----------------------------------------------------------------------------------------------------

 

 

 

           Department of Systems Engineering and Engineering Management

 

 

 

                                   The Chinese University of Hong Kong

 

 

 

----------------------------------------------------------------------------------------------------

 

 

 

Date: Friday, June 5, 2026, 4:30pm to 5:30pm HKT

 

Venue: ERB513, The Chinese University of Hong Kong

 

Title: Loyalty, Talent, and Dynamics of Nurturing

 

Speaker: Professor Siguang Li, HKUST (Guangzhou)

 

 

 

Abstract

 

A principal allocates resources to an agent whose talent and loyalty are both initially unknown. Resources accelerate talent discovery but also build the agent’s portable strength, which a disloyal agent can deploy competitively. The resulting optimal control problem is three-dimensional, with beliefs about talent and loyalty evolving jointly with accumulated strength. We show that an algebraic cancellation in the loyalty-learning terms makes the value function affine in the loyalty belief within each investment regime, reducing the three-dimensional free-boundary problem to one dimensional characteristic integrals that admit closed-form solutions. Under a solvable benchmark, we obtain explicit cell values and an analytical invest-wait frontier. The principal optimally sequences information: by throttling resources, she slows talent discovery while passive loyalty learning continues, letting loyalty resolve before strength becomes dangerous. Four main results emerge: (i) loyalty buys time—higher loyalty beliefs extend the investment phase; (ii) talent without loyalty is dangerous—the loyalty premium admits an erosion-contraction decomposition in strength; (iii) strategic delay arises endogenously and expands with accumulated strength; and (iv) the principal always underinvests relative to the social optimum, with deadweight loss concentrated among the most talented agents whose loyalty is least verified. We apply the framework to digital platforms managing content creators and to AI capability governance, and explore extensions including non-compete clauses, screening, signaling, and output driven empowerment. This talk is based on a joint with William Cong (Nanyang University of Technology).

 

 

Biography

 

 

Everyone is welcome to attend the talk!

SEEM-5202 Website: http://seminar.se.cuhk.edu.hk