Utilisation — How do we electrify DeFi? (pt. 1)

Jun 9, 2023

TLDR;

Utilisation is the mechanism underpinning how today's DeFi lending protocols set their interest rates.
The primary inefficiency of today's DeFi lending markets is wide bid-offers between lending and borrowing rates.
These wide bid-offers plaguing lending protocols today are not considered tenable for 'institutional finance.'

We discuss the concept of Utilisation and how it currently ‘sets’ interest rates on decentralised borrow & lend platforms such as Aave and Compound. This article accompanies (What is Utilisation?) which draws out the importance of Utilisation and how the current DeFi ecosystem could grow 10x overnight.

Let’s get started. What is Utilisation?

It is the ratio of borrowers vs. lenders, in the context of DeFi. For example, if we had $50 of lenders and $100 of borrowers, the ratio would be $50/$100 = 50%. Though, it is important to note that this ratio does not take into account the absolute value of the numbers, in our case: borrowers and lenders. An example:

Borrowers: $500 vs. Lenders: $1,000 = Utilisation is 50%
Borrowers $5 million vs. Lenders $10 million = Utilisation is 50%

Therefore, we can say that Utilisation is agnostic and independent to the absolute value of the numbers — it is simply a ratio (between borrowers and lenders). Why is all this important?

Utilisation is the mechanism underpinning how DeFi lending protocols set their interest rates.

Interest rates are the ‘rate’ at which borrowers pay and lenders receive a yield. Thus, to understand Utilisation is crucial towards understanding how DeFi borrowing and lending protocols are built.

Relationship: Utilisation and Interest Rates

If we plot a graph (screenshot from the video above), with the ‘Y’ or vertical axis as the Interest rate, and the ‘X’ or horizontal axis as Utilisation, as Utilisation increases, the Interest rate increases.

In this case, if we set the number of lenders to $100, then as the number of borrowers increase, from $10 to $25 to $50, the Interest rate would simultaneously increase, as Utilisation increases from 0.10 to 0.25 to 0.50.

As Utilisation increases, Interest rate increases.

The corresponding increase in Interest rates (to an increase in Utilisation) is designed to attract more lenders as they would be able to ‘capture’ a higher yield, which then results in the Utilisation decreasing (all else equal).

As Interest rate increases, Lenders are incentivised to lend more.

However, given the relationship between Utilisation and Interest rate, how do lending protocols strike a balance between the supply of and demand for borrowing? There are two scenarios lending protocols are concerned about:

High Utilisation, say 70% and above; and
Low Utilisation, say 30% and below.

When Utilisation is high, interest rates could increase faster than that of Utilisation. For example, we may see the following relationship between Utilisation and Interest rate:

As we can see, after 70% Utilisation, for every increase of 10%, the increase in Interest rate grows faster, in order to attract Lenders to lend and thereby decrease Utilisation. In our example above, we may see the total lenders of $100 increase to $125 or $150 in order to capture the higher Interest rate, and thereby decrease Utilisation.

Conversely, when Utilisation is low, where basically no one is borrowing, or very few people are borrowing, we need interest rates to be much lower to incentivize or create a reason for borrowers to come and start borrowing.

Therefore, the ‘sweetspot’ for Interest rate within the current Utilisation environment is between 30% to 70%, perhaps we could refer to this as the Goldilocks dilemma, where outside of this range, we experience imbalances with the Interest rate either being ‘too high’ or ‘too low’.

Based on the above diagram, with Utilisation at 70% and the Borrow Rate (blue line) of 5%, the Lending Rate (orange line) becomes 3.5%, with the difference (1.5%) resulting in a ‘deadweight loss’ (discussed in our follow-up article). Or put simply,

Borrow Rate = 5.00%, a function of Utilisation (70%)
Lending Rate = Borrow Rate (5.00%) x Utilisation (70%) = 3.50%

However, what happens when Utilisation drops? Since the Lending Rate is a function of Utilisation (per above), the Lending Rate drops more than the Borrow Rate — think of it as a multiplier effect, given a lower Borrow Rate x lower Utilisation.

