Sharp Pencils vs. Sharp Politics: The 2015 Social Security Trustees Report

The long run balance of the U.S. Social Security system has improved, but almost all of the improvement comes from improved methods and data.  That’s the sharp pencil part, explained in Section IV.B.6, pages 74-79, of the recently released 2015 Social Security Trustees Report.  Put another way, Social Security is in better shape than we thought last year not because of anything that has changed in the world but just because we weren’t thinking exactly right back then.

But the Disability Insurance bit is quite definitely in short term trouble, as the DI Trust Fund is projected to hit zero in 2016.  The Trustees recommend changing the law so that the DI Trust Fund can draw from the larger OASI Trust Fund, which seems reasonable on its face, whereupon the next moment of trouble is reached only in 2034 when the joint Trust Fund hits zero.  The fix for that one requires either increased income (meaning taxes) or decreased cost (meaning benefits) or a bit of both.  Is that the sharp politics part?

Not really.  Something that is only projected to happen twenty years from now is unlikely to focus political decision making in the coming year.  Normally we would expect this particular can to be kicked down the road.

Maybe not.

The closing Actuarial Opinion by Stephen Goss (255-257) points to a more immediate point of likely political conflict, concerning the way the Social Security system is treated in unified budget accounting, which differs from the trust fund accounting in the Report.  His concern, apparently, is that Congress will ignore the sharp pencil Trustees Report and apply instead their own accounting framework.

There are two problems.  First, “unified budget accounting treats redemptions of trust fund reserves as an addition to annual federal deficits”.  That redemption has already begun in the DI Trust Fund, obviously, but it hasn’t added to the annual federal deficit because the OASI Trust Fund has been increasing by more.  Nevertheless, the moment when the joint trust fund begins to decline is coming up fast, projected in 2020.  And there is a presidential election on the way.  Nuff said.

Second, “budget analysis frequently refers to both trust fund reserve redemptions and trust fund obligations not payable under the law after reserve depletion as factors that increase the federal debt held by the public in the future”.  After all budget deficits increase debt, don’t they?  And after the trust funds hit zero in 2034, unified budget accounting assumes that the difference between income and cost, which can no longer be met from the trust fund, will be met instead from general government revenues.  More deficit, more debt, more politics.

In a way, the whole point of Trust Fund accounting was to keep Social Security separate, to give it the look and feel of a contributory pension scheme even though it is actually Pay As You Go.  The Trust Fund measures the excess of past revenues over past expenses, hence saving for the future, right?  Original sin, breaking down the wall between Social Security and the general budget, happened in 2011 and 2012, when a temporary reduction of the payroll tax (Social Security) was implemented as part of the fiscal stimulus package (general government).  The money was more or less paid back as a contribution from the General Fund, but the separation was broken.  No wonder Goss is worried.

I don’t know much about politics, but I do know a bit about the economics of retirement security.  Put aside the politics, and the accounting.  The basic facts we are confronting are demographic.

The aging of the American population, as a consequence of the long term trend toward lower fertility and declining mortality, poses the most significant challenge to our retirement system in the coming decades. We are shifting from an economy in which the dependent population consisted mainly of children, to one in which the dependent population will be predominantly the elderly. The difficulty of managing that shift arises from the fact that whereas children are supported in large degree by direct income transfers within the family, grandparents are more typically supported by indirect transfers using either the government (i.e. Social Security) or financial markets (i.e. pension funds and housing wealth) as intermediary.

The shift toward a relatively larger elderly population thus necessarily requires an expansion of either government or financial market distribution mechanisms, or both. In the future, a
higher fraction of total consumption will necessarily be paid for with “unearned” income,
and a lower fraction with “earned” income, simply because a higher fraction of total
consumption will be flowing to the elderly. The demographic challenge is to devise
distribution mechanisms to effect that transfer most efficaciously.

Twenty years ago, when I read my first Trustees Report, the push was on to privatize Social Security, which seemed to me not only not a good idea but also basically impossible as a matter of economics.  The point I emphasized then is that the projected aging of the population is putting pressure on all the mechanisms through which non-workers gain access to current income, not just Social Security. The economic (and political) question was how that pressure should be spread among the various component parts of the retirement system.  Eliminating the existing Social Security system, as the privatizers were urging, would just put that much pressure on everything else.

In this regard, and coming back to the present, it is significant that so-called dependency ratio (over 65s relative to working age 20-64) is projected to rise from .247 currently to .438 in 2090.  At the same time the ratio of beneficiaries (cost items) to covered workers (income items) in the system is project to rise from 35 currently to 51, or about 1:3 to 1:2.  But the fraction of GDP that Social Security is projecting to be channeling toward beneficiaries is projected to rise only from 5% currently to 6.2% over the same period.  Not to put too fine a point on it, but we are really not asking Social Security to bear very much of the burden imposed on us by demographics.

So why does it feel like such a burden?

In my view twenty years ago, which remains my starting point today, the main problem is that the tax base to support the problem has been systematically eroded.  The biggest culprit has been the rise on non-wage benefits (mainly health insurance) as a fraction of total employee compensation.  Another smaller culprit was the cap on maximum income subject to the tax, which combined with rising inequality to cause less and less of total wage income to be subject to the tax.

From this standpoint, I am happy to see that the current report projects a leveling off of the non-wage benefit fraction, though at a high level.  I am unhappy to be unable to find data on the second culprit, though I observe that the cap remains.  What exactly is being assumed about inequality?

There is lots more in this report that will teach the close reader much about the actual structure of the U.S. economy.  Fertility, mortality, immigration, are central drivers.  But there is also fascinating information about the changing structure of the labor market, including the participation rates of the over-65 (which is another margin along which we can adjust to aging).  I dearly wish that everyone would read the details, and learn about the problem we are facing.  I fear that sharp politics will keep us from learning what the sharp pencils have found out.

One comment on “Sharp Pencils vs. Sharp Politics: The 2015 Social Security Trustees Report

  1. tom michl on said:

    Nice post. I always thought your take on social security as a “social mutual fund” was spot on.

    Not sure this is precise though:

    “And after the trust funds hit zero in 2034, unified budget accounting assumes that the difference between income and cost, which can no longer be met from the trust fund, will be met instead from general government revenues. More deficit, more debt, more politics.”

    Just to be clear, the social security system cannot use general revenues to pay benefits, so this would require new legislation I think.

    FYI (and FYreadersI), there is a piece in the last JEP on the actuarial estimates of fertility/mortality that claims they are systematically off (too pessimistic) over the last couple of decades.