Economics for Health Policy Wonks
This blog is one of the few health policy blogs that approaches its subject from an economic point of view. We believe that almost all the problems in health policy arise because of perverse economic incentives and that to solve those problems we have to get the incentives right.
In a way that’s too bad. That is, it’s too bad there are so few of us. What’s most missing from some otherwise good sites is…well…economics. Examples from Aaron Carroll, Austin Frakt and Sarah Kliff below the fold.
Dedicated to all the teachers in the world.
A friend who taught me
Right from wrong
This is from Aaron Carroll:
One thing to note is that the top 5% of spenders (some of the sickest among us) account for about half of all health care spending. More significantly, the bottom half of spenders (i.e., the healthier half) account for less than 3% of all health care spending…
When we talk about consumer directed health care, we’re talking mostly about healthy people…The problem is that healthy people consume so little care to begin with. If we could incentivize the healthier half of people to forgo all their personal health care spending, we’d spend $36 billion less out of a total $1.259 trillion in personal health care spending. That would be a drop in the bucket. And no one — no one at all — thinks we can get people to stop all their health care spending.
As I pointed out in the comment section of his blog, the patients in the top 5% this year are not the same as the 5% next year or the year after that. If they were the same people year after year, then our health care spending problems would boil down to the treatment of a small group of people. That, however, is not the case. That’s why you need good economic incentives throughout the entire distribution. And when all the employees face good incentives, the Rand Corporation finds that consumer directed health care plans lower the cost of health insurance by as much as 30%.
This is from Austin Frakt defending (apologizing for?) health insurance coverage for contraceptives:
If you’re in a hurry, all you really need to know is that health insurance is redistributive…
Taking contraceptives as a case study, it is very likely that almost no men use covered contraceptives. Same goes for most women above or below child-bearing age…The non-users subsidize the users, who clearly benefit with a lower net cost of contraceptives than they would otherwise pay without insurance.
Ah, but real insurance is a fair gamble. You and I pay the same premium, say, for a service but neither of us knows in advance which of us will actually get the benefit. On the other hand, if I force you to pay the same premium as I pay for a service and we both know in advance that I will use the service and you will not, that is not insurance. That’s theft. By the way, the vast majority of people in health policy don’t know the difference between insurance and theft.
This is from Sarah Kliff:
If you want to understand the problem of how we currently finance our health care system, there’s a great case study to be had in HCA, the largest for-profit chain of hospitals in the country. A New York Times investigation out Tuesday morning found HCA cardiologists in Florida to be performing unnecessary procedures on patients who, in some cases, did not even have heart disease in the first place.
Why perform an invasive procedure, which comes with risks to the patient, if it’s unnecessary?…[T]he doctors may have performed unnecessary procedures because there was a financial incentive to do so.
Get that? Now try this:
The health reform law does take some steps to address these unnecessary procedures, by creating new payment models where providers would essentially lose money by performing procedures they don’t need to. That’s the whole idea behind the health law’s Accountable Care Organizations, where doctors band together and take a lump-sum for covering a set number of patients. If they can deliver care for less — while hitting certain quality metrics — they pocket what’s leftover. Doctors that perform unnecessary care, and go over that set amount, will find themselves in the red.
It’s really hard to know where to begin. Almost all services in our economy are purchased fee-for-service. And usually without a hitch. Did you know that the iPhone repair shop down the street doesn’t perform unnecessary procedures on my iPhone? Are you surprised by that fact?
The problem is not fee-for-service payment; it’s third-party payment — as all loyal readers at this site already know. Any time you have third-party payment, you will have distorted incentives of one sort or another. If you pay doctors to overprovide, you are going to get overprovision. If you pay doctors to underprovide, you will get underprovision. ACOs are going to underprovide. Big time.
Finally, here is Aaron Carroll again, answering this question: “Will ObamaCare raise the price of your pizza?”
Let’s circle back around to Papa John’s, though. Part of what the act does is mandate that companies start providing health insurance to their employees or pay a penalty. Since some don’t do that already, this will cost them money. They could take this out of profits or reduce the salaries of their executives, but they will probably do what every business does: They’ll pass it on to the consumer.
This is as it should be. Some companies probably keep costs down by not providing comprehensive health benefits to their employees. Now, they will have to. I imagine some companies already do, which probably increases their costs, and now they will be on a more level playing field. Regardless, Papa John’s is telling you that people who order its pizza will now bear the cost of its employees’ health insurance.
