Home / Risk Metrics / Correlation
Portfolio metricIntermediate

Correlation

Correlation is a standardised measure of how strongly two return series move together, defined as their covariance divided by the product of their standard deviations, and it ranges from −1 (perfect opposite movement) through 0 (no linear relationship) to +1 (perfect same-direction movement).

Quick answer: Correlation is a standardised measure of how strongly two return series move together, defined as their covariance divided by the product of their standard deviations, and it ranges from −1 (perfect opposite movement) through 0 (no linear relationship) to +1 (perfect same-direction movement).

In simple words

Correlation measures whether two positions tend to move in the same direction, opposite directions, or unrelated. It runs from minus one to plus one: plus one means they rise and fall together, minus one means one rises exactly as the other falls, and zero means their moves have no linear connection. It is the number that decides whether diversification actually helps, because combining low-correlation positions smooths the portfolio while combining high-correlation ones does not. The catch is that correlation is not fixed; it drifts, and in a crisis it tends to jump toward one.

Purpose

This page defines the correlation coefficient, explains why it, not the number of positions, governs diversification, and stresses the crucial limitation that correlations are unstable and converge in crises.

Visual explanation

Correlation

A correlation matrix showing the pairwise correlations across a set of holdings, from which portfolio diversification is judged.

Correlation MatrixNiftyNiftyBankNiftyBankNiftyITITGoldGold1.000.820.55-0.200.821.000.48-0.150.550.481.00-0.05-0.20-0.15-0.051.00Low or negative correlation is what diversification needs

Professional explanation

What the correlation coefficient measures

The correlation coefficient ρ takes the covariance between two return series and standardises it by dividing by the product of their standard deviations, which strips out the scale of each series and leaves a pure number between −1 and +1. A value near +1 means the two returns rise and fall almost in lockstep; near −1 means they move oppositely; near 0 means there is no linear relationship. Because it is standardised, correlation is directly comparable across pairs regardless of how volatile each asset is, which is why it is the natural language of diversification. It measures only the strength and direction of a linear relationship, not the size of the moves.

Correlation is why diversification works

Diversification reduces portfolio risk precisely because the holdings are not perfectly correlated: when one falls, another may not, so their combined swing is smaller than the sum of their individual swings. The lower the average correlation across a portfolio, the greater the reduction in volatility for a given set of positions. This is why the number of positions is a weak proxy for diversification and correlation is the real driver; ten highly correlated positions are barely more diversified than one. Correlation is the input that turns a collection of bets into a genuinely diversified portfolio, or exposes it as a concentrated one in disguise.

Correlation is not causation, and only linear

A high correlation says two series moved together historically, not that one causes the other, and both may simply respond to a common factor such as the overall market or interest rates. Correlation also captures only linear relationships, so two assets can be strongly related in a non-linear way, for example both crashing together only in extreme moves, while showing a modest ordinary correlation. This tail dependence is invisible to a single correlation figure computed over normal periods, which is one reason the number understates the risk of joint extreme losses. A correlation of zero means no linear link, not independence.

Instability: correlations move, and converge in crises

The most dangerous property of correlation for a risk manager is that it is not stable. Estimated over different windows it can change sign and magnitude, and it systematically rises during market stress, when investors sell everything at once and assets that normally move independently all fall together. A portfolio diversified on calm-period correlations can therefore lose its diversification exactly when it is most needed, in a crash. Any risk model built on a fixed correlation matrix, including portfolio volatility and parametric Value at Risk, inherits this fragility, so correlations should be stressed toward one rather than taken at face value.

Correlation, covariance and beta

Correlation, covariance and beta are related but distinct. Covariance measures co-movement in raw units and is unbounded and hard to interpret; correlation is covariance standardised to the −1 to +1 range; beta is the covariance of an asset with the market divided by the market's variance, so it measures sensitivity rather than tightness. Correlation is the standardised cousin that answers how tightly two things move together, while beta answers how much one moves when the other does. In portfolio construction correlation is used to judge diversification, and beta to judge market exposure, and confusing the two leads to sizing errors.

