Crypto Yield Sources: Not All Yield is Created Equal

One of the main advantages of decentralized finance is that it is open for anyone to participate, from anywhere, at any time. With this, comes the opportunity to earn yields as a DeFi participant that would otherwise be difficult or impossible to access in the traditional finance realm.

The permissionless and open-source nature of crypto has turned DeFi into a complex ecosystem that is both wide and deep, evidenced by the endless protocol mechanism designs and the 2,000+ total protocols that exist today. As a result, discovering the types of yields that exist and navigating the underlying tradeoffs between them is a difficult task.

This article will cover:

• The major use cases in DeFi
• The definitions of the major principal and yield categories with relevant considerations for each one
• A comprehensive map of the yield sources that exist
• Things to consider moving forward considering the ongoing “Real Yield” narrative

What are the major use cases in DeFi?

There are 8 major use cases in DeFi that currently exist:

• Liquidity: Accessing liquidity for crypto assets in a permissionless and instantaneous manner
• Swapping: buying and selling tokens in an efficient and seamless fashion
• Directional trading: executing a trade based on a specific view about a market’s or asset’s price direction whether it is longing or shorting
• Borrowing & Leverage trading: borrowing capital instantly and at an efficient rate to increase exposure to a particular asset or market
• Yield Farming: deploying assets passively or productively in order to earn a return
• Staking: delegating capital to a project or validator with the expectation of receiving a portion of generated yields
• Depositing capital: deposit assets into protocols and/or pools to generate a return as a liquidity (or counterparty liquidity) provider or lender
• Storage: securely storing crypto assets without sacrificing custody

In order to serve these use cases, capital, infrastructure, and various services are needed, and in return these suppliers earn a reward.

What is yield?

Yield is the percentage return derived from deploying capital into particular strategies. There are two critical parts to defining yield: the principal that is put up and the yield that is earned. In general, there are two types of principal assets and two types of yield.

Principal: The capital that is put up at the very beginning of an investment period. Alternatively known as the initial investment. For example, depositing $1,000 USDC into AAVE.

Yield: The returns that are earned over a period of time. For example, earning $15 (1.5% APR) in USDC over one year.

Total Return: The net gain/loss in principal value + the realized yield.

What are the different types of principal?

There are two main types of principals in DeFi: price-stable and price-fluctuating.

Price-Stable Principal:

• Definition: Principal that broadly does not fluctuate in value and thus contains no price risk and no dilution pressure.
• Examples: Stablecoins. USDC, DAI, USDT.
• Implications: The primary consideration is the opportunity cost of holding and deploying price-stable principal rather than a loss in principal value, especially during bull runs. Worst-case scenario is that the total return is minimal, for example only earning 0.5% on deposited USDC in AAVE over a year. Best case, returns are low double-digit percentages.

Price-Fluctuating Principal:

Definition: Principal that broadly fluctuates in value thus contains price risk. The implications of price-fluctuating tokens vary greatly by sector and specific project.

Examples:

• Layer 1: ETH, SOL, MATIC, AVAX
• Application: SUSHI, CRV, GMX, SNX
• Web3 Infrastructure: LPT, AR, POKT, FIL
• Governance: UNI, FF
• Meme: DOGE, SHIBA

Implications: All types of price-fluctuating principal suffer from price risk. The worst-case scenario is that yields are low and the principal token’s price drops significantly. The best-case scenario is that yields are sustained, and token price appreciates significantly. There are a few DeFi projects with significant cash flows, while other projects have pure staking mechanisms that rely on dilution or inflation. Many projects rely on a mix of both cash flows and inflation, especially within the middleware sector. Many layer 1 projects have relatively steady mid-to-high single-digit yields.

Below is a table to compare relative tradeoffs between the two types of principal and relevant definitions:

• Principal Price Risk — the possibility that the value of the principal deployed falls due to a fall in token price.
  Price-stable principal protects from total return downside attributed to price changes.
• Yield Return Potential — the degree of yield that is attainable by deploying the principal.
  In general, yield return potential is much higher for price-fluctuating tokens, especially during the early stages when dilutionary tokens are being distributed.
• Yield Rate Predictability — the degree to which level of yields can be forecasted.
  It is more simple to predict yields of price-stable principal tokens because there are less variables to consider and there is an inherent ceiling on the yields.
• Total Return Potential — the degree of overall profitability while factoring in the value changes of the principal.
  Price-fluctuating principal tokens are able to achieve higher returns because the principal can appreciate in value, but this is a dual-edged sword given that principal value can also theoretically drive towards zero.
• Principal Dilution Risk — the possibility that the value claim of one token becomes diluted due to increasing circulating supply over time. One after-effect is that the market perceives each token to be overvalued, thus there is higher-than-normal sell pressure.
  As an example, a DEX has a low initial circulating token supply and allocates a large portion of tokens towards liquidity mining. Over time, more users receive these tokens and some decide to stake to earn swap fees, while others sell onto the market and never use the DEX again. The former dilutes the value claim for each token, while the latter lowers the total value claim for stakers in the first place.
• Counterparty Default Risk — the possibility that a counterparty is unable to pay back a loan.
  On undercollateralized lending platforms, if a lender fails to repay a loan, the lender may lose a portion or all of the initially deposited principal. Alameda is likely to default on a few undercollateralized loans set to expire next month, and Gemini Earn recently halted withdrawals due to Genesis liquidity issues.
• Counterparty Trade Gain Risk — the possibility that a counterparty trades correctly which in turn diminishes returns for liquidity providers.
  As an example, if all GMX traders shorted $ETH at the top, then $GLP providers would need to pay out from the pool and take a loss. $GLP providers act as a direct counterparty to GMX traders.

