While I can't mention an example where above doesn't apply, it isn't - as far as I am aware - a fixed rule. There is an unwritten rule that scientific names as far as possible should elude the obvious risks of confusion, i.e. the spelling shouldn't be too similar to other species/taxa in the genus/species and they shouldn't rime. This is not a fixed rule, but most people that are involved in the naming of taxa today try to follow it. It would be a bit problematic if it was a fixed rule due to the frequent change of genus and/or status of various taxa (there's already too many people that are trying to argue for the upgrading of the junior synonyms of various species!). BTW, I don't have the time (or possibly, will) to check if I can find any examples of subspecies that are named the same as a species in the same genus, but if someone should feel like checking I'd guess genera like Pachycephala, Turdus, Anthus, Troglodytes and Synallaxis would be worth a try. I am fairly sure an example or two should be out there, but, as observed MMSLouis, it certainly isn't a common occurrence.