Sorry. That won’t work, as should be clear by examining the equation of exchange:
MV = PQ
The amount of money in circulation (M) times the average number of times each dollar is spent (V) is equal to the sum total of all spending, or total output (Q) times the price level (P). Clearly, ObamaCare doesn’t change M or V. So total spending must remain the same. Conceivably, it could lower output by causing workers to lose their jobs. But Carroll implicitly assumes that everyone keeps right on working. So what happens to P? Nothing. One price can go up only if another price falls. But prices cannot on average go up or down as long as M, V and Q remain unchanged.
So who is going to pay for health insurance at Papa John’s and all the many other businesses that will face increased health insurance costs (which, by the way, will be almost all of them)? Answer: the employees. The economics literature is unambiguous on this point: employee benefits substitute dollar-for-dollar for money wages. For family coverage, Papa John’s employees are going to have to receive about half their compensation in the form of health insurance instead of wages — whether they like it or not.
Think about that the next time you hear an administration spokesperson ticking off all the goodies offered by ObamaCare. Free wellness exams, free contraceptives, free this, free that. The cost of all those freebies is going to come right out of the worker’s pocket.
Great analysis, as usual.
Super post! Kudos.
Ah, but real insurance is a fair gamble. You and I pay the same premium, say, for a service but neither of us knows in advance which of us will actually get the benefit.
That’s something that otherwise intelligent health policy wonks fail to understand. I’ve presented at numerous conference where advocates for socialized health care systems repeated the mantra… “…health coverage cannot be affordable until everyone is covered…”
What does this even mean? It’s nonsense. But what proponents think it means is that everyone should buy insurance that charges community-rated premiums so people who are poor risks get implicit subsidies (i.e. poor young people should subsidize rich old folks).
This violates the tenants of economics. How can an insurance company compete on price if (while quoting a high price) it explains how your premium is really good considering it would be lower if not for the fact your premiums are subsidizing those less healthy than you.
If middle class wages have stagnated over the past few years, the next few years will result in even less take home pay. Surely the middle class will revolt, but will they figure out that it’s because their health care premiums are directly affecting their bottom line?
To me the key line is that the “majority of people in health policy don’t know the difference between insurance and theft.”
The majority seem also not to understand the importance of targeting one’s policy for effectiveness reasons. On the latter, an example:
I see that obesity is a social problem. My response? Let’s ban the manufacture, sale, purchase, and possession of forks. We can start by restricting the sale of forks to those with three or fewer tines.
What is missing from otherwise good sites is economics? Well, not exactly. It seems that with this statement you are rejecting normative economics, which is ironic in that this site relies much more heavily on normative economics than on purely scientific economics (if such even exists within the realm of health care). The disagreement is not on the science of economics but rather on the value judgements to which economic theory is applied.
I am amazed at how many health “economists” appear not to have read Ken Arrow’s seminal paper of 1963 on the health care market.
Arrow argued that the chief characteristics of medical markets that makes the application of straightforward market principles problematic are (1) a variety of uncertainties whose implied risks cannot be traded in private markets, (2) asymmetric information, besides uncertainty, and (3) a distinct distributional social ethic that differs from that implied in pure market forces. In a nutshell, Arrow argued that there is a difference between health care and widgets, the mythical commodity for which textbooks are written.
Now your readers may say, “Who the hell is Ken Arrow?”, but that is their problem.
Julie:
I think there’s reason to be hopeful that the middle class will come to realize that wages and benefits are fungible. Many employers (mine included) already provide annual “total compensation” statements which break down the components of total compensation (e.g., cash wages + social security taxes + benefits + etc.).
My wife’s employer (she’s a teacher at a private school) went a step further this year, telling the teachers that due to increasing health insurance premiums, increases in cash wages would be lower than they otherwise would have been, and pointing out that total compensation was still increasing by about 3%.
I like the idea of a total compensation statement.
Professor Rheinhardt makes a very compelling point.
Maybe John and Uwe could comment and critique Arrow’s paper?
“Free” is a term used often by the Obama Admin. Not hard to figure out that entitlements or healthcare, are not “free”.
