Thread by @MapleLeafCap

0 – In the past, I have expressed my anticipation on elegant solutions re: non-callable lending – and expect many teams to throw their hats into the ring. @RulerProtocol is what I believe to the 1st valid attempt to take a crack at it (if you know others, dm me! @nftfi maybe?)

1 – The scope of the problem re: “why do we need non-callable lending” is clear – many assets don’t have continuous prices (i.e. trade by appointment, OTC, etc) and can’t really have oracles. Even for less liquid ERC20s, the quality of oracle starts becoming questionable.

2- With current price-centric liquidation model (i.e. loan is called by smart-contract lender when price of collateral dips below X), lend / borrow for long-tail, non-continuous price assets simply doesn’t work (also part of reason why lending protocols pick collateral carefully)

3 – In real life, imagine urself taking out mortgage and buying house. The house won’t have a continuous, “live” price (but instead appraisal is done or only have a record of transactions) – so in the US, banks can’t typically call the mortgage as long as borrowers don’t default.

4 – In essence, the mortgage lender (aka bank, or owner of the MBS) bears the ultimate “off-loading” risk when the borrower defaults – which typically may happen during economic hardship and/or when the property value dips far below the loan amount.

5 – There are several ways to “flex” it such that long-tail assets can work as CT suggests – either via appraiser oracles (@hal2001), liquidation via alternative price proxy (@AlphaKetchum), or pooling like-kind long-tail assets into fungible ERC-20s to be “more liquid”.

6 - … but it feels to me the most “elegant” way for the time-being would be to (a) have the lender be willing to take custody of collateral upon default, and (b) not involve any price-oracle. It’d be like a fine art collector happy to lend against a Picasso painting.

7 – You may ask why anyone would want to borrow this way (yes, it would be more expensive, more on this later) – answer is simple: (a) it’s the only option for this collateral and/or (b) don’t want to ever get liquidated, ask the poor MKR guys who got rekt on March 2020.

8 – There’s an obvious challenge around solving this problem – namely liquidity fragmentation and user-experience. To settle on yes/no re: default, non-callable loans are by-nature “term-based” – i.e. with maturity. If the # of collaterals blow out…

9 - … one could see an almost option-like web with multiple collateral types, multiple terms, and multiple health ratios – each under ERC20 construct require separate pools (that sets the interest rate). It could get unwieldy and very costly quickly.

10 – I don’t have a good solution yet, how @Ruler approaches the MVP is to (a) lock type of assets (BTC and ETH only to start), (b) lock maturity – monthly rolling, (c) lock amount of loan offered per unit of collateral, and (d) let market set the interest rate.

11 - …it’s effectively a curated offering for now: if you really need to pawn 1 WBTC, you can only borrow <25k DAI against it, and you have to pay it back by March 31st, 2021. Precisely how much <25k? That’s decided by the Sushiswap pool (effective interest rate).

12 – For the lenders, by LP-ing in Sushiswap, they would either earn their respective principle + interest (via pull-to-par) or get WBTC if borrowers default at maturity. If you are into options, the lender is effectively selling a put on collateral…

13- …with matched maturity + strike being the set-forth loan-amount. OTM & short-maturity – so cheap to start but one can imagine interest rates blowing out as collateral dips towards mint ratio, thus basically implies @Ruler’s rate has to be higher

14 - …than equivalent Aave / Compound / Yield rates, as the rate = normal rate + short-put proceeds (not taking farming yield into account, inflation could drive this borrowing rate way down). W/ inflation currently rates could be negative (i.e. get paid to borrow kek).

16 – As one can imagine, cost-of-capital for USD is high these days @ 8-25%. So initially bootstrapping the liquidity (and they have to turn on auto-roll) would mean giving inflation to USD liquidity, all while doing RULER:ETH pool 2.

17 - For the Saylors of the world with spare USD, this could actually be kind of a neat deal – you get to farm $Ruler, earn a decent yield on USD, while effectively selling a put on WBTC (and one should be happy to buy it at a steep discount).

18 – With initial circulating of ~3,333 and likely month-1 circulating of 25-30k $RULER (vs. 1 mm RULER total), the 30 bp – 3% early float is tiny (w/ big inflation). I’d be careful aping in (for instance, $1.5k+ Px today = 40 mm month 1 mkt cap but 1.5Bn + FDV)

19 – you can run your own math on farming rewards re: USD lending and $RULER:ETH LP, it’s designed to be intentionally juicy to attract decent TVL. It’s now a sprint & iterate to something better. If you have a brilliant idea, please speak up!

20 - I don’t love the LP design choice currently (having standard AMM curve causes too much interest slippage & low USD utilization), migrating to CRV 3pool makes sense (since most of the time the rctoken would be between 0.8-1)…

21 - …whereby I’d assume the team can also bundle multi-maturity, multi-strike into 1 CRV pool against USDC / USDT / DAI (very similar to Hegic peer2pool). I could also easily see an auto-roll feature for both lenders and borrowers in next version.

22 - Assuming the team can hopefully solve issues discussed above (+ make the tx cheap, need to go to L2 or other L1 probably), I could see the protocol finding decent traction, with longer-term incentives around USD provider still a bit of an issue.

23 – I would be curious if the team can expand into (a) other ERC20 / NFT pools, (b) longer / odder maturities, (c) mortgage-like pay-off structures, and (d) somehow turning fragmented USD liquidity into more composable pools. Tranching would be interesting too.

24 – My hunch is there will be natural demand re: borrowing (especially when rates are cheap enough thanks to inflation initially). It’d be a critical step I think for the team to retain TVL, iterate quickly, and bootstrap community in the first 3-6 months.

25 – It’s experimental for sure (let alone them also carrying $COVER also, ha), but I love that $RULER is taking a crack at it trying. Like how AMM emerged victorious re: L1 DEX, I eagerly await designs that nails this. Plus non-liquidatable loans is just sweet.