REPLY

Howell E. Jackson

I am grateful to all of the commentators for their thoughtful and probing responses to my Article. Their analyses clarify both the advantages and challenges of moving toward a system of modified accrual accounting for Social Security and our budgetary processes more generally. In the same spirit, I would like to reply to a few of the points raised.

First, Professor Clark asserts that reforms in the 2003 Trustees Reports obviate the need for the reforms I propose.[1] As the Clark comment stresses and as I noted in my Article, the 2003 Trustees Report included for the first time alternative presentations of trust fund solvency, including estimates of the trust funds’ open-group and closed-group liability. Unlike Professor Clark, I regard this change as modest and of little importance to the public debate over Social Security reform.

Let me first address the modesty of the 2003 changes. As my Article noted, the change appeared sixty pages into the 2003 Trustees Report and was limited to two and a half pages of analysis. The information presented in the analysis falls short of the FASAB disclosure requirements for Social Insurance programs—that is, the disclosures do not satisfy generally accepted accounting standards for the federal government. In addition, the disclosures do not contain any estimates of accrued liabilities, which my proposal would emphasize and which Professor Clark admits are “an important and useful indicator of the financial status of Social Security.”[2] Moreover, if one reviews the introductory materials to the 2003 Trustees Report, the inconsequential nature of the reforms becomes apparent. While the materials do report for the first time the net present value estimate of the open-group seventy-five-year horizon liability, the information is imbedded into a format that is identical to past reports and conveyed in a way unlikely to attract the attention of anyone but Social Security aficionados.[3] Neither the 2003 Trustees Report itself

*** Top of Page 222 ***

nor the press release accompanying its release in March of 2003 made any claim that the content of the report had been altered to any noteworthy degree.[4] Even more telling, I have found no general press accounts after the release of the 2003 Trustees Reports that have made any mention of the fact the report included important new materials, much less that the change was in some way significant.[5] Indeed, press coverage of the 2003 Trustees Report tracked the traditional measures of performance that the trustees’ reports have included for many years.

Second, several of the commentators disagree with me over the desirability of distinguishing between accrued benefits and benefits to-be-accrued in the future. I personally find it useful to quantify the cost of retirement benefits that the system has incurred to date and will not restate my reasons here. I would, however, like to address some of the political aspects of quantifying these liabilities. As Professor Clark notes and as I acknowledged, we have reduced accrued benefits in the past, most notably in the 1983 reforms.[6] To a large degree these reductions undid retroactive enhancements in benefits from the preceding decade. Without revisiting the wisdom of the 1983 reforms, these retroactive reductions in benefits have complicated reform efforts in recent years. Seniors are now suspicious that “Social Security Reform” is just a code word for retroactive reductions of benefits. Professor Clark, for one, recognizes this difficulty when he notes “virtually all reform proposals be-

*** Top of Page 223 ***

gin by stating that current retirees will not be affected,” but then he and others resist my suggestion that the financial statements of Social Security should recognize these obligations of liabilities. One way to address the legitimate concerns of seniors and to facilitate sensible reforms is to recognize their claims as financial liabilities of the system and then move on to other issues.

To a large degree, I suspect, it is the recognition of accrued liabilities to mid-career workers that the commentators really oppose. As several comments noted,[7] the recognition of accrued liabilities for mid-career workers may make it more difficult for politicians to reduce these benefits. Accrual accounting would not, however, constitute an absolute barrier to retroactive adjustment in accrued benefit of mid-career workers, and such reduction could make sense particularly if these workers were given access to other forms of retirement savings such as individual accounts. To be sure, there is some tension between my proposing the recognition of these liabilities and then leaving open the possibility that they might be adjusted. The goal of liability recognition, however, is to convey the accretion of probable economic sacrifices over time.[8] Mid-career workers are earning statutory entitlements under the Social Security Act and, with the passage of time, these entitlements will mature into politically binding obligations once these workers reach retirement age. Any accounting system that does not recognize the gradual accrual of these obligations over time is substantially misleading, at least in my view.

