Kent Smetters[*]
I agree with Professor Jackson’s assessment of the problems arising from the traditional budgetary treatment of the Social Security system.[1] The Trust Fund’s exhaustion date, which was recently projected as 2042 under the Social Security Actuary’s “Intermediate Assumptions,”[2] has received most of the attention in the past.[3] After the Interim Report of The Commission to Strengthen Social Security, more attention is now also given to the date future Social Security benefits are expected to exceed future Social Security contributions, which was most recently projected as 2018.[4] This so-called “cross-over date” could have more relevance than the Trust Fund exhaustion date if past Trust Fund surpluses have been spent, rather than used to reduce the debt held by the public, and will likely continue being spent in the future.[5] These measures, however, do not provide adequate information concerning the financial problems facing Social Security. Policymakers can manipulate both of these measures, and neither will indicate whether we have actually increased national saving sufficiently to address baby-boomer demographic problems.
The current focus on cash-flow budgeting, deficits, and debt in the federal government’s accounting system not only understates the problems facing the Social Security system, but this focus biases the choices of policymakers toward minimizing short-term deficits at the cost of long-term imbalances. A more relevant accounting framework would include both short- and long-term shortfalls.
*** Top of Page 174 ***
Jackson instinctively approaches the problem as a lawyer well versed in the approach used under private pension systems: accrual accounting. He asks why the government should not be held to the same standard that it imposes on the private sector.[6] This intuition certainly has some rhetorical appeal. Indeed, in testimony before Congress, I once made the off-the-cuff remark that current federal budgetary practices would be illegal in the private sector.[7]
Accrual accounting is reasonable for a private defined-benefit pension system in which firm shareholders cannot assume that the firm will exist forever. In this environment, pension rules must be constructed to determine how much money a company must set aside in order to meet its pension obligations in case the firm goes bankrupt. Retirement benefits in private pension systems are inherently “back loaded,” because benefits are based on the last few years of earnings prior to retirement, and some fairly ad hoc assumptions must be used to calculate the accrued liability; vesting requirements, in turn, necessitate additional assumptions. Except for vesting requirements, these particular technical problems are less of an issue for Social Security, in which retirement benefits are calculated using a worker’s best thirty-five years of indexed earnings.[8]
More significantly, accrual accounting is less attractive for government accounting than for private accounting where it is more natural to assume that the government will last well into the future. Accrual accounting tells us nothing about Social Security’s long-run sustainability. Whether Social Security will need parametric changes in order to remain sustainable depends on the infinite-horizon open-group obligation, a concept discussed by Jackson.[9] This obligation was recently estimated by the Social Security Trustees as $10.5 trillion in present value—a significant shortfall.[10] In other words, benefits must be cut or taxes increased by $10.5 trillion in present value for Social Security contributions to cover scheduled benefits.
While the infinite-horizon open-group obligation is fairly consistent with the system’s accrued liability as calculated by Jackson,[11] this result is coincidental. In fact, if Congress were to decrease benefits or increase taxes enough to eliminate the long-run imbalance, the infinite-horizon
*** Top of Page 175 ***
open-group liability would become zero. In this case, the system would be in long-run balance, even though it continued to transfer resources from one generation to another. The accrued liability, however, would remain positive, unless Social Security were fully privatized or fully pre-funded so that each generation paid for its own benefits. So these two concepts really are unrelated.
At first glance, it might appear that the accrued liability measure would indicate the amount of “implicit debt” that people alive today and in the past are passing to future generations. Accrued liability does not accurately measure “implicit debt,” however, because accrued liability is inherently backward-looking, as it is calculated based on previous contributions. The advantage of being backward-looking is that it is quite easy to calculate. This ease of calculation makes it potentially suitable for certain privatization experiments when the government cannot estimate the individual productivity of each worker, as in the optimal income tax literature.[12] Alternative approaches, however, can be used to finance transitions to a privatized social security system that are less expensive than the costs suggested by accrual accounting.[13]
The cost of the simplicity of the accrued liability calculation is that it overestimates the true liability that the Social Security system is passing to future generations because the accrued measure fails to net out the future taxes that will be paid by younger and richer workers in excess of their future benefits. The more inclusive closed-group obligation measure gives the correct measure of the obligations that are being passed forward in time. This measure also corresponds more closely to the negative impact that a pay-as-you-go Social Security system has on household saving decisions and national saving.[14]
Jackson agrees that the standard accrued liability measure would overestimate the obligations that the Social Security system is passing to future generations.[15] As a result, he presents a modified accrued liability measure that appears to be very similar, if not identical, to the more standard closed-group liability measure.[16] While the usefulness of his modified measure is clear, one wonders why he then even uses the language of “accrual accounting.” Why not focus the paper on the merits of the closed-group obligation measure? To be sure, the positive difference in size between the accrued liability and the closed-group liability measure reminds us that many generations alive today will receive less money in
*** Top of Page 176 ***
present value from Social Security than they are projected to pay into the system. But the difference is spread across many living cohorts, so it is not particularly indicative of the amount of money that, for example, an individual born into the system will pay in excess of future benefits.
As noted above, the accrual accounting approach might have some rhetorical appeal. This spin does not come free. The key problem with accrual accounting is that it gives a sense that Social Security benefits accrued to date are somehow “bonded,” that is, that they represent legal liabilities in the same sense as government debt.
