Portfolio Correlation: Are Your Holdings Really Diversified?
Ben Ghabili · Published
Several holdings can leave you exposed to the same company, sector or economic risk. Counting tickers does not reveal how many genuinely different exposures you own.
Correlation helps describe how returns have moved together. A holdings audit helps explain why they might do so. Use both before calling a portfolio diversified.
What does portfolio correlation measure?
Correlation measures a relationship between paired observations. For investments, Pearson correlation is commonly calculated on returns over matching intervals to describe their linear co-movement. It ranges from −1 to +1 when defined, but does not establish causation or guarantee a future relationship.
Positive correlation means the sampled returns tend to move together relative to their averages. Negative correlation means they tend to move in opposite directions relative to their averages. A value near zero indicates little linear relationship in that sample, not proof of independence or an absence of shared risks.
State the return frequency, sample dates and return convention. Daily price returns over a year and monthly total returns over several years are different measurements. Non-matching trading calendars and stale prices can also distort comparisons.
If either return series has no variation, ordinary Pearson correlation is undefined. Do not automatically interpret a blank result as zero correlation.
Does high correlation mean equal returns or losses?
No. Correlation describes co-movement, not the size of gains and losses. Two investments can have perfect positive correlation in a sample while one moves much more than the other.
Consider four fictional consecutive sessions, using same-currency price returns. All inputs are invented and exclude dividends and costs.
| Session | Investment A | Investment B |
|---|---|---|
| 1 | +1% | +2% |
| 2 | −1% | −2% |
| 3 | +2% | +4% |
| 4 | −2% | −4% |
B's return is exactly twice A's in each observation. Both sample means are zero, so their Pearson correlation is +1. They have identical directions in this example but different movement sizes.
Four invented observations demonstrate the calculation, not a reliable estimate of a real relationship. Nor does the example establish that this pattern continues outside the sample.
Can low historical correlation guarantee protection?
No. A historical correlation estimate describes the chosen sample. Relationships can change, and a low overall estimate can conceal simultaneous losses during particular periods. Diversification can reduce some risks without eliminating market losses.
For that reason, complement an aggregate coefficient with actual joint outcomes. What happened when the portfolio's largest risk materialised? Did both investments lose value then? Were the losses magnified by leverage, currency exposure or liquidity constraints?
Do not assume an asset is a hedge simply because its average correlation was low. A hedge claim requires a defined risk, relevant observations and an understanding of how the exposure behaves when that risk occurs.
How should you assess diversification?
Begin with what you own, then examine how those exposures behaved. This sequence keeps a statistical result connected to an economic explanation.
- 1.Calculate portfolio weights. Use current values, a common currency and a common valuation time.
- 2.Look through funds. Identify overlapping companies and major concentrations. Use dated holdings rather than assuming the fund name explains its exposure.
- 3.Group shared dependencies. Consider sectors, geography, borrowing conditions, currencies and major business drivers. Different classifications do not necessarily mean different economic risks.
- 4.Measure comparable returns. Align dates and definitions; disclose missing observations and sample length.
- 5.Inspect difficult periods. Review joint losses as well as the full-sample coefficient. Avoid treating a tiny subsample as a precise prediction.
- 6.Relate findings to your constraints. Liquidity needs, holding period and leverage affect what a loss would mean in practice.
A portfolio concentrated in a sector may be intentional. The issue is whether that concentration is understood and consistent with the objective, not whether every investor must own a fixed number of funds.
How does sector analysis fit into a portfolio review?
Sector analysis can help identify shared market exposures, but a sector label is not a complete risk description. Companies in different sectors may depend on the same spending cycle, while companies within a sector can have different balance sheets and revenue sources.
Start by recording your holdings' actual exposures. Then compare the relevant groups' performance using Tapelab's sector page. That page supplies market context, not a holdings-level correlation calculation or a diagnosis of your portfolio.
A diversification conclusion you can defend
Replace “I own several funds, so I am diversified” with a specific observation:
In this fictional portfolio, 25% of starting value is exposed to Company A through direct shares and two funds. The number of holdings understates that single-company concentration.
That conclusion follows from the stated inputs. It does not promise a future loss or prescribe a trade.
For your next review, choose your largest holding and trace any additional exposure through funds. Then examine comparable return relationships. A look-through audit and a correlation estimate answer different questions, and neither should stand in for the other.
Nothing here is investment advice.