A quantile-based bivariate distribution

Abstract: In this paper we present a flexible bivariate distribution specified by a quantile function. The distribution contains as special cases new bivariate exponential, Pareto I, Pareto II, beta, power, log logistic and uniform distributions and also can approximate many other continuous models. Various $L$-moment based properties of the distribution such as covariance, coskewness, cokurtosis, $L$-correlation, etc are discussed. The distribution is used to model two real data sets.