Wednesday, December 08, 2004

Is the Social Security Trust Fund worth less than zero? Are the "savings" in it actually worsening budget deficits as we speak?

New empirical analysis says it may be so! But first some necessary background before we get to the main issue...

[If you already know all about the trust fund and the ways it may have economic value, or not, skip to the bottom.]

The Social Security trust fund is claimed by various people to help finance future Social Security obligations in two ways.

(1) The first, naïve claim -- though believed by a great many of the workers who pay the taxes that fund the trust fund -- is that the US government bonds held by the trust fund are assets the government can liquidate in the future to pay the costs of Social Security.

That is, the government will be able to sell them off to pay future retiree benefits, in lieu of collecting taxes or borrowing more money to pay the benefits -- just as you will be able to sell the assets you own in your IRA or 401(k) to fund your own future retirement, in lieu of having to keep working and earning a salary to live on.

But a moment of thought reveals this is not at all true.

The bonds in the trust fund are issued by the government to itself. So if one counts them as assets of the government, one must also count them as liabilities of the government in an identical amount. And since assets and liabilities of equal amount net to $0, that is the liquidation value of the bonds for financing Social Security: $0.

To grasp this simply, imagine that your IRA or 401(k) holds not stocks or bonds but only IOUs written by you to yourself. Go ahead and write yourself an IOU (or IO-ME) for $100,000 ... or $1 million. How much of your retirement expenses will it pay? If only it were that easy!

Alas, when that IOU to yourself matures what you collect from yourself on it will of course be exactly offset by what you pay to yourself on it, net: $0. Thus, you will end up continuing to be entirely dependent on your salary and other income to meet your expenses, just as before, as if the IOU never existed.

The situation is no different with the government's IOUs-to-itself in the trust fund. When it goes to liquidate the bonds for cash it will have to collect the bond proceeds from itself, net: $0. Thus, it will continue to be entirely dependent on raising taxes and increasing borrowing to finance its growing Social Security obligations, just as before, exactly as if the bonds never existed.

In future years, after about 2016, Social Security's expenditures on promised benefits will exceed its payroll tax revenue. In any such year it will thus run a deficit, let's call that amount $D. In order to pay all promised benefits the government will have to make up this shortfall...

* Without the Social Security trust fund the government would have to cover this shortfall using funds obtained from other general revenue sources (income taxes, new borrowing, cuts in other programs, etc.) in the amount $D.

* With the Social Security trust fund the government will cover the shortfall by redeeming bonds in the trust fund in exchange for cash in the amount of $D -- and to obtain this cash needed to redeem the bonds, it will have to use funds obtained from other general revenue sources (income taxes, new borrowing, cuts in other programs, etc.) in the amount $D.

Since $D = $D it is clear the trust fund makes no difference whatsoever in the funding of Social Security.

The government doesn't pretend anything different. To quote its Analytical Perspectives on the 2005 Budget (.pdf) on the subject of the trust fund bonds....

"At the time Social Security ... redeems these instruments to pay future benefits not covered by future income, the Treasury will have to turn to the public capital markets to raise the funds to finance the benefits just as if the trust funds had never existed. From the standpoint of overall Government finances, the trust funds do not reduce the future burden of financing Social Security."
[p. 199, my emphasis]

Those are the government's own words and they couldn't be plainer.

The trust fund bonds do not reduce the future burden of financing Social Security. And when the government turns to liquidating the trust fund bonds to pay Social Security benefits, it will have to pay for the benefits just as if the trust funds had never existed -- imposing tax increases, issuing new debt, and/or cutting other programs, exactly as if there was no trust fund.

OK -- since the trust fund does all of exactly zero to help finance Social Security this way, the question becomes whether or not it has helped finance Social Security in any other way.

(2) The second and more plausible claim is that the trust fund's "savings," as represented by the value of bonds in the fund, help finance Social Security indirectly by reducing the national debt. This is the actual claim that has been made all the way back to 1935 by proponents of trust fund financing.

The logic runs this way: In any given year, absent Social Security, the federal government will run a deficit (or surplus) of $X. This results in its increasing its borrowing by $X (or reducing its total borrowing by paying off debt of $X).

Now add Social Security to the picture and assume its tax receipts are exceeding expenditures by some amount-- a Social Security surplus -- with this surplus being "borrowed" by the Treasury's general operations account from its Social Security account and spent on general government operations: paying for paper clips, B-52s, agricultural subsidies, civil service salaries and whatnot.

Looking at the government as a whole, the argument goes, the surplus coming from Social Security is simply extra tax revenue for the government. As such it lets the government run a smaller deficit (or bigger surplus) than it would have otherwise. The government's total national debt as evidenced by bonds issued to the public thus is reduced by the total accumulated Social Security surplus.

No, the bonds in the trust funds indeed are not "assets" that can be liquidated to finance Social Security in the future in lieu of collecting taxes -- they instead are a tally of national savings, they show the amount by which the national debt has been reduced by the Social Security revenue surplus.