By example, at low levels of Utilisation at (say) 20%, with the Borrow Rate dropping to (say) 2.5%, the Lend Rate drops to 0.5%, with the difference (2.0%) being a deadweight loss — let’s refer to this as ‘spread’.

I think we can both already sense there is an inefficiency baked into how the mechanic is built but let’s address this in our follow-up article.

Putting theory into practice: Aave

Let’s pull up Aave’s rates (Aug. 2022) and see this in practice.

Above, we have Aave’s Ethereum assets market sorted by Total Supplied, listed by assets on their protocol that can be both lent and borrowed. In this example, let’s take the stablecoin USDC:

Total USDC (ERC20) supplied $1.22B
Total USDC (ERC20) borrowed $487.37M
[Utilisation is ~40%]
A lender would earn 0.67%, while a borrower would pay 1.79%
[Spread is ~1.1%]

The first major inefficiency of today’s DeFi markets (the deadweight, or spread, is too wide)

Another example of inefficiency on Aave is Wrapped BTC:

Total WBTC (ERC20) supplied $645.99M
Total WBTC (ERC20) borrowed $24.49M
[Utilisation is ~38%]
A lender would earn 0.02%, while a borrower would pay 0.47%
[Spread is ~0.45%]

The wide bid offer in Crypto (at the moment) seems to be accepted as common practice. However, for most retail investors and retail banking clients who are not as familiar with the interest rate markets, this is a bid-offer / spread not considered tenable for ‘institutional finance’.

Taking spreads to the extreme

When Utilisation is high and spread is low, as Utilisation drops, the spread widens to a point which is untenable for usage by institutional finance. As an analogy, the following hypothetical example of BTC relates to the wide (perhaps considered ‘insane’) spreads.

For argument’s sake, let’s say BTC was trading at $25,000 today. In the previous example, WBTC on Aave’s protocol had a Lending Rate of 0.02% and Borrow Rate of 0.47%. If BTC traded at $25,000, this would be the equivalent of being able to only sell BTC for $2,000 and buy BTC for $47,000.

Sell BTC for $2,000, Buy BTC for $47,000. This is not how real markets work.

Effectively, this highlights that if we compared the wide bid offers we see in the Interest rates market to the kind of token or swapping market, this is what it would be. However, this type of inefficiency would not survive in the world of traditional ‘institutional’ finance. While the spot market in Crypto had relatively wide bid offers to start, their market is now much tighter as the exchanges and protocols have developed to a point where you don’t see much wide bid offers.

In reality, we see tighter spread where you could sell BTC today at $24,990 and buy at $25,010, as opposed to selling at $2,000 and buying at $47,000 like in our hypothetical example.

The wide bid offer spread is the biggest problem with Utilisation in today’s lending and borrowing protocols in DeFi. The current model is based on the concept of raising Interest rates when there is some scarcity of lending (supply), and it lowers the Interest rates to encourage borrowing (demand) at a certain rate.

In conclusion, Utilisation in the current protocols is based on this concept of uneven notional/principal matching (discussed in our follow-up articles).

How do we evolve to DeFi 2.0?

Let’s call Aave and Compound as representative of lending and borrowing protocols in DeFi 1.0. Their Utilisation, as much as it determines the Interest rate, facilitates uneven/notional matching. This leads to wide bid-offers / spreads.

Wide bid-offers / spreads is the main problem plaguing the current inefficient ecosystem.

This is not right. We need to evolve.

The current lending protocols are great as a proof of concept 1.0 where everyone has learnt how to swim at the shallow end of the pool, but at the end of the day, we need to evolve from builders in a room drawing a couple of lines saying this is how interest rates are going to be determined.

Why is everyone who wants to participate beholden to wide spreads?

DeFi 1.0 is not an environment where traditional institutional finance as ‘real money’ allocators can feel comfortable participating in. At the end of the day, these interest rates are deterministic and arbitrary, where the rules of the game means everyone must subscribe to these inefficient protocols.

Can we come up with a better mechanism? One where there is no bid-offer or spread?

The short answer is yes.

In our follow-up article, (part 2) we will be talking about quantifying the spread in today’s protocols and demonstrate how much bigger the borrowing and lending market could get with a more efficient model.