Formula

ρ = Cov(X, Y) ÷ (σ_X × σ_Y), with −1 ≤ ρ ≤ +1

ρ = the correlation coefficient between return series X and Y. Cov(X, Y) = the covariance of the two return series, the average product of their deviations from their means. σ_X, σ_Y = the standard deviations (volatilities) of X and Y. Dividing the covariance by the product of the two standard deviations standardises it so that ρ always lies between −1 and +1: ρ = +1 is perfect same-direction movement, ρ = 0 is no linear relationship, and ρ = −1 is perfect opposite movement.

What each correlation level means for a portfolio

Correlation ρMeaningDiversification effect
+1Move together perfectlyNone; behaves like one position
Around +0.8Move together strongly (e.g. Nifty and Bank Nifty)Weak; little risk reduction
0No linear relationshipStrong; risks largely offset
−1Move exactly oppositeMaximum; risks can cancel
Rises in crisisEverything falls togetherDiversification disappears when needed

Practical example

Illustrative example (Indian market)

A trader with ₹5,00,000 believes they are diversified across a Nifty position and a Bank Nifty position. Measuring daily returns over the past year gives a correlation of about 0.85, because both are Indian equity indices driven by the same macro forces. At that correlation the two positions provide very little diversification: the portfolio volatility is only marginally below that of a single position, so the trader is effectively running one large equity-index bet, not two independent ones. To genuinely diversify, they would need exposures with low or negative correlation to equities, but even those tend to converge toward the equity market in a sharp sell-off. The lesson is that counting positions overstates diversification when their measured correlation is high.

Across NSE, most liquid large-cap stocks, sectoral indices and the Nifty share high positive correlations because they respond to common drivers such as FII flows, rupee moves and global risk sentiment. A retail portfolio of several such names has far less diversification than its position count suggests, and in an event-driven fall they typically drop together.

Advantages

  • Standardised to −1..+1, so it is comparable across any pair of assets
  • Directly quantifies the diversification benefit between two positions
  • Feeds portfolio volatility, Value at Risk and hedge construction
  • Exposes hidden concentration when positions are secretly correlated
  • Separates same-direction, opposite-direction and unrelated relationships

Limitations

  • Blind spot: it captures only linear co-movement and misses tail dependence, so assets can crash together while showing modest ordinary correlation
  • Correlations are unstable and rise toward one in crises, when diversification is most needed
  • Correlation is not causation; both series may follow a common factor
  • A correlation of zero means no linear link, not statistical independence
  • Estimates depend heavily on the window and frequency of data chosen

Why it matters in practice

  • Decides whether a multi-position book is truly diversified or concentrated
  • Its instability is the reason diversification can fail in a crash

Common mistakes

  • Assuming many positions equals diversification regardless of correlation
  • Trusting calm-period correlations that converge toward one in a sell-off
  • Reading correlation as causation between two assets
  • Treating zero correlation as proof of independence or of no tail link
  • Using too short a window, producing a noisy, unstable correlation
  • Ignoring that a modest average correlation can hide strong joint tail moves

Professional usage

Risk managers monitor the full correlation matrix of a book, not just position counts, and they deliberately stress correlations toward one to see how the portfolio behaves when diversification fails. They prefer exposures whose correlations are genuinely low across regimes, treat any historically low correlation as fragile, and re-estimate over multiple windows to gauge stability. Above all they design for the crisis case, in which correlations converge, rather than the calm case in which the reported matrix flatters the diversification.

Key takeaways

  • Correlation measures how two return series move together, from −1 to +1
  • It equals covariance divided by the product of the two standard deviations
  • Correlation, not the number of positions, drives diversification
  • Correlations are unstable and rise toward one in a crisis, defeating diversification

