False Diversification: All Correlated Assets Don't Reduce Risk

A trader owns 10 technology stocks. He tells himself he is diversified. He holds Apple, Microsoft, Nvidia, Tesla, Meta, Alphabet, Amazon, Broadcom, Marvell, and Advanced Micro Devices. All different companies, all different sectors within tech. Then the Fed raises rates and the entire technology sector crashes 20% in a week. His portfolio drops 20%. He is not diversified; he is just holding the same bet 10 times.

This is the false diversification trap. Diversification is not about the number of holdings; it is about low correlation among holdings. Owning 10 correlated assets provides almost no risk reduction compared to owning one. During crises, correlation between assets rises toward 1.0, meaning everything falls together. A portfolio that looks diversified in calm markets reveals itself as a concentrated bet during stress.

> Quick definition: Correlation measures how two assets move together. A correlation of 1.0 means perfect positive movement (both rise and fall together). A correlation of 0.0 means no relationship. A correlation of -1.0 means perfect negative movement (one rises when the other falls). True diversification requires low correlation.

Key takeaways

• Correlation ≠ diversification: Ten correlated stocks provide almost the same risk as one stock; you need low correlation to diversify.
• Crisis correlation: In market crashes, correlation between equities rises to 0.8–0.95. Every stock falls together, revealing that they are not diversified at all.
• Same sector = same risk: All tech stocks, all healthcare stocks, all financial stocks move together because they face the same economic drivers.
• False diversification by name: A portfolio that looks diversified on paper (Apple, Amazon, Microsoft) is concentrated in US tech, a single bet.
• True diversification requires assets that behave differently: Stocks, bonds, commodities, real estate—assets with low or negative correlation in normal and crisis periods.
• Correlation is time-dependent: Two assets might have 0.3 correlation over 10 years but 0.8 during a crisis; historical correlation is misleading.

Understanding correlation: the mathematics of false diversification

Correlation is measured from -1.0 to +1.0. Two tech stocks might have a correlation of 0.85, meaning they move together 85% of the time. An investment-grade bond and a stock might have a correlation of 0.15. Gold and stocks might have a correlation of -0.2.

Let's say you own a $100,000 portfolio split equally between two assets:

Portfolio A: $50,000 in Apple, $50,000 in Microsoft (correlation 0.88)

If Apple drops 10%, it loses $5,000. If Microsoft drops 10% (as it likely will due to the high correlation), it loses $5,000. Total portfolio loss: $10,000 or 10%. The two holdings move in lockstep.

Portfolio B: $50,000 in Apple stock, $50,000 in US Treasury bonds (correlation 0.15)

If Apple drops 10%, it loses $5,000. US Treasury bonds do not move much in a stock selloff; they might actually rise 2% as investors flee to safety, gaining $1,000. Total portfolio loss: $4,000 or 4%. The portfolio is more resilient.

Portfolio C: $50,000 in Apple, $50,000 in US Treasury bonds during a crisis

During a severe financial crisis, the correlation might rise to 0.5. Apple drops 25%, losing $12,500. Treasuries rally hard, up 10%, gaining $5,000. Net loss: $7,500 or 7.5%. Still less painful than Portfolio A.

This is the benefit of true diversification: low correlation reduces portfolio volatility and drawdowns, especially during crises.

Real example: the 2008 financial crisis reveals false diversification

In 2008, a retail trader held a portfolio he thought was diversified:

• 40% large-cap stocks (S&P 500 index)
• 30% tech stocks
• 20% financial stocks
• 10% real estate (REITs)

This looks diversified. But all four are equity-like asset classes. During the crisis, all four crashed:

• S&P 500: -57% (from peak)
• Tech: -60% (higher than S&P)
• Financials: -70% (nearly wiped out)
• REITs: -68% (highly levered, equity-like)

The portfolio fell approximately 60% on average. The false diversification provided no buffer.

A truly diversified portfolio in 2008 might have been:

• 40% US stocks: -57%
• 30% long-term bonds: +14% (bonds rallied as rates fell)
• 20% gold: +5%
• 10% cash: 0%

Portfolio return: approximately -28%. Still painful, but half the loss of the "diversified" portfolio. The difference was true diversification (stocks, bonds, commodities) rather than false diversification (four equity-like bets).

The illusion: why more holdings feel like diversification

Humans have an intuitive bias called the "diversification bias" or "splitting bias." If you offer someone $100 to split between two investments, they split it 50-50, thinking they are diversified. If you offer them $100 to split between five investments, they split it 20% each, thinking they are even more diversified. But the actual diversification depends on correlation, not the count.

