FROM THE OFFICE OF PUBLIC AFFAIRS

I. Introduction

It is a pleasure to be here with you tonight. It is particularly gratifying to have the opportunity to speak to representatives of and experts about an industry that will be at the leading edge in creating solutions to the demographic challenge of the retiring baby-boomers.

Back before the dawn of the modern financial era - that is, before the Treasury had issued price-level-indexed securities - I did the best I could as a low-level member of the staff at the Federal Reserve Board to agitate in favor of the Treasury taking this important step. As you know, in January 1997, the Treasury did in fact begin issuing these securities (though I strongly suspect that the correspondence between my agitation and the actions eventually taken by the Treasury Department was more coincidence than causal). Today, more than $100 billion of the price-level-indexed securities have been sold, with maturity dates ranging between 2002 and 2029. Moreover, I might note that the new inflation-protected I-bond accounts for about a third of our sales of savings bonds.

Nothing has changed my mind about the wisdom of Treasury having issued price-level-indexed securities. However, it is the case that one of the arguments that I among many others made in support of Treasury issuance of the new securities has not come true. In particular, I argued that Treasury issuance of price-level-indexed securities, by finally providing insurers with a means of hedging their price-level risk, would allow the introduction of a new class of retail products offering ironclad protection from inflation. Perhaps the most important of these new products, I speculated, would be price-level-indexed life annuities.

Nearly four years after the introduction of TIPS, it still hasn't happened. True enough, there have been one or two exceptions or near-exceptions - perhaps most notably the variable annuity offered by CREF, with a payout that is tied to the performance of a portfolio invested entirely in TIPS. This contract comes very close to providing payouts with constant purchasing power, though as I shall note below, the buyer response not exactly been overwhelming.

So what I would like to do tonight is to tell a detective story about the non-emergence of price-level-indexed securities, more or less in the manner of an Agatha Christie story. When John Shoven asked me to give this talk, I thought it might be difficult to come up with plausible explanations for this "crime." But I'm happy to report that the authorities have been able to round up a bevy of suspects. So I would like to parade these suspects before you, examine the evidence on each one, and see if we can convict a culprit.

Let me set the stage by being clear about exactly what the mystery is that we are investigating. The essential characteristic of the "missing" product that I have in mind is that the purchasing power of the payments it would provide would be unaffected by inflation surprises.

This is to be contrasted with - among other things - a fixed annuity in which the distributions to the policyholder are specified as a fixed number of nominal dollars regardless of their purchasing power, or a variable annuity in which the distributions depend on the investment performance of an underlying portfolio.

It is often claimed that a variable annuity tied to the performance of the stock market protects the annuitant against inflation risk, but history suggests otherwise. Brown, Mitchell, and Poterba, for example, find that from 1930 through 1997, the correlation between unanticipated inflation and unanticipated changes in nominal stock prices is actually negative. In other words, when the inflation news is bad, nominal stock prices tend to fall - just the opposite of what an inflation hedge should do.

Overall, the facts of annuities seem to be these: a goodly fraction of retirement savings is not annuitized, and most of the annuities that are purchased are in the form of either fixed or variable annuities. Almost none of it is in the form of purchasing-power-guaranteed annuities. At TIAA-CREF, for example, 41 percent of the amount in 1999 that was annuitized was taken in the form of nominal annuities; another 44 percent was taken in the form of variable annuities, 14 percent in the form of their so-called graded annuity, and a more 0.2 percent in the form of their inflation-protected annuity. And that is the puzzle.

It is interesting to note that real annuities are available in a few other countries, including Chile, Israel, Australia, and the UK. Chile and Israel have plenty of experience with high inflation, so it is not particularly surprising that their financial institutions might be more oriented toward dealing with inflation risk. And the UK began issuing index-linked gilt securities in the early 1980s, so its financial sector may be at a later stage of development in this respect than our own, and might possibly offer us a glimpse of our own future.

3. Rounding Up The Suspects

With that as background, then let us move to the trial phase of this enquiry. I proceed by calling each suspect to the dock.

A. All innovation takes time, and this one is no different

A first possibility is that innovation takes time in the financial world, just as it does in any other realm of life. By this hypothesis, in other words, "the check is in the mail," though probably not - as some clever person quipped - in the email. I should say that we ourselves are no strangers to lags in the development of new financial products. The State of Massachusetts created the first indexed debt in 1780. And less than three centuries later, the US Treasury followed suit! Anecdotally, about a decade passed in the UK before a deeper appreciation of the usefulness of index-linked gilts took firm root.

B. The non-availability of price-level-indexed corporate debt

A second possibility is that the introduction of Treasury inflation-protected securities was a good first step, but that the essential precursor for price-level-indexed annuities is the emergence of a meaningful market for corporate indexed debt.

The logic of this view runs as follows: The typical portfolio backing a conventional nominal annuity is mainly invested in corporate debt and mortgage-backed securities. Backing the annuity obligation with private securities allows the annuity provider to earn 100 or 150 basis points above the Treasury rate, and therefore to offer a richer annuity contract. By comparison, a conventional nominal annuity backed entirely by nominal Treasury securities of the same duration would appear to offer a poor return. And so, by analogy, would a price-level-indexed annuity backed entirely by TIPS.

