We have engaged in a variety of work over the years on the economics of data and information goods. We’ve been particularly interested in the question of mechanism design to incentivize data generation and sharing. As we’ve approached the problem of mechanism design in a number of contexts now, we’re consolidating the outputs of our work as a first step to creating a general guide to mechanism design in the data economy.
When it comes to mechanism design, particularly in the context of data, it can be easy to overcomplicate things. This is complex, technical stuff and the solutions should be too, right? Well, not really. Sure, the fine detail needs a degree of technical proficiency and expertise but really, if the basics of your model are unintelligible to a competent adult then you’ve likely missed the point somewhere along the way. As much as shiny new and “innovative” models are alluring, there really aren’t any magic bullets when it comes to incentives. Getting people to do things they wouldn’t otherwise has been one of the fundamental questions of economics and politics since the dawn of civilization, and even as technical complexity of our societies has increased the basic principles remain fairly commonsensical. We hope this brief summary can show that there are only so many ways to tackle this problem, and that while they can all be tweaked and varied depending on context any claims to be able to bypass or innovate past these fundamental principles are likely just noise.
So, what exactly are the potential ways of incentivizing the creation and generation of data?
Here is the summary of models we outlined in two of our major projects in the area: iMed, which focused on pharmaceutical innovation, and our work on environmental data sharing for the EU Green Deal. In each case we’ve linked our final reports along with summaries of the models they reference. As you can see, there are lots of overlaps despite the differing applications.
iMed Horizon Scan: Grants, Monopoly Rights, Prizes and Remuneration Rights
The iMed Horizon Scan report outlined 4 models for funding medical R&D:
- Grants: unconditional, upfront funding for research.
- Monopoly rights: where an innovator may obtain a temporary monopoly (a patent) for a new discovery (such as a drug), and which allows them to set prices at the profit maximizing point rather than health impact maximizing point or linked to cost of research and manufacture.
- Prizes: pre-defined payments for specific one-off innovations or interim achievement, for example a new vaccine for HIV, or a new drug for Hepatitis C. Often, prizes require the winner to make their discovery public and open.
- Remuneration rights: where innovators receive a “remuneration right” entitling them to payment from a dedicated fund on a predefined basis related to desired outcomes (or, possibly outputs) – for example proportional to the number of lives saved by their innovation relative to all other innovations covered by the fund. In return, innovators make their discoveries public and open.
The report includes the following table, giving more detail on these models:
|Mechanism||Description||Example proposal types|
|Grants||A mechanism which funds medical R&D through unconditional, upfront funding. Grants are unconditional in that there is no direct conditioning of payment on outputs or outcomes, i.e. the researcher is paid irrespective of what their research produces. They are upfront in that the researcher is paid or promised payment prior to the work being done. Almost all university research is grant-based, and grants account for a good portion of all research funding and the majority of funding for basic research.|
|Monopoly rights||A mechanism which funds medical R&D by granting innovators a temporary monopoly (a patent) for a new discovery (such as a drug), and which allows them to set prices at the profit maximising point rather than health impact maximising point or linked to cost of research and manufacture. There are many small variations that can be made to improve an essentially monopoly rights based mechanism. We have included all such proposals under monopoly rights.|
|Prizes||A mechanism which funds medical R&D through pre-defined payments for specific one-off innovations or interim achievement, for example a new vaccine for HIV, or a new drug for Hepatitis C.|
|Remuneration rights||A mechanism which funds medical R&D through the allocation of payments upon creation of a product. Innovators receive these payments according to some set of principles or conditions based on outcomes, usually including health impact and access provision. In return, innovators make their discoveries public and open.||Prize funds|
Report on Incentivizing Environmental Data Sharing: Eight Models
Our report at Life Itself Labs on incentivizing environmental data sharing is a little more extensive with 7 models, but as you can see there are many similarities with the above list:
- User-specific licensing: custom access licenses negotiated on a case by case basis, creating incentives through fees (comparable with monopoly rights outlined above).
- Subscription based preemptive licensing: paid subscriptions which carry predetermined access and use conditions, creating incentives through fees.
- Mandating/Regulation: incentivizing data sharing by making withholding costly, either through regulation or making data sharing a condition of trade
- Public grants and subsidies: upfront unconditional payments which incentivize data sharing by “covering” the associated costs.
- Prizes: data sharers are incentivized by financial remuneration for achieving some specific outcome, such as devising improved estimates of the environmental stage of a given area
- Remuneration rights: data sharers are incentivized by gaining entitled to a share of a central funding pool, which is allocated based on a set of outcome criteria
- Data commons models: data sharers contribute to a central pool of data which functions as a public good, and without receiving “hard” value in return
- Complementary goods/services models: data sharers are incentivized to voluntarily share data as this enables them to sell goods or services which are complementary to the data
Below are illustrations used in a prior workshop for how 5 of these models function in practice:
Non-fonancial value exchange