EducationMarket structure~19 min readUpdated 29 September 2026
The short answer
Correlation measures how two markets move together. +1 means they move identically. −1 means they move opposite. 0 means no relationship. EURUSD is heavily negatively correlated with DXY (roughly −0.95). Gold is negatively correlated with US real yields. USDJPY is positively correlated with US Treasury yields. The tradeable insight is not the correlation itself — it is that two correlated trades double your risk. Open EURUSD long and GBPUSD long, and you have one trade at double size.
Why correlation matters more than most traders think
Retail traders think in terms of individual trades. Professionals think in terms of portfolio heat — the total risk if every position moves against you at once. Correlation is what determines whether your three "separate" trades are actually three trades or one.
Open long EURUSD, long GBPUSD, and long AUDUSD. You think you have three positions. You have one bet against the dollar at triple size. If the dollar rallies, all three lose together. This is the single most common reason retail traders blow up: they think they are diversified when they are not.
T
Written by the Trade To The Top team|Reviewed 29 September 2026
Correlation coefficients cross-checked against the BIS Triennial Survey (2022), published Federal Reserve research on USD cross-currency basis and Treasury yield relationships, the World Gold Council's published research on gold and real yields, and the rolling 90-day correlation data published by TradingView, Bloomberg and Refinitiv. The correlation regime framework follows the published research of the Intermarket Analysis school (Murphy, since 1991).
Markets do not move independently. They move in a web of relationships, some stable, some shifting week to week. The relationships are not signals. They are risk-management tools. Understanding them tells you when you have one position and when you have five. It tells you when your "hedge" is real and when it is not. It tells you which direction the dollar is really moving.
Key takeaways
Correlation ranges from −1 to +1. Zero means no relationship. Either extreme means a strong one.
EURUSD vs DXY is roughly −0.95. The strongest, most stable correlation in all of forex.
USDJPY vs US 10-year yield is positive. Higher US rates → stronger USDJPY.
Gold vs US real yields is negative. Higher real yields → weaker gold.
Risk-on / risk-off regimes drive correlated moves across stocks, AUD, NZD, EM currencies and gold.
Correlations are not stable. They shift over time. A pair that was 0.8 last quarter can be 0.3 this quarter.
Correlated trades double your risk. Long EURUSD + long GBPUSD = one trade at double size.
Correlation is a portfolio tool, not a trade signal. Use it to size positions, not to find entries.
In a crisis, correlations go to 1. Everything sells off together, even the hedges.
Inter-market analysis is the study of how bonds, commodities, stocks and currencies inform each other.
Correlation is a statistical measure of how two series move relative to each other. It is expressed as a coefficient between −1 and +1:
The correlation coefficient
Correlation coefficient (r) = covariance(A, B) / (stdev(A) × stdev(B))
Where:
A = first time series (e.g. EURUSD daily returns)
B = second time series (e.g. DXY daily returns)
covariance = how much A and B move together
stdev = the standard deviation of each series
Range: −1 (perfectly opposite) to +1 (perfectly identical) Zero: no linear relationship Rule of thumb: |r| > 0.7 = strong. 0.3–0.7 = moderate. < 0.3 = weak.
Three things to know before you use any correlation figure:
Correlation is calculated over a window. A 30-day correlation is different from a 90-day correlation. Both can be useful; mixing them silently is a mistake.
Correlation is not causation. EURUSD and DXY are negatively correlated because they share the euro and dollar components — not because one causes the other.
Correlation is not stable. It drifts. Regimes change. A correlation that held for two years can break in a single week of news.
EURUSD VS DXY · NEAR-PERFECT INVERSE CORRELATION
Two panels · same period · mirrored price action
EURUSD and DXY. The strongest inverse correlation in all of forex. They are the same trade expressed two ways.
Common mistake
Treating a correlation coefficient as fixed. Correlation is a rolling statistic. A 90-day correlation between EURUSD and DXY of −0.95 does not mean the correlation will be −0.95 tomorrow. It can drop to −0.6 in a week of unusual flows. Check the correlation, do not assume it.
The four core correlations
Four relationships dominate forex and intermarket analysis. Every trader should know them by heart.
Pair
Typical r
Why
EURUSD ↔ DXY
−0.90 to −0.98
Euro is 57.6% of the DXY basket. Same underlying, opposite direction.
USDJPY ↔ US 10Y yield
+0.60 to +0.85
Higher US rates make USD assets more attractive. Carry flows drive USDJPY.
Gold ↔ US real yields
−0.55 to −0.85
Gold pays no yield. Higher real yields raise the opportunity cost of holding it.
AUD ↔ risk-on regime
+0.50 to +0.80
AUD is a commodity and risk currency. Rises when equities rise.
Ranges are approximate. The exact coefficient varies by window and by regime. But the signs are stable. EURUSD will not become positively correlated with DXY for any meaningful period. USDJPY will not persistently decouple from US yields outside of specific Bank of Japan interventions.
