Since we are interested
in comparing different values of an inequality measures such as
the difference between pre-payment and post-payment Gini coefficients,
we adopt the following tests: let V1 and V2 be vectors of
BTS values of measure (G) in the periods 1 and 2, for instance,
pre-payment income and post-payment income measured over
the same households in a particular year, then D= V2 −V1. The test
rejects the null hypothesis that the coefficients for the two periods
are the same; i.e., the equality between coefficients, if the estimated
confidence interval for the difference of the Gini coefficients does
not include zero.