What are the different types of yield?

When principal tokens are put into action, there are generally two types of yields that can be earned in DeFi: token-agnostic yields and token-specific yields.

Token-agnostic yields are agnostic to whether or not there is a project token as there is an organic and direct exchange in value between parties. One party is giving up value up front to access a specific use case, and another party is receiving the value in return for providing capital or a service. Analogous to cash flows in web2.

Token-specific yields must be issued directly from a project treasury. The exact structure, whether it be initial supply, inflation, or burning mechanism, is determined by the holders or founders. Token-specific yields contribute to an ongoing circulating supply that dilutes holders. It would not be possible to control and distribute token-specific yields without a token. Analogous to marketing and/or customer acquisition costs in web2.

Below is a deeper view into the flows of the two main yield categories.

Token-Agnostic Yields

There are 4 main types of token-agnostic yields:

• Network Staking Fees: Delegating tokens to or running validators who secure and moderate blockchain middleware, including blockchains and web3 infrastructure
• Lending: Providing a token and allowing others to borrow it
• Liquidity Provisioning: Providing token(s) and allowing others to trade/utilize it
• Counterparty Liquidity: Taking one side of a ‘trade’, such as shorting volatility, and earning yield or losing principal depending on the outcome

Liquidity provisioning has high principal price risk as impermanent loss amplifies the potential principal value loss. Moreover, total return potential for liquidity provisioning is capped given that LPs purchase into the cheaper asset as the other one grows in price. The yield return ceiling tends to be lowest with lending, however they are easier to predict. Lenders are also susceptible to borrower default. By being an LP or acting as counterparty liquidity, there is risk that arbitrageurs or profitable counterparties will acquire a portion of deposited principal.

Yield Properties: The rate of yield is determined by demand and supply. The more that users want to access a use case, the higher the rates for capital deployers. There is no uniqueness to the yield-earning mechanism, for example overcollateralized lending yields are earned and distributed in broadly the same manner across various platforms. Yield rates are also broadly similar across platforms.

Token-agnostic yields are realized over events (e.g. transactions) or epochs (e.g. blocks), and can be accessed by either directly being a capital deployer, or holding a project token that has claim to protocol cash flows.

Examples of Token-Agnostic Yields

Token-Agnostic Yields Using Price-Stable Principal

• Lending: Deposit USDC into AAVE and earn a variable interest rate from borrowers
• Liquidity Provisioning: Provide USDC-DAI liquidity into Uniswap and earn swapping fees
• Counterparty Liquidity: Deposit USDC into Ribbon that sells ETH puts, or Cega which sells exotic derivatives to market makers

Token-Agnostic Yields Using Price-Fluctuating Principal

• Network Staking: Stake ETH to a validator to earn network fees through base fees and tips or, delegate LPT to an orchestrator that earn fees through transcoding services and receive ETH rewards
• Lending: Lend BTC or ETH into Euler and earn a variable interest rate from borrowers
• Liquidity Provisioning: Supply ETH/UNI liquidity into Uniswap
• Counterparty Liquidity: Purchase GLP and earn trading fees while acting as counterparty liquidity

Token-Agnostic Yields Through Protocol Earnings Distribution

Although not directly tied to a DeFi use case, fee distribution to token holders also presents yield opportunities:

• Stake SUSHI to earn platform swapping fees
• Switch BTRFLY to rlBTRFLY and earn ETH yields from platform fees
• Stake GMX to earn platform trading fees as well as esGMX rewards

Specific Cases of Token-Agnostic Yields:

• Bad case: A user deposits 1 ETH worth $1,000 into a decentralized options vault. The user was advertised 52% APR over the course of one week. The user earns 0.01 ETH because the option expired out-of-the-money, but the price of ETH dropped 25% during that time frame. The user’s portfolio value in USD terms has now dropped 24.3%. Token-agnostic yields are still vulnerable to negative total returns. In this case, price risk became realized.
• Bad case: A user lends $500 USDC to a borrower who collateralizes their NFT worth $1,000. At the end of the loan, the borrower refuses to repay the loan because the NFT is now worth $250. It is still possible to lose money despite not being exposed to price risk. In this case, counterparty default risk became realized.
• Good case: A user deposits USDC into AAVE lending pool and earns 5% on their deposit of $1,000. The realized return at the end of the year is 5%.
• Good case: A user provides $1,000 of DAI-USDC liquidity in Uniswap on Arbitrum. The realized return at the end of year is 5%.

