Thursday, November 3, 2011

Dumbest. Anti-Keynesian economics. Article. Ever!

Whoever claimed that there are 3 kinds of lies: Lies, Damn Lies, and statistics really got it.  I ran across a silly blog post over at seeking alpha.  It was old, but it was so stupid it was worth bringing up.  The writer was trying to make an argument against classic keynesian economics.  His principal argument was the government spending causes unemployment.  Since most keynesians advocate fiscal policy to maintain aggregate demand, that would be quite a blow if he could prove it.

Unfortunately for the writer, his entire premise was based on a chart he made that showed the unemployment rate and government spending as a percentage of total GDP, graphed over time.  The result was a  direct correlation between unemployment and increased government spending - as a percentage of GDP.  Both went up and down at roughly the same time.  Game-set-match.  The statistics prove it!  /snark.

A few objections come to mind.  A statistician would jump in and say that "correlation doesn't prove causation".  That is to mean, just because they happen at the same time doesn't mean that one necessarily causes the other.  However, pointing out that logical fallacy isn't even the strongest argument against this guy's point.

The second thing to disprove this non-sense was mentioned in the articles comments.  I'll call it the "aspirin" factor.
The above article is a brilliant piece of false logic. I’m sure there is an equally close relationship between headaches and aspirin consumption. Which proves that aspirin does not ameliorate head aches. In fact, if you want to be totally and completely illogical, you might deduce that aspirin actually CAUSES headaches.

In the past, governments have often responded to recessions with stimulus spending.  What that means is that the unemployment caused the spending, not the other way around.  Thanks to many "automatic stabalizers" like unemployment and such, some of this spending increases now happens without even political intervention, but it's still in response to the unemployment, not the other way around.

So we've had the "lies" and "damn lies" part.  Now time for the "statistics".  For the graphs, the person completely rigged the results by not using total spending and instead he used "spending as percentage of GDP".  Why is that important?  Well, to get the number, you take total spending and divide it by total GDP.  What that means is that if dollar spending stays the same, but GDP gets smaller, that spending number would rise even though the total dollar amount stayed the same.  Well guess what happens during a recession?  By definition, the GDP gets smaller.   Therefore, even if total dollar spending stayed flat, it would appear that government spending increased along with the oncoming recession.  The data tells a lie that you would only believe when looking at a picture graph of the data.

So there you have it.  All in one article, lies, damn lies, and statistics.
Leave a Comment