1
Business 33032-01/81/85Canice Prendergast
Managing the Workplace Spring 2010
Topic 8
Executive Compensation
Here we consider executive compensation, both because it has received a lot of attention recently and because it is a useful way of drawing together some of the material that we addressed in earlier lectures.
I deal with this topic in the following steps:
- The level of CEO compensation
- Possible mechanisms of corporate control
- How sensitive are executive rewards to performance?
- Do these compensation plans work?
- Are they designed efficiently?
Why are CEOs given much more explicit incentives than other workers in organizations?
I: The Level of Pay
Let us now consider how CEOs are paid in the US compared to other countries. The way we do this is to consider how much higher the wage for a CEO (in large companies: $30 billion in sales or more) is compared to that of the average worker’s wage.
How much higher is the average CEOs wage in Japan?
In France and Germany?
In the US?
The differences for small companies are smaller; here European or Japanese CEOs earn about a third to two thirds as much as an American CEO.
Where does this difference principally arise? Most actually comes in the form of stock payments or options, which are much rarer in Europe and Japan.
How is Pay Set?
This is part of the reason that we focus on executives when we consider explicit incentive schemes.
Before we actually begin talking about the extent to which CEOs are given incentives to exert effort, let us consider how compensation is actually determined. Typically, the package is determined by the Compensation Committee, which is made up of the outside members of the board of directors. The CEO is also a member of the Compensation Committee but has to leave the room for the time that the CEO’s compensation is determined.
Who typically make up the outside members of the board?
CEO Compensation
Because promotion no longer acts as an incentive device, companies use explicit incentives for their CEOs. Before we get to talking more about incentive plans, let me clarify one point. Suppose that a compensation consultant does a regression, seeing how CEOs are paid according to changes in stock price, level of earnings, and sales. Typically the most important explanatory variable is sales.
Some people have interpreted this to mean that CEOs are given the wrong incentives. They should be given incentives to increase stock prices (it is the shareholders’ money he is using after all) rather than empire build, which would be a problem if you pay people on the basis of sales. Is this conclusion right?
Hence we must be careful in inferring compensation plans from regression results. This point gives rise to what is known as the magnification or cloning effect.
Magnification Effect
1
Consider a typical hierarchy, which is a game tree with a series of nodes. The longer the chain of the hierarchy, the more magnified are the decisions made by a CEO. This implies that the larger the chain of the hierarchy, the more magnified are his decisions. As a result, it becomes particularly important that individuals with the most talent are allocated to the places where their decisions have most effect. This implies that the largest companies (those with greatest sales) will be controlled by the most able individuals. Consequently, independent of any incentive effects we would expect companies with the greatest sales to have the best-paid CEOs. This is probably what lies behind the regression results that I described above.
The typical result you get from these regressions is that if sales are 1% higher, compensation of the CEO increases by about .3%.
Incentives
We now turn to the issue of providing incentives to CEOs. One important point to remember here is that when we talk about giving incentives to CEOs to exert effort, we are not talking about the same kinds of things as with shop floor workers or middle managers, where typical issues dealt with are people turning up late, taking long lunch breaks or chatting to co-workers when they should be working. Instead, we are talking about CEOs carrying out activities that are not in the interest of the shareholders. For example, Ford was recently sitting on about $13 billion in cash. The obvious thing to do with the money is to give it out to shareholders. However, one of the things that Ford is thinking of doing is using the cash to buy another company. There is significant skepticism that there are genuine efficiency reasons for this; instead it looks like empire building on the part of the executives at Ford. This is the kind of issue we are dealing with here.
Another example is that CEOs may not have incentives to take the right kinds of projects. Instead they may realize that the only way they get fired is if things turn out to be catastrophic; what effect will this have on the CEO’s incentives?
What compensation plans are usually implemented to overcome this problem?
However, note the problems associated with this compensation plan. What other agents had a similar plan earlier in the course?
II: Mechanisms of Corporate Control
There are a number of mechanisms for controlling CEO behavior, not all of which involve explicit compensation schemes involving stocks or options or whatever. These include
1. Reputation - this is similar to the career concerns story that we told earlier.
Why might this not work?
1
2. Monitoring by the Directors and Post-retirement Events
It might be possible to induce the CEO to act in the interests of the company by threatening not to put him on the board of directors. This threat does not appear to have much teeth, however. In particular, we find that over 80% of the CEOs end up on the board, which is pretty high given that many of the CEOs may retire due to illness. [You need to be a bit careful here, though; it could conceivably be such a good monitoring device that it never needs to punish a CEO.]
Interlocking directorships:
Sarbanes Oxley (2002):
(a)Requires executives to report sales of stocks in two days – previously ten,
(b)Changes in accounting rules to make it harder to misreport earnings,
(c)Audit committees made more independent,
(d)Increases board responsibility for misreporting.
Other changes –
(i)Board selection is now typically done by the nominating committee, not the CEO,
(ii)More equity compensation for directors.
Evidence on the role of boards for the US – look at the effects of
(a)board size, and
(b)outside directors.
Very marked difference between the available evidence and recent legal changes in say the UK.
3.Large Stakeholders
The benefits of stakeholders with more at stake must be traded off against
(a)Worse diversification,
(b)Lower liquidity,
(c)Effects on minority shareholders.
But why should institutional holdings matter?
Evidence from the United States:
Not much evidence that large stakeholders affect performance.
The “curse of liquidity” in the United States.
Despite this, voting stock typically trades for more? Why?
4. The Stock Market (Takeover Threat)
The idea here is that it may be that explicit incentive plans are not necessary as the stock market provides discipline through takeovers. Basically, the idea is that if CEOs shirk or do not act in the interests of shareholders, they are likely to find themselves taken over and replaced.
The reason why the stock market is necessary rather than simply using shareholders is due to a free-rider problem. If you look at holdings in companies, the typical holding is pretty small, as investors try to diversify their holdings. As a result, there is a free-rider problem: individual shareholders have no incentive to get involved in fixing up a company since they have little at stake individually. As a result, what is needed for monitoring is a large shareholder, and usually takes the form of an outsider buying up large chunks of stock in a takeover attempt.
What are the potential problems with using the stock market as a disciplining device?
Common Anti-takeover Defences:
(i)Supermajority Amendments
(ii)Staggered boards
(iii)Poison pills – allows management to issue more shares at a low price to existing shareholders if one owns more than x%.
Why do shareholders agree to these?
(i)Too difficult to coordinate action,
(ii)Monopoly pricing problem.
Empirical evidence on the effect of these on performance and on CEO pay.
4. Explicit Incentives
Here we address the use of explicit compensation plans where we see how sensitive a CEO’s compensation is to changes in the stock price of his company. Let us begin by describing Jensen and Murphy’s results on CEO compensation. Suppose that we begin by looking at data from a famous study by Jensen and Murphy for the 1980s. (More recent data: see Figure 1.)
The look at CEOs of Fortune 500 companies, and consider what happens if the value of the company rises by $1,000.
How much does the CEO’s salary increase by?
1
How much does the CEO’s salary + bonus + stock options rise by?
How much does the CEO’s salary, bonus, options and stock holding rise by?
III: How Sensitive is Reward to Performance?
Do you think that this provides reasonable incentives for the CEO? Jensen and Murphy clearly do not think so. Take an example of a funded chair to a university in the name of the CEO which costs $1m. How much does the CEO have to value the chair before he goes ahead with the plan?
My opinion on this example is that you cannot use incentive plans to stop small expenditures like this. Other forms of monitoring are required to hold the CEO in check on issues like this.
More recent data suggests that the sensitivity has increased, and that there is more reliance of options than before.
Now let us look at the issue a little closer:
Consider a typical Fortune 500 CEO like CSX. The CEO has an annual salary of about $700,000. His company has sales of about $7bn. Suppose that the stock price changes by 1%. How much does the CEO’s compensation change by?
Is this a large or small number?
Hard to tell, but it seems pretty large to me.
Finally, let me address the issue a slightly different way. Jensen and Murphy look at LBOs. They find that the sensitivity of the executives’ compensation is about $64 rather than the $3 we saw above. They claim that this must be better as it gives CEOs more incentives. Do you agree? What problems could $64 cause?
We mentioned above that CEOs will typically care about how they are perceived. This effect is likely to become less pronounced as CEOs get close to retirement. What would you imagine this implies for the compensation of CEOs as they get close to retirement? Would you expect compensation plans to get more sensitive or less sensitive?
1
Gibbons and Murphy.
IV: Do CEO compensation plans work?
This is a difficult issue to test. Jensen and Murphy look at a subsample of their data and argue that those with sensitive compensation schemes typically have better stock market performance than those with less sensitive compensation plans. Graef has looked at Jensen & Murphy’s full sample and finds that on the whole sample, no such effect can be seen.
Why might this not be a good test of the effect of compensation plans? [Hint: Think of what the Efficient Market hypothesis in finance tells us.]
Hence an appropriate measure is to look at how stock prices change when a new (sensitive) compensation plan is implemented. (This is known as an “Event Study.”) Typical finding: stock prices rise.
Are there any problems with concluding that sensitive compensation plans are liked by the stock market?
V: Are CEO plans designed efficiently?
(a)Rewarding for things beyond their control
(b)Little evidence of “relative performance evaluation”
(c)Few restrictions on hedging out
(d)Boards typically offer insurance against being sued
(e)Why aren’t CEOs required to buy stock?