I came up with this around ten years ago to put a point on the logical trap Social Security privatizers (often unwittingly) find themselves in. It goes like this:
To understand the trap we need to do some parsing on ‘privatization’ and ‘crisis’. Now traditionally ‘Social Security crisis’ was equated to ‘Trust Fund Depletion’, which is the point in the future when the Social Security Trust Fund balance projects to go to zero. Over the last twenty years of Social Security reporting the date of Depletion has been projected by the Trustees at various points between 2019 and 2042 (and the reasons for the variation are interesting) but are in recent Report years put in the mid to late 2030s. ‘Depletion’ has often been sold in terms of ‘Social Security won’t be there’ in the sense of ‘no check for me’, particularly to the under 40s but a few seconds of thought shows that ‘no check’ is not a possible outcome as long as payroll tax is being collected, that is benefits can be paid out right up to the amount allowed by then current income. Instead we are talking about a benefit cut, and one relative to the current law baseline (and examining that baseline is interesting as well-subject for later posts). The amount of that benefit cut has been projected variously between 22% and 25% of the ‘scheduled benefit’ or to flip it around a payout of between 75-78%.
So in our first sum-up we have ‘Social Security crisis’ = ‘Trust Fund Depletion’ = ‘25% benefit cut’. In these straightforward terms the solution to ‘crisis’ is to prevent ‘benefit cuts’ and this can be done only with some combination of the following three methods: a direct increase in contributions (i.e. tax increase), an improvement in those economic numbers that contribute to solvency (mostly employment and Real Wage), or a better return on contributions than the current combination of Pay/Go and Trust Fund investments provide.
And it is here that the jaws of the logic trap start to close on privatizers. Detailed discussion below the fold.
The annual Reports of the Trustees of Social Security define ‘crisis’ in terms of ‘actuarial deficit’, or the gap between ‘scheduled benefits’ and ‘payable benefits’ and express that gap alternately as percentage of payroll, or of GDP, or in ‘present value’ dollar terms. Taking ‘percentage of payroll’ first, this is given in terms of the amount of tax increase that would be needed immediately to backfill the actuarial deficit over the 75 year projection period or in terms of the amount needed at the point of Trust Fund Depletion. And these alternatives are spelled out in the Conclusion of the Report Summary, or the first section of the Social Security Report itself: http://www.ssa.gov/oact/tr/2012/II_E_conclu.html
Short version: immediate increase of 2.61% of payroll vs increase in 2033 starting at 4.3% and ultimately reaching 4.7% or a total of 17.1% combined compared to today’s 12.4%. What Social Security doesn’t do, but CBO does (using slightly different assumptions), is to score intermediate approaches that would phase in these increases, which would split the difference with an ultimate increase of around 3.5%.
If an immediate increase of 2.61% is the bitter medicine what then would sweeten it somewhat? Well one approach, that of phased increases has just been referenced and is also the methodology of the Northwest Plan for a Real Social Security Fix (the work product of three Angry Bear reader/commenters led by Dale Coberly).
It turns out there are two different possible sweetening agents, one being economic growth and the other pursuing better returns on investment (ROI). To understand how growth alone can save the day we need to back up and examine the three different economic models used by the Trustees to project solvency or actuarial deficit. These three ‘Alternatives’ are ‘Intermediate Cost (IC)’, ‘Low Cost (LC)’ and ‘High Cost (HC)’.
Intermediate Cost represents the mid-point of economic expectations and is backed up in the Reports by a variety of probability studies designed to prove it is a good faith effort. As such almost all economic reporting and most policy analysis simply assumes IC as their point of departure. For example the Northwest Plan explicitly assumes IC numbers even though some of the authors have private doubts about either the economic or demographic assumptions, doubts that by the way would drive the gap in different directions. Meaning that for this particular planning purpose it is perfectly reasonable to accept IC as a baseline set of assumptions.
