It was noted by Stuart Anderson that applying the kappas has a drastic effect on h(t) around the cal line frequencies, as seen from ASD ratio plots of h(t) with and without kappas applied (see attached plots). It turns out that this is due to the fact that the DARM line is injected into the DARM loop right after the digital filters, affecting its phase in the residual and control chains in a way that causes them to partially cancel when added. See the subway map:
https://dcc.ligo.org/DocDB/0122/G1501518/014/darmschematics-v14.pdf
Applying the kappas affects the degree of cancellation, as they multiply the residual chain and the tst and pum/uim stages of the actuation separately. To see quantitive analysis of the data that produced the plots below, refer to the pdf attachment.