Diagnosing the Dynamics of Social Security

by Bruce Webb

Social Security is typically discussed in static terms, that is it ‘will’ or ‘is projected’ to get to this state or that at some fixed point in the future. This I think is an artifact of ‘crisis’ being seen as the result of a known fixed event, namely Boomer Retirement. And it is certainly true that what was until recently the largest demographic cohort in American history is scheduled to retire between now and 2031. That is we know the wave is going to crash the only question is how rough the surf is going to be when it does.

It is this sense of inevitability derived from a fixed or static event that has led us to decision making drawn from ‘will’ rather than ‘if’. But in actuality Social Security is not fixed in stasis by the demographics, it is instead a dynamic system on any number of axes and should be studied as such in our search for proper policy choices.

One dynamic axis is that of Report time, in that the projected outcomes vary from Report year to Report year. If the crisis was really a static one fixed in place by the demographics you would expect that any year of inaction would make the ultimate gap between cost and income that much harder to fill and/or make the onset of the gap be sooner than projected. This is turn would lead naturally to assertions like ‘sooner is better than later’ and ‘the longer we wait the greater the shock treatment needed’. And indeed this was the actual outcome of Report Years 1993 to 1996, Trust Fund depletion came ever closer in time and the cost of the fix needed continued to rise. But a funny thing happened starting in 1997, the arrow reversed and TF depletion started retreating back in time and the cost of an immediate fix started to shrink. Suddenly what had been a crisis fixed by demography became a dynamic system. Which should have changed the analytical question from ‘what should we do’ to ‘why is it moving the way it does’ and more importantly ‘can this go on’.

That analytical shift never seemed to happen, to the extent that anyone even noticed the reversal it tended to be seen just as variation around the known result of Boomer Retirement, because whatever the numbers showed Boomers were not getting any younger. But for those who chose to ask the ‘why’ and ‘can’ question the result was kind of stunning. Social Security was not static at all, its outlook changed depending on particular projections that varied dynamically not only within each Report but between them. 1997’s Intermediate Cost alternative was not necessarily identical to 2000’s which in turn were not those of 2004. Once you break through the idea that you are fundamentally facing a static crisis fixed by the demography you end up open to some new analytical avenues.

One is just to examine the various projections in sequence, what caused the 1998 result to be so much better than the 1997, why did you get the deterioration between 2005 and 2006 Reports and the improvement between the 2007 and 2008? Are these movements just variations around a trend line firmly aimed at crisis or is the trend line pointing somewhere else? And the answers to these questions show that the SSA’s Office of the Chief Actuary is wrestling with a live system where the inputs never quite come in exactly as expected and the assumptions remain open for modification. For example there was a change in projected depletion date from 2034 to 2037 between the 1999 and 2000 Reports with a corresponding drop in cost of an immediate fix from 2.02% of payroll to 1.92%. On inspection the cause of this dramatic change was a big change in input from actual 1999 economic performance, future numbers were if anything moved in a more pessimistic direction. Which is to say that 10% of ‘Crisis’ simply vanished and I might add never really came back. Further dynamic movements in inputs and assumptions have moved that date to 2041 and the payroll gap down to 1.7%. (And we can expect the marked slump in receipts first visible in early Fall to move those numbers at least slightly the other way).

More analysis of the dynamic (if not dynamic analysis) below the fold.

Once you break free of the stasis you are free to examine the dynamics of Social Security from several perspectives. One is the diagnostic, what caused the models to move in the first place? And is this improvement likely self-sustaining, i.e.will the patient simply get better on its own? Or is the improvement simply transitory with a reversion to full blown crisis ahead? Or perhaps the improvement is real but plateaued, just leaving us with a smaller problem perhaps requiring a different fix? Alternatively you could take a pro-active stance. Having learned that the models can move and having determined why they moved are there steps we could take to assist that movement?

All of these questions come almost automatically once you make the shift from a static projection of demographic crisis to a dynamic projection taking in all the moving parts and determining which ones are moving on their own and which need some exterior push. But it all starts with noticing that the system is moving at all.

Back in 2000 the answer to these questions was pretty easy. From 1997 to 2000 the Reports had significantly under estimated near term growth, moreover then wise men like Greenspan were busy assuring us that these growth rates were totally sustainable, that we could crank out 3% productivity forever. Which if true meant that there would be no Social Security crisis at all, if anything that meant the system was over-funded going forwards and the real discussion should be the amount of timing of FICA tax cuts. Well much of the bloom went off that particular rose when productivity crashed in Q4 2005 and with it stalled the steady improvement in outlook we had experienced right through the Spring of that year (which in retrospect explains the ease of the victory of the ‘ There Is no Crisis’ pushback on the Bush plan). And flashing forward to today is truly sobering, recent numbers show OAS having stalled and DI actually in retrograde. (Biggs has the February numbers in his post Slowdown Slashes Social Security Surplus and they are not pretty). On the other hand the 2009 Report is due out next week and so will give us a new moving target to examine.

(And speaking of moving target it will be very interesting to compare the GDP projections in the Social Security Report to those in the Budget to those in the CBO analysis of the Budget. The 2008 Report had Intermediate Cost projecting 2.3% Real GDP by 2017 and never improving from there over the actuarial window, while Low Cost had Real GDP at 2.8% with ultimate 2.9% long term. My bet is that we will end up splitting the difference.)

Much of the typical argumentation around Social Security was formed in place via a snapshot of where the system was at the time, typically from the early 1990’s when the projected collision of ‘deficits as far as the eye can see’ and Boomer retirement set up a fairly grim scenario. For example people often point out that Krugman was fairly pessimistic about Social Security in 1996. Well he had a right to be concerned the dynamics of SS from 1993 to 1996 were awful. But the earth moved under his feet by 1999 Baker and Weisbrot were fully justified in calling it a Phony Crisis, and this was still true in 2004. Given the totality of what we knew then and the Bush Administrations own projections of economic growth from his tax cuts there simply was no crisis at all. And there still isn’t, not really, even the big shock of the current recession/depression will probably be weathered fine by Social Security. But the situation is fluid, it is dynamic and needs to be discussed in those terms. We need to shelve the ‘will be’ and embrace the ‘may be’ and the ‘if so’.

(To anticipate some commenters. The current deficits projected for the next decade don’t create the same scenario as those of the late eighties, early nineties because the date of TF depletion has been time shifted forward in time in a way that is irreversible, it has now more or less been disconnected from the point of maximum Boomer strain on the system rather than coinciding with it as it used to project).