Atlas Shrugged: The Mocking

Wednesday, March 9, 2011

Megan McArdle's Greatest Hits

I'll be out of the office and traveling today (which in my case means running errands within a two mile raidius), so here are a few posts explaining just why any article Megan McArdle writes about bankruptcy studies is probably full of errors, lies and misdirection.

Tom Levenson
R. J. Erskow
Mike Konczal

Bonus comment from McArdle's post:

A) It matters whether the arguments in favor of policies are true, even if the policies themselves are good.

B) I happen to care about bankruptcy statistics.

C) I didn't mount a case against Obamacare in this post. Had you read it more closely, you would have seen that I am now arguing against a study which theoretically supports "my side"--i.e., the study purports to have found that RomneyCare did not cause medical bankruptcies to fall substantially, even though this was one of the major selling points for ObamaCare.

By the way, this mess all started when McArdle stupidly/dishonestly said that Warren was being dishonest because she said the number of medical bankruptcies was rising, when Warren actually said the percentage of medical bankruptices was rising.

rortybomb 1 year ago
I don't understand this. Comparing total 2001 and 2007 numbers aren't an apples to apples comparison; the law has been radically changed to a higher threshold, so we expect 2007 to be lower. If it wasn't then something would be up.

Also, 2001 was the busting of the tech bubble and 9/11; it was a terrible year for the economy. 2007 was at the height of the boom. It also suffered was an overhang, where many people strategically declared bankruptcy in 2005 (600K in October alone) to avoid the new law, so many people who would have limped to 2007 called it quits there. Comparing the peak to the valley and only looking at total numbers seems like a sneak.

The relevant question for total would be are bankruptcies rising from 2006-2008, and they clearly are. You should revisit this in 6 months. My projections are at least 1.5m for 2009, beating 2001 even with the new law in effect.

rortybomb 1 year ago
"The relevant question for total would be are bankruptcies rising from 2006-2008, and they clearly are." And she brings this up, in the passage you quote.

Megan McArdle 1 year ago in reply to rortybomb
No, it wouldn't. The relevant question is, are bankruptcies being driven higher by medical bills, as their paper very clearly implies. The relevant answer is, they are lower, not higher. A higher proportion of medical bills may or may not be medically related; their data collection method isn't great. But the increase in that proportion could be--indeed, is more likely to be--driven by the fall in other kinds of bankruptcies, than by the rise in medically driven bankruptcies.

Susan [Yes, that's me] 1 year ago in reply to Megan McArdle
By "driven higher," do you mean an increase in the number of bankruptcies?

Megan McArdle 1 year ago in reply to Susan
The number of medical bankruptcies *fell*. Their paper implies, while never quite saying, that it *rose*.

Susan 1 year ago in reply to Susan
Doesn't the paper say the percentage of medical bankruptcies rose, not the number? The study says "Since 2001, the proportion of all bankruptcies attributable to medical problems has increased by 50%. Nearly two thirds of all bankruptcies are now linked to illness." (bolding is mine)

jstrummer 1 year ago in reply to Susan
This post is totally bizarre. The study is not misleading on its face or obviously wrong. It's talking about what are the major drivers of growth in bankruptcies since 2005. Obviously the 2005 reform cut down on the number of filings. That's not at issue. At issue is what is causing the new growth post-2005. The study talks about percentages and proportions. It's not comparing absolute number of medical bankruptcies today with those 5 years ago. That would be apples to oranges.


ifthethunderdontgetya™³²®© said...

One is tempted to conclude that she's vapid, lazy, and entitled.

Susan of Texas said...

One would be correct.

I feel guilty for not breaking down her post, but just don't have enough energy and time today. I did enjoy seeing her link to her debunked work as proof of her thesis, however.

Dave said...

If the abolute number of medical BK decreased significantly , that does not suggests there is a huge problem with rising medical costs, regardless of what happened to the proportion.

But the real problem with this study is that medical costs aren't driving a large percentage of BK anyway. Look at the proportion of BK debt that is medical and this becomes pretty obvious.

Susan of Texas said...

Nice straw man you have there.