Relevant and even prescient commentary on news, politics and the economy.

Social Security: the Shape of Solvency

2011 II_project_IID6 (1)
Well lets try again (two hours of draft having just disappeared)/
The above figure is II.D6 from the 2011 Social Security Report. The reasons why I didn’t use the 2013 version will become clear later, short version is it doesn’t show ‘The Shape of Solvency’ while the longer version is, well, long.

This figure shows the projections of three different economic and demographic models for Social Security expressed in terms of Trust Fund Ratio where 100 = next year’s projected cost. Now Trust Fund Ratio is in some ways an odd beast because it is to some extent after the fact yet also prospective. This is because it reports Trust Fund balances AFTER the calculation of current revenue minus current cost and does so in terms of NEXT year’s cost. For example it is possible for Trust Fund balances to increase in a given year and have Trust Fund Ratio decline, that is SS can be in surplus and yet weakening by the metric deployed by the Trustees. And we can discuss the arithmetic of this in comments, for now let me just assert that there is a good reason to measure Solvency in terms of Trust Fund ratios even in a system that operates on a Pay-Go basis.

The Trustees measure ‘solvency’ in terms of ‘actuarial balance’ over periods of time including ‘short term’ (10 years), ‘long term’ (75 years) and ‘infinite future’. Figure II.D6 shows a combination of a prospective ‘long term’ 75 years and twenty retrospective years for 95 years total. This 95 year period also happens to cover the entire lifespan of nearly everyone in the workforce today. The three graphed lines going forward represent three different economic and demographic models where I represents the projection of ‘Low Cost’, III the projection of ‘High Cost’, and II the projection of ‘Intermediate Cost’.

In this figure from the 2011 Report both High Cost and Intermediate Cost project to intersect the bottom line representing a Trust Fund Ratio of 0%. Which is to say a Trust Fund with no remaining principal. Now under a Pay-Go system where all or most benefits are paid from current revenue, or in SS terms where ‘revenue’ entirely or mostly meets ‘cost’, a Trust Fund with a zero balance/Trust Fund ratio doesn’t equate to “No check for you”. But it DOES mean no reserve funds to maintain Scheduled Benefits. Instead under current law the intersection of the graphed line of II and III would require a reset from Scheduled Benefits to Payable Benefits of greater or lesser severity. On the other hand if the graphed line NEVER hits zero then Scheduled Benefits can be paid in full and so ARE Payable Benefits.

But for very good reasons the Trustees are mandated to not allow things to get that close to disrupting the relation between Scheduled and Payable and instead have set up a test point that serve as the Canary in Social Security’s Coal Mine. And that test point is a Trust Fund Ratio of 100, or one year of next year’s cost for any given year tested. If that tested or projected ratio is set to stay above 100 for each year of the selected time window then Social Security is ‘solvent’. Which gets us back to our Figure II.D6.

In 2011 Figure II.D6 showed High Cost and Intermediate Cost not only busting downwards through the 100 mark and so failing the test for solvency in that year, but also hitting the zero bound of Trust Fund Depletion. On the other hand Low Cost or line I only nudges the 200 mark and then rebounds from there. Making this curve A shape of solvency. But many other curves would meet that test and in fact any such curve that exited the 75 year window between 100 and 200 would equally be A shape of solvency and in some important respects THE Shape of Solvency. Because odd as it may seem an overfunded Trust Fund would be a bad thing. On the other hand so preventable as to not even be a concern. Questions about that can be deferred until comments.

Backtracking a little bit, there are good reasons to call Intermediate Cost projections ‘realistic’ and Low Cost and High Cost ‘artificial’. But this does not have to do so much with a committment to ALL of Intermediate Cost projections as hitting the EXACT median point but instead to a methodology that has Low Cost having all its variables come in in a way positive for solvency while High Cost has them all coming in negative even where there is no reason a priori to believe that all dozen or so will move in the same direction. So while it is tempting to call Low Cost ‘optimistic’ and High Cost ‘pessimistic’ this is only true as narrowly applied to Social Security Solvency as defined. For example people living longer, healthier and happier lives is taken in isolation bad news for Social Security solvency. Yet most of us wouldn’t consider a wave of untreatable pneumonia that wiped out 90% of all residents of nursing homes overnight ‘Good News!’. But it would do wonders for Social Security Solvency and after the immediate medical expense for Medicare as well.

On the flip side most of the economic numbers that undergird Low Cost and nearer to Low Cost outcomes would be positive for society as a whole. For example lower levels of unemployment coupled with higher real wage would go a long way towards restoring Social Security to solvency even if that better resulting standard of living meant more workers living to be Great Grandpas and Grandmas. So we can with a clear conscience root for Full Employment while avoiding construction of Soylent Green facilities or ramping up Ice Floe manufacturing to launch Gramma on a one-way trip into the Artic.

