Blue
Ravenous Bugblatter Beast of Traal
If I can build a Rogue that's better at Survival than the Ranger or Druid every will be or a Rogue better in Arcana than a Wizard, something is seriously wrong.
Expertise allows me to do either (though probably not both at the same time). In my opinion it's a problematic ability.
Great, we have a hypothesis, let's test it.
Okay, so we're starting with a mundane character that is specifically putting part of their features into being educated and skills at arcane or survival. From an in-game narrative, let's call them the savant and the scout.
At 1st level, a wizard probably has a +3 Int and a +2 Proficiency, while the savant has +4 from proficiency. They aren't better unless they also have a 4+ INT modifier, which is possible but means that they are really focusing on being good at the skill, investing heavily in INT as well as a class feature to be good at Arcana.
At 5th proficiency goes up, but the wizard now has a +4 INT mod so the they both have gone up by +2 and it's the same status quo.
At 9th is the same deal - proficiency goes up but an ASI has already happened. Just to review, wizard is a +5 INT +4 proficiency = +9, while the savant is +8 expertise plus INT. So it really depends on how much the savant is investing in intelligence if it's lagging or not.
Mind you, the wizard has invested no class features into improving Arcana, so there is less opportunity cost for the wizard here already. They are improving their primary ability score which they would be doing anyway for spell attack and DC. The savant has invested a class feature, and needs to invest in a ability score that otherwise has no class synergies if they want to pull even, or invest heavily to pull ahead.
This holds true up through 12th. So, the character who has invested heavily can keep up with the generic wizard, in the +/-1 range for levels 1-12, which is where we see most actual play.
At 12th we have another ASI. If there was a feat this would be a chance for the wizard to invest, just like the savant has. Unfortunately the game is short on feats to improve Arcana, about the only one being Prodigy from XGtE, but that's racially limited. Too bad, we can't really evaluate wizard vs. savant with both of them investing in being good at arcana. Just see that the uninvested wizard and rogue who has spent to be good at it are in the same neighborhood
Now, only you can say if that is problematic at your table, I hope looking at it closer that they perceive window of superiority is less than expected at levels of actual play, and does come with a cost.