Yield Farming Basics

Yield farming is a loose label for a set of mechanisms that pay a return for supplying capital to a DeFi protocol. This page explains how those mechanisms work and where the numbers come from. It is not a plan for doing it, and it does not point you at anywhere to go and do it. If you want to know why the site is written that way, our editorial and advertising policy sets it out.

Understanding the machinery is useful whether or not you ever use it. Most of the confusion around yield farming comes from treating a single advertised percentage as though it meant one thing, when in practice it is several different payments bundled together, each with a different source and a different reason to stop.

What the term actually describes

"Yield farming" is not a product. It describes the general activity of moving capital between DeFi protocols in pursuit of whatever return is currently on offer, and it covers several unrelated mechanisms that happen to produce a payment.

The common thread is that you deposit an asset into a smart contract, the contract does something with it, and you receive something back over time. What the contract does, and where the payment comes from, varies enormously. Lumping them together under one word is the source of most beginner misunderstanding, because the risks are not the same and neither are the reasons the return might disappear.

If the underlying vocabulary is unfamiliar, key DeFi concepts covers tokens, smart contracts and gas, and what is DeFi covers the wider picture.

Where the yield comes from

There are only a few genuine sources. Every advertised figure is some combination of them.

Trading fees. An automated market maker charges a fee on each swap and distributes it to whoever supplied the pool's liquidity. This is revenue: it comes from users paying to trade, and it exists because the pool performed a service. It scales with trading volume, so it rises and falls with how busy the market is. Understanding liquidity pools covers the mechanism in detail, and how to use Uniswap shows the swap side of the same system.

Borrowing interest. A lending protocol matches depositors with borrowers and pays depositors a share of the interest borrowers pay. This is also revenue. The rate is usually set algorithmically against utilisation, meaning the proportion of deposited funds currently borrowed. When most of a pool is lent out, rates rise to attract more deposits and discourage further borrowing. When little is borrowed, rates fall towards nothing.

Consensus rewards. Proof-of-stake networks pay validators for producing and attesting to blocks. That payment is newly issued currency plus a share of transaction fees, and it exists to pay for network security. Staking derivatives pass a version of it through to holders.

Token emissions. A protocol prints its own token and hands it to depositors. This is the source that causes the most confusion, and it deserves its own section.

Emissions are not revenue

When a protocol pays you in a token it created, nothing has been earned in the ordinary sense. No user paid a fee, no borrower paid interest. The protocol issued new units of something it controls and gave them to you, and the value of that payment depends entirely on what someone else will pay for the token later.

This is a deliberate mechanism, not a trick. A new protocol has no users, and no users means no fees, and no fees means no reason to supply it with capital. Emissions solve the bootstrapping problem by paying people to show up before the revenue exists. That is a coherent thing to do.

But it has consequences built into the design. Emissions dilute existing holders, because supply rises. They are usually scheduled to decline, so the advertised rate at launch is not the rate later. And the headline percentage is calculated by valuing the emitted token at today's market price, which assumes you could sell all of it at that price. If enough recipients try to, you could not.

The practical point is this: when you see a very large advertised return, the first question is which of the four sources it comes from. A figure built mostly on emissions is a forecast about a token price. A figure built on fees or interest is a share of money that actually changed hands.

Why the advertised number moves

Two numbers are usually quoted and they mean different things.

APR is the simple annualised rate, with no assumption that anything is reinvested. APY assumes rewards are collected and redeposited at some frequency, so compounding is included. For the same underlying return, APY is the larger number, and the more often the calculation assumes compounding, the larger it gets. Comparing one protocol's APY against another's APR is not a comparison.

Both are extrapolations from current conditions. They take whatever is happening right now and project it across a year. A fee-based rate reflects this week's trading volume. A lending rate reflects this moment's utilisation. An emissions-based rate reflects today's token price and today's emission schedule. None of these is a commitment, and all of them can change without anyone doing anything wrong.

There is also a structural reason rates fall. A genuinely high return attracts capital, and the same payment shared among more depositors is a smaller payment each. High advertised rates tend to be self-correcting, and the rate that existed when you read about it is often not the rate by the time capital has arrived.

What happens mechanically when you deposit

Understanding the receipt is what makes the rest of the system legible.

When you deposit into most DeFi protocols you do not simply have a balance recorded somewhere. You send your assets to a smart contract and the contract mints you a token representing your claim. In a liquidity pool that is an LP token; in a lending protocol it is an interest-bearing deposit token; in liquid staking it is a staking derivative.

That receipt token is the claim. Holding it is what entitles you to withdraw. Losing it, or approving a contract that takes it, means losing the position. It also usually accrues value in one of two ways: either its balance grows, or its redemption rate against the underlying asset grows.

