DAOs (decentralised autonomous organisations) are online-native decentralised organisations with self-enforcing rules, governed by their members, and use distributed ledger technology, such as blockchain. DAOs use distributed ledger technology, and often use third-party blockchain platforms, such as Ethereum.
Most DAOs run on a third party’s blockchain, such as Ethereum, but some DAOs, such as the Dash DAO run on their own blockchain. Thus DAOs are certainly not limited to Ethereum. Currently, while most DAOs control digital assets, they will increasingly also be used to control real world assets.
A DAO’s rules are contained in the DAO’s software (its computer code), in effect smart contracts. The members of DAOs can be people (humans) and/or other organisations, including other DAOs. The members may be anonymous or identifiable or a combination of both. People become members of a DAO by owning the DAO’s tokens (a form of cryptocurrency).
Decentralised organizations are not new. But DAOs are a new type of decentralised organisation and are different to what has gone before. As Adam Greenfield explained in Radical Technologies: The Design of Everyday Life (at pages 161-162) “[T]he DAO is that genuine rarity: a new thing upon the Earth, something that really could not have been conceptualized before the technologies underlying it were in place”. In this case the technologies are blockchain/decentralised ledger technologies.
Why could DAOs not have existed prior to blockchain?
The use of blockchain is key to DAOs because blockchain is being used to create organisations that must run according to their rules and those rules are self-enforcing.
The following example explains the difference between DAOs and traditional organisations. Take a company with a rule that the CEO has authority to enter into transactions up to $50,000 without seeking further approval. If the CEO ignored that rule (or did not realise the rule existed) and purchased a luxury car worth $100,000, the company may be bound by that contract and required to pay the $100,000. The company cannot simply refuse to pay by saying that one of its internal rules was broken (whether the company would be bound by that contract would depend on a number of factors too long to go into here). Rules and also law do not prevent wrongdoing from occurring, all they can do is provide a framework for determining who is liable/accountable after the wrongdoing has occurred. And that is assuming that the wrongdoing is noticed, and someone has the time and resources to attempt to hold the wrongdoer to account. Even if the wrongdoer is held to account it is often too late and the victim cannot be restored to the state they would have been in had the wrongdoing not occurred.
In contrast, a DAO is unable to operate outside of its rules (its code/smart contracts). If a person within a DAO (and DAOs are unlikely to have CEOs), attempted to purchase something not permitted by its code they would be unable to do so. While this may feature may appear relatively minor it is a paradigm shift. No longer do checks have to be made that rules and/or laws are being followed. Simply, an action cannot be done if it is not permitted by the DAO’s code. That is ex-ante limitations can be imposed, instead of the normal ex-post monitoring and enforcement. However, because a DAO can only do that for which it is programmed, if something unforeseen occurs the DAO is unable to act, unless its rules are changed. In extreme situations an unforeseen event or a bug in the DAOs code may mean that token holders lose control of the DAO and its assets. In contrast, with a traditional organisation, if the organisation’s rules do not apply in a given situation that does not prevent the organisation from taking a course of action.
What does “decentralized” mean?
As the use of “decentralized” suggests, there is no CEO or equivalent calling the shots. And, unlike many organisations where permission is required to become a member, with most DAOs there is no need to apply for membership. It can be as easy as purchasing tokens from a cryptocurrency exchange or even receiving tokens of a DAO through an air drop [cross reference] without even realising it. There are exceptions, however, as some DAOs require an application to become a member.
It might be thought that “decentralized” means that every member of a DAO has an equal say in the governance (decision making) of the DAO. The amount of decentralisation, however, depends on the individual DAO. If a DAO was truly decentralised it would allow, for example, each member to make suggestions (proposals) about what the DAO should do and how assets are used. In turn every member would be entitled to take part in the discussions and ultimate decision about which proposals to adopt. Few are as decentralised as this. There may be limits on who and how proposals are made, and different people may have a much greater say in how the DAO is operated. DAOs therefore operate along a continuum from radically decentralised through to more centralised.
