United States Senate
February 4, 1999
1737 21st Street, N.W.
Washington, D.C. 20009
Dear Dean and Mark:
I appreciate your letter of January 28th. Clearly, we are in agreement that additional detail is necessary about all Social Security proposals. I look forward to receiving the paper referenced in your letter and will indeed circulate it to my colleagues if it contains the specific information that I have requested. As you know, many proposals depend on the Social Security Trust Fund growing to a total of trillions of dollars, but are mum on where the general revenues will come from in order to redeem it. I take it from your letter that your proposal does not do this, and that you have detailed the general tax increases that would be required to pay it off at whatever size it reaches, in all years that Social Security exhibits a cash-flow deficit. It would be very useful for Senators to see the levels of general tax increases that are required in order to redeem the Trust Fund in years of cash-flow deficits, and if you have prescribed which non-Social Security taxes should be raised and by how much in order to do this, this would be helpful information.
I was intrigued by your finding that my proposal would cause workers to "lose benefits." As you may know, this contradicts the findings by CRS under both the Ball assumptions on rates of return and under conservative rate of return assumptions. Tables I and 4 in the December CRS report show that the present-value of benefits for a low-wage worker retiring at age 65 in 2030 would increase from $87,900 to $98,073 under our plan under conservative assumptions, and to $108,622 under the higher assumptions. For an average-wage worker the growth would be (table 2) from $145,280 to $146,867 under conservative assumptions, and to $170,3 10 under the higher assumptions. While you may take issue with CRS's methodology, they are certainly an objective source with no axe to grind in this debate.
I assume that you would not in your paper repeat the common mistake of comparing our solvent plan to a level of "current law" benefits that does not reflect current law at all, but assumes that substantial payroll tax increases have been enacted in order to meet benefit promises. Clearly, any assumption that one scenario gets the benefit of an infusion of revenue, should apply that same infusion of revenue to the other. If we assume equal levels of revenue infusions for each approach, the advantages of our plan are even greater. As seen on table 7 of the CRS report, a 2% addition to our proposal would increase the present value of a low-wage worker's benefit from $115,568 to $180,112 if that revenue were applied to our personal accounts instead of to shore up the current unfunded system.
As a further analytical point, of course, an objective analysis would also have to account for the impact of general tax increases from 2013-2032 upon the ultimate retirement income of Americans. Our plan would eliminate the vast majority of these pressures on general revenue. Under current law, however, the general tax increases necessary to pay off the Trust Fund would be the equivalent of a hike of 2.26 points in the payroll tax in 2020, 3.69 points in 2025, and 4.66 points in 2030 (see Table III.A.2, 1998 Trustees' Report.) A fair analysis of which plan results in greater retirement income would require the effects of these tax increases to be subtracted from the income of the Americans working in the relevant years, and assumptions made about how this ultimately affects their levels of retirement income. I assume that your paper takes these factors into account when comparing our plan's benefits to "current law.
If your paper succeeds in detailing the size and form of the tax increases that are required to pay off the Trust Fund during the "drawdown" period, and shows that retirees wind up with more income under the traditional system than our plan when all the relevant factors are taken into account, I would be pleased to circulate it among my colleagues, and to forward an accounting of stock return rates in short order. If, however, it makes misleading comparisons between unfunded "current law" and a solvent system, or leaves out the critical information about who would pay off a multi-trillion dollar Trust Fund, it would include far more sizable gaps in analysis than any details about the components of stock returns.