This article is reprinted with permission from the
May 23 , 2005
edition of
New York Law Journal.
2005 NLP IP Company.


Report: "Putting Executive Pensions on the Radar Screen"

By Joseph E. Bachelder

Harvard Law School Professor Lucian A. Bebchuk has taken aim again – this time at pensions for CEOs of major U.S. corporations. Last fall Professor Bebchuk, together with Jesse M. Fried, a Professor of Law at the University of California, published Pay Without Performance (Harvard University Press), a book that is critical of the lack of linkage between top management pay and performance.

Now, Professor Bebchuk, together with Robert J. Jackson, Jr., an Olin Fellow in Law and Economics at Harvard Law School, has published a report entitled “Putting Executive Pensions on the Radar Screen.”[1] (This will be referred to as “the Report.”) The Report is likely to provoke significant attention and discussion. And deservedly so. It discloses that, for many CEOs, the portion of their pay represented by pensions is larger than many have realized.

Defining “CEO Pay”

Before examining the conclusions of the Report, it may be useful to review what we mean by “CEO pay.” CEO pay is the subject of annual surveys like those published by The Wall Street Journal and The New York Times as well as in business magazines such as Business Week, Forbes and Fortune. There are three “basic bricks” in the reporting of CEO pay. The first two bricks are salary and annual bonus (together, they often are described as “total cash compensation”). The third brick, frequently described as “long-term incentives,” is made up of stock options, restricted stock and other long-term awards. These three “pay bricks” are the focus of annual surveys of CEO pay.

Professor Bebchuk and Mr. Jackson raise a legitimate question: shouldn’t pensions be included in the annual surveys and comparisons of CEO pay? In the analysis of CEO pay this creates a “fourth brick” (a term used here, not in the Report).

“Pensions,” as used in the Report, mean defined benefit pensions.[2] A defined benefit pension is based on a percentage (such as 2% multiplied by the number of years employment) of covered compensation (such as the average of the final 3 years’ salary and bonus). This pension is paid over the life of the executive or, if the executive elects (at a somewhat reduced amount), over the joint lives of the executive and the executive’s surviving spouse. The surviving spouse, under such alternative, generally receives annually for life a fraction, such as 50% or 75%, of the annual amount paid to the executive during the executive’s life. (Some plans provide the executive with yet another election: to take a lump-sum payment of the actuarial present value at retirement.) In the following discussion, unless indicated otherwise, “pension” will refer to a defined benefit pension.

Two groups of executives were surveyed by Bebchuk and Jackson. The first group is made up of CEOs who, at the time of the study (2004), had retired from their positions during 2003 and the first half of 2004. The second group is made up of CEOs who, at the time of the study, were still serving and were at or near retirement age (all CEOs between ages 63 and 67 in the ExecuComp database at the end of 2003). All of the executives were or had been CEOs of S&P 500 companies. There are 28 executives with pensions included in the first group and 23 in the second group.[3]

The “Fourth” Brick

One can appreciate the significance of the fourth brick to those 51 executives with pensions from the following two findings (paraphrased from the “abstract” at the beginning of the Report):

•   The executives’ pensions had a median actuarial value of $15 million.

•   The executives’ accumulated pensions (actuarial value) had a median ratio of 34.5% to the executives’ total compensation during their service as CEO.

“Total compensation,” for purposes of calculating the 34.5% median, is defined by Messrs. Bebchuk and Jackson as the sum of cash compensation (salary and bonus) plus long-term incentives (both equity and non-equity, but excluding the pension value itself, which forms the numerator of the ratio).

