Friday, November 26, 2004

Boiled Down to Basics

Take the following two columns of Social Security years: what are the implications?

1996 2030
1997 2030
1998 2032
1999 2034
2000 2038
2001 2039
2002 2041
2003 2042
2004 2042
2005 2041

Okay, if you like you can ignore the stuff below, it was written before the release of the 2005 Report. But the recalculation to achieve 2041 required a couple of things, first some dirty work reconfiguring the mortality tables, second simple acceptance of a ridiculous growth number for 2005. Trust Fund exhaustion would have been pushed out under any realistic growth figure, 2.0 for 2005 is a joke.

The number on the left is easy, it is the Report Year of the Annual Reports of the Trustees of Social Security. But consider the second column. Something is being moved out twelve years, moreover it is moving out a rate of more than a year per year. If this number moves over the next nine years like it did the last we would expect it to be something like:

2013 2056 and the nine after that
2022 2068

In reality the second column represents the dates of exhaustion of the Social Security Trust fund under the Intermediate Cost alternative (the standard one reported in all coverage).

Without delving into the numbers and the reasons for the change, we can see that the outlook for the Trust Fund, in the absense of any reform at all, is improving nonetheless at an average of 1.33 years/year. In 2068 the youngest Boomer will be 104, the Trust Fund will have done the job assigned in 1982, it will have successfully handled the demographic bulge of the Baby Boomers.

It seems to be that it is up to the Privatizers to identify why this number is moving out in the way it has been, and make a case why it won't continue to improve. That will require actually grappling with the various Economic Assumptions under the Three Alternatives and examining their impact on Trust Fund health going forward Trust Fund Ratios under the Three Alternatives. And most importantly producing the economic projections underlying their own privatization models.

If Social Security Privatization is Necessary, it won't be Possible. If Social Security Privatization is Possible, it won't be Necessary.

Links a-plenty from the main page The Bruce Web or from the Intro page Social Security is not Broke: by the numbers

Thursday, November 25, 2004

Social Security is not broke: by the numbers

This site aims to put the debate over Social Security privatization firmly on a numeric basis. There is a lot of rhetoric flowing, but little focus on the economic assumptions underlying the talk of the Trust Fund going broke in some future year, "requiring" some effort to fix the system. Is it really "broke"? Does it even need a fix? The numbers suggest not.

First the numbers. Everything starts with the Reports of the Trustees of Social Security. Every number or factoid ever written or posted comes directly or indirectly from the Annual Reports. What I have done here is to provide links to the HTML versions of the reports from 1995 to 2005 and broken out the key tables and figures. I have also included a link to PDFs of reports from 1942 to 1994. The following represents a diagram of this site with links. As I add new essays I will put up a new link.

New Intro Social Security - Is it about solvency?

Trust Fund Exhaustion: a Personal Journey

Payroll tax vs productivity: what would it take

Trust Fund Ratio explained

The Three Alternatives

Economic Assumptions under the Three Alternatives

Trust Fund Ratios under the Three Alternatives

The Lost Cost Alternative: What does it mean?


The Reports
2005
2004
2003
2002
2001
2000
1999
1998
1997
1942 to 1996

Trust Fund Exhaustion: a Personal Odyssey

Who the hell am I to be lecturing economists on Social Security? Well let's take a visit down memory lane.

In the 1980's the accepted figure for Trust Fund Exhaustion was 2023, a date I found amusingly ironic as it coincided with the year that I would be eligible for full retirement under the reformed system: in my case 66 years and 4 months. But it fixed the date in my mind.

In 1993 I was watching some show or another or browsing through some paper and someone commented that the Trust Fund was due to run out in 2026. Had I misremembered the earlier date? No, the date had been officially changed. Which taught me Lesson One of the Social Security Trust Fund: The exhaustion date is not fixed in stone, it moves with the economy.

In 1996 I was able to get back on line and started actively seeking information on Trust Fund exhaustion. I came across articles like this one from July/August 1995 (the date may be a typo, the article refers to a Dec 1995 article in LBO) The Myth of Social Security's Imminent Collapse Which taught me Lesson Two of the Social Security Trust Fund: The exhaustion date is not determined independently. Everyone is drawing from the same set of numbers: (link to) The Annual Reports of the Trustees of Social Security. And by the way, the exhaustion date was now set at 2030. Which taught me Lesson Three: the exhaustion date can move out fairly quickly, in this case 4 years over a 3 year period. At that rate the Trust Fund would never run out.

