US TIPS market suffered significant losses during post-

http://www.ssga.com/library/povw/Inflation_Linked_Bond_Attractive_Mauro_Tarlow_4.22.09CCRI1242256443.pdf

SSgA CAPITAL INSIGHTS viewpoints
Inflation-Linked Bonds Appear
Attractive on a Breakeven Basis
Recent performance in the inflation-linked market suggests that
the asset class offers favorable relative value.1 In this paper, we
demonstrate that inflation-linked sectors are sensitive to liquidity
risk and that we are currently in a period of increased liquidity risk
aversion. The implication is that breakeven inflation hurdles are
artificially low. We conclude by describing strategies which could
help gain exposure to inflation-linked assets.
Inflation-Linked Bond Observations
Broad scale de-levering after the collapse of Lehman Brothers in
September 2008 catalyzed forced selling across fixed income,
equity and commodity asset classes. Nominal Treasury bonds
and, in particular, on-the-run Treasuries outperformed risky
assets. The prices of inflation-protected bonds, whose credit
quality is identical to that of nominal Treasuries, behaved as if
their cash flows were actually risky. In other words, a reduction
in inflation expectations is not enough to explain the performance
of inflation-linked bonds.
The US TIPS market suffered significant losses during post-
Lehman period as investors sold into a backdrop highlighted by
deflation fears and diminished balance sheets. Chart 1 details the
US TIPS index returns versus a basket of nominal comparators.
Real yields are usually less volatile than nominal yields, so we
also include beta adjusted returns.2 TIPS underperformed
nominals by -18.4% (-11.2% Beta adjusted) during 2008. This
represents their worst return (vs. nominals) since inception of
the market in 1997. Although deflation was a highly publicized
theme we think technical selling pressures and the flight to
nominal Treasuries exacerbated this return differential.
. TIPS Nominals Nominals
(Beta Adj)
vs.
Nominals
vs.
Nominals
(Beta Adj)
YTD 09 5.20% -2.10% -1.20% 7.40% 6.50%
Q408 -2.70% 11.20% 6.40% -13.90% -9.00%
Q308 -3.60% 2.60% 1.60% -6.30% -5.20%
Q208 -0.30% -2.60% -1.60% 2.30% 1.30%
Q108 5.10% 4.90% 3.00% 0.20% 2.10%
2008 -1.70% 16.60% 9.50% -18.40% -11.20%
2007 11.80% 9.90% 6.00% 1.80% 5.80%
2006 0.50% 2.40% 1.50% -2.00% -1.00%
2005 2.70% 3.70% 2.30% -1.00% 0.40%
2004 8.70% 4.90% 3.00% 3.80% 5.60%
2003 8.20% 2.30% 1.30% 5.90% 6.90%
2002 17.00% 15.20% 9.10% 1.80% 7.90%
2001 8.00% 6.10% 3.70% 1.90% 4.30%
Chart 1: Total Returns
Source: Barclays, SSgA as of 3/31/2009.
Past performance is not a guarantee of future results. The above chart is used for
illustrative purposes only.
by James Mauro, Senior Portfolio Manager, Global Fixed Income
and Benjamin Tarlow, Ph.D., Quantitative Research Analyst,
Advanced Research Center
2
SSGA Capital Insights | Tviitlewpointse Goes Here
The carnage in inflation-linked markets drew in long-term investors
based on attractive valuation levels. In addition, we have
seen continued investor interest in TIPS for diversification benefits
within an asset allocation framework. These forces helped
TIPS stabilize towards year end and outperform nominal bonds
by +7.4% YTD (+6.4% beta adjusted).
While there has been an impressive recovery thus far in 2009, we
still feel TIPS are attractive relative to nominal treasuries or on a
breakeven basis. We highlight this relationship by comparing the
cumulative total returns of the TIPS index versus the cumulative
total return of a basket of nominal comparators (Chart 2). The
comparison of these index levels highlights the extreme underperformance
of breakeven returns.
