Which of your holdings are secretly the same bet?
Correlation shows up in a spreadsheet as a number. In real portfolios it usually shows up as an overlap nobody noticed. Here is where it hides.
The five common overlaps
1. Fund inside fund. Two broad index funds from different providers tracking similar large-cap universes are, for risk purposes, close to one holding. Two funds is a marketing fact; one exposure is the financial fact.
2. Top-heavy indices. A market-cap-weighted index is not evenly spread. When the largest handful of companies grow into a large share of an index's value, holding that index plus those same companies individually doubles an exposure already owned — often without anyone noticing that the index itself had become concentrated.
3. Employer overlap. Salary, bonus, career prospects and share holdings all sourced from one company is the structure that made the Enron story a permanent teaching example. The correlation between "my job" and "my savings" is close to +1 in exactly the scenario where it matters most.
4. Currency and country. Holdings across five industries that are all denominated in one currency and sold into one domestic economy share a macro engine. Home bias is real and near-universal; it is worth seeing, whatever anyone chooses to do about it.
5. One factor, many costumes. A set of unprofitable fast-growing companies from five different industries can still be a single bet on interest rates staying low, because long-duration cash flows reprice together. Sector labels can hide a shared factor completely.
The diagnostic question
For each holding, ask: what would have to be true in the world for this to fall 30%? Write the answer in one clause. Then read the list of answers. If four holdings produced the same clause, that is the real position size — not what any single row says.
This is a strictly descriptive exercise. It tells you what a portfolio is, which is a fact. It does not tell you what it should be, which is a decision this course never makes for anyone.
A rounded illustration
An investor lists eight holdings and feels well spread. Written out, the "what breaks this?" clauses read: rates rise, rates rise, rates rise, consumer spending falls, rates rise, rates rise, oil falls, rates rise. Eight rows, three stories, one dominant one. Nothing here says the portfolio is wrong — but the word "diversified" was doing work the contents don't support.
In the data
Inside a fund wrapper the overlap can be counted, because every fund publishes its largest holdings with their weights. An S&P 500 fund and a Nasdaq-100 fund:
Two funds with different names and different stated categories can share the same first five companies. The summed weight of a company across the wrappers, not the number of wrappers, is the real position size.
Try it now
Write the "what would make this fall 30%?" clause for every name in your watchlist, then group the identical clauses.
Make overlaps 1 and 2 countable rather than felt. In the two lists above, how many companies appear in both, and what do their weights sum to across the two? That sum, not the number of wrappers, is the position size.
Count your distinct clauses. That number, not the row count, is what Unit 1's formula was really asking for.