At this point I hear the murmering "No you fool, the Boomers are going to start to retire putting impossible stress on the Fund". Well no, the Annual Reports are fully aware of this impending demographic bulge, that in large part is what they are attempting to address, and every date and factoid you have ever seen is drawn from their pages. God is not making more Boomers, the demographic problem is fixed in stone, all of these forecasts are working with the same pool of retirees.

So lets look at some more numbers. Every year the Trustees present The Three Economic Alternatives and based on those projections give us some key dates and numbers. I want to focus on two. First the date of Trust Fund exhaustion under the Intermediate Cost Alternative if we do nothing, and second the amount of immediate payroll tax increase to fix it. But first a word from the other side.

From our good friends at the Cato Institute from June of 1998 The Myth of the 2.2 Percent Solution
"The 2.2 percent figure assumes that the reform would take effect at the beginning of calendar year 1998. Thus it is already out of date. Each year Congress waits, the magnitude of the tax hike required to balance the Social Security trust funds rises. "
This argument is followed by a series of very dubious assertions, based on from what I can see nothing at all. But let's test this particular assumption: 'by doing nothing the magnitude of the tax hike would have to rise'.

Report Year / Trust Fund Exhaustion Date / Amount of Payroll Gap (and so tax increase needed)
1996 / 2030 / 2.19%
1997 / 2030 / 2.23%
1998 / 2032 / 2.19%
1999 / 2034 / 2.07%
2000 / 2037 / 1.89%
2001 / 2038 / 1.86%
2002 / 2041 / 1.87%
2003 / 2042 / 1.92%
2004 / 2042 / 1.89%

These numbers can be found in each reports "Overview" section. 2004 Report: Overview. Graphic representations of the Trust Fund ratio (zero equalling exhaustion) can be found at Trust Fund Ratios under the Three Alternatives

Now we can argue the numbers in any given year, and the motives of the Trustees (and I will), but a look at these numbers shows that the central premise of the Cato Institute is flat out wrong. We did nothing to address the gap. If they had been correct the exhaustion date should have stayed steady and the amount of tax needed should have gone up. Leaving 2.19% of my paycheck in my pocket every year from 1998 to date instead of devoting it to the Trust Fund should have left that Fund relatively worse off. It didn't. Actually the reverse is true. By doing nothing Congress left thousands of dollars for me to spend or invest over that span, and now the Trustees are reporting that not only did they not need that money after all, they actually need a smaller percentage going forward.

There is a world for a problem that left untreated retreats into the distance and requires an ever shrinking fix. That word is not "Crisis".
(And though it may look like progress is stalled, that has more to do with the increasingly gloomy numbers posted in the 2004 report. As the real numbers for 2004 are reported, and a more realistic number for 2005 is inserted, that 2042 date should lurch forward 2004: Economic Assumptions).

Saturday, November 20, 2004

Payroll tax vs productivity: what would it take.

Can we tax ourselves or grow the economy in such a way that we don't need to "fix" Social Security? Any reader of the Op-Ed pages over the last fifteen years or so would say "Of course not". But let's take a look at the numbers, let us establish the parameters here. Let us ask the two key questions. One, how much would we have to raise payroll tax today to solve the whole problem even given the fairly dire predictions of the Intermediate Cost alternative (the standard model, see below)? Two, if we do nothing at all what is the rate of economic growth that would be required to save the day anyway? From the Reports:



1996- 1997- 1998- 1999- 2000- 2001- 2002- 2003- 2004- 2005- Report Year
2.19% 2.23% 2.19% 2.07% 1.87% 1.86% 1.87% 1.92% 1.89% 1.92% Payroll tax increase (Intermediate Cost)

2.1% 2.2% 2.2% 2.1-2.2% 2.2-2.4% 1.8% 1.9% 1.9% 1.9% 1.9% Productivity Growth in out years to fix (Low Cost)


What this tells me is that we can take the Trustees' Reports at face value and still conclude there is no crisis here. Inaction shoud have led to an increase in the fix or required a higher level of economic performance to save the day. That 2.19% of payroll that didn't flow into the Trust Fund since 1998 is still in my pocket, and now they are telling me that not only didn't they need it, they only need 1.92% going forward. Up from the 1.89% in the 2004 Report, but then again abandoning all pretense at projecting realistic growth for 2005.

