RAI and inflation

[timdaub]

All over the news is currently the ghost of inflation. They say that the USD had e.g. 6% annually. So naturally, I think that it can make sense to think about what to do in inflationary times; especially when considering that this could just the beginning.

I’ve watched a youtube video the other day and there it was suggested that a way to deal with inflation or even to gain profits was to take on debt now, and pay it back after inflation has returned to normal.

To me that makes intuitive sense, as it’d mean taking on an e.g. USD liability now that could be paid back more easily after say a year of inflation with 6%. It’d be because now this liability would be 6% more affordable to pay back as the USD lost that value respectively in buying power.

But for RAI, I don’t get this. What I understand is that RAI is built fundamentally with a PID controller and so from my limited understanding, that means that it is self-referential and hence just tries to minimize the amplitude of swings. So naturally, it settles around an arbitrary value: As as we can observe from the price chart against e.g. RAI/USD, it works.

But what I don’t get and where I’d like to get feedback or help. Or where I’d like to see a discussion happening is on RAI and its properties when it comes to e.g. USD inflating.

Questions:

• If let’s say I put $100 into RAI today and from now on we have an annual inflation of 6%; what would those $100 of RAI be worth next year?
• Also: Regardless of using the PID, I was also wondering: Shouldn’t inflation be observable through e.g. the RAI/USD chart? If USD has less and less buying power: Shouldn’t RAI start to very gradually move up since the USD is losing buying power against RAI? Has anyone observed such a thing?
• Is e.g. inflation seeable by looking at e.g. RAI/EUR? Here there seems to be a clear and slow uptrend. Does that indicate that the eurozone had more or less inflation?

see RAI/EUR: Rai Reflex Index (RAI) Kurs, Grafiken, Marktkapitalisierung | CoinMarketCap (it goes up! see next post for picture)

Also: To all you coders that make this project possible: Awesome work! Super interesting!

[timdaub]

This forum limits my ability to post more than two pictures, so I’m posting the RAI/EUR rate from coinmarketcap as a picture here:

[nikolai]

@timdaub

What I understand is that RAI is built fundamentally with a PID controller and so from my limited understanding, that means that it is self-referential and hence just tries to minimize the amplitude of swings.

It is not self-referential because for Rai the external reference asset is USD. The “error” for the controller is the market price of RAI in USD compared to the system’s internal par price, which is also denominated in USD. Hopefully this explains the clear difference between RAI/USD and RAI/EUR that you see.
My latest take is that Rai is a wrapped USD perpetual swap. It is USD with a ‘drift’. This ‘drift’ might be proportional to CPI inflation, though likely not a 1:1 measure.

[timdaub]

@nikolai

It is not self-referential because for Rai the external reference asset is USD. The “error” for the controller is the market price of RAI in USD compared to the system’s internal par price, which is also denominated in USD.

Thanks for your contribution. Your comment clears up a lot of questions I had. It’s true what you say in that for a human to observe a coin as a stable we have to look at it through the reference of e.g. another asset. So essentially RAI is a stable coin if you consider USD as the volatile asset.

That’s super cool and also explains why my RAI position (I live in the EUR zone) is up compared to EUR.

@nikolai

This ‘drift’ might be proportional to CPI inflation, though likely not a 1:1 measure.

mind_blown.gif

But then now that it has become clear that USD price is indeed used for referencing the error of RAI, isn’t the marketing a bit misleading?

E.g.

I mean, it’s pretty clear that RAI directly engages with the US dollar, so of course, it is directly affected by the US Federal Reserve Monetary Policy. Indeed, as it may be proportional to CPI inflation, there’s a direct influence.

@nikolai Do you have further resources or thoughts on this? E.g. we could go and compare RAI/EUR to CPI in time series and e.g. calculate correlations.

[stone]

Re: “directly affected”, RAI is still small so its behaviour won’t really be affected that much by macro forces. Scale this baby to billions and then maybe we see it affected by more than farms or integrations.

[nikolai]

Every USD derivative is tiny relative to ‘macro scale’, they are all still impacted by fed policy…

[timdaub]

Actually, upon further reflection. I’m still not sure I’m understanding it all.

@nikolai: You said that RAI is like USD with a “drift”. Would you mind elaborating on what you mean by that?

And then regarding my observation that RAI/EUR grows slightly upwards. I put forth the assumption that maybe by looking at RAI/EUR, we can somewhat see USD inflation assuming that the EURs inflation is close to zero.

