• Proposal
  • TAP-27 - Cooler Loan Configuration Vote

Voting against the configuration as I see this is game-theoretically suboptimal.

In a limited capacity outcome, all else equal, holders benefit in backing appreciation. Borrowers, however, benefit from both backing appreciation AND returns on borrowed capital (dual value accrual). This splits community into "haves" and "have-nots"; since interest is collected on repayment & Cooler has built-in roll functionality, this inequality can be continued in perpetuity. Here is how I see the payoff matrix between borrowers and holders:

  1. In $69M cap + 0.5% interest rate scenario, borrowers benefit bigly (3), holders lose out on $69M capital allocation (1).

  2. In $69M + 3.3% interest rate scenario, deal is less sweet for borrowers (2) but still competitive to any other lending market. They also benefit in backing increase from 3.3% interest, as do holders (2). There is a way to further optimize this in favor of holders which I’ll share below.

  3. In full capacity + 0.5% scenario, facility is open to all so everyone has same payoff (3,3)

  4. In full capacity + 3.3% scenario, higher interest rates are less attractive to borrowers but benefit holders (2,2).

The current configuration is not a Nash equilibrium (as the holder and borrower have a better payoff state). I don’t see how the protocol can continue to function under this fundamental dynamic of unfairness.

So what’s the solution? Full capacity is one solution that solves this unfairness. However, the implications are not fully thought out and this will fundamentally change OHM’s intended use case from the direction we are headed now.

Another solution can be a variation of borrowers paying holders. In this case, value transfer flows from borrowers directly to holders rather than to the protocol. This removes the dual value accrual property and creates incentives for holders.

Other solutions probably exist that get us out of this sub-optimal state.

I’m interested in exploring all of these pathways but think prudence and careful consideration must be considered. I will go on record to say that the artificial urgency created by the community for this proposal is impulsive and reckless.

To address remaining parameters: 

  • I care less about LTV as it’s useful insofar it’s accretive to protocol & remaining backers. A low LTV balances lender risk and borrower appetite.

  • I’m for shorter tenors to increase observation frequency and flush non-believers as soon as possible. A smaller aligned community is better than a large non-aligned community. 

  • I want to find ways to realize revenue upfront. Value should accrue instantly rather than taking on time risk.

    Thanks ser - I would have the following comments:

    unbanksy33 In a limited capacity outcome, all else equal, holders benefit in backing appreciation. Borrowers, however, benefit from both backing appreciation AND returns on borrowed capital (dual value accrual).

    This proceeds on the basis that the Treasury is somehow accruing value - the Treasury has, and will likely continue, to lose value in real terms - the dominant strategy for the treasury is to stick DAI in the DSR which has been earning 1% and will now move to 3.49%, well below CPI. There is no value accrual, there is just a slower decline in true value - this will continue to be the case and the protocol cannot be expected to accrue value without speculating on volatile assets (something better done by holders)

    unbanksy33 This splits community into "haves" and "have-nots"; since interest is collected on repayment & Cooler has built-in roll functionality, this inequality can be continued in perpetuity. Here is how I see the payoff matrix between borrowers and holders:

    Fees can be brought upfront and distributed to holders through bonds - that's fine at the 0.5% rate

    unbanksy33 In $69M cap + 0.5% interest rate scenario, borrowers benefit bigly (3), holders lose out on $69M capital allocation (1).

    In $69M + 3.3% interest rate scenario, deal is less sweet for borrowers (2) but still competitive to any other lending market. They also benefit in backing increase from 3.3% interest, as do holders (2). There is a way to further optimize this in favor of holders which I’ll share below.

    In full capacity + 0.5% scenario, facility is open to all so everyone has same payoff (3,3)

    In full capacity + 3.3% scenario, higher interest rates are less attractive to borrowers but benefit holders (2,2).

    This ignores that 0.5% at $3000 is actually good for holders because it is the best chance for us to regain mindshare, builders and create token sinks across defi - with the likely effect of gaining a monetary premium (3,3)

    unbanksy33 The current configuration is not a Nash equilibrium (as the holder and borrower have a better payoff state). I don’t see how the protocol can continue to function under this fundamental dynamic of unfairness.

