In an article on testing reciprocity failure in Photo Techniques magazine in July/Aug 2003, Howard Bond found that modern films have less increase in contrast, and thus less need than older style films for reduced development with the long exposures required to correct for reciprocity failure. TMY and 100 Delta showed no increase in contrast at 4 minutes exposure, and Tri-X only a small increase. The Kodak recommended development reduction for Tri-X of 10%, 20%, or 30% at different extended exposure times didn't fit Bond's findings at all.
I've also seen this reported elsewhere in a more anecdotal fashion, without such rigorous testing as Bond applied.
So I think the best advice with modern emulsions is to test and/or critically examine your results as you work, and not necessarily take conventional wisdom based on older style emulsions as always valid with current films.
Of course many night scenes (especially in cities with artificial lighting) tend to have a much greater dynamic range than typical daylight shooting, which might call for reduced development. But this isn't a case of reciprocity failure increasing effective contrast.
Lee