Robert C. Pozen[*]
Professor Jackson has written an excellent critique of the accounting methods currently applied to Social Security. In short, he articulates two major concerns.[1] First, including the annual cash-flow surplus of Social Security in calculations of the annual U.S. budget surplus or deficit misleads the public and confuses policy makers.[2] Second, emphasizing cash-flow accounting numbers for Social Security creates the illusion that its short-term finances are sound, although its long-term financial problems are serious.[3] In response to both these concerns, Jackson proposes a modi-fied form of accrual accounting, which would show large deficits for Social Security in the short term as well as the long term. Jackson concludes that if Congress recognized the large annual deficits run by Social Security under accrual accounting, it would have a strong incentive to exclude Social Security from calculations of the annual U.S. budget surplus or deficit.[4]
This Comment begins by commenting on the two main concerns discussed by Jackson and his proposals for resolving these concerns. Next, it will discuss the implications of his proposals for evaluating various plans to reform Social Security. In this regard, I will draw upon my experience as a member of President Bush’s Commission to Strengthen Social Security [hereinafter the “Bush Commission”].
Congress is fond of making the U.S. budget look more positive than it actually is by including the annual surpluses of Social Security based on cash-flow accounting. During the last few years of the Clinton administration, this device was used to inflate current budget surpluses by
*** Top of Page 200 ***
$100 to $200 billion per year.[5] During the Bush administration, this device is being used to reduce the size of the annual budget deficit by similar amounts.[6]
The inclusion of cash-flow surpluses from Social Security in the annual numbers for the total U.S. budget is unjustified as an economic matter and inconsistent with relevant legislation, as Jackson demonstrates.[7] Additionally, by systematically understating the size of the current budget deficit, the inclusion of Social Security’s cash-flow surplus is bound to mislead the public about the extent of U.S. government spending. What if the public were told that the official 2003 federal budget deficit of $400 billion is actually more than $550 billion without the addition of an ephemeral cash-flow surplus of more than $150 billion from Social Security? The size of this accounting distortion makes Enron look like a minor accounting error.[8]
Fortunately, relief is in sight. By 2007, the annual cash-flow surpluses of Social Secruity will begin to decline; by 2017, they are projected to turn negative.[9] Congress, therefore, will be inclined over the next decade to adopt Jackson’s suggestion that Social Security be unconsolidated completely from the U.S. budget. Under this approach, any transfers from the U.S. government to the Social Security system would be treated in the same manner as U.S. government outlays to third parties.[10]
The placement of the Social Security system totally outside the U.S. budget would also help clarify the nature of assets held by the so-called Social Security trust fund. The status of the Social Security trust fund has been the subject of vigorous academic and political debate. In reality, the trust fund is an accounting conduit: it represents a set of claims by current and future retirees on the U.S. Treasury. By keeping the Social Security trust fund off-budget, Congress will see more clearly that Social Security claims of workers are not self-financing; rather, these claims are akin to Treasury bonds already held by public investors. In both cases, the U.S. government can pay out monies owed to third parties only if
*** Top of Page 201 ***
Treasury sells new bonds to the public or appropriates funds for this purpose.[11]
As Jackson explains, annual cash-flow accounting is inappropriate for a retirement plan such as Social Security that increases its long-term liabilities each year, even though these liabilities will not be paid out for several decades.[12] Instead, he suggests a modified form of accrual accounting, based on the way private pension plans file their annual reports under Generally Accepted Accounting Principles (“GAAP”).[13] His proposed methodology would focus on the addition of retirement benefits earned by current workers reduced by the excess of taxes to be paid by such workers over the additional benefits they will accrue over the rest of their working lives.[14] If the government were to follow the accounting method proposed by Jackson, the net accrued liabilities of Social Security would have grown by roughly $370 billion in 2002, as compared with the $160 billion surplus recorded for Social Security under cash-flow accounting.[15]
The change from cash-flow to accrual accounting would have a dramatic impact on the politics of Social Security reform. Under cash-flow accounting, there is little political incentive to deal with unpopular issues such as payroll taxes and benefit formulas.[16] When members of the Bush Commission pressed for prompt action on Social Security reforms, they were often met with the response that there was no rush because the system would be running a surplus for the next fifteen to twenty years. Many Senators and Representatives said they were not willing to incur the political risks involved with advocating Social Security reforms since they would no longer be in office when the Social Security surplus turned negative. Under accrual accounting, by contrast, the Social Security deficit is growing by leaps and bounds each year. If a politician delays Social Security reform for even one more year, the Social Security system will be billions of dollars further in the red.
