Phase control is measured by the "power factor," which should normally be close to 1. This is not an issue for the relatively small and simple loads that households normally put on the grid. Industrial customers must provide a load that is resistive, neither inductive nor capacitive, and they pay a penalty if this is not achieved. For example, if a factory has a lot of electric motors, it will have to pay a power factor penalty because they are not resistive loads.

If you're supplying electricity to the grid, you must follow the same rules--in reverse, sort of--and if you don't then the system gets messed up. So as I understand it, the issue for the electric company when net metering is used, is to make sure that the electricity supplied to the grid meets the phase requirements--plus a bunch of other rules. That's something that they normally don't need to worry about with residential customers, so it's an incremental burden on them.

Here is a list of rules that one utility requires you to follow. They're pretty complex, and "somebody" has to make sure, in a net metering environment, that they're being follwed.
http://www.chelanpud.org/Snap/Interconnection.htm