Quadratic funding favours projects that are voted on by the greatest number of people. High-risk high-reward projects are less likely to be funded in a model where funding is dependent on consensus from the most number of people. It is far worse to fund projects that seem obvious to work on, then to not fund high-risk projects that can shift the way humanity operates due to risk tolerance.
Highlighted in this article is a description of the current state of funding, quadratic funding models, and how complementing existing funding models with a retroactive results oracle will increase probability of paradigm shifting work.
Current State of Funding
Science and technology remains to be heavily funded by academic institutions and government agencies. Other mechanisms of capital allocation that have become more popular in funding science include venture funding, corporate agencies, non-profits, family offices, and alternative science agencies.
Many capital allocators consider themselves to be “principled decision makers”, when really the model of funding projects is highly probabilistic and largely independent of the funding body for future success. Teams who understand the probabilistic nature of funding science and are motivated to push the frontiers of human progress, have adopted a higher risk tolerance in the projects they fund.
The concept of “high-risk high-reward” is a funding framework that many Advanced Research Project Agencies (ARPAs) use. Some sample organizations that use high-risk high-reward models include DARPA, ARPA-E, NIH, NSF and most alternative science agencies (ex. Astera, Convergent, & Actuate). Early stage venture funds take a very similar approach to funding, whereby the expected return of a portfolio is potentially positive even if 1/n startups succeed (n >> 1). It is the appetite for this type of risk tolerance that has led to the development of some of the most useful technologies humans interact with today.
Constant amongst all of the mechanisms of funding science highlighted above is centralized decision making. When it comes to funding highly risky projects, having a controlled, centralized group of people with decision making power that does not require mass consensus, increases the probability of funding riskier projects and expediting the time to fund them.
What would a model of decentralized funding early stage projects look like? What happens when you have a pool of capital that is deployed at the discretion of everyone involved in the organization? How can we ensure we are still funding high-risk high-reward projects?
Current Decentralized Public Good Funding
Decentralized public good protocols raise money that are deployed toward project proposals that are submitted to the network. In a decentralized funding model, it is no longer the responsibility of just the core team to decide how money is allocated, but rather all of the individuals engaged in the network that can impact how projects are funded.
The mechanisms by which protocols “raise” money is through inflationary rewards, philanthropy, NFT sales, sequencer fees, transaction fees, or general protocol wealth through other means of revenue. The ways in which these funds are distributed can vary upon the type of decentralized funding mechanism used: quadratic funding; retroactive public goods funding; MolochDAO; Web of Trust; Universal Basic Income; Aqueduct.
Some notable projects that have utilized the growth of their network to fund other projects include: Ethereum Foundation; MetaCartel; MolachDAO; OpenGrants; The Graph; Uniswap grants program; Balancer; mStable.
Of the funding distribution mechanisms commented on earlier, quadratic funding has become one of the most popular in the Ethereum ecosystem with Gitcoin grants dominating the funding of public good projects. Gitcoin has a pool of capital that it is ready to distribute towards projects that are submitted to the network. Individuals submit proposals to Gitcoin and other users of the network can endorse or donate towards submitted projects. There is a non-linear relationship towards the number of endorsements a project gets and the amount of funds that Gitcoin will match from its pooled reserve. Another project that utilizes quadratic funding is clr.fund.
This model of funding is novel and enables funding projects to be done in a completely permission-less, global, immutable, open, and transparent way. Quadratic funding rewards projects that have a high endorsement from the network users and ensures that the matched funds are deployed with the best interest of the most people in the network. This aligns very well with the ethos of decentralized treasury management and deployment.
Problems with current model of decentralized funding
Some of the largest funding agencies in the world have transitioned from consensus based funding to probabilistic funding models that have resulted in an increased risk tolerance due to a fundamental understanding of what it takes to fund projects that will push forward progress.
From a design perspective, funding via quadratic methods are not well aligned to fund highly risky research that often leads to paradigm shifting outcomes for humanity. If we are going to deploy capital via decentralized models, we can not be reliant exclusively on quadratic methods to do so.
In image 1, the y-axis represent the likelihood of endorsement from the public and the x-axis represents risk of projects based on probability of success. Quadratic funding models favour non-risky projects as it is in the best interest for donors to fund non-risky projects for expected individual future return. As such, risk projects, which often require higher support of funding are undervalued in a quadratic funded model.
Funding High Risk Project Via Retroactive Models
As Vitalik Buterin noted in his article, “it is far easier to agree on what was useful versus what is going to be useful”. A strong retroactive funding model enables builders to work on creating meaningful projects and are rewarded if the project is successful.
In the inception stage of a project, they can be funded by any decentralized distribution mechanism: quadratic, donation etc. However, upon contribution of inception donations, contributors are made aware of “results oracles” that protocols have created to measure utility of a project and rewards projects retroactively for their contribution [Vitalik]. The analogy to traditional venture markets is the results oracle rewarding a startup for exiting.
This model does not have to be exclusive for public good projects that are funded on donations / grants. A similar model can be applied to decentralized investing protocols where consensus is needed to be made on where funds are to be deployed. In an investing setting, a protocol can incubate projects in a very similar manner they do now with a portion of the capital pooled in their or other teams results oracles that will retroactively reward useful projects.
How does this benefit highly risk projects?
When an individual is making a contribution to a proposal, they no longer need to take into account the mutual contribution made from peers in the inception stage of a project but rather the utility a project will have upon success, measured in a results oracle. This will change the way individuals vote on deployment of capital and become more risk tolerant as the expected return is far higher with the support of the results protocol.
Next
Any team focused on decentralized public good funding should not only have a funnel to incubate projects, but also think about the results oracle that can shift the risk distribution of projects funded. If we are truly trying to solve coordination failures like climate change, insecure digital infrastructure, healthcare, we need to be funding projects that can truly shift the way we think about this. Often, these projects are very risky and current decentralized funding models do not favour their proposals.
Projects that are able to incorporate this two sided funding model will be able to create a global, immutable, transparent, and programmable funding model that leads to paradigm shifting innovations.