A crisis that left unaddressed requires a flat line fix is just not a crisis.

Trust Fund Ratio explained

Given the numbers of people directly affected by it, either as tax payer or retiree, the Social Security Trust Fund may be the most misunderstood of all major economic instruments.

But it is simple really. It is a checking account. Payroll checks are paid in, disability and retirement checks are paid out. Like any checking account as long as you have money to cover the checks everything is cool. And for most of its history it was truly pay as you go: between the years between 1957 and 1985 year end balance never went below $19.8B and never went above $45.9 billion. What it is not, and never has been is a Savings account.

You do have a Retirement & Disability account with Social Security, it just has the most indirect relation with the Trust Fund. Your retirement and disability checks are indeed computed based on your lifetime earnings and contributions, but your money is not working for you, it went out the door to cover others retirement checks.

In 1982 the country as a whole looked at cash flow to this giant Checking Account. The largest demographic cohort in US history, the Boomers born between 1946 and 1965, were coming into prime earning years, but literally that could not last forever, by 2008 some would start becoming elibible for partial retirement and by 2030 or so all surviving members of the cohort would likely be drawing benefits and it appeared there would not be enough workers coming up behind to comfortably cover the checks. So it was decided to increase the payroll tax and set the money aside in interest earning bonds, hoping to use that money as a bridge loan.

This required a great leap of faith, that the government would have the money to redeem the bonds when the time came, that they would use this new infusion of cash in a responsible way, say by paying down other debt. That didn't happen under Reagan and Bush, they spent all this new money and built up huge public debt besides. But then again the big money had yet to flow to the system, the entire accumulated balance only hit $109.8 billion by year-end 1998.

Trust Fund Ratio. This is the Trust Fund balance expressed as numbers of years of reserve to pay beneficiaries. In 1988 this ratio was 41, which translates to 4.92 months of reserve, by 1992 it hit 96 or almost a year of reserve. But then due to a combination of factors the money started rolling in. By 1995 we had $491.6 billion or a ratio of 1.28 (1 year 2 months), by the end of 2003 we had tripled that to $1.5 trillion and a Trust Fund Ratio of 288 (almost 3 years).

Now the 1996 Report had not predicted this, or at least not directly.
Trust Fund Ratios under the Three Alternatives
The standard numbers, the Intermediate Cost alternative (II) had predicted it would top out at 240, but not before 2010, and then sink to exhaustion by 2030. But the more optimistic Low Cost Alternative ( I) produced a ratio very close to this one, and interestingly showed the Trust Fund ratio soaring to 400 by 2010 and then never dipping below that ratio. In Trustees' jargon this meant it was in "long-term actuarial balance". In my terminology this would mean that Social Security was "fixed", and fixed to the point that it would not be necessary to redeem the actual bonds, though it might be necessary for the General Fund to pay some cash interest into the system.

As we progress through the Trust Fund Ratio figure year by year Trust Fund Ratios we get a graph with a consistent shape, but in each case with a newer, higher starting point corresponding to the actual number scored in the year at hand. By the time we get to the 2004 Report Trust Fund Ratios we have our exhaustion dates for the High and Intermediate Cost pushed out pretty far, and the Low Cost a Trust Fund Ratio practically in the stratosphere.

We can draw two conclusions from this immediately. One, any year we beat the Intermediate Cost alternative the exhaustion date gets pushed back. Two, if we consistently hit or beat the numbers of the Low Cost alternative, our problem is solved.

But there is a third conclusion we can reach, one that moves us from the realm of the economic to that of the political. What is the real meaning of Low Cost?

The Three Alternatives

To understand Social Security you have to understand the three Alternatives: the Low Cost, the Intermediate Cost, and the High Cost. Each is a set of economic and demographic assumptions that purport to show "a range of possible outcomes" with the Intermediate set "designated as alternative II, reflects the Trustees' best estimate of the trust funds' future financial outlook." (2004 Report p.1). Note that there is no suggestion that this estimate is conservative, biased by the natural caution of actuaries, this is supposed to be the midpoint estimate, attested to by the political masters of Social Security.

