> Here's a purely economic argument: if there is an expected profit from self-funding but it is less than that from receiving the Medicare funding, and there is a perception that the Medicare funding only exists because the argument has successfully been made that self-funding is not profitable, then making the smaller profit from self funding would put at risk the possibility of receiving the larger profit through Medicare funding, and administrators might determine it is not worth the risk.

No, that doesn't add up. Hospitals only receive $80,000 per resident from Medicare. If residency programs were profitable at level P, they could increase them from N residents to M residents, where (80000 + P)N < P M[0].

Furthermore, hospitals that currently don't receive any money from Medicare would simply expand self-funded programs, because they wouldn't be losing anything by doing so[1].

There's also no way that hospitals would be doing so much to preserve a mere $80,000 stipend, because increasing the number of physicians is in their best interest - it allows them to decrease their expenses (physician salaries) in the long run.

> That's all speculation, just like your purely economic argument is just speculation.

No, it's not speculation; it's exactly what hospitals, government employees, elected officials, and industry analysts have pretty much all been saying for decades. And it's supported by the actual evidence at hand, including all of the financial figures that they publish.

[0] Of course this doesn't work if P is a decreasing function of either N or M, which is the entire point - it is decreasing, and in fact, is already negative for the current value of N.

[1] Except, of course, if P is negative - which it is.

As we've gone on in this thread, you've provided an increasing amount of actual details on how this works (thanks!) and now express knowledge of external sources that back this up (though citations to those would be useful). Your initial claim read to me as just, "it's simple free market incentives", which is not the same as the more full picture we've gotten as we've gone on, which now includes more analysis of how the Medicare incentive might play out for different decision makers. I feel like that was peoples' point (at least it was mine): an indirect "follow the money and don't worry about why it works the way it does" argument was not sufficient. So thanks for taking the time to fill in a bunch of gaps!

I still think you may be downplaying the impact of the risk calculation hospitals have to make regarding their ability to receive the stipend now or in the future. It may be only 80k, but clearly that 80k is enough to incentive many hospitals to have residents, so it must be material to them to some extent.

Your point about increasing supply of doctors being in hospitals' interest in the long run is interesting, but this is the same training conundrum everyone has: it is often difficult to make the decision to invest in the near term when the payoff is not realized until much later.

I'm sure you're right about all this in general - training people is a tricky and expensive problem for every industry.