DeFi cannibalize PoS security? Increase opportunity cost and reduce risk

In a network where tokens provide voting rights (such as DAO or PoS), we can calculate the cost of borrowing the token on the secondary lending market during the voting period to determine the cost of each vote. This idea highlights the important role of time variables in the governance process, that is, the longer people need to borrow tokens to vote, the more expensive the interest paid. If someone really borrows tokens to vote, then we can use this to design a more robust governance system. The agreement cannot control the interest rate in the secondary market, but it can influence the "governance costs" by manipulating voting time.

The ecosystem of the on-chain and off-chain lending market is constantly developing, and anyone can lend or borrow tokens. Users can earn interest by lending tokens, and can also leverage (collateralize interest payments) on collateral.

However, when contemporary currency has some kind of governance power on its network, it can influence governance decisions: including agreement upgrades, community funding proposals, consensus improvements, and so on.

To explore this, this article will take ZRX tokens with 0x governance voting as an example.

There are approximately 6.5 million ZRX tokens for voting at each decision of 0x, and the price of each token is about $ 0.20, which means that the amount involved in each decision is about $ 1.3 million. However, this number is not the actual cost of reaching a decision.

On Compound, users can borrow ZRX at an annual interest rate of 3.30%, and the interest for one day is about 590 ZRX. Under similar conditions, the total cost of making a decision is about $ 120 (calculated at $ 0.2 / ZRX), and the marginal cost per ticket is $ 0.000018. Regardless of whether the user needs to borrow from DeFi, they will need to spend this money, because non-borrowed tokens also have an opportunity cost (that is, the user could have lent this money for interest).

This number is surprising because the cost of influencing decisions is very low for someone who has enough collateral (or credit). Tarun Chitra also investigated this, and he emphasized that if the benefits of lending tokens exceed the benefits of entrusted tokens, this will threaten the security of the POS system.

Tarun Chitra is the founder of Gauntlet (Gauntlet is a crypto network evaluation platform), and previously worked at the Vatic Lab quantum research and development laboratory.

However, the security of the system may be more valuable to most holders than additional benefits. This is a network characteristic, not a network error: This allows us to measure the cost of voting more accurately, because protocols and DAOs can influence the cost of governance by changing the time required for voting. This can provide an important variable for us when designing the system.

The secondary market influences voting costs through token prices and interest rates. Crypto networks themselves cannot control these factors, but they can control how long users must hold tokens. The network raises or lowers the cost of voting by affecting the cost of interest that hackers or users need to pay to borrow tokens to participate in voting (also applicable to non-borrowers in the form of opportunity costs). Part of the reason why 0x governance is so cheap is that votes are counted in a short period of time (specific block heights), but can be transferred immediately before and after. But if the voting time is longer, the cost will be higher. For example, in the PoS layer of Decred, you can get a vote by locking a certain number of DCRs (about 28 days) before voting. This means that the cost of borrowing DCR to affect the system is much higher because you pay a month's interest (plus collateral costs and volatility risk).

Manipulating governance costs to vote by changing the time that tokens must be locked in may be the ideal solution to many problems.

Systems such as 0x and MakerDAO that use tokens for voting can mimic Decred. By voting for a long time, the tokens must be locked within N days and then returned to the voter, thereby increasing governance costs. There may be an appeal process within that time frame. Developers can also add terms based on the actual situation to make the agreement more flexible. For example, if a decision is particularly important or controversial, a longer lock time is required.

There is also a view that increasing the cost of voting necessarily increases the value of voting, because in an economic sense, in an equilibrium state, cost = price = value. Therefore, the higher the cost of governance, the more valuable the decision to maintain balance. However, simply increasing the cost of voting does not necessarily bring additional governance security. If the cost of governance is too high (that is, the cost of the decision is higher than the value of the decision), then the turnout rate may be low. If the price is too cheap, the system may be vulnerable.

conclusion

This assumption also needs to consider a large number of practical issues, making it difficult to use as a reliable measure of the governance cost in the network. For example, the more tokens that can be lent in the DeFi system, the lower the lending interest rate, and the interest rate will change according to demand and supply-end liquidity. In contrast, few tokens have an active secondary market that we can analyze. . Again, none of this resolves Tarun's concerns. The author is less worried, partly because the market usually responds to changes, and believes that successful networks will find a security requirement that meets Tarun and reach a new balance, even if some of these systems are destroyed in the process (the author's personal opinion, (People being able to lend and borrow voting rights can actually be a good thing, and this argument is left to the future). The author emphasizes that governance costs can be manipulated by changing the time required to complete the vote, but it is difficult to make such calculations very accurately when the market is still immature. But if it works, it can at least provide a workable way to maintain security.