Correlation Between Markets: Why Three Trades Can Be One Position

Markets are not independent. Buy three pairs that all move with the dollar and you have not diversified; you have taken one position at triple size and told yourself a comforting story about it.

Correlation Between Markets at a glance

What it is and why it works

Every market is connected to others through the flows and the economics that drive them. The US dollar sits on one side of most currency pairs, so anything that moves the dollar moves all of them at once. Commodity exporters’ currencies follow what they export. Equity indices in different countries are exposed to the same global growth expectations and the same interest rates. None of this is coincidence; it is the same underlying forces reaching several instruments simultaneously.

Correlation is the statistic that describes it. A reading near +1 means two markets have been moving together; near −1 means they have been moving in opposite directions; near 0 means no dependable relationship over the window measured. That last phrase does most of the work, because a correlation is always a measurement over a specific period, not a property of the instruments. Change the window and you frequently change the answer.

Here are the relationships that genuinely matter, described honestly as tendencies rather than laws. The US dollar index against most currency pairs: since the dollar is one leg of the majors, dollar strength tends to push EUR/USD, GBP/USD and AUD/USD down together and USD/JPY, USD/CAD and USD/CHF up together. This is close to arithmetic rather than correlation. Gold against real yields and the dollar: gold pays no income, so when inflation-adjusted yields rise the opportunity cost of holding it rises and it tends to struggle; it is also dollar-priced, so a stronger dollar is a headwind. That said, gold has long stretches where safe-haven demand or central bank buying overwhelms both. Oil against the Canadian dollar: Canada is a major exporter, so sustained moves in crude tend to be reflected in CAD, most visibly in USD/CAD. The Australian dollar against China and industrial metals: Australia exports iron ore and other bulk commodities predominantly to Asia, so Chinese growth expectations and metals prices tend to drive AUD, with the New Zealand dollar behaving similarly. Equity indices against each other: global indices share exposure to growth and rates and move together most of the time, though the technology weighting of the Nasdaq makes it far more rate-sensitive than a broader index.

Above all of these sits the risk-on / risk-off regime. When participants are confident, capital moves towards equities, commodity currencies and higher-yielding assets. When they are frightened, it moves towards the dollar, the yen, the Swiss franc, government bonds and gold. In a strong risk-off episode almost everything correlates, because one factor, fear, is driving all of it. Which brings us to the warning that matters more than any table of coefficients: correlations break down precisely when you are relying on them. A hedge built on a historical relationship fails in the crisis it was meant to protect against, and a portfolio that looks diversified in calm conditions turns out to be one bet when it counts.

How to trade it, step by step

1. List your open positions and mark what each one really is. Long EUR/USD, long GBP/USD and short USD/CHF is not three trades. It is one short-dollar position, three times over. Write the underlying exposure next to each ticket (dollar, risk appetite, oil, rates) and the duplication becomes obvious immediately.
2. Set a cap on total exposure to any single driver. Decide in advance that your combined risk on correlated positions cannot exceed, say, twice your normal single-trade risk. That one rule stops the most common way accounts blow up while every individual trade looked well sized.
3. Measure correlation on a rolling basis, not once. Check a rolling window (twenty days for tactical decisions, longer for structural context) and watch it change. A relationship that has been weakening for a month is telling you something a static table never will.
4. Use the dollar as your first filter on any forex position. Before taking a trade on a dollar pair, ask what the dollar itself is doing and why. Much of what looks like a EUR/USD setup is a dollar move wearing a euro costume, and knowing which you are trading changes both your target and your invalidation. If the base and quote structure is not yet second nature, see how currency pairs are quoted.
5. Trade the leg with the clearest story, not both. If AUD and NZD are both responding to the same Chinese data, take the one with the better structure and cost profile rather than doubling into the theme. The second position adds risk far more than it adds expected return.
6. Treat a broken correlation as information rather than a mistake. When two markets that usually move together diverge, something instrument-specific is happening. That is often the most valuable observation available: it isolates a genuine idiosyncratic driver from the general market noise.
7. Stress-test your book against a risk-off shock. Ask what happens to every open position if fear spikes tomorrow morning. If the honest answer is that they all lose together, you are not diversified and you should size the whole book as though it were a single trade.
8. Never build a hedge that depends on a correlation holding under stress. If you need protection, hedge with the instrument itself or reduce the position. Cross-instrument hedges are convenience trades, and the convenience disappears in exactly the conditions that made you want the hedge.

Size every one of those entries with the position size calculator and check the trade is worth taking with the risk/reward calculator before you commit.

The conditions it needs

As a risk-measurement tool

This is where correlation earns its keep. It answers the question “how much am I really exposed to this theme?” more honestly than a list of tickets ever can, and it does so before the market forces the answer on you.

