Dynamic Interest Rates: Monetary Policy in the Crypto World

Thanks to Roman Croessmann and Sandra Kumhofer for their feedback and review.

This article is the fourth and last article of a blog post series describing how I think about the future of money and regenerative finance. The third part was about stability and monetary policy. This part is about dynamic interest rates and how they can be implemented into stable asset protocols.

Recapping from the last part, any freely tradable asset is priced by supply and demand. For a stable asset, suppliers are people who enter CDPs and borrow the stable asset into existence and demanders are people who want to use the stable asset.

A stable asset remains stable if the supply matches demand in the open markets. Usually, supply and demand will not match, but vary based on market conditions and user expectations. But, there are ways to influence supply and demand with one key mechanism being interest rates.

Interest rates can be seen as a cost of stability or risk compensation and can be imagined as a fee that is transferred between the supply and demand side. The direction of the payment is dependent on which side is stronger at the moment.

The goal of an interest rate mechanism is to counter the market forces that determine the price and stabilize the stable asset price by offsetting the supply-demand imbalance over time.

When stable assets are created with volatile collateral assets, there seems to be a trade-off between costs and the time horizon of stability. Short-term stability becomes expensive to maintain while ‘loosening the peg’ decreases the costs.

Three Modes of Stability

With the definition of stability being the maintenance of purchasing power, there seem to be three categories of stable assets in the trade-off between time and stability:

1. Shorter-term: Pegged to fiat currencies to secure daily purchasing power. Relative high cost of stability, less likely to be stable long-term due to inflation.
2. Mid-term: Floating peg stable assets like Reflexer RAI or the upcoming Mento ReFi stable asset. Replicates a lower volatility version of volatile productive assets. A compromise between cost, risk, and stability time horizon.
3. Longer-term: No stable asset, but a diversified portfolio of volatile productive assets like stocks and some crypto assets themselves. Basically no stability cost, but short-term fluctuations. Likely most purchasing-power-maintaining in the long run.

Shorter-term stable assets are usually stablecoins soft-pegged to fiat currencies, as most stable assets today. Mid-term stable assets would be stablecoins soft-pegged to economic indicators, like consumer price indices as a proxy for purchasing power and inflation. Another part of that group are floating-peg stable assets, which are not pegged to anything, but represent a time-weighted average price of their collateral assets by smoothing out volatility.

A floating-peg stable asset can also be seen as analogous to the largest fiat currencies in the world. The Chinese Yuan, US Dollar, and the Euro are not directly pegged to anything, but are backed by their respective economic output and vary in exchange rate based on supply and demand in the international currency and debt markets. Their value can be seen as a time-weighted average of their economic output. In contrast, smaller fiat currencies like the West African Franc often intend to peg to a fraction of a larger fiat currency, like in the West African Franc's case the Euro.

Floating-peg stable assets will be the focus of this article. Floating-peg stable assets are essentially inflating and deflating in response to the market forces of demand and supply. How does this work? Let us first look at how interest rates can be implemented in a stable asset protocol.

Two Ways to Implement Interest Rates

For any kind of stable asset, there are two ways that interest rates can be implemented:

1. Balance changes
2. Price changes

In case of balance changes, the actual amount of stable tokens in a wallet would decrease or increase over time. This is the approach that for example, Maker DAI uses. Users can deposit their tokens into a savings contract and accumulate the interest rate. However, there is one important limitation to this: It prevents effective negative interest rates, as nobody would be willing to deposit funds into a smart contract only to lose part of it over time. One special case might be Liquity’s approach, where users pay a one-off interest rate instead of an accumulation over time. Although Liquity does also not allow for negative interest rates, this approach could work better. But with one-off interest rates come new economic problems which might make the situation worse for a non-fiat-pegged stable asset. Maker DAI solved the negative interest rate problem by introducing a centralized fiat-backed collateral asset, USDC, effectively reducing the risk to borrow DAI into existence, making a negative interest rate not necessary anymore. Liquity just accepts the fact that in bear market conditions, their stablecoin will trade above the target peg of one US Dollar.

Allowing for continuous balance changes in the token smart contract would also destroy wallet integrations and user experience, and does thus not seem like a good option.

Here comes possibility number two, price changes. Instead of the balance, a stable asset protocol can simply change its price over time. This allows for both a positive and negative interest rate without any unfavorable technical and usability implications. In the case of a positive interest rate for stable asset users, the price decreases over time and vice versa. The stable asset protocol which pioneered that approach in crypto is Reflexer RAI.

Reflexer RAI

RAI is a CDP-type stable asset protocol where users can borrow the stable asset RAI against Ether as collateral. There is no soft peg, instead, there is a continuously adjusted protocol internal price, called redemption price. The protocol effectively devalues and revalues the stable asset, counteracting market forces and thereby letting the stable asset replicate a lower-volatility, time-averaged version of the collateral Ether price. How does that work?

