Letter to the Editor ¦ March 2004
Scrapping Social Security's Intermediate Cost Projections

By David Langer

The September edition contains explanatory text and principal economic and demographic assumption tables from the 2003 Annual Report of the Board of Trustees of the OASDI Trust Fund.

The following is my critique of these assumptions, which are selected by the trustees for the use of the Social Security actuaries in making the projections that appear in the annual reports.

Note that while the trustees state the future assumption factors are "inherently uncertain," this does not deter them from declaring the three sets of factors used (low cost, intermediate cost, and high cost) to be "plausible," and the "intermediate set represents the Board's best estimate of the future course of the population and the economy."

Therein lies my concern; I do not believe it is possible to make plausible assumptions over 75 years. However, recognizing social security law requires 75-year projections be made, the Actuarial Standards of Practice (ASOPs) for making social insurance cost projectionsmust accordingly be scrupulously observed to prevent undue subjectivity and political bias from playing a role. With regard to the latter, bear in mind that the trustees are high-level political appointees, including many not enamored of Social Security.

I will first present the results of a simple but revealing test I developed on the accuracy of the assumption factors. I presented this as a panelist at the June meeting of the Actuarial Society of Greater New York. I compared the actual projections of the assets as of 12/31/02 made at the end of each of the 10 years 1992 through 2001 with the actual assets on 12/31/02. The results appear in the chart at the right:

Projected Assets Compared to Actual Assets as of 12/31/02
(Actual assets equal $1378 billion)

Projected assets at 12/31/02
from each Annual Report (Billions)
Discrepancy at 12/31/02 between
projected and actual assets
Calendar Year Low Cost Int. Cost High Cost Low Cost Int. Cost High Cost
1992 1537 1120 671 12% -19% -51%
1993 1392 1048 669 1% -24% -51%
1994 1432 1153 874 4% -16% -37%
1995 1284 1068 845 -7% -22% -39%
1996 1214 1109 968 -12% -20% -30%
1997 1295 1225 1148 -6% -11% -17%
1998 1350 1297 1278 -2% -6% -7%
1999 1424 1407 1350 3% 2% -2%
2000 1410 1397 1353 2% 1% -2%
2001 1379 1372 1363 0% 0% -1%
Average Discrepancy
Cal. Years Low Cost Int. Cost High Cost
1992-1994 5% -20% -46%
1995-1997 -8% -18% -28%
1998-2001 -1% -1% -3%
1992-2002 0% -11% -24%
Prepared by David Langer, 9/23/03
Data from SS Trustees Annual Reports

The chart suggests that (1) the high cost projection is so far off it deserves to be discarded, (2) the intermediate cost projection should be redesignated as high cost, (3) the low cost projection, since it is on target, merits promotion to the intermediate level, and a new low-cost basis needs to be developed.

Consider the ramifications. The intermediate basis currently projects the assets to run out in 2041, while the low-cost basis develops a surplus of $18 trillion. At the end of the 75-year projection period, the difference grows to zero vs. $83 trillion.

Under low cost, there is also a never-ending annual surplus. One political implication is that the annual surplus can be "borrowed" by the U.S. Treasury for a great many years without the need for repayment, so long as benefits can be paid in full. A second implication, based on calculations I have made, is that normal retirement age 65 can be restored (age 67, based on an ammendment in 1983, is now being phased in).

Consider, too, the significat effect of the puzzling tendency of the trustees, in setting the intermediate assumptions for 75 years, to make those for the last 68 years or so more conservative than the first eight. One marked example is the key assumption, the annual rate of increase in labor force, which drops from 1.1 percent to below 0.3 percent. Another important drop is in the rel interest rate, from 3.3 percent to 3 percent.

There is additional evidence that the low-cost set of assumptions deserves upgrading to intermediate status, because of ASOPs have not been followed. See my article in Contingencies, "Social Security Finances are in Fine Shape," May-June 1999 and my op-ed in the Christian Science Monitor, "Cooking Social Security's Deficit," Jan.4, 2000. Visit my Web site,, or contact me for copies.

The understanding of the actuarial dynamics of social security takes some effort, but I have been studying them for about eight years, and have been amply rewarded by the vistas that have been opened, including economics, federal budgets, history, the essence of the privatization movement, and, of course, politics.

I urge more actuaries to join in the quest for a professionally sound actuarial basis for valuing Social Security.

-David Langer, A.S.A., E.A.