A Market for Papers

Republished from https://andrewjudson.com/projects/2021/03/08/papers.html

Last year I did some exploratory work on creating a market for scientific papers. This market would be implemented by issuing a sellable NFT to the authors of a paper every time a paper was published. The goal of this project was to provide a viable additional funding source for researchers, and to create an incentive for individuals to seek out novel research that was undervalued and support it. Scientific funding can be considered as a sort of portfolio optimization problem - you have a constraint in terms of how many resources are available, and you want to allocate it to different researchers in order to maximize a utility function. This can depend on the skills of the specific researchers, the “value” of their research field to you, and the nonlinearity of capital in scientific research output - for example, giving one university a professorship and a nuclear reactor may be worth more than funding 10 professorships and spreading them out to other universities. The hope was that the investors in this market would help optimize funding by distributing their money in a way that is closer to the solution to the optimization problem than the current grant giving institutions.

I abandoned this project for three reasons. The first was that I could not figure out a way in which this market would be viable beyond just as a novelty where people paid a lot for famous papers. There are always additional papers being produced, so there is inflationary pressure that needs to be counteracted for people to want to invest. Most papers would end up being worthless, and so people would invest in known winners - defeating the whole point of the market. The second reason is that I was unconvinced that the value of papers in this market would track their actual causal impact. Papers are not the only thing that spurs innovation (consider informal conversation and mentoring) and bibliometrics are notoriously flawed. The existing institutions have failed to track this sort of thing - why should I expect investors in the market who share the same human biases to do better? The third and most important reason was that through my research I came to believe that such a market would have a negative impact on the overall scientific ecosystem. The book “Science-mart” by Mirowski details the problems that adding markets has brought to scientific research. The main thesis is that scientific output in America has gone down since we changed our funding model from the Cold War version (command and control, force sharing between competitors in order to defeat the Soviets) to the post Cold War version (Reaganomics, add markets for everything). Adding patents, for example, created additional friction because scientists need to get permission from the patent holder to use their IP in research. The holder may not grant permission, or they may on condition that the scientist gives them a cut of what they produce, and there is a whole ecosystem of university bureaucrats that now exist just to navigate patents. The end result is that even if the researcher gets permission, there is a huge additional friction that slows them down. Creating a market for papers would do something similar, and I believe any slight gain in funding optimization would be counteracted by the viral nature of markets in science that add additional friction and frustration for researchers.

Given the recent explosion in popularity of NFTs, I believe it is inevitable that someone tries to implement this. I also believe it will fail due to the inflationary aspect of paper publishing and revert to just a novelty that has no effect on the funding system. In this diminished capacity, I hope that it does not have the same negative effects that patents have.