It's purchase price.
Anyone in the game can have expertise. Cost: one level of rogue. Anyone can have Rage. Cost: one level of barbarian. There is a cost as far as delaying progress on your base class.
Not liking: everyone gets expertise. Cost: warlord on party with right subclass. Everyone gets Rage. Cost: warlord on party with right subclass. That means the characters benefitting don't spend a resource (even a spell slot) to get what would cost a level otherwise to get. The only cost even remotely close is if the warlord player didn't want to be a warlord with that subclass; if he did, the cost is practically nothing.
True uniqueness is rare, but that doesn't mean we should erode niche protection further, especially at such a low barrier to entry...
Is that actually what's being described, though?
Expertise is a hell of a lot better than "minimum roll = one person's Int mod." Even if this were a level 1 feature (and nothing I've seen suggests that people want it to be), never rolling less than 3/4/5 is not anywhere near as good as adding twice your Proficiency. I also don't recall anyone suggesting "everyone gets Rage," though for that matter you could be simply mirroring my examples, in which case I guess my response is "I'm not saying these things
should be Warlord-grantable stuff." Just that your argument seems weak when so little is actually, functionally unique.
I also think there's more of a cost to "someone
else in the party picks a specific class and subclass, with specific stat requirements" than you seem to. Otherwise, it would seem like every magic-user ever is just about the most broken thing to exist--because my Fighter can get crazy buffs from the Cleric or Wizard, and
not even need to concentrate on them!
All that said, I do hear and understand your point. A class whose thing is picking one class, and making everyone else half as good as that class, is probably not a great idea. I'm just not sure that that's actually what people are suggesting. I think, instead, it's more like, "Hey, it's really cool that the 4e Warlord was great at making everyone better in tactical combat. Wouldn't it be even cooler if, since 5e's balance is a lot more 'open' (I would call it 'loose at best' but that's just me), there could be subclasses that made the group, as a whole and individually, better at things that
aren't combat?"
The example mechanics might be poor ideas. But to strike down the entire thing--"No that's horrible, you're destroying any reason to ever play anything else!!"--seems a tad excessive. "No, try again" is how I'd respond to that--perhaps with an added, "and try to make it
synergize with other classes, rather than simply nicking their stuff." I do think it's possible. I think it's entirely possible to have an "Arcane Warlord" who can do things like "maintain concentration on a spell so the Wizard/Cleric/Druid/etc. doesn't have to." More mechanics would be needed to flesh out the concept completely, but that basic mechanic sounds like a plausible starting point. It's a totally valid design space that I'd never have thought of myself, but that has all sorts of interesting potential. I think it's entirely possible to have a "Covert Warlord" that does
something with stealth--I just don't know what, because stealth really isn't my
thing. I would utterly love to see a "Scholar Warlord" that, as others have said, forks off all the "knowing stuff really well"/"drawing logical conclusions" side of Intelligence, without carrying any of the spellcasting.
I know these aren't what you are looking for--you want hard mechanics. I just don't trust myself to
make hard mechanics--and I'm hoping someone else in the thread
does, and will come along and help out. Maybe if I took a couple hours to study the stealthy classes (mostly Rogue and Shadow Monk, but Ranger too I guess) I could come up with something sufficiently-meaty. But it would be nice if every. Single. Proposal. weren't met with an immediate "That's
horrible, you're destroying everything good about X!" or "That's not even a
feat, let alone a whole
class!"
Edit:
Since a lot of this stuff is being presented as subclass material, though, it would probably be better to start thinking of
how much subclass matters for the Warlord, because that will determine how alike or unalike two Warlords of different subclasses are as well as how much heft subclass can hold. For Wizards and Barbarians, subclass is
mostly flavor with a few mechanics attached (some strong, some not so strong, but not really "carrying" the class). For Warlock, Cleric, and Druid, subclass (if you allow Patron+Pact to count as a two-part subclass choice) is huge, defining several elements of playstyle. For Fighter and Rogue, it's somewhere in between--I'd argue that Fighter is more heavily class-focused, while Rogue is more heavily subclass-focused.
Once *that* is decided, you can start to figure out how powerful these "facilitating things that aren't combat" features can be. If the Warlord's
stuff is 80% base class, 20% subclass (e.g. like the Wizard), yeah, giving out serious bennies would be excessive. On the other hand, if the Warlord is more like 40% base class, 60% subclass (e.g. like the Warlock) then having major, method-defining elements as subclass features isn't
as big a deal. You still need them to be balanced, and to not invalidate other players' choices, but there
is design space for it if we're willing to really push what 5e can do.
Edit II:
One simple example for a
single, small mechanic that a "covert subclass" Warlord could have just came to mind. "Armored Stealth: Through rigorous, careful training, observation, and the application of common materials in a unique way, you can help your allies overcome the clumsiness and noise of heavy armor. After completing a short or long rest, select a number of allies up to (number--possibly Int or Wis mod, possibly Proficiency score, possibly something else--exact balance would need testing), which can include yourself. The selected creatures ignore any disadvantage on Stealth checks caused by their armor for the next 30 minutes. Other sources of disadvantage on Stealth checks still apply normally. You must complete another short or long rest in order to apply these treatments again--they require time, patience, and minimal distraction to apply."
This strikes me as a perfectly fine, singular ability provided as part of an over-arching subclass associated with facilitating stealth. It's not a huge benefit, and most lightly-armored, high-Dex parties won't benefit from it, but in the right group it could be transformative. I'm sure it could be edited down to be less cumbersome (I tend to be verbose), but it seems to me like a perfectly cromulent, initial stealthing ability that could either be built upon later (perhaps, at high level, granting advantage on Stealth checks if wearing armor that doesn't cause disadvantage?), or applied alongside other sneaky/stealthy/disguise-y mechanics.