USDJPY VS US 10-YEAR YIELD · POSITIVE CORRELATION
Both series plotted over the same period · moves are aligned
USDJPY and the 10-year yield. Higher US rates attract capital to the dollar. The yen, as the funding currency of the carry trade, weakens.
Risk-on vs risk-off regimes
Beyond the four core correlations, there is a regime that moves almost everything at once: risk-on or risk-off. When the market's risk appetite changes, whole asset classes move together.
Risk-on
StocksUP
AUD, NZDUP
USD, JPY, CHFDOWN
GoldMIXED
RISK APPETITE UPCarry trades work
Risk-off
StocksDOWN
AUD, NZDDOWN
USD, JPY, CHFUP
GoldUP
FLIGHT TO SAFETYSafe havens outperform
The risk-on / risk-off regime is the single most important macro context for correlated trades. When risk-on, the AUD, NZD, and equities move together. When risk-off, the USD, JPY, CHF, and gold move together. If you are long AUDUSD and long NZDUSD and the regime flips to risk-off, both lose together — they are effectively the same trade.
Why safe havens are safe havens
The dollar is a safe haven because the US Treasury market is the deepest and most liquid in the world. When risk-off hits, institutions need a place to park size — and Treasuries are the only market that can absorb it without moving.
The yen is a safe haven for a different reason: Japan is the world's largest net creditor. Japanese institutions hold trillions in foreign assets. When risk-off hits, they repatriate capital — selling foreign assets and buying yen.
Gold is a safe haven because it has no counterparty. In a crisis, that matters. In a real panic, gold sometimes sells off first because leveraged holders need to raise cash — but it recovers faster than anything else.
When correlation breaks
Correlations are not permanent. They break. And when they break, they break fast. Four causes:
The four correlation breakers
01
Central bank intervention.
When a central bank intervenes directly — like the Bank of Japan buying yen, or the Swiss National Bank capping the franc — standard correlations break for weeks. USDJPY can fall even as US yields rise, if the BoJ is intervening.
02
Regime shifts.
A move from a low-volatility regime to high-volatility regime can flip the sign of some correlations. Historically, gold and equities were negatively correlated. In 2022, they became positively correlated during the rate-hike cycle. Regime changes correlate behaviour.
03
Unique drivers.
EURUSD and DXY are tightly linked because they share components. But when a euro-specific event hits (ECB policy shift, Italian political crisis), the correlation temporarily weakens. Local news overrides global correlation.
04
Crisis.
In a real crisis — 2008, March 2020 — all correlations go to +1. Everything sells off together. The only asset that holds is cash (and often the dollar). Diversification fails exactly when you need it most. This is the great flaw of correlation-based risk management.
Correlated trades double your risk
This is the practical section. Correlation is not a signal — it is a risk-management tool. The lesson to internalise is simple: correlated positions are one position.
THE SAME BET, THREE WAYS · WHY DIVERSIFICATION FAILS
Three "diversified" trades that are actually one trade at triple size
Three "diversified" trades. All three are the same bet: short dollar. When the dollar rallies, all three stop out at once.
Correlated trades are not three trades. They are one trade at three times the size.
Worked example — the same trades, two correlation scenarios
Account size
$10,000
Risk per trade (nominal)
1% = $100
Trades
Long EURUSD, GBPUSD, AUDUSD
Correlation (scenario A)
Low (0.2)
Correlation (scenario B)
High (0.85)
Scenario A — Low correlation (0.2).
The three trades move independently. Worst-case outcome: all three stop out together, but the probability of that is low.
Expected worst-case loss: $150 – $200 (1.5% – 2% of account).
Scenario B — High correlation (0.85).
The three trades move as one. Dollar rally = all three stop out together. This is not a low-probability scenario — it is the normal behaviour.
Realistic worst-case loss: $300 (3% of account).SAME NOMINAL 1% RISK. THREE TIMES THE ACTUAL RISK.
The correlation matrix
The correlation matrix shows the pairwise coefficient between several markets at once. It is the single most useful tool for portfolio-level risk management. Read it as a heatmap.
CORRELATION MATRIX · MAJOR FX AND CROSS-ASSETS (90-DAY ROLLING)
Green = positive. Red = negative. Darker = stronger.
The correlation matrix. Read as a heatmap. Strong negative correlations are the ones that hurt when combined — they look like hedges but fail in a crisis.
Common mistake
Thinking you have a hedge because two positions are negatively correlated. A negative correlation is not a hedge. A hedge is a position that profits when your main position loses. A negative correlation means the two move oppositely on average — but the risk of both losing at once in a crisis is real. In 2008 and 2020, positions that were −0.8 correlated still lost together when forced liquidations hit.
How to use it in practice
The four practical uses of correlation
01
Cap total correlated risk.
Before opening a second position, check the correlation to your first. If r > 0.7 in the same direction, treat the two positions as one. Reduce size accordingly. Total correlated risk should never exceed 2%.
