Tragedy Of The Commons

[vs77bb]

Decentralized Public Goods Incentivization

Forked directly from Ethereum Foundation - Problems (old).

Problem: Come up with and implement methods for incentivizing public goods production in a decentralized environment.

One of the challenges in economic systems in general is the problem of "public goods". For example, suppose that there is a scientific research project which will cost $1 million to complete, and it is known that if it is completed the resulting research will save one million people $5 each. In total, the social benefit is clear: if everyone contributes $1, then each individual person will see a benefit of $5 - $1 = $4 for $4 million total. However, the problem is that from the point of view of each individual person contributing does not make sense - whether or not you contribute has close to zero bearing on whether enough money will be collected, so everyone has the incentive to sit out and let everyone else throw their money in, with the result that no one does.

So far, most problems to public goods have involved centralization; some large organization, whether a big company or a government, agrees to offer some of its private services only to those individuals who participate in paying for the public good. Often this is done implicitly: for example, some of the money from each purchase of an iPad goes toward research and development (some of which is a public good, and some of which is an excludable "club good"). At other times, it's more explicit, as in the case of taxation. In order for decentralized economic systems (we'll refer to decentralized economic systems that somehow rely on cryptography and/or cryptocurrency as "cryptoeconomic systems") to be effective, ways of incentivizing production of public goods relevant to that system are required. A few possible approaches include:

• Assurance contracts - the idea behind an assurance contract is that N people may or may not put their funds into a pool, where that pool pays to produce a public good if and only if at least $X in total is contributed. Otherwise, the pool pays everyone back. If the pool creator acts optimally, the tipping point will be right at the top of the bell curve that is the probability distribution for how much other people might contribute, meaning that the chance that one user with their contribution of X/N will be pivotal should, by central limit theorem, approach ~1/sqrt(N), creating a sqrt(N)-sized amplifying effect on their donation.

• Dominant assurance contracts - a special type of assurance contract, called a dominant assurance contract, involves an entrepreneur that pays all contributors back slightly more than 100% of what they put in if the fund fails to reach its target (and takes profits if the fund succeeds); this provides an incentive for someone to create optimally targeted assurance contracts.

• Currency issuance - a cryptoeconomic system can contain its own currency or token system which is somehow necessary or useful in some part of the system. These currency units can then either be generated by the system and then sold or directly assigned to reward contribution. This approach gets around the free-rider problem because no one needs to pay the $1 explicitly; the value arises out of the emergent value of the network which is does not cost people to support.
  Status goods issuance - a status good can be defined as a good that confers only relative benefit to its holder and not absolute benefit to society; for example, you may stand out in the public if you wear an expensive diamond necklace, but if everyone could trivially obtain such a necklace the situation would be very similar to a world with no diamond necklaces at all. A cryptoeconomic system can release its own status goods, and then sell or award them. One example of a status good is a "badge"; some online forums, for example, show a special badge beside users that have contributed funds to support the forum's development and maintenance. Another important example of a status good is a namespace; for example, a decentralized messaging protocol may be able to fund itself by selling off all of the 1-4 letter usernames.

• Recursive rewarding - this is in some ways a mirror image of the concept of "recursive punishment" that arguably underlies a large number of social protocols. For example, consider the case of tax-funded police forces. In natural circumstances, there often arise opportunities to take actions which are beneficial to the perpetrator, but ultimately harmful to society as a whole (eg. theft). The most common solution to this problem is punishment - an act which is harmful in itself, but which shifts the incentives so that attacking is no longer beneficial to the perpetrator. However, there is a problem: there is no incentive to participate in the punishment process. This is solved by making punishment obligatory, with non-participation (in modern society by paying taxes) itself punishable by the same mechanism. Recursive rewarding is a mirror image of this strategy: here, we reward a desirable action, and people who participate in the rewarding mechanism (eg. by giving reward recipients a discount in shops) are themselves to be rewarded.

Many of these approached can arguably be done in concert, or even simultaneously within one mechanism.

Additional Assumptions And Requirements

• A fully trustworthy oracle exists for determining whether or not a certain public good task has been completed (in reality this is false, but this is the domain of another problem)
• The agents involved can be a combination of individual humans, teams of humans, AIs, simple software programs and decentralized cryptographic entities
• A certain degree of cultural filtering or conditioning may be required for the mechanism to work, but this should be as small as possible
• No reliance on trusted parties or centralized parties should be required. Where some kind of "supernode" role does exist, the protocol should provide a way for anyone to participate in that function with a mechanism for rewarding those who do it well
• The mechanism should ideally be able to handle both public goods which everyone values and public goods which are only valued by a small portion of the population (eg. the production of a freely available book or video on a specific topic)