CAHAPTER 5 RISK, RETURN, AND THE HISTORICAL RECORD

CAHAPTER 5
RISK, RETURN, AND THE HISTORICAL RECORD
 1

CAHAPTER 5 RISK, RETURN, AND THE HISTORICAL RECORD

5.1 Measuring Returns Over Different Holding Periods

We start with simplest security,a zero coupon bond. This is a bond that
pay its owner only one cash flow, for example, $100, on the maturity
date. The investor buys the bond for less than face value (face value is
just the terminology for the ultimate pay off value of the bond,, in this
case, $100), so the total return is the difference between the initial
purchase price and the ultimate payment of face value.

If we call P(T) the price paid today for for a zero-coupon bond with
maturity date T, then over the life of the bond, the value of the
investment grows by the multiple 100/PT. The percentage increase in
the value of the investment over this holding period is
R(T) = 100/P(T) - 1
We can equivalently write rate of return over the holding period as
Holding period return = r(T) = Price increase+Income/P(T)
= 100-P(T)+0/P(T)

We typically express an investment return as an effective annual rate

(EAR), defines as the percentage increase in funds per year.

Table 5.1 illustrates the technique.

Horizon, Price, r(T) = 100/P(T) - 1 EAR over Given
T P(T) Horizon
Half- $97.36 100/97.36-1=0.0271=2.71% (1+.0271)2-1=.0549
year
1 Year $95.52 100/95.52-1=0.0469=4.69 (1+.0469) -1 = 0.0469
25 Year $23.30 100/23.30- (1+3.2918)1/25=.060
1=3.2918=329.18%

Annual percentage rates:

Rates on short term investment are in practice often annualized
using simple interest that ignores compounding. These are called annual
percentage rate(APR).
1 + EAR = (1+ APR/n)n
Continuous compounding:
 2

CAHAPTER 5 RISK, RETURN, AND THE HISTORICAL RECORD

The differences between APR and EAR grows with the frequency
of compounding. We effectively approach continuous compounding
(CC), and the relation of EAR to the annual percentage rate, denoted by
rcc for the continuously compounded case, is given by the exponential
function

1 + EAR = exp(rcc) = ercc

5.2 Interest Rate and Inflation Rates

Forecasting interest rates is one of the most notoriously difficult parts of

applied macroeconomics. Nonetheless we do have a good
understanding of the fundamental factors that determine the level of
interest rates:
1. The supply of funds from savers, primarily households.
2. The demand for funds from business to be used to finance
investment in plant, equipment, and inventories (real assets or
capital formation).
3. The government’s net demand for funds as modified by actions of
the Federal Reserve Bank.
4. The expected rate of inflation.

Real and nominal rates of interest

An interest rate is a promised rate of return, usually denominated in a

specific currency, for example U.S. dollars. But even a guaranteed rate of
return specified in dollars does not necessarily lock in the increase what
you can buy with your investment proceeds. It is possible that in the
course of the year, prices of goods also will increase, in which case the
increase in your purchasing power will be less than the increase in your
dollar wealth.
Distinguish between a nominal interest rate the growth rate of you
money and a real interest rate the growth rate of your purchasing
power. More precisely, we find the proportional increase in purchasing
power by dividing the growth of invested funds by the growth of prices.
If we call rnom the nominal interest rate, r real the real rate, and i the
inflation rate, then we conclude
 3

CAHAPTER 5 RISK, RETURN, AND THE HISTORICAL RECORD

1 + rreal = 1 + rnom / 1 + I

In words, the real rate rate of interest is the nominal rate reduced by the
loss of purchasing power resulting from inflation.

The exact relationship in equation can be rearranged to

rreal = rnom - I / 1 + I
Which shows the approximate rule overstate the real rate by the
multiple 1 + I

The equilibrium real rate of interest

Three basic factors - supply, demand, and government actions -

determine the real interest rate. The nominal interest rate is the real
rate plus the expected rate of inflation.

The supply curve slopes from up from left to right because the higher
the real interest rate, the greater the supply of household savings. The
assumption is that at higher real interest rates, households will choose
to postpone some current consumption and set aside or invest more of
their disposable income for future use.
The demand curve slopes down from left to right because the lower the
real interest rate, the more business will want to invest in physical
capital. Assuming that businesses rank projects by the expected real
return on invested capital, firms will undertake more projects the lower
the real interest rate on the funds needed to finance those projects.
 4

CAHAPTER 5 RISK, RETURN, AND THE HISTORICAL RECORD

Equilibrium is at the point of intersection of the supply and demand

curves, point E in the figure.