DAOs are not all radically decentralised because decision making is difficult at the best of times. That is why in most organisations, decisions are normally made by a small group and/or by one person. Occasionally the wider membership makes decisions, but that happens rarely and often only for some types of decisions required by law. In a DAO, if anyone can make an unlimited number of proposals, members may be flooded with proposals. Many, if not most, members may not have the knowledge, skills and time to properly evaluate, discuss and then vote upon each proposal. For this reason limits are often placed on how and who can make proposals that go forward for discussion (see “How can proposals be made?” below) and there may even be gatekeepers checking the proposals to ensure they are suitable. Also, because reaching a consensus is time consuming and may be impossible, DAOs normally require members to vote upon proposals. Most DAOs use voting schemes where it is one token-one vote (which is similar to one share-one vote with companies).
Another example of what decentralisation means in DAOs is that instead of a traditional organisation deciding what type of work and how much they are willing to pay for that work, and often hiring employees to carry out the work, with a DAO people pitch work (in the form of proposals) that they believe the DAO needs and the price and other conditions under which they will perform the work. The token holders therefore are presented with a range of different proposals for things they never asked for and have to work out whether those things are needed at the price being offered. Also they need to assess whether the proposers have the ability to carry out what they have promised.
If addition, if decisions need to be made relatively quickly, for example, there is a bug which requires urgent fixing, it is not likely to be feasible for a proposal for the fix to be created and then wait for weeks for the voting to occur. Thus some DAOs can have a small technical or other group able to make urgent changes to the DAO’s code. While allowing a small group to make unilateral changes is arguably counter to decentralisation, it is a practical mechanism and without it a DAO could cease to operate and/or become ungovernable and its assets frozen forever.
What does “autonomous” mean?
The term “autonomous” is misleading. DAOs are owned and controlled by its members (people and/or other organisations) and those members decide how the DAO operates. Entities that make their own decisions, such as an autonomous car that decides who to pick up and books its own servicing etc, are better called AI DAOs. To be sure, early in the DAO journey some thought that once a DAO was created and its rules set, the DAO would run autonomously according to those rules and from further human intervention. However, it is extremely difficult, if not impossible, to predict and code for every eventuality. Changes in rules are often necessary, especially if there are errors in the code. An example of an error in a DAO’s code was seen early in the development of DAOs with The DAO hack (see “The DAO Hack” below).
Autonomous, in terms of a DAO, means that, depending on the DAO, once a decision is made through on-chain voting it is implemented automatically, no person needs to action it. For example, a proposal is made to send the equivalent of $1000 NZ in a cryptocurrency to a charity after a natural disaster and the proposal gathers sufficent votes to pass. Once the vote finishes the $1000 is sent automatically to the charity. (The proposal would contain the amount sought and the charity’s public address [cross reference] – the only way the charity would not receive the cryptocurrency would be if the DAO had insufficient cryptocurrency or the charity’s public address in the proposal was not correct so the transaction could not go through. In which case, the smart contract should be written so that the cryptocurrency in question could be accessed by the DAO and not sit stuck in the smart contract.)
Just as there is a wide range of organisations, from companies to incorporated societies to trusts, there are a wide range of DAOs. There are also a wide range of motives behind creating a DAO. While some intend their DAO to be a for-profit venture where people and/or organisations pool funds to invest, akin to a venture-capital (VC) firm, others want to create not-for-profit organisations. There have even been suggestions that a country or an area within a country be run as a DAO.
DAOs are not just limited to people who may even be anonymous creating an organisation. Increasingly companies and other entities are joining forces and instead of creating a new company or non-profit society/association they are creating a DAO to manage common resources.
Use cases for DAOs
For a good website which covers the largest DAOs, see www.deepdao.io.
