Total reflection cannot occur with a negative delay time

It was recently predicted [ in ibid., vol. 33, p. 519, 1997] that the frustrated Gires-Tournois interferometer exhibits a negative delay time for reflection. Given its 100% reflectivity, this appears to contradict causality. We demonstrate that an additional, positive, contribution comes from consideration of the transverse dimension. We prove that this contribution is always large enough to enforce a positive total delay.