Bitcoin shows a different price on every exchange because each exchange is its own separate market. The price you see is the last trade, or the best current bid and ask, in that exchange's order book. Those trades depend on that venue's buyers, sellers, liquidity and quote currency. Arbitrage traders push prices back towards each other. They can only do it as fast as money and coins can move between venues, and as far as fees and rules allow. Inside one country the gaps are usually small. Between countries with capital controls, research has measured gaps of tens of percent.
There is no single official crypto price
Bitcoin trades on many independently owned venues in parallel, and no central exchange sets a reference price. Makarov and Schoar, in the Journal of Financial Economics (2020), describe cryptocurrencies as traded on "many nonintegrated exchanges" that are independently owned and run side by side across countries.
Every centralised exchange keeps its own order book, a live list of buy orders (bids) and sell orders (asks) from its own users. When a bid meets an ask, the trade sets that venue's latest price. A second exchange may have different orders at the same moment, so it shows a different price.
Price aggregators publish a blended figure rather than one exchange's price. CoinGecko's published methodology works like this:
- It ranks every active trading pair for a coin by volume and keeps the top 600.
- It removes prices far from the majority using the median absolute deviation.
- It combines the remaining tickers into a volume-weighted average, so higher-volume venues count for more.
This average is a useful reference point. You cannot trade at it.
What drives the gaps
Liquidity and order book depth
Liquidity is how much you can buy or sell without moving the price. On a venue with thin order books, a large order uses up the best offers and fills at worse prices further down the book. A deeper venue absorbs the same order with less movement. As a result, smaller venues can show larger and more sudden gaps.
Who is trading, and when
Makarov and Schoar split exchange order flow into a common part and a venue-specific part. The common part explained 80% of bitcoin returns. The venue-specific parts helped explain the arbitrage spreads between exchanges. Put simply, buying or selling pressure from one venue's own users can push its price away from the rest of the market.
Capital controls and borders
This is the largest documented cause. Makarov and Schoar studied 34 exchanges in 19 countries, focusing on the 15 most liquid, from January 2017 to February 2018. They found that price deviations were "much larger across than within countries", and they linked this to capital controls that slow the movement of arbitrage capital.
- South Korea (the "kimchi premium"): the daily average price ratio between the US and South Korea over the study period was below 15%. Between December 2017 and February 2018 it reached 40% on several days.
- Japan and Europe: the ratio was about 10% for Japan and 3% for Europe.
- Duration: these deviations lasted "from several hours to days and weeks".
- Size: the authors estimate that at least $2 billion of potential arbitrage profit went untaken between December 2017 and February 2018.
They also found that cross-country gaps moved together and widened when bitcoin rose sharply.
Quote currency and stablecoins
A BTC/USD price and a BTC/USDT price are not the same thing. The first is quoted in US dollars. The second is quoted in a stablecoin whose market price can move slightly away from $1. Here is a hypothetical example: if BTC/USDT is 97,500 and one USDT trades at $0.998, that BTC price equals about $97,305. Before comparing two venues, check that both use the same quote currency.
Fees and transfer friction
The displayed price is a trade or quote price. Trading fees, the bid-ask spread and withdrawal fees come on top of it. These costs do not create a gap by themselves, but they decide whether a gap is worth closing. If the cost of buying on one venue, moving the coins and selling on another is higher than the price difference, the gap can stay open.
Moving coins also takes time. Bitcoin's difficulty adjusts every 2,016 blocks, with a target of two weeks (1,209,600 seconds). That works out to one block about every 10 minutes on average, and exchanges usually wait for confirmations before crediting a deposit. The price can move while the coins are in transit.
Why arbitrage rarely works for you
In theory, arbitrage is simple: buy where it is cheap and sell where it is expensive. In practice, the opportunities have shrunk. Crépellière, Pelster and Zeisberger (Journal of Financial Markets) found that crypto arbitrage opportunities still exist. Their size, however, fell sharply from April 2018, and the authors conclude that the remaining gaps are barely profitable to trade. They also link more informed trading to fewer arbitrage opportunities.
Most gaps you can see today are therefore either:
- too small to cover fees, spread and transfer costs, or
- large but hard to reach, because of capital controls, slow withdrawals or local account requirements.
How these strategies work is covered in our guides to [crypto arbitrage](/blog/arbitrage-cryptocurrency/) and [whether crypto arbitrage is legal](/blog/is-crypto-arbitrage-legal/).
Is that gap only the stablecoin?
Enter a BTC/USDT price, the dollar value of one USDT and a BTC/USD price to see how far apart the two prices really are once both are in US dollars.
| Bitcoin price on the stablecoin pair (BTC/USDT) | Market price of one USDT in US dollars | Bitcoin price on the US dollar pair (BTC/USD) | Gap between the two prices after converting USDT to dollars |
|---|---|---|---|
| 97,500 USDT | 0.998 $ | 97,000 $ | 0.31 % |
| 97,500 USDT | 1 $ | 97,000 $ | 0.52 % |
| 97,500 USDT | 0.998 $ | 97,305 $ | 0 % |
- All three prices are your own inputs or the article's hypothetical example, not measured prices from any venue.
- A positive result means the stablecoin pair is still higher after conversion; a result near zero means the stablecoin's price alone explains the gap.
This only converts the quote currency: it leaves out trading fees, the spread, withdrawal fees, order book depth and transfer time, and cannot show whether a gap exists or lasts.
A worked example (hypothetical)
Bitcoin trades at about $97,000 on Exchange A and Exchange B. A very large market buy order hits Exchange A, works through its order book and lifts its price. Exchange B's order book has not changed yet, so for a short time B is cheaper. Arbitrage traders buy on B and sell on A until the two prices come close again. The gap closes only as far as the costs in the previous section allow.
The [market makers](/blog/what-is-market-maker/) who place quotes, and the [automated market makers](/blog/what-is-amm-automated-market-maker/) used by decentralised exchanges, each feed into this process in different ways.
Practical checks before you trade
- Compare with an aggregate price. A volume-weighted average from an aggregator shows whether one venue is unusually far from the rest.
- Look at depth, not only the last price. The price shown is not guaranteed for a large order. Investor.gov notes that in fast-moving markets, the price at which a market order executes "often deviates from the last-traded price". Parts of a large order may fill at different prices. A limit order caps your price but may not execute.
- Count the total cost. Add trading fees, the spread and any withdrawal fee before you treat a lower price as cheaper.
- Match the quote currency. A USD pair and a stablecoin pair can differ only because the stablecoin is not exactly at $1.
Frequently asked questions
Which exchange has the "real" bitcoin price? None. Each venue's price is real for that venue. Aggregated prices, such as CoinGecko's volume-weighted average, are reference figures, not prices you can trade at.
How big are the differences? Within one country they are usually small. Across borders, the research above measured an average US–Korea price ratio below 15% in 2017–2018, with peaks of 40% on several days. Later research found that arbitrage gaps shrank sharply after April 2018.
Can I profit from price differences? According to Crépellière, Pelster and Zeisberger, the remaining gaps are barely exploitable once costs are counted. This article explains how prices form and is not trading advice.
Limitations: the measured gap sizes come from 2017–2018 data and one later study; current gaps between specific venues were not measured for this article.
