 countries joined in 2004 the EU didn't have as many powers as it does now. One of the criteria for them joining in their accession treaties were that they signed up to the EU Constitution, now in law as the Lisbon Treaty.]

Before Greece joined, Cyprus joined, before Finland and the UK there weren't the powers in the original documents as there are now. I know that the Tories use the reason for supporting expansion as saying that the pond will become wider and less deep but it's a nonsense. It's the reason they're using for their support of Turkey joining and countries like Serbia, but it's a fucking nonsense.

What I'm trying to establish in my own mind is that are the Tories, and Cameron in particular, being mendacious in their statements or are they being ignorant?

Answers on a shoebox, please.


Jackart said...

Rule: Correlation does NOT imply causation.

The two effects could both be affected by the same factor independently of each other.

Or it could be coincidence... and so on.

Knowing stats is only part of understanding them.

Trixy said...

no, it doesn't which is why we end up doing lots of tests and we analyse.

One could get perfect correlation between phases of the moon and the purchase of custard but we'd hope that the person doing the test was either

a) bored
b) trying to show that one had to look at the stats one was regressing and not just relying on one number.

The fact is that there is a direct correlation and so saying that increased membership of the EU causes less power grab and central control is not true. it's just not fucking accurate.

McGonagall said...

I don't believe we need stats to understand we're being shafted. EU expansion, integration, and the democratic deficit are not necessarily dependent on each other. They happen regardless.

View from the Solent said...

Trixy, in your qestion why do you assume that the two possibilities are mutually exclusive?

Mark Wadsworth said...

Re what J says, logic usually tells us whether it is causation (and what causes what) or correlation (in which case, there must be a third factor that causes both) or coincidence.

It is quite clear that increasing the amount of EU regulations cannot be the cause of EU enlargement. what attracts new entrants is all the bribes they can get for the first decade or so.

So either EU enlargement causes more regulation (1), or they both stem from a third factor (2) or it's coincidence (3).

1) It is easily observable that the larger the state or organisation, the more tiers of bureaucracy there are. This is a diseconomy of scale, which is why empires fall apart and businesses only ever reach a certain size. So possibly, EU enlargement causes more regulations.

2) The third factor (which underlies the first two) is that the EU love imposing rules on people (at the behest of various lobbyists); the more rules that they can impose on more people, the better.

3) It's not coincidence, is it?