Third-Party Research & Methodology Only

This section shares summaries of third-party academic research and descriptions of quantitative models. The content represents the findings of the original researchers, not the opinions or recommendations of Foxholm Financial. Foxholm Financial does not publish hypothetical or backtested performance metrics on its quantitative research pages. All content is restricted to methodology, signal construction, factor logic, and risk architecture. SEC rules require that investment advisers not present misleading performance data, and our methodology-only approach reflects that standard and the firm's fiduciary obligations.

Correlation

Statistical Measure Portfolio Construction Risk Analysis

Correlation measures the strength and direction of the linear relationship between two variables. It ranges from −1 to +1 and is one of the most important concepts in portfolio construction, because it determines how much diversification benefit investors actually receive when combining different assets.

A correlation of +1 means two assets move in perfect lockstep. A correlation of −1 means they move in exactly opposite directions. A correlation of 0 means there is no linear relationship between them at all. In practice, most asset pairs fall somewhere in between, and the correlation between them shifts over time.

Definition

The standard measure of correlation is the Pearson correlation coefficient, typically written as r or the Greek letter rho. It quantifies how closely the returns of two assets track each other in a straight-line pattern.

Pearson Correlation Coefficient

r = Cov(X, Y) ÷ (σ_X × σ_Y)

Where Cov(X, Y) is the covariance (a measure of how two variables move together), σ_X is the standard deviation (a measure of spread) of variable X, and σ_Y is the standard deviation of variable Y.

+1: Perfect positive correlation. The two variables move together in the same direction, proportionally.
0: No linear relationship. Knowing the value of one variable tells you nothing about the other.
−1: Perfect inverse correlation. When one variable rises, the other falls by a proportional amount.

The formula divides covariance by the product of both standard deviations, which normalizes the result to fall between −1 and +1 regardless of the units or scale of the original data.

How to Interpret Correlation Values

Correlation Range	Interpretation
−1.0 to −0.7	Strong negative: assets tend to move in opposite directions
−0.7 to −0.3	Moderate negative: some tendency to move in opposite directions
−0.3 to +0.3	Weak or no linear correlation: movements are largely independent
+0.3 to +0.7	Moderate positive: some tendency to move in the same direction
+0.7 to +1.0	Strong positive: assets tend to move together closely

These ranges are general guidelines, not strict rules. In portfolio construction, even small differences in correlation can have meaningful effects on overall risk. A portfolio of assets with pairwise correlations of 0.3 behaves very differently from one where all pairwise correlations are 0.8.

Why Correlation Matters for Investing

Correlation is the mechanism through which diversification (spreading investments across multiple assets) historically has reduced risk. When assets have low or negative correlation, their gains and losses partially offset each other, reducing the overall volatility of the combined portfolio.

This is not just a theoretical concept. A portfolio holding two assets with a correlation of +0.2 will experience smaller drawdowns and smoother returns than a portfolio holding two assets with a correlation of +0.9, even if the individual assets have the same expected return and volatility.

Correlation plays a central role in several key investment decisions:

Diversification: The lower the correlation between portfolio holdings, the greater the reduction in overall portfolio risk. This is the mathematical basis for the familiar advice to "not put all your eggs in one basket."
Portfolio risk reduction: Total portfolio volatility depends not only on the volatility of each asset but also on the correlations between every pair of assets. Adding a low-correlation asset to a portfolio can reduce total risk even if that asset is individually volatile.
Asset allocation: Strategic decisions about how much to allocate to stocks, bonds, international equities, and other asset classes depend heavily on the expected correlations between those classes.

Correlation Is Not Causation

A high correlation between two variables means they tend to move together, but it does not mean one causes the other to move. Two assets may be correlated because they are both driven by a common underlying factor, such as interest rates or economic growth, rather than because one directly influences the other.

This distinction matters for portfolio construction. If two seemingly uncorrelated assets are both secretly driven by the same risk factor, the diversification benefit may disappear precisely when it is needed most. Understanding the economic drivers behind observed correlations is just as important as measuring the correlations themselves.

Known Limitations

Limitations to Keep in Mind

Correlation is not stable over time. The correlation between any two assets can and does change across different market environments. During financial crises, correlations across asset classes tend to spike toward +1, a phenomenon sometimes summarized as "correlations go to 1 in a crash." This means the diversification benefits measured during calm markets may partially vanish during the periods when they are most needed.
Only measures linear relationships. Pearson correlation captures straight-line relationships. Two variables can have a strong nonlinear relationship (for example, a U-shaped pattern) and still show a correlation near zero. A low correlation does not necessarily mean the variables are independent.
Sensitive to outliers. A single extreme observation can significantly shift the calculated correlation. This is especially relevant in financial data, where large moves (crashes, squeezes, flash events) are more common than a normal distribution would predict.
Backward-looking. Correlation is calculated from historical data and reflects past relationships. There is no guarantee that historical correlations will persist into the future, particularly if the economic or policy environment changes.
Time period dependence. Different measurement windows produce different results. The correlation between stocks and bonds over a 5-year window can differ substantially from the correlation measured over 20 years. Always consider what time period was used when interpreting correlation estimates.

Practical Example

Stocks and bonds have historically exhibited low or moderate correlation over long periods. This is the foundational reason why the classic balanced portfolio (a mix of stocks and bonds) has been a staple of investment management for decades. When stocks decline, bonds have often held steady or risen in value, cushioning overall portfolio losses.

However, this relationship is not constant. During certain periods, such as the inflationary environment of the 1970s and parts of the early 2020s, stocks and bonds moved in the same direction (positive correlation), reducing the diversification benefit of the traditional mix. Investors who assumed the historical low correlation would always hold were caught off guard when both asset classes declined simultaneously.

This example illustrates two important lessons: correlation is a useful guide for portfolio construction, but it should never be treated as a fixed constant. Regular reassessment of correlation assumptions is a necessary part of sound portfolio management.