The Report cites some striking examples. One such example is Dr. Henry A. McKinnell, CEO of Pfizer, Inc. The Report states:

At sixty-two, McKinnell is three years away from retirement. Assuming conservatively that his compensation will not increase before his retirement – and using the pension tables provided in Pfizer’s annual proxy – we estimate that Dr. McKinnell will receive an annual pension of $6.5 million upon his retirement. It is worth noting that Pfizer’s proxy statement discloses neither the actuarial value of Dr. McKinnell’s pension benefits nor the amount of the annual payment. Report at 8.[4]

The Report then estimates the actuarial value of such a pension. It values the pension at $71.5 million as a single life annuity. It values the pension at $83 million if the pension is a joint and survivor annuity with 50% to the survivor (assuming Dr. McKinnell’s spouse to be the same age as Dr. McKinnell).[5]

As a result of the Bebchuk/Jackson observations, the author took a random look at pensions accruing for CEOs at other S&P 500 companies. One such example (not included in the Report because he is not retired or between the ages of 63 and 67) is Kenneth D. Lewis, CEO of Bank of America Corporation. His pension starting at age 60 (the current proxy statement gives his age as 57) is estimated in the proxy statement for Bank of America to be approximately $3.5 million per year. Assuming a 75% joint and survivor annuity (indicated by the proxy statement) if his spouse survives him (and assuming she is the same age as Mr. Lewis) we estimate the actuarial present value of such a pension at age 60 to be approximately $48 million. Based on the same proxy statement information (the 2005 proxy for the 2004 fiscal year) we estimate the aggregate actuarial present value for the top four Named Executive Officers at Bank of America to be in the range of $80 million to $85 million.[6]

Following the merger of FleetBoston Financial Corp. into Bank of America in 2004, Charles K. Gifford, CEO of FleetBoston at the time of the merger, retired in January 2005 with a pension of approximately $3.1 million per year. Assuming a joint and survivor annuity similar to those of the Bank of America Named Executive Officers noted above, the actuarial value of Mr. Gifford’s pension is in the range of $40 million to $45 million. In selecting this illustration, we are not suggesting that Bank of America is alone in providing large pensions. The Report contradicts any such suggestion.


In analyzing the survey data, Bebchuk and Jackson make a number of important observations.

The median ratio between the executive pensions and total executive compensation in our first sample was 35.3%; in our second sample, 27.8%; and in the overall group of executives in our study, 34.5%. Table 7 [the related table in the Report] therefore indicates that the executives’ pensions represented a considerable proportion of the executives’ total compensation during their service as CEO. Report at 21.[7] (Bracketed statement added.)

[T]he significant variance among executives with respect to the relationship between pensions and total executive pay indicates that analysis of executive compensation that excludes pensions is likely to lead to substantially inaccurate comparisons among CEOs. Report at 22.

In sum, in any ranking of executives’ total compensation, the exclusion of pension values leads to significant underestimation of the relative position of executives with substantial pension values and overestimation of the relative positions of executives with low pension values or no pension plan at all. Report at 22.

Ignoring the Fourth Brick

Why has the “fourth brick” of executive pay – pensions – been so ignored in the analysis of executive pay?

1.   It is difficult to assemble the information.

Salary, annual bonus and long-term incentives are available in the Summary Compensation Table and associated stock option and long-term incentive tables of each issuer’s proxy statement. The Pension Table in the proxy statement is difficult for the average reader to understand and gives no assistance in determining the actuarial present value of a pension in order to compare it to other components of CEO pay.

2.   There are complexities of valuation in making comparisons among CEOs.

For example, how does one compare a pension for a CEO in his or her 40s with one in his or her 50s or 60s? Even if their pension formulas and other compensation were exactly the same, the current pension values will be worth less for the CEO in his or her 40s due to the age difference. This will be so even if the ultimate accrued pensions turn out to be exactly the same.

3.   What if one CEO works 30 years at one employer for his or her pension and another CEO has a contract right to the same amount of pension based on only 10 years’ employment?

How do you compare the two?

4.   Is it really appropriate to relate a pension based on 25 or 30 years of service to compensation being paid over a relatively short period as CEO?

(Table 7 of the Report indicates there may be a difference, but not a really significant one, between basing the ratio on the total compensation as CEO and basing it on compensation for the CEO period plus the period prior to becoming CEO.)

Despite these complexities, it would be very helpful to have better disclosure of pension information in proxy statements. The Report suggests improvements that might be made in proxy statement reporting of pensions.

(A)   Displaying the estimated annual pension benefit for each Named Executive Officer.

