Social Security Trustees and Funds

Joel writes:

Instead of relying on a phony source like Don Luskin for numbers about when Social Security will be in trouble, you ought to take a moment and look at the Social Security Trustees' actual report. If you did, you would notice that Luskin's numbers are completely phony. I don't know where he got them, but they do not represent what the Social Security financial reports say.

Joel is probably responding to this article where I said Bush's plan to reform Social Security was fascist. I think it's the only time (recently, anyway) that I've linked to anything of Luskin's.

I'm confused by Joel's letter. He's criticizing me for using Luskin's numbers, but Luskin was himself citing the Social Security Trustees:

According to the latest annual report of the Trustees of the Social Security Trust Funds, the surplus in 2004 was $64.4 billion dollars. It will be higher this year — at $87.7 billion. The surplus will keep getting bigger and bigger through 2008, when it will reach $108 billion. Each year, that's more and more money that the federal government won't have to raise from the world capital markets. It's a captive audience of bond buyers — and a growing one.

But in 2009, just 5 years from now, the surplus will start to shrink. In 2009 it will fall to $103.7 billion, and in that year the federal government will have to go to the capital markets to raise $4.3 billion that it didn't have to raise the year before. That's not a lot of money in the grand governmental scheme of things. But it's an important turning point for Social Security — it's the year the crisis begins.

If the Trustees' report is good, and Luskin's numbers are bad, but Luskin's numbers are from the Trustees' report… well, maybe Joel is the one who is confused.

This page of the Trustees' report has some lovely graphs that illustrate the problem. You can see for yourself in Figure II.D2 that 2009 is the year when the Social Security surplus begins to shrink. (Alan Greenspan, in his recent testimony before Congress, actually cited 2008 — but I'm not going to quibble over a single year.)

Luskin's numbers are unassailable, if you believe the Trustees' report. Only his assertion that there's a crisis could be arguable, but Joel curiously doesn't make that argument.

Back to Joel's letter:

Instead of spreading garbage in you column, check you facts instead of swallowing righ twing [sic] nonsense whole.

Stop right there. I take offense at the implication that I'm a right-winger. My preferred epithet is "procapitalist extremist." Right-wingers are emphatically not advocates of capitalism.

I always read Luskin with caution, and I have had some minor disagreements, but I do read him regularly. He's absolutely a better source than, say, Paul Krugman. [you're just inviting flames, you know - ed]

So what did the reports actually say? That income exceeds pay outs until 2028 and assets are not used up until 2042 and that Social Security can continue to pay out 69 to 75% of benefits, with no fix, in the 75th year depending on whether the percentage reduction begins in 2042 or in the 75th year.

No, the report does not say that income exceeds payouts until 2028. The year is 2018 — a decade earlier. See Figure II.D4 (same link as before).

The year 2042 is correct, but betrays a fundamental misunderstanding of the nature of the Social Security Trust Fund. Considering the Social Security system on its own, it is correct to say that it's funded until 2042. But the bonds in the trust fund are an obligation of the government, and Social Security is part of the government, so the bonds in the trust fund are the economic equivalent of owing money to oneself. The value of the bonds as assets to the Social Security system is exactly offset by the liability of the bonds to rest of the government. Considering the government as a whole, those bonds are worth exactly nothing. That's why the "lockbox" is such a howlingly funny joke.

If you "owe yourself" a thousand dollars, the only way to pay it is to earn a thousand dollars. You have to get that money from somewhere. The government's options are to tax, borrow, inflate, or spend less on other things — that's it. In 2018 it has to start diverting general revenue to Social Security. But I agree with Luskin's conclusion that 2009 is the year that really matters, because that's when Social Security stops contributing more and more every year to the general fund. That's the year when the government's budget deficit will, ceteris paribus, get worse every year automatically.

Incidentally, the roughly 30% cut in benefits forecast for 2042 should not be tossed aside lightly. Joel, if you had invested in a company that went bankrupt and could only get 70 cents on the dollar back, wouldn't you be upset? Do you think beneficiaries in 2042 will roll over and peacefully accept a 30% benefit cut? That's a lot of money.

UPDATE 2005-02-26 23:23:04 UTC: Joel responds:

First, your answer demonstrates that even if you go to the Social Security financial reports, finally, that you don't know how to read them. The 2018 number you reference is if you ignore income on the SS assets. Perhaps you could explain what a good deal private accounts would be if you ignored the income expected from them. The honest number is the one that compares income with pay outs and counts interest income. Gosh, if you want to make SS look like it is in really bad shape, why don't you just do your comparison on the assumption that SS income will be only half of what is expected?

Joel, I have a thought experiment for you. How would total government revenues and total government spending be impacted over the next forty years if the interest rate on the Social Security bonds was (a) 0%? (b) 10%?

If you understand that the answer is "total revenues and spending would be completely unaffected in either case", you understand the economic nature of the Social Security trust fund. Otherwise, you don't.

If you might better learn this lesson from satire, I recommend this. The Social Security system could be made "solvent" by increasing the interest rate on the bonds. But that doesn't help the government's financing problem in the slightest. It is this financing problem, involving the total government budget, that is the Social Security crisis. The crisis is not evident in the Social Security system when considered in isolation from the rest of the government.

If I spend $1000 but write myself a bond promising to pay myself back $1000 at 10% interest, that interest income doesn't help me at all. It's irrelevant. The interest, like the principal, is simultaneously an asset and a liability that exactly cancel each other. Simply put, the interest doesn't help me because I have to pay it!

This situation is inapplicable to private accounts because in that case the investment income is a transfer between two different people.

But instead of recognizing the facts, you suggest that I am just confused. Excuse me, but you are the one who is insulting. I hope your readers actually look at the financial reports that are readily available at the SS site. Some will certainly be smart enough to notice what you have overlooked.

Your confusion is that you attacked Luskin's numbers — you called them "completely phony" and that they "do not represent what the Social Security financial reports say" — but in fact they were the numbers from the Social Security Trustees' Report.

I wholeheartedly endorse Joel's recommendation that my readers look at the Trustees' Report. Indeed, I think these graphs make the situation very clear. The important dates — 2009, 2018, and 2042 — are all clear from Figure II.D2.

As for saying you are a right wing nut, look again at my note. That is not what I said. I said you should have checked your facts and not taken right wing garbage as fact. That still seems like a fair and relevant comment. What I did not anticipate was your inability to look at a graph and properly interpret it.

I do not think I am misinterpreting Figure II.D2 in any way. The Trustees' have done the numbers correctly by excluding interest income on the Social Security bonds in the calculation of the date when payments exceed revenues.