In Fleming v. Nestor,[17] however, the Supreme Court recognized that a Social Security benefit represents the policy of Congress and not an express or implied contract between beneficiaries and the Social Security Administration. The Social Security Administration, therefore, picks its terminology very carefully. For example, in the Trustees Reports, future projected benefits are referred to as “scheduled benefits” instead of “promised benefits.”[18] Moreover, the open- and closed-group statistics are reported as measures of “obligations” instead of “liabilities.”[19] Jackson seems to brush aside the importance of these distinctions, almost taking for granted that Congress would never reduce accrued benefits. In doing so, Jackson restricts the degrees of freedom that Congress has available when making policy. Indeed, the President’s recent Commission to Strengthen Social Security proposed three plans, all of which would have reduced the accrued benefits of some people in exchange for personal accounts.[20]
In contrast, the infinite-horizon open-group and closed-group obligation measures make no pretense of bonding. Instead, they simply describe the implications of the continuation of current policy. The infinite-horizon open-group obligation indicates whether current policy can continue without changes. The closed-group obligation indicates the value of obligations that are being passed to future generations if current policy were to continue. Together, these two measures make a powerful combination, and they can be integrated easily with the rest of the federal budget.[21] Both measures can also be applied to the U.S. Medicare system, which faces a long-run imbalance that is about five times larger than the Social Security system.[22] The nation’s Medicaid program also faces big problems.[23]
*** Top of Page 177 ***
In fact, using the same economic and policy assumptions in the President’s 2004 Budget, the U.S. federal government faces a $45 trillion fiscal imbalance across all expenditure categories, including Social Security, Medicare, Medicaid, defense, transportation, and education.[24] In other words, the current level of debt held by the public plus the present value of all future projected expenditures exceeds the present value of all future projected taxes and other receipts by $45 trillion.[25] Jackson is correct that it is important that the government make these shortfalls more transparent in the federal budget. His paper presents a very persuasive set of arguments toward this end, but accrual accounting fails to reflect adequately the current state of the Social Security system.
[*] Assistant Professor of
Insurance and Risk Management, The Wharton School, University of Pennsylvania.
Research Assistant, Public Economics, National Bureau of Economic Research
(NBER). Ph.D., Harvard University, 1995; M.A., Harvard University, 1992; B.S.,
Ohio State University, 1990.
[1] See Howell E. Jackson, Accounting
for Social Security and Its Reform, 41 Harv. J. on Legis. 59 (2004).
[2] See 2003 Bd. of Trs. of the Fed.
Old-Age and Survivors Ins. and Disability Ins. Trust Funds Ann. Rep., H.R. Doc.
No. 108-49, at 2, available at http://www.ssa.gov/OACT/TR/TR03/tr03.pdf
[hereinafter 2003 Trustees Report].
[3] See, e.g.,Christian E. Weller,
Editorial, Securing Social Security, Wash. Post, Aug. 20, 2002, at A12;
Michael Barone, A Knack for Reframing, US News & World Rep., Sept. 8,
2003, at 23.
[4] See Kent Smetters, Is the Social
Security Trust Fund Worth Anything?, 94 Am. Econ. Rev. (forthcoming 2004)
(manuscript at 11, available at http://irm.wharton.upenn.edu/WP-security-smetters.pdf
) (last modified June 2003).
[5] See id.
[6] See Jackson, supra note 1, at
Part I.A.2.a.
[7] See Balanced Budget Amendment: Hearing on
H.R. Res. 22 Before the Subcomm. on the Constitution of the Comm. on the
Judiciary, 108th Cong. 26 (2003) (statement of Kent Smetters, Assistant
Professor, The Wharton School, University of Pennsylvania).
[8] See Soc. Sec. Admin., Examples of
Benefit Calculations for Workers Attaining Age 62 in 2004 (last modified Oct.
15, 2003), available at http://www.ssa.gov/OACT/ProgData/retirebenefit1.html
(last visited Nov. 7, 2003). Social Security has its own technical problems,
however; unlike the linear retirement benefit formulas in the private sector,
Social Security’s benefit formula is progressive, requiring some arbitrary
choices when calculating accrued liability.
[9] See Jackson, supra note 1, at
Part II.C.3.
[10] See 2003 Trustees Report,
supra note 2, at 62.
[11] See Jackson, supra note 1,
at Part II.B.1.
[12] See generally J. A. Mirrlees, An
Exploration in the Theory of Optimum Income Taxation, 38 Rev. Econ. Stud.
175 (1971).
[13] See generally Kent Smetters
& Jan Walliser, Opting Out of Social Security, 88 J. Pub. Econ.
(forthcoming 2004), available at http://irm.wharton.upenn.edu/WP1-Opt-Smetters.pdf
(last modified Aug. 7, 2002).
[14] See id. at 5.
[15] See Jackson, supra note 1,
at Part II.C.5.
[16] See id.
[17] 363 U.S. 603, 610–11 (1960).
[18] See, e.g., 2003 Trustees Report,
supra note 2, at 8.
[19] See, e.g., id. at
44–61, 62, tbl.IV.B7, 63, tbl.IV.B8, 64–73.
[20] See President’s Comm. to
Strengthen Social Security, Strengthening Social Security and Creating Personal
Wealth for All Americans, at 119–42 (Dec. 2001) available at http://www.csss.gov/reports/Final_report.pdf.
[21] See Jagadeesh Gokhale & Kent
Smetters, Fiscal and Generational Imbalances: New Budget Measures for New Budget
Priorities 27 (2003).
[22] See id.
[23] See generally Jan Ellen Rein,
Misinformation and Self-Deception in Recent Long-Term Care Policy Trends,
12 J.L. & Pol. 195 (1996); William Alvarado Rivera, A Future for Medicare
Managed Care: The Lessons of California’s San Mateo County, 7 Stan. L.
& Pol’y Rev. 105 (1995-96).
[24] See Gokhale & Smetters,
supra note 21, at
27.
[25] See id.
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.