One lesson that emerges from the foregoing debate is the impossibility of articulating a fully unbiased estimate of the present value of future Social Security benefits.[9] If one focuses on absolute legal entitlements, then future benefits have no present value, because Congress could eliminate the program completely.[10] Instead, one could focus on accrued benefits, as I propose, on the ground that this system best comports with statutory formulae and best approximates political reality. A third option, following President Bush’s Commission,[11] analyzes payable benefits—that level of current benefits that the projected revenues would cover over the next seventy-five years or some other period of time. Fourth, as traditionalists prefer,[12] one could focus on scheduled benefits,

*** Top of Page 224 ***

because that is the level of benefits that the statutory structure currently mandates. All are defensible approaches and each has an element of bias.

Finally, the overarching question at issue in this debate asks which financial format the trustees should use to report the financial performance of the trust funds. The traditional format features the system’s net change in assets as a short-term measure of performance along with the seventy-five-year actuarial deficit and trust fund exhaustion dates as measures of long-term solvency. Modified accrual accounting highlights the system’s annual profit or loss and the level of implicit trust debt as a percentage of GDP. The accompanying table presents these measures under both approaches for the past six years.[13] Which format most effectively communicates the financial condition of the trust funds?

Under the modified accrual accounting approach, the message for the general public is clear. The Social Security trust funds are incurring annual substantial losses—on the order of hundreds of billions of dollars a year—and the level of implicit trust fund debt is increasing as a percentage of GDP. The take home lesson to the general public is that Social Security is losing ground fast and its implicit debts are mounting rapidly. The case for reform is clear.

The traditional measures convey a very different message. Net assets have increased markedly in each of the last six years, rising $164.4 billion in 2002. The seventy-five-year actuarial deficit, measured as a percentage of total payroll, has declined (that is, improved) in four of the last six years, and the deficit at the end of 2002 was actually a good deal lower than it was at the of 1997. Over the same period, the projected trust fund exhaustion date has actually been pushed back in each of the last six years. At year-end 2002, the exhaustion date was a full decade later than it was five years earlier. One reason the trustees have had such difficulty in communicating the seriousness of the crisis in Social Security finances is that the financial statement they produce each year highlights financial measures that imply the systems finances are, at the very least, holding their own. The crisis does not seem imminent.

With financial statements, there must be a bottom line. The bottom line for Social Security’s current financial statements should be substantially and unambiguously negative. The traditional presentation does not satisfy this criterion. A modified system of accrual accounting would.

*** Top of Page 225 ***

	Traditional Measures			Modified Accrual Accounting
Year-End	Net Change in Assets ($ billion)	Actuarial Deficit (% of payroll)	Year of Trust Fund Exhaustion	Annual Loss ($ billion)	Implicit Debt (% of GDP)
1997	+ $ 88.6	– 2.19%	2032	– $515	97.7%
1998	+ $107.0	– 2.07%	2034	– $287	95.8%
1999	+ $133.7	– 1.89%	2037	– $550	96.8%
2000	+ $153.3	– 1.86%	2038	– $761	98.9%
2001	+ $163.1	– 1.87%	2041	– $535	101.1%
2002	+ $165.4	– 1.92%	2042	– $370	101.7%

[1] See Robert L. Clark, Liabilities, Debts, Revenues, and Expenditures: Accounting for the Actuarial Balance of Social Security, 41 Harv. J. on Legis. 161 (2004).
[2] See id. In the Spring of 2003, I wrote Professor Clark in his capacity as chair of the Technical Panel on Assumptions and Methods, outlining the additional information that the trustees’ report would need to include to allow outside analysts to derive more complete accrual accounting statements for Social Security. See Memorandum from Howell Jackson to Robert Clark (May 13, 2003), available at http://www.law.harvard.edu/faculty/hjackson/projects .
[3] Reproduced below are the two key sections of the introductory materials of the 2003 Trustees Report:

The combined OASDI Trust Funds are projected to become insolvent in 2042 under the long-range intermediate assumptions. For the trust funds to remain solvent throughout the 75-year projection period, the combined payroll tax rate could be increased immediately by 1.92 percentage points, benefits could be reduced immediately by 13 percent, a transfer of $3.5 trillion in general revenue (in net present value) could be made, or some combination of approaches could be adopted. Significantly larger changes, would be required to achieve solvency beyond 75 years . . . .

