How to convert a funding rate to APR, and why the number can mislead you
To annualise a funding rate, multiply the rate per settlement by the number of settlements in a day, then by 365. A rate of 0.01% every eight hours becomes 0.01% × 3 × 365 = 10.95% a year. This puts contracts on different schedules onto one scale, which makes them comparable. It does not tell you what you will earn or pay over a year, because it assumes today's rate never changes.
Annualised funding = rate per settlement × (24 ÷ hours between settlements) × 365. Use it to compare contracts at one moment. Do not read it as a yield: funding rates change at every settlement, and the largest annual figures tend to be the shortest-lived.
The formula, step by step
Annualising takes two conversions, and the first is the one people skip.
- Step 1, get a daily rate: daily rate = rate per settlement × 24 ÷ hours between settlements.
- Step 2, get an annual rate: annual rate = daily rate × 365.
Here is how the common schedules work out for a rate of 0.01% per settlement:
- Every 8 hours: 3 settlements a day, 0.03% daily, 10.95% annualised.
- Every 4 hours: 6 settlements a day, 0.06% daily, 21.90% annualised.
- Every 1 hour: 24 settlements a day, 0.24% daily, 87.60% annualised.
The same 0.01% on the screen covers an eightfold range. Skipping the first step and assuming eight hours for everything is the most common error in funding comparisons. It is also an easy one to make, because some tools report a raw rate per settlement without showing the schedule. Check the schedule for the exact contract: 206 of the 1052 contracts we read settle on different schedules depending on the exchange.
Why not compound it?
You may see two different annual numbers for the same rate. A simple annualisation multiplies. A compounded one treats each payment as reinvested: (1 + rate)^n − 1. At 0.01% every eight hours, simple gives 10.95% and compounding gives about 11.57%.
We use simple multiplication, and so do most funding tables, because funding does not compound by itself. A payment lands in your margin balance as cash. It does not increase your position unless you actively add to it. Position value, not account value, sets the next payment. If you want to model reinvesting, that is a trading decision with its own fees and should be calculated as such. Whichever method a site uses, check that you are not comparing a simple figure from one source against a compounded figure from another.
What the annual figure is good for
Its one honest use is ranking different contracts on a single scale at a single moment. It answers questions such as:
- Is the BTC rate on exchange A higher than the rate on exchange B now, once schedules are accounted for?
- How far apart are the highest and lowest rates for this contract, in percentage points?
- Is this rate unusually high compared with what we normally see?
For differences, use percentage points. If one exchange annualises at 12% and another at 8%, the gap is 4 percentage points. Saying "4% higher" would suggest 12% is 4% more than 8%, which would be 8.32%. This distinction matters when you are deciding whether a gap is worth acting on.
Why it is not a yield
A true yield describes what you expect to receive. An annualised funding rate describes what would happen if a single reading held for a year, which is not how rates behave. There are four reasons not to read it as income.
- The rate changes at every settlement. A contract showing 80% annualised today might show 10% tomorrow. The rate reflects a premium, and the payment itself attracts traders who close that premium.
- High figures are the least stable. The largest annual numbers tend to come from short bursts, thin contracts and the moments just after a large move. Those are the readings least likely to continue.
- Direction can reverse. Positive funding can turn negative, at which point the side that was receiving starts paying.
- The base is position value, while the risk is on margin and price. A 20% annualised funding income on a position that moves 5% in a day is not comparable to a 20% savings rate.
A better habit is to look at what the rate has actually done. An average over the past 30 days is far more informative than the latest reading, though it still does not predict anything. On contract pages such as ETH, we show the 30-day annualised average per exchange where enough readings exist.
Turning an annual figure back into something real
Annual numbers are hard to feel. Translating back to money for your actual holding period is usually more useful.
Expected funding = position value × annualised rate × days held ÷ 365
A 15,000 USDT long at 18.25% annualised, held for 10 days, costs 15,000 × 18.25% × 10 ÷ 365 = 75 USDT if the rate stays constant. At 3× leverage the margin is 5,000 USDT, so that is 1.5% of margin in 10 days. Now the number can be weighed against something: a 0.5% price move on the full position is also 75 USDT. Viewed this way, a rate that sounded alarming or attractive often turns out to be small next to ordinary price movement, or large next to the profit you were expecting.
Common errors when annualising
- Assuming eight hours. A one-hour contract annualised as if it settled every eight hours is understated by a factor of eight.
- Mixing percent and decimal. An API may return 0.0001 meaning 0.01%. Multiply by 100 for a percentage once, and only once.
- Using a running estimate as a settled rate. A figure shown before settlement can change before it applies.
- Comparing different moments. Two annual figures read hours apart are not a fair comparison during a fast market.
- Dropping the sign. Negative annualised funding means shorts pay. An absolute value hides the most important fact about it.
