![MathJax Logo](/templates/jsp/_style2/_tandf/pb2/images/math-jax.gif)
Ebaid et al. (2020) proposed the following new extended FGM copula
(1)
(1)
Ebaid et al. (2020) claimed that, the admissible range of the parameters is
(2)
(2)
Based on the admissible range Equation(2)
(2)
(2) , Ebaid et al. (2020) claimed that the Spearman’s rho,
and Kendall’s tau,
Unfortunately, these ranges of the correlations, as well as the given admissible range Equation(2)
(2)
(2) , are wrong as the following example shows:
Counter example
Pick out two admissible values b = 0.9 and
clearly
which doesn’t make sense?
Pick out the admissible value
It can be shown that the probability density function of the copula Equation(1)
(1)
(1) at
and v = 0 is negative for all admissible values
Remark.
Now the important question is: Does the copula Equation(1)(1)
(1) (of course with the true admissible range which was not determined for the two parameters a and b in Ebaid et al., 2020, or even in this correction) increase the maximum positive correlations for the FGM copula
Unfortunately, the answer is no, as the following example shows: Consider the FGM family with the two marginals
and
(3)
(3)
Clearly, the ranges of the Spearman’s rho,
and Kendall’s tau,
for the FGM family (3) are respectively
and
On the other hand,
Thus,
and