You are on page 1of 3

C

L I E N T

The Mystery of

E T T E R

A R T N E R

TIPS*1

Robert D. Arnott
Chairman, First Quadrant, LP

C T O B E R

2003

E S S A G E

How Reliable is the Real Growth of Dividends?


How reliable is the index-tracking for stock dividends the real
coupon that we clip from indexed stock ownership? In Exhibit 1, we
can see that best-fit measures of dividend growth, that
dampen the effect of market cycles over long 40-year spans,
Best-fit measures
are almost always between zero and two per cent. Earnings
of dividend growth
growth follows a similar path, albeit with a wider range.

are almost always

A version of Treasury Inflation Protected Securities (TIPS), or inflationindexed bonds, has been with us for a long, long time. Not merely
since 1997, but since the dawn of the capital markets. After
all, TIPS arent really all that different from stocks!
TIPS arent really
all that different
What does a well-diversified equity portfolio do for us? It
from stocks!
provides income which grows typically slightly faster than
inflation, thereby providing a real yield roughly indexed to
inflation. It also rises if the dividend yield (proxying for the real cost
of capital) falls.
What do TIPS do for us? They provide income which rises with
inflation, that thereby provides a real yield tightly indexed to inflation.
It also rises if the governments real cost of capital, the TIPS yield, falls.
The biggest difference is that TIPS have an expiration date and stocks
dont. A secondary difference is that stocks provide imperfect inflation
tracking. In other words, stocks can be seen as inflation tracking
corporate TIPS consols2, with an unreliable growth kicker averaging
about one per cent per year.
*
1

Published in the Financial Analysts Journal, September/October 2003.


In the interests of full disclosure, I should acknowledge that my own personal
investments include a significant commitment to TIPS, as do the mutual
funds that I subadvise for PIMCO.
Consols are perpetual bonds, without an expiration date.

Where are the sustained 5% real earnings growth rates that


between zero and
Wall Street likes to forecast for the broad market averages?
two per cent.
The highest ever, 4.1%, was in the span from 1931 (a most
auspicious starting point for any measures of real earnings
growth!) to 1971. What of the much-vaunted new paradigm of the
late 1990s? It appears as a tiny and brief uptick in 40-year best-fit real
earnings growth in the late 1990s, to a lofty 1.2% annual growth
rate above inflation; this was never mirrored in dividends, which
continued a slow decline to a current level of 0.4% real 40-year growth.

5%

Exhibit 1. 40-Year Real Dividend and Earnings Growth


(Jan 1842-Marg2003) (Log-Linear Best-Fit Growth Rate)
40-Year R eal
E arnings Growth

4%
3%
2%
1%
0%
40-Year R eal
Dividend Growth

-1%
-2%
1840

1860

1880

1900

1920

1940

1960

1980

2000

PARTNERS MESSAGE: THE MYSTERY OF TIPS i

L I E N T

E T T E R

A R T N E R

E S S A G E

TIPS clearly define the real cost of capital to the government. Before
the launch of TIPS, we needed to infer that cost. For an estimate of
the real cost of capital for the government, we uses long-government
bond yields, less a model for long-term prospective inflation, in Arnott/
Bernstein [2002].3
Finance literature suggests that the real cost of capital should reflect
expected real productivity growth, plus some premium for default risk
and/or illiquidity. Since productivity growth closely tracks real percapita GDP growth, this would point to a real cost of capital for the
government (assuming zero default risk, which isnt quite true)
approximately matching real per capita GDP growth, which has
averaged 1.4% since World War II. Equities should have a risk
premium, reflecting both the volatility of equities and the uncertain
rate of real earnings or dividend growth.
Figure 2 shows that this has been an inaccurate picture, with the
government cost of capital sharply exceeding these levels (particularly
3

Arnott, Robert D., and Peter L. Bernstein, What Risk Premium is Normal?
Financial Analysts Journal, March/April 2002.Sons; 1st edition (April 4,
2003), pg 23.

