What is the essence of the problem with incentives based on traditional economic models, according to Palepu and also the lecture about the problems with economic models?
They are not mathematically rigorous enough.
The payments schedules they explicitly recommend, which increase with some kind of measure of performance, are exactly the kind of schedules that have been observed to result in employee "gaming" and dysfunctional activities.
They result in incentive that are not strong enough for real situations.
They recommend using penalties, but not reward