Composability, and why risk stacks

Because those receipts are ordinary tokens, they can be deposited into other protocols. A liquidity position can be staked somewhere else. A staking derivative can be used as collateral to borrow. That borrowed asset can be deposited somewhere else again.

This is what people mean by composability, and it is the genuinely novel property of the system. It is also where risk accumulates in a way that is easy to miss. Each layer adds the failure modes of a further contract, and they are not independent of one another. A problem at the bottom of the stack propagates upward through every position built on it. A return quoted for the whole arrangement is a return on the combined risk of every layer, not on the last one.

The risks, as mechanisms

These are not warnings bolted onto the end. They are properties of how the system is built.

Smart contract risk. The contract does exactly what its code says, including when the code is wrong. An audit is a review at a point in time by people who may miss things, not a guarantee. Contracts that can be upgraded can be changed after you deposit, and contracts that cannot be upgraded cannot be fixed after a flaw is found. Both are trade-offs rather than one being safe.

Approval risk. Depositing usually requires granting a contract permission to move your tokens, and that permission persists after the transaction until it is revoked. This is the mechanism behind a large share of real losses, and it is covered properly in DeFi security best practices and common DeFi scams.

Impermanent loss. Supplying a pool of two assets means holding a changing ratio of them as their relative price moves. The pool rebalances by being traded against, so you end up with more of whichever asset fell. The resulting position can be worth less than simply holding both assets would have been, and fees earned may or may not cover the difference. It is called impermanent because it reverses if prices return, and that name understates it, because nothing guarantees they do.

Liquidation. Borrowing against collateral means the protocol monitors a ratio and sells your collateral if it breaches a threshold. This is automated and does not wait for you. It is triggered by a price feed, so the behaviour of that feed, its sources, its update frequency and its resistance to manipulation, is part of your risk whether or not you ever look at it.

Exit risk. Withdrawing requires that there is something to withdraw into. A pool that is thin, or heavily utilised on the lending side, or subject to a withdrawal queue, may not let you leave at the moment you want to. Some positions have explicit lock-ups; others have effective ones that only become visible under stress.

Governance and custody. Someone can usually change parameters, and sometimes pause or upgrade contracts. Whether that power sits with a multi-signature wallet, a token vote or a single key is a real difference in what you are exposed to.

Automated strategies

Aggregators and autocompounding vaults exist because claiming and redepositing rewards manually costs a transaction fee each time, which at small position sizes can exceed the reward. A vault batches that work across many depositors so the cost is shared.

Mechanically, the vault is another contract holding your assets and issuing you another receipt. It typically takes a fee on performance, on withdrawal, or both. So it adds a contract layer and a cost layer in exchange for removing a manual step. Whether that is worthwhile is an arithmetic question about position size and transaction costs, and it is not the same question as whether the underlying return is worth having.

Realised return versus advertised return

The gap between the two is systematic, not bad luck, and it comes from several places at once.

Transaction fees are paid on entry, on each claim, on each compound and on exit, in the network's own currency and at whatever it costs at that moment. The advertised rate applies only from the moment capital is actually deposited. Emitted tokens are valued at the price you can sell them for, not the price quoted when you deposited. The underlying assets change in value independently of any yield. And for pooled positions, the ratio you withdraw is not the ratio you deposited.

A return calculated properly compares what you could withdraw today against what you put in, measured in the same unit, with all costs included. Very few advertised figures are calculated that way, and many are not intended to be.

UK tax and regulation

Both matter and neither is covered properly by recalling a number from memory, because the thresholds and the rules move.

HMRC publishes its position in the Cryptoassets Manual, which is the right source to read rather than any summary. Different activities are treated differently, record-keeping obligations are real, and the treatment of a payment can depend on its nature rather than what a protocol calls it. If the amounts are meaningful, this is a question for an accountant rather than a website.

On the regulatory side, the UK regime is changing. The FCA's authorisation timetable is covered in our news piece on the authorisation window for UK crypto firms.

What this page deliberately does not do

It does not tell you which protocols to use, it does not rank them, and it does not lay out a plan for getting started. That is a settled editorial position rather than an omission, and the reasoning is published in full in our editorial and advertising policy.

The short version: explaining how a mechanism works is useful, and steering readers toward a speculative activity is not something this site does.

Where to read next

Understanding liquidity pools is the natural next page, since pooled liquidity underpins most of what is described here. Key DeFi concepts fills in vocabulary. DeFi security best practices and common DeFi scams cover the approval and phishing mechanics that cause most real losses.