By this time most readers of the various venues of TPM know the basic outlines of Social Security 'Crisis': small problem decades in the future. But for those of a certain inclination the real interest is in the numbers behind the numbers. Because when studied in context they tell a different story than you think.

First thing to know about Social Security. The Trustees don't present one set of numbers, they present three. These are called Low Cost, Intermediate Cost, and High Cost. They are supposed to represent the probable range of outcomes with Intermediate Cost providing a median. The Trustees even supply a 'Stochastic Analysis' that purports to show that Intermediate Cost does represent the median, but really it does nothing of the sort, it never actually tests the underlying assumptions.

What an examination of the numbers actually shows is that Intermediate Cost consistently under projects growth and has for most years in the last decade. Moreover in the face of new numbers better than projected the tendency is to adjust future number years down with little to no motivation. This seems odd until it suddenly strikes you. Low Cost is Out There. The whole system of projections is actually being controlled by the Low Cost alternative.

To see this you have to examine the Reports in order. In each case since 1997 you will find that Low Cost returns the same operational result, it always results in a Social Security system that is fully funded over the standard 75 year window but never overfunded. Low Cost is always like Baby Bear's porridge in the Goldilocks story, never too hot and never too cold. While it is barely possible that the top amount of variation from a neutral projection always has the same result more or less by coincidence, mathematically such a result is pretty unlikely. Wielding Occam's Razor to slice away to the most likely cause of this phenomenon, you can only conclude that this is deliberate, that Low Cost is designed to give this same result year in and year out.

Now people familiar with the Reports will reliably reject this as paranoia. Well they need to face the fact that I made this claim on my own blog in 2004, and it got tested with the 2005, 2006, and 2007 Reports. And passed. Accidental or not what Low Cost really is is the answer to this question: "What set of economic and demographic assumptions would fully fund Social Security with no changes in payroll tax, retirement age or benefits?" Within this frame Low Cost becomes the 'Fix'.

What does this mean for the model as a whole? Well it automatically relabels Intermediate Cost as 'Crisis' and forces that model to be fixed at some point more pessimistic than Low Cost. Because if Crisis ever rises to Fix it is Game Over.

So back to the Reports. When you reexamine them through the spectacles of Fix and Crisis you see a constant tension between the models. Every year of better than projected growth serves to grind down Fix, numerically Solvency is just that much easier to achieve. To compensate for that requires removing more growth from the model in the out years and the better the current year numbers the more drastic that revision has to be. In really good years like the stretch from 1998 to 2000 this reaches perverse proportions, good news was just killing the models.

What does this mean for Intermediate Cost, aka Crisis? Well it means it has to be placed at some plausible level below Low Cost, if it is allowed to get too close to Fix then people are going to shrug shoulders and say 'Problem Solved. On the other hand if the numbers are so pessimistic as to be implausible to anyone then it's Game Over for privatizers anyway.

On close examination you can see that the models are under tremendous strain, Fix on balance getting ever cheaper, Crisis getting to be ever more based on pessimistic numbers. At some point the pressure will blow and Crisis and Fix will intersect. When will that be? Well in 2004 I predicted 2006, but the Productivity meltdown of late 2005 and now the Housing meltdown have postponed that, the bad news being good news for proponents of Crisis. But the numeric reality hasn't changed much, the economy still returned numbers at Intermediate Cost levels, Crisis was only able to tread water.

The key number to watch is Q4 Real GDP. If it comes in above 2.0% we beat Crisis. We won't beat Fix numbers, that would require a screaming 5.6% quarterly number, but we don't need to beat Fix in the current year, every time we beat Crisis it reduces the cost of Fix in the next Report. This grinding down process can be quick like it was in the late 90's, or it can more or less stall some as it did in 2004 and 2005, but absent multi-year recession it can't be reversed, not in the long run.

If Bernanke is right about Q4 2007 and the economy performs at the lower end of the Feds range then 2007 will be seen as another year of stall. But the numbers in the medium term are simply not that hard to reach. Low Cost is Out There. And he definitely has your number.

Actual numbers can be found at http://bruceweb.blogspot.com/2006/07/guide-to-bruce-webb.html