The Path Forward for DeFi Fixed Income and Yield

Tokens
Premium · AUG 17TH, 2022

Aaron D.

Compound Element Notional

Aave MakerDAO
Finance Finance Finance

DISCLOSURES: THE AUTHOR OF THIS REPORT IS A DISCORD MODERATOR

FOR PENDLE. THE AUTHOR OF THIS REPORT HAS NO MATERIAL OPEN
POSITIONS IN ASSETS MENTIONED IN THIS REPORT. THE AUTHOR HASNOT
PURCHASED OR SOLD ANY TOKEN FOR WHICH THE AUTHOR HAS MATERIAL
NON-PUBLIC INFORMATION DURING THE RESEARCH AND DRAFTING OF THIS
REPORT. DELPHI VENTURES HAS INVESTED IN APWINE (APW) AND DYDX.
THESE STATEMENTS ARE INTENDED TO DISCLOSE ANY CONFLICT OF
INTEREST AND SHOULD NOT BE MISCONSTRUED AS A RECOMMENDATION TO
PURCHASE OR SELL ANY TOKEN, OR TO USE ANY PROTOCOL. THE CONTENT
IS FOR INFORMATIONAL PURPOSES ONLY AND YOU SHOULD NOT MAKE
INVESTMENT DECISIONS BASED SOLELY ON IT. THIS IS NOT INVESTMENT
ADVICE.
Key Takeaways

Yield derivatives have floundered inside DeFi. Despite the primitives going
live over 18 months ago, they boast just $200M in cumulative TVL and under
10% market share of DeFi derivatives.
For yield tokenization protocols, their AMM design plays a key role in how
effective they can be. Fragmented liquidity has been a major factor to their
lack of traction thus far.
TVL in most of these protocols has fallen by over 90%, leading to a natural
contraction in volume.
Notional, the largest DeFi fixed income product, currently offers LPs an
organic yield that is negative.
Most of these products rely on Aave and Compound stablecoin rates, which
have fallen from 10%+ at the start of 2022 to under 1% now.
At the heart of the issue lies the problem of finding the target user. The ideal
customer is here, but yet to begin utilizing these products in a serious way.

Introduction

Since early 2021, crypto derivatives have taken off at a wild pace. But the
success of DeFi derivatives’ has been restrained to a select few exchanges. Most
derivatives traded in crypto today are tied to the price of a crypto asset. While
there is obvious market-fit for these instruments, there’s an entire world of
derivatives beyond single-asset instruments. An interesting type of derivative,
ones centered around interest rates and yields, started to pop up in DeFi early
last year. But their existence has gone broadly unnoticed by the market.

While extremely popular in TradFi, yield derivatives and fixed income products
have struggled to attain usage. When looking at an aggregate of the DeFi
derivatives market, we can see that yield derivatives – including fixed income –
are largely overshadowed by asset perpetuals on protocols like dYdX.

On the surface, this is surprising. At around $340T, interest rate derivatives (IRDs)
are the largest market in TradFi. IRDs include instruments like interest rate swaps
(IRS), forward rate swaps, and zero coupon bonds. Meanwhile, fixed income
products, like T-bills and corporate bonds, are some of the most popular financial
instruments, with a cumulative size of roughly $120T. In fact, these markets are so
large that there is even debate on how to measure them.

DeFi fixed income and yield derivatives seemed like a sure bet to those who
believed DeFi trends would converge with TradFi soon. But their total TVL is less
than $200M cumulatively despite being live for more than 18 months.

This report will briefly survey a select group of yield derivative and fixed income
protocols, explain their mechanics, and analyze their recent performance. We
will hypothesize as to why these protocols have struggled with adoption, which
will allow us to understand what existing or new protocols need to do in order to
grow this market.

TradFi IRS and Fixed Income Refresher

Before we get into the nitty gritty of DeFi interest rate derivatives, let’s briefly
gloss over how interest rate derivatives work.

The most common variation of rate derivatives in DeFi is the interest rate swap
(IRS). APWine, Pendle, and Element are all variations of an IRS. However, IRS in
DeFi and IRS in TradFi work very differently.

In TradFi, Traders don’t trade IRS like perpetual contracts or options. People
aren’t “degen trading” these instruments – for the most part. IRS are almost
always traded OTC with custom terms. Additionally, these specialized derivatives
are most often used by commercial banks, corporations, or other large
institutions to hedge interest rate volatility or speculate on where rates will go.

