Proof-of-Deposit – initial proposal and request for feedback

My last reply for the night. I really need to sleep. Appreciate your engagement.

Decentralisation is safeguard against any form of tamper resistance. Although reputable validators controlling the majority may well be ok normally, it makes the network at the mercy of their “good will” which can change due to regulation, government overreach, etc.

Alright, good night, I think I pointed out the recipe for how someone can rug the entire network via POD.

It will be up to the foundation and other community members to weigh that risk against the benefits of POD.

Last last reply. Fair point (subject to more consideration from my side). I would like to see some stats on the prevalence/profitability of such a strategy though.

With Foundation still being a main voter for smaller validators at least the near term, big validators valuing their reputation, and MEV revenue likely low compared to CELO inflation, I don’t see this being an issue. In the absolute worse case, certain types of transaction fees can be excluded from the block rewards that PoD distributes.

Defi trades usually occur with a slippage tolerance of 50bps, the UI allows you go to down to 25bps, so MEV can eat that proportion of the trades that run through their blocks.

But if the goal is not MEV but to halt the network, then someone can juice their blocks as much as they are willing to invest in a short period.

The gains from halting the entire mainnet is capped by how big of a leveraged short position you can build up in private via various perpetual futures and options contracts. This number will be very lucrative.

@ying_chan @John.Fletcher, I think it would be beneficial at this point to come up with a digest of all the points that have been raised in this long and exciting thread together with responses and potential mitigations. It would be really helpful for others to follow the discussion as the barrier to entry has become rather high

2 Likes

Yes, good idea. I’m writing it up.

1 Like

If you get rid of block gas fees from the reward calculations altogether then it closes the loophole that allows rebating above the market rate.

That’d fend off someone subsidizing their reward rates to overtake the network

3 Likes

Seems like a good point @diwu1989 . I don’t see why block gas fees need to or should be included. Thanks

1 Like

Near Protocol recently launched $USN with similar idea to provide yield. A separate DAO launched the stablecoin backed by USDT & Near. Minting USN will put Near in their reserve, which in return will be staked for approx 10% apy. The staking reward will be used to provide the yield through different defi platforms.

here is the whitepaper

@inlumino - many thanks for sharing. This is fascinating.

This is very relevant to the discussion on adding stables to the Celo Reserve, so I have left related comments there: Upcoming CGP: Increase target level of stable holdings in the reserve - #10 by Pinotio.com

With respect to interest rates

  • I don’t believe that Near have launched their interest rate mechanism yet.
  • In general, I’m skeptical of any programs that advertise headline rates (Anchor, Near, Valora Supercharge) because fixed rates are not sustainable long term. [However, I’ll give the BIG caveat that because Valora use phone verification for each account, and limit deposits on Supercharge to $1,000, this ends up effectively being a type of UBI program (that incentivises savings) - and that makes more sense to me in line with Celo’s mission. This might seem like a small feature, but I think it makes the Supercharge program very different in practise to these other protocol-wide approaches.]

I don’t know the specific mechanism for funding USN interest (please post here if you do). One possibility is that the NEAR in the reserve is being staked/locked, and those rewards (instead of going to the NEAR in the reserve) are being paid out to USN. What is happening there is that reserve collateralisation is being reduced (because the NEAR in the reserve is not earning its staking rewards, and so NEAR in the reserve is being diluted relative to NEAR outside the reserve) in exchange for paying those NEAR rewards to cStables. It’s a trade of reserve collateral for payment of interest to stables. I’m not sure if this is a good or a bad idea overall, but it means the NEAR reserve will be under pressure to hold a high fraction of NEAR in the reserve so that there are more staking rewards that can be used to pay USN interest, and this will reduce the fraction of stables in the reserve, which increases chances of a death spiral.

Big Picture Thinking on Near/Celo

  • The objective of the reserves are to provide safe collateralisation of stables. In principle, this is a service to the stable-holders.
  • The income of the reserve is whatever can be earned as return on reserve assets, plus any exchange (Mento) fees.
  • The costs of the reserve are interest paid to stable holders (currently zero) + impermanent loss on Mento exchanges + dilution of it’s CELO holdings (because CELO in the reserve is not currently locked and does not earn locked Celo rewards, so it is being diluted relative to locked CELO outside the reserve).

Right now, the return and collateralisation strategy of the CELO reserve is currently to be invested 94.5% in a basket of volatile cryptos (BTC, ETH and CELO) and 5.5% total in DAI and cMCO2.