Sarah Kliff wrote —
“If you want to understand the problem of how we currently finance our health care system, there’s a great case study to be had in HCA, the largest for-profit chain of hospitals in the country. A New York Times investigation out Tuesday morning found HCA cardiologists in Florida to be performing unnecessary procedures on patients who, in some cases, did not even have heart disease in the first place.
“Why perform an invasive procedure, which comes with risks to the patient, if it’s unnecessary?…[T]he doctors may have performed unnecessary procedures because there was a financial incentive to do so.”
This argument is always a hoot. Good ol’ Jack Wennberg made the same argument about hysterectomies in Lewiston vs. Wiscassett Maine. Problem with it is that the incentives in Lewiston were PRECISELY THE SAME as they were in Wiscassett. Ergo, the financial incentives did NOT drive behavior. But these folks are so smug that they never bother to look into what DOES drive behavioral differences.
If HCA cardiologists in Florida behave differently than HCA cardiologists in Kansas City, and reimbursement policies are identical, Ms. Kliff should be looking at in what ways patients are different in the two areas.
Greg,
I think the incentives are different in the case of a for-profit hospital like HCA, where you have the investors and administrators breathing down your neck for not performing enough procedures, compared to working in private practice or a hospital, where only personal profits are at stake.
Two good points from @Don and @Uwe. I am fascinated that, every time someone disagrees with him, John Goodman assumes they are not taking an economic approach to the problem (whatever that means) but must be applying some other, lesser standard. In reality, John’s classical approach requires assuming things such as information symmetry that do not exist in the real world.
In the words of Adam Savage, Goodman rejects reality and substitutes his own. Why is anyone surprised that he comes up with non-solutions?
As an economist who works in health policy, I spend my days trying to convince people that the economic principles that apply to other industries don’t (for some unexplained reason) suddenly stop working with health care. Granted, there are normative factors to consider in the delivery of medical services, as Dr. McCanne reminds us. As a society, we don’t want people to drop dead on the street because they’re having a heart attack and lack health insurance or sufficient funds to pay for medical care. Yet, if society were to decide health care should be organized as a pure public good devoid of price rationing, the principles of economics should still not be dismissed as meaningless. In a system where price rationing is prohibited, another form of rationing will necessarily take its place. For instance, if the price of medical care was set to $0 at the point of service, the quantity of medical care demanded would exceed the quantity supplied (resulting in waiting lines).
Many countries with National Health Systems limit technology and use rationing by waiting to smooth demand to a level deemed affordable by society. This is something most Americans would not tolerate. While many Americans routinely answer polls saying that nobody in need should be deprived of medical care, the same people resist higher tax burdens and oppose any move that might eliminate their own private health insurance.
The degree to which health care in our society is a normal good, a luxury good – or a public good – is a delicate balance that policy makers must navigate. However, rejecting economics as a tool to efficiently allocate health care resources is something that policy wonks do at their own peril (and ours).
David,
Nonsense. Give an example of a market in which there is not information asymmetry. What do we do when faced with a system with extreme asymmetry? We hire an agent. That agent might be an attorney, an accountant, an engineer. But in every case (except health care) that agent works for us and is accountable to us. In health care, ALL the agents work for the third-party payer.
Seems we need a critique of Arrow’s paper. It seems possible to me that we have a lot of experience since 1963 that might show that Arrow’s insights, though perhaps true, are not sufficient to direct us to the correct type of health care system.
Kenneth Arrow’s article, Uncertainty and the Welfare Economics of Medical Care
(December 1963), is an interesting article for discussion. But it does not make the definitive case that health care doesn’t adhere to any of the principles of economics. The RAND Heath Insurance Experiment illustrated that health care utilization does respond to incentives.
Greg is correct that other areas of our economy have information asymmetry. The information asymmetry (principal / agent) problem in health care is made worse by the way medical providers are licensed and medical care is paid for.
The information asymmetry argument is even more interesting given that nearly 50 years have passed since Kenneth Arrow wrote his article. More information (and misinformation) is now available on the Internet than Kenneth Arrow could ever have dreamed of back in 1963. When Arrow wrote his article, patients received virtually all of their information from their doctor. Today, patients spending a few hours on the Internet can dig through a treasure trove of information that their doctor would never have time to peruse.