Token-Specific Yields

There are three main token-specific yields:

• Token Holder Rewards: Offering yields to those who stake or hold that same token (usually DeFi or governance-based projects)
• Participatory Rewards: Offering yield to those who are using the project
• Network Staking Emissions: Offering yields validators and/or delegators who are contributing to proper functioning of blockchain middleware (Layer 1 or web3 infrastructure projects)

Yield Properties: Token-specific yields are dependent on a specific project, and thus realized returns can vary tremendously. Yield accrued over time is completely hypothetical and denominated in the project’s token price. The yield is only realized by selling from one party to another. Therefore, there are underlying assumptions to realizing the yield, such as the ability to sell on the open market, zero price movement, and proper functioning of underlying mechanism. Sometimes, token-specific yields contribute to the downfall of a project by diluting token value with heavy emissions, or by acting as an inflationary mechanism for a failed project structure. Other times, project-specific yields help bootstrap real demand which eventually leads to a thriving network.

Examples of Token-Specific Yields

Token-Specific Yields Using Price-Stable Principal:

• Participatory Rewards: Deposit stablecoins into Compound and earn $COMP tokens, or deposit stablecoin-pair liquidity into a new DEX and earn native tokens, or earn airdrops for participating in networks early

Token-Specific Yields Using Price-Fluctuating Principal:

• Token Holder Rewards: Lock CRV for veCRV to get boosted power and rewards or, Stake $APE to earn more $APE tokens
• Participatory Rewards: Earn airdrops for participating in networks early or, deposit wBTC/renBTC into Curve and earn $CRV rewards

Specific Cases:

• Bad case: A user bought and staked 100 $MOON tokens and began earning 50% APR (denominated in $MOON) over a year. The user now has 150 $MOON tokens. However, the price of $MOON at the start of the year was $1, and dropped to $0.50 by the end. The user portfolio value has dropped from $100 to $75. In order to realize the loss, the user sold all tokens onto the market with 10% slippage, and obtained $67.5. Although the advertised yield was 50%, the total return was actually -32.5%. In theory, if the price of $MOON stayed constant and the user sold with zero slippage, the user’s return would have been 50%. Advertised yields lead to misconceptions about risk. In this case, price risk and liquidity risks became realized.
• Good case: A user deposits capital into an AAVE lending pool on Optimism and earns $OP tokens. The user earns 5% APY in $OP tokens by the end of the year and sells onto the market for a 5% return, on top of the regular lending yields.
• Good case: A user stakes 1,000 $UP (hypothetical token worth $1), and earns 20% APY in emissions. At the end of the year, $UP token price appreciates 500% because the project has gained strong traction and is distributing strong revenue. The user now owns 1,200 $UP tokens worth $6,000 for a gain of 6x.

In summation, below is a chart that compares the two major types of yields and deeper implications:

Token-agnostic yields are generally more predictable, but will always have lower return potential: it is much easier to sell shovels in a gold rush than to pick the one metal that appreciates in value by 1,000x. Misconception risk is greatest with token-specific yields, as exemplified by Bancor and general pool2 farming yields. Bancor makes yields attractive during times when the mechanism functions, but creates disastrous outcomes when the mechanism does not — volatility precedes aggressive $BNT minting used to pay out impermanent loss. Token-specific yields are comparable to discovering metals during a gold rush, and realizing the gain by selling these metals onto the open market. These gains are only possible if another party is willing to be a buyer, meanwhile there is downward pressure on the unit price as people find more of the same metals.

What about Yield Farming?

Yield farming is the act of maximizing yield relative to risk, mainly through market inefficiencies.

Yield aggregators like Yearn Finance perform strategies that maximize value for their depositors by farming tokens of specific projects, which in turn hurts token holders of said projects. The Yearn USDC vault has one particular strategy that deposits $USDC onto Stargate, farms $STG for providing liquidity, and dumps $STG onto the market in order to return yield to the vault depositor. This is otherwise known as the dilution pressure problem for $STG token holders. This process could be portrayed as rate maximization.

There are also new yield aggregators that are focused on different types of yield farming actions, and these categories can be categories into three main buckets:

Rate arbitrage — borrowing assets at rate x, and lending assets elsewhere to earn higher than rate x. Or, holding spot amount x and shorting the equivalent to earn a funding rate spread.

