I'm posting this on pre-ideas because it needs a little ironing out.

The problem? Large, collaborative, ongoing efforts to provide public goods - such as open source projects - are underfunded.

There have been several proposed attempts to remedy this. The "street performer protocol" is the suggestion that you solicit pledges of donations, then produce the good when the pledges add up to a sufficient amount. This has been implemented in the form of crowdfunding, yet it is not perfect. There are not always enough incentives to donate, and it encourages competition instead of cooperation to produce the good.

There have been proposals to fix both of these problems. To encourage donors, there is the "rational street performer protocol", where donors encourage each other by making their pledged donations conditional on a sliding scale; and "dominant assurance contracts", where (a) centralized entity(ies) take(s) a profit on contracts and uses some of it to pay "consolation" money to those who make pledges which go unfulfilled. Yet both of these fail to give any incentive to donate retroactively - after the good already in the public domain.

To encourage cooperation, there is the "wall street performer protocol". This is an attempt to help organize individual producers to cooperate by creating a futures market in the pledges - the holders of the futures contracts would be motivated to promote any productive cooperation. However, in my mind, the weakness of this proposal is the initial allocation of the futures contracts to a single party. It is true that this party will be motivated to help (financially or otherwise) the production of the good, but they could also just sit back and wait. Moreover, any system which is perceived as unfair will actually discourage altruistic contributions.

My idea is to combine these into a single "public goods marketplace". The goods in question are assumed to have the following characteristics:

- Once they exist, they can provide benefits to a large population (for instance, for an open source software product, the users).

- It is impractical or undesirable to exclude people from these benefits.

- Some of them can be improved over time, or need maintenance. This improvement or maintenance should be funded too.

The market would have the following characteristics:

- Each good would be covered by one "project", which could be divided into any number of "subprojects".

- Each subproject or project would have "shares". Subproject shares would also count as shares in the containing project.

- Donors could make either direct donations, general pledges, or subproject pledges to a given project.

- Direct donations and general pledges would be divided among project "shareholders" by number of shares.

- General pledges would be payable when a project reached its next generally-agreed-upon quality milestone. Milestones would be set to be attainable; thus, a mature project's milestone could be something like "still has at least 10K users on date X 6 months from now"

- Specific pledges would be payable when specific features were attained. A portion of the money would go only to shareholders in that subproject, and another portion would go to the project as a whole. The relative proportions would depend on the amount of money that has gone to each; a small specific pledge on an immature main project would go largely to the subproject, whereas a large specific pledge on a mature main project would go as much as 2/3 to the main project.

- Pledges would expire after a certain time. Donors who had their pledges expire would get a "consolation prize". The size of the "consolation prize" would depend on how likely it had been at the time of donation. Highly funded, easy projects would have large "consolation prizes"; underfunded, hard ones would have nominal "consolation prizes", which could come in the form of privileges rather than money.

- Each project could determine its own form of share allocation, and whichever model was most successful would win out. Democracy and equity would be desirable aspects in these schemes. For instance, a project could assign 1 share for each hours' work in contribution, with extraordinary contributions rewarded with (limited numbers of) extra shares, and then reduce the value these shares by 2% for every dollar they paid out (so that later contributors did not have their shares too diluted). Note that, for open source, marketing and documentation are contributions as important as coding. (Many details to work out)

- Shares would be transferrable, allowing shareholders to "cash out" at any time. Note that speculators would still have an interest in encouraging cooperation.

Some mechanics:

- There would be ways to pay shareholders irrespective of geography.

- There would of course be widgets to get donations and pledges, which people could put on any web page. Pledges would be kept in escrow if the pledger were not fully trusted for some reason. Pledge amounts would be public knowledge, though of course they could be made anonymously.

- The house would take a cut, both for profit and to provide for "consolation prizes". This cut would be kept minimal and announced up-front.

- Existing software projects could be covered, and they would have to work out share allocation. To keep things from getting too rancorous, these "retroactive shares" would be of a relatively small number of "hours" relative to the actual investment in a project, and would go to a relatively small number of universally respected project members (basically, the "founder" and "maintainer".

- This would work best if hosted by an existing, respected open-source company, or, even better, by a coalition of them. Obviously, there are "network effects".

Posted: September 12, 2008, 4:35 pm

Actually, I misspoke a bit. "Dominant assurance contracts" do not pay a consolation prize if the good is not produced; they only provide the consolation prize if the pledges do not reach a predetermined "sufficient" amount. This avoids the perverse incentive to hamper success.

To generalize to the case of ongoing donations rather than pledges: you could offer double-money-back prize to anyone if the month starting with (and including) their donation has under a third of the donations of the month prior to their donation. This would encourage people to keep donating, with the assurance that they will not be the only (last) one to do so. Altruism still plays a part, but it is supplemented by rationality (and by a gambler's mentality).

Posted: September 12, 2008, 6:30 pm