With the rising level of real yields available on TIPS, the rationale for including them in a fixed-income portfolio has become compelling. TIPS not only offer very attractive yields both historically and relative to nominal bonds, they also possess unique characteristics not found in other bonds. Their inclusion in a traditional fixed-income portfolio promises to improve overall efficiency and increase total returns.

TIPS offer compelling value.

At 4.4%, the real yield currently available on TIPS is very high compared to their initial yield three years ago. In a broad historical context, TIPS real yields fall near the high end of the range of observed real yields.

The next chart shows the reconstructed (ex-post facto) real yield on treasuries going back almost 30 years. Yields were only higher during the disinflationary 1980s and when the Fed was tightening aggressively to combat 5% inflation in the 1987-90 period.

The above chart reconstructs real corporate yields (to serve as a proxy for unavailable government yields) going back 125 years, and again they are near the high end of their range. Real yields were only higher during times of deflation (in the late 1800s, the early 1920s, and the 1930s) or sharp disinflation (the early 1980s).

Relative to alternative investments, TIPS are also quite attractive. The next chart shows the history of real yields on inflation-indexed bonds from Canada and the U.K. The highest real yield ever recorded on a government-issued inflation-indexed bond was 5.0%, in Canada. UK inflation-linked bonds have never traded with a real yield higher than 4.6%. Real yields have presumably been higher in Canada because of that market's limited liquidity ($9 billion outstanding) vs. the U.S. ($106 billion) and U.K. ($217 billion).

TIPS have cheap imbedded options.

One of the original rationales for the issuance of TIPS was to lower the Treasury's cost of borrowing. It was thought that investors would be willing to pay a premium to buy bonds with built-in inflation protection. That is, in exchange for this protection, they would be willing to accept a total yield (the sum of the real coupon and the inflation rate) somewhat lower than that offered by nominal bonds. Since TIPS’s running yield is currently greater than the yield on treasuries of comparable maturity (i.e., 4.4% real yield plus 2.7% CPI = 7.1%, vs. treasury yields of 6.8%), it would appear that Treasury's expectations have not been met, and that investors are unwilling to pay any premium for inflation protection. TIPS, in other words, carry with them a free option.

TIPS actually have two imbedded options that add to their valuation appeal. The first is obvious: the inflation call option. If inflation rises, TIPS will directly benefit since their principal, and the coupon which is paid on the appreciating principal, will rise in line with inflation. If inflation remains steady on the other hand, TIPS will deliver a running yield which is likely to be equal to or even somewhat in excess of current treasury yields. Inflation would have to decline from last year’s 2.7% pace to about 2.3% before TIPS returns would break even with treasury returns.

The second option imbedded in TIPS is the deflation put option. TIPS could suffer initially from a deflation because their principal would fall below par, but since they will be redeemed at par in the event of a sustained deflation, they would recover as they neared maturity. If held to maturity in a deflationary period, TIPS would deliver an effective real yield equal to their initial real interest rate plus the rate of realized deflation, less a possible reinvestment risk should real yields decline. In other words, their total real returns would be greater than their initial yield in times of deflation; this extra would be the payoff from the put option.

The next chart reminds us of the range and volatility of U.S. inflation in this past century. Each red bar represents the high and low mark for inflation in each decade. The yellow line represents the average for the decade, and the blue bar is the average plus or minus one standard deviation. The message is that current inflation is quite low historically, and exceptionally nonvolatile. This may well continue indefinitely, but if inflation were to again exhibit any of its past characteristics, TIPS would benefit handsomely. Yet these potentially valuable options can be had for free.

TIPS improve the risk/reward characteristics of a fixed-income portfolio.

The perfect diversifying asset is one that has low or negative correlation to other securities in a portfolio, low volatility, and equal or higher expected returns. TIPS fit this description nicely. To begin with, and as mentioned above, the running yield on TIPS currently is higher than that of comparable maturity treasuries.

As the next chart shows, TIPS have exhibited both positive and negative correlation to treasuries over the past three years. From the time of their issuance until October 1998, TIPS actually were negatively correlated to treasuries (–0.5). Subsequently, the correlation of TIPS and treasuries rose to 0.9. For the entire period since TIPS were first issued, their correlation to treasuries has been approximately zero.

Another attractive feature of TIPS is that they exhibit low volatility. The above chart illustrates the point: 10-year TIPS have had about one-half the yield volatility of 10-year treasuries over the past three years. This point is further emphasized by the total returns for TIPS and treasuries in 1999 in the table which follows.

