Correlation and Portfolio Risk in Multi-Asset Trading

A trader with five open positions is not necessarily running five independent risks. Depending on how those positions are constructed, they may be running something closer to one very large risk expressed through five different instruments, or they may be running a genuinely diversified set of exposures that partially offset each other. The difference is not visible from a position list. It requires understanding of correlation.

Correlation is the degree to which two instruments tend to move together. A strong positive correlation between two positions means they tend to rise and fall together. A high negative correlation means they tend to move in opposite directions. A correlation close to zero means the movements are largely independent. The portfolio risk created by holding multiple positions simultaneously is determined not by the sum of their individual risks but by the correlations between them.

The articles on position sizing and overnight financing in this series address risk at the individual position level. This article addresses risk at the portfolio level: how the aggregate exposure created by multiple simultaneous positions can differ substantially from the sum of the parts, and how correlation is the mechanism that drives that difference.

What Is Correlationin Trading?

Correlation measures the statistical relationship between the price movements of two instruments over a defined period. It is expressed as a number between -1 and 1.

A correlation of +1 indicates that the two instruments move in perfect lockstep: when one rises by 1 per cent, the other rises by 1 per cent. A correlation of minus one indicates that they move in perfect opposition: when one rises by one per cent, the other falls by one per cent. A correlation of zero indicates no relationship: the movements of one provide no information about the movements of the other.

In practice, correlations between financial instruments are never exactly plus or minus one, and a correlation of exactly zero between any two actively traded instruments would be unusual. Most pairs of instruments that share common economic drivers will exhibit some degree of positive correlation. Instruments with opposing economic sensitivities will exhibit a negative correlation. The correlations that matter most for portfolio risk are those that are sufficiently strong to create meaningful co-movement between positions under the conditions most likely to be encountered.

The correlation coefficient is a backwards-looking measure, calculated from historical price data over a specified lookback period. A 30-day correlation between Apple Inc and Nvidia calculated today tells you how closely those two stocks moved together over the last 30 days. It is an informative starting point, but it is not a guarantee of how they will move over the next 30 days. The most important characteristic of correlation in trading, and the one with the most practical consequences, is that it changes over time and changes most dramatically precisely when its consequences are most significant.

Why Apparent Diversification Can Be Illusory

The intuitive appeal of holding multiple positions across different instruments is that gains on others will offset the losses on some. This is the logic of diversification, and it is sound when the positions involved are genuinely uncorrelated.

The problem is that many positions that appear to offer diversification share common underlying drivers that cause them to move together under stress conditions, precisely when diversification is most needed.

Consider a trader holding three positions simultaneously: a long EUR/USD, a long European equity index CFD or Future, and a long position in crude oil. Assessed individually, these appear to be three distinct exposures across a currency pair, an equity index, and a commodity. In a conventional diversification framework, the losses on one might be expected to cushion losses on the others.

Under a broad risk-off episode, however, all three positions tend to move in the same direction simultaneously. Risk-off sentiment drives capital away from European currencies and into the dollar, pushing EUR/USD lower. The same risk aversion drives equity selling, pushing the index lower. Demand concerns reduce oil prices, pushing the crude position lower. Three apparently distinct positions have converged into a single directional bet on global risk appetite. The diversification was real under normal conditions and largely absent under the conditions that produced the largest moves.

This is the characteristic of correlation that most directly affects practical portfolio construction: correlations between risky assets tend to rise during stress. Positions that were relatively independent in normal market conditions become more tightly co-moving when markets are under pressure. Diversification is most available when it is least needed and least available when it is most needed.

Common Correlation Pairs and What Drives Them

Understanding the drivers of correlation between specific instruments enables us to anticipate when correlations are likely to strengthen or weaken, rather than simply observing them after the fact.

EUR/USD and GBP/USD. Both pairs use the US dollar as the quote currency, meaning they share significant price movements driven by dollar dynamics. When the dollar strengthens broadly, both pairs tend to fall. When it weakens, both tend to rise. The correlation is typically high and positive, often in the range of 0.7 to 0.9 over short to medium periods. A trader long both EUR/USD and GBP/USD is not holding two independent forex positions. They are holding two expressions of a broadly similar directional view on the dollar, with the EUR/GBP cross as the additional variable that differentiates them.

