Trevon Capital runs a long-short book at 400% gross on Rs 100 crore of equity, with a prime broker requiring equity of at least 25% of gross. Longs and shorts are equal. How far can longs fall with shorts flat before a margin call, and what happens if longs fall while shorts rise?

Trevon is already at the limit, so any loss triggers a margin call. If longs fall 5% with shorts flat, equity is Rs 90 crore against Rs 390 crore of gross, 23.1%, and Trevon must cut Rs 30 crore of gross: three rupees of trading for each rupee lost. If longs fall and shorts rise 3%, it must cut Rs 48 crore, and in crowded names that selling moves prices further against it. Running at 300% gross would give room for an 8.3% move in both legs.

A borrower whose bank requires the loan to stay below 80% of a flat's value, and who borrows exactly 80%, is in breach the first day the flat's price slips. Trevon's equity is exactly 25% of its gross exposure, so it has no buffer: every rupee lost comes out of equity in full, while gross falls by the same rupee from a base four times larger. After a 5% fall in the longs with shorts flat, equity is Rs 90 crore and gross Rs 390 crore, a ratio of 23.1%. The limit allows gross of four times equity, Rs 360 crore, so Trevon must cut Rs 30 crore, half by selling longs and half by buying back shorts.

This is the move that hurts long-short books: the popular longs fall and the popular shorts rally at the same time. A 3% move in each leg costs Rs 12 crore while leaving gross unchanged at Rs 400 crore, because the longs shrink by as much as the shorts grow, so the ratio falls to 22% and Trevon must cut Rs 48 crore of gross, four rupees for every rupee lost. The picture shows how the two kinds of move compare, and how much room a 300% book would have had.

Because Trevon sells what everyone else holds. Selling its longs pushes them down; buying back its shorts pushes them up; both widen the very spread that caused the call. Assume each Rs 10 crore of forced trading moves both legs a further 0.5%: the first round of cuts, Rs 48 crore, loses another Rs 8.4 crore, which forces a second round, and equity ends near Rs 67 crore after the original Rs 12 crore shock. The impact figure is an illustration, but the shape is the one seen whenever crowded long-short books deleverage together: losses arrive in rounds, each smaller than the last, and the fund sells at the worst prices of the move.

Keep a buffer sized to the move you expect to survive. At 300% gross, Rs 150 crore a side, equity starts at 33.3% of gross; the book can take an 8.3% move against both legs, or a fall of more than 20% in the longs alone, before a call. Choose the buffer from how the book has moved in its worst weeks, not its average ones, and measure crowding: the more of your book other funds own, the larger the second-round losses. Say the limit of this analysis too: prime brokers can raise margin requirements in a stress, which moves the line itself up just when the ratio is falling.