Setting up a full-time mining rig is a tremendous resource commitment – and you either have to be all-in or go home if you want to ever actually see any return (yeah, I know – I'm ignoring syndicate/collective mining).
On the other hand, running a full node that doesn't mine is really not that expensive a thing to do – and provides all of the transaction validation, consensus voting, and block propagation that a mining node does.
For developers (and perhaps for users who just, well, use Ethereum a lot) there is a direct benefit to running such a node, but I wonder if there could be some way to provide incentive for other people to run these sorts of nodes. When I think about it I'm not seeing anything obvious and/or elegant.
I guess a corollary is whether or not it would even be a good idea. Seems to me the more nodes the better, though.
Best Answer
Currently there is indeed no monetary incentive to run a full node. It's done by enthusiasts that just do it for the fun of it, or by projects who are building on top of Ethereum and need access to network data (also wallets run it at the moment because light clients aren't done yet, they will certainly switch away s soon as they can).
The inherent problem with light nodes is that they leach the network: why run a full node when you can get away with a light node? To address this we're considering a model where light nodes actually pay for data they request from full nodes. If we can work out a meaningful model for this, then this would be the incentive for people to run full nodes: they can serve light nodes, and be compensated for it.
Also we're working on Swarm, which would allow anyone to rent out their disk capacity for others to use as a distributed file system, which could also net full nodes some extra income for their efforts.