Continuous-time incentives in hierarchies

TL;DR

This paper develops a continuous-time hierarchical principal-agent model using advanced stochastic control techniques, revealing how incentives can be optimally structured in complex multi-level organizations.

Contribution

It introduces a novel continuous-time framework for hierarchical incentives, utilizing second-order backward stochastic differential equations to handle drift and volatility control.

Findings

Managers control volatility of agents' utilities, not just their drift.

The model simplifies the principal's problem to a fixed-dimensional space regardless of hierarchy size.

The approach enables comparative statics and large-scale hierarchy extensions.

Abstract

This paper studies continuous-time optimal contracting in a hierarchy problem which generalises the model of Sung (2015). The hierarchy is modeled by a series of interlinked principal-agent problems, leading to a sequence of Stackelberg equilibria. More precisely, the principal can contract with the managers to incentivise them to act in her best interest, despite only observing the net benefits of the total hierarchy. Managers in turn subcontract with the agents below them. Both agents and managers independently control in continuous time a stochastic process representing their outcome. First, we show through a continuous-time adaptation of Sung's model that, even if the agents only control the drift of their outcome, their manager controls the volatility of their continuation utility. This first simple example justifies the use of recent results on optimal contracting for drift and volatility control, and therefore the theory of second-order backward stochastic differential equations, developed in the theoretical part of this paper, dedicated to a more general model. The comprehensive approach we outline highlights the benefits of considering a continuous-time model and opens the way to obtain comparative statics. We also explain how the model can be extended to a large-scale principal-agent hierarchy. Since the principal's problem can be reduced to only an $m$-dimensional state space and a $2m$-dimensional control set, where $m$ is the number of managers immediately below her, and is therefore independent of the size of the hierarchy below these managers, the dimension of the problem does not explode.