Among the many vague buzz words this cycle, “capital efficient” has really stood out to me. And why wouldn’t we want capital efficiency? That means we can have deeper liquidity and earn more with the same amount of money... Right?

As you may have guessed, it’s not as simple as it sounds, and in many cases “capital efficiency” is synonymous with “leverage”, whether that leverage manifests itself in an amplified risk/return profile, impermanent loss, price risk of a project token, or blowup (solvency) risk.

Let’s look at a few examples where “capital efficiency” = “leverage”, and then a few actual cases of capital efficiency without increased risk profiles.

Where “capital efficiency” is leverage

Rehypothecation

Perhaps the most simple example is yield farming. Most ‘strategies’ involve depositing collateral into a pool in return for an LP token, which can be used again as collateral for another purpose. Most traders intuitively see this as leverage, but new DeFi users should be warned that most single vault services abstract the rehypothecation steps without elaborating on the compounded risk profile.

Impermanent loss in Uni v3

A concentrated liquidity position provides leveraged trading fee returns in return for leveraged impermanent loss.

Following an example from a 2021 Impermanent Loss in Uniswap v3  report, with an initial x price of 100, and an 80-120 range, we effectively have around 9x leverage! And yes, there is a kind of ‘stop loss’ in place with upper and lower bounds but this also requires monitoring to readjust.

There is unfortunately no easy way from Uniswap to visualize this risk associated with this leverage. The result has been many reports of substantial losses across smaller retail LP positions.

Algostables

Are algostables “ponzis”? No, not technically. While many dish out incentives that continue to compound the risks, they do not directly pay newcomers with the last investor’s money or restrict early investors from exiting.

Are they prone to blowup risk? Yes. Algostables in general leverage up on their treasury at the expense of bank run risk.

This risk might be viewed as similar to knowing exactly where a significantly leveraged trader will be margin called, causing a cascade of liquidations and flash crash. If market participants are able to predict the threshold needed to trigger it, they will do so.

We all felt the Terra Luna risk, but it took a substantial amount of money (or fear) to cause a bank run.

Idle Asset Lending

Bancor v2 and many other protocols are tapping into their deposits to lend amongst other time-tested protocols. While the risk of a hack while a portion of the TVL is being lent out may be low, it is likely reflected accurately in the low additional yield that is generated.

Token-Reliant Models

I view Bancor and THORChain’s token liquidity models in a similar light to algostables. The pairing of their tokens against all pooled assets sounds “capital efficient”, but again comes with amplified risks.

Both protocols provide impermanent loss protection that may be seen as an incentive for LPs in addition to the fees they generate. The IL protection is important when sourcing deep liquidity against a volatile protocol token.

If LPs were to lose faith in IL protection and/or all rush for the doors at once, we could see value in the LP pools evaporate quickly. While there are some protections in place here (i.e. THORChain’s slippage fees), it is at the expense of greater friction in pricing.

These are highly innovative protocols and we have yet to see whether they survive the test of time. Proceed with caution if providing liquidity with any large portion of your capital.

Low Collateral Requirements

Liquity has much lower collateral requirements than MakerDAO. This is due to their ability to conduct faster liquidations while cutting out unnecessary features. While this is great during times of stability, it likely amplifies risk during volatility (i.e. the possibility that their stability pool cannot absorb any inefficiencies in liquidations). So has Liquity made itself more capital efficient? Or has it simply shifted the risks to another aspect of the protocol?

Where “capital efficiency” is not just leverage

Capital efficiency should mean some combination of “the ability of liquidity to flow with minimal friction”, or “the degree to which price discovery is allowed to take place”, or the “ability to access leverage on any position”.

Below are a few examples of “real” capital efficiency.

Potential Energy

Capturing the potential supply/demand within a market increases capital efficiency immensely. I believe this to be one of the most under-explored territories within both traditional and decentralized markets.

A simple example is knowing at what prices someone would be willing to buy a multitude of assets. For example, I may have 1000 USDC and be willing to buy UNI at $2 (i.e. a limit order), short a GOOGL perp if the funding rate exceeds 15% annually, lend on Aave if the variable APY exceeds 5%, etc.

By permitting greater access to supply and demand across assets and across protocols, we can deepen liquidity, provide on-demand liquidity when burstiness limits scaling, and provide greater price discovery (and thus stability) to assets.

What’s fascinating about DeFi is that it provides us with universal settlement layer(s) where we can make this happen without double spending!

Aggregation & Interconnectedness

Aggregation and interconnectedness increase capital efficiency by incentivizing best pricing. This is important in facilitating efficient conversation between protocols (in the form of liquidity or potential liquidity).

The aggregation of liquidity across multiple networks into single, deep, pools is something we have not seen succeed yet. There are some great experiments happening in this space with L2s accessing L1 liquidity and vice versa. Crosschain bridges and projects like Thorchain may provide an answer, but as of yet have not figured out how minimize contract, centralization, and tokenomic risk. It is more likely that we see aggregation around ecosystem-specific hubs like Ethereum, Polkadot, etc.

Zero-risk Use of Idle Capital

Flash loans are a great example of utilizing reserves for a second purpose in a relatively low-risk manner (okay, some risk, but just smart contract risk as the assets are returned within the same block). Compound and Aave both permit flash loans from their pools.

Capped pools

In a Uniswap pool, there is no limit to how many LPs can join. This is important because it facilitates deep liquidity, but the implication is that deep liquidity will only become centered around assets that get high trading volume (and thus enough fees to compensate a large pool of LPs).

Creating an upper bound to the TVL (and thus dilution) of a pool is something we have been seeing more and more of.

Hashflow achieves this by letting market makers use deposits to provide realtime quotes without locking up capital. Clipper focused on providing low-fee swaps to small traders, thus allowing themselves to optimize for APY by capping their pools (small swaps do not cause much slippage so they do not need deep liquidity to fulfill their value proposition).

Atomizing risk/reward

Breaking ownership into smaller components can increase capital efficiency. A good example is with Convex, which splits the CRV token’s claim on governance from its claim on revenue.

This allows market participants to invest more specifically in what they believe to be valuable; the granular price discovery provides a more accurate description of Curve as a platform, and also allows more accurate pricing of the Curve token.

To sum it all up...

If a protocol is advertising itself as a “capital efficient solution to ____”, it’s worth asking yourself what the tradeoffs are. It is very possible that the capital efficiency is simply built-in leverage, promoted without the equal and opposite risks!

That being said, as the DeFi space has matured, true capital efficiency in the context of pricing and low friction has increased dramatically. It will continue to increase as the level of connectedness rises and as more market information becomes available to protocols and market participants.

// Jason Garland