Henry Aaron on Health Care Costs
Nov 15 2007

In this bonus middle-of-the-week podcast, Henry Aaron of the Brookings Institution talks with EconTalk host Russ Roberts about health care costs. Researchers in a New England Journal of Medicine article have estimated that the US could save $209 billion if the US went to a single-payer system like Canada. Is this number reliable? Aaron takes a deeper look at the estimate and discusses the relevance of such estimates for health care policy.

This is a special mid-week podcast. It's a follow-up to an earlier podcast with Arnold Kling that raised the issue of administrative costs and potential savings from going to a single-payer system. It also ties in with recent discussions here at EconTalk about the challenges of accurate measurement in the social sciences. We hope you enjoy it. If not, come back Monday when our regular schedule resumes.
Arnold Kling on the Economics of Health Care and the Crisis of Abundance
Arnold Kling of EconLog talks with EconTalk host Russ Roberts about the economics of health care and his book, A Crisis of Abundance: Rethinking How We Pay for Health Care. Kling discusses whether we get what we pay for when...
Arnold Kling on Hospitals and Health Care
Arnold Kling of EconLog talks with EconTalk host Russ Roberts about the death of his father and the lessons to be learned for how hospitals treat patients and our health care system treats hospitals.
Explore audio transcript, further reading that will help you delve deeper into this week’s episode, and vigorous conversations in the form of our comments section below.


Nov 15 2007 at 3:08pm

I’m a hospital administrator in Canada, though I take a very dim view of the social-insurance model. I remember sitting in the staff room reading the NEJM article in 2005 when a doctor (an American) came in, read the title over my shoulder, and immediately began crowing about the superiority of the Canadian system. Soon others joined in with her, until I started pointing out some of the same flaws that Prof. Aaron indicates. Most Americans (and Canadians) don’t know that competition between American health insurers is severely limited by interstate trade laws, that labor supply is articially constrained by professional regulation (and in other countries, but that doesn’t matter), that principal-agent/public-choice conflicts incentivize adminsitrators and politicians to overspend on technologies that drive costs at a rate far beyond productivity growth in healthcare, and on and on.

Most people also do not recognize that there are some very good things about those parts of US healthcare that are truly competitive. This includes–and let this be settled once and for all, please–the fact that Americans actually pay less for medication, because the higher price of their patented medications is easily offset by the very low cost of their generic medications, which are produced by a highly competitive and efficient industry.

The obsession with payer identity (apparently a trichotomy between state, insurer, or me) is also a great way to ignore the more costly and politically entrenched issue of professional regulation. Robyn Dawes, a former president of the American Psychological Association and accomplished researcher, published an excellent work called House of Cards, in which he debunks the relevance of professional regulation on healthcare quality. Is it possible under government-sanctioned monopoly regulations to err multiple times and still practice? Amazingly, yes. Is it possible that most of us just need the services of professionals who are average rather than Cadillacs, and that we would also be better off paying Prius prices rather than Cadillacs-for-all prices? Painful and humbling, but yes. There are dozens of other studies out there to prove the same point.

As for the willingness of the “average” taxpayer to part with more income in exchange for the promise of a universally accessible system of care, I think that says more about the auction value of political rhetoric than the taxpayer’s perception and measurement of value-for-money. What a wonderful way to make informational asymmetry even worse.

Nov 15 2007 at 3:14pm

P.S. That was the last time anyone at the hospital ever asked me for my political opinions.

Nov 15 2007 at 4:11pm

Small example: Women in the US choose their OB very carefully, they sometime interview several of them until they find the right one. Then they plan their pregnancy and delivery-strategy with the doctor they have chosen. In some European countries, women don’t have a choice, they have to show up at the hospital for the various routine check-ups and they are seen by the doctor on-call whoever that is, same goes the day of the delivery. Sure you could say that the second method saves money. But cost shouldn’t be measured in dollars only.

paul roman
Nov 15 2007 at 4:28pm

I guess that looking at executive salaries
Advertising costs, 1.3 billion DTC pharma.
profits to shareholders is where I would start looking before blaming the consumer for the problem.

John Alcorn
Nov 16 2007 at 7:11am

Prof. Roberts, You asked for feedback about the bonus, mid-week podcast to follow up in more depth a particular issue. Your talk with Henry Aaron was a helpful supplement to the previous discussion with Arnold Kling. High caliber scholars, constructive conversations. Today’s talk branched out usefully frrom the issue of administrative costs. I would have liked to see you dig deeper about what Henry Aaron callls cost-sharing. (I’m thinking of Arnold Kling’s point that insurance against catastrophes has morphed into insulation from routine costs.) Your podcasts are very helpful to non-economists like me who want to understand policy issues. Thanks!