Leveraged yields — borrowing an asset, converting it to another asset to increase productivity and therefore the overall yield.

Delta-Neutral strategies — hedging out price risk of underlying assets in order to gain sole exposure to the yield itself.

Examples of Yield Farming Strategies

Rate Arbitrage

• Collateralizing an asset, borrowing USDC for 2% on AAVE, and depositing in an undercollateralized lending platform such as Ribbon/Maple/TrueFi to earn 10%+
• Borrow DAI on AAVE, and deposit into dAMM to earn a surplus through dAMM native token-specific rewards
• A cash-and-carry vault that longs spot and shorts perps to collect a funding rate

Leveraged Staking

• Deposit BTC, borrow AVAX for sAVAX, and repeat
• Lever up on stETH yields through recursive ETH borrowing and staking through Index Coop

Delta-Neutral Strategies

• Liquidity aggregation layer on top of GMX that accepts single-sided deposits in $USDC
• A vault that hedges $GLP price exposure through shorting

In general, both rate arbitrage and leveraged staking yields contain relatively higher liquidation risk, for example many users were caught when stETH came off its theoretical peg. Cross-platform rate arbitrage yields can contain very high smart contract risks, given that new yields are often sourced from new non-battle-tested platforms. In terms of delta-neutral farming, there is no guarantee that the hedging mechanism will perfectly cancel price exposure (risks including liquidations, deviations in derivative price vs. spot price).

For a full breakdown of the types of yields that exist, feel free to check out the linked spreadsheet.

Given a deeper understanding of the yield landscape, what are some things to consider?

A particular inspiration for this article is the ongoing “Real Yield” narrative on Twitter (example threads: DeFi Edge, Miles Deutscher). Most “Real Yields” can be categorized as token-agnostic yields earned through supplying any of the two main principal tokens. While it is encouraging to see that projects are generating fees based on real user activity and there are some examples of cash flows being quite lucrative, “Real Yields” are a dangerous narrative because advertised APRs can change quickly given general reflexivity. If a protocol sees less usage, not only will the yields drop, but negative public perception can cause sell pressure and therefore total return can turn negative.

Another interesting question to ponder is whether or not protocols should be distributing earnings in the first place, as typically, revenue earned during a project’s early stages should be reinvested into growth. There are very few protocols that have reached a mature stage, if any. Moreover, in crypto, anything to do with yield has effectively attracted farmers, thus it is easy to optimize for short-term price gains due to hype rather than long-term fundamentals. While it may be challenging, understanding the founders’ motivations and intentions can better signal if a project is gearing for the long-term.

Explorable Yield Opportunities

Lending to NFT Holders (Price-Stable or Fluctuating Principal, Token-Agnostic Yields): Yield on lending to borrowers with NFT collateral can be extremely lucrative given its relative volatile nature. The monthly average loan APR floats well into the 30–50% range, and can be three times as high during periods of high borrowing demand.

Lenders can find offers that are priced in both DAI or ETH. Main risk is borrower default risk due to a collateral price drop and illiquidity risk if the collateral is acquired. Higher APRs compensate for higher NFT risk.

NFT Liquidity Provisioning (Price-Fluctuating Principal, Token-Agnostic Yields): For risk-seeking yield farmers, providing liquidity on NFTs is relatively quite lucrative if prices of the NFTs stay constant. There are currently two major platforms to explore: NFTX and Sudoswap.

NFTX allows users to have one-sided exposure to the NFT alongside providing two-sided liquidity. As of November 21st, several notable collections have double-digit liquidity provisioning APRs, including Mooncats, Miladies, Squiggles, and Forgotten Runes. Sudoswap allows users to set their own curves and fees.

Other opportunities:

• Undercollateralized lending to on either Goldfinch (non-crypto borrower focus) or Maple (crypto-native institutions)
• Providing stablecoin pair liquidity on new EVM chains to earn trading fees from bridgors, farmers, and experimenters
• Earning emissions tokens through platform deposits & borrows on new lending platforms
• Earning trading fees while acting as counterparty liquidity through tokens such as $GLP and Gains Network’s DAI vault
• Hedging or monetizing the premium on steady cash flows from blockspace through Alkimiya
• Shorting volatility and earning yields by depositing into Opyn’s Squeeth Crab Vault

In crypto, risk appetite correlated directly with yield and total return potential. The higher the upside, the higher the downside. The most important aspect is that users are given optionality, and we envision that more diverse yield sources will continue to grow as more primitives and opportunities come on-chain. If you’re building something in the DeFi space, don’t hesitate to reach out!

Disclaimer: Both the author of this piece, and the investment fund 1k(x), are not financial advisors. This piece and its content should not be construed as investment advice.

Many thanks to Dmitriy Berenzon for the extensive feedback on this piece.