[ergothus]

As others have said, you cant ignore the purchase price of the research, and the buyer has that because they sell high priced drugs

...that said, it makes a good point about how the incentives work. If the govt did Xprize-like rewards for medical goals, it would work similarly: you stop paying for failed research and only pay for success, just like the examples you cite.

All told, I think too many electrons are spilled arguing over the need to pay for R&D (direct or indirect) and not enough answering the title here: "how much?" I would not be surprised to discover a dramatic reduction in prices could occur with minimal impact on results - which doesnt deny the need to pay for R&D. I'd rather people with expertise focus on how much,as well as figuring out how to shift incentives towards completeness over the current incentive for lockdown.

[testpostpls]

I don’t quite understand your point about govt-funded Xprize-like rewards—who is paying for failed research under such a regime? Is a govt-provided IP monopoly not itself an “Xprize-like reward?”

[ergothus]

The concept that has been upheld for the current system in the past is: Companies do R&D, some fails, and some succeeds, and the profits from the successes pay for both the failures and successes.

The poster I was replying to points out that this concept isn't actually what is happening (at least, anymore, for some areas) - instead, smaller companies do R&D - some fail, and the investment is lost; some succeed, and the profit from being BOUGHT rewards the investment. The reward for success is great enough that some investors will gamble on these small companies (while the big companies investors are investing with lower risk, lower rewards).

I pointed out that this is very similar to the XPrize - private investment was done in many companies, but only one (ish) would win. (This is actually a bad analogy, since the XPrize was much more about status - the prize wasn't large enough to truly be significant to a company that could claim it). My (unspoken) point was that the govt could stipulate controls on the resulting product - companies could research with greater confidence of what payout they'd get (on success, of course), but the incentive for price gouging is removed.

The difference between today's IP-monopoly and the XPrize is that the prize was a known amount, and in the part I didn't state, minimal profit would be had past that, compared to today's "you get as much as you can gouge, for as long as you can abuse rent-seeking".

Really I just did a terrible job of explaining my point.