How we calculate it on this site
Our tables take each exchange's own published settlement schedule for each contract and apply the simple formula above. Where the schedule cannot be confirmed from the exchange, we show that separately and do not present an annual figure as if it were reliable. Current comparisons use readings less than 30 minutes old. The full rules are on our methodology page.
Common questions about annualised funding
Is APR or APY the right term for funding?
APR is the closer term, because the calculation is simple multiplication with no reinvestment. APY implies compounding, which funding does not do by itself. Some sites use the labels loosely, so check the arithmetic rather than the label. If you multiply the rate by the settlements in a year and get their figure, it is an APR whatever they call it.
What is a normal annualised funding rate?
On Binance, a contract sitting at the default of 0.01% every eight hours annualises to 10.95%. In calm markets, many large contracts on many exchanges sit close to that level or below it. It is common to see much larger figures on thinly traded tokens, and these are usually short-lived. Comparing a contract with its own past average tells you more than comparing it with a fixed "normal" number.
Can I add funding income to a staking or lending yield?
Be careful. Staking and lending have their own terms and risks, and funding is not a contracted rate at all. If you are holding spot and shorting the perpetual, the two income streams come from different positions with different risks, and the funding part can turn into a cost at any settlement. Model them separately and use a conservative historical average for funding, not the latest reading.
Why does my calculated figure differ slightly from the one on a website?
Small differences usually come from rounding, from reading the rate at a slightly different moment, or from one source using 360 or 365 days. Larger differences usually mean a different schedule was assumed, which is worth checking, or that one figure is an estimate and the other is a settled rate. A site that shows its formula and its schedule makes this easy to check.
Should I look at the rate per settlement or the annual figure when placing a short trade?
For a short trade, use the rate per settlement and the time until the next settlement, because those tell you what you will actually pay. The annual figure is built for comparing contracts and says little about a position that lives for a few hours. If your trade will cross a single settlement, the cost is one payment, whatever the annualised number happens to say.
A quick way to check any annual figure
- Find the rate per settlement and the schedule for that exact contract.
- Multiply by settlements per day, then by 365, without compounding.
- Keep the sign.
- Compare with a historical average, not just the latest reading.
- Convert back to money for your position and holding period.
Related reading: How to compare funding rates across exchanges · What a funding rate is · Funding rate arbitrage
Sources
Other guides
What is a funding rate in crypto perpetual futures? A plain explanation
A funding rate is a regular payment between traders holding opposite sides of a perpetual futures contract. It exists because a perpetual contract never expires, so something else has to keep its price close to the underlying asset. When the rate is positive, longs pay shorts; when it is negative, shorts pay longs. The exchange sets the formula but, on the major exchanges, the money moves between traders.
How to compare funding rates across crypto exchanges: 8h vs 4h settlement
You cannot compare two funding rates by looking at the percentages alone. A rate only means something once you know how often it is charged, whether you are long or short, and whether the two readings describe the same contract at the same moment. The same +0.01% costs twice as much per day on a four-hour schedule as on an eight-hour one, and it earns a short exactly what it costs a long.
What does a negative funding rate mean? Who pays, and what it does not tell you
A negative funding rate means traders holding short positions pay traders holding long positions at the next settlement. It usually happens when a perpetual contract trades below its index price, which often reflects heavy shorting or hedging. It does not mean the price is about to fall or about to rise, and it is not free income for longs, because the price can move against them by far more than the payment.
Funding rate arbitrage explained: how it works and what eats the profit
Funding rate arbitrage means holding two offsetting positions in the same asset so that price moves roughly cancel, and collecting funding payments on one side. The classic version buys spot and shorts the perpetual. A second version shorts the perpetual on the exchange with the higher rate and goes long on the exchange with the lower rate. It is often described as risk free. It is not. The risks are simply different from the risk of an ordinary directional trade.
Mark price vs last price vs index price: which one liquidates you?
Perpetual futures show three prices that are easy to mix up. The last price is where the most recent trade happened. The index price is an average of spot prices across several exchanges. The mark price is an exchange's estimate of what the contract is fairly worth, and on the major exchanges it is what decides your unrealised profit and whether you are liquidated. Knowing which price does what explains why a position can be liquidated even though the chart never touched your liquidation level.
How much does a market order really cost? Fees, spread and slippage
The fee on an exchange's fee page is only part of what a market order costs. The rest comes from the price you actually receive, which depends on the direction of your trade, your order size and the depth of the order book at that moment. To compare exchanges fairly, measure the average fill price against one reference price, then add your trading fee on top, and be careful not to count the spread twice.
How to estimate slippage from order book depth before you trade
Slippage is the difference between the price you expected and the average price you actually got. You can estimate most of it before placing an order, because the order book shows how much is available at each price. Add up the levels on the side you will consume until they cover your size, calculate the average price, and compare it with the mid price. The estimate is only as good as the snapshot it came from, and books thin out at the moments you are most likely to need them.