I guess it’s true but also a bit convoluted. After looking at it a second time, I think that comparing buying power of a fiat currency is probably reflected in its price being referenced by other fiat currencies.

E.g. I can compare the buying power of the EUR against USD and so if EUR goes up then it means that its buying power is increasing relatively. It’s because after an increase in buying power I need to send less EUR to the US for buying a thing.

So e.g. we can take a look at USD/EUR and we also see a slight uptrend over the last year:

I didn’t know where to correlate these prices online but what I’d like to understand is e.g. the Pearson correlation between:

• USD/EUR and RAI/EUR
• What does a par time-series chart look like
• Is RAI painting a completely original signal or is it just copying e.g. the USD plus adding +1.5?

Will RAI = USD + 1.5 hold indefinitely? So is its function just mirroring the USD but with +1.5 on the y axis? Does anyone have all the data that was put into par? Can someone reference to those sections in the code?

[ameen]

It is not self-referential because for Rai the external reference asset is USD.

This is somewhat semantic, but I disagree that RAI is not “self-referential”. The reference asset for RAI might be USD but the target price isn’t some fixed USD price (e.g. $1), it’s the USD price of RAI.

So even if RAI is “USD with a drift”, it references its own drifting USD price, not some fixed USD peg. If RAI’s target price was. fixed at $1, I would call it DAI and say it isn’t self-referential

Generally agree though that the drift is likely to be proportional to inflation and not 1:1 with CPI.

[ameen]

Why I believe RAI is inflation resistant

Having received pushback on this topic recently from notable project supporters, I thought it would be helpful to lay out the entire argument for RAI being inflation resistant as a reference for future RAI supporters.

https://twitter.com/tbr90/status/1476503099137138694

Martin Köppelmann 🦉💳

How RAI’s redemption price works

The first thing you need to understand is how the redemption price of RAI works. The redemption price in RAI/USD is used for determining your liquidation price when you borrow RAI against ETH and also for global settlement in case of an emergency shutdown to determine how much ETH each RAI holder should receive. It can be thought of as “the price the system wants RAI to be”, and as also called the target price. As you can read in the following thread, RAI’s redemption price floats to balance supply and demand for RAI.

Ameen Soleimani

RAI Supply and Demand

RAI can only be minted with ETH, so the supply of RAI is determined by the demand for ETH leverage via RAI. The more people want to leverage their ETH collateral to mint RAI the higher the RAI supply will be.

RAI demand is determined by the demand for holding RAI, including demand to use RAI as collateral for farming & liquidity mining. The more people want to hold RAI, the greater the demand for RAI.

https://twitter.com/ameensol/status/1469383949814157313

To balance the RAI supply and demand, RAI’s algorithmic controller sets the redemption rate (the instantaneous rate of change of the redemption price) to proportionally oppose market price movements away from the redemption price.

When the market price of RAI is greater than the redemption price, the controller takes that as a signal that RAI demand is greater than the RAI supply, and sets the redemption rate negative. This disincentivizes holding RAI and incentivizes minting RAI, and continues until the market price deviation is corrected and RAI supply and demand are once again in equilibrium.

https://twitter.com/ameensol/status/1469383952213307395

Local Dynamics vs Global Dynamics

It would be naive to look at RAI, observe its price movement from inception ($3.14) to present ($3.04) and assume it is not only not inflation resistant, but actually inflating faster than the USD (as a result of RAI’s purchasing power declining faster than USD over the same period).

https://twitter.com/LWarho/status/1476221929568837635

The reason is that RAI is subject to both local and global supply / demand dynamics. During the first 3 weeks of RAI, our team had created skewed liquidity mining rewards that heavily favored buying RAI over minting RAI. The demand for RAI was so high the redemption rate went to -80% and RAI’s redemption price went down by 5% in just 3 weeks. And when we balanced the lopsided liquidity mining incentives by requiring LPs to also mint the RAI they were LPing, the market price deviation was also corrected and RAI’s redemption price went mostly flat.

https://twitter.com/ameensol/status/1476224477268611073

In other words, the entire price move from $3.14 → $3.00 in the first 3 weeks of RAI had absolutely nothing to do with the larger context of “USD inflation as evidenced by CPI growth” and everything to do with apes bidding up RAI to LP and earn rewards.