Snark aside what this means is that we don’t have to commit without reservation to every particular projection that goes into Low Cost, just enough of them to bring the tail of Intermediate Cost first above a TF Ratio of 0% and then ideally to a level that never hits 100 from the upside. We DON’T need a tail going out through the 75 year window at the 300 level and trending up. On the other hand a curve that bottoms out at say 128 and stays steady from there would be almost ideal. Though to be safe one would want to have that tail tick up a bit. But there are any number of ways to accomplish that. One way is just to assume Intermediate Cost economic and demographic numbers and make up the difference on the revenue side. This is basically the position of the ‘Raise the Cap’ folks. Another would be to make a conscious and sustained effort to postively change those economic and demographic numbers of Intermediate Cost in a way that moves them towards Low Cost without resorting to putting Gramma on the Ice Floe. That is basically the little known position of the ‘Rosser-Webb’ folks. Or you could go for the ‘First Do No Harm’ position of baking revenue increases into the pie while having a mechanism to back off if ‘Rosser-Webb’ actually comes through.

Which is in a nutshell the meat of the Northwest Plan. It seeks to achieve as close to a perfect Shape of Solvency as possible by introducing the right changes to the formula at the right time in reaction to real time data as it comes in. The various authors of Northwest have varying opinions as to the merits of say Intermediate Cost vs Low Cost but in the end it matters not. Because the methodology adapts to the data and targets the Shape.

Tags: , , , Comments (23) | |

2013 Northwest Plan for a Real Social Security Fix

The Northwest Plan for a Real Social Security Fix was first introduced and revised in 2009. Obviously it was never adopted. Still the basic mechanism remains the same as does the rationale and working assumptions and in what follows I am going to assume a certain knowledge of the mechanics of Social Security finance and reporting. For those who get lost I can only suggest revisiting the 2009 version and all the posts surrounding it or just place your questions in comments.

 The NW Plan can be described in a variety of ways but here I want to present it as the answer to a question: “Under the Social Security Trustees Intermediate Cost Alternative and given current law Scheduled Benefits and Cap Formula what would be the minimally disruptive revenue only fix to deliver those benefits with no changes in retirement age?” Note the questions that are not being addressed: “Is Intermediate Cost a realistic mid-point projection?” “Are current Scheduled Benefits adequate? too generous? equitable?” “Why NOT adjust the Cap Formula?” While these are all important questions and ones the authors of the NW Plan have plenty of opinions on, they just are not the question at hand.

For the purposes of establishing a baseline for further discussion the NW Plan simply assumes IC and current law benefit and cap formulae, and also adopts the Trustees tests for ‘adequacy’ ‘solvency’ and ‘actuarial balance’. For the Trustees Social Security is in actuarial balance if it is projected under IC assumptions to end each year of the short term window and the last year of the long term window with an asset reserve equal to 100% or more of the next year’s projected cost. The short term window is 10 years and so coincides with the standard budget window used by CBO and OMB. Which in turn means that anytime Social Security is in ‘short term actuarial balance’ it properly has no NECESSARY role in budget talks. This BTW is doubly true if all proposed ‘fixes’ start operating outside that window.

 On the other hand the Trustees also take a longer view. Under their definition a ‘current participant’ in Social Security is anyone 15 and older and their long term window is 75 years. Meaning long enough that anyone who could reasonably be contributing today has their retirement interests taken into consideration for the entire working and retirement lives of all but the longest lived of todays teenagers. Who of course are for older workers their grandchildren. The NW Plan provides for ALL those cohorts. Details below.
The NW Plan starts by examining the following Table VI.F8.—Operations of the Combined OASI and DI Trust Funds, in Current Dollars, Calendar Years 2012-90 and fixing the point at which combined OASDI projects to fail the Short Term Test. By interpolation TF Assets will fall below Total Cost around 2027. But since the Test covers all years to then Year Ten test FAILURE projects to occur in 2018 and the NW Plan adopts the former year as the action point. This even though the combination of revenues and reserves in 2018 would still be sufficient to pay full scheduled benefits for 15 more years. Because that would be small comfort for the over 40s among us.

 The NW Plan takes this 2018 failure point and calculates a series of 0.1% FICA increases that would under IC assumptions keep 2027 and immediately following years in actuarial balance (Trust Fund ratios in excess of 100). And for good measure does the same for the entire long term window. Not because we have any real confidence in numbers outside the short to medium term but because once engaged on the exercise of addressing Short Term it is trivial to suggest numbers to address the largely theoretical gaps of the Long Term. One just has to take the numbers supplied by the Reports and mechanically insert some FICA rate changes.

 In short the NW Plan imposes a permanent (actually 75 year) fix given best available information in the Plan year. On the other hand that best available information changes with each Report year. But here is the key, by accepting a diagnostic that places ‘failure’ 13-15 years in advance of the real world consequences and starts the fix at the earlier date almost any conceivable change in outlook can be addressed by tweaking the FICA numbers at the tail, in almost all conceivable circumstances you have that 10-13 year lead time and an existing fix that at worst addresses the vast majority of the new projected gap. Which in most cases will not have actual incidence until 30 or 40 years down the road. In posts to come I will be putting up results of the new NW Plan spreadsheet based on the 2012 Report (the latest available). This post on the other hand is designed to let some of the conceptual objections be aired before diving into the actual numbers and assumptions.

Tags: , , Comments (17) | |