Proposal to Update SPOT BondIssuer Tranche Ratios to [33/67]

Short Description

Proposal to update the tranche ratio parameters in Spot’s BondIssuer from [A/Z] = [25/75] to [33/67]

Proposal Details

The tranche ratios balance safety with capital efficiency.

Increasing the A ratio from 25 to 33 means:

  • With 100 AMPL, one could mint 33 SPOT instead of 25
  • For A tranches to be negatively impacted, AMPL supply* would need to decrease by only 67% instead of 75%, over the course of the 28 day bond.

* Note: “supply”, not “Market Cap”

Decreasing the Z ratio from 75 to 67 means:

  • Z tranche rebase multiplier increases from 1.33x to 1.5x (multiplier = 1 + A / Z)
  • Higher Z tranche multiplier means higher effective multiplier on the Rotation Vault [edit 1/24: See notes below]
  • Rotation vault can sustain a higher SPOT supply for a given TVL size

To read more on Spot’s configuration, see About SPOT Configurations in the docs.

SPOT has been live for just over a year. It was launched with a conservative configuration of [20/80], with the assumption that it would be relaxed over time. Last May, it was updated to [25/75].

The most AMPL supply has contracted within 28 days, since the Sigmoid rebase curve was instated on March 26th 2022, is 40%.

It’s also worth noting that the rebase multiplier of the Rotation Vault would not change instantaneously with this update. It would rotate in slowly over the course of a month as new bonds come into the system and old bonds leave the system.

6 Likes

I think this kills two birds with one stone.

It allows for more Spot to be minted per Ampl, which is important to scale Spot

And it gives the vault more leverage, which in turn attracts more deposits.

Spot also becomes a way to dampen long negative cycles with the new arbitage enhancements coming soon.

With that said, I think 67% / 33% is a safe configuration. If for any reason it isn’t we can always revert it back anyways.

I am in favor of this proposal.

Brandon,

I am in favor of this proposal. As a heavy believer in what the Amplforth team has built SPOT has been the most impressive innovation we’ve seen in a really long time. I believe being able to draw out more SPOT liquidity from a users AMPL will be helpful in helping grow SPOT supply and deepen liquidity on the market.

Is there a target ratio we want to hit that makes theoretical sense? Or are we going to keep adjusting based on history? The scenario I would like to avoid is adjusting too much based on empirical evidence and ending up in a situation where we have to reactively snap it back up because we didn’t hit a market situation that reveals that a configuration is too tight. There is not too much history of the sigmoid curve, but it’s definitely promising. Just want to play devil’s advocate and not fly too close to the sun before we are ready.

1 Like

I’m also in favor of this proposal.

And I’m also with @kbambridge to not be too lighthearted when adjusting the ratios more loose.

A negative rebase of the supply by 67% translates to around -3.88% avg. per day for 28 days (0.9612^28). I think that might still be fine, but any more than that and I would start to be concerned.

1 Like

This is the right question to be asking, @kbambridge and @Togenkyo , so I’m glad you’re both raising it. While the answer is ultimately up to the DAO, not needing to “reacting to” or “predict” markets is closer to the original ethos of the project. So I’m hopeful we can avoid those scenarios and slowly close in on the right resting configuration that’s sustainable without meddling.

Thankfully, we have two footholds–empirical evidence from operating history, but also as @Togenkyo recognized, the sigmoid rebasing curve gives us nice analytical bounds.

This is one great property of building on top of AMPL–it’s not subject to havoc caused from one red candle.

2 Likes

After ruminating on this a bit more it’s becoming clear that trying to create a model for this is pretty difficult. The chance of SPOT tranches being impacted by negative rebases is a function of total AMPL liquidity and net total market cap market sold within the bond time period. So yeah it seems we have to make changes based what we observed and what we know about the mechanics of the system. Consolation is that even if we do hit the threshold in one of the tranches, it’s not a positive feedback loop, so risk is mitigated because of SPOT’s design.

Having said all that, I’m in favor of this proposal.

Thinking along this worst case scenario, I’d like to call to attention the uneven distribution of Z tranches in the vault. 50% of the Z tranches mature at the same time on 31/01/2024. The vault has a greater market risk, and if the market goes down near 31/01/2024, the total AMPL value may decrease more than if the vault had an even distribution of Z tranches.