Another way to illustrate the financial shortfall of the OASDI system is to examine the cumulative value of taxes less costs, in present value. Figure II.D4 shows the present value of cumulative OASDI taxes less costs over the next 75 years. The balance of the combined trust funds peaks at $2.3 trillion in 2017 (in present value) and then turns downward. Through the end of 2077, the combined funds have a present-value unfunded obligation of $3.5 trillion . . . .

The open group unfunded obligation over the 75-year projection period has increased from $3.3 trillion to $3.5 trillion.

2003 Trs. of the Fed. Old-Age and Survivors Ins. and Disability Ins. Trust Funds Ann. Rep. 3, 10, 12 [hereinafter 2003 Trustees Report]. Some information that Professor Clark cites as new to the 2003 report actually appeared in earlier reports. For example, prior reports had routinely reported how much benefits would need to be cut after the trust fund exhaustion date in order to equal projected revenues. See, e.g., 2002 Trs. Of the Fed. Old-Age and Survivors Ins. and Disability Ins. Trust Funds Ann. Rep. 16–18 [hereinafter 2002 Trustees Report].
[4] See Press Release, Social Security Administration, Social Security Not Sustainable for the Long Term (Mar. 17, 2003), available at
http://www.ssa.gov/pressoffice/pr/trustee03-pr.htm.
[5] See Howell Jackson, Accounting for Social Security and Its Reform, 41 Harv. J. on Legis. 59 (2004).
[6] See id. at Part II.C.2; Clark, supra note 1.
[7] See Kent Smetters, Inadequacies of Accrual Accounting in Social Security, 41 Harv. J. on Legis. 215 (2004); Peter A. Diamond & Peter R. Orszag, Accrual Accounting for Social Security, 41 Harv. J. on Legis. 173 (2004).
[8] Contrary to Professor Clark’s suggestion, see Clark, supra note 1, recognition of liabilities under accrual accounting does not require certainty of payment.
[9] But see id.
[10] Professor Smetters seems to advance this view, see Smetters supra note 7, although his preferred measures of fiscal imbalance rely on present value calculations of scheduled benefits and revenues over the infinite horizon.
[11] See generally The President’s Comm’n to Strengthen Soc. Security, Report of the President’s Commission: Strengthening Social Security and Creating Personal Wealth for All Americans (2001).
[12] See, e.g., Clark, supra note 1; Diamond & Orszag, supra note 7.
[13] For the source of the data in the table, see 2003 Trustees Report, supra note 3, at 2, 4; 2002 Trustees Report, supra note 3, at 6, 16; 2001 Trs. of the Fed. Old-Age and Survivors Ins. and Disability Ins. Trust Funds Ann. Rep. 2, 4; 2000 Trs. of the Fed. Old-Age and Survivors Ins. and Disability Ins. Trust Funds Ann. Rep. 4, 5, 24; 1999 Trs. of the Fed. Old-Age and Survivors Ins. and Disability Ins. Trust Funds Ann. Rep. 2–4, 23; 1998 Trs. of the Fed. Old-Age and Survivors Ins. and Disability Ins. Trust Funds Ann. Rep. 2-4, 23.

If you have any questions or comments about JOL, please contact us: hlsjol[at]law[dot]harvard[dot]edu
Last modified: September 16, 2005.

End of content. Navigation bar follows