Diversification and correlation: five tickers can share one risk
Calculate two equally weighted assets and separate the number of names from common risk exposures.
Diversification spreads investments across different sources of risk. Five funds may own the same companies or depend on one industry or currency. The number of rows in a holdings list does not establish how different their risks are.
Look through the labels
Compare fund holdings, largest-company weights, industries, currencies and interest-rate exposures. Funds from different managers can overlap substantially. A fund focused on a single industry does not automatically provide broad diversification.
Correlation describes how consistently comparable returns moved together across multiple observations. It is not similarity in the height of price charts. The estimate depends on dates, frequency, currency and data; a past value does not promise the same relationship during stress.
One period with explicit weights
Each table row is a separate invented scenario for the same period, not a sequence of trades. Start with 100, split 50 into A and 50 into B. Weights are 50/50; there are no flows, distributions, debt or costs. A row’s return is 0.5 ×rA+0.5 ×rB.
When both assets lose 10%, the portfolio loses 10%. With+10% and−10%, the changes offset: final value 100 and return 0%. If they fall 10% and 20%, final value 85 means a 15% loss. These are weight calculations, not estimates of correlation or scenario probabilities.
Our own teaching example; all figures are hypothetical.
r = 0.5 × rA + 0.5 × rB; 0.5 × 10% + 0.5 × (−10%) = 0%
| Scenario | A;50% weight | B;50% weight | Portfolio | End |
|---|---|---|---|---|
| Joint decline | −10% | −10% | −10% | 90 |
| Different directions | +10% | −10% | 0% | 100 |
| Reverse directions | −10% | +10% | 0% | 100 |
| Both decline | −10% | −20% | −15% | 85 |
What this cannot establish
Dependence estimation requires return series over multiple periods and an explicit method. Even low correlation does not mean independence, no common exposure or protection against every loss. Weights drift as asset values change unless reset; that is a separate rebalancing question.
Record the common-risk hypothesis and conditions for checking it instead of selecting an attractive number. Elliott does not optimise portfolios or recommend asset weights. A practical next step is to preserve observations and limitations in the scenario journal.
With 80% in A and 20% in B, A returns+10% and B−10% in one period. What is the result before costs?
0.8 × 10%+0.2 ×(−10%)=6%. An initial 100 becomes 106. This calculates one scenario; it does not recommend an allocation.
Sources
Sources checked: 2026-10-21
This material and its translations were prepared with AI assistance. Educational calculations were checked programmatically. Sources explain concepts; examples and files were created for this article.
Related reading
Next step
Record holdings, weights, dates and the testable hypothesis in the learning journal.
A market-scenario journal: test an idea without hindsight →