But Jackson stops short of banning cash-flow accounting for Social Security; he wants to introduce modified accrual accounting alongside cash-flow accounting. While it is difficult to argue against presenting multiple perspectives on a complex set of issues, it is critical to establish
*** Top of Page 202 ***
a primary accounting approach that best captures the economic implications of Social Security. In this view, cash-flow accounting is inferior to accrual accounting not only because cash-flow accounting conflicts with GAAP for pension plans,[17] but also because cash flow accounting systematically understates the real economic cost of Social Security to taxpayers. One useful analogy is to consider how the SEC’s restrictions on a company’s presentation of non-GAAP measures of its financial performance—for example, the reporting of quarterly profits excluding acquisition charges—would apply to any government report on Social Security. While the SEC allows a company to announce non-GAAP measures, it must display the GAAP numbers prominently and specify which material items are omitted in the non-GAAP measures.[18] Following this SEC analogy, all government reports on Social Security would be required to use accrual accounting as the primary method of presentation. If cash accounting were used by a government report on a particular subject, then the report would also have to display prominently the results under accrual accounting and compare those to the results under cash-flow accounting.
More fundamentally, some critics argue against accrual accounting because Social Security benefits do not constitute a legally binding entitlement; they may, in theory, be reduced or eliminated by Congress.[19] In response, Jackson limits his modified accrual accounting system to Social Security retirement benefits already earned by workers, as compared to retirement benefits they may earn in the future.[20] While Jackson recognizes that Congress may retroactively change almost any aspect of Social Security, he points out that already earned benefits should be given preference over future benefits from the perspective of political realities and by analogy to private pension accounting.[21]
As Jackson recognizes, his proposal for modified accrual accounting on earned Social Security benefits is very similar to the closed-group unfunded obligation (“CGUO”), recently included in the back pages of the Social Security Trustees Annual Report (“Report”).[22] That Report, however, does not specify the year-to-year increase in the CGUO; Jackson computes this increase by analyzing a set of data supplied by the Social Security Administration to experts on request. The Report also includes an estimate of the current open-end group liability for Social Security—another form of accrual accounting that includes future as well as current
*** Top of Page 203 ***
workers.[23] Again, however, the Report does not specify year-to-year increases in open-end liabilities of Social Security.
The main objective should be for Social Security to move from cash-flow accounting to some form of accrual accounting as soon as possible. This move is critical to counteract the illusion that Social Security is in good financial shape for the next fifteen years and, as discussed below, to allow an accurate assessment of specific proposals to reform Social Security. Nevertheless, accrual accounting will be resisted by commentators who believe it is inappropriate for governmental programs where benefits or tax rates can be changed by Congress from year to year.[24] Although this general belief may have some validity, cash flow accounting on a yearly basis is unsound for retirement plans where the government is promising in the current year to make large payouts over thirty to fifty years in the future.
The CGUO should become the primary measure of the financial status of Social Security because the methodology for this form of accrual accounting is already established and because CGUO focuses on benefits already earned by current workers. CGUO is a more useful form of accrual accounting than open-end liability, because the latter includes projected Social Security benefits for people who are not currently workers but who are likely to become workers during the next seventy-five years. These projections would introduce unnecessarily a potentially large and speculative element into Social Security accrual accounting. The good news is CGUO for the current year is now included in the middle of the lengthy annual report by the Social Security trustees. The key is to move CGUO into the executive summaries of all official reports on Social Security, which should prominently display year-to-year changes in CGUO.
As Jackson indicates, the advantages of accrual over cash-flow accounting for Social Secruity can readily be seen in the debate about the creation of a personal retirement account (PRA) for Social Security participants.[25] In the carve-out form of a PRA, workers would invest a relatively small portion of their Social Security payroll taxes (i.e., 3% of the 12.4% total in Social Security payroll taxes) in a portfolio of securities.[26]
*** Top of Page 204 ***
In return, workers would agree to a so-called offset—lower Social Security benefits than currently scheduled by an amount reasonably related to the portion of their payroll taxes invested in PRA.[27] Thus, such workers would receive two types of retirement benefits—lower scheduled benefits from Social Security plus the investment returns from their PRA.