Nov 16 2007 at 9:43am

If we get only 55 percent of the care that we should and yet when you remove accident and homicide deaths we live longer than they do in other rich countries, you have to ask how much care do they get? Of course maybe it is not that, maybe we live longer because we are fatter than they are.

What great ironic data has been coming out lately. It seems we do not know much about this subject. I am loving it.

Hey pass that twinky over here.

Nov 16 2007 at 10:01am

Some insurance company should offer insurance where an attempt is made to equalize the payments over a life time. The insured’s early payments would go half to insurance and half to the annuity. The deducible on the insurance part would rise over time with the value of annuity keeping the premium down. The annuity money would be used by the insured at his discretion to cover his own medical expenses up to the deductable causing him to economize. The insured would get the annuity in cash when he becomes eligible for Medicare.

Also I think that more self care should be allowed and encouraged. All of those “see you doctor before you do anything” messages IMO tends to scare people from acting on their own. Some pre-scription drugs should be dispensed to anyone how passes a little test.

Nov 18 2007 at 8:47pm

A great podcast and I enjoyed the extra information. I have one question however. It seemed to me that the study’s authors were saying that by having government provided insurance forms, nurses and other hospital staff would have an easier time filling them out. However, being an accounting graduate student, I can’t help but compare this idea with the current situation with our income tax forms. Sure, they all come from the government, but what is to say that we won’t have a scenario where hospital staff would use a 1040-style form for this type of patient while this other patient type is eligible to use the 1040EZ-style? I guess I am just skeptical that everyone could use the exact same form without some sort of rent-seeking shenanigans creating ever more confusing alterations and form types, even though they are still all government based insurance forms.

Nov 21 2007 at 12:11am


This was a wonderful addition. Keep these kind of supplemental sessions coming. One would wish that Members of Congress would listen to both the Kling and this session.

Comments are closed.


About this week's guest:

About ideas and people mentioned in this podcast:Articles:

Podcasts and Blogs:



