Thursday, March 03, 2005

Social Security and interest rates.

Interest rates are very important to the whole debate about Social Security.

Two new, interesting things to note:

#1: Alan Greenspan testified before Congress again yesterday ...

Greenspan again endorsed the key part of President Bush's Social Security overhaul to set up private accounts .... The administration estimates those accounts will require about $745 billion in new borrowing over the next decade .... He said it is entirely possible that the impact on interest rates will be "zero,"... [AP]
How can $745 billion in borrowing over 10 years have zero effect on interest rates?

First, $75 billion a year isn't a whole lot by historical standards. Added to current projections, in GDP terms it still leaves debt a good deal lower than in the Reagan and following years.

Second, and maybe more to the point, it is not new debt. The government has already promised to pay all these unfunded Social Security benefits and is on the hook for them -- so funding them as it has to do sooner or later doesn't add any new debt for it any more than your deciding to pre-pay an obligation you owe adds to your debt, or any more GM's borrowing last year at low rates to pay off its unfunded pension obligations added to its total debt.

As Greenspan explained on an earlier occasion,

A critical consideration for the privatization of social security is how financial markets are factoring in the implicit unfunded liability of the current system in setting long-term interest rates.

If markets perceive that this liability has the same status as explicit federal debt, then one must presume that interest rates have already fully adjusted to the implicit contingent liability. (emphasis added)
In other words, since the debt already exists why should simply admitting it affect interest rates?

The fact is that General Motors was lauded for strengthening its balance sheet by borrowing at low rates currently to fund its previously unfunded pension liabilities. Why would the situation with Social Security be any different?

In any event, Greenspan's statement sure is a far cry from Krugman's screech about how all this immense, new debt will be ruinous -- and from Kinsley's bizarre theory (endorsed by Krugman) that this $75 billion a year (compared to stock market capitalization of over $33 trillion) will somehow result in the difference in return between stocks and bonds being bid away(!).

#2: Jeremy Siegel and Bill Gross are both cited as saying the 3% real return on Treasury bonds that is projected for the future by the Social Security actuaries is far too high, in an article in the LA Times.

Prof. Siegel of Wharton, the authority on the last 200 years of stock and bond investment returns, says, "Three percent is way too high", and notes that the real return on Treasury bonds from 1946 through 2004 has been only 1.6%.

Mr. Gross, who has become known as the "Warren Buffett of bonds" running the investment portfolio at Pimco, says that to get a 3% real return on bonds "you would have to invest in Mexico or Russia". The current real yield on US bonds is about 1.3%.

This is important because a too-high estimate of future bond returns makes private accounts look worse than they really are, makes the finances of Social Security look better than they really are, and makes the government's unfinanced future liability for entitlements look much smaller than it really is (and it looks plenty huge even using the 3% rate to discount it.)

If Siegel and Gross are right, then not only are private accounts a much better deal than they appear using the actuaries' and White House's numbers, but it is also more important to start a reform like this (among others) now because the coming fiscal crunch of future decades is going to be a lot worse than it looks today using the actuaries' numbers.

More explanation of the difference that a too-high interest rate makes in regard to the White House's specific proposal for private accounts is in an earlier post -- and a fuller explanation of how such an interest rate skews projections for Social Security, and of the current value of future government liabilities, is in a Tech Central Station column by Arnold Kling.

But there is an obvious question: why did the actuaries pick an interest rate that is so high?

In the same LA Times story the Chief Actuary explains....
Stephen Goss, the chief actuary at the Social Security Administration in Washington, said he considered 3% "a reasonable place to be" given that the real return on government bonds since 1980 has been more than twice that number.
But my gosh, what sense does that make? Talk about picking your starting point to skew your end result!

That starting year, 1980 was the very height of the oil-crisis inflation. Nominal interest rates were at an all-time high -- the rate on the 30-year bond went over 14% in 1981.

Return on bonds consists of two parts: the real interest rate (the nominal interest rate paid by the bond minus inflation) and price appreciation as the nominal market interest rate falls (or depreciation as the rate rises). Bonds bought around 1981 had the highest nominal interest rate ever, and had an extra-high real rate as an "inflation risk premium".

If you bought 30-year bonds in 1981, locked in that extra high real rate, then also took all that appeciation in the bonds' value as the nominal market interest rate fell from 14% to 4%, of course you've had great returns all the way through to today!

But now nominal interest rates are at their lowest since the 1960s. They can only go up -- meaning bonds lose capital value, offsetting the interest paid on them. Which is why Siegel says he is pessimistic on bonds and foresees 0% return on them for some years ahead.

Yet that 3% future return on bonds projected by the actuaries is estimated by looking at what happened since 1980, when what happened since then is impossible from now on. And I do mean impossible -- for it is literally true that we may not live to see nominal rates this low again, in which case bonds can only depreciate in capital value from today's interest rate levels.

That being the case, returns going forward should logically be estimated at below the historical average (and it took those huge returns from appreciation since 1980 to get the average return since 1946 up to 1.6%!) other things being equal.

Of course, not all other things will be equal, and as Arnold Kling points out in the same column mentioned above, one of those things expected to be unequal in the future is a projected GDP growth rate that drops from the long-term average of about 3% to about only 2% as the boomers start retiring (due to the decline in the growth rate of the work force). And this decline in GDP growth rate implies a lower real interest rate for the future.

So there are two reasons to believe returns from bonds will be lower in future years than in the past -- depreciation in bond values as nominal rates rise from today, and a lower future real interest rate -- yet the actuaries project future returns from bonds to be almost twice as high as they were over the last 58 years!

I don't know what more to say about that.

But even if returns from bonds are projected at just the post-WWII historical average, the government's finances in general and Social Security's in particular are much worse than currently reported -- and the economics of private accounts are much better.