Equity indices across major markets. The S&P 500, FTSE 100, DAX, and other major equity indices exhibit significant correlation, particularly during periods of shifts in global risk sentiment. When US equities sell off sharply on a macro development, European and Asian indices typically follow suit. The correlation between the S&P 500 and the FTSE 100 over medium-term periods is has been seen in the range of 0.6 to 0.8. A trader long both US and European index CFDs simultaneously is running a position with considerably less diversification than the separate account entries suggest.

Gold and the USD. Gold is priced in dollars, which creates a mechanical inverse relationship: a stronger dollar makes gold more expensive for buyers holding other currencies, reducing demand and typically lowering the dollar price of gold. The correlation between gold and the dollar index is generally negative, ranging from -0.4 to -0.7. A long gold position alongside a long dollar position provides some natural offset between the two. A long gold position combined with a short dollar position compounds the same directional exposure rather than diversifying it.

Oil and risk sentiment. Crude oil tends to be positively correlated with risk assets because oil demand expectations are tied to global economic activity. When global growth concerns rise, oil demand expectations fall, and oil prices tend to fall alongside equities. When growth expectations are strong, both tend to rise together. The correlation is not constant and is heavily influenced by supply-side factors that operate independently of risk sentiment. Still, the growth expectations channel is sufficiently consistent to be a meaningful portfolio consideration.

Safe-haven assets under stress. The Japanese yen, Swiss franc, US Treasuries, and gold tend to be positively correlated with each other during broad risk-off episodes because they attract capital flows as investors reduce risk exposure across the board. Multiple long positions in safe-haven assets during a stress event tend to perform similarly rather than diversifying.

How to Measure Correlation Exposure Across a Portfolio

Traders who want to understand the correlation structure of their open positions do not need sophisticated software. A practical working framework requires three steps.

Map the common drivers. For each open position, identify the one or two primary factors that drive its price: monetary policy expectations, risk sentiment, commodity supply dynamics, and currency differentials. Positions that share the same primary driver are likely to be correlated. Positions with different primary drivers are more likely to operate independently.

Check historical correlations for key pairs. For positions where the driver mapping suggests potential correlation, checking the recent correlation coefficient between the instruments provides a quantitative basis for the assessment. Most charting platforms provide this calculation. A correlation above 0.6 between two positions held simultaneously should prompt consideration of whether the combined exposure represents the intended risk level.

Aggregate the directional exposure. If a trader holds a long EUR/USD, a long GBP/USD, and a short dollar position across a third instrument, all three positions effectively express a broadly similar view of dollar weakness. Expressing the combined exposure as a single directional bet on the common driver makes the actual risk more visible than three separate position entries do.

The Correlation Breakdown During Stress

The most practically consequential aspect of correlation in portfolio construction is its behaviour during market stress, which warrants specific treatment rather than being absorbed into the general discussion.

During normal market conditions, correlations between risky assets are moderate and variable. EUR/USD and the S&P 500 might show a correlation of 0.4 over a quiet month. A long position in both carries meaningful but not extreme co-movement risk. Volatility is contained, liquidity is deep, and the drivers of each instrument operate with a reasonable degree of independence.

During a risk-off episode, the correlation between these same instruments typically rises sharply, often above 0.7 or 0.8 within days. The reason is the same mechanism discussed in earlier articles on liquidity under stress: when broad risk sentiment shifts, investors reduce exposure across multiple asset classes simultaneously, and the resulting selling creates co-movement that was absent when each instrument was driven by its own idiosyncratic factors.

The practical consequence is that a portfolio constructed with moderate correlation risk under normal conditions can develop very high correlation risk within a short period when conditions deteriorate. A position-sizing framework calibrated to normal-condition correlations may significantly understate the portfolio’s true risk during periods when losses are largest.

This is not an argument against holding multiple positions. It is an argument for sizing each position with some awareness of how the portfolio’s correlation structure might change under stress, and for maintaining a sufficient margin buffer to withstand a scenario in which all correlated positions move adversely together rather than independently.

Correlations That Work in the Trader’s Favour

Correlation is not only a risk amplifier. When positions carry meaningful negative correlation, the portfolio benefits from a natural hedge: losses on one position are partially or fully offset by gains on the other.

The classic example is a long equity position combined with a long gold position. Under normal conditions, this pairing offers moderate diversification benefits because gold and equities share only a modest correlation. During a sharp equity sell-off driven by risk aversion, gold typically rises as safe-haven flows increase, providing a partial offset to the equity losses. The portfolio’s drawdown during the stress event is lower than the equity position alone would produce.

