When Positions You Thought Were Independent Move Together

A trader builds a book of five positions, each one sized as if it were an independent bet with its own risk budget. On paper, the book looks diversified: a long in a regional bank, a short in a homebuilder, a long in an airline, a short in a shipping company, a long in a semiconductor supplier. Five different sectors, five different theses. Then a single macro headline — a sharper-than-expected rate move — hits, and all five positions lose money on the same day, in the same direction, at the same time. The book wasn't five independent bets. It was one bet on interest-rate sensitivity, sized five times over without anyone noticing.

Why this keeps happening

Correlation between positions is usually estimated from history, and history is a poor guide to exactly the days that matter most, because normal-day correlations and stress-day correlations are often not the same number at all. A bank stock and a homebuilder stock might show a modest 0.3 correlation across a calm year, both drifting on their own sector news most of the time — and then move to 0.9 for the one week that a rate shock hits, because it turns out both were quietly the same trade: "this business does worse when rates rise fast." The 0.3 was real. It just wasn't the number that mattered.

A scenario

A trader's book carries five positions, each individually sized to a $60k risk budget, for $300k of total intended risk if the positions are genuinely independent. The trader ran a rough correlation check when each position was added and found nothing alarming — none of the pairwise correlations exceeded 0.4 in the trailing year of data.

Position	Sector	Rationale
Long regional bank	Financials	Cheap on book value
Short homebuilder	Consumer discretionary	Overpriced on forward earnings
Long airline	Industrials	Post-pandemic demand recovery
Short shipping company	Industrials	Overcapacity
Long semiconductor supplier	Technology	New product cycle

A surprise rate hike hits. The bank falls because higher funding costs squeeze its margin outlook faster than expected; the homebuilder short works because higher mortgage rates hit housing demand — that's one loss and one gain, roughly offsetting. But the airline falls on higher fuel-hedge financing costs, the shipping short fails to work because a broad risk-off move drags every industrial name down together regardless of individual fundamentals, and the semiconductor long falls because growth stocks broadly re-rate lower when rates rise. Four of five positions lose money on the same day. The realized loss is close to $220k — not the $300k maximum the risk budget nominally allowed for five independent trades going wrong simultaneously, which the trader had judged to be extremely unlikely, but far more than the correlation estimates used at entry would have predicted for one macro print.

How to catch it before the surprise

The check that would have caught this isn't a better correlation matrix on the individual stocks — it's asking, for every position, "what single macro variable, if it moved sharply tomorrow, would hurt this trade?" and then looking at how many positions share the same answer. Four of the five names in the scenario above share the answer "a sharp move up in rates." That's the real position size, and it was four times what any single risk budget assumed.

What a desk does about it

Once the shared driver is identified, the fix isn't necessarily to close positions — it's to size the book as if it were one $300k bet on rate sensitivity rather than five $60k independent bets, and decide deliberately whether that concentrated size is still something the desk wants to carry. Sometimes the answer is yes, at a smaller total size. Sometimes it's cutting one or two of the positions specifically to reduce the shared exposure, keeping the ones with the best individual thesis and the least overlap with the rest.