Arnold Kling  

Health Care Costs and Bankruptcy

My Alternate Opening Statement... Biggest Flaw in Labor Theory o...

I wrote on the famous study when it first came out. You should read that article. I have to say that I cannot think of any other study that has had such a high ratio of political impact to substantive validity.

Comments and Sharing

COMMENTS (8 to date)
Floccina writes:

That is truly a great article I read it back when it was first published and it opened my mind to some things. I like the idea of event-based insurance but because the incentive to fraud is big I think that it would have to evolve slowly. (BTW in my life I have known many people who fraudulently collected SS disability and workman's comp. Fraud seems rampant in those programs.)

Also, I have read that "medical bankruptcy" is a problem in some countries that have socialized medicine. It seems that nothing substitutes well for prudence and good family relations.

I would love hear from supporters of socialized medicine how much of own wealth people should be required to spend to care pay for their medical bills before the state kicks in.

My compromise plan would be:

My compromise plan:

The state would provide insurance to all Americans but the annual deductible would be equal to the family’s trailing year adjusted income minus the poverty line income (say $25,000 for a family of 4) + $300. So a family of 4 with a trailing year adjusted income of $30,000 would have a deductible of $5,300. A family of 4 with a trailing year adjusted income of $80,000 would have a deductible of $55,300.

We expect some people live at the poverty level, should we expect formerly middle class people to live at the poverty level after an illness?

Ted Craig writes:

From front-line experience, I can say medical costs are a main cause of financial hardship for low-income individuals. But divorce is a close second. The underlying issue in both cases is living on the edge financially. I leave it to others to debate wether that's a circumstance they created or were thrust into.

Brian Schwartz writes:

Has anyone looked at what John Goodman, Linda Goodman, Devon Herrick & Robert Sade wrote about medical costs and bankruptcy? (OK, maybe only one of them wrote this small section.) It's published by the National Center for Policy Analysis.

The first reference [54] is to the Himmelstein et al study in Health Affairs, which Arnold references in the TCS Daily article.

Here's the relevant section:

A study claiming that more than half of all bankruptcies are caused by medical debt [54] is often cited, but the claim conflicts with four decades of economic research. The label "medical bankruptcy" was applied if out-of-pocket medical bills exceeded $1,000, even though out-of-pocket expenses of the average US household were $2,182 in the year studied.[55] Recalculating the study’s data, Dranove and Millenson conclude that only 17% of the sample "had medical expenditure bankruptcies." [56] Well-designed economic studies have found no statistical link between bankruptcies and health problems.[57] In fact, household consumption is largely unchanged even in the face of very large medical bills.[58]
Source, with references to footnotes above:

Thanks for posting the TCS article, Arnold. I'll surely take a closer look.

Steve Roth writes:

Thanks Brian Schwartz for answering my question: is there more-current (and more-convincing) research?

I haven't read it all, but did follow footnote 57 to:

And am curious about that paper's footnote 23: "Income and the reduction in income are omitted [from the regression] because they are highly correlated with all three of the adverse events variables."

I wonder how a Bayesian should think about this, and their model in general. IOW, WWAGS? (What would Andrew Gelman say?)

John Goodman writes:

The link to your piece on event based insurance isn't working. Is there another way to access it?

Arnold Kling writes:


tcs changed its system so the links all have changed

Brian Clendinen writes:

@Brian S. great info, so the executive summary really should say " Medical cost where a contribution factor to almost half of all bankruptcies ."

So I still disagree with Dranove and Millenson classification based on the U.S. average. The classification Medical bankruptcies should be based on a given % of a person income which I would put into the 20% to 25% range (an alternative would be to use multiplier of 4 or 5 as I explain below).

They should also add a contributing factor class based on the average % of out-of-pocket medical expenses a family spends in their given CPI region. It should then be further broken down to indicate how much it contributed to the bankruptcy. This should be based on multipliers of the % of income. (

I think that information would give some useful information for policy purposes and how to design effective new types of insurance. Although I do not know how to take into consideration Net worth.

Mr. Econotarian writes:

Canada also has plenty of medical bankruptcies.

Why (when medical care is "free")?

Because when you get sick enough, you can't work, and if you don't get paid, you can't pay your mortgage, car loans, and you often end up in bankruptcy to save some of your assets.

In truth, medical bills are probably one of the easier bills to put off paying. As of right now, no one is going to repossess your new liver, but your new car is fair game. And your home can be foreclosed, and the sheriff will evict you. But doctors and hospitals have little option but to get an order to garnish your wages to pay your unpaid bills, which if you are not working is not going to help them get payed.

Comments for this entry have been closed
Return to top