On the minimum and maximum of bivariate lognormal random variables (Q2488472): Difference between revisions

Let \((u,\nu)\) be a bivariate Gaussian vector, \(R=\min(e^u,e^\nu)\), \(Z=\max(e^u,e^\nu)\). The author investigates monotonicity properties of \(\mathbf{E}R\), \(\mathbf{E}Z\) and Var\(R\) as functions of such parameters of \((u,\nu)\) as marginal means, variances and \(\rho=\text{corr}(u,\nu)\). E.g., \(\mathbf{E}R\) and \(\mathbf{E}Z\) are monotonically increasing functions of \(\rho\).