Week 9 - Subjective performance evaluation

Question 1

Main point about subjective performance evaluation

Answer 1

Instead of address uncertainty ex ante, subjectivity addresses it EX POST by changing the EX ANTE incentive arrangement
ie. can make necessary adjustments after the performance period from the info gathered during the performance evaluation period; can now estimate its impact relatively reliably
- subjectively OVERRIDE the CONTRACT agreed at the start

• recall that the issue is that there is a lot of UNCERTAINTY - unanticipated events that are uncontrollable by employees - that could happen in between CREATING the incentive system and the COMPENSATION payments

Question 2

Why are Promotion decisions IDEALLY subjective?
esp. for jobs that require different skills than current job

Answer 2

1. promotions often play an “INCENTIVE” role, ie. work hard to get promoted
   - measures need to be informative about EFFORT
2. but… they also play a “SORTING” role, ie. the firm wishes to assign individuals to jobs based on (future) ability
   - measures need to be informative about (future) ABILITY
3. otherwise, firms encounter the “Peter Principle” (Peter and Hull, 1969)
   » at some point of promotion, employees are actually INCOMPETENT at their promoted if not evaluated subjectively
   - The best worker is not always the best candidate for manager. (Benson et al., 2019)

• the more DIFFERENT the next job, the less important CURRENT performance is, and the more important ABILITY is, ie. the more you should rely on subjectivity

Question 3

Discretionary bonuses/penalties vs Subjective weights
(Höppe and Moers)

Answer 3

Discretionary bonuses/penalties
- ie. ex ante option to EX POST OVERRIDE formula-based metrics
- used when there is NOISE in the performance METRIC
(eg. Covid-19 distorts measured performance)
» noise = variance that is not controllable by employees’ actions
-> REDUCES RISK in CONTRACTS b/c know that pay can adjusted ex post

Subjective weighting
- ie. the ex ante absence (at the beginning of performance evaluation period) of EXPLICIT weights
- used when there is ENVIRONMENTAL UNPREDICTABILITY
(eg. we don’t exactly know where to focus ex ante)
-> ADDRESSES PROBLEMS of NONCONGRUITY: you can change weights along the way and steer effort in the most congruent direction

^make the weights more vague and flexible in the beginning; not stuck by the incentive system designed in the beg. & allows firm to make adjustments to the weights to be most optimal for the org.
-> allow more congruent behaviour / more aligned actions to org’s objectives

Question 4

5 costs of subjectivity

Answer 4

1. EXCUSE CULTURE/politicking
   eg. employees complaining for more bonus when sth bad happens (but not when sth good happens)
2. Fairness concerns, eg. FAVOURITISM
   » even when ppl are aware, biases still happen
3. VAGUE EXPECTATIONS/directions
   » will probably cause more costs than benefits for lower level employees
4. EXPENSIVE in management TIME
   eg. double checking each other’s ratings is costly
5. Performance BIASES
   ie. leniency bias & centrality bias

Question 5

2 performance evaluation Biases (Moers, 2005)

Answer 5

1. LENIENCY bias
   ie. higher performance ratings {skewed to the right}
   » managers tend to prefer giving higher ratings b/c it looks good on themselves
2. CENTRALITY bias
   ie. compressed performance ratings
   » ratings have similar distributions

Both are problematic: less useful ratings -> personnel and future incentives issues
ie. if everyone will get the same rating regardless of performance, why bother to work hard in the first place?

Question 6

3 ways to address the performance evaluation Biases

Answer 6

1. Avoid unnecessary subjectivity
   eg. the “overall aggregation” subjectivity in the Johansen’s case
2. FORCED RANKING systems
   » to force more normal distributions, but also have problems
3. Have an INDEPENDENT party ANALYSE RATINGS -> “calibration committee” (used by 54% of firms)
   - typically, other teams’ or higher-level managers
   eg. make sure that ratings across diff. managers are similar

Question 7

Calibration committees & the PE biases

Answer 7

• decreases leniency bias, but increases centrality bias
  » doesn’t solve both problems simultaneously so need to know which is more problematic
  » probably centrality bias b/c cannot distinguish who is high or low performer
  » leniency bias is less problematic as long as can still distinguish employees, although all ratings are probably INFLATED

Question 8

Calibration committees - Who gets adjusted?

Answer 8

• above (below) average, more (less) likely to get DOWNWARD ADJUSTMENT
• The “CLOSER” the CC is from the employees being evaluated, the MORE LIKELY to make adjustments!
  » b/c they know the employees better
  » thus, ideally should put CC who are closer to the employees so that they know them better and feel more comfortable making adjustments

> > best suggestion is to rotate the managers in the CC

Question 9

Calibration committees can help to address biases through EXPLICIT adjustments.
What are 2 ways that CC can also provide IMPLICIT INCENTIVES for evaluator to provide UNBIASED ratings?
ie. can change the behaviour of the evaluator

Answer 9

1. If the evaluator shows a LACK OF SKILLS,
   - they are less likely to get promoted
   - thus, incentive to IMPROVE SKILLS, which leads to better ratings OVER TIME
2. If the evaluator shows OPPORTUNISM,
   - CC will lower evaluator’s own rating
   ie. lower compensation and less likely to get promoted
   - thus, incentive to DECREASE BIAS, which leads to better ratings OVER TIME

Question 10

Vitality Health case
In the old performance evaluation system, managers preferred to assign UNIFORM RATINGS, which did not reflect “true” performance.
^root cause of the problem with Vitality’s old performance measurement system
1. What were the consequences?
2. Why did managers do so (Agency problem)?

Answer 10

Consequences
1. performance system did NOT IDENTIFY top/poor performers
2. over time, top performers will LEAVE the firm and it will be difficult to attract new talent

Reasons for assigning uniform ratings / Agency problem
1. careful reviews take TIME that can be spent on better-rewarding activities
2. managers do not want to face SOCIO-EMOTIONAL COSTS of differentiating among employees {don’t want to hurt their feelings}

Question 11

Vitality Health case
Forced distribution ranking systems are a natural but IMPERFECT solution (doesn’t solve all problems).
What are 3 suggestions for improvements?

Answer 11

1. use FLEXIBLE rating distributions
   ie. in GOOD TIMES of economy, can consider increasing the % of high performers; decrease in bad times
2. ex post differences in rating distributions depending on division performance
   » make adjustments between teams, eg. no one with Top rating if whole team performed poorly
3. change managers’ rating preferences/incentives
   - by using CALIBRATION COMMITTEES to “discipline” evaluators, eg. lowering their bonuses

Question 12

Vitality Health case
3 main takeaways

Answer 12

1. Rewarding top performers and punishing poor performers
   requires that BOTH TYPES of performers be identified
   - for many employees, performance measurement depends on subjective ratings by direct managers
2. AGENCY PROBLEM: managers prefer to assign UNIFORM RATINGS
   (see previous flashcard for 2 possible causes)
3. FORCED RANKING systems are a natural, but IMPERFECT, solution
   - b/c creates a sense of COMPETITION, ie. could encourage less teamwork and only work on rewarded activities
   - managers still play “rating games”, eg. some don’t distinguish, some rotate ratings
   - WITHIN a team, managers have to distinguish too (even if the team did really well -> between-team “inconsistencies”)
   » eg. top performer of one team may be worse than low performer of another team