Innocent Bystanders

“New” Krugman Sinks While Clinging to Weak Study July 6, 2010

[This may be a bit dry, but I'd do anything to push that picture down the page]. King Bainan over at Hot Air had a nice post on the 2 faces of Paul Krugman: “unemployment benefits increase the duration of unemployment” Paul; and “unemployment benefits have a negligible effect on unemployment” Paul. I.e., “New” Paul vs. “Old” Paul.

Apparently Dr. Krugman was basing his thinking on a recent ‘study‘ by a couple of people from the Federal Reserve Band of San Francisco. I put ‘study’ in scare quotes, because there’s not much to their work, and what little there is is poorly done.

How can I be so blithely critical of the experts? Here’s exactly, precisely how.

The Federal Reserve guys looked at the categories of “Reason for Unemployment” that are tracked by the Bureau of Labor Standards:

• Job Losers
  • Temporary Layoff
  • Permanent Job Loss
  • Completed Temporary Job
• Job Leavers
• New Entrants
• Reentrants

They decided to compare “Job Losers,” who are eligible for unemployment insurance, to Job Leavers and New Entrants, who are not. And they came up with this chart, which shows that there’s virtually no difference in length of unemployment between the two groups.

From Valletta & Kuang, 2010

Looks like New Krugman was correct, doesn’t it? But you know, the thing about bundling similar categories together is that you need to make sure that they’re really similar. And if you’re really looking to isolate a difference between two scenarios, you ought to throw out all the subcategories that might obscure that difference.

So, for example, you should make a chart like this, which takes a look at the data from June 2010, and compares the various categories (I couldn’t get similar data for 2009 off the BLS website, dang it):

What a mess!

Oh dear. None of the categories look alike, suggesting that different influences are at work on each category. So let’s simplify it.

New entrants may be sleeping on their parents’ couches, so we’ll get rid of them, since they have a private form of unemployment insurance. People who are temporarily laid off do not get to choose when they go back to work – they’re just waiting for word from their employer. So we get rid of them, and of people who completed temporary jobs, since their lot is subject to the whims of fate.

That leaves us with only people who lost their jobs vs. people who left their jobs. That’s the purest comparison we can make to find out whether unemployment insurance makes a difference in peoples’ behaviors. Unfortunately I can’t get the average term of unemployment w/o getting more info from the Bureau of Labor Statistics, but I can simplify the chart above to directly compare the two categories.

Now we can see that New Krugman and the Federal Reserve Board are all wet – there’s a dramatic difference in unemployment duration between people who have unemployment insurance, and people who don’t. People with insurance are unemployed for much longer times than people without, as common sense would lead you to expect.