A portfolio with 50 correlated stocks provides less actual diversification than a portfolio with 5 uncorrelated assets (stocks, bonds, commodities, real estate, cash).

This is why many retail traders feel protected by owning 10 tech stocks but get wiped out when the entire tech sector sells off. The number of holdings is not the variable that matters; the correlation is.

Same sector = same systemic risk

All financial stocks are sensitive to interest rates, credit spreads, and economic growth. When the Fed raises rates, they fall together. All consumer discretionary stocks are sensitive to consumer sentiment and recessions. When recessions loom, they fall together. All energy stocks are sensitive to oil prices. When oil crashes, they fall together.

A trader who owns five financial stocks owns five correlated bets on one thing: the financial sector. If the sector is hit by a crisis, all five fall. If the sector is fine, all five rise. There is no diversification; there is concentrated sector risk.

True diversification means owning different asset classes, not just different companies. You need stocks, bonds, commodities, and cash. You need domestic and international exposure. You need growth and value, cyclical and defensive. Only then do you have genuine diversification.

Correlation during crises: the big risk

In normal times, correlation between stocks is typically 0.4–0.6. This is reasonable. But in a crash, correlation spikes to 0.8–0.95 within days. Everything falls together. This is the moment when diversification should matter most—but false diversification fails because all the holdings are equity-like.

A trader with a portfolio of 10 uncorrelated stocks, bonds, commodities, and cash will suffer much less in a crash than a trader with 10 correlated stocks. The bonds will cushion the fall, the commodities might rally, and the cash allows for rebalancing at lower prices.

The 2008 crisis, the 2020 COVID crash (brief), and the 2022 market selloff all demonstrated that correlation is highest exactly when you need diversification most.

Correlation check

Real-world examples: correlation disasters

The 2008 financial crisis: Investors who thought they were diversified (stocks, REITs, commodities) discovered that all equity-like assets fell 50–70%. Bonds were the only safe harbor. False diversification by asset class (all equities) led to massive losses.

The 2020 COVID crash: On March 16, 2020, the S&P 500 fell 12%. On that day, VIX spiked to 82.69, and every equity index fell. A portfolio of US stocks, international stocks, and emerging market stocks all fell together. False diversification within equities was useless. Bonds provided the cushion.

The 2022 bear market: Rising interest rates hit both stocks and bonds in 2022, a rare event. A traditional 60/40 stocks-bonds portfolio fell 16% that year. Investors discovered that bonds and stocks can correlate when inflation is the common enemy. True diversification would have required commodities and real assets.

Crypto correlation in 2022: Bitcoin, Ethereum, and 100 other cryptocurrencies all crashed together from 2021 to 2022, correlation approaching 0.98. Owning 10 different cryptocurrencies provided zero diversification. Investors who thought crypto was diversification away from stocks learned an expensive lesson.

How to measure correlation: checking your portfolio

Most brokers and portfolio analysis tools show correlation matrices. You can also calculate correlation using Excel or Python using historical returns. For a portfolio to be truly diversified, you need:

• Average correlation between holdings below 0.5
• Ideally some holdings with negative correlation (Treasury bonds, gold, or puts can be negative correlation to stocks)
• Diversification across asset classes, not just holdings within one class

A portfolio of 50 tech stocks has an average correlation near 0.85. A portfolio of 10 stocks from different sectors might have correlation near 0.6. A portfolio of stocks, bonds, commodities, and real estate might have correlation near 0.3.

The lower the average correlation, the lower the portfolio volatility.

Building a truly diversified portfolio

Diversification across asset classes:

• Stocks: provide growth, some inflation protection
• Bonds: provide stability, rally in crises
• Commodities: provide inflation protection, rally in stagflation
• Real estate: provide income, some inflation protection
• Cash: provides optionality, allows rebalancing at low prices

Diversification within stocks:

• Domestic large-cap (S&P 500)
• Domestic mid and small-cap (Russell 2000)
• International developed (EAFE)
• Emerging markets
• Different sectors (tech, healthcare, financials, energy, industrials, utilities)

Diversification by trading style:

• Growth and value
• Cyclical and defensive
• Momentum and contrarian

A simple 80/20 rule works for most investors: 80% of diversification benefit comes from diversifying across asset classes (stocks, bonds, commodities). The remaining 20% comes from diversifying within stocks.

When false diversification kills your returns

A portfolio of 20 correlated stocks delivers lower returns than a portfolio of 5 uncorrelated assets because all the correlated stock holdings are "wasting" diversification slots on positions that move together.

