5 - Imperfect Performance Measurement Flashcards
What are the elements of a model in multitasking and imperfect performance measurement?
- eᵢ ≥ 0: A’s non‐observable effort in task i (i= 1, 2)
- c(e₁, e₂) = 1/2 e₁²+ 1/2 e₂²: A’s effort costs
- A is risk neutral and his reservation utility is zero
- y ∈ {0, 1}: tasks’ contribution to firm value
- p ∈ { 0, 1 }: verifiable performance measure
What is y and what is problematic about it in the imperfect performance measurement model?
- y = tasks’ contribution to firm value
- y is non‐verifiable and thus cannot be part of an incentive contract
How is y defined in a model of imperfect performance management?
y ∈ {0, 1}: tasks’ contribution to firm value
- Pr [y= 1 | e₁, e₂] =f₁e₁ + f₂e₂
- fᵢ ≥ 0 is task i′s expected marginal productivity
fᵢ (i = 1, 2) is such that all probabilities remain below 1 at the first‐best and second‐best solution
How is p defined in a model of imperfect performance management?
p ∈ { 0, 1 }: verifiable performance measure
- Pr [p = 1 | e₁, e₂] =g₁e₁ + g₂e₂
- gᵢ ≥ 0 is task i′s expected marginal impact on the performance measure
- in general fᵢ ≠ gᵢ
gᵢ (i = 1, 2) is such that all probabilities remain below 1 at the first‐best and second‐best solution
What is the timing of an imperfect performance measurement problem?
- P offers bonus contract (a,b)
- A can accept (-> stage 3) or reject (reservation utility 0)
- A chooses effort e₁, e₂ in both tasks
- Performance measure p is realized (y could be realized later)
- A is paid
What are remedies for problems of imperfect performance measurement?
a) Additional performance measures
b) Job design
What are advantages and disadvantages of additional performance measurements?
- Additional performance measures can mitigate or even resolve multitasking problems.
- However, collecting performance measures is costly.
Give examples of jobs & imperfect performance measures?
- Production worker: quantity + quality
- Performance measures: quantity + some form of quality control
- Sales agent: selling the product + customer service
- Performance measures: sales + customer satisfaction
- Professor: research + teaching
- Performance measures: Publication record + student evaluations
How can job design help to solve imperfect performance measurement problems?
- Assume P can hire two identical agents k = I, II and separate tasks
- Agent I (II) performs task 1 (2); only one performance measure p
Assigning the tasks to different agents allows the principal to individually adapt incentives for each task, which solves the multitasking problem.
What are potential pitfalls to separating tasks in response to imperfect performance measurement problems?
- Not all tasks can be separated (e.g., quantity and quality in production)
- Hiring additional agents is costly (e.g., positive reservation utilities, minimum wages)
- If tasks are complements, separating tasks increases effort costs (e.g., professor: research and teaching)
- Fairness concerns in case agents have social preferences (e.g., inequality aversion)
- Coordination problems
Why do firms use relational contracts?
Main advantage: relational contracts can incorporate non‐verifiable performance measures that better reflect the worker’s contribution to the firm (e.g., y in our model)
What are relational contracs?
- agreement that cannot be enforced by a court because it is based on non‐verifiable variables
- promises and understandings that cannot be formalized by legal documents, e.g., promotion or salary increase for a “job well done”
- only works when the parties find it in their own interests to adhere to the contract
- E.g., reputational concerns in a repeated relationship can make a relational contract self‐enforcing.
What is the main change in modeling a relational contract in an imperfect performance measurement setting?
- that principal and agent can observe y, but no third party can (y is still nonverifiable)
- With probability y ∈ (0, 1) , P and A work together in the following period -> potentially repeated employment relationship
Why are relational contracts self-enforcing with repeated employment?
- If the firm reneges on its current promises, it may find that its employees expect it to renege on future promises as well.
- Employees are unwilling to exert extra effort in the future.
- If the firm reneges on its promises it profits in the short term, but its damaged reputation will cost it in the long term.
- The relational contract is self‐enforcing if the (expected/discounted) long‐term losses exceed the short term profits.
What are examples for relational contracts based on subjective performance evaluation?
- 360‐degree peer reviews: an employee’s supervisor, coworkers, and subordinates provide information on that employee’s performance
-
Management by objective systems: employee and supervisor construct a set of (non‐verifiable) goals for the employee
- at the end of some specified period, the two meet to review the employee’s performance on those goals
- Supervisor can take into account (non‐verifiable) factors that may have made it easy or difficult to reach the goals
-
Merit rating systems: employees are given numerical scores
- Based on their subjective evaluations, supervisors are required to allocate a fixed pool of points among employees