Glossary
Correlation measures how closely two assets move together, from +1 (identical) to -1 (opposite). Diversification works by combining low correlations - and fails in panics, when the correlations rise toward one.
Correlation is the statistical measure of how two assets move together, scored between +1 and -1. A correlation of +1 means the assets move identically, 0 means no relationship, and -1 means they move opposite. Diversification works on the number: combining assets with low or negative correlations produces a portfolio less volatile than any of its parts, because the falls land at different times and different sizes.
The catch is that correlations are not constant - they rise in panics. In March 2020, correlations across asset classes spiked toward 1 as everything sold together, and the diversification a portfolio counted on evaporated for exactly the weeks it was needed. The Australian manifestation is the index itself: ASX-listed stocks correlate heavily with each other because the index is concentrated in banks and miners, and international shares add a different driver - different cycles, different currency, different central bank.
The portfolio arithmetic is the first mechanism. Two assets each with 15 percent volatility and a 0.3 correlation combine into a portfolio with roughly 12.1 percent volatility - lower than either asset alone, which is the diversification working. Push the correlation to 0.9 and the same pair combines into roughly 14.6 percent volatility - the diversification evaporated, and the portfolio behaves like a single asset with a bad day built in. The correlation calculator runs any pair.
The panic is the second mechanism, and the structural one. Correlations between risky assets rise in sell-offs because the selling is driven by the same forces: deleveraging, margin calls and redemptions. In March 2020 the correlation between shares and almost everything else spiked toward 1 - the exception being government bonds, which usually hold the line but wobbled even there in the worst week. A portfolio built on 2019's correlations owned 2020's correlations, which is the trap: the number is measured on calm data and needed in the panic.
The index is the third mechanism, and the Australian one. ASX stocks correlate with each other at high levels because the market is concentrated - banks and miners dominate the capitalisation, and a resources slump takes most of the register with it. Adding international shares introduces different drivers - different cycles, currency exposure, a different central bank - and lowers the correlation of the whole portfolio. The diversification entry covers the split, and the portfolio rebalancing entry covers keeping it honest.
For an Australian reader, correlation is the number behind the advice to diversify beyond the ASX. A portfolio of ten ASX stocks is less diversified than it looks, because the stocks correlate with each other at high levels through the banks and miners that dominate the index. International shares, bonds and cash run on different drivers, and the correlations between them are lower in calm markets - which is the diversification that holds. The portfolio diversification calculator runs the arithmetic on any mix.
A colleague named Brent (yes, that Brent) held ten ASX shares across five sectors and called it diversified. The March 2020 panic repriced all ten together - the correlations he had counted on spiked toward 1, and the portfolio fell like a single stock. The highest-leverage thing most Australians can do is size the diversification on the correlations that hold in panics, not the ones measured in calm years - and that usually means the whole market, some of it international.
Two assets each with 15 percent volatility and a 0.3 correlation combine into a portfolio with roughly 12.1 percent volatility - lower than either part, because the falls land at different times. The same pair at a 0.9 correlation combines into roughly 14.6 percent - a portfolio that behaves like one asset. The falls land at different sizes and times; the portfolio smooths them. The correlation calculator runs any pair at any correlation.
In the five weeks of March 2020, correlations across asset classes spiked toward 1 as deleveraging and redemptions sold everything together. Portfolios built on calm-market correlations fell like single stocks; government bonds held the line better than shares but wobbled in the worst week. The correlations held in 2019 were not the correlations of the panic. The diversification that was counted on was measured in 2019 and needed in the panic.
Ten ASX shares across five sectors carry high correlations with each other because the index is dominated by banks and miners - a resources slump or a rate cycle takes the register together. The same ten-stock portfolio with half the weight in international shares runs on different drivers: different cycles, currency, a different central bank. The sectors looked different; the correlations were not. The drivers, not the sector labels, decide the correlation. The portfolio diversification calculator prices any mix.
Mistake: Calling ten ASX stocks a diversified portfolio.
Why it bites you: The stocks correlate with each other at high levels through the banks and miners that dominate the index, and the portfolio falls like a single asset. Brent's five sectors were one market in a panic.
Mistake: Measuring diversification on calm-market correlations.
Why it bites you: Correlations rise toward 1 in panics - March 2020 sold everything together - and the diversification counted on evaporates when needed. The number is measured in calm and needed in the storm.
Mistake: Assuming bonds always protect the portfolio.
Why it bites you: Government bonds usually hold the line and wobbled even there in the worst weeks of March 2020; corporate bonds and credit sold with the shares. The hedge has its own correlation, and it rises in panics too.
Mistake: Reading a correlation as permanent.
Why it bites you: Correlations drift with cycles and regimes, and the pair that diversified for a decade can converge for a year. The number needs re-measuring, which is why the rebalancing exists.
I ran two-asset portfolios through correlations of 0.3, 0.6 and 0.9 and re-ran the March 2020 panic against the same mixes so you don't have to. The pattern is consistent: diversification works on the calm numbers and is tested by the panic numbers, and the only defence is owning things with genuinely different drivers.
Brent held ten ASX stocks across five sectors and called it diversified; March 2020 repriced all ten in the same five weeks. The correlation is the number behind the advice to own the whole market and some of it offshore - not because offshore is better, but because it moves for different reasons. Diversify on drivers, not on sector labels.
The lower the better - assets with correlations below about 0.5 meaningfully reduce portfolio volatility, and negative correlations reduce it further. The catch is that correlations rise in panics, so the calm-market number overstates the protection.
Correlations across asset classes spiked toward 1 in the panic as deleveraging and redemptions forced selling of everything at once. The diversification built on calm-market correlations evaporated for exactly the weeks it was needed.
Yes - different cycles, currency exposure and a different central bank lower the correlation to an all-Australian portfolio. The ASX is concentrated in banks and miners, and international shares run on different drivers.
It compares the returns of two assets over the same period and scores the relationship between +1 and -1. The number is historical, drifts with regimes, and needs re-measuring as the market changes.
Yes - a correlation of -1 means the assets move opposite, and genuinely negative correlations are the strongest form of diversification. In practice, reliable negative correlations are rare and usually limited to specific hedging instruments.
Disclaimer: This information is general in nature and does not constitute financial or legal advice. Always consult a qualified professional for your specific situation.

Financial Chaos Analyst
Ivy Sinclair-Wren is a Financial Chaos Analyst covering investing, AI, wealth psychology, and the emotional consequences of opening finance apps during market crashes. Based in Melbourne, she specializes in demystifying the Australian tax code and helping users navigate the intersection of spreadsheet logic and human irrationality.