    This is only true in a world envisioned where:

    1. There is not enough capacity to for everyone who wants to to access - i.e. we keep deploying until there are no stables left in the treasury (incredibly unlikely)
    2. We do not trade above backing (then what are we doing here)

    Of course the very thing likely to happen is that there isn't demand for full capacity and the mechanism has the effect that we gain monetary premium and head above backing. In that circumstance those who do not participate are just deciding to not engage in a core mechanic of the protocol as is their right - and can be somewhat compensated through fee direction

    unbanksy33 OHM’s intended use case from the direction we are headed now.

    This is complementary because it is a loan at backing - OHM's purpose is to gain monetary premium this helps achieve that end

    unbanksy33 I care less about LTV as it’s useful insofar it’s accretive to protocol & remaining backers. A low LTV balances lender risk and borrower appetite.

    There is no lender risk if volatiles are managed and loan is at backing

    Don_G_Lover I just don't understand how lower rates would encourage fewer defaults.

    Esp. if borrowers end up paying the interest upfront. If that would be the case, the reason for them paying it back is purely based on their willingness to keep their exposure to OHM right?

    Even if you look at the PoV of let's say a protocol using these loans to access cheap credit to fund some sort of development/product, I still do not see a strong reason for them to pay it back.

    Can you go further into the risk of default?

    Each of the selected choices are problematic by themselves and especially problematic in combination. I see no reason for this to be approved.

    • 69m DAI capacity: Creates scarcity which leaves holders who don't act quickly stuck both without a productive treasury and no way to make the assets productive themselves. Capacity should be uncapped.
    • 2,850 DAI per OHM: With hindsight, it should be clear how using a fixed value is a bad idea—3,000 made sense when the conversation started and now it seems silly. Sell volatile assets, pull liquidity and allow RBS and third-parties to replace it, figure out what actual supply is, base the value on the final numbers with some wiggle room. Vote on the wiggle room.
    • 3.3% interest rate: I have a hard time understanding why there's an interest rate at all. Free the treasury, let people buy OHM when they see it at an attractive LTV. The "interest rate" worth talking about is premium above the loan value. In a lot of ways, this replaces the relegated OHM bonds by "locking" the spread between the loan value and the price of OHM at the time of the loan.
    • 12 month term: Opens a massive existential risk to the protocol for little to no benefit. Lower tenor spreads out default rates as price and environment will have made smaller moves and whatever decision one might make will be more nuanced.

    The combination of all of these choices means that holders unable to access liquidity can become disenfranchised and exit conventionally, stalling price appreciation. True believers who want to play the RBS game by looping and unwinding their investment with price movements are more likely to externalize their investments in assets that offer enough yield to counteract the interest—the lower the LTV and the higher the compounding rate of looping, the less value there is in re-investing. These disincentives point to stagnation a year from now, where default is attractive when looking at a 3.3% fee.

      I have spoken and debated my views extensively in the "General - Policy Discussion" channel on discord, but outline a summary of thoughts here for good order.

      Firstly, I think it is important to understand what Cooler is as I see widespread confusion on certain elements. It is a native feature of the protocol which leverages the unique design of Olympus as a backed (not algo or pegged) currency. Cooler facilitates a preservation of the simplicity and low risk nature of the Olympus backing by distributing the risk to holders that wish to access the backing of their OHM to pursue use cases and investments further out the risk spectrum. In doing so, it alleviates the pressure the DAO and treasury face to introduce (further) risk into the backing assets of the treasury. Cooler is furthermore a potent tool alongside the lower RBS cushion / wall to backstop the OHM price at stable backing. In this sense it can operate as a powerful supply sink to remove OHM from circulation when market conditions are depressed (such as now).

      This framing is important because it supports my thinking on how the parameters of Cooler should be set.

      Capacity: I favour a level that allows access to those who want it, but does not over-allocate to this facility and thereby reduce flexibility for how the treasury is managed. I believe 33m would have been adequate, but 69m was decisively voted for in TAP-25. I therefore think we should close the book on this and settle on 69m as the initial capacity. If the facility is tapped out, then we can expand. If it is under-utilized then we can right-size.