Presumably this would be a supplemental statement adjacent to the currently required Pension Table in the proxy statement. The Pension Table itself does not identify the exact amounts of the pensions estimated for each of the Named Executive Officers. Information as to years of service and corresponding pension amounts, as set out in the Pension Table, usually enable a reader willing to take the time an opportunity to “guess” at what the pension might be for a particular executive. This is not always possible when the pension formula (including covered compensation to which the pension formula applies) is not presented clearly.

(B)   Displaying the actuarial value of the pension (that is, the total current value of the pension, not just the estimate of future annual pension payments) for each Named Executive Officer.

This evaluation cannot be done without information regarding each executive’s age, whether a joint and survivor annuity is provided (in which event marital status and age of spouse would be necessary) as well as other factors (including interest assumptions) necessary to the calculation of an actuarial present value. For most proxy statement readers, the actual calculation, not just the assumptions, would be necessary to understand the actuarial present value. This also could be put in the same section as the Pension Table.

(C)   Displaying in the Summary Compensation Table the amount by which the actuarial value of the pension for each Named Executive Officer has increased over the preceding year.

This could be accomplished by adding a column to the Summary Compensation Table or by including this amount in the column for “All Other Compensation” together with an explanatory footnote.


Professor Bebchuk and Mr. Jackson have focused attention on the “fourth brick” in CEO pay. At the very least, the Report should encourage discussion on whether pension disclosures in proxy statements are adequate. More broadly, the Report suggests a rethinking of how we view pensions in valuing CEO pay.


[1] The Bebchuk/Jackson Report is available at

[2] Not included in the study on which the Report is based are other (generally less valuable) forms of retirement benefit such as defined contribution plans. Defined contribution plans are plans in which a certain percentage of current compensation (salary or, sometimes, salary and bonus) is set aside each year and then that amount, whatever it becomes, is available to fund a retirement benefit.

[3] It appears that the original sample included 77 CEOs and that this was reduced to 51 (28 in the first group and 23 in the second group) by eliminating those CEOs who have no pension plan. Thus, the first group (recently retired CEOs) had 41 executives which reduced to 28 after taking out 18 without pension plans. The second group (current CEOs near retirement) had 36 executives which reduced to 23 after taking out 13 without pension plans. Consequently, the medians and percentages in the Report are higher (being based on 51 CEOs with pensions) than they would have been if based on the original sample (assumed here to be 77 CEOs including both those with pensions and those without pensions).

[4] In preparing the column, we have assumed that Dr. McKinnell’s final average compensation (the average of his highest 5 years of salary, bonus and Long Term Incentive Plan (LTIP) payouts) will be approximately $11.5 million. This assumes his compensation remains unchanged for the next three years and he retires at age 65. A 1.75% accrual rate multiplied by his 35 years of service (the maximum years of service allowed) would result in an annual single life annuity of approximately $7.0 million. If Dr. McKinnell elects a 50% joint and survivor benefit, we assume the annual annuity amount for the single life annuity would be reduced so that the actuarial value of the benefit to be paid over two lives would approximate the actuarial value of the single life annuity. On this basis, our estimate is that a joint and survivor benefit of approximately $6.5 million (approximately 8% less than the single life annuity) would be paid each year. (One half of this, or $3.25 million, would be paid to Dr. McKinnell’s spouse if she survives him).

[5] The Pfizer pension plan adjusts the “regular” pension, a single life annuity, in the event a joint and survivor annuity is elected. We do not know this adjustment factor. For present purposes, as explained in footnote 4, we have assumed a reduction factor of 8% from a single life annuity for a 50% joint and survivor annuity. With regard to the single life annuity, based on our assumed mortality (1983 GAM) and interest (6%) factors we estimate the actuarial present value at age 65 for Dr. McKinnell’s pension to be approximately $78.2 million. With regard to the 50% joint and survivor annuity, assuming as we do a lower annual pension than for the single life annuity, the actuarial present value should be the same as that for the single life annuity.

[6] According to the proxy statement, one of the Named Executive Officers, James H. Hance, Jr. retired January 31, 2005.

[7] According to footnote 46 of the Report, “[t]o calculate the CEO’s total compensation, we used ExecuComp’s total compensation data including the value of stock options and restricted stock at the issuance date and adjusted each value to 2003 dollars using the Consumer Price Index.” Report at 21.