The perverse incentives created by third-party payment means that both patients and their doctors are complicit in driving utilization. The principal / agent problem in 1963 may have been the problem of doctors knowing more than their patients and not being a reliable agent. The principal/agent problem has since become a problem of doctors and patients having more information than the insurers (payers), and becoming less prudent in their consumption of medical care.
Here is my take on why the Carroll/NIHCM analysis is probably not right and the Rand research (both old and new) on the cost containment potential of moderate amounts of upfront cost sharing is probably right. Rand showed that the main effect of cost sharing is to reduce the rate at which people initiate episodes of care. Some of these episodes end up being very expensive, thus plunging someone into the top 5% of all spending. But, not only is that not known at the beginning whether an episode will turn out to be high or low cost, it appears that cost sharing discourages treatment that ends up being high cost (think of the endless spending to track down an allergy). How else explain that it cut inpatient use by 30%? So, while high cost sharing does not affect most people who are high cost (because they blow through the deductible), it need not prevent the start of very many high cost episodes to save a lot of money. We obviously need much more of the detail here but I bet this is what happens. Plus most of the NIHCM analysis of “high cost” patients is on people over 65 and refers to Medicare spending. I would have more worries about a HDHP for people (like me) about to go on Medicare (though I could see it as an option), but the real applications are to the under 65s where long term chronic conditions that have to be costly are much less of a phenomenon and even most expensive illness tends to resolve after 3 or 4 years (in either recovery or death).
@Greg, should we start hiring doctors to go with us to the doctors office?
With reference to Professor Reinhardt’s comment on Arrow’s 1963 paper, it is worthwhile noting that subsequent work by Professor Pauly (1968) and others suggest that Professor Arrow may have been unduely pessimistic about the uniqueness of medical care and the welfare gains from having government “undertake insurance in those cases where this market, for whatever reason, has failed to emerge.”
In a 1968 paper, Professor Pauly showed that “even if all individuals are risk-averters, insurance against some types of uncertain events may be nonoptimal. Hence, the fact that certain kinds of insurance have failed to emerge in the private market may be no indication of nonoptimality, and compulsory government insurance against some uncertain events may lead to inefficiency.”
Professor Arrow commented on the paper. He wrote that “Mr. Pauly has convincingly shown that the optimality of complete insurance is no longer valid when the method of insurance influences the demand for the services provided by the insurance policy…In the theory of optimal allocation of resources under risk bearing it can be shown that competitive insurance markets will yield optimal allocation when the events insured are not controllable by individual behavior. If the amount of insurance payment is in any way dependent on a decision of the insured as well as on a state of nature, then the effect is very much the same as that of any excise tax and optimality will not be achieved either by the competitive system or by an attempt by the government to simulate a perfectly competitive system.”
Professor Arrow ends his comment by writing that “the lesson of Mr. Pauly’s paper is that the price system is intrinsically limited in its scope by our inability to make factual distinctions needed for optimal pricing under uncertainty. Nonmarket controls, whether internalized as moral principles or externally imposed, are to some extent essential for efficiency.
Which is in essence a trivial comment, except for the fact that a large segment of that generation of economists automatically assumed that government officials and their pet experts had the secret sauce, and saintly motivation, necessary to construct pareto optimal external controls.
I would submit that subsequent events have shown that this view is erroneous. Furthermore, there is no “distinct distributional social ethic” and almost all markets suffer, to one degree or another, from asymmetric information.
@david: We already hire doctors to send us to other doctors. And then many people go to a different doctor and ask what they would do and compare the advice. And, anyone with a doctor in their family will usually take the family member’s recommendation over any number of paid agents (or at least over having to pay agents).
So, I guess that means, yes, you should hire (or ask) someone you trust to guide you through the health care system.
@Uwe Reinhardt:
1. Presumably you realize that the origins of the “agents must have perfect information to make rational decisions” is a profoundly undrealistic condition necessary to make the computational exercises as the heart of neoclassical DSGE models mathematically tractable – not something that actual humans have ever required to make decisions in the real world, which are always conducted under conditions of incomplete information and uncertainty.
2. I’ll see you Arrow’s 1963 paper and raise you Hayek’s “The Use of Knowledge in Society.” (1945)
It’s not clear to me how one can uncritically cite Arrow’s paper in defense of substituting centralized bureaucratic mechanisms for coordinating supply and demand and allocating resources if one has read Hayek’s paper, but I’m certainly willing to read through any attempt that you’d care to offer.
