The Research Of Quadratic Compensation Incentive Mechanism Based On Multi-task Principal-agent Model

Today competition is more and more fierce and innovation of compensationdesign is pushing forward, along with the increasingly development of economy. Inthis paper, multi-task principal-agent model is reconstructed from the perspective ofoptimize compensation design.First introduce the classic multi-task principal-agent model. And discusses itagain in which performance expected is a linear transformation of efforts.Then, proposed quadratic compensation hypothesis, which with the function ofsmooth income volatility, as linear compensation hypothesis in the classical modeldose not agree with the phenomena of nonlinear compensation rules in reality.Deduce the formula to calculate agent’s certainty equivalent income under quadraticcompensation, set up two multi-task principal-agent model under it, give thenecessary conditions that the optimal solutions of the model have to satisfied.Third, compare linear compensation model with quadratic compensation specialmodel from multiple angles. Combine them with the example of Amway (China)direct sales representative incentive under multi-level direct selling, numericalsolution of the two models is given. And get the conclusion that the more agent riskaversion is, the higher agency cost under the classic model, however the lower underquadratic compensation design. This shows that quadratic compensation design isbeneficial to reduce agency costs when the agent is risk aversion.Finally, the Amway (China) direct sales bonus system is described in detail, itsadvantages and disadvantages is included. And improve it by quadratic compensationdesign. Show that the quadratic compensation design significantly reduces moralhazard phenomena of the direct sales representative stock up impact performance withthe numerical calculation results.