Global Liquidity and Global Breakeven Inflation Rates
The breakeven inflation rate is the difference in yield between
a nominal Treasury bond and a comparable inflation-linked
Treasury bond. The principal amount of a typical inflation-linked
Treasury bond rises with realized inflation. Roughly speaking,
if inflation were to rise at the breakeven rate, then a hold-tomaturity
investor would be indifferent to having held a nominal
bond or an inflation-linked bond. The investor’s usually lower
yield on the inflation-linked debt would be exactly offset by rising
principal value.
Given this close link between inflation realizations and relative
returns, market participants tend to associate breakeven inflation
with market-implied inflation expectations. This market-adopted
terminology of “breakeven inflation” implicitly rules out other
primary determinants of breakevens. In this article, we will argue
that liquidity plays an under-appreciated role in determining
breakeven rates.
The topic of liquidity risk is worthy of its own dissertation, but
our aim here is to simply show the explanatory power of one
commonly used measure of market liquidity risk. In particular,
we will show a strong relationship between the on-the-run yield
premium3 in the US Treasury market and world breakeven
inflation. We will say that a security has liquidity risk if its price
falls during times of stress due to reasons unrelated to changes
in expected cash flows. If prices are low for reasons other than
cash flow risk, then the implication is that expected returns rise
in terms of stress. In our opinion, the inflation-linked market may
be in such a period.
Measuring world breakeven inflation is somewhat challenging,
as we need to compare nominal and inflation-linked bonds
across countries and across the yield curve. The Barclays Capital
World Inflation-Linked Index is unique in that Barclays keeps a
separate Treasury index of nominal comparators. For every bond
in the inflation-linked index there is a corresponding issue in
Barclays Treasury index. We can therefore directly compare the
yields on these two indices. This measure of world breakeven
inflation does not depend entirely on any particular country or
any particular point on the maturity curve.
250
100
150
200
03/97 03/99 03/01 03/03 03/05 03/07 03/09
A
B
A TIPS
B Nominals
US Treasury Total Return Index Levels
Chart 2: US Treasury Total Return Index Levels,
February 1997 = 100
Source: Barclays, SSgA as of 3/31/2009.
Past performance is not a guarantee of future results. The above chart is used for
illustrative purposes only.
3
SSGA Capital Insights | Tviitlewpointse Goes Here
Importantly, we are not concerned with measuring illiquidity
associated with the start-up costs of beginning an inflation-linked
program. For example, most academics and market participants
agree that real yields were too high early in the history of the US
TIPS market. Investors may have worried that the TIPS program
would be abandoned by the Treasury and the potential lack
of any new issuance might cripple secondary trading. To help
mitigate the influence of start-up costs, we begin our study
in January 2000. Chart 3 shows historical values for world
breakeven inflation and Treasury specialness.
Evidence from World Breakeven Inflation
Our goal in this section is to demonstrate the sensitivity of world
breakeven inflation to liquidity. Liquidity, of course, is not the only
driver of breakeven inflation and we just want to highlight that
liquidity explains breakevens to a surprising degree.
Our analytical strategy is to run a simple regression of breakeven
inflation on Treasury specialness. Regression is a standard
analytical tool used to quantify the relation between one dependent
variable and one or more explanatory variables. Using this
procedure we can estimate the sensitivity of the dependent
variable to each explanatory variable (beta). In addition, we
can compute the so-called R2, which measures the correlation
between the dependent variable and its value implied by the
estimated relationships with the explanatory variables.
In a regression of breakeven inflation on specialness, we look for
two results. The first result we want to show is that the beta, or
sensitivity, on specialness is negative, which means that larger
specialness spreads imply lower breakevens. The second result
we want to find is a reasonably positive R2. In tests, we actually
do find a negative beta and a positive R2.
Chart 12: US Noninvestment-
Grade Corporate Bonds Yield-to-Worst
Source: SSgA, Credit Suisse.
3.0
0.5
1.0
1.5
2.0
2.5
2000 2005 2010
Percentage Points
Chart 12: US Noninvestment-
Grade Corporate Bonds Yield-to-Worst
0.7
0
0.1
0.2
0.3
0.4
0.5
0.6
2000 2005 2010
Percentage Points
Chart 3: Daily World Breakeven Inflation and Treasury Specialness
Source: SSgA, Bloomberg as of 3/31/2009.