@ Buster and Chuck:
Agree
@ Julie Sias and Keith:
Let’s hope so.
@ Earl Grinos:
Why limit the ban to forks? We could make eating even more difficult if we banned all silver ware and all plates.
@ Don McCanne
I disagree. What we have here are violations of basic economic theory.
@ Uwe Reinhardt, Jim Morrison, Mike Ainslie, David Bob Blanford, and JayB
As Greg Scandlen, Linda Gorman and Devon Herrick point out, nothing really follows from the fact that there is asymmetry of information. Or from any other mareket “imperfection,” for that matter. Is there some reason to believe there are fewer imperfections in the politcal system, thus establishing a prima facie case for government intervention? In most cases the answer is no.
@ Mark Pauly
Excellent analysis. Very insightful point.
@ Linda Gorman
Thanks for that bit of history, which I suspect most economists are unaware of.
“Is there some reason to believe there are fewer imperfections in the politcal system, thus establishing a prima facie case for government intervention? In most cases the answer is no.”
Couldn’t agree more.
To be fair – I think that the incapacity to distinguish between the formalisms at the heart of neoclassical econ (the map) and the reality inhabited by real humans(the territory) that characterizes virtually all of the output issued forth by academic health economists is also characteristic of most modern academic macro-economists.
Scan the field from Samuelson to Stiglitz and you’ll see towering figures in the field using basically the same imperfect/asymetric information argument to argue that market outcomes can never be pareto-optimal and justify government intervention on those grounds – presumably by technocrats who are always and everywhere operating with superior information, superior rationality, superior incentives, superior ethics, and superior knowledge of each individual’s true preferences. That, and the information constraints cited by Hayek (most of the consequential information that economic actors rely upon is too disperse, implicit, time-sensitive, etc to be used effectively by remote actors) evidently don’t apply to them either.
Arnold Kling’s one-sentence summary of Atul Gawende’s worldview applies equally to virtually all of them.
“…you have to understand that the concept of consumers and consumer choice is alien to Gawande. He is deeply rooted in folk economics. He trusts individual agents, not market processes. He has a magical-thinking model of government.”
http://econlog.econlib.org/archives/2012/08/atul_gawande_on.html
@ JayB
I agree. And thanks for the nifty quote from Arnold.
John,
An economic point of view is fine, but only if the providers of health care are taken to task becuase their major goal is to scam the system to fine undeserved revenue. The economic perversity is physician and institutional based, only to increase revenue, not to provide the best care, which most of the time is the least expensive. So don’t use economics as the driver. Somehow get rid of the greed and the economics will follow. Get primary care back in the driver’s seat, and you will see how well it works. But you must be aware and mindful of the fact that primary care has it flaws, but should be easier to confront and manage.
Dr Bob Kramer
@Zack, I never said we didn’t or couldn’t hire other doctors, but that it won’t be cheap or efficient, especially if we’re going to have this god-awful doctor shortage Goodman warns about.
@John Goodman, let’s be serious. If 8 years of medical school doesn’t qualify as asymmetry of information, then no such thing exists. I agree that asymmetry of information doesn’t necessarily justify government intervention, but that isn’t the point. Right now that asymmetry is mitigated by insurance companies. They have people with more knowledge who determine whether medical products and services are worth the money the insurance companies pay for them. That’s something you pay for when you buy insurance–for someone else to decide whether a given product or service is worth the money.
When most people walk into a doctor’s office, they are complete at his or her mercy, and many doctors take advantage of that.
@JayB, the government is by definition a more powerful market actor and thus often has better information. For instance, the SEC has a better grasp of a company’s financial information because they can determine what must be provided, set accounting standards, etc. And if the problem is remote actors trying to make decisions with imperfect information, wouldn’t that problem apply equally to a corporation? The only difference is that a corporation’s market motivation is profit, whereas government’s is political.
You may argue that the profit motive is “better” than the political one (you’d be wrong, but that’s another question entirely), but the information asymmetries and dispersian apply just as equally.
Uwe writes: “there is a difference between health care and widgets, ” … “Now your readers may say, “Who the hell is Ken Arrow?”, but that is their problem.”
I love this elitist attitude.
Widgets aside, there is a difference between talking about health care and actually providing it.
Perhaps if Ken Arrow was in charge we would still be doing open heart massage.