The vault serves to automatically rotate fresh A tranches into SPOT, in a sense it gives SPOT some stability. If the vault operates well for a long time, market participants who used to rotate SPOT manually may stop, increasing concentration risk of manual rotators.

If the DAO doesn’t keep these risks in mind, then the impact of a market downturn on SPOT may be more than expected.

I think this is misleading. Remember that matching As and Zs in the correct proportion are equivalent to holding the underlying AMPL. You can see that there are also many As that mature on 31/01/2024.

The vault has two main functions:

  1. Deploy, which tranches free AMPL, sends the As into SPOT and receives As from SPOT.
  2. Recover, which reclaims the AMPL from matching tranche pairs it holds.

The upkeep calls on the vault call “recover & deploy” on a weekly basis. So in the meantime, there may be some leftover tranches which haven’t been recovered yet. If you call recover right now (it’s freely callable for anyone who wants to cover the gas), you would see many fewer Z tranches.

In so far as the vault is the only rotator, then the assets the vault holds will be the complement to what SPOT holds. Here is the SPOT asset breakdown:

I do agree with the points you raise though! It’s healthy for the system to have an even balance of asset maturity dates. This is one reason there is a “per-tranche” mint cap. But, I’d say the asset makeup right now is very healthy.

4 Likes

I’ve posted the signal vote, linked below. The voting window opens 24hrs from now.

https://signal.ampleforth.org/#/proposal/0x8f2d50068e66e41529d0ffc100a53dc1e551b8ef13eedb7a4e32949ec7afc06f

Note that since SPOT is still managed by the Kennel Club multisig and hasn’t yet been handed off to the Governor contract, this is the final voting step.

Hi - I will use a hypothetical scenario to illustrate why I oppose this proposal. I fear that some people are missing the bigger picture and either getting a false sense of security from the relative market stability during SPOT’s first year, or are thinking too much of short-term gain (ie leveraged vault rewards) rather than long-term viability / robustness.

Here’s the scenario:

By the end of 2025, AMPL’s market cap has ballooned from ~$100m to $2bn following steady adoption of SPOT and growing market confidence in the Ampleforth ecosystem. The vault is in high demand thanks to the appeal of extra income on AMPL, which users widely see as a proven / safe asset. As such SPOT’s collateral set is almost fully immature, and a 67% (or even 50%) tranche ratio would be more than adequate to keep the value of 1SPOT pegged to 1futureAMPL for the short-term.

There then follows a sustained 18-month downturn, during which high-profile industry failures or equivalent bad news sows panic in crypto markets. By mid 2027, the vast majority of alts have either retraced 90-95% from their cycle highs or have failed altogether. Excess leverage & sustained bearish sentiment has forced market participants to pull out of good and bad projects alike. Users are fearful & despondent.

As this scenario unfolds, Ampleforth’s supporters, like supporters of all alts, will be asking themselves if their pet project is really all it’s cracked up to be. The stakes are especially high for stablecoins/flatcoins, as (historically, at least) there’s no middle ground for their survival paths. They can’t just “do ok” - they either successfully preserve their function as a unit of account, or they fail and enter a death spiral.

So, my fear is that two things will happen to the ecosystem in this scenario: first, SPOT’s collateral set will gradually become more and more mature - ie will contain a rising proportion of (supply-volatile) raw AMPL, and a falling proportion of healthy recent tranche vintages - as users pull funds from the vault, either due to panic or based on a rational judgment that AMPL is overvalued and due a major downward correction. Second, at some point towards the end of the 18-month downturn, AMPL’s market cap will have contracted by a market-average of 90% down to a low of ~$200m (mostly due to supply contraction).

Unless I’ve misunderstood something basic, these two outcomes would inevitably mean that the number of SPOT in circulation will be too high to maintain the CPI-adjusted target price with the current 75% (or even the former 80%) ratio split. The collateral set would have shrunk in value so much that the proportional fully-backed value of 1SPOT may be just $0.50 instead of the target price of $1.25. (And the only way to mitigate this would be to have pushed the tranche ratio higher to 90%, as then the 90% collapse in market cap would have been fully absorbed by the older Z-tranches / the raw AMPL in the collateral set.)

