Marginal Utility

by Tom Bozzo

A couple days ago, my mother pointed me towards this pro-privatization op-ed from the Wilmington News Journal by professors Stacie Beck and Eleanor Craig of the University of Delaware economics department. Mom wanted to know what I thought about it — request-blogging lives!

It not uncoincidentally happens that I hold a B.A. in economics from Delaware, and moreover was hired by Eleanor Craig in 1995 to teach three sections of the course that in my days was called "Money, Credit, and Banking" but had by then been revved up to something like "Introduction to Monetary Economics." My personal experience of Eleanor Craig was a nice person and good boss. I don't know Prof. Beck. I will try to be tough only on their arguments below.

They both are conservative partisans to the extent of having signed a letter in early 2003 endorsing Bush administration budget policies as fiscally responsible, along with two other Delaware profs. That obviously hasn't turned out so well. Since the department has 27 full-time profs, that's pretty high uptake (several much larger top-tier departments only provided one or two signatories each); the Delaware econ department was on the conservative side.

----------------------------------------

The op-ed itself is of the variety where privatization will make everyone richer, smarter, and keep rabbits from eating their perennial gardens — unless you're over 55 in which case the important thing is that you should be reassured that you're getting none of it.

Beck and Craig lead off with one of my favorite privatization howlers:

You could own your own retirement assets if Congress passes President Bush's plan to guarantee the solvency of Social Security and establish personal retirement accounts.

Wow, I could own my own retirement assets? Makes me wonder what all that mail I get from Vanguard is about. Naturally, nearly everyone can own their own retirement assets through the relatively simple expedient of opening a traditional IRA or Roth IRA and putting some money in it.

Of course, there is doom-mongering:

[Social Security] is projected to start losing money in 2018, and will become bankrupt in 2042.

The "worthless IOU" canard is implicit, otherwise the tapping of the trust fund, projected to happen in 2017 by the Social Security Trustees (evidently, the News Journal deadline hit prior to the 2005 Trustees' Report), is the sort of thing that any prefunding anticipates — the need to draw down previously acquired assets. Not mentioned is the latest Congressional Budget Office reading, which puts the dates for trust fund tapping and trust fund exhaustion at 2020 and 2052, respectively.

All bankruptcy is relative, though:

Keeping these promises will require drastic increases in payroll taxes (from 12% to 18%) or in other taxes. The alternative will be for Congress to reduce promised benefits by 25%.

The second sentence contains a material sin of omission: scheduled benefits when the trust fund is exhausted in the Social Security Trustees' "intermediate cost" scenaro are considerably higher in 2005 dollars than current retirees' benefits. They are so much higher, in fact, the benefit cut needed to avert "bankruptcy" would leave future benefits higher in inflation-adjusted terms than they are now.

Well, back to worthless IOUs:

If workers could use Social Security taxes to fund personal retirement accounts, their money will finance their own retirements rather than having it spent by Congress, which is the case right now.

Josh Marshall notes today that the incremental public debt incurred under the actual and projected Bush budgets is large enough to have "paid back" the entire Social Security trust fund. Much of that debt has been incurred to the (maybe temporary) benefit of wealthier income tax payers. Hmmm. See also DeFazio on the House floor, via Atrios.

I also wonder if Profs. Beck and Craig have any U.S. government debt in their own private retirement accounts...

On the other hand, the long-term rate of return on stocks has been nearly 7%. A conservative estimate of the return on a balanced portfolio of stocks and bonds is 4.6%, after inflation.

That rate of return assumption is only "conservative" by assertion. There is ample scholarship suggesting otherwise, notably Shiller's recent work suggesting that forward-looking equity returns are not likely to match past performance, and Baker, DeLong, and Krugman [PDF] on the inconsistency of forward-looking returns with the long-range economic assumptions used to set the trust fund expiration date. Other things equal, stock returns of historical magnitudes imply that the market price-to-earnings (P/E) ratio would increase to levels greatly exceeding the 2000 peak of the bubble in 2000 around mid-century and increase from there without bound.

Now they move on to a numerical example. Their numbers don't tie to the Social Security Trustees' projections, so there are considerable nitpicking opportunities, but I'll try to focus on the more substantive problems.

For example, an average person entering the work force today gets a benefit of $19,600 a year in today's dollars. If he put one-third of his payroll taxes in a personal retirement account, his government check would be $12,600, plus he would have assets of $165,000 in an account.

Placing those assets in an annuity would add $18,200 a year to retirement income, for a total of $30,800 -- more than 50% greater than Social Security would provide.

Assume that the "average person" is a 22-year-old earning the "average wage index" amount and retiring at 67 in 2050. The expected annual current-law benefit for that scenario (intermediate cost) is $24,440 in 2005 dollars according to the Social Security Trustees' Report. So if Beck and Craig intend to refer to current-law benefits, they're materially understating what Social Security would offer.

If the Trustees' intermediate trust fund balance forecasts are right, the payroll tax will generate sufficient revenue to pay 75.2% of OASDI benefits in 2050. Per CBO, the benefits would be fully payable in 2050, with the payroll tax covering 78% of current law costs in 2053. It's anyone's question how benefit formulas would be adjusted, but clearly the benefit could be at least $18,379 (with rounding, 75.2% of $24,440; $19,584 under CBO's less pessimistic assumptions), again in 2005 dollars. So if Beck and Craig meant "payable" benefits under the Trustees' assumptions, in the absence of any other program changes, they're more-or-less on the mark.

But it should be noted that the "bankrupt" system can pay 20% more, in real terms, than today's system, which provides a yearly benefit of $15,335 in this scenario. People wondering whether Social Security can "be there" in the future should take note.

I used the CEPR Accurate Benefits Calculator (ABC) to check on the private account figures, along with some back of the envelope calculations of my own. In the ABC, I specified a single person, born in 1983 and retiring in 2050 (at 67), earning the default wage (which is insignificantly different from the Trustees' 2005 average wage index forecast) added 0.25 percentage point to the return to match Beck and Craig's 4.6%, and de-selected the "enhanced low-earner benefit" option. Per the ABC, the private account would grow to $167,213.76 gross of fees at the worker's retirement.

If the account is half debt (at the CBO's 3.3% real interest rate) and half stock, the implied 5.9% real return on stocks implies, other things equal, a market P/E of 63.9 in 2050.

Significant problems arise with the claimed private account benefits.

First, $165,000 cannot be turned into a $18,200 yearly annuity payment at present market terms. A single participant in the federal Thrift Savings Program, which according to CBO offers better annuity terms than the open market, could today turn $167,213.76 into $15,372/year without inflation adjustment (inflation would erode the value of that payment to the equivalent of less than $9,000 after 20 years), or $11,580/year with an inflation adjustment. Those are 84% and 64%, respectively, of the amount claimed by Beck and Craig.

Second, in addition to materially overstating the private account annuity, they likewise understate the effects of guaranteed benefit offsets and reductions in the Bush "plan." There are two expected sources of benefit reductions: a change in the benefit formula to freeze "real" benefit levels (not present in plans such as Ryan-Sununu that don't address the actuarial balance issue), and a benefit offset to pay back the borrowed money that goes into the accounts.

The $7,000 guaranteed benefit reduction is not quite enough to pay back the borrowing to fund the private account. Using the CBO's interest rate projections, the ABC figures the required benefit offset for the borrowing at $9,293/year. Beck and Craig are closer at the Social Security Trustees' lower real interest rate assumption (3% versus 3.3%).

An additional benefit reduction whammy of roughly equal magnitude arises from the benefit formula change, another $8,986 per the ABC. Between the two, the guaranteed check is not $12,600, but rather $6,829. (Note, this subtracts from the ABC's $25,108 figure for current-law benefits, which is the CBO-assumption analogue to the $24,440 from the Trustees' Report, not from Beck and Craig's $19,600.)

Add back an inflation-protected annuity (to be comparable with Social Security benefit terms) and the grand total benefit from this privatization scenario is $18,409 — 40% lower than the figure Beck and Craig cite, 26.6% lower than the current law benefit, and just about what the Social Security payroll tax can pay on its own without other changes to the system. Reducing the benefit offset rate and lowering annuitization costs from the TSP's low levels could get back maybe half the difference.

There is no privatization miracle, though, even under the aggressive assumptions regarding the private account returns. More plausible stock returns would decimate the privatization case.

And there's the free lunch:

Finally, the long-run budget impact is zero or positive. The cost of diverting payroll tax dollars into personal accounts today will cover [sic] future retirement benefits.(*) Because the new plan will increase savings, investment, productivity and economic growth, in the long run there will be more government tax revenue for other programs or tax cuts.

This, I'll admit, is the theory. However, the plan will involve no net additional savings in the near term, since every dollar "saved" in a private account will be offset by a dollar of borrowing from the public. With no net savings (or tax reductions), there's no reason to expect additional investment and associated economic growth in the near term. The market has to believe that today's new debt truly will be paid with future spending reductions, and future Congresses can't bail out seniors who fare poorly under the plan or do anything else that turns thirteen or fourteen figures' worth of privatization transition costs of privatization into a reality. Alternatively, pigs might fly.

So, Mom, the bottom line is that if you overstate the benefits by at least half, and assume away the transition difficulties, privatization looks great. In the real world, this sort of argument should not get a very good grade in an undergraduate economics class.

--------

(*) This seems to be an artful way of saying that the cost of diverting payroll tax dollars today will be covered by future reductions in government-paid retirement benefits.