At a high-level, IRS swaps occur when one entity wants to lock in fixed returns on
its position, and another wants access to variable yields because they think rates
will increase. In this situation, they negotiate an agreement where the variable
rate buyer commits to pay a fixed interest rate to the other party. In contrast, the
fixed interest buyer commits to pay a variable rate. The fixed side is short rates
and the variable side is long rates.
If rates go up, the variable buyer makes money; if rates go down, the fixed-rate
buyer is in profit. Notably, only the difference in the returns changes hands at the
agreement’s maturity. If, at the end of the term, the fixed rate earns $10K and the
variable rate earns $15K, only $5K changes hands.

Fixed Income

Everyone is aware of fixed income products. These are most commonly

corporate or government bonds, popular with pension funds, the risk-averse, and
grandparents. The products pay set rates over the life of the agreement, returning
the principal when done.

It is important to note that fixed income here pays a fixed rate and returns the
principal when done. DeFi fixed income is a different beast altogether. Read on
to understand how DeFi diverges from TradFi in fixed income markets.

DeFi Yield Derivatives Market

As we mentioned earlier, the DeFi yield derivatives market is small, with only a
handful of prominent protocols. Of those protocols, Notional Finance, which
focuses on fixed rates, makes up around 60% of the market. The second largest,
Element Finance, focuses more broadly on fixed and variable rates and makes
up around 20%.
The Tokenizers: APWine, Element, and Pendle

APWine, Element, and Pendle all function similarly with some differences in their
AMM design. These yield tokenization protocols’ work by allowing a user to ‘split’
a yield-bearing token in two:

1. A token representing the principal of the position [PT/OT]

2. A token representing the future yield on the position [FYT/FT]

These tokens are then traded through the respective protocols’ AMM. The tokens
have a contract maturity date, where the yield and principal can be claimed. For
example, the protocol may allow users to split an aUSDC position into PT and
FYT tokens, which a user can redeem on December 1, 2023. Pendle has some
additional nuance here which we will discuss later.

Some of the trades yield tokenization allows for are:

Fixed Rates: Users can immediately sell the yield on their deposit positions
to lock in a set yield.

Zero-Coupon Investing: A trader can buy discounted PT tokens and redeem

them for the principal when they mature. Because the final price of the
token is known, they function akin to zero-coupon instruments.

Leveraged Yield: A user can get levered exposure to interest rates by

purchasing FYT tokens with a loan.

How These AMMs Differ

All of the above protocols allow users to trade their yield and principal tokens
through their AMMs. But the difference in their AMMs changes how users interact
with the protocol.

APWine and Element use modified Balancer AMMs to create markets for the PT
and FYT/FT tokens. But, APWine and Element differ in which assets trade against
each other in the pools. APWine creates a pool where the underlying asset trades
against PT, and a pool where PT trades against FYT.

Element creates pools where PT and FT both trade against the underlying asset.
The underlying asset, in these cases, refers to the base asset for the deposit. For
example, USDC is the base asset for aUSDC, so the pools would be PT-USDC
and YT-USDC. The design of these two protocols means that users must wait for
the maturity of the FYT/YT token before they can claim the yield accrued by the
tokens.
Pendle, meanwhile, uses a dynamic curve AMM for their Yield Tokens [YT] and
outsources the Ownership Tokens [OT] trading to SushiSwap (Ethereum) and
Trader Joe (Avalanche). OT represents the principal of the deposit and trades
against the underlying asset. The YT tokens also trade against the underlying
asset, but users holding YT can claim the accrued yield at any time.

Claiming the yield creates a tricky situation for Pendle. The user who holds YT
accrues and retains the yield, even if they sell the token later. This means that the
value of YT declines as the holder claims it. If they used a traditional AMM for YT,
LPs would face near infinite impermanent loss as the value of YT fell to zero at
maturity. Hence, they created a custom AMM that uses “dynamic curve shifting”
to reduce the IL caused by the time decay.

Digging Into the Data

Unfortunately for the Tokenizer protocols, they have generally struggled to find
market fit for their products. The protocols generally have low TVL, creating
problems for LPs and users. The low TVL has led to low volume, which has also
generally led to low fee accrual for the protocols and LPs.

TVL Drops

APWine’s Interest Bearing Tokens [IBT] – tokens deposited in the protocol – only
briefly broke above $10M in TVL, and have fallen below $5M now. As APWine
requires separate pools for each PT/YT asset, liquidity is highly fragmented. The
low liquidity may contribute to APWine earning less than $5K in fees on
Ethereum Mainnet and under $10 of fees on Polygon.

