A friend of mine asked me to give a talk on DMing, so I've been writing and rewriting endless cycles of notes on that. One theme that I know I want to hit is putting players mostly-in-control of the difficulty of the combat that they face, by giving them free rein to scout the dungeon and avoid dangerous monsters. Another theme is that really strong play often requires submission of the individual party member to the interests of the party - the wizard has to hold on to his one spell carefully rather than using it early for personal satisfaction, the thief has to expose himself to substantial personal risk in order to find traps so that they don't kill anyone (or everyone) else, and fighters are sometimes called upon to die for the party so that everyone else can live. Clerics, of course, are clerics, and have always operated been expected to operate "in service" to the party as a whole.
But when you take these two ideas, players collectively in control of their destiny and subordination of the individual to the party, and combine them with XP for treasure, a funny possibility arises.
Under a certain reading of the rules, you earn 1 XP per GP that you receive from the adventure. If the party as a whole determines (by vote, say) that some player failed to contribute and to pull their weight, by (say) fleeing from combat, the party as a whole need not allocate any GP from the treasure to that player, and hence he will earn little XP - just the monster XP portion, which is around 20% of XP on most adventures. A persistent problem player who regularly earns 20% XP will eventually end up about 2 levels behind the rest of the party. A player allocated only half a share of gold will tend to end up one level behind. This is, obviously, not a form of censure that should be used lightly, but it is one that should be considered. I suspect that, as with PC death, this is not something that needs to happen often for it to modify player behavior substantially. The knowledge that it is a possibility encourages cooperation.
The reverse also applies - if you have a new player join a group and he's playing a class that you really need, the party can allocate greater than a share to him. If someone's fighter died in a heroic rearguard action to cover the party's retreat and now they're playing a henchman and a level behind, you can allocate them greater than a share until they come up to level parity. The trivial, common case is that a PC is 10 XP from leveling with standard shares, and a tiny reallocation might push them across a threshold.
Unfortunately, henchmen complicate this XP allocation scheme, for one because they receive an odd share size, and for two because traditionally they receive their shares straight from the party pool, but if a player receives an odd-sized share, do his henchmen too? In the case of a problem player, it seems reasonable to dock his henchmen as well; in the case of a hero player, it seems reasonable to award his henchmen extra as well. But this all complicates the accounting, or raises again the specter of "the pay for your henchmen should come out of your share of the treasure", which I suspect most sensible parties will reject on the grounds that a player with many henchmen is of great value to the party as a whole.
At the end of the day, one of the interesting facets of this game not present in more modern editions is that it is in many ways an experiment in small-scale self-governance, civics, and organizational behavior. And yes, the gaming community seems to have settled on very regular norms regarding the allocation of treasure and XP, but one of the joys of transgressive/retrogressive gaming, allowing things like PvP and uneven allocation of treasure, is an opportunity to reinvent those norms (with good stories and object lessons about why we have those norms), or to arrive at new and strange norms. And isn't that what the OSR is all about? Rolling back the clock and seeing the other possibilities that could have been, not merely in the rules of the game, but in the rules of the group?