Podcast Episode Highlights
0:36Intro. Markets versus the political process. Do markets work as well as people think? We can only get what we want if a lot of people want what we want. If you have unusual tastes, market doesn't provide the goods. By contrast, political process is subject to Mill's tyranny of the majority. But markets also are subject to a similar tyranny of the majority, though not for tie color. Fixed costs. For some goods, whatever level is provided is what is received by everyone. Only one level of army protection (whether privately or publicly supplied). Post office, by contrast, or public schools, are different kinds of goods and different levels can be received. Tyranny of the majority is a choice made by the government in those cases.
5:12Role of fixed cost. How many people out there affects choices. If fixed costs are large but not huge then it's not enough for me to want something for the market to supply it. Others with similar preferences to mine help me, but those who do not share my preferences don't hurt me. Radio industry, 20 stations on the dial, big enough city has lots of choices. In more extreme examples, like daily newspapers, maybe one in each major city. Abstraction, but that product has to choose how to target itself. Tyranny of the majority occurs. New York Times and USA Today are available nationally, though, and could compete with local papers. Natural solution to the tyranny problem is reducing the costs or reducing the market size. Trade increases the market size. In the trade example with newspapers, at first it looks like a new option. But as the NYT becomes available it causes the local papers to reposition toward the local clientele. Nothing wrong with that but it's not a smooth continuum of increase in products. Moving from small town, Fergus Falls, to big city Minneapolis, to very big city NYC, you get more restaurants and more diversity in restaurants. Larger cities have more choice. But it doesn't work that way for newspapers. Is it attributable though to fixed costs as the book says? Or is it that additional consumers are relatively cheap to service--economies of scale? Newspapers have high fixed costs. If people do not agree on the product, then a firm can make a product "better" with higher fixed costs but no higher marginal costs. John Sutton, in industries with fixed costs, they need not fragment as they get large.
14:26Quality. It's not that the marginal costs of quality are low. It's the marginal cost of producing another newspaper is low. In cities with only one newspaper, you get a better paper covering a wider array of stuff, but because the advertising revenue is high relative to subscription rate, paper's incentive is to widen their coverage. Washington Post has different sections for different neighborhoods. They do cater to individual audiences, but not the way we usually think of it. A newspaper is a bundle of articles. But take language--newspaper is written in a specific language. It is a complicated product. If you measure quality by number of newspapers it looks pretty thin, but if you measure it by quality of information available, say to the Hispanic or religious or other individual, the market has never had such a variety of information. Cable, non-local information, many languages and groups targeted. It's not the whole story, though. Local information versus national information. Ethnically targeted newspapers, tends to be weeklies rather than dailies. Hispanic consumer in Fergus Falls doesn't get same level of information, but is that the tyranny of the market? You do need to get to a critical mass if you are a small minority. But does adding people unlike you make your life harder? Look at African American newspapers, cross-effects are negative. High fixed cost product and more people of one type actually harm those of another type by pushing the newspapers out of providing information specific to the different groups.
22:37French movies. If you have Hollywood movies, it would be crazy for the French to restrict imports because it would restrict choices. But with big fixed costs it could displace French cinema. The usual explanation, though, for why the French restrict imported movies is protectionism, to enrich French filmmakers--rent-seeking. High fixed costs: there are high fixed costs of making a blockbuster. Requires big budget, so it had better attract a large audience. In books, Harry Potter; some publishers specialize in blockbusters. Yet there still seem to be plenty of books in small niche markets, directed at specialized markets. Movie costs have come down dramatically, financially viable to reach a small audience. Doesn't that cut into the French cinema argument? People's preferences over movies are very varied. Fragmentation around the fringes. Blockbuster end of the industry conforms to the fixed cost model but there is an explosion of movies at the other end. Chris Anderson on The Long Tail, technology has lowered even stocking choices. Digital divide affects African Americans, but technology of the internet is to the rescue.
29:42Role of fixed costs: they certainly can reduce consumer choice. Looking for an Afghan restaurant in Dayton, Ohio might disappoint. If you don't want a back seat with your car, you don't have that choice. Costly to tailor cars to individual consumers. Get more choices now, but don't get everything you want. But should we do something about it? Is it a problem? It could be efficient for people to be somewhat unhappy. What makes this more is that when fixed costs are big we don't expect markets to do all the things they ought to do. Markets accept products where revenue exceeds costs. If there is perfect price discrimination available, products will get tailored to everyone; but in reality that's not possible. So there may be a range of products where it's big enough that the value to society exceeds the cost but not big enough that the revenue exceeds the cost. Efficient, to economists, means there is a net benefit. (Several meanings for the term, this is one.) Deadweight loss, forgone net benefit is inefficient. You want to add up all the benefits, subtract all the costs; treat everyone equally, just a sum, not a weighted sum. But somebody may really value the good a lot, say really wants to eat at Afghan restaurant a lot, but only a few like that, a single price might not be able to cover the fixed costs of the restaurant even though there would be enormous value to everyone who even wants to eat there a little. Inability to price discriminate--or if the fixed costs were lower--there would be a net gain. But isn't that just a fact of life? Does it mean you have to do something about it? In large markets you can even have too many products. Empirical question--imperfect world. What could we do tangibly to make the world better? We have a fair number of government subsidies in markets like that. Air travel to small markets, public broadcasting, telecommunications in small markets. Do we abolish these things or say that maybe they are serving some efficiency-enhancing function?
38:30Mason City, IA, small town, far from any airports, so you have a drive a few hours, catch a connecting flight. City gets a subsidy of $1 million per year so an airline will serve them. Is that efficiency-enhancing? How widespread are those airline subsidies? As soon as a city gets big enough, it's no longer eligible. Something like 50 cities. Flat fee charged for everyone, nice for those being subsidized, paid for by the rest of us. Political process ignores efficiency and focuses on redistribution. How would you ever know if these inefficiencies are there as opposed to just being politically attractive? Could get data. Do high-value demanders get represented well in demand curve estimates? No, empirically hard to determine where demand curve hits the axis. Let's gather at least some evidence, though.
43:49Economic education. Book argues that textbook treatment of markets is overly optimistic, rosy. Is the economics education glass half full or half empty? How many economics majors are actually taught that markets work perfectly, versus how many are taught that markets are flawed? Supply and demand curves are good at some things but bad at capturing the dynamics of innovation over time. Perfectly competitive model is a model of homogeneous products, miss differentiation. Good intellectual benchmark and good first approximation, but it obscures other things. Teaching business students--unintuitive to them when economists say "taking quality as given." But in modern world quality is critical decision, which market should I be in. In economics classroom, it's how much should I make. Our models are very poor at the quality question, hedonics, Sherwin Rosen, Lancaster, Gary Becker have worked on it, but not a lot. John Sutton. Heterogeneity of people, winners and losers. Now data are more easily available.