Asymmetrical Information - Megan McArdle

03.01.13

Can the Government Control Health Care Costs By Fiat

Bureaucrats are not Captain Jean-Luc Picard. They cannot just "make it so".

A couple of days ago, I did an interview with economist Bart Wilson on the role of prices in markets.  We were riffing off the commentary on Steve Brill's 22,000 word opus on hospital pricing, but some readers complained that we weren't really getting down in the weeds of the health care market.  

Well, the dark and weedy recesses of the health care market are Martin Gaynor's specialty, and he's written a great piece for the health care blog on cost control: 

. . . there are some countries that use rate setting, such as Australia, France, Israel, and Italy that have lower growth rates than the US, and some such as Canada, Finland, and the UK that have higher growth rates. The US is below the OECD average, whereas Finland is above, as is The Netherlands. While I wouldn’t put much weight on anything we see in cross-country differences (there are way too many differences across countries besides price controls), nonetheless nothing striking emerges from these numbers.

. . . So what do we conclude? My answer is that we don’t know what the impact of rate setting (price controls) would be on health care spending in the US. It’s possible that rate setting could prevent some of the most egregious practices recorded in the Brill article, but that depends on what’s enacted and how it’s enforced. Whether rate setting would substantially slow the rate of growth of health care spending isn’t clear. Further, the question that must be asked is what is the alternative? There’s evidence to suggest that robust price competition, such as we had with managed care during the 1990s, can perform very well in controlling costs. Unfortunately there has been a tremendous amount of consolidation in health care markets since the 1990s, raising serious challenges to competition. Whether the US decides to go with competition or with regulation, we have some serious work to do to make the system we choose work effectively.

Medicare does seem to show somewhat lower average growth rates than the private sector in recent years.  But we should be cautious about assuming that this could be extended economy-wide.

Source: The Health Care Blog ()

As I've written before, Medicare pricing exploits the difference between the marginal cost and the average cost.  Say you have a hospital with 100 rooms.  That hospital costs you something, say $100,000 in mortgage every month.  On top of that, you have to heat it, provide telephone service, pay to have the lobbies cleaned, maintain at least a skeleton staff of doctors and nurses and check-in clerks and billing specialists, put flowers in the lobby, and so forth--all before you let the first patient through the door.  That's your fixed cost, and in our fictional example (numbers chosen less for accuracy than for easy division), it's $3,00 per room per month.

Now on top of that fixed cost you have a marginal cost: the amount that it costs you to actually put a patient in a room.  Patients require a nurse to watch over them, an orderly to wheel them around, more doctors, food, laundry, medicine, electricity, heat and so on.  Let's say for simplicity's sake that our marginal cost is also $3,000 a month, for a total cost of $6,000 a month, or $200 a day.  

In order to make my hospital pay, I need each room to average $6,000 a month.  But that doesn't mean that every patient should pay exactly $200 a night.  Probably, not all of my rooms will be filled every night--in fact, we don't want them to be, because then where will we put emergency admissions?  If they're filled 25 days out of 30, then I'll need to charge at least $240 a night to break even.

But let's that I'm a hospital adminsitrator at this hospital, averaging 5 empty days a month per room, and a new government program comes to me.  This program plans to give the uninsured health insurance, so all the new patients will be people we weren't treating before.  They'll guarantee us another 100 patient-nights a month.  But there's a catch: they only want to pay $150 a night.  Should I take this deal?

Absolutely!  I've already got the fixed expenses covered by the existing patients.  And it only costs me an extra $100 per night to actually care for the added patients, so they're giving me a clear $50 a night profit.  I might even be able to give all the other patients a slight discount.  Everyone wins.  

That's why it's possible that Medicare is a great deal for hospitals--and that nonetheless, we cannot save money by puting everyone on Medicare.  (At least, not without pushing a bunch of hospital groups into bankruptcy--or raising Medicare rates so that they aren't so cheap any more.)  Everyone wants to be the marginal cost consumer. In fact, most people think they deserve to be the marginal cost consumer. But this is mathematically impossible.  Someone has to pay for the walls and the air conditioning.

So the disparity we see in growth rates between Medicare payments and those in the private sector may simply reflect cost transfers: the government declares by fiat that it will not pay, and so more and more costs are transferred to the private payers.  Medicare costs grow more slowly; private sector costs grow faster.  We have not ended cost inflation, just redistributed it.

Source: The Health Care Blog ()

The experience in other countries that use rate setting should give us some pause.  Implemented at a national scale, fiat pricing does not seem to be a sure-fire strategy for holding down cost growth.  

But wait, I hear you cry--Europe has lower costs than we do!  Indeed they do.  But they don't have lower cost growth.  The US experienced a unique burst of health care cost-inflation in the 1980s. Growth rates since then have been well within OECD averages, but we're growing from a much higher base, and over time that's added up to a giant disparity.  

However, no one is really talking about using fiat pricing to actually lower reimbursements.  It's a lot easier to curb increases than it is to do actual cuts, for the same reason that you'd find it easier to stomach a year with no raises than you would a sudden 10% decrease in your salary.  If countries that use rate setting can't even reliably hodl their growth rates down below ours, I find it hard to believe that rate setting is going to somehow enable us to make large absolute cuts.

All of which, I think, reinforces Wilson's core argument: ". . . the simple solution sounds like, let the government take it all over. But it is only the statement that is simple."