Merck & Co., Inc. (A, B, C)
Teaching Note*
I. Background
1. Merck & Co., Inc. (A,B,C) was presented as the first case in a four-session module focusing on
performance measurement and appraisal. The Simon School does not have a required HRM
course, and this case marked the introduction of many concepts and institutions, including
the Hay point system, performance measurement, forced-distribution systems, and relative
performance evaluation.
2. The students are well versed in agency theory and the importance of incentives in
organizations, but have given little (if any) thought to the performance measurement and
appraisal system that drives any incentive compensation system. There are several concepts I
want students to have at the conclusion of the case:
a. “Rewarding top performers and punishing poor performers” requires that performers be
identified.
Key to any successful incentive compensation program is the system that defines performance
and identifies high and low performers. Divisional or corporate-level profitability may serve
as suitable performance measures for divisional managers or top-level executives. But, for the
vast majority of lower- and middle-level managers, performance measurement depends on
subjective performance ratings assigned by direct superiors in the corporate hierarchy. The
success of company-wide incentive compensation programs therefore depends on the
effectiveness of managers in monitoring and appraising the performances of their
subordinates.
b. There is an important agency problem between the supervisor/manager and the owners
of the firm: for a variety of reasons, managers prefer to assign uniform ratings to
employees regardless of performance.
Self-interested managers have little incentive to invest in performance evaluation, both
because careful appraisals take time away from better-rewarded activities, and because
managers face large nonpecuniary costs from disgruntled employees with mediocre or low
evaluations. The agency costs associated with managerial preferences to assign uniform
1
, 491-008 Merck & Co., Inc. (A, B, C)
ratings are large, since uninformative performance appraisals preclude writing effective
incentive contracts with other workers in the organization. Indeed, the difficulty in getting
managers to reveal what they know about subordinate performance may explain the relative
paucity of successful incentive compensation systems in large organizations.
c. Forced-distribution ranking systems are a natural, but not perfect, way of countering
managerial preferences to assign uniform ratings.
Forced-distribution rating systems—in which managers are forced to adhere to a given
distribution of performance ratings—mitigate managerial tendencies to assign uniform
ratings but may generate important counter-productive side effects. The side effects are
reduced by increasing the size of the reference group and by allowing the forced distribution to
vary depending on group performance.
d. Does the cost of structuring an effective incentive compensation outweigh the benefits?
This is, in most cases, the students’ first exposure to design and implementation issues.
Changing to the new system creates losers as well as winners. Employee “buy-in” is
important: they must understand that the change is not just a transfer of wealth from
employees to employers. Many employees are unhappy under the new system, but (as it turns
out) less so than were unhappy under the old system.
3. The Merck forced-distribution rating system also gave me the opportunity to introduce and
discuss the costs and benefits of relative performance evaluation, which would become more
important later in the course (specifically in conjunction with executive compensation).
4. The Merck discussion takes approximately one and a half 80-minute class sessions. In the
remainder of the second session, we discussed the American Cyanamid case, which involved
a firm switching away from a forced-distribution system. (Gellerman, Saul W. and William G.
Hodgson, “Cyanamid’s New Take on Performance Appraisal,” Harvard Business Review,
(May-June 1988): 36-41.) The parallels and inconsistencies of Merck vs. Cyanamid are
discussed below in the teaching plan.
II. Sketch of the Basic Theory
1. Managerial tendencies to assign uniform ratings.
a. Exhibit A2 shows the distribution of performance ratings assigned by Merck supervisors
and managers in 1985. Most of the 6,734 employees (93%) received ratings from 3 to 4+;
only 96 employees (1.4%) received 5 or 5- ratings, and only eight received the lowest
ratings of 1+ or 1. Moreover, the company-wide distribution depicted in Exhibit A2
overstates the dispersion within divisions: some divisions would assign uniformly higher
ratings than others, making the overall distribution substantially more dispersed than
divisional distributions. An underlying hypothesis in this teaching note is that the
tightness of the ratings distributions in part reflects ineffective performance appraisal by
self-interested managers and supervisors.
b. Performance appraisals are typically conducted by the managers who work closely with
the employees on a day-to-day basis, since these managers have the specific knowledge
required to evaluate their subordinate’s performance. The time and effort line managers
allocate to performance appraisal depends in part on the rewards they receive for
conducting appraisals relative to their rewards from engaging in other activities. Since
performance appraisals are often a required but unrewarded managerial task, it is
rational for managers to spend no more than the minimum acceptable time and effort in
evaluating subordinate performance.
c. Moreover, managers bear a disproportionately large share of the non-pecuniary costs
associated with performance appraisal, which leads them to prefer to assign uniform
ratings rather than to carefully distinguish employees on the basis of their performance.
