OK then that naturally leads to a slightly different hypothesis: take the candle problem, but have 3 groups: a control group that is paid nothing, a group that is paid "decently and fairly", and a group that is paid according to incentives.
Salary, in this case, is not the type of incentive we're talking about. Having a decent salary is just to get the employee willing to work - not as an incentive to work at max capacity.
What we're talking about, really, is already past salary: is a group of fairly paid people more efficient and creative when motivated financially (say, by extra bonuses or salary increases) or by promises of autonomy (say, 20% time).
Thus, the candle problem, when placed in the real world, would have three groups: a control group which is paid 'decently and fairly', and a group that is paid 'decently and fairly' and is given financial incentives beyond that, and a group that is payed 'decently and fairly' and is given autonomy as an incentive.
That's fine, but you still need a control group that is given no compensation at all as a baseline.