About Houserules and Handling Balance
@mxyzplk’s answer states that “balance is overrated” – well, that depends on how highly it is rated in the first place. It is certainly not the be-all, end-all consideration of game design. But it should not be ignored, either. Running the game exactly as written out of a mistaken belief that it is necessary to do so to maintain balance is a mistake; you’re paralyzing yourself. Running things however you feel like it without regard to the consequences, however, is also a mistake.
In a cooperative game, like the overwhelming majority of role-playing games, the most important form of balance is “intra-party,” that is, within a party. For the most part, no one should be dead weight, and no one should be doing everything. Ideally, everyone is generally competent, so that people are not side-lined by scenarios that differ from their forte, but everyone has specialties where they can shine.
It is worth noting that Dungeons and Dragons 3.5 and Pathfinder do not protect you from either situation. Wizards can do everything, if the player is sufficiently knowledgeable and motivated to do so; Monks can be dead-weight, if played naïvely and the rest of the group is competent. So houserules and “gentlemen’s agreements” are necessary, even with the basic rules, to achieve intra-party balance. For the most part, this does not even have to be codified; it’s simply a matter of all the players being on the same page and having roughly the same expectations and knowledge of the system.
So for houserules, you need to be wary of anything that might exacerbate the intra-party balance issues, but if you’re already successfully playing 3.5 or Pathfinder, you’ve been “naturally” handling some very severe intra-party balance issues as it is. This should make you fairly comfortable with the idea of modifying the rules, at least with respect to balance: the rules are already rather imbalanced anyway.
I do recommend reading JaronK’s Tier System for Classes. It’s for 3.5, but Pathfinder did not change the game that much: the Core classes listed there are in the same tier in Pathfinder as they are in 3.5 (excepting maybe the Paladin, who might have been bumped up), and more importantly, the reasons that a given class is in one tier or another remain the same. Knowing these can allow you to recognize houserules that will be fine, and houserules that will be problems (mostly, you don’t want to nerf mundane classes and you don’t want to buff spellcasting classes; the divergence of their power levels is already quite extreme).
By which I mean, the party vs. the rest of the world. This is generally a much lesser issue: as a DM, you can throw anything you feel is appropriate at the party. The CR rules from 3.5 are close to useless anyway, which means a good DM needs a very good idea of what his party can do and what his NPCs can do, and how an encounter is likely to go. And, of course, the PCs always have the option of retreat or what have you. Finally, houserules that are applied equally to PCs and NPCs are unlikely to significantly affect balance anyway.
In this regard, I would not worry about balance much at all, or at least, not for the sake of gauging a houserule. You always have to consider balance if you are aiming for a particular encounter to hold a certain level of challenge for your PCs, but houserules don’t change that.
I will add one caveat here, however: rules that increase the randomness, applied equally to PCs and NPCs, massively favor NPCs. PCs have to survive every encounter, usually around four per day. Team NPC, on the other hand, only has to get lucky once. Even if something has a 1/400 chance (i.e. a nat-20 followed by a nat-20 to confirm), it’s going to happen to a PC at some point. They’ll do it to NPCs too, but that matters less because there are always more NPCs.
So, for example, critical fumble rules (i.e. you don’t just miss, you do something actively bad to yourself or your allies), or super-critical hits, are very much bad for the game because sooner or later, a PC is going to die to something they had no way of protecting themself from or responding to – because someone just happened to get that very-unlikely roll. This is bad.
Your Particular Case
In your case, you are talking about giving all of the players a benefit. If all of them get it, then intra-party balance is not affected significantly, particularly if the benefits are greater for mundane characters than spellcasters (as is often the case with Teamwork benefits).
The inter-party balance is less important, and you can easily cancel out any effect you dislike by giving enemy parties the same deal. Even if you do not do this, the Teamwork benefits are small, which means you give the players a small edge – exactly what you stated you were out to do, considering the general lethality of the world. That’s entirely reasonable.
So I say go ahead.