From the perspective of many workers, especially younger ones, the concept of a carve-out PRA should have much appeal. The projected real return after inflation for Social Security contributions over the next few decades is low (i.e., 1% to 2% per year),[28] relative to the real return estimated by Social Security actuaries for a balanced portfolio invested 50% in an equity index fund and 50% in a bond index fund (4.6% annually after administrative expenses).[29] Moreover, many younger workers are deeply skeptical that they will actually receive the level of Social Security benefits scheduled for their retirement.[30] Having participated in 401(k) plans, younger workers may like the idea of controlling the investment of some portion of their payroll taxes in a PRA, rather than having all their payroll taxes funneled into an arcane system in Washington, D.C.
Additionally, carve-out PRAs can reduce Social Security’s long-term deficit if the offset rate is set appropriately. While a portion of payroll taxes is allocated to PRAs during working years, workers will receive lower Social Security benefits during their retirement years. This trade-off is visible immediately under accrual accounting because the amount of earned Social Security benefits by current workers with PRAs would be significantly lower than those under the current system. An accrual accounting system would also drive home, as Jackson notes, the transitional costs of moving from the current system to a new system with lower Social Security benefits and PRAs.[31]
In contrast, the merits of PRAs are obscured by cash-flow accounting. The lower cash flows into Social Security from workers with PRAs are apparent immediately, yet the lower benefits due to such workers are not apparent for forty or fifty years. This deficiency in cash-flow accounting is accentuated by the conventional use of a seventy-five-year period to measure the long-term Social Security deficit. Consider a pro-
*** Top of Page 205 ***
posal for young people entering the workforce in 2037 to allocate 3% of their Social Security payroll taxes to PRAs. This allocation of 3% to PRAs would reduce the annual cash flows of Social Security every year from 2037 onwards, but the reduction in benefits accepted by these workers will not help the annual cash flows of Social Security until after 2077. Thus, under cash-flow accounting, there is no positive impact on today’s estimate of the long-term Social Security deficit.
Similarly, the consolidation of the Social Security trust fund into the total U.S. budget precludes a fair evaluation of any Social Security reform proposal involving investments in the securities markets. Under current accounting practice, claims by the Social Security trust fund on the U.S. Treasury are considered intra-governmental transfers with no effect on the total U.S. budget.[32] Accordingly, if some portion of the Social Security trust fund were invested in publicly traded securities, this portion would be deducted from U.S. budget totals because that portion would no longer constitute an intra-governmental transfer. As Jackson explains, however, this accounting method distorts public debate because the amounts held by the Social Security trust fund would remain the same immediately before and after the securities investment.[33] Instead, public debate should focus on the differences, other than budget accounting, between investments in diversified securities portfolios and intra-governmental claims on the U.S. Treasury.
For example, suppose the Social Security trustees proposed to invest a small portion of trust assets directly in a broad-based index of the U.S. stock market like the Wilshire 5000.[34] The debate on this proposal should focus on the expected returns and risks of this index investment, as compared to the economic returns and political risks of holding only U.S. Treasury bonds that must ultimately be repaid by congressional appropriations. This legitimate debate is only confused by the argument against the proposed investment of Social Security assets in a stock market index on the ground that such an investment would automatically increase the consolidated budget deficit. This artificial argument would be eliminated by keeping Social Security trust assets separate from the rest of the U.S. budget.
The move from cash-flow to accrual accounting would enhance the quality of public debate on two other types of Social Security reform proposals: addition of participants and changes in benefit formulas.
*** Top of Page 206 ***
A number of commentators have suggested that Social Security can be financially bolstered by bringing new participants into the system. One common suggestion is to bring into the Social Security system the many state and municipal workers who are currently exempt from it.[35] Another proposal is to bolster the finances of Social Security by increasing immigration of workers into the United States.[36]
Under cash-flow accounting, an addition of Social Security participants looks like a winning idea because annual revenues begin to increase immediately and higher benefits do not materialize until years later. The addition of any new cohort of workers may turn out, however, to be a net loser for Social Security over time if the total retirement benefits paid to such workers exceed their lifetime contributions plus interest. This could easily happen with low-wage workers because they will receive a relatively high level of benefits for each dollar they contribute to Social Security.
