1

Suppose we have a DAO that wants to issue “debt,” which we will define as (i) a claim that must be paid and if it isn’t the the debt claimants have a right to force an event called “bankruptcy”, (ii) in the event of a “a bankruptcy” this claim must be paid before “equity.” (Note that this notion of debt is distinct from what is currently contemplated by others; most other notions of debt are implicitly secured debt collateralized by a financial asset and here we are contemplating an unsecured debt claim). Here “equity” is understood to be the token that allocated the resource being created by the DAO and with which the “management team” (eg, keepers) are compensated.

The following are several “attacks” that occur in “normal life” corporate settings that have been fixed by making them illegal, but do not appear to have a tractable solution in decentralized governance. I want to discuss how these attacks can be avoided:

1) Frauds to the Estate: suppose this DAO has a “fee” that is used to pay for costs of goods sold that cannot readily be handled by one party. An example is a large advertising budget or fees to use to pay an external vendor. Once we allow the DAO to make decisions on spending, how do we ensure that keepers do not defraud the DAO by spending on themselves and not things that benefit the DAO? Note that token ownership is never sufficient to stop this. (If I buy a $1mm jet with company and own 1% is the company that bought me the jet, then I just got a jet for 99% off pretty much.) Moreover, the law is never sufficient to stop this (if it is truly decentralized, we cannot ensure the fraudster is in any particular legal jurisdiction nor is it clear how this case would be prosecuted).

2) Value Leakage to Equity: there are multiple ways to conduct this attack. One way is that equity holders can capitalize the company with 50% debt and thereby raise a lot of funds. then they rewrite the credit agreements of the company such that they are now allowed them to pay themselves a special dividend of all the cash they raised. Then they file for bankruptcy and the equity token and debt ends up worthless, but the equity holders got the cash from their special dividend and thereby screwed creditors.

3) Existential Effects of Distress: normal corporations have value either because it will issue or currently issues dividends, because it can be liquidated for its underlying assets or because it will be bought by another company (which will realize value in one of the two previous ways). In bankruptcy, distress will force equity holders to get none of the value and all the value goes to the more secure of creditors. It’s difficult to see how this would work with tokens, but I can imagine a couple ways to create a similar property for a DAO. Curious to hear what others have thought of.

  • Could you elaborate on DAO in this context? Give an example of a DAO and the rules defining it? – Sentinel Feb 22 '18 at 7:26
  • @Sentinel suppose we create a decentralized file hosting service where the coin that allocates hosting is the equity-like token that also creates governance rights. My understanding is this is how filecoin works (I could be wrong), so if you want to assume we are filecoin that’s fine. Suppose Dropbox sues the entity, asks for money as damages and wins (let’s pretend this is enforceable somehow). One way to pay is to take out corporate debt. If you are filecoin, it does not make sense to create secured debt. My question is how to we create a trustless debt that solves attacks above. – griggah Feb 22 '18 at 13:17
  • To explain for example, whenever filecoin tries to pay for the legal fees, somebody now has to be trusted. How do we ensure they don’t just transfer money to fund a private jet? How do we ensure the filecoin token holders don’t rewrite the protocol such that they use the cash raised by the debt to pay themselves then immediately liquidate (which in many cases could work in defrauding folks). Etc – griggah Feb 22 '18 at 13:23
  • I am having some trouble wrapping my head around this scenario. One thing is suing a legal entity that operates the filecoin-like smart contract, another thing is attempting to sue the smart contract. I don't see how that latter case is possible without legal infrastructure already on chain, and the former is nothing new. – Sentinel Feb 22 '18 at 16:23
  • I think I must be missing something here. Can you clarify further? – Sentinel Feb 22 '18 at 16:25
1

If I understand the question correctly, I would say that what you are looking for is on-chain legal infrastructure.

I think that what you might be looking into, or even creating, is a seminal legal infrastructure that would be defined as a meta-protocol.

So for example, if ERC20 is a higher level protocol that defines how tokens can behave and interoperate, what you might be looking to define is a general protocol for 'promises' or 'futures' that smart contracts (like ERC20) can implement.

Then, if a smart contract follows those norms, it can provably follow promises. For example, let us define a class of smart contracts called "US Government Smart Contracts" An instance of such contract might be called "US Financial Crimes Contract". All other contracts might implement an interface called "Cashflow" that your new protocol recognizes. A "suable" contract in your new protocol will be one in which the contract allows itself to divert cashflow to the Crimes Contract, if the Crimes Contract says it must.

In short, I think what you are looking for is infrastructure that allows this. The smart contracts would voluntarily subscribe to that infrastructure. So, those that avoid the possibility of corporate attacks, might also suffer from not being able to attack, and not be able to exploit protection from the US government. In a way, the government infrastructure on chain would have to be useful, or it won't get created.

  • I think that’s a the beginnings of a good answer. I agree that a meta-protocol is a necessary level of abstraction and that a lot of work needs to be done to create that. In any tractable solution, you’d need a very detailed set of standards that folks understand that create what would otherwise normally be in corporate law. The question is whether that type of meta protocol could truly be constructed. Id need to put a lot of thought into how to do that. – griggah Feb 22 '18 at 20:56
  • I suspect that the scope is so huge, it is essentially redesigning the world into a decentralized, auto-audited, ledger based new one entirely. I suppose the way forward would be to pick a niche and get funded. It seems the money is there and a lot of will to build this new world is floating around. Agreed, this is only a nugget of a thought of an answer. Needs brainstorming and refinement. – Sentinel Feb 22 '18 at 21:05
  • @griggah A thought. There is the Enterprise Ethereum Alliance. Perhaps you could contact them and help them with the state/legal side. – Sentinel Feb 22 '18 at 21:07
  • @griggah One more thought. It might be the case that the whole concept of debt and the scenarios you describe become obsolete. I think though that the notion of lending against guaranteed percentages of future cashflow would, if implemented as a protocol, become the key ingredient in any future concept of unsecured debt, insofar as 'unsecured' has meaning when the 'cash' is basically as much cash as tangible asset (cryptocoin). Futures, derivatives, etc., could all be based on top of it maybe – Sentinel Feb 22 '18 at 21:16
  • thanks all good ideas. I agree that a tractable solution will take a lot of work. I think a lack of fundraising is not a problem in this space; optimal capital structure and having debt is not the issue. There’s far simpler problems now that need to be fixed first. I’ll look at the enterprise ethereum alliance. I don’t think debt and equity will ever be done away with, but I think we can improve governance a great deal. – griggah Feb 22 '18 at 23:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.