2
Teaching Note*
I. Background
1. Merck & Co., Inc. (A,B,C) was presented as the first case in a four-session module focusing on
performance measurement and appraisal. The Simon School does not have a required HRM
course, and this case marked the introduction of many concepts and institutions, including
the Hay point system, performance measurement, forced-distribution systems, and relative
performance evaluation.
2. The students are well versed in agency theory and the importance of incentives in
organizations, but have given little (if any) thought to the performance measurement and
appraisal system that drives any incentive compensation system. There are several concepts I
want students to have at the conclusion of the case:
a. “Rewarding top performers and punishing poor performers” requires that performers be
identified.
Key to any successful incentive compensation program is the system that defines performance
and identifies high and low performers. Divisional or corporate-level profitability may serve
as suitable performance measures for divisional managers or top-level executives. But, for the
vast majority of lower- and middle-level managers, performance measurement depends on
subjective performance ratings assigned by direct superiors in the corporate hierarchy. The
success of company-wide incentive compensation programs therefore depends on the
effectiveness of managers in monitoring and appraising the performances of their
subordinates.
b. There is an important agency problem between the supervisor/manager and the owners
of the firm: for a variety of reasons, managers prefer to assign uniform ratings to
employees regardless of performance.
Self-interested managers have little incentive to invest in performance evaluation, both
because careful appraisals take time away from better-rewarded activities, and because
managers face large nonpecuniary costs from disgruntled employees with mediocre or low
evaluations. The agency costs associated with managerial preferences to assign uniform
1
, 491-008 Merck & Co., Inc. (A, B, C)
ratings are large, since uninformative performance appraisals preclude writing effective
incentive contracts with other workers in the organization. Indeed, the difficulty in getting
managers to reveal what they know about subordinate performance may explain the relative
paucity of successful incentive compensation systems in large organizations.
c. Forced-distribution ranking systems are a natural, but not perfect, way of countering
managerial preferences to assign uniform ratings.
Forced-distribution rating systems—in which managers are forced to adhere to a given
distribution of performance ratings—mitigate managerial tendencies to assign uniform
ratings but may generate important counter-productive side effects. The side effects are
reduced by increasing the size of the reference group and by allowing the forced distribution to
vary depending on group performance.
d. Does the cost of structuring an effective incentive compensation outweigh the benefits?
This is, in most cases, the students’ first exposure to design and implementation issues.
Changing to the new system creates losers as well as winners. Employee “buy-in” is
important: they must understand that the change is not just a transfer of wealth from
employees to employers. Many employees are unhappy under the new system, but (as it turns
out) less so than were unhappy under the old system.
3. The Merck forced-distribution rating system also gave me the opportunity to introduce and
discuss the costs and benefits of relative performance evaluation, which would become more
important later in the course (specifically in conjunction with executive compensation).
4. The Merck discussion takes approximately one and a half 80-minute class sessions. In the
remainder of the second session, we discussed the American Cyanamid case, which involved
a firm switching away from a forced-distribution system. (Gellerman, Saul W. and William G.
Hodgson, “Cyanamid’s New Take on Performance Appraisal,” Harvard Business Review,
(May-June 1988): 36-41.) The parallels and inconsistencies of Merck vs. Cyanamid are
discussed below in the teaching plan.
II. Sketch of the Basic Theory
1. Managerial tendencies to assign uniform ratings.
a. Exhibit A2 shows the distribution of performance ratings assigned by Merck supervisors
and managers in 1985. Most of the 6,734 employees (93%) received ratings from 3 to 4+;
only 96 employees (1.4%) received 5 or 5- ratings, and only eight received the lowest
ratings of 1+ or 1. Moreover, the company-wide distribution depicted in Exhibit A2
overstates the dispersion within divisions: some divisions would assign uniformly higher
ratings than others, making the overall distribution substantially more dispersed than
divisional distributions. An underlying hypothesis in this teaching note is that the
tightness of the ratings distributions in part reflects ineffective performance appraisal by
self-interested managers and supervisors.
b. Performance appraisals are typically conducted by the managers who work closely with
the employees on a day-to-day basis, since these managers have the specific knowledge
required to evaluate their subordinate’s performance. The time and effort line managers
allocate to performance appraisal depends in part on the rewards they receive for
conducting appraisals relative to their rewards from engaging in other activities. Since
performance appraisals are often a required but unrewarded managerial task, it is
rational for managers to spend no more than the minimum acceptable time and effort in
evaluating subordinate performance.
c. Moreover, managers bear a disproportionately large share of the non-pecuniary costs
associated with performance appraisal, which leads them to prefer to assign uniform
ratings rather than to carefully distinguish employees on the basis of their performance.
2
