Arnold Kling  

The Chess Game of Health Care Regulation

Laboratories of Socialism... News Item: Half of American Fa...

Daeho Kim writes,

To identify the impact of PPS on cardiac procedures, I exploit the discontinuity in Medicare eligibility at the age of 65. Utilizing data from before-and-after the PPS [prospective payment system] reform, I find a discontinuous change in CABG [coronary artery bypass graft] use at age-65 after the reform that implies an increase of 50 to 60 percent. Nearly all of the increase is driven by a composition change in the patients who receive CABG, with treatment expanded to patients who are observably healthier (i.e., fewer grafts or no comorbidity). Possible competing hypotheses do not exhibit changes at the age-65 threshold (e.g., disease incidence, insurance rates). The increased CABG use was not cost effective - the lower bound estimate of the cost per quality-adjusted life year was over one million dollars. The average cost payments of PPS provided incentives for hospitals to expand the use of technologies that have high fixed costs; an expansion that increased health care costs with possibly little health benefits.

Pointer from Reihan Salam, via Tyler Cowen.

The idea of PPS was to reduce costs. It did not work. The idea of the health insurance mandate in Massachusetts was to reduce costs (of uncompensated emergency room care). It did not work. The great technocratic hope these days (e.g., David Cutler) for reducing costs is pay-for-quality. It has already been tried in the UK. It did not work.

The problem is analogous to the problem of financial regulation that I described as a chess game. The regulator looks at existing behavior and sets up a set of regulatory incentives to try to achieve a result. However, the regulated entity does not respond passively. Instead, the health care provider (or the bank, if that is what we are talking about) figures out how to achieve its objectives within the new regulatory framework. The net result is that the intended consequences of the regulation are not achieved, while unintended consequences crop up.

Noah Millman finds that if you compare by decade the growth in per capita health care spending in the U.S. and other countries, the 1980s is the outlier period. Before the 1980s, we are within the ballpark of other countries, and after 1990 our spending grows at the same rate as other countries. We attained our status as a high spending in the 1980s. Tyler Cowen provided the pointer.

Comments and Sharing

COMMENTS (5 to date)
aretae writes:


I think you're generally brilliant on this topic, and that you know more than I do, but I worry that you're overly quick in dismissing pay-for-quality.

I know of roughly 4 healthcare systems operating in the rich world today:

1. US -- Government-subsidized, Employer-paid insurance.
2. Europe/Canada -- Single Payer
3. Germany/Kaiser -- HMO with Pay-for-quality (pay for no repeat visit).
4. Singapore -- Health Savings Accounts + Catastrophic coverage.

While Singapore's system is clearly the winner among systems for cost, I had understood that the German system was a strong 2nd place finisher. Am I mistaken?

Jehu writes:

This isn't a chess game---unless you're thinking of the infamous human 'chess' game from History of the World Part 1.
The regulated moves SO much faster than you do that it's like they take 5 or 6 moves for every one you take as the regulator. Getting into a manuever contest when your opponent has to work to NOT get inside your agency's OODA loop is pure insanity. And that assumes they don't just spend the effort and resources to capture you.

Caleb writes:

@ aretae.

I think you are not mistaken at all in you framing of the issue. However, the question still remains as to what mechanism establishes the best cost-for-value system overall. What you have to account for is not only cost v befit in the abstract, but how that calculus effects each legally defined society overall. You are right in identifying the effective systems of health risk management in modern societies. However, there is still the question of how each governance mechanism applied to each society in question.

Robert Easton writes:

aretae's list misses a model - The UK system: Single payer, single provider. There's a heavily regulated private market as well but few people use it.

Arnold says "The great technocratic hope these days for reducing costs is pay-for-quality. It has already been tried in the UK. It did not work.", which seems like it could not be more wrong to me. Government healthcare spending in the UK is similar to in the US, and with virtually no private spending that makes overall spending half what it is in the US. There are many problems with the UK healthcare system but cost is certainly not one of them.

Arnold Kling writes:

I was referring to a specific experiment in the UK, where they shifted to a "pay for quality" model. Studies showed that doctor compensation rose by 20 percent, but actual medical practice changed very little. The doctors managed to game the system.

Comments for this entry have been closed
Return to top