In stable, identifiable macro regimes

Relationships driven by real economic linkages (oil and CAD, iron ore and AUD, rates and the Nasdaq) are more durable than statistical ones because there is a mechanism underneath. They still shift, but they shift for reasons you can follow.

When you can explain <em>why</em> two markets are linked

A correlation with a causal story behind it is far more useful than one found by scanning data. If you cannot explain the mechanism in a sentence, you have probably found a coincidence over your chosen window, and it will not survive contact with a new regime.

For spotting divergence

The most tradeable use of correlation is when it fails. If an index is rising while its usual peers stall, or a commodity currency ignores its commodity, something specific is happening. Divergence localises the story in a way that following the correlation never does.

When it fails

• “Trading several correlated pairs spreads my risk.” It concentrates it. Three positions on the same theme, each risking one per cent, is a three per cent bet on one idea, and one that will lose all three at once. This is the single most common way a disciplined-looking trader takes reckless risk.
• “These two always move together.” Nothing always does. Correlations are regime-dependent and can invert. Gold has traded with the dollar rather than against it during specific episodes; risk assets have fallen alongside bonds when the driver was inflation rather than growth. Any statement containing “always” is a position waiting to be surprised.
• “Correlation tells me which way to trade.” It is a description of past co-movement, not a directional signal, and it carries no information about which market leads. Building entries on a correlation coefficient generally produces trades with no edge and doubled exposure.
• Confusing correlation with causation. Two markets often move together because a third factor drives both. Removing that factor removes the relationship, which is why correlations discovered by data-mining rarely survive. Insist on a mechanism.
• Hedging one instrument with a correlated one. A hedge that relies on a relationship holding will fail when the relationship breaks, and relationships break under stress. You have not removed risk; you have swapped directional risk for correlation risk and made it harder to see.
• Ignoring that correlations converge in a crisis. In a genuine panic, almost everything moves the same way as positions are liquidated indiscriminately. A book that looked balanced on Friday can be one enormous bet on Monday morning, and no adjustment made after the fact will help.

Where you will see this most clearly

• USD/CAD: The cleanest commodity link in forex, sustained oil moves show up here directly.
• AUD/USD: The market’s liquid proxy for Chinese growth and industrial metals demand.
• Gold (XAU/USD): Tied to real yields and the dollar, with long periods where safe-haven flow overrides both.
• NAS100: Moves with other indices but with far more rate sensitivity because of its technology weighting.
• EUR/USD: So dominated by the dollar leg that it is often a dollar trade rather than a euro one.

For different levels of experience

If you are brand new

Different markets are connected. If you buy EUR/USD, GBP/USD and AUD/USD at the same time, it feels like three separate trades, but all three go up when the US dollar goes down. You have made one bet against the dollar, three times.

That matters because you probably chose your position size assuming each trade was independent. If you risk one per cent on each, you are actually risking three per cent on a single idea, and when the dollar goes the other way all three lose together.

The habit to build is simple. Before you open a new trade, look at what you already have on and ask: if I am wrong about the dollar, or about markets generally, do all of these lose at once? If the answer is yes, either take the best one and skip the rest, or make each position smaller.

A few relationships worth knowing: the dollar is on one side of most major pairs, so it moves them together; the Canadian dollar tends to follow oil; the Australian dollar tends to follow Chinese growth and metals; and share indices around the world usually move in the same direction.

If your results are inconsistent

The classic intermediate error is not a bad entry; it is a book that is secretly one position. Every trade is sized correctly on its own, the journal looks disciplined, and then a single dollar move takes out four trades in an afternoon. Reviewing your worst days will usually show exactly this shape.

Fix it with a total-exposure limit rather than a per-trade one. Group your open trades by driver (dollar direction, risk appetite, oil, rates) and cap the combined risk in each group. Most traders find that this single rule changes their equity curve more than any change to their entries.

The second adjustment is to check what is actually driving the pair before you take a technical setup on it. A beautiful structure on EUR/USD is worth much less if the move is entirely a dollar story that is about to be resolved by a US release. Knowing which leg is in charge tells you what to watch and what would invalidate the idea.

Finally, treat divergence as the real signal. When correlated markets stop agreeing, something specific has changed, and that is usually more tradeable than the correlation itself.

If you are experienced

The operative issues are non-stationarity and tail dependence. Rolling correlation estimates are noisy and regime-sensitive, and linear correlation systematically understates joint tail behaviour, which is precisely where portfolio risk is realised. A book optimised on average-period correlations will be materially more concentrated than it appears during a liquidation episode.

Practically that argues for treating correlation as a constraint rather than an input to optimisation: cap gross exposure by factor, use conservative assumptions for stress correlation, and combine it with the volatility scaling described in volatility explained, since volatility and correlation both rise together in a risk-off event. The compounding of those two effects is what turns a well-sized book into a drawdown.

