Yield Vaults: What You Underwrite When You Deposit

A yield vault takes deposits and routes them through strategies that earn a return, rebalancing as opportunities change. The convenience is genuine: it removes the work of finding, entering and compounding positions. What it does not remove is the risk of those positions, and it adds two of its own — the vault contract itself, and the authority to change the strategy after you are in.

Four layers of exposure

The vault contract is the first. It holds your deposit, and a flaw in it loses the funds regardless of how sound the underlying strategies are.

The strategy contracts are the second: the code that moves capital between venues, each with its own logic and its own bugs.

The underlying protocols are the third, and usually the largest. A vault farming a lending market is exposed to that market's bad debt, its parameters and its governance. A vault across five protocols is exposed to five sets of all of that.

The fourth is the one that gets skipped: the authority to change the strategy. On most vaults, a strategist or governance can redirect capital into protocols you have never assessed, without your consent and without a withdrawal window. You did not underwrite that protocol. Your capital is in it.

On curated lending vaults that authority usually belongs to a curator or risk manager, and those firms have records a depositor can read. Our ranking of DeFi risk managers scores each one on the losses that happened under its mandate and on how it responded.

The composition question

The useful mental model is not a savings account but an actively managed fund with no prospectus. The natural questions follow from that framing.

What does the strategy hold, and can I see it. Some vaults publish current allocations on-chain in a form anyone can read. Others describe the approach and leave the positions opaque. The difference is enormous and rarely reflected in the yield.

Who decides the allocation, and what constrains them. A vault with a published mandate — permitted protocols, maximum concentration, minimum audit standard — is a different instrument from one where a strategist has discretion. Both exist and both are described the same way in the interface.

What is the worst realised drawdown, and how was it handled. A vault that has never lost money has either been well run or been lucky in a benign period, and the way to tell is whether the operator publishes the incidents it has had.

Fees, and why net return is the only number

Vaults charge in two ways and both need to be in the comparison.

A management fee accrues on assets regardless of performance. A performance fee takes a share of the return generated. The second is the larger drag in most cases, and because it applies to the return rather than the balance, it does not appear as a deduction anyone notices.

The comparison that matters is not the vault's advertised rate against another vault's advertised rate. It is the vault's net-of-fee return against the return you would earn holding the underlying position directly, which is the alternative you actually have. On expensive chains, auto-compounding frequently justifies the fee outright because manual compounding costs more in transaction fees than the vault charges. On cheap chains that argument weakens considerably, and what you are paying for is the strategy selection rather than the mechanics.

Where auto-compounding earns its keep

This is worth isolating because it is the one part of the value proposition that is arithmetically checkable.

Compounding a position requires claiming rewards, swapping them and redepositing, each of which costs a transaction fee. Doing this optimally requires knowing the point at which the gain from compounding exceeds the cost, which changes with position size and network conditions. A vault does it across pooled capital, so the fixed cost is shared.

For a small position on a network with meaningful transaction costs, this alone can exceed the vault's fee, and the strategy selection is a free extra. For a large position on a cheap network, the maths inverts and you are paying primarily for convenience.

Before depositing

• List every protocol the strategy touches. If you cannot, you cannot scope the exposure.
• Find out who can change the allocation and whether there is a timelock or a withdrawal window before changes take effect.
• Check whether the vault publishes per-strategy risk assessments rather than only a yield figure.
• Compute the net-of-fee return against holding the underlying directly, including what compounding would cost you manually.
• Look for a record of past losses and how they were handled. Silence is not the same as an unblemished record.

The vaults worth using tend to be the ones that make this analysis easy, because publishing allocations and risk assessments is a choice an operator makes when the answers are defensible. Opacity is rarely an accident of engineering.