The following are just a tiny fraction of the DAOs in operation:
- Running a blockchain: Dash DAO
- Charities: Moloch DAO (https://www.molochdao.com/) – gives grants for development in the Ethereum ecosystem
- DeFi]: Uniswap (Decentralised Exchange) (https://uniswap.org/), SushiSwap (https://sushi.com)
- Stablecoins: MakerDAO (https://vote.makerdao.com/)
- NFT platform: Rarible (https://rarible.com/)
- Metaverse: Decentraland (https://dao.decentraland.org/)
- Art club investing in NFTs: Flamingo DAO (https://www.flamingodao.xyz/)
- For a proposal for governing real world assets and rights, ie water, see Brett Miller, “Application of Blockchain Capabilities to the Management of Water Rights and Markets” (November 2021)
A (very) short history of DAOs
The first DAO, BitShares, was created in 2014 by Daniel Larimer. BitShares created BTS, an early form of stablecoin and a platform (exchange) to exchange BTS and other tokens. BitShares, however, became unworkable as the token holders did not have the time and/or the skills required to make decisions, in fact, fewer than ten percent of token holders took part in voting. Token holders also were wary of spending any of BitShares assets to keep the platform operational. So that decisions could be made, changes to BitShare’s governance structure were made and twelve proxy holders, voted in by token holders, were tasked with making the decisions.
The Dash DAO, created in 2015, was the next DAO of significance and at the time of writing in February 2022, it is still in operation. The Dash DAO runs the Dash blockchain. Instead of allowing any token holder to vote only masternodes only can vote. 1000 dash tokens are required to run a masternode, plus there is a requirement to provide computing resources to run the Dash blockchain. There are around 4,000 masternodes, although one person or organisation can own more than one masternode. )People can also pool their dash tokens and obtain part ownership in a masternode.) Any token holder can submit proposals upon which the masternodes can vote. To cut down the number of proposals made, a fee (payable in dash tokens) must be made.
The first high profile DAO, The DAO, was created and failed in 2016. Since The DAO many DAOs have been proposed and many created. See www.deepdao.io for a current list and information about the largest DAOs.
The DAO hack
The DAO Hack is mentioned here because The DAO was the first high profile DAO, and its rise and fall were spectacular. For those reasons The DAO remains the best known “DAO”, and no doubt its name contributes to its notoriety. Many people still believe that The DAO was the first DAO, but as we have seen above, this is not the case.
The purpose of The DAO was to create a form of a VC fund. People would contribute funds into The DAO, which would fund projects – with The DAO token holders receiving profits (if there were any) from those projects in proportion to their contributions to The DAO. People wanting funding for their projects would submit proposals containing the details of their projects. The DAO token holders were expected to evaluate the proposals and vote which ones to fund. The DAO was built on Ethereum.
Initially The DAO was successful, perhaps too successful. It raised over $US100 million in ether, a form of cryptocurrency [cross reference]. Before The DAO was able to make any distributions a “hacker” exploited a flaw its code draining over $US50 million from The DAO by moving it to another account within The DAO. However, the hacker could not abscond with the “stolen” ether because The DAO’s rules (its smart contract) contained a 28-day holding period preventing the further movement of the funds. The hacker’s exploits generated a large debate. On side argued the hacker was following The DAO’s code, so they had done nothing wrong and once the holding period elapsed the hacker should be able to take the ether sitting in the account. Others argued that the hacker’s actions were against the spirit/purpose of The DAO and the hacker should not be permitted to steal the ether. After much discussion and controversary it was agreed that the hacker should not be allowed to walk off with the ether. To prevent the theft it was decided to fork the Ethereum blockchain. (Note: DAOs using a third party’s blockchain—as was the case with The DAO—cannot simply decide to fork the underlying blockchain, the decision to fork Ethereum was made by Ethereum.) The fork saw the creation of Ethereum Classic, so there were now two blockchains: the original Ethereum and the new Ethereum Classic. On Ethereum the hack was effectively reversed, as if it never happened. With Ethereum Classic the hacker was able to abscond with the ether (now Ethereum classic) they had stolen. The forking of Ethereum meant the end of The DAO. (The decision to fork Ethereum to protect The DAO token holders was highly unusual and will likely never happen again.)