On the other hand it is true that ‘Low Cost is Out There’. In operational terms Low Cost is rather simply defined: it is precisely that set of economic and demographic assumptions that would produce a fully funded Trust Fund through the 75 year actuarial period. Mind you this is not how the Trustees define it, instead they present it as a best case scenario hovering at the outside of the probability bound. But whether you accept that or not the numbers produce the outcome they do as seen in the following Figure in the 2012 Report where ‘I’ represents Low Cost:
Figure II.D6.—Long-Range OASDI Trust Fund Ratios Under Alternative Scenarios
Under Low Cost the Trust Fund Ratio dips perilously close to zero around 2075 but shows as slowly rising through the end of the projection period. As it turns out this result DOES NOT meet the Trustees’ test for ‘Long Term Actuarial Balance’ or ‘Sustainable Solvency’, that would require TF ratios never dropping below 100. On the other hand Low Cost WOULD deliver 100% of scheduled benefits with no changes in FICA tax rates.
The differences between Intermediate Cost and Low Cost (and the more pessimistic High Cost) Alternatives are set out in a series of six Tables from V.A1 to V.B2 showing selected demographic and economic assumptions for all three models. 2012 Report List of Tables The interactions between these numbers are complex and to some extent produce contradictory results, or example improvements in Real Wage serve to boost future income but also increase future cost in nominal terms. On the other those improvements also move the baseline, so that benefit cuts might be more or less in percentage terms but still produce a better result. But these complications can be hashed out in comments and future posts. I want to return to the Ditty and the Logic Trap.
If Low Cost numbers happen (they are by definition WITHIN the probability bound) then there will be no Trust Fund Depletion and so by the terms used here no Social Security Crisis. Meaning that neither a tax based fix or privatization would be necessary, we would have dodged the bullet.
But that bullet may have ended up lodged in the heart of privatization. Because Privatizers explicitly assume Intermediate Cost in their projections of ‘Crisis’, every single scary number they produce whether that be in terms of benefit cuts at depletion or ‘unfunded liability’ over 75 years or over the ‘Infinite Horizon’ derives directly from the specific economic and demographic numbers deployed by that particular model. Meaning that any different economic assumptions they insert arithmetically move those ultimate numbers away from IC projections. And to the precise point of this post the closer that any such numbers get to the supposedly improbable ones in Low Cost the less necessary privatization becomes as a means of solving ‘crisis’ AS DEFINED.
My contention here, and asserted in bare form so as to invite specific responses using real numbers is that privatizers can’t deliver. Not without using numbers that would fix Social Security along the way. So to parse the ditty:
“If Privatization is Necessary” meaning needed to avert ‘crisis’ and so benefit cuts at Depletion it has explicitly (although silently) endorsed Intermediate Cost economic assumptions. My assertion is that no privatization solution based on the spread between equities and bonds and any ancillary effects will work under the employment and wage assumptions of Intermediate Cost. Meaning “Privatization Won’t be Possible”.
On the other hand “If Privatization is Possible” meaning among other things requiring historical rates of return on equities, “It Won’t be Necessary”. Because any rate of GDP growth and wage and employment improvements needed to fund what is in the end worker funded retirement accounts starts bumping up against Low Cost numbers which as the Figures and Tables in the Reports show deliver 100% of scheduled benefits anyway.
Economist Dean Baker posed this question is slightly different form in Nov 2004 under the title ‘No Economist/Policy Analyst Left Behind’ Challenge (NELB) and was backed in that by Paul Krugman, all of which I blogged about right here back in 2008 with this AB post Double books and the ‘No Economist Left Behind’ Challenge
To my knowledge no one has met this challenge and few have tried. Those that have tend to rely on models that produce better ROI by suppressing Real Wage growth going forward and/or relying on returns from overseas investment. Which ignores the fact that worker funded retirement accounts have to be funded by workers, proving that the 1% can make out like bandits doesn’t translate to a privatization solution to the CRISIS AS DEFINED. And of course neither do current attempts to simply slash benefits starting immediately, not if we equate ‘crisis’ with ‘benefit cuts at TF depletion’.
So privatizers can use scare tactics based on benefit cuts or ‘no check for me’ but in order to avoid charges of being big, fricking liars they need to show that they can produce better results than IC at lower increased costs to workers than a simple payroll tax increase along NW Plan lines would require OR produce better returns than IC without resorting to economic assumptions that trend towards LC. Can’t be done. Or else prove it CAN be done.