Why Investors Need to Think About Inflation
A Conversation with Harvard Economist John Y. Campbell
The idea of inflation-indexed bonds has been around since 1780, and
they've been available in the United Kingdom for 15 years. Just three months ago, two Ivy
League economists -- John Y. Campbell of Harvard and Robert Shiller of Yale -- told their
colleagues at the National Bureau of Economic Research these bonds were important to
investors.
John Y. Campbell is the Oto Eckstein professor of applied economics
at Harvard University. He spoke recently with John Byczkowski, editor of Deloitte &
Touche OnLine.
Why are inflation-indexed bonds needed? Aren't
there other inflation hedges available?
Campbell: The short answer to whether
there are other inflation hedges is that if you have a short investment horizon, you have
a number of options. Basically, with a short investment horizon, inflation is not such a
concern, because there's only very limited uncertainty about inflation over the next few
months. You can hold a conventional Treasury bill or a money market fund and get a return
that is known in advance in current dollar terms, and is pretty close to being known in
advance in real terms as well.
The problem comes if you have a long-term investment
horizon. We think most people should have long horizons. When people retire they should be
thinking in terms of 20 years, and not thinking in terms of five years or under. When
somebody retires they ought to realize they're going to need a stream of income for 20
years.
Your real income needs, in purchasing power terms, are
going to be fairly constant. You may need a nest egg for medical expenses. What you want
is constancy of real income over long periods of times.
No asset out there really delivers that. Certainly
conventional long-term Treasury bonds do not. How a fixed nominal income can be affected
over 20 years by inflation, it's just dramatic, even at low rates of inflation. That
really is not a good option.
For that reason, I find it incomprehensible that so many
financial advisors urge their clients to shift out of stocks and into long-term
conventional bonds as they get older. That advice is based on the premise that long-term
nominal bonds are safe investments in a relevant sense. They're not; they're only safe in
dollar terms. That's meaningless. What a retiree needs is purchasing power and not green
pieces of paper.
The stock market has certainly turned in a very
attractive performance recently. Stocks are attractive because they have high returns on
average. And their long-run risk may not be as great as their short run risk. Even so,
stocks cannot be described as safe investments. People at the moment are too mesmerized by
the favorable experience of the 1980s and '90s, and they've forgotten the '70s and the
periods in the '50s and '30s when stock returns were very bad for long stretches of time.
So, while there's always going to be a role for stocks in
people's portfolios, there's also a role for something that's really safe in real terms,
and that's what these bonds are going to offer.
Back to the Top
The Tax Issue
Under current laws, the owners of these
inflation-indexed bonds would have to pay taxes each year on the markup. What affect does
that have on the bonds' desirability and their usefulness?
Campbell: What it means is the bonds are
going to be most attractive for people who have tax-deferred retirement accounts. If you
could shove the thing in a 401(k) or your pension fund buys them on your behalf, that's
going to be the best option. If inflation goes up, you'll get this inflation adjustment,
but then you'll be taxed on some of that (adjustment), and your real return will be
negatively affected, just through the tax effect.
In my view it would be better if the Treasury would design
something that would offer a tax-exempt real income, where the inflation-adjustment would
not be taxed. But I think what they're doing is certainly better than nothing.
Conventional Treasury bonds, you're taxed on the full
nominal interest, even though some of that is merely compensation for inflation, which is
eroding your real principal. Yet you're taxed on all of what is labeled as interest, even
though not all of it is real. This is a problem with the whole tax system. We have to live
with it, and be realistic, and I think these bonds area big step forward, even though
they're not perfect.
In your paper, you and Dr. Shiller talk about the
fact that in some other countries, these bonds are tax free.
Campbell: In the UK for example, they
have always had a system whereby bond interest is taxable, but capital gains on bonds are
not. They treat the inflation adjustment (of inflation-indexed bonds) as a capital gain.
But it's in the context of a whole tax system that works differently (compared to the US
system).
How likely is the United States to adopt that
taxation method?
Campbell: It's not very likely. I don't
think we're going to see special treatment for the inflation-indexed bonds. It's
conceivable that sometime in the future the whole (US tax) system might be indexed.
Back to the Top
How big a premium over inflation should these
bonds pay?
Campbell: If you look at the UK
experience, the real yield of these bonds -- think about a 10-year zero – the real
yield has been somewhere between 3% and 4%. I would expect – and this is a guess,
(because) we would have to see what happens at the auction – but I would expect to
see a real yield of about 3% above inflation.
