Esta entrada ainda não foi traduzida. Estamos exibindo a versão em inglês por enquanto.
- Também conhecido como:
- tail dependence, crisis correlation, downside co-moves, joint tail risk
Tail dependency asks a simple question: do two assets crash together on the worst days?
It's like correlation, but focused only on the left tail (bad days).
Why it matters
Two assets can have low average correlation and still melt down together in a crisis. Tail dependency captures that.
How we approximate it at Gale Finance
We look at shared dates and count days where each asset has a big downside move (by default, a z-score below -2 on its own log return series). Then we compute conditional overlap:
- "When asset B has a tail day, how often does asset A also have one?"
That's a practical, readable proxy for tail dependence. We display the simple returns on those dates for readability, but the threshold is based on log-return z-scores.
Use it with
Tail dependency is most useful alongside correlation and explicit tail metrics like VaR and Expected Shortfall.