It suffices to show that the following relation holds
for any m \in {\mathbb Z}_{\ge 1}:
\begin{aligned}
\sum_{\lambda \in {\mathcal P}}
S_{[\lambda]}(X|Y)S_{\lambda }(T^{(m)})
&=\prod_{j=1}^{m}
\left(
\frac{\prod_{i} (1-t_{j}y_{i})}{\prod_{i} (1-t_{j}x_{i})}
\right)
\prod_{1 \le i \le j \le m}(1-t_{i}t_{j}),
\end{aligned}
where T^{(m)}=\{t_{1},t_{2},\dots, t_{m}\}.
Substituting the Schur function
\begin{aligned}
S_{\lambda }(T^{(m)})=
\frac{\det_{1 \le i,j \le m} (t_{i}^{\lambda_{j}+m-j})}
{\det_{1 \le i,j \le m} (t_{i}^{m-j})}
\end{aligned}
into the left hand side of (Proof), and multiplying
the Vandermonde determinant
\begin{aligned}
\det_{1 \le i,j \le m} (t_{i}^{m-j})
=\prod_{1 \le i < j \le m}(t_{i}-t_{j})
\end{aligned}
on both sides, we obtain
\begin{gathered}
\sum_{\lambda \in {\mathcal P}}
S_{[\lambda]}(X|Y)\det_{1 \le i,j \le m} (t_{i}^{\lambda_{j}+m-j})
=
\\
=\prod_{j=1}^{m}
\left(
\frac{\prod_{i} (1-t_{j}y_{i})}{\prod_{i} (1-t_{j}x_{i})}
\right)
\prod_{i=1}^{m}(1-t_{i}^{2})
\prod_{1 \le i < j \le m}^{m}(t_{i}-t_{j})(1-t_{i}t_{j}).
\end{gathered}
In order to prove this, we need the following identity
\begin{aligned}
\det_{1 \le i,j \le m} (t_{i}^{m-j}-t_{i}^{m+j})
=\prod_{i=1}^{m}(1-t_{i}^{2})
\prod_{1 \le i < j \le m}^{m}(t_{i}-t_{j})(1-t_{i}t_{j}).
\end{aligned}
By setting t_{i} \to t_{i}+t_{i}^{-1} in (Proof), we obtain
\begin{aligned}
\det_{1 \le i,j \le m} ((t_{i}+t_{i}^{-1})^{m-j})
=(-1)^{\frac{m(m-1)}{2}}\prod_{i=1}^{m}t_{i}^{1-m}
\prod_{1 \le i < j \le m}^{m}(t_{i}-t_{j})(1-t_{i}t_{j}).
\end{aligned}
The (i,j)-matrix element of the left hand side of (Proof)
can be expanded as
\begin{gathered}
(t_{i}+t_{i}^{-1})^{m-j}
= \sum_{k=0}^{m-j}\binom{m-j}{k}t_{i}^{2k-m+j}
\\
=
\begin{cases}
1 & \text{if} \quad m-j=0
\\
t_{i}^{m-j}+t_{i}^{-m+j}+
\sum_{k=1}^{\frac{m-j}{2}-1}\binom{m-j}{k}
(t_{i}^{-2k+m-j}+t_{i}^{2k-m+j})+\binom{m-j}{\frac{m-j}{2}} & \text{if} \quad m-j \in 2\mathbb{Z}_{\ge 1}
\\
t_{i}^{m-j}+t_{i}^{-m+j}+
\sum_{k=1}^{\frac{m-j-1}{2}}\binom{m-j}{k}
(t_{i}^{-2k+m-j}+t_{i}^{2k-m+j}) &
\! \! \! \!
\text{if} \quad m-j \in 2\mathbb{Z}_{\ge 0}+1
\end{cases}
\end{gathered}
Thus, in the left hand side of (Proof),
by repeatedly performing the operation of subtracting
\binom{m-j}{k} times the (j+2k)-th column from the
j-th column of the matrix, for k=1,2,\dots , (m-j)/2
if m-j \in 2\mathbb{Z}_{\ge 1}, and
for k=1,2,\dots , (m-j-1)/2
if m-j \in 2\mathbb{Z}_{\ge 1}+1,
starting from j=m-2 down to j=1,
we arrive at the following relation:
\begin{aligned}
\det_{1 \le i,j \le m} ((t_{i}+t_{i}^{-1})^{m-j})
&=\frac{1}{2}\det_{1 \le i,j \le m} (t_{i}^{m-j}+t_{i}^{-m+j}).
\\
&=
\frac{(-1)^{\frac{m(m-1)}{2}}}{2}\det_{1 \le i,j \le m} (t_{i}^{j-1}+t_{i}^{-j+1}).
\end{aligned}
Note that the prefactor 1/2 in the right hand side of
(Proof) is necessary since
(t_{i}^{m-j}+t_{i}^{-m+j}) becomes 2 at j=m while (Proof) becomes 1.
Multiplying (-1)^{\frac{m(m-1)}{2}}\prod_{i=1}^{m}t_{i}^{m-1}
(1-t_{i}^{2}) on both sides of (Proof), we obtain
\begin{gathered}
(-1)^{\frac{m(m-1)}{2}}\prod_{i=1}^{m}t_{i}^{m-1}
(1-t_{i}^{2})
\det_{1 \le i,j \le m} ((t_{i}+t_{i}^{-1})^{m-j})
\\
=
\frac{1}{2}\det_{1 \le i,j \le m} (t_{i}^{m+j-2}-t_{i}^{m+j}+
t_{i}^{m-j}-t_{i}^{m-j+2}).
