Correlation and Concentration Risk: When Different Trades Are One Bet

বাংলায় পড়ুন

Correlation and Concentration Risk: When Different Trades Are Really One Bet

Learning objective: learn how to spot positions that appear diversified but are driven by the same market factor, then set portfolio risk with that overlap in mind.

Owning several instruments does not automatically create diversification. A portfolio can contain a UK equity index, a US technology fund, Bitcoin and an oil producer yet still depend heavily on one broad outcome: stronger risk appetite and easier financial conditions. If that shared driver reverses, several positions may lose together.

Correlation helps describe this relationship. It does not predict the next move, and it is not fixed. The practical purpose is to stop traders from accidentally multiplying the same risk across different tickers.

1. Understand what correlation measures

Correlation is commonly expressed between -1 and +1:

• +1 means two return series have moved in the same direction and proportion over the sample.
• 0 means there was no consistent linear relationship in that sample.
• -1 means they moved in opposite directions and proportion.

Traders should calculate correlation from returns rather than comparing price levels. The answer also depends on the timeframe and lookback. A 20-day daily-return correlation can differ greatly from a 250-day measure or an intraday estimate.

Correlation is descriptive, not permanent. It can rise sharply during a market shock as investors sell multiple risky assets at once.

2. Find the common risk driver

Before using a matrix, describe why each trade should work. Useful driver groups include:

• global growth and risk appetite;
• interest rates and bond yields;
• the US dollar;
• energy prices;
• inflation expectations;
• one industry, country or company supply chain.

Two positions can have a modest historical correlation and still share an event risk. A long airline and a short oil trade, for example, both express a view on fuel costs. Several gold-mining shares add company and equity-market risk to the gold-price exposure.

3. Measure exposure, then stress it

Create a simple table with each position, cash risk to its stop, direction and main driver. Then review return correlations over more than one window, such as 20 and 60 trading days. This is a diagnostic, not an instruction to optimise around a single number.

Next, test a stress scenario: what happens if equities fall, the dollar rises, volatility jumps or oil gaps? Assume correlations can move towards +1 for assets exposed to the same liquidation pressure. The stress test often reveals concentration that calm-period statistics hide.

4. Use a portfolio risk budget

Position risk should be considered both individually and as part of a cluster. A trader might limit the combined cash risk of several highly related ideas rather than allowing each one a full standalone allocation.

The exact limit must reflect the strategy, liquidity, holding period and tested loss tolerance. Percentage-risk conventions are examples, not universal rules. Stops can slip, gaps can exceed planned losses and correlations can change before an exit is filled.

Worked example

Suppose a hypothetical £25,000 account has three planned long trades:

Position	Cash risk to stop	Main driver
US technology ETF	£250	growth and falling yields
Bitcoin	£250	risk appetite and liquidity
Semiconductor share	£250	growth, AI demand and falling yields

Each trade risks 1% of the account, but the combined planned loss is £750, or 3%, if all stops are reached. The labels differ, yet all three may suffer when yields rise and risk appetite weakens.

One possible process is to treat them as a single risk cluster. The trader could reduce each position so the cluster risks £375 in total, select only the strongest setup, or add a genuinely different exposure after testing how it behaves. This does not guarantee a smaller loss; gaps and changing relationships still matter.

Now assume the three 60-day return correlations were only moderate, but all fell together during the last two sharp risk-off sessions. The stress behaviour deserves more weight than the calm-period average for a short-term downside plan.

Common mistakes

• Counting tickers instead of identifying shared drivers.
• Treating a past correlation coefficient as a permanent law.
• Mixing daily and intraday data without matching the strategy horizon.
• Ignoring direction: two positively correlated assets can offset when one is long and the other short.
• Believing a hedge is exact when its size, basis or trading hours differ.
• Adding several trades after the same news catalyst and assessing each in isolation.
• Forgetting that stops may fill beyond their intended prices.

Practical checklist

Before placing related trades, ask:

1. What economic or market driver makes each setup work?
2. Which positions could lose from the same surprise?
3. What is the total cash risk inside each driver cluster?
4. Do short- and medium-window return correlations tell the same story?
5. What happened to these relationships during recent stressed sessions?
6. Does position direction create overlap or offset?
7. Are gaps, liquidity and different market hours included in the risk plan?
8. Would the portfolio still be acceptable if correlations rose suddenly?

The Money Man view

For daily-timeframe trading, correlation is most useful as a portfolio warning system. Start with the trade thesis and cash risk, use statistics to challenge that view, then stress the whole cluster. The analysis is invalidated when it assumes relationships will remain stable or treats several correlated stops as independent protection.

This tutorial is educational information, prepared with AI-assisted research and writing. It is not personalised financial advice, a trade recommendation or a promise of returns. Markets can gap, liquidity can disappear and losses can exceed planned stop risk.