Skip to main content
diversification.com
← Back to L Definitions

Legal_title

What Is Correlation?

In finance, correlation is a statistical measure that quantifies the degree to which two assets move in relation to each other. It belongs to the broader field of Portfolio Theory and is a fundamental concept for understanding investment behavior. Correlation ranges from -1.0 to +1.0, where +1.0 indicates a perfect positive correlation (assets move in the same direction), -1.0 indicates a perfect negative correlation (assets move in opposite directions), and 0 indicates no linear relationship. Investors use correlation to evaluate the potential for risk reduction within an investment portfolio and to inform asset allocation decisions. Understanding correlation is crucial for effective diversification.

History and Origin

The concept of correlation as a key element in portfolio construction gained prominence with the development of Modern Portfolio Theory (MPT) by economist Harry Markowitz in the 1950s. Markowitz's seminal work highlighted that investors could optimize their portfolios not just by selecting assets with high individual returns, but also by combining assets whose returns were not perfectly positively correlated. This insight revolutionized investment management by demonstrating how combining assets with varying correlations could reduce overall portfolio volatility for a given level of return, or increase return for a given level of risk. Markowitz was awarded the Nobel Memorial Prize in Economic Sciences in 1990, partly for his contributions to financial economics, which underscored the critical role of correlation in diversification.

Key Takeaways

  • Correlation measures the statistical relationship between two asset returns, ranging from -1.0 to +1.0.
  • A correlation of +1.0 signifies assets move in the same direction; -1.0 means opposite directions; 0 means no linear relationship.
  • Negative or low positive correlation between assets is vital for effective portfolio diversification, helping to reduce overall portfolio risk.
  • Correlation coefficients are calculated based on historical data and can change over time, especially during periods of market stress.
  • Understanding correlation helps investors make informed decisions about combining different asset classes.

Formula and Calculation

The correlation coefficient, often denoted as ρ (rho), between two assets, Asset A and Asset B, is calculated by dividing their covariance by the product of their individual standard deviation.

The formula for the correlation coefficient ((\rho_{A,B})) is:

ρA,B=Cov(A,B)σA⋅σB\rho_{A,B} = \frac{Cov(A,B)}{\sigma_A \cdot \sigma_B}

Where:

  • (Cov(A,B)) = Covariance between the returns of Asset A and Asset B
  • (\sigma_A) = Standard deviation of the returns of Asset A
  • (\sigma_B) = Standard deviation of the returns of Asset B

Interpreting the Correlation Coefficient

Interpreting the correlation coefficient provides insight into the likely behavior of assets within a portfolio:

  • Positive Correlation (0 to +1.0): When two assets have a positive correlation, they tend to move in the same direction. A correlation close to +1.0 implies that if one asset's price increases, the other's is highly likely to increase as well, and vice-versa. While this might seem desirable for assets expected to perform well, it offers limited diversification benefits against downside risk.
  • Negative Correlation (-1.0 to 0): Assets with a negative correlation tend to move in opposite directions. A correlation close to -1.0 means if one asset's price rises, the other's is highly likely to fall. This relationship is highly valued in portfolio construction because it provides significant risk reduction. If one part of a portfolio declines, another part may increase, offsetting losses.
  • Zero Correlation (0): A correlation of 0 suggests no linear relationship between the returns of the two assets. Their price movements are independent of each other. While not offering perfect hedging, assets with zero or near-zero correlation still contribute to diversification by not moving in tandem with existing portfolio holdings.

Investors evaluate correlation alongside expected return and risk tolerance to build a robust portfolio.

Hypothetical Example

Consider an investor, Sarah, who has a portfolio consisting entirely of technology stocks. She observes that during market downturns, all her tech stocks tend to decline simultaneously, due to their high positive correlation. To enhance her portfolio's diversification, she decides to add an asset that traditionally exhibits a low or negative correlation with technology stocks, such as gold.

Let's assume:

  • Technology Stock Fund A has an annual return of +10% and a standard deviation of 15%.
  • Gold ETF B has an annual return of +5% and a standard deviation of 10%.
  • The covariance between Technology Stock Fund A and Gold ETF B is -0.005.

Using the formula:

ρA,B=−0.005(0.15⋅0.10)=−0.0050.015≈−0.33\rho_{A,B} = \frac{-0.005}{(0.15 \cdot 0.10)} = \frac{-0.005}{0.015} \approx -0.33

The correlation coefficient between Technology Stock Fund A and Gold ETF B is approximately -0.33. This negative, albeit not perfectly inverse, correlation suggests that when tech stocks perform poorly, gold might perform relatively better or decline less, providing a buffer to Sarah's overall portfolio. By adding gold, Sarah aims to achieve better risk management through diversification.

