Ok let's explore this a little bit, I don't care about Federer's efficacy in dispatching early round opponents so I'll focus on just the top 10.
Looking at equivalent periods, Federer in 04-07 held 87.8% of the time, and in 2014-2017 he held 88.5% of the time. So there is a slight edge to 2014-2017, it's less than 1% - there's a full 5% drop in return games won though. Which is significant in a game of fine margins as you call it. The proportion of clay matches probably has something to do with both those stats as well, 7 matches in 14-17 vs 14 in 04-07. If we look at just HC the most represented surface, he actually held more versus the top 10 in 2004-2007 (90.4% vs 88.6%). Though I guess you could call that the Djokovic factor in 2014/2015? I don't think it's clear at all though that you'd want Older Fed serving for you if your life depended on it against an elite opponent. On grass he held serve better in 2014-2017, holding serve 91.3% of the time versus 89.8 but there's a monstrous 10% difference in return games one 24.3% versus 14%...Not sure how exactly you'd quantify Federer's decline but outside of specific tournaments and runs it's quite obvious IMO. Massive decline is too far in a game we agree is often decided by fine margins but it's enough to flip some results in 2004-2007 the other way IMO, or turn four set wins into hard five setters etc...
I'll do a quick look at Djokovic, but I don't want to include the back half of 2016, so 2011-2016.5 he held serve against the top 10 83.9% of the time (broke 29.9%), on HC this 85% and 32.2%. From mid 2018 to 2023 he was at 86%/23.1, and 88.4%/23.5% on HC. I would think the quality of the top 10 opponents in this timeframe versus the earlier one influences these but I guess you'd say the reverse for Federer. So on face value he did have a significant uptick in hold percentage, with a slightly bigger drop in return effectiveness. What I would say here though is that if we look at say 2015, he held serve 86.7% of the time and broke 32% of the time, versus 2023 where was at 88.5% holds and 22.5% breaks - which again looks like a significant dip to me.
I do (rightly) think the serve is the most important shot but even Fed was never a serve bot - not enough to offset the drop in return quality.
If you look at the physicality of the 2012-2013 Djokorray matches at the AO I don't see how you could favour 2021/2023 Djokovic for example. Perhaps 2019 was good enough to avoid that kind of attritional match but that's just one year. The best of the next generation went down so easily at the AO in 2021/2023, USO 2023 etc...I just don't see how you could argue the level tennis hasn't regressed. Not when we have dozens of matches where second tier talents at least took sets or even beat the Big 3 when they were in or closer to their primes.
If I put my salt shaker down for a moment I will say that the weakness of the nextgen doesn't invalidate Djokovic's achievements nor does it mean he doesn't have a strong GOAT case. For me it just means there's still plenty of room for debate, which is not the opinion of many. So I do push back perhaps a little too hard.