Exhibit 2. Corportate and Government Cost of Capital


(Jan 1960-Mar 2003)

Stock Yields and TIPS Yields, The Real Cost of Capital


Simplistically stated, stock dividends have (1) a real yield, (2) a default
premium for the risk of bankruptcy, and (3) a real growth expectation.
For a broad market index, the latter two factors converge on
a real growth rate of about 1%, so that the real cost of
The real cost of
corporate capital is approximately the dividend yield plus 1%.
equity-based
This implies that the average real cost of capital for the broad
corporate capital
market averages, when funded through the equity markets,
is approximately the
lies about 1% above the average dividend yield. Surprisingly,
dividend yield
corporate bonds are a weaker measure of the real cost of
plus 1%.
corporate capital, due to the unknowable future rate of
inflation.

2003

C T O B E R

10%
9%
8%
7%
6%
5%
4%
3%
2%
1%
0%
1960

1965

1970

1975

E st. Corporate Cost of Capital

1980

1985

1990

E st. Government Cost of Capital

1995

2000

T IP S Yield

in the 1980s, as investors factored in a large premium for


expected inflation, which failed to materialize). The corporate
cost of capital swung sharply below the governments cost
of capital at the peak of the 1998-2000 bubble. The cost of
capital for the government and for the corporate arena (defined
in this fashion) have also seen sharply negative correlations
since late-1996, just before the launch of the TIPS market.

The corporate cost


of capital swung
sharply below the
governments cost of
capital at the peak
of the 1998-2000
bubble.

So, Whats the Mystery?


TIPS are more similar, in their mechanism for producing investment
returns, to stocks than to bonds. An analogy is that we have corporate
bonds and government bonds; we also have corporate stocks and
government stocks, which we call TIPS.

PARTNERS MESSAGE: THE MYSTERY OF TIPS ii

L I E N T

E T T E R

A R T N E R

If they are seen as, in effect, government-issued equities, why is


it that TIPS are negatively correlated to stocks?

If they are seen as,


in effect,
government-issued
equities, why is it
that TIPS are
negatively correlated
to stocks?

Given that stocks are like infinite-maturity TIPS, with some


wiggles around the inflation-tracking, given that both are
real-return assets which directly measure the real cost of
capital, why is it that TIPS have yielded as much as four
times the stock yield (the circumstance in early 2000),
implying a government cost of capital far higher than the
cost in the corporate arena?

The only explanation that I see is that much of the


investment community still thinks of TIPS as an alternative to
conventional bonds rather than as an alternative to stocks (or, perhaps
more accurately, as a different beast entirely!).
Table 1. Asset Class Correlations (t-Statistics), Since TIPS
Inception (Feb 1997 Mar 2003)
Assets

Correlation

Stock vs Bond Returns

-0.204 ( -1.9)*

Stock vs TIPS Returns

-0.217 ( -2.0)

Stock vs TIPS Yield

-0.637 ( -9.0)

Bond vs TIPS Returns

+0.597 (+7.9)

C T O B E R

2003

E S S A G E

Why Does any of this Matter?


TIPS and stocks are the two best ways to defease real
pension liabilities, and real spending for endowments and
foundations. Remarkably, much of the investment community
studied (i.e., largely ignored) TIPS in early 2000, when the
yield was as high as 4.3%, focusing their real asset
exposure on stocks with a yield of 1.1%. For stocks to merely
match the performance of TIPS from that starting point, let
alone provide a risk premium, we would have needed real
dividend and earnings growth of 3.2% above inflation, a rate
of growth, which is without precedent apart from the recovery
from the Great Depression.
TIPS yields and stock market yields (plus a 1% growth
premium) are, in my view, the best measures of the cost of
capital for the government and for the corporate world,
respectively. The relative comparison of the two is, in my view,
the best measure of the relative attractiveness of stocks and
bonds, given the remarkable (and largely unnoticed) parallels
in the ways that the two assets deliver returns to their
investors.

TIPS and stocks are


the two best ways to
defease real pension
liabilities, and real
spending for
endowments and
foundations.

TIPS yields and


stock market yields
(plus a 1% growth
premium) are, in my
view, the best
measures of the cost
of capital for the
government and for
the corporate world,
respectively.

*As compared with +0.078 correlation since the end of World War II.

PARTNERS MESSAGE: THE MYSTERY OF TIPS iii

You might also like