I recognise that this in itself doesn’t mean the protocol fails. SPOT is designed to bend its price (not fix it with a hard peg) and to dampen collateral risk (not make it magically disappear). SPOT can and at some point probably will have to recover from a slump from its broad target price range. However, it’s surely a fact that this is a bad scenario which will undermine confidence in the protocol, and that this is therefore something to be actively mitigated against?

I believe keeping the tranche ratio high offers the best mitigation. Given what we know about crypto markets dumping 80-95% in most cycles, I can’t see how a 67% buffer is considered adequate?

The arguments I’ve read in favour of the proposal seem to hinge on two presumptions that either AMPL won’t dump hard for a sustained period, or that the collateral set will always contain a healthy mix of recent vintages. I don’t see either of these as givens - the protocol’s design clearly recognises that cryptos are a risky collateral set capable of major crashes in value (hence the need to tranche “safe” bonds); and also that instilling value in a native flatcoin requires the active consent of users (hence incentives for rollovers / bribes to keep replenishing the supply dynamic so that SPOT sufficiently dampens future supply volatility of its collateral).

If my presumptions about how the above scenario would impact SPOT are wrong, I’d love to hear why - i.e. specifically how SPOT would maintain its target price despite a drawn-out and deep downturn for AMPL that coincides with greatly reduced SPOT rotations. If not, it seems the debate is an ideological one about what degree of risk is acceptable while still being able to claim that SPOT is a robust flatcoin that preserves real value over the long-term.

Given that >99.99% of people alive today haven’t heard of SPOT - and the majority of them wouldn’t understand or trust it even if it was explained to them - I see a heavy burden on the ecosystem to go the extra mile when it comes to value preservation and robustness / redundancy. Lifting the volatility profile of A-tranches is moving in the wrong direction IMO.

I’m in favor of this proposal.

@MLR too many conditional probabilities have been chained together in this presumption ($2B MC is a bit low x 2025 end timeline is very conservative x what the Ampleforth supporters will be asking x and more).

Regarding the adequacy of the 67% buffer, perhaps we might all benefit from an interval in which A-Tranches are affected if the supply contracts enough. SPOT is likely to behave better than conventional stablecoins in this scenario as well, because the “liquidation” of A-Tranches will be graded and contained as opposed to violent liquidations of price elastic assets.

This is a good thought experiment to go through. I’ll try to break down how I reason about it.

It’s helpful to think in terms of two time spans:

  • short, within the timespan of individual bonds, and
  • long, arcs over the course of months or years, covering many bond cycles

The tranche ratios play into both of these to varying degrees.

The short timespan effects are the most tangible. If AMPL supply drops by >67% over 4 weeks, senior tranches may become impaired. I think this case is fairly well understood, with the caveat that supply moves much slower than MC does because it’s bound by the sigmoid function. This is what most of the discussion has been about so far.

The case you raise is worthwhile, though, because it’s related to long-term trends in the markets which may play out over months or years. The short term danger is tranches being impaired. The long term danger is rotations going unfilled.

A first observation is that, for any stable asset, it must have the ability for the supply to increase and decrease with demand. If it doesn’t have this ability, then it can’t claim to have longterm and sustainable stability.

For this scenario to unfold, it must be the case that there’s a sustained drop in demand for AMPL, but no commensurate drop in demand for SPOT–even in the face of all the risk that’s developing over those 18 months. One nice property of Spot is that all this risk is transparent and readily available onchain. This allows the market to price in risk (and thus, supply) as it develops.

Another thought experiment, from the other direction:
If there is, in fact, still demand for SPOT even in the worst of market turmoils, do you not think it would also validate the AMPL ecosystem as a result? If demand for SPOT stays strong, that could be incredibly confidence inspiring. Even if you don’t think so, can you imagine that someone else would?
This isn’t a mathematical proof, but perhaps it updates your probabilities on all the pieces required for this scenario to play out.

Can the market be temporarily irrational? Yes, and we should respect that. If we assume too much rationality and dial things too tightly, I agree that we open ourselves up to perturbations that aren’t necessary. So, I do respect this train of thought to help guide the configuration boundaries.

