In the Netherland's stream, I posted the following which is applicable here.
"It is indeed more work to create and organize the category structure, but that can be done over time. I hope you will leave open the possibility of having English-language geographic categories to mirror the Dutch ones."
It is not only extra work to create and organize the category structure, but also a great deal more work to maintain the multilingual structure as it increases the number of categories that must be watched to make sure that the extra categories are created and organized correctly. This is no small burden, especially for smaller projects and projects which are already struggling with an insufficient number of bilingual people to do all of the other work requiring bilingual fluency.
Spreading the extra work over time is often no answer at all. For smaller projects or those with a limited number of members willing and able to work on creating categories, all the extra time we have is already needed just to create single language categories for all the places we need to categorize all our profiles properly.
Projects need the option of deciding what work is most important for meeting their goals. Limiting locations categories below a specified level to the native language to improve consistency and conserve people resources so more effort can be put into putting profiles in landing level location categories often makes a great deal of sense for regional projects.
Making the aka template available and setting guidelines for how to handle multiple language category streams for projects which find it worth the effort and want to do it is one thing. Pushing projects toward establishing guidelines which support multi-lingual language strings in all or most levels of location categories is an entirely different matter. There are valid reasons for restricting location categories below a specified level to a single native language. Having an authoritative resource to consult for place names in only the native language which can be used to improve consistency in category names is one such valid reason.
Leave it to the regional projects to decide whether limiting categories to a single language below a certain level is beneficial or not in their geographic area.
Additionally, introducing so called equivalencies perpetuates some misconceptions about administrative divisions. For example, Herred or Herreder is often translated as Hundred and Amt or Amter as County or Counties and Sogn or Sogne as Parish or Parishes. This encourages the perception that these administrative divisions are the equivalent of counties, hundreds and parishes in England or the United States, which they are not.
Such translations can tend to move some people toward the idea that if such translations exist, then the units are sufficiently similar that we can move toward standardizing the format of all or most location categories along the lines of town/city, county, state/province with sometimes country tacked on the end. In some places and times whatever gets translated as "county" for example may or may not be a significant administrative divider.
Sticking to the native language can be a reminder to pay attention to the context of the country and its administrative history as well as what is genealogically important in deciding what works best for a geographic area.
Although parallel language structures add a great deal to the understanding of what categories are intended to cover in topical and non-regional categories, place names generally don't receive the same benefit from multilingual translations. Geographic terms are the one area where both tourists and genealogists have to adapt to the locality and be able to deal with the native language to function at even the most basic level. I think it is reasonable for a genealogy site to expect all users to be able to function with place names in the native languages.
Again, leave this choice to the projects. It is one area that does not need to be standardized and in fact benefits greatly from variations.