C. The fiscal outlook

Another possibility is that potential providers of price-level-indexed annuities have been taking on board the tremendously favorable fiscal news of the last few years. The latest of those projections, released last summer, showed the debt held by the public being paid off altogether by 2012, potentially leaving would-be providers of price-level-indexed annuities with no means of hedging their inflation risk.

Let me offer two reasons for taking that projection seriously. First, the economic assumptions that underlie the projection remain conservative, as they have been throughout the Clinton Administration. Second, the Social Security surpluses alone will be large enough, according to our most recent projection, to get the job done by about 2015.

D. Nominal illusion

Turning to the demand side of the equation, a fourth admittedly prosaic possibility is that potential annuitants suffer from what economists refer to as "nominal illusion." Nominal illusion is a form of confusion that derives from an inability to evaluate dollar amounts, interest rates, and other financial magnitudes in inflation-adjusted terms. For example, when confronted with a choice between receiving a fixed nominal payment of, say, $1,000 per month for as long as they live, versus - say - $750 per month indexed to the price level, many people would choose the former option simply because it involves the higher initial payment. Nominal illusion will make indexed annuities a tough sell, and will require sustained effort to overcome.

E. Adverse selection

Another demand-side possibility - rather contrary in spirit to the preceding hypothesis - is that potential annuitants really are sophisticated, and understand that the payments on a real annuity will be more back-loaded than under a nominal annuity, and therefore subject the real annuity to even more adverse selection than the nominal one. The presence of greater adverse selection implies that, for any given potential annuitant, the real contract will look like a worse deal - leaving aside, of course, long-term price-level risk - than the nominal one.

And indeed, there is some international evidence of just this sort. For example, a variety of studies conclude that the "money's worth" of a typical real annuity in the UK is lower than the money's worth of a typical nominal annuity.

F. Absence of price-level risk

Another possibility that is often mentioned to me is that individuals simply don't believe that there is enough price-level risk to warrant the purchase of purchasing-power insurance. Let me be clear that, by "price-level risk," I mean to refer to surprises in either direction in the purchasing power of a dollar.

It seems to me that this "suspect" runs into two important difficulties. First, the relevant uncertainty pertains to surprises in the price level at very long forecast horizons - potentially as far ahead as 20 or 30 years. I don't know a single macroeconomic forecaster who believes that he or she can predict the price level at that horizon with any confidence. Therefore, it seems to me that the premise of low price-level risk - even given today's low inflation environment - is questionable at best.

Second, suppose we were to grant the premise of low long-term price-level risk. Then it seems to me we would be driven to the conclusion that insuring against that risk would be very cheap - not that we should take out the insurance. On the whole, I am skeptical that this explanation has much to say about the matter.

G. Tax effects

Lastly, there is the possibility that price-level-indexed securities might be treated less favorably by the tax code than their nominal cousins. While I am not in a position to deliver a definitive statement on this issue, at least as a first cut, I can say that if they ever issued inflation-protected annuities, insurance companies probably would issue them out of their general account, so that the profits from issuing the contracts would be taxed under the same regime that applies to so-called "fixed" annuities. There are no federal tax-law barriers in this treatment. There may be state law questions to be answered.

As I mentioned at the outset, when I agreed to give this talk, I worried for a while that I might not be able to identify any plausible suspects. Now, I feel a bit more like Hercule Poirot on the Orient Express, with a room full of potential perpetrators of the crime. As Poirot concluded, it may be that many suspects indeed have had a hand in the crime.

Here is what I take away from the trial we have just conducted: I now suspect that two further developments may be required for a market for price-level-indexed annuities to become well established.

• On the supply side, notwithstanding all the puzzles that it raises, I am inclined to think that corporations may have to begin issuing price-level-indexed debt, and individuals may have to borrow to a much greater degree through price-level mortgages.
• On the demand side, we are going to have to undertake a major public education effort in order to overcome nominal illusion in particular, and improve financial proficiency generally. After all, a major social marketing campaign was required to persuade the American public of the wisdom of wearing one's seatbelt while riding in a car or truck - and that in a context where the adverse consequences of failure to comply were all to graphically evident.

In this connection, I should note that the Treasury Department has been pleased to participate in the launch of a new organization called the National Partners for Financial Empowerment, or NPFE. The mission of the NPFE is to bolster and support the work of the many groups around the country that are promoting the importance of personal financial management to all Americans.

In that undertaking - the effort to educate the public about the wisdom of prudently providing for one's long-term financial security in retirement - I solicit your continued commitment to a cause that is essential to helping the nation prepare for the enormous demographic adjustments ahead.

I look forward to addressing this group again in ten years, long after a market for price-level-indexed annuities has taken root and flourished, and having the opportunity to talk with you then about how your clients finally have a wonderful new financial tool for making their lives in retirement more financially secure.