What is correlation in forex trading? How paired currencies move together

What correlation in forex actually measures

Correlation in forex is a number between -1 and +1 that measures how two currency pairs move together, and it is the single most useful risk concept a retail trader can learn. A correlation of +1 means two pairs move in lockstep, 0 means they move independently, and -1 means they move in opposite directions (Myfxbook; Dukascopy).

I think of correlation as the answer to a question I ask before every second trade: am I already long this? If the new pair is highly correlated with one I hold, the honest answer is usually yes.

The reason it matters is that currency pairs share currencies, and shared currencies mean shared risk. Two pairs both quoted against the dollar are never fully independent of each other.

How the correlation coefficient works

The coefficient is calculated over a window of recent price data, commonly 30, 50 or 100 periods, and it expresses how tightly two pairs moved together over that window. The closer to +1 or -1, the stronger the relationship (Babypips; Investopedia).

A positive coefficient means the pairs rose and fell together, a negative one means one rose while the other fell, and a reading near zero means no consistent relationship. The strength matters more than the sign, because a +0.3 correlation is noise while a +0.9 correlation is effectively the same trade.

I check the coefficient on a tool like Myfxbook or Babypips before I add a second pair to a position. The number takes ten seconds to read and saves me from doubling up on a single idea.

One nuance worth knowing: the coefficient is historical, not a guarantee. It tells you what pairs did together recently, which is useful, but it is not a contract about what they will do on the next candle.

Classic forex correlations: the pairs that move together and apart

A few correlations are famous because they are stable and large, and they are the ones every trader should know before sizing a second position. The table below sets out the classic relationships with their typical coefficients (Myfxbook; Dukascopy).

The logic behind each one is structural. AUD/USD and NZD/USD move together because Australia and New Zealand are neighbouring commodity economies whose currencies rise and fall with similar forces (Dukascopy).

EUR/USD and USD/CHF move opposite because the Swiss franc and the euro share a region, so when the dollar weakens against the euro it tends to weaken against the franc too, flipping USD/CHF lower (Dukascopy; Investopedia). I keep these five or six classic pairs in my head as a quick filter, because they explain most of the accidental concentration in a typical retail forex book.

Why correlated positions are really one position

This is the trap correlation exists to expose. A trader who is long EUR/USD and long GBP/USD thinks they hold two positions, but with a +0.8 correlation between them, they effectively hold two copies of the same dollar-short bet (PriceActionNinja).

If the dollar rallies, both stop out together, and the loss is roughly double what the trader planned for. The two positions felt diversified because the pairs have different names, but the risk was identical.

I size correlated positions as if they were one. If my rule is to risk 1% on a trade, two +0.9-correlated longs together risk closer to 2%, which is why I either drop one or halve each size.

Concentration is not always bad, but it should be a choice rather than an accident. Correlation turns the accident into a visible number you can manage.

A worked example: reading your book's correlation

Suppose you are long EUR/USD at your normal 1% risk and you now want to add a long on AUD/USD. The two share no currency directly, so they feel independent, but both are dollar-shorts in a risk-on market.

You check a correlation tool and find EUR/USD and AUD/USD sitting at +0.7 over the last 50 periods. At +0.7 the two pairs move together most of the time, so adding the AUD/USD long at full size roughly doubles your dollar-short exposure.

I would either skip the AUD/USD trade, take it at half size, or wait for the correlation to drop. The point is that the +0.7 changed my decision, which is the whole reason to look the number up.

The same logic applies to a quote currency you already hold long elsewhere in the book. If the shared side is the same, the positions are partial clones of each other.

Using negative correlation to hedge

The flip side of the trap is the hedge. A negatively correlated pair moves opposite to your main position, so holding both can smooth the equity curve when the market goes against you (Investopedia).

A long EUR/USD position paired with a long USD/CHF position is a partial hedge, because the two tend to move opposite. It is not a perfect hedge, since correlations are never exactly -1, but it dampens the swings on a rough day.

I use negative correlation for risk reduction, not for profit. The point of a hedge is to cap the damage on the main position, and a -0.8 correlation does that reasonably well without requiring a separate short on the same pair.

The danger is that the correlation breaks exactly when you need it most, which is the next section's problem.

Correlation is not causation, and correlations break

Correlations are not laws of physics, they are habits the market has fallen into, and habits change. A coefficient that read +0.9 for a year can collapse to near zero in a single risk-off shock (Investopedia; PriceActionNinja).

Central-bank divergence is the most common breaker. If the Fed and the ECB head in opposite directions, EUR/USD and GBP/USD can decouple sharply even though they usually move together, because the euro and sterling react to different rate paths.

Flight-to-safety events are the other big one. In a panic, the dollar strengthens against almost everything, which inverts the usual relationships and catches traders who hedged on yesterday's coefficient.

I re-read my correlations monthly and never assume the current number is permanent. The coefficient is a snapshot, and the regime around it can shift faster than the math updates.

Correlation changes over time

A correlation coefficient is a snapshot, and the relationship it describes drifts as the market regime changes. Even the strongest links shift over months, so a reading you banked on last quarter may no longer hold by this one (Investopedia; PriceActionNinja).

This is why a single reading is dangerous. A coefficient of +0.9 today does not promise +0.9 next month, and the break usually happens at the worst time for anyone hedging on it.

I check correlations on a rolling basis and watch for divergence between a short and a long window. When the short-window reading drops well below the long-window one, the relationship is breaking, and that is the moment to stop trusting the hedge.

Treating correlation as a dynamic number rather than a fixed rule is what separates traders who get caught by a break from those who see it coming.

Correlation across timeframes

The correlation between two pairs also depends on the timeframe you measure it on, which is a subtlety many traders miss. Two pairs can be tightly linked on a daily chart and loosely linked on an hourly chart, because short-term noise drowns out the underlying relationship (Babypips; Myfxbook).

A swing trader should care about the daily correlation, because that is the horizon of the trades. A scalper should care about the intraday correlation, which may be far weaker and noisier.

I match the correlation window to my holding period, so I am sizing and hedging against the relationship that actually governs my trades. Measuring the wrong timeframe gives a number that looks precise but answers the wrong question.

The same two pairs can tell different correlation stories on different charts, which is why stating the timeframe is as important as stating the coefficient.

How to use forex correlation in your trading

Use correlation to manage risk before you use it to find trades. The first job is to stop accidentally doubling up, and the second is to build a book of positions that are not all the same bet.

Before I add a pair, I check its correlation to what I already hold, and I treat anything above +0.7 as the same trade for sizing purposes. The stop distance on each is its own number, but the portfolio risk is what correlation exposes.

For diversification, look for pairs with low or negative correlation to your core position, so a single currency move cannot take out the whole book at once. The goal is positions that zig while others zag.

Finally, treat every correlation reading as conditional on the current regime. When volatility spikes or central banks diverge, the relationships you trade on can break, and the hedge that worked last month can become the loss this month.