VII. Pay for Performance

1
VII. Pay for Performance
2
Why not pay by output?

• 1st principle of optimal insurance the party who
  can more easily absorb risk (ie the party that is
  most risk neutral) should insure the other party.
  
• employees are likely to be more risk averse than
  the firm. The firm should not avoid risks if
  this leads to lower profits. Stockholders can
  diversity risk across investments. They want the
  firm to earn the highest profit possible.
  Therefore the firm should insure the employees
• the firm can pay lower wages if it accepts the
  risk
  
• firms that pay by output will need to pay higher
  compensation. If they do not get greater
  productivity, it is a waste
  

3
Why not pay by output?

• Sometimes output is difficult to measure
• it is possible that variations in output are due
  to external factors that are not controllable by
  the employee. Unless the firm can condition on
  these factors (ie use RELATIVE measures), it may
  be difficult to give the proper incentives.
• If the work of the employee depends upon
  coordination with other employees, it may be
  difficult to single out their individual
  contribution. It may be necessary to pay on a
  team basis.
• There may be multiple tasks. Incentives must be
  balanced across different tasks. If it is
  difficult to measure the performance on one task,
  strong incentives on other tasks will result in
  the employee ignoring that task. There can be a
  severe problem of distortion of incentives.

4
Why not pay by output?

• It may be difficult to set the rate.
• Workers have incentives to slow down when you are
  attempting to set the appropriate rate.
  
• Management has the incentive to increase the
  required rate when the workers easily surpass
  their old levels of productivity. This is called
  the ratchet effect. Since workers realize that
  the ratchet effect exists, they will be reluctant
  to work to their full capacities or they may try
  to force other workers to work at less than full
  capacity.

5
Why not pay by output?

• It may be costly to monitor output
• Paying by output can reduce the quality of
  output.
  
• It is possible that workers try to sabotage
  others. There is no incentive to cooperate.
  
• Changing the pay mix will typically result in an
  increase in turnover.

6
Why choose output based pay?

• Positive influence on hiring and retention
• also known as adverse selection
• low ability workers will look for jobs that pay
  by input rather than output
  
• workers who do not want to work hard are likely
  to leave the firm. This is good.

If you can sell 5, you go to the company with
output based pay. If
pay
7
Why choose output based pay?

• Motivates current workers to work hard.
• also known as moral hazard or shirking
• Shirking does not imply that workers are bad
  persons.
  
• If monetary incentives are not given, then
  something else must be put in their place
  
• This is not just a result of agency theory, but
  could be predicted by expectancy theory or
  reinforcement theory.

8
1. How Strong Should Incentives Be?

• Our goal motivate employees to act like owners
  maximize firm value
  
• Lets use the most common form of pay for
  performance (linear) Pay a bPM
  
• a base salary b commission rate
• PM performance measure
• We want to think about how to set a b

9
1a. Level v. Shape of Pay
Pay

• If the employee works harder, what happens to
  pay?
  
• dPay/dPM b slope
• does not depend on the intercept
• the employee has greater incentives if
• effort has greater effect on productivity
• the commission rate b is higher
• Incentive is driven primarily by b, not by base
  salary a
  
• since our focus is on incentives, we care more
  about what determines b

a2 bPM
a1 bPM
PM
10
Implications

• b is often referred to as the incentive
  intensity
  
• larger b Þ stronger incentives
• we discuss more elaborate pay-performance
  shapes below
  
• a overall level of pay has less effect on
  incentives
  
• depends on labor market value of employees
  skills
  
• also must compensate for employees effort cost
  C(Sei), riskiness of pay
  
• Use a 2-stage process in designing pay plans
• 1. focus on getting incentives right
  performance evaluation b
  
• 2. adjust overall level of pay a to attract
  retain appropriate talent

11
1b. Marginal Costs Benefits of Effort

• Suppose the employee provides a little more
  effort on some dimension of performance
  
• This imposes a cost on the employee, DC/Dei
• the firm has the same marginal costs of extra
  effort as the employee
  
• must compensate the employee for working harder,
  so the firms costs must go up by the at least
  the same amount
  
• similarly, any increase in pay risk to the
  employee must be compensated
  
• They do not necessarily have the same marginal
  benefits of extra effort, however
  
• employees benefit is (DPay/DPM)(DPM/De)
  b(DPM/De)
  
• firms is DQ/De
• if these are not equal, we have a conflict of
  interest imperfect incentives

12
Perfect Incentives

• We concluded last lecture that incentive problems
  arise b/c of conflict of interest (different
  benefits) between the worker firm
  
• what incentive intensity b gives the employee the
  same interests as the ownermaximize profits?
  
• e.g., consider giving employee either 25 or 50
  of profits they create
  
• which will give stronger incentives? what is the
  limit of this argument?
  