Interestingly the Intermediate Cost Alternative is the only one that is ever cited in the media, and is generally presented as a firm prediction: The Trust Fund "will" start running out in year X and will be totally depleted by year Y. There is nary a suggestion that it is just the mid-point of a range and that the Trustee's allow that the economy might outperform those numbers and no conversation about the future state of the Trust Fund if it does.

The biggest driver for the Trust Fund is Productivity Growth. If the economy grows more than projected the date of Trust Fund exhaustion gets pushed back. Moreover early years are more important than later years. And now the data. (2004 Report p.88) Economic Assumptions

Productivity Growth Under the Intermediate Cost Alternative
2004 2.7% 2005 1.8% 2006 1.9% 2007 1.9% 2008 1.8% 2009 1.8% 2010 1.7% 2011 1.7% 2012 and years forward 1.6%.
At first glance these are pretty pessimistic, at second glance they are ridiculous. The best guess of the Trustees is that productivity will shrink from the 3.4% it achieved in 2003 to an ultimate rate half of that? Is there a single other model out there that shares these numbers? (And a little challenge to would be privatizers: how are you going to return 7% on stocks with these anemic economic numbers.)

Productivity Growth Under the Low Cost Alternative
2004 2.8% 2005 2.1% 2006 2.2% 2007 2.2% 2008 2.1% 2009 2.0% 2010 2.0% 2011 1.9% 2012 and years forward 1.9%
For guys that are pushing Tax Cuts as the engine to drive a roaring economy, this belief that 1.9% is the best we can expect longterm seems to go against every other prediction I have seen. We can do better than that, though as we will see, to save Social Security we don't have to.

And no matter how you slice it, we already beat the 2004 number in both models.

Economic Assumptions under the Three Alternatives

(Author's note. In February 2006, the Social Security Administration took down the HTML versions of the 1997 to 2000 Reports. Those links will no longer work. The Reports are still available in PDF form and it is possible that we will get them mirrored somewhere else on the web, at which point I will fix the links.)
The tables linked to here show the economic projections under all three alternatives from the 1996 Report to the 2005 Report. Note how the numbers projected for the out years (after 2010) change over time. Why does the crystal ball get increasingly cloudy?

2005 Report
2004 Report
2003 Report
2002 Report
2001 Report
2000 Report
1999 Report
1998 Report
1997 Report
1996 Report

Trust Fund Ratios under the Three Alternatives

((Author's note. In February 2006, the Social Security Administration took down the HTML versions of the 1997 to 2000 Reports. Those links will no longer work. The Reports are still available in PDF form and it is possible that we will get them mirrored somewhere else on the web, at which point I will fix the links.)
The Trust Fund Ratio is the measure by which the Trustees measure the health of the Trust Fund. It represents the amount of years that the Trust Fund could fund Social Security without any other income. Given that there will always be some income flowing in in the form of payroll taxes any persistant positive Trust Fund ratio means that Social Security is "fixed" under that projection. Contrawise just because the Trust Fund ratio goes to zero doesn't mean checks stop. It just means that some external form of funding must be found to continue benefits at the projected level.

What we have here are graphic representations of the Trust Fund ratio under the three alternatives from all Reports 1996 to present.

2005
2004
2003
2002
2001
2000
1999
1998
1997
1996

Two important things to note here. First alternative I, the Low Cost Alternative always, magically, results in a positive yet flat Trust Fund ratio. That is it shows the classic Goldilocks scenario: Social Security funded with a steady level of reserve. Not too hot and not too cold. Too hot is unthinkable, the notion that we may actually be overfunding Social Security would require too drastic a narrative change. Too cold is left to the Intermediate and High Cost alternatives. Second the date of Trust Fund exhaustion under the Intermediate and High Cost alternatives is steadily being pushed out. The 1996 Report shows it running dry in 2030 under the Intermediate Cost alternative, the 2004 Report sometime in 2041. (The 2005 Report shows it being pulled back by about five months into 2040, the result of simple surrender on 2005 numbers, see discussion at that post).

There is a word for a problem that improves itself eleven years over a nine year period after being neglected. That word is not "Crisis".