      Tenor: The tenor parameter for me is a balance of administrative burden on the user (i.e. frequency at which your facility drawdown must be rolled) versus a trigger point for the protocol to have an updated view on how much supply remains sunk into the facility. I favour less friction on the user, and this again was decisively voted for in TAP-25 as a 12 month preference. Importantly however, to preserve the substance of Cooler, the user must preserve the right to roll this drawdown on the same terms. If not, the attractiveness and utility of the facility is materially diminished because users will not be able to confidently use their backing to invest in longer duration assets.

      LTV: I am fundamentally neutral between 2850 and 3000. I think 2850 is adequate to deliver the substance of Cooler, but 3000 would likely assist in reigniting the flywheel.

      Interest rate: This is the most important term for the success of Cooler. It should be nominal / 0.5% to remove friction associated with using the facility. This is not a loan to be struck on market based terms or some fancy new income stream for the treasury. It is just a facility to access the backing of each individual's OHM. We want to encourage use of the facility by people who are considering selling their OHM to invest in higher risk assets (i.e. have your cake and eat it too). We want to encourage use of the facility by people that believe in the Olympus roadmap and want to increase their exposure to OHM (i.e. consolidate supply in stronger hands). 3.3% is simply a tax on active users.

      Now as it relates to TAP 27.

      It is clear from TAP-25 that people want Cooler. So a vote against these combined parameters should be interpreted as a vote against the interplay of the combined parameters. In which case I think we should be coming back to present 4 combined options (all at 69m capacity and 12 month tenor) for final vote and implementation;

      1. 2850, 0.5%
      2. 3000, 0.5%
      3. 2850, 3.3%
      4. 3000, 3.3%

      Regarding 2850 v 3000, it would be helpful to have from the DAO in conjunction with this vote a clear view on what is our backing per gOHM today including a list of assumptions (e.g. stable v volatile), and what actions will need to be taken to facilitate $69m capacity at both 2850 and 3000 LTV.

        pottedthings What's your response to the (wrappedDAI) argument?

        i.e. if capacity is uncapped, all treasury assets would have to change into DAI making us some form of wrappedDAI and thereby lose our floating nature.

          Yella The implicit question is whether or not OHM is pegged to backing? My answer is no. OHM is currently pegged to selling, which is why it's running along the lower cushion. Or is the implicit question whether or not the protocol should drive backing down in order to create a premium? My answer is no. The protocol's investment into volatile assets have been again and again a detriment to backing.

          WAGMIcapital Thank you for this, I'm some of the few that voted to get the cooler loans and I couldn't get why there was so many vote for a no. ( I recently came back to crypto after being disgusted by the FTX affair)

          Viewing so many votes for no and a lot of my own ohm holdings into an loan and not accounted for was very depressing. But now I get where most people come from.

          unbanksy33
          "I’m interested in exploring all of these pathways but think prudence and careful consideration must be considered. I will go on record to say that the artificial urgency created by the community for this proposal is impulsive and reckless."

          if there's anything that comes out of this whole exercise… I hope it's this one. I will add that the "artificial urgency" seems to have come from both the community and the DAO (by rushing previous vote to snapshot without having thought it through).

          I want to take this back to first principles for a moment. I think WAGMIcapital does a good job laying out the What of Cooler, but I think equally important is the Why and the Where to.

          At the core of the original proposal is an attempt to derisk the protocol. This should manifest in several ways, some of which have been more discussed than others.

          First, it removes the need for treasury management. There is no need to compose a basket of treasury assets or deploy them into various protocols if individuals can do these actions on their own. I have not, and still do not, see any rational argument against this position in a world in which an individual can access a cooler loan and diversify their backing as they see fit. The one argument I do understand, that of protecting the vanity metric of volatile (and possibly appreciating) backing because of ie ETH holdings, is not one that I agree with and not one that holds any water, in my opinion.

          Second, it ensures holders are as protected from financial and smart contract risks as they want to be. Worried about USDC? Hold your backing in something else. Worried about Ethereum? Move to a different chain or off-chain. You might think: but if any of those risks were to manifest, the protocol is still rekt — and you’d be right. But the status quo is not that there would be any less harm on the protocol level, but simply far more harm on the user level.

