We are not convinced, however, that a
one-year period is the appropriate time frame to judge long-term
investments. So, we’ve smoothed the annual returns of stocks and
bonds by computing three-year trailing averages. These averages
are shown in Charts 3 and 4. Note that a three-year average does
not change the average annual return.
On a three-year basis, stocks have had
one (two) down period(s), and bonds have had three (three) down
periods. So, if your definition of risk is the probability of
losing money, the difference is small, but it favors stocks.
When people on Wall Street talk about
risk, they really are talking about the variability of returns,
not the probability of losing money. Wall Street maintains that
stocks are riskier than bonds simply because there is a greater
variation in the one-year return.
Chart 1 (Updated)
Chart 2 (Updated)
Chart 3 (Updated)
By Wall Street’s definition, even if returns
were positive each year and had the same pattern for stocks and
bonds, but stocks varied between 0% and 20% while bonds varied
between 0% and 10%, stocks would be considered riskier because
the variation was greater. We think the problem is in the definition.
Over 41 (51) years, stocks averaged 11.7% (11.2%) per year—7.5%
(7.3%) over inflation—and netted a total of 17 (33) times the
original purchasing power. Bonds averaged 5.6% (6.8%) per year—
1.4% (2.9%) over inflation—and netted less than two (four) times
purchasing power.
Total returns from stocks consist of the dividends received and
the change in price. Total returns from bonds consist of the interest
received and the change in price. People seem to have forgotten
that when interest rates go up,
bond prices go down, and investing in bonds can lose you money.
By 1980, 5% bonds bought in 1966 were worth less than $0.50 on
the dollar. So, if you can predict interest rates, you will know
when to own bonds.
We have found, however, that you don’t need to predict interest
rates to know when to own bonds. You only need to know whether
current “real” returns are attractive.
Chart 4 (Updated)
Get the “Real” Story
At Muhlenkamp & Company, we define risk as the probability of
losing purchasing power over time. When we look at bonds, we subtract
the current inflation rate from the current yield level to get
the expected “real” return. We then set a hurdle rate of a 3%
“real” return before we are willing to lend money by buying bonds.
We’ve plotted nominal long-term interest rates in Chart 5 and
real long-term interest rates in Chart 6. We have also indicated
our 3% hurdle rate in Chart 6.
Chart 5 (Updated)
Chart 6 (Updated)
Chart 6 makes it apparent why, from 1968, when I entered the investment
business, to 1981, I never invested in long-term bonds. Bonds
didn’t meet the 3% hurdle rate for real returns except for a brief
period. It also illustrates why I was very comfortable investing
heavily in bonds from 1981 through 1986. Then, current real returns
on bonds were 6–8%, versus a hurdle of 3%. At various times during
that period, corporate interest rates were higher than corporate
returns on equity, meaning it was unprofitable for companies to
borrow money and interest rates had to fall. But Chart 6 demonstrates
more than that.
Chart 6 demonstrates three distinct periods in the returns available
from bonds. From 1952 to 1965, bonds promised average nominal
returns of 3.5% and real returns of 2.1%. Hindsight, and Chart
2, shows that they actually provided nominal returns of 2.1%.
We consider this period “normal,” at least in comparison to what
followed. From 1966 to 1980, bonds promised nominal returns of
6.75% and very poor real returns of 0.1%. They produced nominal
returns of 2.6%. From 1981 to 1993, bonds promised nominal returns
of 9.5% and unusually good real returns of 5.0%. They produced
nominal returns of 13.6%. (From 1994 to 2002, bonds promised nominal
returns of 6% and real returns of 3.9%. They produced nominal
returns at 9.2%—because rates declined). We consider both the
1966 to 1980 period and the 1981 to 1993 period to be unusual
and not likely to be repeated any time soon. (For an explanation
about why this happened, see our essay “And the Climate Is...”
Muhlenkamp Methods #6.)
If we also calculate the stock returns for these three periods,
we find that in the 1952 to 1965 period, when bonds averaged a
2.1% return, stocks averaged 14.5%. In the 1968 to 1980 period,
both stocks and bonds did poorly, but stocks did better than bonds:
6.7% compared to 2.6%. In the 1981 to 1992 period, when bonds
did very well, stocks did better: 14.7% compared to 13.6%. (In
the 1993-2002 period stocks did 9.5% compared to bonds 9.2%.)
To understand why this is so, read our essay “Why the Market Went
Down” available on our website: www.muhlenkamp.com.
At Muhlenkamp & Company, we believe the reason stocks perform
better than bonds is not because they are “riskier” but because
corporate management works for the stockholder and against the
bondholder.
Chart 6 also shows that, after an unusually good decade, “real”
long-term interest rates have returned to “normal” levels of roughly
3% over inflation. Rates are now 6–7% (4–5%). Returns greater
than 6–7% (4–5%) will require a continued decline in interest
rates, either because inflation continues down or because public
enthusiasm for bonds causes an overshoot beyond fair value. While
we believe each of these possibilities have a slightly greater
than 50% probability (now completed), bond returns will no longer
be driven by the unusually high real interest rates of the last
decade. The time to be heavily invested in long bonds has just
come to an end (even more true today).
Chart 5 also can be read as the return investors expected from
their purchases of bonds in each of the past 41 (51) years. In
fact, these returns were guaranteed. Chart 2 shows the returns
investors actually received. Realized returns were well below
guaranteed returns until 1981.
Conclusion
Although we haven’t yet constructed a chart for stocks similar
to Chart 6, we judge the average stock to be priced to return
9–10% (8–9%). The caveat is that stocks are normally more sensitive
to public hopes and fears than are bonds, so corrections of 5–15%
can occur at any time. Partly, this is because stock prices are
reported on the news every day and played up by the media and
the brokerage community, while bond prices are largely ignored.
Frankly, the most likely trigger for such a correction in stocks
in the current environment would be an “uptick” in interest rates,
and therefore, a decline in the price of bonds. When we put all
of the above together, we see the following current conditions:
