Trends in Corporate Governance

1
Trends in Corporate Governance

• Benjamin E. Hermalin
• UC Berkeley

2
What trends?

• In US, last twenty-five years have seen
  significant shift toward more outsider
  representation on the board.
• In US, trend toward more external hiring of CEO.
• Similar trends emerging in UK.
• In US, trend toward greater CEO compensation
  (both contingent and non-contingent).
• In US, trend toward shorter CEO tenures
• In US, renewed efforts at reform SOx, NYSE, etc.

3
Questions?

• How do these trends relate?
• What do they tell us about organization?

4
Thinking about governance

• The role of directors
• Hire a CEO.
• Monitor him
• (make assessments).
• Replace him if necessary.

5
Model Timing
Board hires new CEO. Internal (I) or External (E)
6
Model Timing
Board hires new CEO. Internal (I) or External (E)
Board monitors with intensity p that is,
acquires signal, y, about CEOs ability with
probability p.
7
Model Timing
Board hires new CEO. Internal (I) or External (E)
If signal acquired, Board makes decision to keep
or fire incumbent CEO.
Board monitors with intensity p that is,
acquires signal, y, about CEOs ability with
probability p.
8
Model Timing
Board hires new CEO. Internal (I) or External (E)
If signal acquired, Board makes decision to keep
or fire incumbent CEO.
Board monitors with intensity p that is,
acquires signal, y, about CEOs ability with
probability p.
Earnings, x, realized.
9
Preferences and ability

• Earnings, x, are distributed normally with a mean
  equal to the ability, ?, of the CEO in place at
  the end (i.e., the initial hire or his
  replacement).
• Board likes x, but dislikes monitoring effort, p.
• Assume behavior of board can be aggregated to
  that of a single decision maker with utility
  function ?x c(p), where c(?) has usual cost
  function properties and ? is a parameter that
  reflects diligence.

10
Informational assumptions

• CEOs ability, ?, is fixed throughout his career.
  It is unknown, ex ante, by anyone, but it is
  common knowledge that ? is the draw from a normal
  distribution with mean ? and precision ?. Recall
  precision 1/variance
• The signal, y, which board receives with
  probability p, is distributed normally with mean
  ? and precision s.

11
Thinking about incumbent ability
distribution of true ability.
value (ability)
12
Thinking about incumbent ability
expected ability ?
distribution of true ability
value (ability)
13
Thinking about incumbent ability
expected ability ?
expected ability of replacement (which
is normalized to 0)
distribution of true ability
value (ability)
14
Thinking about incumbent ability

• So, absent new information, want to keep original
  CEO (his expected value greater than
  replacements)

15
Thinking about monitoring
distribution of true ability
expected ability
expected ability of replacement
bad signal
good signal
replace incumbent
keep incumbent
16
Thinking about monitoring
highly likely
distribution of true ability
not so likely
expected ability
expected ability of replacement
bad signal
good signal
replace incumbent
keep incumbent
17
Benefit of monitoring
expected ability
expected ability of replacement
distribution of true ability
value (ability)
on average, get rid of low ability CEOs
on average, keep high ability CEOs
18
Analysis

• Proposition 1 The intensity with which the board
  monitors the CEO is
• decreasing with the prior estimate of his
  ability, ?
• decreasing with the precision of the prior
  estimate, ? but
• Increasing with the boards diligence, ?.

19
Who gets monitored?
estimated ability of replacement
value (ability)
more value to monitoring red CEO than green CEO.
20
Who gets monitored?
estimated ability of replacement
value (ability)
more value to monitoring red CEO than green CEO.
21
Monitoring and who to hire
ability external candidate
ability internal candidate
value (ability)
22
Monitoring and who to hire
ability external candidate
ability internal candidate
estimated ability of replacement
value (ability)
monitoring means largely avoid these values
monitoring means largely keep these values
23
Monitoring and who to hire

• Monitoring means willing to trade a higher
  estimated ability for greater uncertainty about
  ability.
• External candidates have an edge.
• But note This result relies on the assumption
  that the CEO will be monitored
• Lower the probability of getting signal of
  ability (i.e., less intensely CEO monitored),
  less willing to make this tradeoff.
• Boards who are more inclined to monitor will have
  a greater tendency to hire external candidates.
• See Proposition 2 for a formal statement of these
  results.

24
Two trends related

• Outside directors are generally thought to be
  more inclined to monitor
• Theoretical reasons (e.g., inside directors too
  closely tied to incumbent manager)
• Anecdotal/field work evidence (e.g., Mace)
• Statistical evidence (e.g., Weisbach).
• So a trend toward greater outsider representation
  on boards should lead to more external candidates
  being hired.

25
CEO tenure

• Recall No monitoring ? Always keep incumbent
  CEO.
• Monitoring ? some CEOs get fired.
• Hence, more monitoring ? shorter CEO tenures on
  average.
• So, more outsider representation ? more
  monitoring of all CEOs ? shorter CEO tenures on
  average.
• Also indirect effect

26
External CEOs more vulnerable
ability external candidate
ability internal candidate
likely to draw bad signal and get fired
estimated ability of replacement
bigger left tail also means greater reason to
monitor
value (ability)
likely to draw bad signal and get fired
27
Effort Model New timing
If signal acquired, Board makes decision to keep
or fire incumbent CEO.
Board hires new CEO. Internal (I) or External (E)
CEO expends effort, e, at cost k(e).
Board monitors with intensity p that is,
acquires signal, ye, about CEOs ability with
probability p.
Earnings, x?(e), realized.
Surviving CEO gets benefit, b gt 0.
28
Effort
distribution of signal
signal
29
Effort
effort shifts the signal to the right, making CEO
seem better if monitored
signal
30
Effort
effort shifts the signal to the right, making CEO
seem better if monitored
signal
But in equilibrium boards not fooled it
subtracts back expected effort when inferring
ability
31
Effort
So even though not fooling anyone, CEO has to
expend effort or look even worse!
signal
But in equilibrium boards not fooled it
subtracts back expected effort when inferring
ability
32
Effort

• Proposition 5 Assume for the relevant parameter
  values that the game with CEO effort has a
  pure-strategy equilibrium. Then the following
  comparative statics hold
• the lower the CEOs estimated ability, the more
  effort he expends in equilibrium (Avis) and
• the more diligent is the board (i.e., the greater
  is ?), the more effort the CEO expends in
  equilibrium.

33
Result ii.

• Result ii. predicts that greater board diligence
  leads to more effort from CEO.
• Might seem a no brainer
• but recall no monitoring of effort per se.
• The greater effort is induced indirectly because
  the CEO is trying to look more able.
• His efforts are for naught in equilibrium, but
  must still work harder.

34
Effort and compensation

• Proposition 6 If CEOs with similar attributes
  enjoy equal expected utility in the equilibrium
  of the CEO labor market, then, controlling for
  attributes, CEOs who work for more diligent
  boards will receive greater compensation than
  CEOs who work for less diligent boards.
• Even in the model without effort, working for a
  more diligent board is less desirable than
  working for a less diligent board.
• Higher compensation as compensating differential.

35
Time series and cross section

• Last predictions might seem at odds with
  predictions of some that it is less diligent
  boards who pay more.
• This cross-section prediction can be reconciled
  with the time-series prediction if CEOs are
  heterogeneous
• monitoring and ability are substitutes, so less
  diligent boards have greater demand for ability
  ceteris paribus
• more able CEOs can demand salary premia over less
  able CEOs ceteris paribus
• hence, in cross section, higher paid-higher
  ability CEOs can work for less diligent boards
  while lower paid-lower ability CEOs work for more
  diligent boards.
• However, over time, as boards on average become
  more diligent, the trend should be toward an
  increase in CEO compensation

36
Putting all the trends together
more effort
more independent boards (more outsiders)
more monitoring
more external CEOs
greater compensation
shorter average tenures
37
Conclusions

• Many of the trends weve been observing in
  corporate governance can be linked via the
  boards monitoring role.
• Some of the good trends (e.g., more independent
  boards) may yield bad trends (e.g., greater CEO
  compensation).
• From the perspective of theory, work remains.