That would seem to "reorder"
investments. Historically, long-term government bonds have tended to pay about two
percentage points above inflation. If inflation-indexed bonds pay three points above
inflation, that would reorder the roles of investments available.
Campbell: The experience of nominal bonds
has been distorted to some extent by the big losses holders of nominal bonds took in
periods of high inflation. It's not clear that bond holders looking forward are expecting
to get only 2% over inflation. They may be looking for more.
It's certainly true that short-term money market
investments are barely paying 1% above inflation. Three percent above would be pretty
attractive.
It's possible that demand drives the yield down, closer to
inflation. But we really have to see.
Back to the Top
Changing
History
You and Dr. Shiller think this would be a
pretty significant event historically. How would this change investing?
Campbell: We think that historically,
inflation has had considerable power to disrupt society and transfer wealth between
people. It happened back in the 1930s, when there was deflation. In the '60s and '70s
there was inflation, and people who own property made out like bandits. There were big
transfers (of wealth) in society, because people make long-term arrangements in dollar
terms (and the purchasing power of dollars was eroded by inflation).
We think this is a bad thing for individuals to be
facing this risk. It's also bad because it undermines faith in the fairness of capitalism.
Now, the significant thing about this (the
introduction of inflation-indexed bonds) is that it will directly
provide a way for people to get a safe return. Beyond that, it will encourage people to
think in real terms. Part of the problem is people aren't systematic enough about thinking
through what it means to think in real terms instead of dollar terms.
For example, many people take out long-term nominal
mortgages, and they think this is the ultimate safe arrangement. They're protected if
interest rates go up, and if they go down, they can refinance. It's a very large scale
arrangement for many people.
If you think about this arrangement in real terms, the
conventional 30-year mortgage is a huge one-way bet on inflation. In real terms, if
inflation goes down, nothing much happens, you refinance. If inflation goes up a lot, you
cash in, you've hit the jackpot. That happened to the generation that bought homes in the
'60s and '70s.
This one-way bet, it does not come for free. The
people who are lending you the money have to charge you for this possibility that you're
going to win big. They charge you for it by raising the mortgage interest rate. The
mortgage interest rate is well above the Treasury bond yield, and that's not just because
it's risky. A lot of it is because you're signing up for a contract, you're taking a
chance at hitting the jackpot. You're buying a lottery ticket, and you pay for that
lottery ticket.
Our best guess is about 125 basis points (or 1.25
percentage points) of the mortgage interest rate is the price of this lottery ticket. If
you would give up the lottery ticket, you would reduce your mortgage interest rate by 125
basis points.
We don't think most people really need or want this
one-way bet on inflation. One of the things that may happen as a result of these indexed
bonds, people may start to see through this, and we may start to see more indexed
contracts – indexed mortgages, for example.
Back to the Top
Which Measure
of Inflation?
One thing the Treasury Department hasn't decided
is how to define inflation for the purpose of indexing these bonds. Your personal rate of
inflation – the rate that's relevant to you – is different than what's average
for the economy overall. What's the best solution?
Campbell: Everybody in a sense has their
own personal rate of inflation. If I buy a lot of books, then I care about the rate of
increase in the prices of books.
But, to be realistic, we're not going have a plethora of
inflation indexes. We have to agree on something.
The best thing to agree on is something that's closest to
the average. An average price index is going to be as close as we're going to get to all
the individual price indexes.
Isn't there some debate about the accuracy of the
Consumer Price Index and other current measures of inflation?
Campbell: There is this argument that the
CPI overstates inflation, because the quality of goods has been rising, and that's not
taken into account. The people constructing the CPI look at a car today and a car last
year, and see today the car is 3% higher in price than last year's car, neglecting the
fact that this year's car is perhaps 1% better. So really, the inflation is only 2%, but
they measure it at 3%.
That's a real problem, particularly for the indexation of
Social Security payments. But I don't think it's much of a problem for inflation-indexed
bonds, because this bias (in inflation measures) – we don't know much about it, but
we think it's pretty constant and pretty predictable. Maybe it's a half percent a year,
maybe it's 1% a year. And it's steady; it's constant over time.
There's not much uncertainty about that. So that's going
to get built into the price people are willing to pay for the bonds.
Remember, the function of the bonds is to protect people
against random, uncertain and unpredictable changes over time, and this bias is not random
and unpredictable. So the shocks of inflation, which the bonds are supposed to protect you
against, are genuine shocks. And so the existence of bias is not a problem. |