          Third, it makes the implementation of on chain governance (in my opinion, the single most important development) far easier. This is not only because less/no treasury management means less complexity and scope of governance. The biggest blocker for OCG to date has been what I will call the “honeypot dilemma”: if the system is fully controlled via token voting, then token voting could be used to send all treasury funds to a single (or small group of) voter(s). Imagine OCG with the ~8% turnout for TAP-25 (the highest in a while) — if a voting bloc has >8% of supply, they could vote to transfer the entire treasury to themselves (12.5x their share). This is the single worst possible outcome for the network and it has proved very difficult to hedge against with a governance system. I used to think that this dynamic would actually serve to ensure a consistent premium for the network: for example, if 33% of supply is needed to pass a rug vote, then OHM would always remain priced at >=3x premium. This hypothesis remains neither valid or invalid; but at the current discounted state it doesn’t really matter. OCG is untenable so long as OHM remains so cheap relative to its treasury liquidity.

          Here is where Cooler comes in though. By lending backing to holders, you remove the honeypot. Rather than focus on premium (controlled by the market and not by governance), we can focus on treasury liquidity (controlled by governance and not by the market). Using our previous example with an 8% governance threshold, trying this attack does not make much sense if only 10% of the treasury is liquid and able to be heisted — a max-25% ROI is not worth the risk you would take on by trying the attack.

          I think that, following the launch and utilisation of Cooler, OCG can be implemented quickly and easily with a Compound/Open Zeppelin governor. The idea of a custom, complex, and untested system taking on such an important role makes me uncomfortable to begin with, and if its not necessary it should not be done. We can take some additional solace in the fact that the Compound/OZ system has a suite of tooling around it, meaning no custom smart contracts, front ends, monitoring, or anything else of the sort. This should sound especially great to those that have been dissatisfied by the governance process around Cooler.

          The thing to understand about the benefits to the implementation of OCG, however, is that it relies on high subscription. I wish I had fully understood this dynamic going into this whole process, because if I had I would have made fewer concessions when it came to capacity. I have heard many share the same sentiment I had around the limit — that if/when its reached, more would be added. However, this seems like a ‘kick the can down the road’ strategy that offers little benefit versus the already tapered deployment strategy (OIP 144) and carries the risk of undermining the entire proposal. The other detriment of limited capacity, which again I did not understand fully until well into this process, is that low interest rates run the risk of harming holders that cannot access the loans. If treasury productivity was 3% but falls to 2% of a result of loans, holders unable to borrow because of capacity limits lose at the expense of those able.

          With the long form out of the way, I will suggest this:

          • Use TAP-27 to reject the TAP-25 configuration (as seems to be occurring already)

          • Start over with a new proposal (I can do this part)

          • Implement the following terms

            • Capacity for all tokens (this can be iterative to hedge smart contract risk)

            • 2,850 DAI per gOHM loan (a wider buffer to backing makes sense)

            • 0.5% interest rate (incentives subscription at no individual loss bc full capacity)

            • 3 month tenor (loans can be rolled over in advance of their expiration, so all this will do is require more frequent interest payments. I get the arguments favouring this and I’m here for it.)

            nicnombre The one argument I do understand, that of protecting the vanity metric of volatile (and possibly appreciating) backing because of ie ETH holdings, is not one that I agree with and not one that holds any water, in my opinion.

            if this is your view of holding ETH and if there's no value in diversifying backing of OHM with additional volatile assets why did we ever offer volatile bonds? What exactly was the thought process behind accumulating (and never selling) tens of millions of dollars worth of ETH? Even Oly Pro was touted as a way to "diversify OHM's backing through a basket of assets" etc. I don't remember hearing any arguments from you (or anyone else in leadership at that time).

            I agree with the remaining parametrization except 0.5% interest rate is still pointlessly low. Somewhat unsure how full capacity will work, but it is definitely more egalitarian.

              z_33 you need the previous sentences to understand what you quoted in context:

              nicnombre There is no need to compose a basket of treasury assets or deploy them into various protocols if individuals can do these actions on their own. I have not, and still do not, see any rational argument against this position in a world in which an individual can access a cooler loan and diversify their backing as they see fit.

              it makes sense when out of the control of the individual, doesn't when its not. I was never big on it in the first place but I am just one of many and I try not to opine unless I feel I need to. Maybe I should have -- spilled milk.