We have much of the same story occurring with Pendle. Liquidity in the AMM has
declined from $40M at the beginning of the year to just above $3M – marking a
92.5% decline.

And, once again, Element Finance is struggling with liquidity, as it fell from $170M
to $17M.

So across the board, we are seeing a considerable drop in liquidity in yield

tokenization products. This is obviously far from ideal, because we know liquidity
is essential for them. Additionally, as each of these protocols offers yield trading
on numerous assets, the meager liquidity is split even further – rendering some
pools unusable.

Note: The Dune Analytics query for Element Finance had some bugs, so we
made some changes we believed fixed this. The numbers may still be slightly off.

Where Has All the Volume Gone?

Now, TVL doesn’t give us the full picture. We need to look at volume to better
understand what’s going on with these products. Unfortunately, APWine did not
have available volume data, but given their low LP fees, one can assume it isn’t a
considerable amount.

With Pendle, we see a spike of activity throughout the first few months of 2022,
but that tapered at a faster pace than QE going into Q2 and Q3. Pendle had six
pools expire during this time frame, with no new ones added to replace the
expiring ones. But the remaining pools still have little volume.
Element experienced a situation similar to Pendle. The year started well, but then
volume experienced a steady decline, barring a massive spike (for stETH) on Mar.
14. The decline is most likely attributable to a decline in Yearn APYs through Q1
and Q2.

The Biggest Name in DeFi Fixed Income

Notional Finance, unlike the yield tokenizers, focuses on fixed lending and
borrowing. They do this through a system called fCash. fCash tokens are how
Notional accounts for the balance of the protocol – and they are minted in pairs
by borrowers or lenders. Users who borrow from the system receive negative
fCash tokens that denote a maturity date and fixed rate repayment amount.
When people lend, they receive positive fCash tokens that denote a maturity
date and fixed return amount.

However, you would think this system only works if Notional can match specific
lenders with specific borrowers – which it doesn’t. Instead, Notional has liquidity
pools where users provide liquidity for cAssets/fCash tokens and receive
nTokens in return. For reference, cAssets are the derivative tokens from
depositing capital into Compound.

The LPs in these pools take the opposite side of each fixed rate borrow or lend
on their platform. If users lend at a fixed rate, LPs borrow at a variable one, if a
user borrows at a fixed rate, LPs lend at a variable rate. Effectively, LPs end up
providing liquidity across the full range of maturities offered by Notional, and
Notional itself points liquidity to high-demand maturities. Liquidity providers who
hold nTokens receive fees – scaled based on maturity – and token incentives.
Also, as they provide liquidity in the form of Compound cAssets, LPs earn the
variable rate on cAsset deposits.

Of the various yield derivative protocols, Notional is the biggest and has the most
traction. In Nov. 2021, they almost had $1B in TVL. That’s now around the $100M
mark – a massive draw down, but still a significant sum of capital.

Systemic Problems in Current Yield Tokenization Schemes

An interesting phenomenon has occurred at Notional, which is indicative of

issues within the current yield derivative/fixed income design space: LPs are
earning a negative variable rate. How does this even happen?

In Notional, LPs do the opposite of what the end user does. They are a
counterparty to every position, and right now positions are skewed heavily
towards lending (60%) at a fixed rate. DAI LPs have become net borrowers at
Notional’s fixed rate (12m at 2.71%) because users have been locking in fixed
lending rates. But the variable rate has fallen so low (under 1% on cDAI) that LPs
are losing money. Incentives are currently boosting LP profitability, but that’s not
a long term solution.
To Notional’s credit, they realized the issues in this design, and the team has
proposed a fix in the forum.

As we can see, the current landscape for yield tokenization and fixed rate
protocols requires tuning to be properly functional, let alone find market fit. There
also seem to be a few critical and systemic issues that affect the entire space.

Interest Rate Risk

As we have alluded to throughout this piece, one of the prevailing issues is that
the protocols shove much of the interest rate risk onto LPs. Often, they are the
ones that take the entire interest rate risk and who shoulder losses if rates are
volatile. The risk causes LPs to experience a level of toxic flow and has a pull
effect on liquidity – causing low liquidity for the products.

