9.4.1 Long-term History
Characterizes Risk and Return
The history of
several U.S. stock markets are captured in Figure 9-2.
In essence this chart captures the effectiveness of capitalism over
the last 81 years. The numbered events in Figure 9-2 are
taken from the historical events in Table 9-2 below
it, titled
“Market Turmoil and the Dow Jones Industrial Average.” Despite
several set backs, capitalism continues to work. Also note that the
value of a dollar scale is a log scale, so each unit increases by
a factor of 10. These are indexes and therefore the growth of a dollar
does not reflect any fees or transaction costs. This long-term history
of quality data allows investors to create the best set of probabilistic
estimates of future performances of these indexes.
Figure
9-2
Table 9-2
IFA Index Portfolios
have also shown tremendous long-term despite the impact of short-term
bear markets. Click here to view a pdf of the growth of a dollar
invested for the last 69 years in Index Portfolios 5, 50 and 100
as well as for the S&P 500.
9.4.1a Effects of Government Intervention on Equity Returns
History shows that after nearly every major economic downturn, questions
arise as to whether the free market system remains an appropriate
way to organize and direct the nation’s resources.
Many individuals may be surprised to learn that government intervention
can play a key role in free market systems. Milton Friedman, widely
known as the most vocal proponent of the free market system cited that
the true cause of the Great Depression was the US government’s
failure to act swiftly to inject capital into the failing banking system.
"The Federal Reserve system stood idly by when it had the power and the
duty and the responsibility to provide the cash that would have enabled the banks
to meet the insistent demands of their depositors without closing their doors," Friedman
stated in “Free to Choose 3: Anatomy of a Crisis.”
The chart below shows the relationship between equity returns and economic
freedom rank. Economic freedom rankings data from Heritage
Foundation awards their rankings in consideration of 10
specific elements.
As the chart shows, the US ranks very high in the area of economic
freedom, while France came in significantly lower. It would be widely
determined then, that the equity returns of a more Socialist-leaning
France would be lower than those of the US. The reality, however, is
quite the opposite. The chart’s vertical axis measures the equity
returns of the countries. It shows that higher returns over the 39-year
period were not always delivered to the countries with the highest
degrees of economic freedom. Notoriously socialist-leaning countries
relative to the US include UK, Canada, Sweden, France, Norway, Belgium
and Denmark. The 39-year annualized returns of each of these countries
defy the presumption that increased returns come from increased economic
freedom.
Figure
9-2A
The figure directly below depicts the annualized standard deviation,
or the Risk, of each of the above countries, plotted against their
annualized return, the Reward, over the last 39 years.
Figure
9-2B
The bar chart directly below depicts the 39-year returns shown in
figure 9-2A..
Figure
9-2C
The bar chart below shows the 10-year returns for countries based
on their economic freedom rankings, as well. As you can see, in both
long-term and short-term data, economic freedom indicators dispute
the commonly held belief that government intervention hampers returns.
While the data presented here may seem surprising, the explanation
is very straightforward. Just as value investments demand a higher
return relative to growth investments to compensate for the higher
risk associated with them, so too should investments in countries with
increased government intervention demand higher expected returns to
compensate investors for the increased perceived risk of investing
in them.
This research, once again points to the simple and profound truth
that investment returns come from investment risk, proving once again
that there is no free lunch — even for perceived free market
economic systems.
The global history
of the size and value effect on stocks is made even more clear by
reviewing
Figure 9-3. Next, Table 9-4 provides
a thorough analysis of many indexes over the 1927 to 2008 period. Both
the chart and table indicate that over the 82-year period, small-value
has outperformed the S&P 500 and large-cap growth. Also, it is
clear that value has higher returns in international and emerging markets,
even though available data only dates back to 1982 for international
and 1989 for emerging markets.
Figure 9-3
Figure
9-4
Table
9-4
Table
9-3
To expand the range of asset
classes to include art, farmland and gold, let’s take a look at
Table 9-3.
It is interesting that over
the 48-year period emerging market public equities outperformed venture
capital, and at a lower risk level. In addition, the S&P 500 outperformed
real estate by more than 50%, although the S&P 500 had about three
times the risk. Figure 9-4 graphs the data from Table
9-3 on the Markowitz risk/return plot and adds in index portfolios
5, 50 and 100 for comparison. Note where venture capital and emerging
markets sit on the plot. Gold and silver are also interesting, reinforcing
the idea that they have lots of risk and returns pretty close to T-bills
and bonds.
