Abstract
This study proposes the
use of Johnson’s (1949) multivariate SU distribution for
estimating the joint distribution among economic well-being variables. The SU
distribution exhibits extremely high flexibility, enabling it to effectively
capture the extreme skewness and kurtosis commonly seen in wealth and income
variables. As this distribution imposes no constraints on variable ranges, it
is well-suited for estimating the distributions of net worth or disposable
income, which often include non-positive values. As an illustrative example, we
employ the bivariate SU model to estimate the joint
distribution of net worth and age using the US survey data.