If you have $100,000 and allocate it to 20 tech stocks, you might achieve 8% annual returns with 18% volatility. If you allocate it to 60% stocks, 30% bonds, 10% commodities across 10 holdings, you might achieve 6% returns with 10% volatility.

The first portfolio feels exciting (potential 8% returns) but is actually riskier and more likely to blow up during a crisis. The second portfolio feels boring (only 6% returns) but is more stable and more likely to preserve capital during downturns.

Most traders prefer the exciting, risky path until a crisis hits. Then they wish they had been boring.

Common mistakes with diversification

1. Counting holdings instead of measuring correlation: "I own 15 stocks, so I am diversified." Check the correlation. If they are all tech, you are not diversified.

2. Diversifying only within one asset class: Ten stocks are more diversified than one stock, but they are all equity risk. Diversification across asset classes is more important.

3. Using historical correlation to predict future correlation: "These assets had 0.2 correlation over the past 10 years." In a crisis, correlation might jump to 0.7. Always assume correlation can spike.

4. Ignoring sector overlap: You think you are diversified, but 40% of your holdings are tech, 30% are financial, so 70% is concentrated in two sectors. That is not diversification.

5. Diversifying during calm times only: A portfolio looks well-diversified in 2023 (low volatility). The same portfolio looks concentrated in 2024 (crisis). You need diversification that works in both environments.

6. Not rebalancing: You allocate 50/50 stocks and bonds. Stocks rally 20%; now you are 60/40. You are no longer diversified according to your original plan. Rebalance quarterly.

Frequently asked questions

What level of correlation is "safe" for a portfolio?

Below 0.5 for an overall portfolio is reasonable. Below 0.3 is excellent. A traditional 60/40 stocks-bonds portfolio has correlation near 0.2 in normal times, rising to 0.4–0.5 in crises. This is generally acceptable.

Should I diversify into asset classes that make lower returns?

Yes. A 6% return with 10% volatility is better than an 8% return with 18% volatility for most investors. The extra 2% annual return is not worth a 60% crash. Most professional investors choose the boring, stable path.

Can I diversify with just a few ETFs?

Yes. A simple portfolio of VTI (US stocks), BND (US bonds), and GLD (gold) gives you exposure to three asset classes with low correlation. This is better than owning 50 correlated stocks.

Is international diversification worth it?

International stocks have historically had 0.7–0.8 correlation to US stocks, lower in normal times but rising in crises. It provides some diversification but not as much as diversifying across asset classes. However, it is still worth 10–20% of a portfolio for true global diversification.

How often should I check my portfolio's correlation?

At least quarterly. Correlation is not static. A portfolio that is well-diversified one year can become concentrated if some holdings outperform others and take up more weight. Rebalance to maintain target correlation.

Is a 100-stock portfolio better diversified than a 10-stock portfolio?

Only if the correlation is lower. 100 correlated stocks are worse than 10 uncorrelated assets. Focus on correlation, not count.

What is the best level of diversification?

For most investors, a portfolio with 5–10 uncorrelated holdings across 3–4 asset classes (stocks, bonds, commodities, real estate) with average correlation below 0.4 is excellent. More diversification than this does not add much benefit and adds complexity.

Related concepts

• Understanding Correlation and Portfolio Risk
• What Risk Means in Investing
• Ignoring Tail Risk Until It Is Too Late
• What is a Stop Loss?

Summary

False diversification is owning many correlated assets and mistakenly believing you are protected against loss. Diversification works only when holdings have low correlation (below 0.5 ideally). During crises, correlation rises to 0.8–0.95, and correlated assets fall together. A portfolio of 10 tech stocks has nearly identical risk to a portfolio of 1 tech stock. True diversification requires different asset classes: stocks, bonds, commodities, real estate, and cash. A portfolio with average correlation below 0.3 is much less volatile and has lower drawdowns than a portfolio of correlated holdings. During the 2008 crisis, investors who thought they were diversified by owning stocks, REITs, and commodities fell 60% because all three are equity-like; those who held bonds fell only 28% because bonds have low or negative correlation to stocks. The key metric is correlation, not holdings count. Measure your portfolio's average correlation. If it exceeds 0.6, you are not diversified. Add asset classes or holdings with lower correlation. Rebalance quarterly to maintain target diversification, because volatility shifts weight and can make a diversified portfolio concentrated.

Next

→ Ignoring Tail Risk Until It Is Too Late