It’s suboptimal that:

  • The Celo reserve collateralises stable assets (cStables) almost fully with volatile cryptos. Obviously, this is a risky collateralisation strategy. Rationally, for outsiders to hold cStables, they will require a very high rate of return (probably on the order of 10%-25% as per Near, Anchor, Valora). If the Celo reserve had higher quality collateral, then outsiders would not need such high interest rates to hold funds in cStables. [Side note: Anchor is even worse than this because Terra is collateralised solely by volatile crypto and Anchor is a lending platform, so there another layer of lending+smartcontract risk as well.]
  • The Celo Reserve investment strategy is 94.5% volatile cryptos, which makes it subject to strong drawdowns when Celo/BTC/ETH price fluctuates. Strong drawdowns greatly affect long term compound returns of a portfolio.

I think the long-term answers to these questions are:

  1. Get assets in the reserve (incl. natural capital) that have a closer correlation with cStables.
  2. Get assets in the reserve that provide a more conservative risk reward trade-off (currently we are taking an almost pure crypto market bet as the reserve investment strategy). This would ideally involve some assets providing yield (as commented by @markbarendt).

The problem is, for KYC and liquidity (especially liquidity on the Celo blockchain) reasons, this is going to take time.

So, in the interim, it seems best to:

  1. At least put more stables in the reserve (helps with correlation and also with taking a more conservative risk-reward investing approach with the reserve).
  2. Aggressively pursue natural capital (via the Climate Collective CGP). [We should ideally also pursue non-ReFi yielding assets, but I don’t think we have bandwidth for this and I think it’s better to focus on ReFi for mission reasons].

I know this is circuitous, but I’m getting back to interest rates and PoD…

Why offer interest?

  1. Interest should compensate for the risks in holding cStables (and also compensate for any opportunity cost for not earning interest on other stables, including just holding fiat). [obviously if we have a safer reserve, we would need to pay less interest.]
  2. In the short term, excess interest rates or payments can be used to incentivise usage/growth of Celo.

What does supercharge do?
In my view, Supercharge rewards is aimed at point 2, and it does so in a nice quasi-UBI way (which would be even nicer if we had more users outside the US and Europe).

I wonder if the Wallet and dApp council @arthurgousset (with experience shared from Valora and @jackie ) could build on Supercharge to extend it to all wallets with phone verification. It seems this could be inline with UBI goals and provide a big incentive for more wallets to adopt cStables. I also think just giving each phone number a flat amount (no saving required) per epoch is worth considering as well (more pure UBI) because it doesn’t incentivise saving. Maybe some combination…

What does Proof of Deposit do?
I see proof of Deposit doing a lot of things that are quite high level and go a lot beyond providing cStables with short term interest payments:
a. Allows us to find a market based way of setting interest rates for point #1 above on “Why Offer Interest”. Seems like something we definitely should be doing (and that no other protocols, strangely, seem to be thinking about). [Note: many blockchains have lending/borrowing platforms, but that is not the same thing, that additionally involves another credit risk.]
b. Allows us improve network security by i) creating healthy competition (which helps decentralisation) between validators [rather than the ad hoc rate of 65k paid per validator today, which incentivises large CELO holders to run multiple validators], ii) diversifying security across not just CELO but other cStables as well. So yeah, PoD helps to create a credible long run interest rate (which, btw, if we had better collateral in the reserve, should result in growth of outstanding cStables and a lowish interest rate in the long run), but PoD is about a lot more too that is worthy of careful consideration.

4 Likes

Good stuff Pinotio,

I am starting to see some possibility in PoD but with several prerequisites.

  1. Regulatory & social complications need to be understood and presented to the community for consideration.

The intent here is to avoid complicating or having the use of the stable coins regulated in a way that would limit the usefulness or market reach a similar generic stable coin would have. For example, I wouldn’t want PoD to put the CELO stable coins in a situation where it takes being an accredited investor to use them in the USA.

Similarly I want to avoid social or religious law problems like the Muslim beliefs around not taking interest on money. To me addressing the need to include Muslim norms suggests that separate classes or statuses (interest bearing and standard) need to be provided to include them.

  1. The interest bearing stable coin account will likely and eventually require regulatory approval.

That approval should be sought or bought wherever a Celo Stable coin would naturally be used. Whether that’s just KYC, or the SEC approval or establishing “The Bank of Celo” or partnering with various and redundant banks in each country or something new developed ‘in house’.

  1. Getting the reserve on track to creating income.

Plan A: Replacing BTC and ETH with stable income generating assets will un-tether the Celo Stable coins from crypto to a great degree. That, plus getting proper on and off ramps, and continued additions to the Celo ecosystem should also un-tether the CELO coin from crypto.
Plan B: is that the income generating reserve could carry the load if PoD’s objective of luring exchanges to list the stable coins is successful and other development is slow. I can see a future where cStable holders get too sticky and the stable coins don’t move enough to make the system profitable and stable.

