sDEBT Discussion

[grantferowich]

Many in the community have raised the idea for a Synth that mirrors the global debt pool. Since Mintrs see their individual debts go up when the global debt pool increases, one strategy is to hedge against changes in the debt pool by roughly mirroring the global distribution of synths on an individual basis.

For example, if sETH is 40% of the pool, sBTC is 5% of the pool, and the remaining 55% of the synths are sUSD, a Mintr might choose to allocate 40% of his Synths in sETH, 5% in sBTC, and 55% in sUSD. In fact this is the approximate distribution of all the synths in circulation.

Source: synthetixstats.com

This is the mirroring strategy that Mintrs can employ to hedge against major changes in the debt-pool. This strategy requires an individual Mintr to rebalance his Synth allocation as the global distribution changes. If sETH goes 5x, the Mintr has exposure to this price growth and won't get rekt on his debt.

In another scenario where a Mintr who has a synth allocation of 100% sBTC, or 100% sUSD, and sETH goes 5x in price while sBTC and sUSD stayed flat, the Mintr will be strictly worse off, and will owe more debt than his synths are worth. See "Why does my total sUSD debt fluctuate over time?" for more examples and discussion.

Issue

Some members of the Synthetix community have therefore suggested: Why is there not an sDEBT Synth that automatically rebalances as the distribution of Synths changes? Some in the Synthetix Community, including the current author, believe this to be an unambiguously bad idea for the system.

sDEBT is untenable with the premise of the Synthetix peer-to-contract, pooled collateral model where Mintrs absorb the risk of debt fluctuations in exchange for an economic claim on fees generated by the Synthetix Exchange.

Summary of Discourse

1. The sDEBT would be a negative expected value instrument, because it's unreasonable to think traders will lose on net before fees in the long term. You cannot expect to quote a market 24/7 and win prior to fees. That's just not how trading works. (Spreek)

I'm also not assuming that traders on net will be profitable. It's quite conceivable that they could lose money on net. A priori, there is nothing indicating that everyone else in the system that does not own sDEBT will be strictly off. right now a majority of synths are long BTC and ETH. If those two drop heavily, on average you'll see traders lose. (ZombieNik3)

It's theoretically possible but it defies common sense. Name me any market maker that quotes 24/7 with no spread. (Spreek)

Yes it does, because sDEBT's gains are the system's positive delta in debt, which is absorbed by the collective mintr base who doesn't own sDEBT. Unless under a few circumstances where you don't own sDEBT but still kill it in trading, and so are net better off anyway. (farmwell)

2. If sDEBT existed - i'd mint completely to it and then just wait for the fees each week - I'm not providing any positive benefit to the system and get free reward. Just don't see it as a viable solution - I don't see why everyone wouldn't do the same meaning we would have no synths being traded. (Munter)

The sDEBT would need rebalancing at a transaction cost/fee to the system (cleese)

3. I agree it doesn't make sense for the reasons already outlined around stakers being required to take on risk to be rewarded. However, one thing that could be interesting is if we provide some sort of sDEBT position that has a daily fee for holding it (or larger fee to enter/exit). That way, you're paying for the right to avoid the risk, and might provide an alternative system debt rebalancing mechanism, because there's the risk while people hold this sDEBT position, the system debt doesn't fluctuate (or even goes down), which means the people in the sDEBT position lose out due to the fees they paid to hold that position. (garth | synthetix)

Put a price on the right to avoid risk and sell it. feels like an insurance policy on first glance. Hmm. The value proposition of sDEBT evolves into: here’s an insurance policy to hedge against changes in the debt pool. (The people underwriting the insurance policy are the mintrs. This starts to feel like the value prop of being a mintr changes from providing liquidity as a pooled counterparty to being a claimant on fees without having to take on risk of your debt being repriced. If the premium is significant it feels like there may be an opening. The premium would be distributed to all mintrs who don’t hold sDEBT? Sounds trick implementation wise, just sell out of sDEBT at snapshot and claim fees paid by people holding sDEBT. (farmwell)

you could do it as rebalancing the sDEBT portfolio paying the .3% exchange fees every period
which is obviously theoretically doable for a bot currently, and we could let people do it and save some gas. but like I said, I would worry that it would be treated as an endorsement and cause people to lose money doing it (especially in a world without minter inflation rewards) (Spreek)

It still feels like a way to harvest inflation rewards while skirting the risk of debt repricing, and the issue there is now you are asking Mintrs to collectively underwrite Global synth debt changes as well as this weird index product that is itself a mirror of the global synth debt. There’s some second or third order effects I can’t quite wrap my brain around here atm. (farmwell)