Total Returns 1999
	TIPS	Treasuries
10 year	3.2%	-8.4%
30 year	-2.3%	-12.6%

As an added consideration, the future volatility of TIPS could be dampened by what appears to be limited downside risk. For one, U.S. government-guaranteed real yields are highly unlikely to move up indefinitely, whereas nominal yields could rise without theoretical limit in a pronounced inflationary environment. The experience of Canada and the UK suggests that 4.5 - 5% is probably the upper limit for TIPS yields. Plus, the inherent attractiveness of real yields at these levels should keep them from going much higher: for example, there has never been a 30-year period in history when U.S. investors could earn the 4.4% government-guaranteed annual real yield currently available on 30-yr TIPS. If nothing else, high real yields at some point should cool the economy, thus prompting the Fed to allow real yields to fall.

The challenge: what is the duration of TIPS?

TIPS are attractively priced, they have free imbedded options, they improve a portfolio's risk/reward characteristics, and they probably have limited downside risk. What more could a portfolio manager want? Unfortunately, the key characteristic missing is the duration of TIPS. We can use mathematics to calculate their modified duration (i.e., their price change with respect to changes in real yields). But how will TIPS behave as nominal yields change?

As the next chart illustrates, this is a very difficult question to answer. That's because the historical correlation between TIPS and treasuries actually has been extremely volatile. Without the ability to predict the correlation, we can't know the duration of TIPS. There are several ways to approach this dilemma, however, which bear exploring.

To begin with, there are some patterns which emerge from a study of the behavior of real yields. First, the correlation of TIPS and treasuries appears to be a function of changing perceptions of monetary policy. As the chart above shows, correlations were low and unstable in the period leading up to October 1998. This period, in turn, was characterized by market expectations that Fed policy would be either neutral or easing. Subsequently, expectations of Fed policy shifted to a tightening bias, and correlations rose and became more stable.

Second, the level of real yields appears to be heavily influenced by these same perceptions, as shown in the following chart. This makes sense intuitively, since the Fed can only tighten or ease by changing the level of real short-term rates; evidently, longer-term real rates are in turn heavily influenced by short-term real rates.

As a rule of thumb, we could therefore expect tighter monetary policy to both increase the correlation of TIPS to treasuries (and therefore increase their effective duration), and to increase the level of real yields. Conversely, neutral or easing Fed policy should lead to lower correlations, lower effective durations, and lower real yields.

As the next chart shows, the market's expectations for Fed policy are extremely pessimistic, having fully priced in a 125-basis-point rise in the Fed funds rate by year end. The combination of this pessimism and the fact that core inflation has fallen over the past several years thus appears to account for why TIPS yields have risen to current high levels.

These observations suggest that the duration of TIPS should follow from a portfolio manager's expectations of Fed policy going forward. If the manager anticipates that the Fed will tighten by more than is already expected by the market, then he should make the duration of TIPS more closely match the duration of comparable maturity treasuries, depending on his level of conviction. If he thinks that neutral or easier policy is the next likely development, then he should make the duration of TIPS closer to zero, again depending on his level of conviction.

An alternative method for assigning a duration to TIPS has been proposed by Rudolph-Shabinsky and Trainer of Sanford Bernstein & Co. They argue that the optimal duration has nothing to do with the observed duration of TIPS. Rather, it should be a function of the portfolio manager's rationale for including TIPS in a portfolio of nominal bonds. If the manager likes TIPS because he thinks real yields are attractive and likely to decline, then the appropriate duration should be zero. In this way, the addition of TIPS to the portfolio will not play a role in the portfolio's overall duration, and thus any decline or increase in TIPS yields will add or subtract directly to the portfolio’s performance regardless of the behavior of nominal bonds.

If, however, the manager likes TIPS because he thinks inflation expectations are likely to rise (which would be reflected by a widening of the spread between TIPS and treasury yields), then the appropriate duration is their modified duration. This will cause TIPS to displace nominal bonds of similar duration, and TIPS will thus add or subtract to performance depending on whether they outperform or underperform nominal bonds.

In the end, of course, it is virtually impossible to know the effective duration of TIPS when they are included in a portfolio of nominal bonds, and guessing wrong can mean unpredictable performance. If it weren't for this nagging problem, TIPS would be the proverbial free lunch.