The DAO hack is often credited in setting back the development of DAOs. On balance, however, The DAO hack was positive as it exposed the frailties of DAOs and made those creating DAOs more careful about how they were designed and the value of assets controlled by DAOs.
In the introduction to this section about The DAO hack, the term “DAO” was used. This is because strictly speaking “The DAO” did not accord with the definition of a DAO as a online-native decentralised organisation with self-enforcing rules, governed by its members. If The DAOs token holders could govern and therefore control the DAO, when the hack was discovered, a proposal could have been made, voted upon, and The DAO’s code changed to prevent the hacker absconding with the ether, all without forking Ethereum and destroying The DAO. Many lessons have been learnt following the rise and fall of The DAO and one of those is that token holders must have the ability to make changes to a DAO’s code/rules. Note: the ability to make changes to a DAO’s code/rules does not mean that DAOs are immune to the often disastrous effects of errors and bugs, but it does help make them slightly less brittle.
How can proposals be made?
Some DAOs allow any token holder to make a proposal. Or to cut down the number of votes, some DAOs, such as the Dash DAO, require a fee to be paid when submitting a proposal. Some DAOs refund that fee if the proposal is successful. Other DAOs will let only certain people submit a proposal. For example, only those who hold governance tokens (if the DAO has separate governance tokens), or only those people who have achieved a sufficient level of reputation within the DAO may be able to lodge proposals. Reputation is gained through actions people perform for the DAO and it may have a decay function, ie overtime reputation decreases and a person must remain active within the DAO to maintain and grow their reputation.
Does anyone check proposals before they go to the vote?
Not all DAOs checks of proposals before they are voted upon. Not checking proposals can potentially be dangerous as they may contain errors or be malicious. For example, a proposal could provide that the DAO sends all of its assets to an external address, in effect, asset stripping the DAO. To be sure, checking of proposals runs against notions of decentralisation as elements of centralisation begin to creep in. Yet, again a balance must be struck between decentralisation and centralisation.
With some DAOs proposals may be subject to discussion and consensus is required to put the proposal up for a vote. Other DAOs have a Council or similar body which checks proposals to decide whether they should go forward to a vote by token holders. It is also possible for a Council to provide an indicative vote, ie whether they support the proposal or not, which token holders can overrule if they wish, by voting through a super majority.
All sorts of ways are being experimented with to limit the number of proposals token holders are required to vote on. For example, one way is to only allow one proposal each round to go forward and a process of game theory is used so that the best proposal goes forward. Also, to have some strategic oversight, the same DAO uses a Council and that Council puts forward a proposal of its own, so token holders have two high quality proposals to vote for.
How does voting work?
As with companies, voting is normally on the basis of one token-one vote. This is in contrast to other organisations such as cooperatives where it is normally one person-one vote and also democracies where each enrolled voter has one vote. It is also possible for some DAOs to have separate tokens and only those with governance tokens can vote. There are many variations, for example, with the Dash DAO, only masternodes are entitled to vote. One reason why one member-one vote is not often used is because that would require verification of a person/organisation’s identity and most DAOs do not want to use such a process.
Most DAOs use a bare majority for voting, ie 50% plus one is required. But there is nothing to prevent the requirement of a higher majority, say 80%. Within a DAO, higher or lower majorities can be used depending on the type of proposal. Important decisions, for example, the use of a significant percentage of the DAO’s assets, or the decision to wind the DAO up, may require a super majority of 75% in favour. Another example of different majorities can occur in the example above of a Council providing an indicative vote, if the Council was unanimously in favour of the proposal the token holders may need only need a bare majority or even lower, but if the Council was not unanimously in favour (there was a split vote), the majority required could raise depending on the split.
