Andrew Samwick says:
Pecuniary Externalities, by Andrew Samwick: For what it's worth, I think Paul Krugman makes some good points about the problems inherent in using the tax code to encourage or discourage the purchase of health insurance in his column today (original here, reposted here). ...
However, I found this statement (highlighted in bold) in Krugman's column to be odd:
While proposing this high-end tax break, Mr. Bush is also proposing a tax increase — not on the wealthy, but on workers who, he thinks, have too much health insurance. The tax code, he said, “unwisely encourages workers to choose overly expensive, gold-plated plans. The result is that insurance premiums rise, and many Americans cannot afford the coverage they need.”
Again, wow. No economic analysis I’m aware of says that when Peter chooses a good health plan, he raises Paul’s premiums. And look at the condescension. Will all those who think they have “gold plated” health coverage please raise their hands?
Is he kidding me? That is almost the definition of a pecuniary externality. Wikipedia describes it as follows:
A pecuniary externality is an externality which operates through prices rather than through real resource effects. For example, an influx of city-dwellers buying second homes in a rural area can drive up house prices, making it difficult for young people in the area to get onto the property ladder.
This is in contrast with real externalities which have a direct resource effect on a third party. For example, pollution from a factory directly harms the environment...
So in the President's defense, there's a very simple argument to be made here. When one person feels inclined, for whatever reason, to purchase more health care services, that puts upward pressure on the price of health care services (if the supply curve is not flat) and thus the cost to everyone else in the market. Normally, we don't pay any attention to this, because that is precisely the mechanism by which a competitive market achieves economic efficiency.
The President is referring to the pecuniary externality generated by a tax distortion in the treatment of health insurance, which interferes with a market achieving economic efficiency and thus should concern us. It goes as follows. Premiums are fully excludable from income tax, but out-of-pocket expenses are not tax advantaged. That favors health insurance arrangements in which there are low deductibles and high premiums. Such arrangements can lead to higher utilization of health services, since the insured faces no financial cost at the margin once the low deductible has been met. (This is just a standard moral hazard argument.) Krugman may not believe that the relevant behavioral effects are large here, but he's on shaky ground with his "Wow ... no economic analysis ..." comment.
For more on pecuniary externalities, I came across this source.
There have also been several comments here on this point.
Andrew says the pecuniary externalities only occur "if the supply curve is not flat." Paul Krugman, via email, says the long-run supply curve for medical services is flat so Andrew's exception applies:
Paul Krugman: Aha - I was wondering if anyone would raise that. I was taking it as true to a pretty good approximation that the long-run supply curve for medical services is horizontal. Unless you think that there's permanently limited supply of medical education, or something, why should we think otherwise?
And I would guess that very few people would read Bush's statement to mean that it's bad if other people have extensive insurance, because it drives up doctors' paychecks.
In comments, Andrew responds with:
Andrew Samwick: Based on Krugman's response ..., we're now in the much more comfortable environment in which this is a few economists talking about the magnitude of various key parameters.
One could point out that if he was "wondering whether anyone would raise this point," then he seems to realize that he was going a bit overboard in claiming that "no economic analysis I'm aware of says that when Peter chooses a good health plan, he raises Paul’s premiums."
On the substantive point, one could assert that almost any market has a long-run supply curve that is flat. Exceptions would be made for markets like diamonds--there is a finite quantity available to be mined. At this juncture, it becomes quite relevant how long we think it will be before we are in the long run.
As evidence against this happening any time soon, I don't think the AMA is going to give up its near-monopoly on certifying medical practitioners. Licensed practitioners will be in short supply for a long time even if wholesale medical prices rise. In order to get more services when prices change over this long run, we have to build a lot of buildings--medical schools and hospitals--and fill them with really expensive equipment. I'm guessing that long run will take a while to get here.
I'd be interested in hearing about academic research on this topic. Anybody know about research on the shape of the long-run supply curve and the speed of the adjustment process?
Update: Alex Tabarrok implies that the long-run supply curve is relatively steep.