In order to counteract market forces, the protocol needs to measure the imbalance of supply and demand in the market. If the market price of the stable asset rises, the protocol needs to make it cheaper to counteract. If the market price of the stable asset declines, the protocol needs to make it more expensive to counteract. RAI does this with its redemption rate adjustment mechanism.

Measuring supply-demand imbalance

In the RAI system, there are two prices for the stable asset:

1. The redemption price. This is the protocol internal accounting price at which all positions are valued at. It is also the price users get when interacting with the protocol directly.
2. The market price. The price at which the stable asset trades on open markets.

The redemption price is also called the target price at which the protocol wants the stable asset to trade at, like one US Dollar for other fiat stablecoins.

Now, when the market price is higher than the redemption price, it indicates more demand than supply. In this case, the protocol counteracts by decreasing the redemption price, incentivizing more supply and less demand, as it becomes cheaper to borrow into existence and more expensive to use, as it loses value over time.

When the market price is lower than the redemption price, the protocol needs to incentivize demand and less supply. It does so by increasing the stable asset redemption price. It becomes more expensive to borrow the stable asset into existence and cheaper to use, as it appreciates in value over time.

How does the protocol increase or decrease the redemption price? It does so with a so called redemption rate. The redemption rate determines the speed and direction at which the redemption price is adjusted upwards or downwards. For the math people among us, it can also be understood as sort of the first derivation of the redemption price.

The redemption rate is propagated over time, based on its deviation from the market price:

where the adjustment term is in the opposite direction of the market price deviation. This redemption rate then determines the change in the redemption price over time. The redemption rate acts as a continuously applied and dynamically adjusted interest rate.

To properly adjust the redemption rate in response to market conditions, RAI uses an established mechanism from control theory, a PID controller. This is the technology behind for example, cruise control or lane-keeping assistants in cars.

RAI has been explained many times, and different explanations probably make it easier for you to understand. If you want to dive deeper into the mechanics of RAI and the PID controller, I recommend the following two articles:

1. Dankrad Feist, RAI -- one of the coolest experiments in crypto.
2. Reflexer Labs and BlockScience, Summoning the money god.

In practice, RAI worked well and has been floating with much less volatility than cryptocurrencies like Ether or Bitcoin. Below you can also see a figure that compares the price development of RAI with those of several fiat exchange rates. As you can see, RAI seems to be a smooth overlay, indicating more mid-term stability by smoothing out volatility and saving stability costs as defined above.

Nobel-prize-winning economist Milton Friedman once said that he would prefer a computer over the FED. “Leaving monetary and banking arrangements to the market would have produced a more satisfactory outcome than was actually achieved through government involvement”. This idea is heavily debated by economists, but I think one thing is for sure: RAI is a super interesting experiment developing exactly that kind of technology and testing it in the real world, contributing to the monetary policy arsenal of tomorrow.

Mento Labs' Approach to Stability

Low volatility assets like RAI would be a good candidate for the Mento Reserve, a diversified basket of cryptocurrencies backing Mento stable assets.

Mento Labs approaches stability from two angles. First, there are several fiat-pegged stable assets that provide preservation of the purchasing power of fiat currencies. In my stability framework above, these can be classified as shorter-term stability in reference to purchasing power. But if inflation hits the referenced fiat currencies, the stable assets will lose their purchasing power along their side.

Mento and Mento Labs are about sustainable digital assets. Sustainability for Mento is defined in economic and environmental terms. This is where the ReFi stable asset as introduced in part two of this blog post series comes in to provide more mid-term stability.

ReFi Stable Monetary Policy

The ReFi Stable Asset, as envisioned by Mento Labs, will utilize the floating peg monetary policy from Reflexer RAI. Supply and demand dynamics are even more unclear with the newer, more illiquid, and smaller collateral assets that the ReFi Stable Asset will rely upon. It is thus important for the protocol to remain flexible and adjust to market conditions, for which the RAI monetary policy is well suited. The ReFi Stable Asset can, with that monetary policy, also be understood as replicating the time-averaged, lower volatility price of the basket of ReFi assets that are backing it.

This will bring a new floating-peg stable asset into crypto and can help solve the jumpstart problem of on-chain ReFi markets. If the asset grows in users and liquidity it might even provide a good collateral asset for other, mostly fiat-pegged stablecoins.

This is the end of the four-part blog post series. I hope they helped you understand money better. How money could develop into something inherently friendly to our planet, and that this is not just some idealized theory, but part of the crypto world is actively working on developing such technology.

If you have any comments or questions, please, reach out.

And if you want to keep up to date with these topics, you can follow Mento Labs on Twitter X or join the Mento community on Discord.