02
Use correlated pairs for confirmation.
If EURUSD breaks a level, check GBPUSD. If DXY confirms, the break is real. If DXY does not confirm, the EURUSD break is suspect. Correlation is a filter, not a signal.
03
Trade the strongest expression, not all of them.
If the dollar is breaking down, do not trade every USD pair. Pick the one with the cleanest setup and trade that. One trade, full size, clean execution.
04
Watch for correlation breakdowns.
When EURUSD and DXY diverge, that is information. When USDJPY and US yields decouple, that is information. Breakdowns often precede reversals.
Worked example — sizing a portfolio with correlated positions
Account
$10,000
Max correlated risk
2% = $200
Position 1
Long EURUSD, 0.5% risk
Correlation with P1
0.78
Position 2
Long GBPUSD, sized at 0.5%
Correlation with P1
0.85
Position 3
Long AUDUSD, sized at 0.5%
Total nominal risk
1.5%
Problem: The three positions are correlated 0.78–0.85. If the dollar rallies, all three stop out together.
Effective risk: 1.5% × 1.3 (correlation adjustment) ≈ 2%.
That is the maximum correlated risk for a single strategy. You cannot open another USD-short trade.
Solution: Reduce each position to 0.4% risk. Total nominal risk: 1.2%.
Effective correlated risk: 1.2% × 1.3 ≈ 1.6%.
Now you have room for one more position if a genuinely uncorrelated setup appears.
SIZE FOR TOTAL RISK, NOT PER-TRADE RISK.
Correlation is a sizing tool, not an entry tool. Use it after the setup, not before.
When this fails
When this fails
In a crisis. All correlations go to +1 when forced liquidation hits. Diversification fails exactly when you need it most. Reduce total exposure before crisis periods, not during them.
Central bank intervention. When the BoJ or SNB intervenes, standard FX correlations break for weeks. USDJPY can decouple from US yields entirely. Check the calendar for intervention risk.
Regime shifts. The gold/equity correlation flipped sign in 2022. The EURUSD/DXY correlation temporarily weakened during the 2020 pandemic. Correlation is not stable.
On very short timeframes. On the M5 and M15, correlations are noisy. The relationships hold over days and weeks, not minutes. Use correlation for position sizing on daily timeframe, not for scalping.
If you remember nothing else: two correlated trades are one trade. Size accordingly.
In one box
Correlation ranges from −1 to +1. Zero means no linear relationship.
EURUSD ↔ DXY ≈ −0.95. The strongest, most stable correlation in forex.
USDJPY ↔ US 10Y yield ≈ +0.70. Higher US rates → stronger USDJPY.
Gold ↔ US real yields ≈ −0.70. Higher real yields → weaker gold.
Risk-on / risk-off drives correlated moves across stocks, AUD, NZD and safe havens.
Correlated trades double your risk. Long EURUSD + long GBPUSD = one trade at double size.
Correlation is a sizing tool, not a signal. Use after the setup.
Correlations change. Check the rolling coefficient before every correlated trade.
In a crisis, everything correlates to 1. Diversification fails when you need it most.
Total correlated risk should never exceed 2% of the account.
Log correlated positions as one trade. Our free trading journal lets you tag entries by correlation cluster — USD-short, risk-on, carry — so you can see your real portfolio risk, not the fake per-trade number. Most traders discover their "3% risk" is actually 5%.
5 questions · immediate feedback · retake any time
Question 01 of 05
What is the typical correlation between EURUSD and DXY?
Correct: B. EURUSD and DXY are heavily negatively correlated because the euro makes up 57.6% of the DXY basket. The coefficient typically ranges from −0.90 to −0.98.
Question 02 of 05
You go long EURUSD and long GBPUSD, each with 1% risk. What is your actual total risk?
Correct: C. EURUSD and GBPUSD are correlated roughly 0.78–0.85, both largely against the dollar. Going long both is effectively one bet against the dollar at double size. Your real risk is close to 2%, not 1%.
Question 03 of 05
What does USDJPY typically correlate with?
Correct: A. USDJPY is positively correlated with US Treasury yields, roughly +0.60 to +0.85. Higher US rates make USD assets more attractive and drive carry flows into the dollar and out of the yen.
Question 04 of 05
What happens to correlations during a real crisis (like March 2020)?
Correct: D. In a real crisis, forced liquidations cause everything to sell off together. Diversification fails precisely when you need it most. This is why reducing total exposure before crisis periods matters more than relying on correlation.
Question 05 of 05
What is the correct use of correlation in a trading plan?
Correct: B. Correlation is a risk-management tool, not an entry signal. Use it to size positions and to cap total portfolio risk. The setups come from structure and price; correlation tells you how much to risk on each one.
Borrow in a low-yield currency, invest in a high-yield one. The trade that quietly drives trillions in FX flow — and how it unwinds when it goes wrong.