The above charts are used for illustrative purposes only.
World BEI US Treasury Specialness
4
SSGA Capital Insights | Tviitlewpointse Goes Here
Since breakevens and specialness have moved so dramatically
over the past year, we ran two separate regressions. The first
regression covers the period from 2000 to present, while the
second regression leaves out the last year of data. Specialness
alone explains about 60% percent of the variation over the full
sample and about 40% of the variation when one excludes the
past year from the analysis. In either case, our opinion is that
inflation-linked debt shows a meaningful sensitivity to liquidity.
Chart 4 shows the actual and fitted values of these regressions.
Regional Evidence
We ran the same regressions that we ran at the world level at
the regional level and we have displayed the results in Chart 5.
Our hypothesis is the coefficient on liquidity should be negative
for all regions and, to varying degrees, breakevens in each of
the regions do show negative sensitivity to liquidity. Note that
the number of observations for each regression varies with
data availability
Chart 12: US Noninvestment-
Grade Corporate Bonds Yield-to-Worst
3.0
0.5
1.0
1.5
2.0
2.5
2000 2005 2010
A
B
A Actual
B Fitted
Percentage Points
Chart 12: US Noninvestment-
Grade Corporate Bonds Yield-to-Worst
3.0
1.8
2.0
2.2
2.4
2.6
2.8
2000 2005 2010
A
B
A Actual
B Fitted
Percentage Points
Chart 4: Actual and Fitted Breakeven Inflation (BEI)
Source: SSgA, Bloomberg as of 3/31/2009.
The above charts are used for illustrative purposes only.
Actual vs. Fitted World BEI Including Last Year: beta = -2.54; R2 = 0.59 Actual vs. Fitted World BEI Excluding Last Year: beta = -1.86; R2 = 0.41
Beta R2 Number of
Daily
Observations
UK -1.01 0.08 2271
Japan -5.80 0.83 1231
Italy -1.04 0.31 1355
US -4.43 0.73 2271
Canada -2.29 0.57 2271
France -1.65 0.38 2271
Germany -2.19 0.60 731
EU -1.65 0.38 2271
World -2.54 0.59 2271
World (excluding
last year)
-1.86 0.41 2019
Chart 5: Regional-Level Regressions of Breakevens on Specialness
Source: SSgA, Bloomberg as of 3/31/2009.
The above chart is used for illustrative purposes only.
5
SSGA Capital Insights | Tviitlewpointse Goes Here
Supporting Evidence from the Inflation Swaps Market
One potential critique of the analysis so far is that if liquidity
is negatively correlated with global expected inflation, then our
previous calculation of the specialness betas may be biased.
What we need is a traded measure of global inflation
expectations that is independent of the liquidity issues, which
may be present in the linkers market.
Although we do not have a traded, non-linkers-based world
expected inflation measure, we can use the information in the
US inflation swaps market to help resolve this issue. The rate
on a zero-coupon inflation swap is a traded measure of inflation
expectations that does not depend on the relative liquidity sensitivity
of the nominal and inflation-linked markets. We have a little
over four years of daily data on the US inflation swaps market.
We ran a regression of US breakevens on a constant, the inflation
swap rate, and specialness. The R2 of this regression is 0.96,
while the beta on specialness is -2.09. We believe the magnitude
of this beta is significant in the economic sense.
Economic significance is a subjective statement on the potential
payoff of a given strategy. Suppose that the on-the-run premium
is currently about 50 basis points and its normal level is approximately
20 basis points. As estimated by the regression, the
sensitivity of US breakevens to specialness is about 2. Based on
these estimates, US breakevens are then approximately 60 basis
points too low (2x (50 - 20) = 60).