The benefit of these national savings is that they make it easier to finance Social Security in the future, in a number of ways. With the government having borrowed less, investors will have invested more in real economic investments, growing the economy larger. Less debt means less debt service and thus lower taxes, also growing the economy. A larger economy means it will be easier to collect more taxes in the future when they are needed for Social Security. And less debt will make it easier to increase debt in the future as well, if needed.

Thus, while the trust fund itself does not at all reduce the amount of future taxes that will have to be collected to finance Social Security compared to if there was no trust fund, the process of accumulating the surplus it represents will in the future make it easier for people to afford to pay those increased taxes. So the argument goes.

But ... very clearly, a condition must be met for this to be true: Congress must actually save the surplus payroll taxes to actually reduce the national debt -- and not simply consume them by increasing spending and/or reducing other taxes.

If in response to receiving $100 billion of tax revenue through a Social Security surplus Congress decides to curry favor among its constituents by increasing spending $50 billion on paper clips, B-52s, agricultural subsidies, civil service salaries, and whatnot, and also cutting income taxes by $50 billion, then the amount added to national savings is $0. The entire mechanism of surplus-and-trust fund then is worthless, no matter how many bonds eventually accumulate in the trust fund.

So which is it? Does Congress save the Social Security surplus or spend it? The argument has raged since the beginning of Social Security itself.

In 1937, at Social Security's very origin, Senator Arthur Vandenberg complained the surplus was already being spent...

"[I]n plain language, payroll taxes for Social Security ... have been used to render painless another billion of current government spending, while the old-age pension fund gets a promise to pay which another generation of our grandsons and granddaughters can reckon with decades hence. It is one of the slickest arrangements ever invented. It fits particularly well when the federal government is on a perpetual spending spree..."
Arthur Altmeyer, the first Commissioner of Social Security, responded...

"The Social Security contributor benefits [from paying increased taxes to fund a surplus] because future government expenditures will be proportionately lower. This is because the government debt in the hands of banks and other private investors will be that much less.... "
They were talking past each other. Vandenberg was saying the surplus was being spent, Altmeyer that it was being saved.

Sixty years later, people were still talking past each other. For example, economist Alan Blinder ...

"These [trust fund] cash-flow surpluses have raised national saving and investment and thus have helped to expand national income. They have bolstered our ability to pay Social Security benefits in the future…"
... and Senator Tom Daschle ...

"Right now we are using Social Security trust funds to pay for other spending in the federal budget, and that to me is unacceptable. That isn't what we should be doing ... No, there is no such fund per se."
With countless other examples. Which side is right? There was no definitive answer (although the number of politicians lined up saying they themselves were actually spending the surplus, from Vandenberg to Daschle, should I think have given any defender of trust-fund financing pause).

Opinions have been subjective and varied, with a "reasonable middle ground" splitting the difference: sure, Congress probably spent some of the SS surplus, but not all, so the trust fund probably represents national savings of some value.

And now, at long last, we get to the point of this post!

A new empirical analysis of the Social Security surplus and its historical effect on government spending ventures a conclusion on the value of the surplus. And it surprisingly lies outside what seemed the logical bounds until now: 100% of the value of the trust fund bonds on the upside, and $0 on the downside.

It is that the value of the accumulated Social Security surplus, represented by the Social Security trust fund, on national savings, is negative.

From Is the Social Security Trust Fund Worth Anything? (.pdf), by Kent Smetters, PhD., Wharton School and National Bureau of Economic Research:

~~ quote ~~

We find that there is no empirical evidence supporting the claim that trust fund assets have reduced the level of debt held by the public. In fact, the evidence suggests just the opposite: trust fund assets have probably increased the level of debt held by the public....

... each dollar of Social Security surplus appears to have actually increased the debt held by the public in the past by $1.76.

At first glance, this dramatic overspending of Social Security surpluses seems entirely implausible. However, the next section presents a simple game theory model that is consistent with this result...

We show how this counterintuitive result can be explained by a simple "split the dollar game" where competition between two political parties exploits the ignorance of voters who don’t understand that the government’s reported budget surplus actually includes the "off-budget" Social Security surplus...

The political game is really interesting but can't be explained here, you'll have to take a look at the paper. Prof. Smetters continues...


... going forward, the results herein suggest that a newer mechanism might be needed in order to prevent Social Security surpluses from being spent elsewhere.

In 1999, the Clinton Administration proposed the creation of a “lock box” whose purpose was to provide a better storage technology. The “lock box” was defined as “broken” if the on-budget surpluses ever turned negative.

However, this definition was somewhat naïve and, for example, could even rule out automatic stabilizers and other countercyclical actions during recessions -- fiscal policies that might have been undertaken even without the availability of Social Security surpluses. Whether Social Security surpluses are spent or saved does not depend on whether the on-budget surplus is negative or even positive.

Instead, this determination depends on how the on-budget surplus (positive or negative) changes in response to the presence of off-budget surpluses.

It is unlikely that a complicated budget rule that incorporated this feedback, as captured in the regression analysis herein, could ever be implemented.

Creating a more safe storage technology, therefore, might require the adoption of personal accounts that augment the current Social Security system...

~~ end quote ~~

Hey, there's $3.5 trillion more scheduled to go into the Social Security trust fund by 2016.

Stop the politicians from spending $6 trillion of it!

Put it in private accounts!!