Interest rates are very likely to rise over the next year or two. To
what degree, how quickly and when that rise in rates might begin is hard
to say. But the economy does seem to be firming, the Fed is not
likely to do another round of asset purchases, and inflationary
pressures may be building. This has led many investors to hold unusual
amounts of cash. It has also spurred interest in products such as the ProShares UltraShort 20 Year-Plus Treasury (TBT - commentary - Trade Now), which each month breaks new records for shares outstanding.
But how risky are bonds really? What are the
odds that a typical diverse bond portfolio will produce steep losses?
Instead of jumping to conclusions, let's put pencil to paper (or perhaps
more accurately, typeface to bandwidth) and do the math. In this week's
Bondlife, we'll take a real bond portfolio and put it through the
rising-rate wringer, and see what comes out the other end. Our
conclusions will be very different depending on what kind of investor
Plus, why you should buy high coupons in some markets and low coupons
in others, and should states be allowed to declare bankruptcy Get ready
to spin the wheel, Monte Carlo style!
Interest Rates and the Fed: Roulette
Bond portfolios are more complicated that you realize, and the
risk/reward cannot be summarized as merely duration vs. yield. To
illustrate what I mean, I did a Monte Carlo simulation. I ran the
simulation over eight quarters, with each quarter having its own
independent simulation run. In each case, the simulation assumed that
all rates would move parallel and that the rate shifts were normally
distributed. The standard deviation was assumed to be equal to the
20-year average for 10-year bonds, and the mean of the rate distribution
was set at +25 basis points. So therefore, on the basis of this
simulation, if all eight quarters were exactly equal to the mean rate
shift, then all rates would rise by 200 basis points over the course of
the two-year simulation.
I assumed a portfolio which was evenly invested
between one-to-10-year maturities and that the whole portfolio has an
average yield spread of 80 basis points over Treasuries. This is equal
to the yield spread on the taxable bond fund which our firm manages. I
then assumed the portfolio remained static over the two years, except
that all income and maturities were reinvested in a new 10-year bond.
While this model employs a bit of a simplification, it's a pretty good
proxy for a real portfolio managed in a real way.
The quarterly Monte Carlo analysis produces results that are much
closer to reality than a simple stress test of a portfolio with a
+200-basis-point shock. The market will not suddenly shift 200 basis
points; instead it will oscillate back and forth, potentially winding up
+200 basis points but following a mixed path along the way. It could be
that it rises 300 basis points and then takes 100 basis points back. Or
it could rally 50 basis points before selling off 250 basis points.
Each of these scenarios creates a very different reinvestment
opportunity. Running a series of Monte Carlo simulations (I ran 1,000)
gives us the chance to see how this portfolio might perform in a variety
The chart below shows the distribution of final 10-year Treasury
rates based on the simulation. The rate is along the bottom, and the
percentage of simulations that ended with a rate in that range is along
the vertical axis.
Not surprisingly, only about 4% of scenarios wind up with interest
rates materially lower than today, and 70% of scenarios result in yields
in the 5% to 7% range, which is consistent with around +100 to +300
from where we are now. Stop here for a moment. If you knew this was the
potential distribution of 10-year interest rates between now and the end
of 2012, you'd be ready to short bonds right? Or at least hold cash.
Yet there are a lot of things going for bonds that
tend to severely dampen the volatility of a portfolio. Look at the
distribution of total returns produced by the simulation.
Only about 13% of scenarios produced returns at or below zero. This
essentially tells you that in all but a small percentage of cases, an
investor would be better off holding the one-to-10-year bond portfolio
than holding cash. How can this be with most of the rate cases resulting
in substantially higher rates?
Bear in mind the following: First, the cash flow is getting
reinvested. Einstein allegedly said that compound interest is the most
powerful force in the universe. Well, the compounding factor actually improves
when rates rise, because the reinvestment happens at higher rates.
Second, when a bond drops in value, its accretion factor improves. That
is, as time passes, every bond price moves toward par. This isn't some
kind of accounting magic, it really is how bonds trade. If you own a
10-year bond at $100 and market forces cause it to drop to $90, that $10
discount to its maturity value gets amortized over the life of the
bond. So if nothing happens, in one year, the bond is worth $91. In two
years it is $92. Another way to think about it is if you start with a
bond with a coupon of 3.5% and market yields rise such that it now
yields 5%, even though you have a capital loss on your initial 3.5%
position, looking forward the bond does indeed yield 5%. Since your
original 3.5% coupon is less than the market yield, the difference is
made up through the accretion of the discount.