Voting on proposals is often left open for two or more weeks, the time will depend on the DAO.
On-chain voting means that the voting is done on the blockchain and the results of the vote are executed automatically. The opposite, off-chain voting, is when voting occurs not on the blockchain but using a tool, such as Snapshot, which allows token holders to vote and that vote is then implemented by a person or body (see this example). Off-chain voting is often used to circumvent the cost of voting in Ethereum as each vote is treated as an Ethereum transaction, with high gas costs.
Factors to think about when designing a voting scheme
Does a proposal need 50% or higher of all the DAO’s tokens cast in favour of the proposal for it to succeed? If such a quorum is required very few proposals will be successful as voting in DAOs tends to mirror voting in other areas of life, such as parliamentary elections where not everyone votes. Or is only a majority of the tokens actually cast required? If no minimum percentage of the DAO’s over all tokens is required a could succeed with only a few percent of the total DAO tokens voting in favour. For example, if a proposal attracted only 5% of the DAO tokens and of that 5%, 60% voted in favour (so less than 3% of the overall tokens), the proposal would succeed. Such a scenario could well occur if any, some or all of the following factors were present: no checks are made on proposals; there are a lot of proposals for people to consider and it is expensive and/or time consuming to vote. For DAOs using Ethereum on-chain voting can be expensive, so the level of voting is often low. For this reason it would be prudent to require at least 20% of token to be cast for a proposal.
Creative voting schemes
One token-one vote is a blunt tool. And it favours those with more tokens given them a disproportionate say in the operation of the DAO. For this reason DAOs have been exploring more creative voting schemes that token holders’ strength of preference and sometimes also time. See, for example, quadratic voting and conviction voting.
Legal Status of DAOs
Many, if not most DAOs, will be considered as general partnerships in many countries. This means that each DAO token holder is considered a partner and jointly and severally liable for the debts of the DAO, and possible, depending on the facts, they are also liable for some of the actions of their fellow DAO token holders. There are all sorts of reasons why general partnerships are not suitable for DAOs. Just one of those reasons is that each time a person or organisation becomes a DAO token holder or divests themselves of their DAO tokens, a new partnership forms and this could occur many times a day.
Some jurisdictions are allowing the registration of some forms of DAOs. For example, Wyoming, Vermont, Delaware all allow DAOs gain registration as LLCs. A Parliamentary Select Committee in Australia has recommended that Australian company law be changed to accommodate DAOs and work is being undertaken on the issue. Note, even if Australian company law is changed, it will not accommodate all DAOs in Australia as not all DAOs are for-profit. Combinations of other legal structures are also possible, for example, the Dash DAO uses a trust and a Delaware C-Corp, thus “hacking” the law.
While the registration of a DAO as a formal legal structure may go against the original conception of a DAO, practically speaking many DAOs if they control significant assets will want to gain the benefits of registration, although that require changes to the structure of the DAO. Also, once a DAO is successfully registered it will be recognised in that jurisdiction and others.
DAOs operate in very different ways, there is no one DAO structure. DAOs can be used to run and govern many different things. DAOs are very much in their infancy and there is considerable experimentation occurring. While the thought of “decentralisation” is appealing to many people, creating a decentralised organisation is difficult, even with the advantages that blockchain provides. This is especially true for traditional organisations who are seeking to create DAOs.
At the time of writing Ethereum is the most popular blockchain to build DAOs upon. The drawback, however, with using Ethereum are the costs associated with its use, especially if on-chain voting occurs.
- Who owns my DAO? by Phillippe Honigman (Hackernoon article)
- Decentralised autonomous organizations (DAOs) with Aaron Wright (podcast)
- Decentralised Autonomous Organisations: Governance, Dispute Resolution and Regulation by Alex Sims (a long read!)
About the Author