[stephencanon]

There are multiple factors at work, only one of which is funding.

Residents are required to handle a minimum number of a large variety of cases by the time they graduate, in order to guarantee that they've seen a representative sample of cases in their field and have knowledge of all of them. E.g. a neurosurgery resident might need to do (completely fabricated numbers) 30 open vascular cases, 50 spine fusions, 40 tumors, etc. This is probably the primary limiting factor for specialist surgery residencies; these residents are profitable (they can handle the bulk of most simple cases fairly autonomously once they're a couple years into their training, and they stick around for 5-7 years), so many hospitals would like to hire more of them, but there are simply not enough patients with the necessary conditions for them to add more trainees.

For non-surgical residencies, the residencies are much shorter (so you have less time from highly-skilled residents), and the residents are less profitable, so funding is a significant limitation.

It's also important to note that residents are competing with mid-levels in the "less expensive practitioners" category, and mid-levels are a far better deal for the hospital in most specialties. They're somewhat more expensive in terms of raw salary, but they remain mid-levels, which means they have the time to develop near-perfect competence at the things they do handle, and they don't leave just when you've trained them up. A few good mid-levels make all the difference in keeping a department running smoothly.