Decreasing performance pay in the risk-aversion of the principal

What is the intuition behind a decreasing performance pay ß if both principal and agent are risk averse compared to a principal being risk neutral and an agent being risk averse?

Actually, intuitively I think it would make more sense if the performance pay to the agent increases, such that the agent would bear more risk, so more risk is relocated from principal to agent if the principal is risk averse as well.

• "...decreasing performance pay ß if both principal and agent are risk averse compared to a principal being risk neutral and an agent being risk averse"---what's the source of this claim? – Michael Jul 20 at 9:51