This was an interesting read. You might be interested to see my own writing on this topic, recently posted as part of the ongoing decadal survey of particle physics: https://arxiv.org/abs/2207.00122 (Section 7). I also consider researcher-directed funding as one possibility worth trying.
The problem with your "true" proposal of "focused research organizations" is that it won't work. You note early in the essay that grants require "changing your research agenda to fit the state’s goals". This is nominally true, and yet it only partially recognizes the real problem. Scientists are very smart and dedicated to their research goals. More often, they change how they *present* their original goals to fit into whatever hot topic is currently getting more funding from the federal government. That means the government virtually never realizes the intended purpose of targeted funding for specific topics, while also forcing scientists to go through more contortions to keep their labs going.
Also, a correction: you state "most of the other half goes through the National Institutes of Health (NIH) and the National Science Foundation (NSF)." In fact, the Department of Energy Office of Science has a budget similar in size to the NSF. See https://www.aip.org/fyi/federal-science-budget-tracker. It's quite harmful to physical science research that the DOE is frequently left out of popular conversations about science funding.
This was very informative, thank you. I took a different approach myself, largely based on the lottery approach you mention. I would love your thoughts on this assay of mine:
Only someone who has zero familiarity with the current academic research environment would suggest that the problem is *too much* external oversight of research directions and *not enough* researcher choice of research direction. The real problems are (1) Far too much mediocre research, coupled with a bean-counting approach to research output, driving even talented researchers to produce incremental "LPUs" as fast as they can, at the expense of quality; (2) an incestuous academic-political environment in which credit and funding is granted based on political power and mutual log-rolling, with skilled politicians and self-publicists being able to generate prestige (and funds) for their own and each other's worthless make-work research while starving rivals; and (3) the fox-guarding-the-chicken-coop approach of funding agencies "renting" researchers to judge their colleagues' worthiness on a rotating basis, ensuring that no external oversight of large research fields is possible.
The fundamental problem, of course, is that because research is inherently speculative, it's difficult to measure the relative (or absolute) value of a given research agenda, so various proxy measures are substituted, all of them woefully inadequate and easily corrupted. An alternative approach would be to reward results rather than fund research directly--i.e., award prizes to researchers (or their employers) based on the impact of their work. Research labs could then fund researchers whose potential they believe in as investments, hoping to profit as a result. Among the benefits: funders would have a stake in funding (only) promising research with a real chance of success; researchers would be encouraged to make higher-risk, higher-reward bets--something that faces massive disincentives under the current system; and the influence of petty politics among research communities would be overshadowed by the effects of concrete achievements.
This is an excellent proposal. It mirrors school choice driven policy - fund the student, not the school. I don't think the research university vs teaching university dichotomy need be the weakness you suggest; in many ways, that distinction is artificially created by the current research funding system. Instead, simply fund everyone who has a PhD in a given field, with the 60/40 stipulation in funding allocation, and let them go where they please. One's prize for obtaining a PhD would then be lifetime funding.
That opens its own gaming strategy of course - there would need to be a considerable degree of policing to prevent fake doctorates from parasitically crowding the system. But, that could be dealt with by allocating funding by dividing the pot by the number of PhDs. The disciplines would then be strongly motivated to limit the number of entrants.
One thing I'm surprised you didn't mention, in terms of rent-seeking that drains resources towards unproductive activities, is F&A. Overheads are a huge unnecessary cost, which additionally drive much of the administrative bloat that is in turn largely responsible for the totalitarian climate of the modern campus. Eliminating overheads would be a massive win in itself. Your proposal naturally leans in that direction: if a given researcher's funding is capped, and he can go where he pleases, organizations would compete to offer the lowest overheads possible.
This was an interesting read. You might be interested to see my own writing on this topic, recently posted as part of the ongoing decadal survey of particle physics: https://arxiv.org/abs/2207.00122 (Section 7). I also consider researcher-directed funding as one possibility worth trying.
