Managerial Incentive Problems: A Dynamic Perspective

Holmström, B., “Managerial Incentive Problems: A Dynamic Perspective,” Review of Economic Studies 66 (1999), 169-182.

Purpose: To show how a manager’s desire to build his resume and firm owners’ desire to earn financial returns can conflict or harmonize.

Findings:

  • Fama (1980) posited that market forces eliminate the agency problem over time.  A manager will choose to maximize shareholder returns because, in the long run, he is concerned about his long-term career prospects.
    • Under some narrow assumptions, Fama is right, but in general, he is not.
  • When the market is uncertain about a manager’s ability, he will work harder to achieve good outcomes and look like a high-ability manager.
    • Ability changes over time and is never fully known, so managers can always substitute effort for ability.
  • Managers’ supply of effort/output in every period is less than the socially optimal level.
  • A risk-averse manager may not want to invest, since investing poorly could reveal low managerial ability, and since it cannot be proved that investments not made would have been successful.

The Basic Model:

  • Effort is unobservable and cannot be contracted, so managers in each period are paid in advance for their efforts.
  • A manager’s ability is revealed over time through the outcomes of his effort.
  • A risk-neutral market pays a manager wages w_t(y^{t-1} = E[\eta | y^{t-1}] + a_t(y^{t-1}).
    • y^{t-1} is the series of historical output through time t-1.
    • w_t is wages paid at the beginning of period t.
    • \eta represents belief about manager ability.
    • a_t \in [0,1] is the manager’s decision rule, or how much effort he gives in period t.
  • The manager maximizes the expected discounted payoff of his current and future wages, minus the current and expected disutility of effort g(a_t), where g(\cdot) is an increasing and convex function.
  • Solving these two equation gives the equilibrium wage and managerial decision rule.
  • When there is uncertainty, there are gains to effort.
    • High effort biases the market’s ability estimating process upward, and lower effort biases it downward.
  • Managerial ability changes over time, and so never becomes fully known.