In the 9 months since we balanced the local imbalance, however, RAI has appreciated from $3 → $3.04 (and avg rate of ~1.5%/year). This could be a reflection of RAI being resistant to the ~5-7% inflation in terms of CPI growth.

https://twitter.com/ameensol/status/1476225988484734977

RAI Equilibrium Bounds

You can somewhat predict RAI’s behavior by understanding the other systems it exists in equilibrium with.

The RAI supply, determined by the demand for ETH leverage via RAI, is in equilibrium with all other sources of ETH leverage. If someone has to pay 5% to borrow DAI on Maker, but observes RAI only has a 2% avg annual redemption rate, then they would save on interest payments by switching their debt over to RAI. Likewise if the redemption rate of RAI is +10%/y, and Maker is only charging 5% to borrow DAI against ETH, then a RAI borrower would save money by switching their debt over to Maker. As such, we would expect the upper bound of the RAI redemption rate to be roughly equal to the ETH borrow rate in Maker and other DeFi products.

The demand for RAI is also similarly in equilibrium with interest rates offered on other stablecoins. If the RAI redemption rate is at a steady -5%, then people can save money by selling their RAI for other stablecoins. Of course, RAI holders farming with their RAI might be willing to stomach a -5% redemption rate if they earn greater than 5% by farming. Indeed this is what took place during the first 3 weeks of RAI, when traders stomached a -80%/y RAI rate because they speculated their LP rewards would be even higher. But all together, simple RAI holders will likely sell RAI if the RAI redemption rate goes negative (because then holding USD stablecoins would be more profitable), and RAI farmers are likely to sell RAI if their net income (redemption rate + farming rewards) is less than the available farming rewards on other stablecoins. As such the lower bound of the RAI redemption rate is determined by the opportunity cost of holding RAI vs other stablecoins.

Why RAI is Inflation Resistant

Ok, now that you have all the relevant background information for the mental model, the main argument is somewhat simple. We’ll argue by contradiction first. Let’s assume the “RAI is pegged to USD” crowd is correct. What happens in a hyperinflation scenario (as defined here as +100%/y CPI growth)?

Well, the first thing that should happen is that traders, expecting the purchasing power of USD to drop significantly, would take out USD denominated loans and dump the USD for anything that seems likely maintain its purchasing power, or possibly appreciate. We would expect the DAI borrow rate against ETH to spike (in an effort by Maker to maintain the 1 DAI = $1 peg), but of course remain well under 100% due to the various risks collateral and liquidity risks involved in locking up ETH and borrowing DAI (which needs to be paid back later). Let’s say the borrow rate for DAI against ETH reaches 30%.

So how might RAI behave if:

1. people assume RAI is pegged USD
2. USD is inflating with 100%/y CPI growth
3. Maker is offering DAI loans against ETH at 30% interest

Assuming RAI starts out at a redemption rate of 0%, we would expect rational traders to mint and sell RAI until the RAI redemption rate is at most 30%, where it would exist in equilibrium with the Maker borrow rate.

There’s a couple things going on here that I’m going to unpack. The first thing is that at t=0, RAI can still be thought as sort of pegged to the USD, since its redemption rate is 0. But the interesting thing is that the more RAI is considered pegged to USD, the more profitable it becomes, in expectation, to mint and sell it in a hyperinflationary environment. If you assume RAI = USD, and the redemption rate on RAI = 0%, then you can save yourself up to 30% borrowing RAI instead of DAI. This works until enough people mint & sell RAI that the redemption rate reaches equilibrium with other ETH leverage markets, and it stops being a profitable interest rate arbitrage opportunity to mint & sell RAI.

So the neat thing that is worth repeating is: RAI can be inflation resistant, even if no one using it thinks it is. In the example above, RAI will still reach equilibrium with other lending markets regardless of the beliefs of RAI market participants, so long as they engage in profitable interest rate arbitrage opportunities. So by virtue of being flexible enough to resist inflation, RAI can do so naturally, and without requiring constant human intervention at the protocol level.

Note - for the purpose of this article I defined “inflation resistant” as RAI appreciating proportional to CPI growth, but not necessarily 1:1. The degree of inflation resistance would be the correlation with CPI, so RAI with 50% inflation resistance would appreciate let’s say 50% of the CPI over the same timeframe. The degree of inflation resistance for RAI is unknown (and likely unknowable) as it depends on the environment that RAI exists in equilibrium with, which is also always changing.

