The goal of the references is to ensure that other scholars can locate the paper being referenced. If there is less standardization, the chances of references that omit information and do not actually allow locating the paper increase.
It would be 'kinder' if journals all followed the same citation conventions - especially when having to re-format manuscripts for submission to other journals after rejection. It's not a level playing field though. Some journals are very strict that you exactly follow their house-style conventions, some are a bit looser as long as you broadly follow their recommended convention - others can sometimes suggest a citation style of your choice - as long as the style is consistent. Any re-formatting is picked up at the copy-editing and proofing stages later down the line. The essence there is that, regardless of the citation style, if it is correct then readers can get to the source easily enough.
As to why there are different styles - I think that there are two reasons. One is pragmatic. Numerical (Vancouver) styles take up less copy space. Articles are shorter and therefore cheaper to publish and/or you can get more articles in a journal volume. The other, I suspect, is just historical i.e. editorial preference; usually the original editor. Some journals don't really know what their style is (or confusion has set in along the way) - and they may say their house-style is an 'adaptation' of a particular style. Of course - if you want to bypass this issue to a certain degree - the use of reference managers (such as EndNote) always help.
I publish a lot with Elsevier Justin - both journals and books. Especially with their journal catalogue - they are one of the largest publishing houses in the world. Depending on how widely you publish across a range of their journals - you will find that house-styles for citation will vary across them - and how strict some are and some are not so.