Versions published in different countries can differ - it's not uncommon for books to be published in one version in the US/Canada, and slightly differently elsewhere. Even a minor difference like binding could conceivably affect page references, so it's important to be clear in that sense. There's also a wider principle of redundancy; if the author does get some part of the citation wrong (say, misspells the author's surname) the more information included in the rest of the citation the better the reader has a chance of recovering the actual item.
With regard to why we use human-meaningful information rather than details like an ISBN, I would venture that dense information like an ISBN is actually quite difficult to read for a human; it's easy for the author to mistype the number when producing the bibliography, and then easy for the reader to mistype it when looking up the reference. It's comparatively harder to get details like publisher name and city wrong. If I'm looking through a bibliography or list of references I much prefer to see information I understand than a collection of opaque numbers.
Having said the above, I'm sure to a large extent it's an historical tradition that has persisted. Certainly before the development of international reference standards like the ISBN system the above arguments held even more weight, and in the days of much less extreme globalisation the various global offices of major publishing houses were much more autonomous. It might conceivably have been quite a different undertaking to be published by, say, Oxford University Press in Oxford rather than Oxford University Press in New York.