(､ᐛ)ﾍ＿/ Nadim's Blog

Pendulum of many pivots

Tried and true, Pendle stands the test of time. I still remember the first time I encountered Pendle, back in 2021. Points weren't a thing yet, liquidity mining was the norm. The accounting wizardry they had just introduced completely escaped me.

It felt convoluted. It wasn't always obvious. It was far from inevitable. Excellent market timing as well as consistent and steady improvements to the product and its presentation have proven sufficient to elicit impressive growth.

Yield splitting, an arcane art masquerading as a science and masterful game of musical chairs, now feels like a natural, necessary abstraction on top of unpredictable and often fictitious DeFi yields.

Where do we go from here?

The Split

Let me try to walk you through Pendle's clever accounting:

In other words: SY is a bond. PT is a zero-coupon bond. YT represents the coupon payments.

It is the balance of YT or PT in a user's wallet that defines their claim at expiry. Now, you might sense a problem here, as this seems to require a number of separate liquidity pools which would create liquidity fragmentation & pricing issues.

In practice, Pendle only maintains one real pool: PT / SY. The other swaps are facilitated by "flash swaps" using the same liquidity pool. This routing is possible because the price of SY will always be equal to the sum of PT and YT (inverse correlation).

This reduces fragmentation and allows all the liquidity to be concentrated in a relatively attractive pool giving exposure to PT yield, underlying yield, swap fees and token incentives.

Because a pure constant product AMM pool is far from ideal for trading yield (or anything really), a further optimisation consists in concentrating liquidity within a range of yield (relatively easy to define for most yield-bearing assets). The tradeoff here is that trading out of the range becomes almost impossible, making it hard to arbitrage any major yield mispricing or volatility through swapping.

There's more to this but we'll leave it there for now. Key points:

The Value Prop

This is compelling for three reasons: (1) DeFi yields are volatile, (2) DeFi rewards (points) are unpredictable, (3) people like to gamble.

It's a fine balance: Volatility and uncertainty creates demand for fixed yield. Speculative interest creates demand for variable yield.

Composability is the cherry on top. Think PT as collateral and leveraged YT exposure. Loops.

The Twists

The success of Pendle has been closely followed, and validated the viability of yield-splitting in an environment where uncertainty around rewards was rising.

Pendle wasn't the first. Nor was it technologically leagues beyond competitors like APWine (now Spectra). Still-- they proved it could work. Now it's happening.

Spectra has experimented with permissionless listings-- but in a world of users confused by yield opportunities, keeping the options constrained has proven competitive; almost desirable.

Hourglass takes an interesting approach. Instead of catering to everyone by offering a wide range of exposures to yield, they primarily offer their YT-equivalent token (PYT) to their users. This is framed as a solution for leveraged airdrop farming. Another twist: market makers provide the liquidity through a Request-for-Quote (RFQ) system, where the order matching happens off-chain but the settlements happens on-chain via Seaport. The market makers can then arbitrage Pendle markets and ensure some level of implied yield parity across equivalent yield opportunities.

StableJack on the other end increases the optionality by offering two other splitting products on top of their Yield Trade product which most closely resembles Pendle's model:

There's much more to all of this, but the key idea here is that new approaches to splitting & liquidity/market making are emerging, each introducing new trade-offs.

The Possibilities

There's an open design space for better yield-splitting. Even more interesting to me is the design space for splitting different things entirely.

Cork is attempting to apply the same model to risk splitting, specifically the risk of a depeg:

A similar mechanism to the one described earlier enables users to swap the risk of depeg by holding either Cover Tokens or Depeg Swaps. There are some other required changes that I won't fully get into, but it is one of the more viable attempts at adapting the yield-splitting concept to another use case.

What else is possible?

Enjoyed this blog post? Consider subscribing: