Max Speaks, and He Listens and Replies

Max Sawicky kindly (and constructively!) responds to my recent post about whether we need to reform Social Security today. His original statement began:

There is absolutely no reason at present to make changes in Social Security, except out of political fear of the Right.

As I noted, first in "Why is Social Security a Campaign Issue?" and then again in the post to which Max is responding, my reason is that we make the problem about $300 billion worse every year that we delay. This is roughly the interest that we accrue on the unfunded obligation of $10.4 trillion in a year when the real interest rate is 3 percent, as in the 2004 Trustees Report's long-term assumptions.

I should reiterate that my preferred solution--raising the age of full benefit entitlement (with some other modifications) for future beneficiaries--does not require new funds to flow into the system today (or ever). And there is no need to do that immediately, except to give people as much time as possible to react to the new law and plan accordingly. If we don't mind giving them one fewer year of notice, then I don't suspect that we would mind enacting the same reforms to plug a $10.7 trillion hole tomorrow rather than a $10.4 trillion hole today. But taking a "wait and see" approach does not strike me as sound fiscal management.

In addition, to the extent that the reforms will include raising revenues in the near-term to provide investment returns to support benefits in the future (whether through personal accounts or the Trust Fund), then delay does have distributional consequences across different age cohorts. According to the way Social Security played out in this year's campaign, neither side was willing to ask more of or give less to those cohorts who are "at or near retirement." If an additional year of delay means one more birth cohort crosses the threshold of being "at or near retirement," and thus exempt from its share of filling the $10.4 trillion hole, then we should reform sooner rather than later.

I don't think Max and I disagree about the math underlying this calculation, but we do disagree about how relevant it is. Max writes:

I don't get excited about measures of the present value of unfunded liabilities from now till forever because I think they are jive. It's a way of ginning up a huge number. If you want to do that, compare it to the present value of GDP.

Fair enough. In "How to Reform Social Security, Part I," I noted:

Once new revenues are being added to the system, then it becomes important to figure out where they should go--the Trust Fund or personal accounts.

Confronted with that choice, I opt for personal accounts. For me, an immediate and permanent contribution of 3.5 percent of taxable payroll into personal accounts for all workers, in addition to the 12.4 percent payroll tax that they and their employers already pay, is preferable to the current system. The contributions are 3.5 percent because that is the amount that the Social Security actuaries say is required to restore solvency even if invested entirely in Treasury bonds. But such a reform, though preferred to the current scenario, is also far from ideal.

The 3.5 percent figure is roughly what Max is asking for in the way to make the comparison. (The small difference is that taxable payroll is projected to fall by about 5 percentage points relative to GDP over the next 75 years.) Max and I can differ on how much confidence we have in the Social Security actuaries' projections. The Trustees Report does contain some sensitivity analyses and additional projections that factor in uncertainty, but not on these two measures. The latter projections do suggest that there is essentially no chance that the system won't be running annual deficits in 75 years. Factoring in the uncertainty suggests to me that there is more, not less, of a need to act sooner rather than later.

Max also wonders:

One thing that is baffling about this whole privatization campaign is why a knowledgeable person like Andrew would prioritize the 2042 problem over the 2018 problem. The latter date is of course about when income tax revenue will be needed to redeem obligations to the Trust Fund and, by extension, Social Security beneficiaries.

In fairness, I have never thought of this as a 2042 problem, or even a 2018 problem. (That was, if I recall, the third paragraph in this post by Max.) To me, it is a 2004 problem that I don't want to mushroom into an even bigger problem, in 2005, 2018, 2042, 2080, or any year at all. Entitlement programs should be projected to be in balance--period. Accomplishing that requires changes in the programs in the future (if not the present). Those changes should be debated and legislated sooner rather than later, so that the changes can spread the burden as evenly as possible across cohorts.

Max concludes with some statements about the politics of reform:

As far as politics goes, you can hardly blame the Dems for declining the opportunity of leading with their chins with tax increases and benefit cuts, in the face of a hugely fiscally irresponsible Administration.

What does he mean? Of course I can blame them, like I did here. And I can also try to point them to "Democrat-friendly" reforms like the Diamond and Orszag plan, like I did here. Social Security's unfunded obligations are not a creation of the Bush administration and have been widely discussed for over a decade. The Democrats and Republicans alike have had ample opportunity to behave responsibly on entitlements. Some have, but most have not, and the sooner the first group can prevail, the better off we'll be.