              Haven't seen meaningful discussion so posting 3 options as alternatives to @nicnombre suggestion:

              1. Fully ETH-backed treasury (no Cooler Loans)

              2. Fully ETH-backed treasury (with Cooler Loans)

              3. Cooler Loans with staking (or variant thereof)

              On (3), some variation of staking acts as a game-theoretic tradeoff between choosing to use Cooler Loans facility and staying in OHM. E.g. revenues and liquidations accrued from Cooler Loans flow from borrowers to stakers. The yield could hypothetically be attractive enough to compete with whatever Cooler Loans parameters are chosen.

              On (1) and (2), whether it's from a narrative perspective or regulatory perspective, a fully ETH-backed treasury is closer to the vision of a digital currency (in my personal opinion). I include option (2) as there's merit in Cooler Loans derisking the protocol. All of nicnombre's points are also addressed by (2). Also, this reduces regulatory risk in better ways that original Cooler Loans - the protocol takes on DAI risk for as long as it has outstanding loans in DAI.

              Keep in mind that none of these ideas are fully fleshed but same can be said for original proposal of Cooler Loans. The current state, which I characterize as a false choice fallacy marked by deep community divide, is an indication of that. I therefore disagree with nicnombre's next steps and looking forward to further discussions with him to find a way forward. For those that feel the urgency of some impending bull run, I invite them to consider utilizing existing (and soon to launch new) lending markets with competitive interest rates.

                unbanksy33 I think (3) is fine to explore but need not hold up Cooler loans to develop the mechanism. But it does lead to the question why is the protocol served by ppl not accessing Cooler loans? If it isn't, why do you need such a balancing mechanism? I think this is the primary problem with the unfairness argument and would be good to understand - I mean we don't have such a mechanism for the essentially risk free arb between the 2.3% staking rate and the DAI/OHM pool on Aura at 14%apy? I would say ppl are less likely to sell their OHM if they have borrowed against it in Cooler and pursued some other strategy or are using a vault product that pursues some yield or OHM related strategy (additionally it creates a product space for builders) - this friction is beneficial to the protocol in a similar light to LPs - but it is so much smaller diff at Cooler theoretical 2.8%apy difference rather than the LP at 11.7%apy difference

                In relation to the LP pools there will continue to be ppl who see the LP options as superior strategies for them who do not access Cooler and just want to LP, particularly if we hold onto some/all our Aura/Convex

                The ETH backed treasury would be a good (separate) project for the DAO to launch with RBS and Cooler (RIP Jimbo) but isn't within the realm of possibilities wrt to stability for OHM because we denominate in USD and will do so for the foreseeable future. With that said fully supportive of the DAO proposing use of some treasury funds to stand up such a project if it is backed by some OHM also maybe 50%OHM/50%ETH backed or somethin

                With all that said I think we are on a good path wrt some urgency to get things done (as things usually take longer than we estimate) and the iterative process of coming up with consensus - keen to see @nicnombre's new proposal and discuss soon!

                  Mark11
                  If we are actually talking about "serving the protocol" rather than serving the holders (which both who access cooler loans, and those who don't, equally are), then those who borrow assets from the treasury at potentially ridiculously low interest rates (compared to market rates) aren't "serving the protocol" either.

                  Having said that, I believe (3) only makes sense in the case that Cooler loans are offered at limited capacity which creates these forced "haves" and "have nots". Won't be needed in case of full capacity as everyone will have equal access.

                    z_33 Have and have nots is a choice - and full capacity is a meme as there is a lot of dead OHM - the protocol and holders are served by supply sinks from Cooler borrowing OHM - 'potentially ridiculously low interest' is far higher than the 0.25% charged by Liquity over two years and we should aim to be near as we can be on par with the best. We are going back to ZIRP soon in any event - this just gives us a chance to get some mind share - as I've said before this is marketing budget

                    Write a Reply...