The problem with your "true" proposal of "focused research organizations" is that it won't work. You note early in the essay that grants require "changing your research agenda to fit the state’s goals". This is nominally true, and yet it only partially recognizes the real problem. Scientists are very smart and dedicated to their research goals. More often, they change how they *present* their original goals to fit into whatever hot topic is currently getting more funding from the federal government. That means the government virtually never realizes the intended purpose of targeted funding for specific topics, while also forcing scientists to go through more contortions to keep their labs going.
Also, a correction: you state "most of the other half goes through the National Institutes of Health (NIH) and the National Science Foundation (NSF)." In fact, the Department of Energy Office of Science has a budget similar in size to the NSF. See https://www.aip.org/fyi/federal-science-budget-tracker. It's quite harmful to physical science research that the DOE is frequently left out of popular conversations about science funding.
This was very informative, thank you. I took a different approach myself, largely based on the lottery approach you mention. I would love your thoughts on this assay of mine:
https://sebastianquezada.substack.com/p/on-the-value-of-investment-in-science
Only someone who has zero familiarity with the current academic research environment would suggest that the problem is *too much* external oversight of research directions and *not enough* researcher choice of research direction. The real problems are (1) Far too much mediocre research, coupled with a bean-counting approach to research output, driving even talented researchers to produce incremental "LPUs" as fast as they can, at the expense of quality; (2) an incestuous academic-political environment in which credit and funding is granted based on political power and mutual log-rolling, with skilled politicians and self-publicists being able to generate prestige (and funds) for their own and each other's worthless make-work research while starving rivals; and (3) the fox-guarding-the-chicken-coop approach of funding agencies "renting" researchers to judge their colleagues' worthiness on a rotating basis, ensuring that no external oversight of large research fields is possible.
The fundamental problem, of course, is that because research is inherently speculative, it's difficult to measure the relative (or absolute) value of a given research agenda, so various proxy measures are substituted, all of them woefully inadequate and easily corrupted. An alternative approach would be to reward results rather than fund research directly--i.e., award prizes to researchers (or their employers) based on the impact of their work. Research labs could then fund researchers whose potential they believe in as investments, hoping to profit as a result. Among the benefits: funders would have a stake in funding (only) promising research with a real chance of success; researchers would be encouraged to make higher-risk, higher-reward bets--something that faces massive disincentives under the current system; and the influence of petty politics among research communities would be overshadowed by the effects of concrete achievements.
Would funding research at the state level reduce the costs of competition for funding, and possibly change incentives around vetomanship?
China invests 2.5% of its $25 Tn economy in STEM research.
We invest half that.
China has 300,000 people with 160 IQs.
We have 10,000.
Before we do anything with our science funding, we should study China.
Chinese research ranked as high as or higher than U.S. work in the top 1% of scientific studies in 2019. This work is considered the most notable published science. https://news.osu.edu/analysis-suggests-china-has-passed-us-on-one-research-measure/
And in quantity and breadth, it's no contest.
This is an excellent proposal. It mirrors school choice driven policy - fund the student, not the school. I don't think the research university vs teaching university dichotomy need be the weakness you suggest; in many ways, that distinction is artificially created by the current research funding system. Instead, simply fund everyone who has a PhD in a given field, with the 60/40 stipulation in funding allocation, and let them go where they please. One's prize for obtaining a PhD would then be lifetime funding.
That opens its own gaming strategy of course - there would need to be a considerable degree of policing to prevent fake doctorates from parasitically crowding the system. But, that could be dealt with by allocating funding by dividing the pot by the number of PhDs. The disciplines would then be strongly motivated to limit the number of entrants.
One thing I'm surprised you didn't mention, in terms of rent-seeking that drains resources towards unproductive activities, is F&A. Overheads are a huge unnecessary cost, which additionally drive much of the administrative bloat that is in turn largely responsible for the totalitarian climate of the modern campus. Eliminating overheads would be a massive win in itself. Your proposal naturally leans in that direction: if a given researcher's funding is capped, and he can go where he pleases, organizations would compete to offer the lowest overheads possible.