Subject: Social Security Reform without benefit cuts
From: Martin Feldstein (msfeldst@nber.org)
Date: Mon Jun 26 2000 - 05:47:33 EDT
Dear Public Economics Program Members:
The attached file presents new calculations of the effects of financing
Personal Retirement Accounts by allocating 2% of existing payroll tax
revenue. Critics have claimed that doing so forces benefit cuts. The
analysis in the attached paper, based on actuarial and economic assumptions
used by the Social Security actuaries, shows that benefits can be
maintained and that the long-run value of the Trust Fund can be positive
and growing.
There is a temporary period when the Trust Fund must borrow to maintain
benefits but this borrowing is repaid with interest. This contrasts with
those other plans that use general government finance as a permanent
supplement or that infuse new bonds that have to be repaid with additional
general government revenue. In short, despite the need for temporary
financing, the Trust Fund remains solvent.
The attached paper is also available as NBER Working Paper 7767 by Martin
Feldstein and Andrew Samwick under the title: Allocating Payroll Tax
Revenue to Personal Retirement Accounts to Maintain Social Security
Benefits and the Payroll Tax Rate
The paper has been published in a slightly revised form in Tax Notes, June
19th, as "Allocating Payroll Tax Revenue to Personal Retirement Accounts,"
pp 1645 to 1652.
The analysis is important in the current policy debate because Governor
George Bush has proposed funding personal retirement accounts by allocating
existing payroll taxes. Critics who claim that such Bush-type Personal
Retirement Accounts require benefit cuts base their conclusion on a paper
by Henry Aaron, Alan Blinder, Alicia Munnell, and Peter Orszag. Although
these authors show that a plan can be designed that requires benefit cuts,
the attached paper shows that such cuts are not necessary. The 3 key
differences between the Aaron et al paper and the Feldstein-Samwick
analysis are described below.
The basic point is that there are many ways in which the Personal
Retirement Accounts can continue to maintain total retirement income and
achieve long-run trust fund solvency. The Feldstein-Samwick paper is not
an attempt to describe "the" Bush proposal but only shows that such a plan
is feasible. An earlier Feldstein-Samwick paper (Tax Notes, May 1998, and
NBER Working Paper 6540) showed a different way in which personal
retirement accounts could achieve the same goals.)
Critique of Bush Plan by Aaron, Blinder, Munnell and Orszag
(Revised - 6-12-2000)
Martin Feldstein
The recent critique of the Bush Social Security plan by Aaron,
Blinder, Munnell and Orszag (www.tcf.org) concludes that the total benefits
(the combination of traditional PAYGO benefits and the annuities from
Personal Retirement Accounts) would be reduced by 20% relative to current
law for a typical individual retiring in 2037 at the age of 65. In
contrast, our conclusion is that the total benefit could be at least
maintained for all future retirees.
The three assumptions in the Aaron et al study that lead to their
conclusion are:
Individuals at retirement shift their Personal Retirement Accounts to a
fixed interest rate annuity. They say that they assume a real annuity but
do not give the real interest rate. They instead state that they assume a
real interest rate of only 3.0 %, slightly more than half of the 5.5
percent return on the "variable annuity" that we assume. This cannot by
itself explain a 20% reduction in the total benefit since the 10.4% annuity
would still be providing substantial benefits. (We have done the
calculation with a 3% real annuity and found that even with this low return
we can maintain benefits. All that happens is the Trust Fund remains
negative for a longer period of time. But the system is solvent; the Trust
Fund eventually returns to a positive value.)
The SS Trust Fund cannot borrow or run a deficit while we do have a period
of borrowing by the Trust Fund. (They note in footnote 14 that they have
considered allowing the Trust Fund to borrow but decided it was
implausible. But they correctly note that if borrowing is allowed the
trust fund could repay the debt and be back at 100 percent of payroll by
2074.) This requires a sharp cut in the PAYGO benefits.
The extra corporate tax revenue is not credited to the Trust Fund.
We note that the
increased national saving leads to additional profits and therefore
additional corporate tax revenue. We transfer this revenue, estimated at 2%
of the assets in the PRAs, to the Trust Fund. Their failure to make this
transfer causes the Trust Fund to be smaller and to require larger cuts in
PAYGO benefits.
It is the combination of the sharp cut in the PAYGO benefit and the reduced
annuity that causes the 20% cut in total benefits.
It is interesting to note that in their analysis they grant, at least for
the purpose of their study, that:
(1) the relevant thing to look at is total benefits from both
sources;
(2) a reasonable return assumption is our 5.9% before administration
expenses (although they discuss the risk that the return would be less than
this);
(3) our 0.4% administration cost is reasonable
The study has an extensive discussion of risk. Instead of looking at the
likely risk, they focus on the extreme case that the individual might
experience the worst 35 years of the past 50. In that case, the PRA account
at retirement would be cut by 38% below the value with 5.5% return. The
best 35 years, they note, would produce a PRA account at retirement that is
three times as large (i.e., plus 200 percent) as the account value with the
5.5 % return.
June 12, 2000
