December 8th, 2004, 3:00 am
This follows directly from the definition of basic statistical functions.Variance(X) = E(X^2) - E(X)E(X)Covariance(X,Y) = E(XY) - E(X)E(Y)Correlation(X,Y) = Covariance(X,Y)/[Variance(X)*Variance(Y)]^0.5For an indicator variable, E(X) = E(X^2) = p. E(XY) is just the joint probability of X and Y both being 1. So if p12 is the probability of joint default, the correlation coefficient is:(p12 - p1*p2)/[(p1 - p1^2)(p2 - p2^2)]^0.5= (p12 - p1*p2)/[p1*p2*(1-p1)*(1-p2)]^0.5