Negative correlations are most valuable when they are stable across conditions rather than appearing only under stress. A position pair in which negative correlation is a structural feature of the relationship, such as the gold-dollar relationship driven by mechanical dollar pricing effects, provides more reliable diversification than one in which negative correlation is situational.

Understanding which pairs in a portfolio exhibit structural negative correlation, rather than incidental negative correlation in the recent lookback period, is the discipline that makes correlation awareness a practical tool rather than a retrospective observation.

Correlation and the Position Sizing Framework

The position-sizing framework in article 9 of this series addresses individual trade risk through the fixed percentage rule. When multiple positions are held simultaneously, that framework needs to be applied with the correlation structure in mind.

If two positions have a correlation of 0.8 and each is sized to risk 1 per cent of the account on the stop-loss outcome, the portfolio’s expected risk in a scenario where both stops are triggered simultaneously is not simply 2 per cent. Because the positions are highly likely to move together, the probability of both stops being triggered in the same adverse event is substantially higher than it would be for two uncorrelated positions. The effective portfolio risk is closer to 2 per cent concentrated in a single risk factor than 2 per cent spread across two independent ones.

A trader who wants to maintain a genuine 1 per cent risk per independent risk factor, rather than per position, should scale down the size of each correlated position to reflect the shared exposure. If two positions share a correlation of 0.8, sizing each to risk 0.6 per cent of the account rather than 1 per cent produces a combined portfolio risk closer to 1 per cent of genuine independent risk exposure. The arithmetic is approximate, but the directional discipline is sound: recognise correlated positions and size them collectively rather than individually.

Applied across a multi-position account, this adjustment keeps the aggregate risk to any single underlying driver within the intended limits, regardless of how many positions are used to express that view.

The Bottom Line

Holding multiple positions simultaneously does not automatically create diversification. It creates a portfolio with a correlation structure that determines whether the risks in those positions offset, compound, or operate largely independently.

Most traders assess their positions individually. The risk they are actually running is the aggregate created when positions move together, which is often higher under stress conditions than the sum of the individual assessments suggests. Understanding which instruments share common drivers, how those correlations tend to change when market conditions deteriorate, and how to size correlated positions collectively rather than individually is the practical application of portfolio thinking to a leveraged trading account.

The next article in this series examines margin requirements in volatile conditions: how brokers calculate the minimum capital required to maintain open positions, why those requirements can increase during market stress, and what a margin call means for the position sizing and correlation frameworks built across this series so far.

Joshua Owen is CEO of Lunaro Financial Services. He has spent over a decade on trading desks at FCA-regulated firms, with a background in risk management, trading, and quantitative finance.

Disclaimer:

Lunaro is an execution-only service provider. This material is a marketing communication and is provided for general information and educational purposes only. It does not take into account your personal circumstances, objectives or needs. Any opinions are those of the author at the time of writing and may change without notice. Nothing in this material constitutes (or should be construed as) financial, investment, legal, regulatory or tax advice, or a recommendation to engage in any investment activity. You should not rely on this material when making investment or trading decisions.

This material has not been prepared in accordance with legal requirements designed to promote the independence of investment research and is not subject to any prohibition on dealing ahead of the dissemination of investment research. Lunaro may deal as principal and/or have an interest in the financial instruments or markets referred to in this material in the ordinary course of its business, including at or around the time of publication, and does not seek to take advantage of this material prior to its dissemination.

While reasonable care has been taken in preparing this material, no representation or warranty is made as to its accuracy or completeness and Lunaro accepts no liability for any loss arising from any use of, or reliance on, this material. Past performance is not indicative of future results.

CFDs are complex instruments and come with a high risk of losing money rapidly due to leverage. Approximately 80% of retail client accounts lose money when trading in CFDs. You should consider whether you understand how CFDs work and whether you can afford to take the high risk of losing your money.

Lunaro Financial Services Limited is authorised and regulated by the Financial Conduct Authority (FCA), Register No. 184333. Registered in England and Wales, No. 03148972.
Lunaro Markets Limited is authorised and regulated by the Financial Services Regulatory Authority (FSRA) of Abu Dhabi Global Market (ADGM), Financial Services Permission No. 200034, Registration No. 000005466.