Practical Applications

Correlation is a fundamental tool across various areas of financial markets and analysis:

  • Portfolio Diversification: The primary application is in building diversified portfolios. By combining assets with low or negative correlation, investors can reduce overall portfolio risk without necessarily sacrificing returns.
  • Risk Management: Financial institutions and fund managers use correlation to understand aggregated risk exposures. For example, in times of market stress, correlations between seemingly unrelated assets can increase dramatically, a phenomenon known as "correlation breakdown" or "correlation contagion." Regulators often push financial firms to stress-test their portfolios against scenarios where correlations shift unexpectedly.
  • Asset Allocation: Strategic asset allocation decisions often factor in historical and anticipated correlations between asset classes like stocks, bonds, commodities, and real estate.
  • Hedging Strategies: In options and financial derivatives, understanding correlation is crucial for designing effective hedging strategies, as the effectiveness of a hedge depends on the correlation between the underlying asset and the hedging instrument.
  • Arbitrage Opportunities: Professional traders may look for temporary discrepancies in correlation relationships to identify potential arbitrage opportunities.

Limitations and Criticisms

While correlation is a widely used and valuable metric, it has several important limitations:

  • Historical Data: Correlation coefficients are typically calculated using historical data, which may not be indicative of future relationships. Market conditions can change rapidly, leading to shifts in correlation that invalidate past assumptions.
  • "Correlation Breaks Down in Crises": A significant criticism is that correlations between assets tend to converge toward +1.0 during periods of extreme market stress or financial crises. This "flight to safety" or widespread panic can cause assets that were previously uncorrelated or negatively correlated to move in the same downward direction, undermining diversification benefits precisely when they are most needed.
  • Linear Relationship Only: The correlation coefficient measures only the linear relationship between two variables. It does not capture non-linear dependencies or tail dependencies (where assets move together only during extreme events).
  • Causation vs. Correlation: Correlation does not imply causation. Just because two assets move together does not mean one causes the other to move. Both might be influenced by a third, unobserved factor.
  • Sensitivity to Time Horizons: The calculated correlation can vary significantly depending on the time period and frequency of data used (e.g., daily, weekly, monthly returns).
  • Outliers: Extreme data points or outliers can disproportionately influence the correlation coefficient, potentially skewing the perception of the true relationship.

These limitations underscore the importance of using correlation as one tool among many in a comprehensive quantitative finance framework.

Correlation vs. Covariance

Correlation and covariance are closely related statistical measures that both describe the relationship between two variables, but they differ in their scale and interpretability. Covariance measures the directional relationship between the returns of two assets, indicating whether they tend to move together (positive covariance) or in opposite directions (negative covariance). However, the magnitude of covariance is not standardized and depends on the units of the data, making it difficult to compare across different pairs of assets.

Correlation, on the other hand, standardizes covariance by dividing it by the product of the assets' standard deviations. This normalization results in a coefficient that always falls between -1.0 and +1.0, making it a dimensionless measure. Because of this standardization, correlation is much easier to interpret and compare across different asset pairs, regardless of their individual volatilities or units of measurement. Therefore, while covariance tells you the direction of the relationship, correlation tells you both the direction and the strength of the linear relationship on a universal scale.

FAQs

Why is correlation important in investing?

Correlation is crucial in investing because it helps investors build diversified portfolios that can potentially lower overall risk. By combining assets that are not perfectly positively correlated, the negative movements of some assets may be offset by the positive movements of others, leading to a smoother portfolio return stream.

Can correlation change over time?

Yes, correlation is not static and can change significantly over time due to evolving market conditions, economic cycles, and investor sentiment. It is particularly known to increase (move towards +1.0) during periods of financial crisis or extreme market stress, reducing the effectiveness of diversification.

What is a "perfect" correlation?

A "perfect" correlation refers to a correlation coefficient of either +1.0 or -1.0. A +1.0 indicates a perfect positive correlation, meaning the assets always move in exactly the same direction and magnitude. A -1.0 indicates a perfect negative correlation, meaning the assets always move in exactly opposite directions with the same magnitude.

Is a low correlation always good for a portfolio?

Generally, a low or negative correlation is desirable for portfolio construction because it enhances diversification and reduces overall risk. However, assets with negative correlation might also have lower individual expected returns. The ideal portfolio balances the benefits of diversification against the potential for higher returns from more correlated assets, aligning with an investor's risk tolerance.

How is correlation related to a scatter plot?

On a scatter plot, correlation indicates how closely data points cluster around a straight line. If points form an upward-sloping line, there's a positive correlation. If they form a downward-sloping line, there's a negative correlation. If points are widely scattered with no clear pattern, it suggests low or zero correlation, which is often examined in regression analysis.

Sources:
The Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel 1990. NobelPrize.org. Nobel Prize Outreach AB 2024. Mon. 29 Jul 2024. https://www.nobelprize.org/prizes/economic-sciences/1990/summary/
Krugman, Paul. "Correlations in the Crisis." The New York Times, November 3, 2008. https://www.nytimes.com/2008/11/03/opinion/03krugman.html
Authers, John. "The perils of correlation." Financial Times, March 29, 2011. https://www.ft.com/content/158c8952-474c-11e0-b6f7-00144feab49a
Alloway, Tracy. "Fund managers to stress test for 'correlation breakdown'." Reuters, September 15, 2014. https://www.reuters.com/article/us-funds-stress-idUSKBN0H61QZ20140915/