What it means
Correlation measures how closely two variables move together. A reading of +1 means they move in lockstep, 0 means no relationship and -1 means they move in exactly opposite directions.
Most real-world assets sit somewhere between these extremes, so a figure like -0.6 indicates a fairly strong opposite tendency. Diversification relies on this idea.
If a portfolio holds assets that do not all fall at the same time, a loss in one can be partly offset by a gain in another, so the portfolio swings less than its parts. Negatively correlated assets provide the strongest offset, which is why fund managers search for them.
Common examples include high-quality government bonds and shares during certain crises, when investors sell shares and buy safe bonds, or an airline's profits and the oil price, since higher fuel costs squeeze airline profit while oil producers gain. The relationship is not fixed, however, and can change with the economic environment.
A business can use the idea in many ways. A company with revenue sensitive to a commodity price might look for a hedge that gains when the price falls, and a fund manager may mix assets with low or negative correlation to smooth returns.
Finance teams also use it when testing whether a proposed hedge actually moves against the exposure it is meant to protect. The crucial caveat is that correlations are estimates from past data and they shift over time, often rising sharply in market stress.
Assets that appeared negatively correlated in calm periods can fall together in a crisis, so correlation should be tested over different periods. Correlation also says nothing about cause.
Two variables can move in opposite directions by coincidence, or because both respond to a third factor, so a statistical relationship should be backed by a sensible reason before it is relied on.
In practice
Real-world examples.
Example
An investment adviser builds a portfolio of global shares and high-quality bonds for a client. Because bonds have often risen when shares fell sharply, the combination has produced a smoother ride than shares alone. The client sleeps better, and is less likely to sell in a panic at the worst moment.
Example
A delivery firm with high fuel costs buys a small stake in an oil producer. Its profits fall when oil prices rise, but the investment gains, so the two partly cancel. The firm sizes the stake so that it offsets only part of its fuel exposure and does not turn into a large bet on oil.
Example
A seasonal business that sells ice cream in summer partners with a hot-drinks business that sells most in winter. Their combined revenue is steadier across the year than either alone. A lender looking at the combined business sees lower swings in cash flow and may offer better terms.
Formula
Calculation
Correlation = Covariance of A and B / (Standard deviation of A x Standard deviation of B)
Worked example: asset A has a standard deviation of 10% and asset B has 8%. Their covariance is -0.0048.
Product of standard deviations = 0.10 x 0.08 = 0.0080
Correlation = -0.0048 / 0.0080 = -0.60
A correlation of -0.60 means the two assets tend to move in opposite directions, though not perfectly.
For a portfolio with half in each, the variance is 0.5^2 x 0.10^2 + 0.5^2 x 0.08^2 + 2 x 0.5 x 0.5 x (-0.0048) = 0.0025 + 0.0016 - 0.0024 = 0.0017, so the standard deviation is about 4.1%, lower than both individual assets.Case study
Seen in the real world.
Sandstone Capital is an illustrative, fictional fund that invested only in technology companies. The returns were excellent in good years, but the portfolio dropped sharply when the sector sold off.
The risk manager measured correlations and found that many holdings moved together. She proposed adding a defensive allocation to government bonds and a small gold position that had historically shown low or negative correlation to the shares.
In the next downturn, the portfolio fell less than before. In this illustrative story, she also warned the board that correlations had risen briefly during an earlier panic, so she kept a cash reserve as an additional protection. The board approved a rule that correlations be recalculated every quarter and stress tested using crisis periods.
Watch out
Common mistakes.
- Assuming a negative correlation will hold in every market condition.
- Confusing negative correlation with no correlation, when zero means no relationship at all.
- Treating correlation as proof of cause and effect.
Questions
People also ask.
What is a good correlation for diversification?
Lower is better, and negative values give the most risk reduction, but a mix of assets with correlations well below +1 already helps.
Is -1 realistic?
A perfect negative correlation is extremely rare in real markets, though some hedging instruments come close. Even then, costs and timing differences prevent them from cancelling out exactly.
How often should correlations be reviewed?
At least annually and after major market events, since relationships change over time.
From the founder's library
Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%