I just think that it is fundamentally a zero sum game between the minter who wants to hedge all his risk on platform and all the other minters, so I don't really see a huge value proposition. (Spreek)

Like, if you can collect inflation rewards and be able to not lose on your debt, what’s the point of the inflation rewards? (farmwell)

yeah that's a good point. perhaps the high gas costs of implementing a hedging strategy are more of a feature to discourage it haha (spreek)

I don’t love the idea of becoming an insurance policy underwriter on global debt changes for not owning sDEBT, when as a mintr I already take on risk in the global debt balance fluctuating. (farmwell)

The oracle for sDEBT would be the debt itself. So I think it would create a circular reference. As debt goes up, sDEBT goes up, so debt goes up, ect. So you would eventually end up with with infinite debt and sDEBT. So implementing this synth would crash the crash the network with a p of 1, given enough time. I think. (deltatiger)

I thought the same @Deltatiger , but i screwed around a bit in excel and it looked like it was working.
I might have been doing something wrong though haha (Spreek)

Great discussion guys! A couple of points/thoughts.

I would be careful talking about insurance not to confuse things further. Everything in these type of enviroments is about allocating, pricing, assuming and handling risks. An insurance actor is someone who assumes risk and makes a business of handling that risk efficiently. Breaking things down to the different factors regarding risk is a better way to discuss not to get stuck in existing ways of thinking.

Regarding Circular reference issue from sDEBT. That is only a problem if you have a realtime, constant rebalancing. If you set up a periodic rebalancing (every 15 min, hourly, daily, or similar) it will not be an issue.

This whole discussion about sDEBT is actually not so much about the hedging feature against the debt-pool exposure but more around by whom, where and how such feature should be implemented. In order to bring this further I would need to start with better understanding of the foundational philosophy of Synthetix:
Is the vision of Synthetix to be a slim layer with features being created on top by third parties? Or does Synthetix want to provide more features/functionality itself?

I see risk allocation and handling as central part for the success of this product but implementation of such can be achieved in a number of ways depending on the foundational philosophy (cleese)

6. Imagine an edge case scenario: all your mintrs buy sDEBT, and sDEBT open interest is greater than 70% of total Synths because it’s really popular among mintrs. Then you have a tiny minority of the mintr vase taking on all the risk of debt fluctuations. Not healthy.

Similarly it’s not the case that everyone could go into sDEBT and the system stays sustainable. Imagine another edge case where sDEBT open interest is really high. Now you have an index of the synth distribution that is itself greater in size than its underlying components, which opens up more oracle price arbitraging risk, because you can buy or sell against sDEBT based on how it’s going to be revalued in 15 minutes or an hour. So the self referentiality problem doesn’t go away, it gets worse the more you think about the product.

While it may be attractive from an individual perspective, the hang up for me is that it’s not the case everyone could buy it and the system is still better off. It would strictly shift risk from one group of users to another, and both groups get the same rewards for taking on different risk? Okay, but this will be fixed with a fee, to be paid how and at what rate? So non-sDEBT people get paid to take on the risk, that’s the only way it works, and it has to be so high you discourage most mintrs from doing it.

And the issue with the risk not being efficiently dealt with by non-sDEBT mintrs is exactly why this group shouldn’t be forced to underwrite the risk of global debt fluctuations for the collective mintr base. (farmwell)

6. I would like to re-iterate - mintable sDEBT would NOT work, but a token representing the global debt allocation like an ETF could. (1053r)

7. Dudes want to complain that sDEBT is a bad idea, yeah OK sure. Just make our own baskets then. (ZombieNik3)

Like I said, I don't have an issue with that, in fact in some sense it's the recommended strategy. (Spreek)

8. You can't do it within SNX. Just think bout it. We would have to do it outside, like with Set Tokens or similar. (maex)

9. Finally an explanation for the risks of sDEBT. (hearthouseflowerssquarequite)

Alternatives

1. You could have an index token, similar to UNI-V1 tokens that represents sDEBT instead. This index token basically works like an ETF where user deposits sUSD inside, and it automatically purchases the various sX equivalent to the current global debt allocations while issuing a token (the index token) that represents the bought sX.

This index token can basically be then redeemed back into sX at any point in time (since there's no slippage but only transaction fees).

Of course the details for rebalancing of the index token allocation would then need to be ironed out. Ultimately, this is only going to increase fees across the network where hedgers are willing to pay for. (1053r)

2. Someone can start an index fund call sDEBT and charge management fees. It's essentially the same as if the system has an sDEBT. (atosm)