Maker and taker fees explained: how to compare crypto exchange trading fees
A maker adds an order to the book that waits to be filled. A taker fills against an order already there. Most exchanges charge takers more than makers, because makers supply the liquidity others trade against. Comparing exchanges by the headline fee alone is misleading: you need to know which side you usually trade on, what tier your volume actually puts you in, and how the fee compares with slippage and funding on the same trade.
How to check if a crypto exchange is down: a step-by-step guide
When an exchange stops working for you, the first question is whether it is down for everyone or only for you. You can usually answer that in a couple of minutes: check the exchange's own status page, check an independent measurement, try a different network, and test the public API directly. Each source sees something different, so one green page does not prove everything is fine, and one failed request does not prove the exchange is down.
Is an exchange really down? How to read API uptime and outage records
Uptime describes the share of successful requests to particular services, over a particular period, from a particular network route. It does not establish whether you can place an order, withdraw funds or trust an exchange with your assets. A figure like 99.9% is useful only if you know what was tested, from where and for how long, so this guide explains how to read those details on our pages and on anyone else's.
What to do when a crypto exchange goes down while you have an open position
If an exchange stops responding while you hold a position, your first job is to find out what is actually broken, the second is to stop making it worse, and the third is to reduce your risk by whatever route still works. Do not keep resending the same order, because earlier attempts may have gone through. Check the mobile app and the API as alternative routes, consider hedging elsewhere if your exposure is large, and record what happened in case you need to raise it later.
Why exchange APIs reject timestamps: clock offset and stale quotes explained
A fast response can contain an old quote. Round-trip latency, clock offset and quote age describe different parts of a request, and mixing them up can make a clock mismatch look like a delayed market feed. This guide explains each one with a worked example, shows how we estimate them, and covers what to do when an exchange API rejects your request because of the time.
Binance error -1021 timestamp: causes and how to fix it
Binance API error -1021 INVALID_TIMESTAMP means the timestamp on your signed request falls outside the window Binance will accept. The message usually reads "Timestamp for this request is outside of the recvWindow" or "was 1000ms ahead of the server's time". In most cases your own clock is the cause, so the fix is to synchronise it, measure the offset against Binance's server time, and handle any remaining network delay properly. A larger recvWindow can hide the problem, but it should not be your first move.
Crypto exchange API rate limits: what error 429 means and how to stay under it
HTTP 429 means you have sent too many requests in a given time, and the server is asking you to slow down. It is not an outage. Each exchange counts requests differently: Binance uses a weight for each endpoint and bans IP addresses that keep going after a 429, Bybit blocks an IP that passes 600 requests in five seconds, and OKX sets limits per endpoint and returns error code 50011. Staying under the limit means reading the response headers, backing off properly and using WebSockets where you can.
Why an exchange API returns 403 or 451: blocks and regions
A 403 means the server understood your request and refuses to carry it out. A 451 means access is refused for legal reasons. On crypto exchange APIs, a 403 can be a web firewall, a request limit, an unauthorised key or a regional block, so the code alone does not tell you which. The response body, the exchange's documentation and a comparison from another network usually do. A real restriction applies to you for a reason, so the right response is to find out whether the service is available where you are.
Before you open a crypto exchange account: a checklist by country
Before you open an account on a crypto exchange, check four things: whether the exchange is allowed to serve people where you live, whether your country's regulator lists it, whether the site in front of you is the real one, and how the exchange has been running lately. The sign-up form takes ten minutes. Recovering from the wrong choice can take months, and sometimes it is not possible at all. This guide works through each check, using Binance and OKX as worked examples because they are the two exchanges readers ask us about most.
How to open and verify a Binance account: a step-by-step guide
Opening a Binance account takes four steps: register with an email address, a phone number or a Google, Apple or Telegram account; complete identity verification; secure the account; and make a first deposit. Registration takes a few minutes. Binance says verification is usually reviewed within 48 hours. This guide follows each step as Binance's own help pages describe it on 8 November 2026, with screenshots of those pages, and points out the details that most often cause trouble later.
How to open and verify an OKX account: a step-by-step guide
Opening an OKX account takes four steps: register with your country of residence, an email address and a phone number; complete identity verification; secure the account; and make a first deposit. OKX says it reviews verification applications within 24 hours. Which OKX company you end up with depends on where you live, and that decides the address you log in at, the fees you pay and the products you see. This guide follows OKX's own help pages as they stood on 8 November 2026, with screenshots, and flags the choices that are hard to undo. OKX shows some help pages differently by region; we read the versions served at okx.com/en-us, and the rules shown in your own account after you log in are the ones that apply to you.