Factor decomposition is more useful than pairwise coefficients for anyone running several positions. Most retail forex and index exposure reduces to a small number of drivers (dollar direction, global risk appetite, rate expectations, commodity terms of trade) and expressing the book in those terms exposes duplication that a correlation matrix can obscure. Lead-lag relationships, where one market genuinely prices information first, are the exception worth studying for signal rather than risk purposes.

Risk management for this strategy

Correlation risk is the risk that your positions are less independent than you assumed, and it is dangerous because it is invisible on any individual trade ticket. The defence is to size the theme, not the trade.

Set a per-driver cap alongside your per-trade risk: for example, no more than two per cent of the account at risk across all positions exposed to the same factor, regardless of how many tickets that represents. Assume correlations rise under stress rather than using their calm-period values, and assume volatility rises at the same time, because it does. Use the position size calculator for each position and then check the aggregate; the aggregate is the number that decides whether a bad day is survivable. And if you would not take the combined position as a single trade at that size, do not assemble it out of pieces.

Where Market Structure Pro fits

Correlation is an account-level problem, and most tools on a chart are instrument-level. Nothing in a standard indicator suite tells you that the clean setup in front of you is the fourth version of a trade you already have on.

Market Structure Pro helps at the point where correlation does the most damage: taking marginal versions of the same trade. Because it delivers one verdict per instrument (TRADE, TRANSITION or NO TRADE) with a confidence percentage and an A/B/C grade, you can compare candidates directly and take the strongest expression of a theme rather than all of them. A B-grade correlated duplicate is easy to decline when an A-grade version of the same idea is on the screen next to it.

Its session and spread awareness matter here too, since correlated setups often differ mainly in execution quality rather than in structure. And its ranging filter keeps you out of the instruments where a theme has not actually produced directional flow. MSP is decision support and does not manage your portfolio: the exposure cap is a rule you have to set and keep yourself.

Frequently asked questions

What does correlation mean in trading?

It measures how consistently two markets have moved together over a chosen period, on a scale from plus one to minus one. Plus one means they moved in lockstep, minus one means they moved in exact opposition, and zero means no reliable relationship over that window.

Which currency pairs move together?

Pairs sharing the same base or quote currency tend to move together because one leg is common to both. EUR/USD, GBP/USD and AUD/USD generally rise when the dollar weakens, while USD/JPY, USD/CAD and USD/CHF generally rise when it strengthens. The dollar is the common driver in each case.

Why is trading three correlated pairs risky?

Because it is one position at triple size, not three diversified positions. If each risks one per cent and all three are exposed to the same driver, you are risking three per cent on a single idea and all three will lose together when that idea is wrong.

Does gold move opposite to the dollar?

It tends to, because gold is priced in dollars and pays no income, so a stronger dollar and higher real yields both raise the cost of holding it. The relationship is a tendency rather than a rule, and safe-haven demand or central bank buying can override it for extended periods.

How is oil connected to the Canadian dollar?

Canada is a major oil exporter, so sustained moves in crude affect its terms of trade and its currency. Rising oil prices tend to support the Canadian dollar, which typically shows up as USD/CAD falling. Short-term moves often diverge; the relationship is clearest over longer horizons.

Why does the Australian dollar follow China?

Australia exports large volumes of iron ore and other bulk commodities, with China as its dominant customer. Expectations about Chinese growth therefore feed into Australian export earnings and into the currency, which is why AUD is often used as a liquid proxy for Chinese demand.

Do correlations ever break down?

Constantly, and usually at the worst moment. Relationships shift with the macro regime and can invert entirely. In a genuine crisis most correlations converge as positions are liquidated indiscriminately, which is exactly when a diversified-looking portfolio turns out to be a single bet.

Can I use correlation to hedge?

Only with caution. A hedge that depends on two instruments continuing to move together fails when the relationship breaks, and relationships break under stress. Reducing the position or hedging with the same instrument is more reliable than substituting correlation risk for directional risk.

What is risk-on and risk-off?

A description of the prevailing appetite for risk. In risk-on conditions capital flows towards equities, commodity currencies and higher-yielding assets. In risk-off conditions it moves towards the dollar, the yen, the Swiss franc, government bonds and gold. Many correlations are really just this single factor showing up in different markets.

Related reading

• Volatility Explained: Volatility and correlation rise together in stress, compounding your real exposure.
• What Causes Price to Move: Correlated markets move together because the same flows reach all of them.
• Market Cycles Explained: Correlation regimes shift with the broader cycle rather than at random.
• Position Sizing: Where the per-trade limit becomes a per-theme limit.
• Instrument Guides: What actually drives each market, one instrument at a time.