Expected fee on open interest

[josojo]

Hello,

these are really interesting papers. Thanks for sharing them.

The paper describes well that the oracle system has to charge a fee on all open interest of the system, in order to support its own valuation and therefore keeping the overall system safe. However, I think the papers are missing one key aspect, an expected value on these fees.

Here, I wrote down my own thoughts. I am looking forward to any thoughts about my calculations and I am very curious about any modeling that was done by you guys on determining the expected fee:

Expected Market Cap of oracle platform =  P&E ratio* Earnings per share * number of share
s
//Assume that P&E ratio will be similar to Nasdaq's average P&E ratio = 20
Expected Market Cap of oracle platform = 20 * earnings per share * number of share
s 
share price * number of share
s = 20* fee per year

share price * number of share
s= 20 * (open interest per year) * fee percentage
   

//In order to keep the system safe, we require:     

Marketcap / 2 >  open interest at any point

share price * number of tokens /2 > open interest



\\Combining both equations
=>
 fee percentage per year >= 10 % per year

[mrice32]

Thanks for the feedback - @cc7768 will be getting back to you soon!

[cc7768]

This is a really useful comment -- We hadn't approached the "back-of-the-envelope" calculations for the expected fee from this direction. I have more training in theoretical work and I think this is a cool empirical thought experiment with a good fiat world counter-part.

I would add three "addendums" to your calculation

1. The 10% fee is on the margin held in the system. In theory, the margin in the system is not the entire value of the contract. If we only have a 10% margin requirement then the tax is effectively a 1% tax on the actual value of the contract rather than a 10% tax -- the more levered the contracts are, the better this looks.
2. The second equation Marketcap / 2 > open interest. This stems from the CoC > PfC equation that we reference in the paper. In lots of our early work, we assumed symmetric contracts which would lead to Marketcap / 2 > open interest / 2. This in turn would lead to the final result being that the fee percentage > 5% rather than fee percentage > 10% -- In lots of the contracts that we are working on now though, this won't necessarily be true. I interpret the computation you did to be a useful "worst-case" of the required fee if the entire margin could be seized
3. The average PE ratio in the Nasdaq is about 20 but there is a pretty high variance in the PE ratios. The question becomes, "which companies are a good comparison?" I don't think we have a great answer for this yet, but I think we probably look more like a tech than a manufacturing company. We hope that this means we can support a higher PE ratio than 20, but it is certainly an open question.

Now to follow up on your question about how we're modeling our future fees. Our approach has been to build a mathematical model of the interactions between margin, price, and fees. We suppose a particular stochastic process for margin growth and then determine the buyback policy (as a function of system margin) that produces the minimum present discounted value of buybacks such that we ensure the PfC > CoC inequality is satisfied. Once we have this sequence of buybacks, we choose a fee policy that can finance these buybacks. This fee policy allows for the collection of extra taxes in times when buybacks are low so that we can finance future buybacks without imposing too large of a fee.

[josojo]

Sorry, I missed the response here.

The 10% fee is on the margin held in the system. In theory, the margin in the system is not the entire value of the contract. If we only have a 10% margin requirement then the tax is effectively a 1% tax on the actual value of the contract rather than a 10% tax -- the more levered the contracts are, the better this looks.

Yeah, this is a good point. This will reduce the oracle fees significantly. Nice!