I’m sure there are other prerequisite things that should be considered also.

3 Likes

Great point! I think it traces back to our understanding what responsibilities, rights and rewards of Celo and cStable holders are. If we consider voting for a validator as a loan, then it is probably considered to be riba and therefore illegal. However, if it is considered as contribution to the network, i.e. strengthening the validator setup by reducing the voting weight per asset, it might be legal. Would be great to get some expert opinion.

3 Likes

There’s some great conversation here, but I do agree with a previous comment about the barrier to entry to this thread being quite high right now.

Can I throw another wrench into the discussion, something I don’t think has been addressed well enough yet: how are validator groups going to work?

If we’re decoupling the maximum epoch rewards you may earn from the number of elected validators (with the proviso obviously that you must have at least one elected validator), there’s no need for you to run more than one validator (potentially).

Or put another way - how are the validator group voting caps intended to work with PoD? Will this be the way we force groups to run more than one validator? By restricting their new cStable deposits until they open up more voting / depositing capacity by running new validators? How will we calculate the equivalent weights of CELO votes versus cStable deposits?

Or are these concepts completely decoupled? (I understand in the PoD proposal that the current election process remains unchanged. Does this include group voting capacity?) Flipping this thought, if there is a move out of locked CELO and into deposited cStables - does this imply your validator group capacity may drop and force some of your validators out of election even though you may have (ultimately) more net $USD value equivalent voting/depositing for you?

3 Likes

Hi @Thylacine, we’re currently putting together materials that collates and summarises the discussion in this thread. We will post it early next week in a new thread.

As a taster, here is our diagram that explains the difference between the current system vs PoD that should answer your questions.

6 Likes

This is great, and clarifies the proposal a lot, in terms of what rewards are being sent where.

Could you also create a before/after chart showing how groups and the election process might or will be affected? I understand the election process technically will work as it does now, however if the switch to PoD is expected to increase demand for cStables, it should theoretically mean some locked CELO is unlocked and converted to cStables deposited against a validator group, perhaps drastically reducing the total votes against them?

Will we see a net reduction in the minimum CELO votes required to be elected? Should we pair the PoD proposal with an increase in validator slots (perhaps to say, 120) to allow some market participants on the edge back into the validating set?

4 Likes

The power of a picture :grinning: Big thanks to @Pinotio.com for suggesting we make one

Let’s discuss your other points on the governance call this Thursday (I assume you’ll be there for the CCF discussion?). These may be potentially be scenarios that we analyse, although models are only as good as their underlying assumptions.

Some loose thoughts: if demand for cStables rises, and demand of CELO rises in response, it seems logical to me that the demand will be met by circulating supply of CELO first, and locked CELO second,
rather than expect significant behaviour change. Currently theres at least 100M CELO in active circulation to absorb demand, and will also rise in value as demand for CELO rises. I also note that as we move to Mento v2, Celo will be closer to the MakerDAO model where collateral providers can dynamically expand/contract cStable supply based on a fixed amount of CELO.

2 Likes

Need people to check my math on this proposal please. Discord

Link back to discord didn’t work for me, see Celo’s discord in Governance>Governance Proposals

I do understand what it is attempting to be. In my own banking, my savings account gets maybe 12 transactions a year, most inbound. My concern remains, the incentive PoD proposes, is to HODL not spend. Until the currency reaches a significant velocity PoD will be slowly draining the reserve. What transaction volume would it take per stable coin to keep the lights on at Celo with PoD in place?

Wow, just did some rough math on this, for the current 65million CUSD alone the cost to the protocol to pay the estimated 5% interest would be $3,250,000/year and that would be essentially a fixed cost of doing business. At a transaction cost of $0.001 it would take 3,250,000,000 transactions to cover just the cost of Proof of Deposit alone. Add the rest of the cost of running Celo…
How in the world do you expect Celo to pay that bill?
On an ongoing basis and forever?

Looks like the EU is seriously considering making interest bearing crypto currencies illegal.

Research on to competition…

http://surveys.kraken.com/jfe/form/SV_3CdoE5uNxeBnFIO?Q_CHL=social&QSocialSource=twitter

Took the survey. I like doing that kind of thing to see what questions they ask, it helps me get a sense of what market they are looking at and how they might market that product.

Has any research been done along these lines for Proof of Deposit?

Dear All,

We have started a new thread on Proof of Deposit here, (as this one has become too long!)