@nikolai

My latest take is that Rai is a wrapped USD perpetual swap. It is USD with a ‘drift’. This ‘drift’ might be proportional to CPI inflation, though likely not a 1:1 measure.

[timdaub]

Hey @ameen thanks for your post.

One thing that hasn’t been considered yet is the connection between buying power and an asset’s utility.

E.g. in your post, you bring up the USD hyperinflation example and you argue that it’d change the system parameters of RAI and DAI in a certain way.

But the reality is that buying power may not need to be connected to e.g. the system’s accounting validity.

E.g. here in Germany as a freelancer, I sometimes get paid in USDC or DAI. But to pay my rent and buy food, what I truly need is EUR. Already, I have a feeling that getting paid in crypto is somewhat of a disadvantage as I’ll have to liquidate it, pay taxes, pay tx fees, have the risk of holding it and my bank hates me for trading crypto.

But in an extreme case where I and others direly needed food/shelter and had both crypto stable coins or fiat for payment, I’d argue that the buying power for EUR would be much higher than e.g. the respective buying power of USDC/DAI. E.g. if I have 0 EURs but 100 DAI, and I’m super hungry; I’d pay e.g. 50 DAI for bread costing otherwise 10 EUR. It’s because I’d want to forgo the action of liquidating my DAI for EUR etc. E.g. in early post-war Germany cigarettes were the de-facto tender; they’re a hard currency as they provide stable intrinsic value for a steady share of the population that is addicted to e.g. 10 cigs a day.

I think an effect like this already needs to be accounted for when looking at the charts of crypto stable coins. It’s because I’m pretty sure that a very small number of individuals are subject to these types of situations around the world already (e.g. ppl that cash out via localbitcoins etc.). I’m saying that small events like this may happen and have happened for RAI already; and they influence the RAI/EUR or RAI/USD or DAI/USD or DAI/EUR price.

The same case is true for assets that e.g. have worse liquidity. E.g. a small-cap stable coin pegged to the USD may have actual less buying power than a big-cap USD stable coin as liquidation comes at a smaller risk/cost.

Dankrad Feist provides good reason for my intuition above: Just because it has a fixed supply doesn’t make it a good store of value | Dankrad Feist
The productivity of an asset decides about it’s future value (buying power). We need a RAIconomy .

[flashback]

Why isn’t this invalidated by your final note? If the “degree of inflation resistance” can be 1%, or is possibly always changing, then is it really inflation resistant at all? Sorry I might be missing something

[ameen]

I agree that asset utility / liquidity / convertibility drive demand for the asset. I think this utility boost applies to both holding RAI and minting it though, since minters need to use or sell their RAI as well.

Driving the RAI economy will be helpful for stability (larger systems are harder to manipulate), but also generally for growth and awareness. The tax stuff is tricky, we’ll have to consider that more…

[ameen]

Why isn’t this invalidated by your final note? If the “degree of inflation resistance” can be 1%, or is possibly always changing, then is it really inflation resistant at all? Sorry I might be missing something

I don’t think so. My main argument is that:

1. RAI exists in equilibrium with other ETH lending opportunities as well as other stablecoin holding/farming opportunities
2. During hyperinflation, I expect rates to borrow USD stables against ETH will rise, because the expected profit on USD debt goes way up
3. As a result, RAI’s redemption rate will rise as well, which should make RAI the asset resistant to the hyperinflation

Of course, the degree of inflation resistance is difficult to predict in advance. If RAI goes up only 1% of the CPI growth, then it is indeed not particularly inflation resistant. But until we see RAI survive a hyperinflationary market and proves its inflation resistance empirically, all we can do is speculate. So for now the argument is sort of like “RAI is in fact a spring, but if you want to know exactly how springy, you’ll have to pull on it”.

My guess would be RAI being about 0.2-0.5 correlated with the CPI. This is roughly extrapolated from Maker ETH borrow rates being about ~3% while CPI growth was about 6-7%.

[Surfaccountant]

Nothings in universe has stable price. It is okay to have a token trying to be always equal to one USD. It is okay for that token to become worthless in inflation and hyperinflation with USD. USD needs a decentralised uncensorable ungoverned infinitely variable interest rate. ETH itself is that in token form. We need it in interest rate form and still don’t have it. Peg a token to USD and don’t stress over de-peg. Just give me decentralised uncensorable ungoverned infinitely variable interest rates on it. This ultimate rate will suck value from USD scam to ETH non-scam more efficiently.