Table
9-5
Venture
Economics, an information provider for equity professionals, compiled
a 20-year data series of various types of private equity strategies
for the period ending December 31, 2005. According to the survey,
venture and private equity strategies generally performed well over
the period. But, the premium relative to public securities appears
rather small considering the higher risk, investment concentration,
absence of liquidity, transparency and daily pricing. The results
are shown in Table
9-5.
9.4.1b Probability of Portfolio Recovery
Rare and severely punishing drops in the stock market can find investors
wondering how long it might take for their portfolios to recover from
a big loss.
The table below
shows the percentage amount of loss for the S&P
500 Index as well as for IFA Index Portfolios 90, 70, 50, 30, and 10
during the 23-month time period from November 2007 through December
2009, as well as the percentage gain that is required to restore each
portfolio to its end of October 2007 high.
The probabilities
of achieving those post-drop recoveries are set forth in the line
graph below the table which shows the probability of each portfolio
recovering within a specified time period from 1-year through 20
years. The probability studies were created using 81 years of historical
rolling period returns data for each Index Portfolio and the S&P
500 Index. The y-axis in the line chart below expresses the probability
that each portfolio’s recovery will occur in
the number of years expressed along the x-axis. For example, the IFA
Index Portfolio 70 has a 91% probability of a full recovery or better
in less than 6 years from the first day of the end of the time period
stated.
Figure
9-4A
Figure
9-4B
Figure
9-4C
Figure
9-4D
9.4.2 Cross Correlation
among Indexes
In addition to the long-term
risk and return of indexes, a third input used to create optimal portfolios
is cross correlation. Cross correlation refers to the extent to which
performances of different asset classes move in relation to each other.
The lower the correlation among different indexes in a portfolio, the
greater the diversification, which means lower volatility of returns.
If indexes are highly correlated,
then their prices are responding to market news in the same direction
at the same time. Market news that affects prices in all markets, include
the overall strength of the U.S. economy, consumer confidence, the level
of interest rates and expectations for inflation rates. A low correlation
means that market prices of different indexes react in different directions
to the same news. These indexes have market price movements that are
not connected, showing a low similarity in movement to each other.
For example,
stocks and fixed income historically have a low correlation. As seen
in Figure
9-5, large company stocks and
one-year fixed income have a very low correlation of 0.02,
which means that there’s almost no correlation between the
market price movements of these two asset classes.
The next best diversifier
of risk is low positive correlation among asset classes in a portfolio.
By designing the proper mix of low correlation index funds, it is possible
to lower a portfolio’s risk and increase its risk-adjusted return
at the same time. More historical data on the correlation among indexes
found in the global financial markets appears in Figure 9-5.
Figure
9-5
Figure
9-6
The data in Figures
9-6, 9-7and 9-8 is attributable to the three
risk factors documented by Eugene Fama, Kenneth French, and Jim Davis.
These factors are used in a multiple regression analysis to risk adjust
returns of other investments and to establish the cost of capital of
firms that sell their equity. Remember that a firm’s cost of capital
is equal to the investor’s expected return. The Fama/French data
indicates that these three factors explain 95% of stock returns in diversified
portfolios. In those calculations, average instead of annualized returns
are used. The average annual returns of these risk factors are known
as the risk premiums.
A Comparative History
of Several Indexes using Rolling Periods
At times investors doubt whether the fundamentals of capitalism and
the relationship between risk and return will hold up in the future.
For example, the August 13, 1979 issue of BusinessWeek featured this
question on the cover: “Are Equities Dead?” After 10 years
of lousy performance, it really must have appeared that way. For the
11-year period of 1969 to 1979, the S&P 500 average annual compound
return was only 4.5%. And, it was even worse, 3.2%, for the more than
seven-year period of 1973 to 1979, just before the article. These kinds
of returns made it seem as if stocks were no longer a viable investment.
Thus, many investors decided to invest only in Treasury bills, which
outperformed stocks for both periods, and avoid the risk of stocks.
Of course, the concern that the fundamental relationship between risk
and return wouldn’t hold up was as ridiculous then as it is now.
An analysis
of multiple year rolling periods offers an interesting way to sort
out these kinds of concerns. For example, if you look at the
flash chart below and select S&P vs Treasury
Bills, you
will see that we have 853 ten-year periods shifting one month at
a time, over the 81 years from 1928 to 2008. Of those 853 periods
the S&P
500 Index outperformed T-bills 85% of the time. In the 733 twenty-year
periods it out performed T-bills 100% of the time. And in one-year
periods, it outperformed only 67% of the time. This brings to
mind Benjamin Graham's famous observation that, “In the short
run, the market is a voting machine, but in the long run it is a
weighing machine.”