Frequently asked questions

What is correlation in trading?
Correlation measures how strongly two return series move together, on a scale from −1 to +1. Plus one means they move together perfectly, minus one means they move exactly opposite, and zero means there is no linear relationship. It is the key input for judging diversification.
What is the formula for correlation?
The correlation coefficient equals the covariance between two return series divided by the product of their standard deviations. Standardising by the two volatilities forces the result into the −1 to +1 range, making it comparable across any pair of assets.
What does a correlation of 1 mean?
A correlation of +1 means the two assets move together perfectly in the same direction. Combining them provides no diversification, because the portfolio behaves like a single position; there is no offsetting of one asset's moves by the other.
What does a negative correlation mean?
A negative correlation means the two assets tend to move in opposite directions: when one rises the other tends to fall. Negative correlation is valuable for diversification and hedging because the assets partly cancel each other's swings, reducing portfolio volatility.
Why does correlation matter for diversification?
Because diversification works only when holdings are not perfectly correlated. The lower the average correlation, the more the positions' moves offset, and the more portfolio risk falls. Correlation, not the number of positions, is what determines whether a portfolio is genuinely diversified.
Does zero correlation mean the assets are independent?
No. A correlation of zero means there is no linear relationship, but the assets can still be related in a non-linear way, such as both crashing together in extreme moves. Zero linear correlation is not the same as statistical independence.
Why do correlations rise in a crisis?
In a sharp sell-off, investors sell many assets at once to raise cash, so assets that normally move independently fall together. Correlations converge toward one, and a portfolio that looked diversified on calm-period data loses its diversification exactly when it is most needed.
Is correlation the same as causation?
No. A high correlation means two series moved together historically, not that one causes the other. Often both respond to a shared factor such as the overall market or interest rates, so treating correlation as causation can lead to false conclusions about why assets move.
What is the difference between correlation and beta?
Correlation measures how tightly two returns move together, bounded between −1 and +1. Beta measures how much an asset moves relative to the market and can exceed one. Beta equals correlation times the ratio of the asset's volatility to the market's, so correlation is one component of beta.
How is correlation used in a portfolio?
Correlation feeds the portfolio-variance formula, so it directly sets portfolio volatility and Value at Risk. It is used to judge whether positions diversify each other, to design hedges, and to detect hidden concentration when supposedly separate positions actually move together.
What is a typical correlation between Nifty and Bank Nifty?
Nifty and Bank Nifty are usually strongly positively correlated, often around 0.8 or higher, because both are Indian equity indices driven by common macro factors. That high correlation means holding both provides little diversification against each other.
Can correlation change over time?
Yes, substantially. Correlations drift with the market regime, can change sign, and systematically rise during stress. Because they are unstable, a correlation estimated over one window may not describe the next period, which is why risk models built on fixed correlations are fragile.
What is tail dependence and why does correlation miss it?
Tail dependence is the tendency of assets to move together specifically in extreme events, even if their ordinary correlation is modest. Because the standard correlation coefficient captures only average linear co-movement, it understates the risk of joint extreme losses, so it should not be the only measure of joint risk.
How many data points do I need to estimate correlation reliably?
There is no fixed number, but very short windows give noisy, unstable estimates that can mislead, while very long windows blend different regimes. A practical approach is to estimate over several windows and treat the result as a range, recognising that any single figure is uncertain and regime-dependent.

Voice search & related questions

Natural-language questions people ask about Correlation.

What does correlation mean for my trades?
It tells you whether your positions move together or apart. If they all move together, you are not really diversified, even if you hold many of them.
What is a good correlation for diversification?
Lower is better for diversification. Positions with low or negative correlation smooth your portfolio, while high correlation near one gives you almost no protection.
Are Nifty and Bank Nifty correlated?
Yes, strongly. Both are Indian equity indices driven by the same forces, so they usually move together with a correlation around zero point eight or higher.
Does zero correlation mean no connection?
Not exactly. It means no straight-line link on average. The two can still crash together in a rare shock, which plain correlation does not capture.
Why does diversification fail in a crash?
Because in a panic almost everything falls together. Correlations jump toward one, so positions that used to offset each other now drop at the same time.
Is correlation the same as one thing causing another?
No. Two things moving together does not mean one causes the other. Often they both react to the same underlying driver, like the overall market.

Sources & references

    Last reviewed 12 July 2026. Educational content only — not investment advice. Markets and rules change; verify current conventions with SEBI, NSE/BSE and your broker.

    ← Previous
    Beta
    Educational content only — not investment advice. Examples use illustrative numbers and simplified models. Risk-management techniques reduce but never remove risk, and trading derivatives involves substantial risk of loss. See our Risk Disclosure and SEBI Disclaimer.