Jackson points out this deficiency of cash-flow accounting for evaluating the economic implications of expanding Social Security participation.[37] Under his approach to accrual accounting, by contrast, the finan-cial effects of adding new categories of workers would be reflected accurately. From the start, the incremental revenues from new participants each year would be offset to some degree by the accrual of Social Security benefits earned in that year by new participants. By including increases in benefit obligations as well as increases in contributions by new participants, accrual accounting would allow policy makers to estimate better and monitor the impact of new participant proposals on the long-term deficit of Social Security.
To reduce the long-term deficit of Social Security, commentators have suggested a variety of changes in the benefit formulas for retirees.[38] For example, some have advocated an increase in normal retirement age as life expectancy rises. Others have promoted the use of price indexing, rather than wage indexing, to adjust initial benefits in order to reflect the growth of the U.S. economy over the worker’s career.
Because any change in benefit formulas is controversial, it may be politically important to apply the change only prospectively. For instance, one of the constraints in the executive order establishing the Bush
*** Top of Page 207 ***
Commission was the maintenance of the current Social Security benefit structure for all workers “at or near retirement.” This phrase was interpreted by the Commission members to mean that any change in the benefit structure would not apply to any worker fifty-five years old or older. Such an interpretation implies that, for workers younger than fifty-five, a change in some aspect of the Social Security formulas might apply to benefits already earned in the past, as well as retirement benefits to be earned in the future.
Jackson’s proposal for modified accrual accounting supplies a particularly useful way to implement any reform proposal involving the grandfathering of benefits for various categories of workers.[39] Because his proposal captures the long-term cost of Social Security benefits already accrued by current workers, it would allow for the application of new formulas only to retirement benefits that accrue after the date the reforms are instituted. For instance, wage indexing could be applied only to initial Social Security benefits accrued in the past, while price indexing could be applied to benefits earned in the future. Such an approach to Social Security reform, as Jackson emphasizes, would have the political virtue of honoring existing commitments, while offering a rapid improvement in the long-term Social Security deficit by reducing the rate of benefit accrual in the years immediately following the adoption of the reform.[40]
In this comprehensive article, Jackson sharply criticizes the cash-flow accounting currently used in the Social Security system and puts forward a thoughtful proposal for a modified form of accrual accounting that should be employed in Social Security accounting. Although the exact details of his proposal can be debated, his argument in favor of accrual accounting over cash-flow accounting is compelling.
Accrual accounting for Social Security is, however, an arcane subject with little popular appeal, and politicians have a vested interest in the overly optimistic picture of Social Security created by cash-flow accounting. If Jackson wants his proposal adopted for Social Security accounting, he should publish a series of one-page editorials in newspapers and magazines.[41] In this article, he explains in detail why cash-flow accounting for Social Security is inappropriate, but useful, to the Washington establishment. Possibly through brief editorials, he can generate
*** Top of Page 208 ***
enough public support to make accrual accounting the primary measure of the financial status of Social Security.
[*] John M. Olin Visiting
Professor of Law from Practice, Harvard Law School. Secretary of Economic
Affairs, Commonwealth of Massachusetts. J.D., Yale Law School, 1972; B.A.,
Harvard University, 1968.
[1] Howell Jackson, Accounting for Social
Security and Its Reform, 41 Harv. J. on Legis. 59 (2004).
[2] See id. at Part I.B.
[3] See id. at Part I.A.
[4] See id. at Part III.
[5] See Congressional Budget Office, The
Economic And Budget Outlook: An Update, at 53 (Aug. 1998).
[6] See Congressional Budget Office, The
Budget And Economic Outlook: Fiscal Years 2002–2011, at 19 (Jan.
2001).
[7] See Jackson, supra note 1, at Part I.B.
[8] On October 16, 2001, Enron announced that it
was reducing shareholders’ equity by $1.2 billion because of improper
accounting. See Report of Investigation by the Special Investigative
Committee of the Board of Directors of Enron Corp., at 2–3 (Feb. 1, 2002),
available at http://news.findlaw.com/hdocs/docs/enron/sicreport/sicreport020102.pdf.Less than one month later, Enron announced a restatement of its
financial statements for the years 1997 through 2000 because of other forms of
improper accounting; these restatements involved a total reduction of a little
more than $2 billion in shareholders’ equity. See id.