• No conflict of interest only if b is set so that
  b(DPM/De) DQ/De
  
• e.g., if PM Q?, b must equal 1
• more generally, b must rescale the PM so that the
  employee gets 100 of incremental profits he or
  she creates
  
• for linear pay, b Q/PM
• if not, incentives are weak compared to
  ownership
  
• But if b Q/PM, pay aQ, firm profits from
  the employee a !

13
Selling the Job

• Our solution amounts to selling the job to the
  employee
  
• profit Q a bPM a
• firm profits 0 only if a
• all marginal profits go to the employee, who pays
  the firm for right to earn them
  
• Examples
• cab drivers
• seats on exchanges
• Berghof
• outsourced sales
• other?
• Many jobs have sell the job aspects to them

Pay
a bPM
b
PM
a
14
Practical Incentives

• Even CEOs ( most other employees), arent
  usually rewarded with 100 of the profits they
  create. Why not?
  
• upcoming reading Jensen Murphy on CEO pay
• Suppose we cannot perfectly measure performance
  PM Q e
  
• pay a b(Qe) a bQ be
• spay bse
• pay is risky if there is measurement error the
  stronger the incentive, the greater the risk
  
• Implications
• pay risk premium thru higher total pay,
• / or incur costs to evaluate performance more
  accurately,
  
• / or give weaker incentives
• Because of measurement error, incentives are
  imperfect (weaker than ownership) in essentially
  all jobs

15
1c. Multitask Incentives

• Suppose the employees job has two dimensions
• e.g., quantity quality
• suppose performance on one dimension (quantity)
  can be measured accurately, but the other
  (quality) cannot
  
• The theory just developed implies that we should
  put a strong incentive on quantity, a weak one
  on quality
  
• but that would distort incentives
• to reduce distortions, we need to balance
  incentives across different tasks
  
• our discussion of narrow performance measures was
  a special case with a narrow measure, some tasks
  are given zero weight
  
• Where balance is difficult b/c accuracy of
  measures on different tasks is very different,
  explicit incentives are often set to zero
  
• implicit incentives based on subjective
  evaluations are used instead

16
1d. SummaryWhen to Give Stronger Incentives

• Sorting for talent is more important
• Effort is more costly has greater impact
• dC/de larger
• dQ/de larger ( Q more valuable in the market)
• Risk aversion is lower
• Incentives are well balanced across important
  tasks
  
• Performance evaluation is more effective
• measurement error s² is lower measurement cost
  is lower
  
• performance measure distorts less can be
  manipulated less
  
• subjective evaluator has better judgment, is
  more trusted

17
2. More Elaborate Pay-Performance SIhapes

• Generalization of incentive intensity b
• These shapes are often encountered in practice
• why / when might you use each?
• what problems might you encounter?
• Identify what they graphs might represent

18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
Summary PayPerformance Shape

• Upside potential / downside risk from employee
  actions Þ reward or punishment shape
  
• Simplicity is a virtue
• All jobs have implicit incentives
• promotions are most important for most firms,
  middle managers

23
Targets

• Floors reduce employees downside risk
• can facilitate stronger incentives
• change incentives for risk taking
• ex. employee stock options
• Caps may reduce gaming, but can hurt recruitment
  / retention
  
• Targets often motivate gaming if performance is
  near the target
  
• Goldbricking

24
The Ratchet Effect

• If an employee earns high pay from an incentive,
  the firm may be tempted to change the plan to
  lower pay
  
• But doing so may reduce future incentives
• b/c of expectation that good performance today Þ
  lower rewards in future
  
• this is often called the Ratchet Effect, is
  something to avoid
  
• Quota Restriction
• Two earlier course concepts help understand the
  issue
  
• the Hold-up problem temptation to change the
  rules ex post
  
• implicit contracting credibility / reputation

25
3. Implementation of Pay for Performance

• Expect turnover
• Pay for performance ( luck)
• Transition gradually
• convert raises to bonuses
• provide initial insurance against downside risk,
  then phase out
  
• start with test cases role models
• Communicate listen
• Clearly reserve rights to adjust system over
  time
  
• Foster an appropriate culture

26
Real World Examples

• Examples
• REMAX - real estate agents pay 20,000 per year
  to REMAX for advertising, etc. They keep all the
  proceeds of the sales themselves
  
• Hairdressers - shop often sells spots.
  Individual keeps all sales.
  
• Taxicabs - often pay a flat fee per day to rent
  the cab
  
• Securities traders Buy seat for 100,000s
• Waiters or waitresses
• Anywhere with a small base salary (opportunity
  cost)
  
• Salespersons with quota

27
4. Economic Ideas

• Selling the job as an intuitive approach to moral
  hazard (agency) problems
  
• Tradeoff of risk v. incentives
• Incentive effects of shape v. level of pay
• Ratchet Effect

28
Other Points

• Selling the job is a natural implication of our
  view of organizational design as an internal
  market
  
• similar to our logic on GHC screening
• To set targets effectively reduce gaming
• invest in effective implicit contracting
• use subjective performance evaluations