2 Comments
User's avatar
Talaat | MacroFragility Game's avatar

I have to admit that my knowledge of copula models is limited. I think that you might have used t-copula and it obviously showed weak correlation during tails. Correct me if I am wrong here, but that would mean that decoupling of bitcoin from the S&P 500 when US imposed tarrifs was still something that is somewhat expected knowing the unpredictable nature of bitcoin. Is that right?

Expand full comment
Orca's Spread's avatar

Gaussian Copula, t-Copula requires better data and using something as simple as max drawdown would be fine. The Copula measures jumps, not very extremes. – think of it as measuring jumps to Default Risk Vs. Jumps to Worst Case Default Risk.

It shows that dislocation was informative and not expected.

Expand full comment