\end{gathered}
Thus, in the right hand side of (Proof),
after multiplying the first column of the matrix by the prefactor
1/2,
we perform a sequence of column operations in strictly descending order of the column index j, starting from j=m down to j=3.
Specifically, for each such j, we add the (j-2k)-th column
to the j-th column, where k=1,2,\dots ,(j-1)/2
if j \in 2\mathbb{Z}_{\ge 1}+1, and
for k=1,2,\dots ,j/2-1 if j \in 2\mathbb{Z}_{\ge 2}.
Through this process, we find that (Proof) is equivalent to
\det_{1 \le i,j \le m} (t_{i}^{m-j}-t_{i}^{m+j}),
and thus (Proof) is equivalent to (Proof).
Let us modify the right hand side of (Proof) as
\begin{gathered}
\text{r.h.s. of \eqref{co-Bm2}}=
\prod_{j=1}^{m}
\left(
\frac{\prod_{i} (1-t_{j}y_{i})}{\prod_{i} (1-t_{j}x_{i})}
\right)
\det_{1 \le i,j \le m} (t_{i}^{m-j}-t_{i}^{m+j})
\quad \text{[by \eqref{mmdet}]}
\\
=\prod_{j=1}^{m}
\left(
\sum_{k_{j}=0}^{\infty} h_{k_{j}}(X|Y)t^{k_{j}}_{j}
\right)
\det_{1 \le i,j \le m} (t_{i}^{m-j}-t_{i}^{m+j})
\quad \text{[by \eqref{genh}]}
\\
=
\left(
\sum_{k_{1}=0}^{\infty}
\dots
\sum_{k_{m}=0}^{\infty}
\prod_{i=1}^{m}
h_{k_{i}}(X|Y)t^{k_{i}}_{i}
\right)
\sum_{\sigma \in \mathcal{S}_{m}}
\mathrm{sgn}(\sigma)
\prod_{i=1}^{m}
(t_{i}^{m-\sigma(i)}-t_{i}^{m+\sigma(i)})
\\
=
\sum_{k_{1}=0}^{\infty}
\dots
\sum_{k_{m}=0}^{\infty}
\sum_{\sigma \in \mathcal{S}_{m}}
\sum_{\xi_{1} \in \{-1,1\}}
\dots
\sum_{\xi_{m} \in \{-1,1\}}
\mathrm{sgn}(\sigma)
\prod_{i=1}^{m}
h_{k_{i}}(X|Y)\xi_{i} t^{k_{i}+m-\xi_{i}\sigma(i)}_{i},
\end{gathered}
where \mathcal{S}_{m} denotes the symmetric group on \{1,2,\dots ,m \}.
By setting \eta_{i}=k_{i}+m-\xi_{i}\sigma(i), we obtain
\begin{gathered}
=
\sum_{\eta_{1}=0}^{\infty}
\dots
\sum_{\eta_{m}=0}^{\infty}
\sum_{\sigma \in \mathcal{S}_{m}}
\sum_{\xi_{1} \in \{-1,1\}}
\dots
\sum_{\xi_{m} \in \{-1,1\}}
\mathrm{sgn}(\sigma)
\prod_{i=1}^{m}
h_{\eta_{i}-m+\xi_{i}\sigma(i)}(X|Y)\xi_{i} t^{\eta_{i}}_{i}
\\
=
\sum_{\eta_{1}=0}^{\infty}
\dots
\sum_{\eta_{m}=0}^{\infty}
\sum_{\sigma \in \mathcal{S}_{m}}
\mathrm{sgn}(\sigma)
\prod_{i=1}^{m}
(h_{\eta_{i}-m+\sigma(i)}(X|Y)-
h_{\eta_{i}-m-\sigma(i)}(X|Y)) t^{\eta_{i}}_{i}
\\
=
\sum_{\eta_{1}=0}^{\infty}
\dots
\sum_{\eta_{m}=0}^{\infty}
\det_{1 \le i,j \le m} (h_{\eta_{i}-m+j}(X|Y)-
h_{\eta_{i}-m-j}(X|Y)) \prod_{i=1}^{m} t^{\eta_{i}}_{i} .
\end{gathered}
If any of the elements of \{\eta_{1},\eta_{2},\dots ,\eta_{m}\} are equal to each other, the determinant above becomes zero, so by rearranging the rows, it takes the following form.
\begin{gathered}
=
\sum_{\eta_{1}>\eta_{2} >\dots >\eta_{m} \ge 0}
\sum_{\sigma \in \mathcal{S}_{m}}
\mathrm{sgn}(\sigma)
\det_{1 \le i,j \le m} (h_{\eta_{i}-m+j}(X|Y)-
h_{\eta_{i}-m-j}(X|Y)) \prod_{i=1}^{m} t^{\eta_{\sigma(i)}}_{i}
\\
=
\sum_{\eta_{1}>\eta_{2} >\dots >\eta_{m} \ge 0}
\det_{1 \le i,j \le m} (h_{\eta_{i}-m+j}(X|Y)-
h_{\eta_{i}-m-j}(X|Y))
\det_{1 \le i,j \le m}( t^{\eta_{j}}_{i}).
\end{gathered}
By setting
\eta_{i}=\lambda_{i}+m-i, we obtain
\begin{gathered}
=
\sum_{\lambda_{1} \ge \lambda_{2} \ge
\dots \ge \lambda_{m} \ge 0}
\det_{1 \le i,j \le m} (h_{\lambda_{i}-i+j}(X|Y)-
h_{\lambda_{i}-i-j}(X|Y))
\det_{1 \le i,j \le m}( t^{\lambda_{j}+m-j}_{i})
\\
=\text{[left hand side of \eqref{co-Bm2}]}.
\end{gathered}
This proves (Proof), and thus (3.2).