This low liquidity leads to low volume, which further causes LPs to leave due to
low fees from reduced trades and so on. We’ve all seen this happen, and it leads
to a vicious cycle until liquidity trends to 0. To try and retain liquidity, teams are
forced to incentivize pools with their native token.

Hypothetically, teams can alleviate this through a mechanism that transfers

interest rate risk back to those using the products to protect LPs. Currently, it
seems that users accessing fixed rates or IRS products are walking away with LP
money – with little risk beyond the usual risks in crypto.
Liquidity

As we have seen with these protocols, liquidity is a perennial issue. Often, these
protocols take a single yield-bearing asset and split it in two. These two assets
then require two pools and two more assets to trade against – spreading liquidity
even thinner.

Current Market Conditions: Low Rates

Unfortunately for these protocols, rates on the money markets they support have
hampered their ability to attract users. Rates have fallen nearly 90% YTD. The
decline in rates leaves yield tokenization protocols in a tough spot.
Who is the Customer?

The final issue in the space is that of the core user of these protocols. In TradFi,
interest rate derivatives and fixed rates markets are vast, owing to a legitimate
need for these products from institutional players. In crypto, however, the
average user is a retail investor.

Interest rate derivative users are primarily corporations and financial institutions
like commercial banks and investment funds looking to hedge rates or speculate
on them. Fixed rates, meanwhile, are prevalent with pension funds, hedge funds,
and other huge entities. DeFi has a lot of institutions investing into funding rounds
and buying tokens, but there’s a dearth of institutions actually using these
products – which is where the trouble is for interest rate derivatives.

The vision of these protocols is one where DAOs or crypto institutions are
actively using their products. But DAOs are struggling with basic treasury
management (with a few exceptions); they aren’t yet sophisticated enough to
begin considering fixed rates or yield derivatives without specialized support.
Meanwhile, crypto institutions will require far deeper liquidity than the protocols
can currently offer. And the main use case would be to hedge on-chain loans
they’ve taken.

The issue begs the question: do these products have active customers in DeFi?
Looking Forward

Though the data on traction and usage doesn’t look great, all is not lost for
protocols building interest rate derivatives. As DeFi matures, interest rate
derivatives will find market-fit at some point. More than anything, this is a timing
issue – it seems to be a bit premature for interest rate derivatives. Before this
segment of DeFi can thrive, it needs to solve its liquidity problem i.e. find a way to
build up liquidity without relying on token incentives. Something we have
touched on numerous times in our research is that liquidity and usage (volume) is
a chicken and egg problem. You need one to achieve the other. Ultimately, these
protocols need a catalyst that will entice capital into one side (liquidity or
volume), which will eventually lead to natural capital inflows into the other side.

Though things look poor at the moment, there are user archetypes that could find
these products compelling once they have achieved a reasonable scale of
liquidity:

Miners and Validators: For PoW miners (like Bitcoin), debt is an important
instrument in their capital formation. If they were to obtain on-chain loans at
variable rates using their existing inventory (WBTC) as collateral, hedging
would make their cost of debt fixed and predictable. Validators could
potentially lock in a fixed return periodically if interest rate derivatives that
track PoS returns become a thing.

Neobanks: As crypto becomes more popular and reaches a more diverse

clientele, we could see fixed rates become more critical to neobanks who
seek to capture customers with fixed rate products. The kicker is that these
fixed rates would be on-chain, transparent, and sustainable – as opposed to
what we’ve seen unfold with centralized crypto lending platforms of late.

Institutions: Hedge funds delving into on-chain leverage would prefer to

hedge their interest rates. For investment funds, predictable cost of capital
is more desirable than floating rates with steep volatility – even if that
means a slightly higher borrowing rate. Additionally, we could eventually
see specialized funds that focus on crypto interest rate derivatives and
 fixed rates. They could become the counterparty for these products by
specializing in market making for yield protocols.

To summarize, the yield derivatives market is an exciting avenue of financial

engineering, but one that has struggled for traction. DeFi has been struggling
under bearish conditions, which has only made the situation worse. We could
see the landscape for these products change if interest rates on
Aave/Compound start to tick up or even just become more volatile.

The technology the teams are developing here is promising, but there is
hesitation surrounding the space due to timing issues. Before we see these
markets take off, we probably need more of the DeFi stack to be built and figured
out – especially liquidity provisioning. And until that’s the case, this section of the
market may continue to struggle.

A big shoutout to the teams building these protocols for their help explaining how
these mechanics work.

delphi-3656-bc6135