# exponential of Covariance of two variables in logs, cov(X,Y)=exp(X)*exp(Y)*cov[ln(X),ln(Y)] [duplicate]

I am stuck in calculating steady state in a model that has covariances in logs. I am wondering in general if the following accurate.

cov(X,Y)=exp(X)*exp(Y)*cov[ln(X),ln(Y)]

if that is accurate could you please let me know which theorem/textbook could be useful?

Thanks a lot and wish you a nice evening

## marked as duplicate by whuber♦Jul 17 '18 at 23:38

• That formula can't be right. For one thing, ${\rm cov}(X,Y)$ is a scalar and the thing on the right hand side is a random variable. – within_person Jul 17 '18 at 20:18
• This is trivially true for degenerate random variables (i.e. constants), since $\text{Cov}(X, Y) = 0$ if they are real numbers ;) – Kevin Li Jul 17 '18 at 20:22