Comparison with Other Studies
A recent article by New York Federal Reserve President William
Dudley4 reviewed the case for continuing US TIPS issuance. If
TIPS are a cost-efficient method for the US Treasury to raise
funds on behalf of taxpayers, then it must be the case that any
illiquidity premium must be currently small. Dudley cites two
quantitative estimates of the illiquidity premium, both of which
show the illiquidity premium declining substantially over time. In
this view, illiquidity has trended lower and now resides, perhaps
permanently, at some tolerable level.
We would agree that the level of the illiquidity premium has
likely declined since TIPS were originally introduced. Our point
of contrast in this note is to show that low levels of illiquidity
premiums need not be a continuous feature of a well-functioning
inflation-linked market. The inflation-linked market may occasionally
have periods of severe illiquidity even if the premium is
lower in normal markets.
Conclusions and Methods for Gaining Exposure to Linkers
We have argued that inflation-linked bonds are more sensitive to
liquidity stress than some investors may appreciate. This relationship
has become pronounced during the current financial crisis
and caused significant price dislocation between real yields and
nominal yields. We recommend owning TIPS on a breakeven
basis and, as a starting point for a discussion, we propose the
following possible method to establish the position.
An investor could purchase a basket of TIPS or a commingled
TIPS fund and hedge out the interest rate risk. Our preferred
hedging method would be to short a basket of nominal comparators
(beta adjusted). We believe that this method allows for
relatively precise hedging of curve risk, but it does require mild
counterparty management. An alternative for hedging the duration
would be to sell exposure on the Treasury futures exchange.
Since the futures market contains some optionality (cheapestto-
deliver), we feel that hedging with futures may be more costly
over time. In addition, the limited number of futures maturity
points limits our ability to match duration across the curve. For
long-only portfolios, an investor could implement this strategy
by shifting part of their Treasury allocation to a (beta adjusted)
portfolio of TIPS.
6
SSGA Capital Insights | Title Goes Here
© 2009 State Street Corporation INST-0014-0509
This material is for your private information.
The views expressed (contained) in this material are the views of SSgA through the period
ended April 22, 2009 and are subject to change based on market and other conditions.
All information has been obtained from sources believed to be reliable, but its accuracy
is not guaranteed. There is no representation or warranty as to the current accuracy,
reliability or completeness of, nor liability for, decisions based on such information. This
document contains certain statements that may be deemed forward-looking statements.
These statements are based on certain assumptions and analyses made by SSgA in
light of its experience and perception of historical trends, current conditions, expected
future developments and other factors it believes appropriate in the circumstances. Past
performance is not a guarantee of future results.
Interest rate increases can cause the price of a debt security to decrease. Increase in
real interest rates can cause the price of inflation-protected debt securities to decrease.
Interest payments on inflation-protected debt securities can be unpredictable.
Investing involves risk including the risk of loss of principal.
SSgA Capital Insights is an integrated thought leadership program
designed to educate clients on timely investment and market
topics. As part of State Street’s Vision Thought Leadership series,
the SSgA Capital Insights program gives clients access to the
expertise and viewpoints of SSgA’s thought leaders and investment
talent via a variety of multimedia channels.
Since 2006, State Street’s Vision Series has been distilling
our unique research, perspective and opinions on key themes
impacting institutional investors worldwide into publications for
our customers around the world.
viewpoints
1 SSgA, April 2009.
2 The beta adjustment is a method for adjusting between the relative volatility of the two
sectors. The magnitude of this adjustment is an empirical question. SSgA estimates this
beta from one-year rolling regressions of real yield change on nominal yield change.
3 We compute the on-the-run yield premium by comparing the yield on the most recently
issued 10-year Treasury with the 10-year fitted yield. The fitted yield is the constantmaturity
10-year yield implied by a group of off-the-run Treasury bonds. The data is from
Barclays Capital. The on-the-run premium is also known as Treasury specialness.
4 Dudley, William, J. Roush, and M. Steinberg Ezer (2009). “The Case for TIPS:
An Examination of the Costs and Benefits.” FRBNY Economic Policy Review.
The Barclays Capital World Inflation-Linked Index is a trademark of Barclays Capital, Inc.
The Barclays Treasury Index is a trademark of Barclays Capital, Inc.