Related to this, there are some features rolling out in the next Vault version that specifically address potential longterm imbalances between demand for stability and demand for volatility. More to come there, but that was the original driver of Vault v2, actually.

I bring this up because, to be most precise, what matters for whether rotations are fully subscribed is not the MC of AMPL, but the amount of AMPL capital in the vault. It also happens that, as the A ratio increases, less AMPL is required in the vault to fully handle SPOT rotations.

So one way to think about it is: increasing the A ratio makes the short term case harder, but the long term case easier.

1 Like

I included this in the original post, but I think it’s not entirely accurate without further detail:

A ratio update allows the vault to have a higher ceiling on it’s multiplier. The Z tranches the vault holds will have higher multipliers, but on the flip side it will need fewer of them for a given SPOT supply.

Similarly, for minters–A user will receive more SPOT for their AMPL, and they will receive fewer Zs but each with a higher multiplier.

An easy way to think about this is that SPOT represents a “de-volatized” AMPL, and the vault takes the complementing side of this volatility position. So the multiplier on the vault is still proportional to the size of SPOT it’s supporting. A higher ratio allows the vault to support more SPOT in circulation, and therefore receive a higher multiplier when it can deploy more of its capital.

Hope that makes sense and apologies if there was confusion. The previous signal vote passed, but I am open to re-running if this changes the calculus for anyone.

1 Like

I think what you are missing here is that, it’s only the change in supply over 28 days that matters and not the marketcap (price x supply).

For 1 SPOT to be worth less than 1 AMPL, the supply has to shrink by 67% over the span of 28 days.

… or the price of AMPL being -38.8% below target price for 28 days straight:
$1.17 * 0.612 = $0.716

Seems unlikely to happen for me.

A 90-95% drop in marketcap over 18 months wouldn’t affect 1SPOT = 1AMPL, as long as there was no such period during that time.

What could happen though is that 1 AMPL being worth less than target price at redemption.

1 Like

@Togenkyo that’s incorrect I’m afraid. You’re describing short-term risk accurately. But as Brandon explained above there is also a long-term risk of there being too many SPOT in circulation to match the value of corresponding / underlying AMPL 1:1. Big picture that’s because AMPL is supply-volatile, whereas SPOT is (or aspires to be) supply-stable. When you use a supply-volatile coin as collateral for a supply-stable flatcoin, you run the risk of a supply side disconnect developing over the long-term. That would happen if rotations halt and AMPL enters a lenthy period of supply contraction (without SPOT being burnt in tandem). If that happens, each SPOT will only be redeemable for its proportional value within the shrinking collateral set (i.e. a given percentage of 1AMPL, not one full AMPL). This is explained in the whitepaper.

I think it’s vital to always acknowledge - and even be proud of - this innate risk, rather than painting a rosier picture for newbies who join the ecosystem during this cycle. SPOT will only succeed if it proves itself to be the grown-up in the room, differentiating itself from forebears who made big claims that made holders feel safe but ultimately weren’t guaranteed. SPOT is designed to fail in an orderly way and that’s a positive trait, as long as we don’t hide the fact that, yes, SPOT can fail (or, more accurately, unwind) if market conditions dictate. All that can be asked of the ecosystem is full transparency and adequate risk mitigation. I’m still digesting Brandon’s comment as it takes me time to fully process these scenarios. My conservative bias remains but I’ll happily admit when I may be wrong :upside_down_face:

1 Like

Very thoughtful comments from our forward thinking community in this discussion.

Because we can never completely eliminate risk but only do our best to reshuffle it, I believe its best to upgrade the tranche ratios as proposed at this stage in the progression of the project. If unexpected outcomes were to unfold, then we can analyze and better contain the associated impacts now rather than later.

1 Like

Rotations can halt for any tranche ratio. Not sure why this is a concern for this ratio?

The big concern with more Z leverage is that it can reduce vault supply rapidly with negative rebases

But, with the arb between Ampl ↔ Spot coming in V2, negative rebases may not even last that long anymore?

I think its just a thing we need to get into the wild, and observe, especially with the new arbitage enhancement.

1 Like

Thanks for the healthy discussion and thoughtful consideration everyone.

The BondIssuer configuration has now been updated to 33/67 and these values will be used starting next week, when the new minting bond is issued.

2 Likes