Correlation measures how two instruments move together. It runs from +1 (identical moves) through 0 (unrelated) to −1 (exactly opposite).
It matters for one reason: correlated positions are not diversification. They are the same trade, sized larger than you think.
Why FX is so correlated
Every currency pair contains two currencies, so pairs sharing one are mechanically linked.
Pair
Typical correlation
Why
EURUSD / GBPUSD
+0.85
Both are "not the dollar"
EURUSD / USDCHF
−0.90
USD is on opposite sides
AUDUSD / NZDUSD
+0.90
Same commodity-linked bloc
USDJPY / US 10-year yield
+0.70
Yen carry is a rates trade
XAUUSD / USD index
−0.60
Gold is priced in dollars
Long EURUSD and long GBPUSD is not two trades. It is one short-dollar bet in two wrappers. If the dollar strengthens, both lose together, at the same time, for the same reason.
The arithmetic nobody runs
A trader risks 2% per position and feels disciplined. They hold five dollar-negative positions:
Short USDJPY
Long EURUSD
Long GBPUSD
Long AUDUSD
Long gold
Five positions, "2% each." Real exposure: roughly 10% on a single view — that the dollar falls.
One hawkish Fed sentence, and all five move against them simultaneously. This is the mechanism behind most "I don't understand, I was only risking 2%" accounts.
Worse, it compounds through margin. Five positions require five lots of margin, so free margin falls five times as fast — which is how a bad day becomes a forced liquidation rather than a bad day.
Correlations are not constant
This is the part that catches experienced traders.
Correlations strengthen in crisis. In normal conditions EURUSD and GBPUSD run around +0.85. In a genuine risk-off event they approach +0.98 — everything becomes one trade, precisely when you most need diversification to work.
They also break without warning. AUDUSD and NZDUSD usually move together; an RBNZ surprise decouples them for a week.
So treat published correlation figures as a rough guide, not a constant. Two rules follow:
Assume correlations are higher than the table says, especially under stress.
Recheck before assuming two positions are independent. Yesterday's relationship is not a guarantee.
Using it properly
Count exposure by currency, not by position.
Before entering, ask: *how much am I already short the dollar?* If the answer is 6%, a sixth dollar-negative trade is not a new 2% risk — it is topping up an existing 6% one.
A simple discipline that works:
Rule
Effect
Max 2% per position
Standard
Max 5% per currency
Prevents five wrappers on one view
Max 8% total open risk
Caps a bad day
Two dollar-negative positions at 2% each is 4% — inside the limit. A third is not.
Genuine diversification means positions that can win independently: EURGBP (a euro-versus-sterling view, dollar-neutral) alongside USDJPY is closer to two trades than EURUSD alongside GBPUSD.
Worked example
$10,000 account. Trader is long EURUSD, long GBPUSD and long AUDUSD, each risking 2% ($200). They believe they risk $600 across three trades.
The dollar rallies 90 pips on a hawkish Fed. All three are dollar-negative and all three move against them at once.
EURUSD hits its stop: −$200
GBPUSD hits its stop: −$200
AUDUSD, more volatile, gaps through its stop: −$260
Total: −$660, or 6.6% in a single session, from one headline. Their risk model said 2%.
Same account with a currency cap of 5%: the third trade is refused. Loss is $400, or 4%. A bad session rather than a hole that invites revenge trading tomorrow.
The difference is not skill. It is one line in a written plan.
What to take away
Correlated positions are one trade, split across several tickets.
Pairs sharing a currency are mechanically linked — most retail portfolios are a dollar bet.
Correlations rise toward 1 in crises, exactly when diversification is needed.
Cap exposure per currency, not just per position.
Real diversification means positions that can win independently.
That completes Fundamental Analysis. Next: why the chart itself tells you more than anything derived from it.