[9] See 2003 Bd. of Trs. of the Fed.
Old-Age and Survivors Ins. and the Fed. Disability Ins. Trust Funds Ann. Rep.,
[hereinafter 2003 Trustees Report] at 8.
[10] See Jackson, supra note 1, at Part II.D.2.b.
[11] See id. at Part II.D.2.c.
[12] See id. at Part I.A.2.a.
[13] See generally, Employers’
Accounting for Pensions, Statement of Financial Accounting Standards No. 87
(Financial Accounting Standards Bd. 1985) [hereinafter FAS No. 87].
[14] See Jackson, supra note 1, at Part II.A.
[15] See id. at Part
II.D.1.
[16] See id. at Part
III.A.1.
[17] Compare cash-flow accounting with the
standards set forth in FAS No. 87, supra note
13.
[18] See, e.g., 17 C.F.R. §§
229.10, 244.100, 244.101 (2003).
[19] See generally, Accounting for
Social Insurance, Statement of Recommended Accounting Standards No. 17 (Federal
Accounting Standards Bd. 1999) [hereinafter FAS No. 17].
[20] See Jackson, supra note 1, at Part II.A.
[21] See id. at Part II.C.4.
[22] See 2003 Trustees Report.,
supra note 9, at 62–63.
[23] See id. at
61–62.
[24] See FAS No. 17, supra note
19, at Appendix A.
[25] See Jackson, supra note 1, at Part III.D.
[26] See Strengthening Social Security
and Creating Personal Wealth for All Americans: Report of the President’s
Commission 82 (Dec. 21, 2001). For a summary of other PRA proposals for SS, see
Liqun Liu & Andrew Rettenmaier, Comparing Proposals for Social Security
Reform, National Center for Policy Analysis Report No. 227 (Sept. 1999),
at http://www.ncpa.org/studies/s227/s227.html.
[27] See id. at90.
[28] See id. at 88.
[29] See id.
[30] A 1997 survey by the Employee Benefit
Research Institute and Matthew Greenwald & Associates asked whether people
had greater confidence that they would receive Social Security benefits or that
alien life would exist in outer space. Among Generation Xers, 63% had greater
confidence in receiving Social Security benefits while 33% had greater
confidence that alien life exists in outer space (4% presumably did not answer).
See The Future of Social Security for This Generation and the Next: Hearing
Before the House Ways and Means Subcomm. on Soc. Sec., 105th Cong. (1997)
(statement of Dallas L. Salisbury, President, Employee Benefit Research
Institute), at http://waysandmeans.house.gov/legacy.asp?file=legacy/socsec/105cong/10-23-97/10-23salis.htm
.
[31] See Jackson, supra note 1, at Part III.D.1.
[32] See id. at Part
III.D.2.
[33] See id.
[34] Several commentators have made this type
of proposal. See generally Henry J. Aaron & Robert D. Reischauer,
Countdown to Reform: The Great Social Security Debate 119–20 (Century
Found. Press 2001) (1998).
[35] See U.S. Gen. Accounting Office,
Social Security: Different Approaches for Addressing Program Solvency 33
(1998).
[36] See Aaron & Reischauer,
supra note 34, at 64–65.
[37] See Jackson, supra note 1, at Part III.B.1, III.C.3.
[38] See generally, U.S. Gen. Accounting
Office, supra note 35, at
38–49.
[39] See Jackson, supra note 1, at Part III.F.
[40] See id. However, Congress may want
the flexibility of changing already accrued Social Security benefits, at least
for younger workers.
[41] During the process of editing this
comment, Professor Jackson did publish such an editorial. See Howell E.
Jackson, It’s Even Worse Than You Think, N.Y. Times, Oct. 9, 2003,
at A35.
Copyright © 2004 by the President and Fellows of Harvard College
Harvard Journal on Legislation (JOL) - Volume 41, Number 1, Winter 2004
If you have any questions or comments about JOL, please
contact us: hlsjol[at]law[dot]harvard[dot]edu
Last modified: September 16, 2005.