In "The
Little Book of Common Sense Investing", page 160,
there is a mention of the difference
between value stocks and growth stocks returns. Here is a table comparing these
returns, growth of $1, and standard deviations of returns over
the periods mentioned and also an 81 year period (all the data available for
these dimensions). Click here for backtested
data sources and disclosures. The links in the IFA Index column will take you
to the IFA
Risk Return Calculator, which was used to calculate the data.
Table
9-A
Table
9-C
Table
9-B
Table
9-D
The
Flash chart below offer numerous comparisons of this kind
of data and they are very helpful in understanding the
comparisons of various indexes. This data shows that large
value does not always outperform large growth stocks. In
fact, the size and value risk factors come and go unpredictably.
This is consistent with the Random Walk Theory of changes
in stock prices. In addition, the cycle of good or bad
returns for small company stocks compared to large company
stocks can last for many years.
Figure
9-9
A Comparative History of Market Cap Deciles
Figure 9-10 clearly lays out the history
of the size effect. The several charts breaks out a
number of time periods in history to illustrate the
diversifying power of small-cap stocks. This chart
is created using CRSP market capitalization data broken
down into one-tenth size buckets, referred to as deciles.
All 10 deciles are then measured and charted in different
time periods. It illustrates that especially in shorter
periods, small company stocks don’t always outperform
large company stocks, but as seen in the top left chart,
over the whole time period of 1927 to 2006, there is
a clear advantage to have some exposure to small companies.
But, in shorter periods anything can happen. For example,
during the five-year period of 2002 to 2006, small
company stocks widely outperformed large company stocks,
while during the seven-year period of 1984 to 1990,
and six years from 1994 to 1999, large-cap stocks were
the king of the hill.
Figure
9-10
The
Returns Matrix
The use of a return matrix is yet another interesting way to
look at long-term data. Figures 9-11 bring
together annual and annualized returns covering every combination
from 1974 to 2008 for an index portfolio 90 (see Appendix A).
This big triangle identifies the years along all three borders.
The intersection of any two years shows the annualized return
over that period. The diagonal lines show one year returns
on the first diagonal and rolling period returns can be found
on each diagonal line below the first one. For example, the
first gray diagonal shows five year rolling periods from 1974
to 2008. The very bottom left hand corner shows the annualized
return over the entire 35 year period, which is 12.5% for index
portfolio 90.
*How to read the Annualized Returns Matrix: You can locate
the annualized compounded rate of return for this simulated
Index Portfolio for a designated time period by following
these easy instructions: Locate the column for the beginning
year of the period. Years are labeled at the top and the
bottom of each column. Then, locate the ending year of the
period on the left-most vertical column. The annualized return
can be found where the first year's column intersects with
the ending year's row. IFA advisory fees of 0.9% per year
and DFA mutual fund expense ratios have been deducted from
these results. The 10-Yr diagonal (highlighted, starting
from far left column) represents the estimated average holding
period for investors who score 90 on the Risk Capacity Survey
at ifa.com. Sources, Updates, and Disclosures: ifabt.com.
9.5
Summary
A good understanding of the
long-term historical risk and return of various indexes enables an investor
to know how to construct an efficient asset allocation according to
risk capacity. Risk and return will work themselves out or revert to
the mean over the long run. In the meantime, the best bet is to diversify
among index funds that are structured for optimal exposure to risk factors
that history has shown to be most rewarding.
9.6
Review
Questions
Please answer the following questions
before moving on to the next Step.
1. Stock markets are best characterized when looking at:
a. 1 year period
b. 5 year periods
c. 80 year periods
d. 3 year periods
2. The long-term
characteristics of indexes are important because:
a. they better reflect the differences between capital and capitalization
b. margin rules are the same throughout history
c. favored industries change with time
d. the law of large indexes is not applicable to market returns
3. Many
high net worth investors try to get allotments of venture capital partnerships.
According to Morgan Stanley, over a 48-year period venture capital
had a 16% return and a 35.4 risk index. Emerging market equities over
the same period had the following:
a. 4.9% return, 26 risk
b. 16% return, 29.6 risk
c. 5.4% return, 6.2 risk
d. 12.7% return, 8.2 risk
4. Many
people look at 80-year risk and return data and say that it is not relevant
to them because they don’t have 80 years to invest. This is faulty
logic because:
a. the basic concept of sampling error means short-term data is worse
than long-term data
b. we have 500 years of stock market returns, so that is better than
80 years
c. none of the above
d. A and B
e. we have 500 years of stock market returns, so that is even better
than 80 years
5. The index
with the highest return since 1928 is:
a. large growth index
b. large value index
c. small growth index
d. small value index
e. total market index
