P.S. Mandel replies that, yes, he agrees with me in general about the problems of ratios where the denominator is a sum of positive and negative components, but that in this particular case, all the major components of regulatory employment change are either positive or a very tiny negative.