Some of the issues you often run into with Japanese subtitles are:

  * Almost all Japanese subtitles include subtitles for every noise made as well as including SDH-like information about sounds. This means that a naive attempt to just match up subtitle timings won't work -- the English translation will have fewer subtitles and you will have to skip Japanese lines that do not have an equivalent in the English transcript.
  * Most translated subtitles simplify things and have to re-organise sentence structures to match the target language, which means that you often have to pick different timings for how a sentence is broken up in English than in the original Japanese. Sometimes a very short Japanese phrase requires two sentences to accurately translate, sometimes a long Japanese sentence can be translated into a fairly short English phrase.
  * Higher-quality subtitles will also provide translations for signs and other on-screen text (ASS supports custom placement, fonts, styles, and colours -- it is powerful enough to the point where some of the really good fansub jobs I've seen make it look like the video was actually localised to English because all of the signs look like they have been translated in the original video). The original timings don't help with this.

I should mention that I have used tools like alass[1] to re-time subtitles between languages before (including retiming Japanese subtitles to match English ones) so this is not an unreasonable idea on its face, but those tools mostly work with already existing subtitle tracks that have correct timings. My experience is that if you have tracks with very different timings (as opposed to chunks of subtitles with fairly fixed offsets) you start getting rubbish results.

[1]: https://github.com/kaegi/alass