Finally, as bonds get shorter, the yield the market demands
naturally declines. For example, five-year Treasuries currently yield
2.05%, whereas four-year bonds yield 1.53% and three-year bonds yield
1.06%. That means that if nothing changed about the curve slope but
yields rose 100 basis points across the board over two years, five-year
bonds would rise from 2.05% to 3.05%, but one who actually bought a
five-year bond two years ago would now own a three-year bond. Under our
scenario, three-year bonds would yield 2.06%, so the two-year-old
five-year bond would basically be unchanged in price.
What are our conclusions here? First, bonds are
more resilient against losses than you might think, especially if you
give them a little time for income to accrue and time to be on your
side. As a corollary, this tells you that if you are holding bonds
because you want a relatively stable portion of a larger portfolio, then
this standard bond portfolio fits the bill. Second, holding cash
becomes very timing dependent. Sure, if yields rise in the very near
term, it can work for you. But unless you have that very specific
outlook, you will probably do better owning an intermediate-term bond
That being said, there remains very little upside opportunity in
bonds. We are remaining very tactical in our positioning, trying to make
a lot of small winning trades rather than making big macro calls. We
believe this can add enough alpha to make bonds reasonably attractive,
but the fact remains that bonds are what they are: fairly stable, but
Treasury Bonds: Getting Nothing From Something
What about shorting Treasuries? As with cash, this becomes very much a
timing game. In my Monte Carlo analysis, 83% of scenarios resulted in
rates rising by at least 1% by the end of two years, and 41% of
scenarios saw rates rise by at least 100 basis points in a single
quarter at some point during the simulation. Clearly, if you are good
enough to time that move correctly, there is good money to be made
shorting Treasuries. But if your timing isn't right, your results can
become very pedestrian very quickly. The curve roll-down, the discount
accretion and the income compounding are all working against you. A
Treasury short should be a trade you make when interest rates look like
they are going to imminently rise. It's not a trade you just hang on to.
Corporates and MBS: Luck Only Lends
In a rising rate environment, you want to own high-coupon
corporates and low-coupon mortgage-backed securities. High-coupon
corporate bonds pay you more money during the life of the bond, and this
gives you better re-investment possibilities as rates rise. Low-coupon
bonds pay you more at maturity, so you have fewer reinvestment
possibilities in the near term.
The opposite is true of MBS. High-coupon MBS have
less natural principal retirement (because more of the monthly mortgage
payment is going to interest). Plus you have to pay up for the high
coupon, but you don't know if the mortgage holder is going to prepay. So
you could be paying up for a high coupon which you'll never actually
get. Low-coupon MBS have much more stable average life profiles, meaning
you know what you are getting and will have a steady stream of
principal payments to reinvest at potentially higher rates. This is
especially true if you buy a shorter amortizing bond, such as a 15-year
MBS. This kind of trade really takes advantage of the compounding
Municipals: Living Dangerously
Should states be allowed to declare
bankruptcy? Several articles are suggesting that Congress is giving it
some thought, most recently in a New York Times piece from
Friday. I think there is a lot of logic to this. It was only after
declaring bankruptcy that the airlines and the auto companies were
finally able to strip out badly structured retirement benefits and
modernize their business models. I'm not certain that every state can
make the cuts needed in their pension benefits without a court forcing
the unions to the table.
But one thing should be clear: No state has a "debt problem" per se.
It isn't debt service payments or even the need to roll over any given
debt that is crushing certain states. In fact, states have a budget
problem, which in my opinion is inherently a political problem, not an economic
problem. In fact, I'd argue that only a handful of states have a
serious pension/benefit problem, and maybe only Illinois really needs bankruptcy. Maybe just the threat of such a thing is enough to bring the unions to the table.
All this talk makes me all the more cautious on buying Illinois
bonds. I mentioned a couple of weeks ago that one could consider the
Illinois 5.1% bonds due in 2033, which are taxable bonds and trade
regularly. These bonds have been trading around $75, which gives you a
lot of protection in the event Illinois winds up trying to settle its
debt for less than $100. If you consider Illinois at all, please don't
make the classic mistake of buying short-term debt (which inevitably is
trading near par), figuring this will protect you. If the worst happens,
all debt will get the same haircut. Buy the lowest dollar price bond