Is there any research how much benefit "hyper targeting" actually gives? Not just "it helps" but quantitatively (e.g. more effect than 30 just mildly target adds? Or just more effect than 2?)

In particular, how does it compare to somewhat targeted offline ads like (local shop advertising in local newspaper) or (car stuff advertised in car enthusiast magazine)?

Also, the question is about having a level playing field. If your competition has access to highly effective advertising, you can't afford to not use it as well. So right now you need it. However, if hyper targeting were forbidden (because the spying and stalking is prohibited by law or because browsers make it infeasible) then you don't need it.