The major consideration is that, by the time you reach level 17, there's nothing in the world that can threaten you. From the time you get there, until the end of the campaign, is what? Five sessions, maybe?
The unlikeliness of anyone dying during that time, in a manner that can't be fixed trivially with a 3rd level spell, makes the supremacy of Wish kind of pointless. Wish is flat-out better than Resurrection, in much the same way that Teleport is better than Dimension Door. When you get to high levels, your tools become way more efficient. That's just how it goes.
I mean, if you're worried about the mere 1000 GP that it takes to cast the spell, then keep in mind that a CR 17 monster is supposed to have thousands of GP, or hundreds of PP. Cash isn't supposed to be a meaningful limitation on Resurrection, by the time you get to level 17.
My concerns include campaign world logic and considerations as opposed to just PC considerations. That said, I am not sure that five sessions gets you from 17th level to the end of a campaign. Not only are there provisions in the DMG for going beyond that, but I don't think PCs tend to level up every 1 or 2 sessions at higher level.
In any event, wish also has the long term problem of avoiding all material component costs. In 3rd edition, Wish had a high cost (5,000 XP at least) and that made the spell a reserve for "oh crud" moments....for emergencies.
Now, it merely becomes the means for a high level wizard or sorcerer to be more effective than a cleric or druid in some respects, especially for effects that work out of combat, where the one wish per day limit doesn't matter. Or to spam spell effects that should not be spammed.
For example, lets look at the Symbol spell.
Symbol is a 7th level spell. So it is high level, and you get it 4 levels before Wish, so it's not that far removed. The spell lasts until discharged, so it can be spammed out of combat easily at 1 casting per day.
Clearly, the Symbol spell is balanced by the need to use a 1,000 gp gem as a consumable material component, to stop some wizard from casting symbol on every square inch of his tower.
In fact, even a 16th level wizard has to pay for that spell, again to keep it in check.
But that same 16th level wizard goes up 1 more level, and suddenly, using Wish, he can spam the heck out of that spell, casting it once per day anywhere he wants at no cost whatsoever. This is a problem. Material component costs were put onto spells to balance them and Wish comes along and discards that balance, as if there were no need for such things at 17th level. I find that troublesome.
And yes, I know I can house rule it (and have), but I think some thought should have been given by the designers to reign in Wish.
What I have done is say that Wish works as written if you supply the material components. That puts a 17th level wizard on par with a cleric, but not ahead of his (except for casting time). A Wish caster can forego a material component, but for each 1,000 gp of components not used, there is a chance that the wizard will become stressed and suffer the effects of stress under the Wish spell,. The chance is very low, but not negligible. It is enough to stop a 17th level wizard from spamming symbols or what have you will-nilly.