What is the Low Cost alternative? What does it mean?

(Author's note. In February 2006, the Social Security Administration took down the HTML versions of the 1997 to 2000 Reports. Those links will no longer work. The Reports are still available in PDF form and it is possible that we will get them mirrored somewhere else on the web, at which point I will fix the links.)
Well there are two answers to that. One is to take the Trustees' at their word that that is just about the best the economy can reasonably be expected to do. The other is to be more empirical and examine that number over time and to see what stays consistent within the model. And the answer is obvious: outcome. Let us take a look at all reports from 1996 to 2005.

As we click through the Reports upwards from 1996 (it may take a couple of seconds to actually move to the graphic from the top of he loaded page) we find a graph that has a consistent shape. High Cost ( III ) and Intermediate Cost ( II ) alternatives that show shortfall dates being pushed out, and a Low Cost ( I ) that always shows us a positive, yet flat Trust Fund Ratio.

Yes in each case the Low Cost alternative results in a flat line, that is Social Security funded, and with a consistent level of reserve. But what you never find is a projection that would result in an ever growing Trust Fund Ratio.

And now a look at the Economic projections in these respective reports. From which I have extracted the following series representing growth in the out years:

1996- 1997- 1998- 1999-- 2000-- 2001- 2002- 2003- 2004- 2005- Report Year

2.1% 2.2% 2.2% 2.1-2.2% 2.2-2.4% 1.8% 1.9% 1.9% 1.9% 1.9% Productivity Growth in out years to fix (Low Cost)

What we have is a curious suppression of growth projected in the out years. Given that the economy was consistently hitting the high end of the forecast, why ratchet down this number? A look back at the Economic Projections for 2004 suggests an answer. A difference of .2% or .3% can result in huge changes in Trust Fund Ratio. If the Trustees substituted the numbers considered perfectly reasonable in 1996 into the charts for the 2005 the result would have been a substantially overfunded Trust Fund.

That would not fit the Privatization agenda at all.

Sunday, November 14, 2004

2004 Report

Entry page
Table of Contents
List of Tables
List of Figures

Economic Assumptions under the Three Alternatives
Trust Fund Ratio under the Three Alternatives

2003 Report

Entry page
Table of Contents
List of Tables
List of Figures

Economic Assumptions under the Three Alternatives
Trust Fund Ratios under the Three Alternatives

2002 Report

Entry page
Table of Contents
List of Tables
List of Figures

Economic Assumptions under the Three Alternatives
Trust Fund Ratios under the Three Alternatives

2001 Report

Entry page
Table of Contents
List of Tables
List of Figures

Economic Assumptions under the Three Alternatives
Trust Fund Ratios under the Three Alternatives

2000 Report

(Author's note. In February 2006, the Social Security Administration took down the HTML versions of the 1997 to 2000 Reports. These links will no longer work. The Reports are still available in PDF form and it is possible that we will get them mirrored somewhere else on the web, at which point I will fix the links.)
Entry page
Table of Contents
List of Tables
List of Figures

Economic Assumptions under the Three Alternatives
Trust Fund Ratios under the Three Alternatives

1999 Report

(Author's note. In February 2006, the Social Security Administration took down the HTML versions of the 1997 to 2000 Reports. These links will no longer work. The Reports are still available in PDF form and it is possible that we will get them mirrored somewhere else on the web, at which point I will fix the links.)
Entry page
Table of Contents
List of Tables
List of Figures

Economic Assumptions under the Three Alternatives
Trust Fund Ratios under the Three Alternatives

1998 Report

(Author's note. In February 2006, the Social Security Administration took down the HTML versions of the 1997 to 2000 Reports. These links will no longer work. The Reports are still available in PDF form and it is possible that we will get them mirrored somewhere else on the web, at which point I will fix the links.)
Entry page
Table of Contents
List of Tables
List of Figures

Economic Assumptions under the Three Alternatives
Trust Fund Ratios under the Three Assumptions

1997 Report

Entry page
Table of Contents
List of Tables
List of Figures

Economic Assumptions under the Three Alternatives
Trust Fund Ratios under the Three Alternatives

1942 to 1996 Reports

1996 Report

1995 Report

1942-1994 (PDF format)