Interest rate and inflation

The nominal rate of return on an asset is approximately equal to the real

rate plus inflation. The nominal rate must rise along with inflation to
maintain the expected real rate of return offered by an investment.
Irving fisher (1930) argued that the nominal rate ought to increase one-
for-one with expected inflation, E(i). The so called fisher hypothesis is
rnom = rreal + E(i)

Taxes and the real rate of interest

Tax liabilities are based on nominal income and the tax rate determined
by the investor’s tax bracket. Congress recognized the resultant “bracket
creep” (when nominal income grows due to inflation and pushes
taxpayers into higher brackets) and mandated index linked tax brackets
in the Tax Reform Act of 1986.
Index linked tax brackets do not provide relief from the effect of inflation
on h taxation of savings. Given a tax rate (t) and a nominal interest rate,
rnom, the after tax interest rate is rnom(1- t). The real after tax rate is
approximately the after tax nominal rate minus theinfltion rate:

Rnom(1 - t) - i = (rreal + i)(1 - t) - I = rreal(1 - t) - it

Treasury bills and inflation, 1926-2018

The Fisher equation predicts that the nominal rate of interest should
track the inflation rate, leaving the real rate somewhat stable. But our
discussion of the determination of the real rate implies that it too will
evolve with some uncertainty. The nominal rate will inherit the
uncertainty of both the inflation and the real rates.
 5

CAHAPTER 5 RISK, RETURN, AND THE HISTORICAL RECORD

Table shows the history of returns 1 month U.S. Treasury bills, the
inflation rate, and the resultant real rate.

Annual Rates Standard

Deviation

The figure shows why we divide the sample period 1952. After that year,
inflation is far less volatile, and probably as a result the nominal interest
rate tracks the inflation rate with fa greater precision, resulting in a far
more stable real interest rate.

5.3 Risk and risk premium

In this section, we will begin thinking about how to measure this risk.

Holding period returns:

You are considering investing in a mutual fund. The fund currently sells
for $100 per share. With an investment horizon of 1 year, the realize
rate of return on your investment will depend on (a) the price per share
at year’s end and (b) the cash dividends you will collect over the year.
 6

CAHAPTER 5 RISK, RETURN, AND THE HISTORICAL RECORD

Expected return and standard deviation:

There is considerable uncertainty about share prices one year from now,
so you can be sure about your eventual HPR. We can organize our belief
about possible outcomes for the mutual fund by posting various
economic scenarios as well as their probabilities.
We will characterize probability distributions of rates of return by their
expected or mean return, E(r), and standard deviation. The expected
rate of return is a probability-weighted average of the rates of return in
each scenario. Calling p(s) the probability of each scenario and r(s) the
HPR in each scenario, where the scenarios are labeled or “indexed” by s,
we write the expected return as

The variance of the rate of return is

Excess returns and risk premiums:

We measure the reward as the difference between the expected HPR on

the mutual fund and the risk free rate, that is, the rate you would earn in
risk free assets such as T-bills, money market funds, or the bank. We call
this differences the risk premium.
The degree to which investors are willing to commit funds to stocks
depends on their risk aversion. Investors are risk avers in the sense that
if the risk premium were zero, they would not invest any money in
stocks. In theory, there must must always be a positive risk premium on
stocks to induce risk averse investors to hold the existing supply of
stocks instead of placing all their money in risk free assets.

5.4 Learning from Historical Returns

 7

CAHAPTER 5 RISK, RETURN, AND THE HISTORICAL RECORD

Time series versus scenario analysis:

A set of relevant scenarios and associated investment rates of return,
assign probabilities to each, and conclude by computing the risk
premium (expected reward) and standard deviation (risk) of the
proposed investment. In contrast, asset return histories come in the
form of time series actual returns that do not necessarily match
investors’ original assessment of the probabilities of those returns; we
observe only dates and associated HPRs.

Expected returns and the arithmetic average:

If there are n observations, we substitute equal probabilities of 1/n for
each p(s). The expected return, E®, is then estimate by the arithmetic
average of sample rates of return:

The geometric (time weighted) average reurn:

The arithmetic average provides an unbiased estimate of the expected
future return.Intuitive measure of performance over the sample period
is the (fixed) annual HPR that would compound over the period to the
same terminal value obtained from the sequence of actual returns in the
time series.

Estimating variance and standard deviation:

Calculating variance by averaging squared deviations from our estimate
of the expected return, the arithmetic average. We again use equal
probabilities for each observation and use the sample average in place
average in place of the unobservable E(r). Therefore, using historical
data with n observations, we could estimate variance as
 8

CAHAPTER 5 RISK, RETURN, AND THE HISTORICAL RECORD

We can eliminate the bias by multiplying the arithmetic average of

squared deviations by the factor n/(n - 1). The variance and standard
deviation then become

Mean and standard deviation estimates from higher-frequency

observations:
Observation frequency has no impact on the accuracy of estimates of
expected return. It is the duration of a sample time series (as opposed to
the number of observations) that improves accuracy.
Ten annual returns provides as accurate an estimate of the expected
rate of return as 120 monthly returns. Because the average monthly
return must be consistent with the average annual returns, the
additional intra-year observations provide no additional information
about mean return.

The reward to volatility (sharpe) ratio:

Investors presumably are interested in the expected excess return they
can earn by replacing T-bills with a risky portfolio, as well as the risk they
would thereby incur. Even if T-bill rate is not constant we still know with
the certainty what nominal rate we will earn in any period if we
purchase a bill and hold it to maturity.
The importance of the trade-off between reward (risk premium) and risk
(as measured by standard deviation or SD) suggest that we measure the
attraction of a portfolio by the ratio of its risk premium to the SD of its
excess returns.

5.5 The Normal Distribution

 9

CAHAPTER 5 RISK, RETURN, AND THE HISTORICAL RECORD

The bell shaped normal distribution appears naturally in many

application. the heights and weights of newborns, the life spans of many
consumer items such as light bulb, and many standardized test scores
are well described by the normal distribution. Variables that are end
result of multiple random influences tend to exhibit a normal
distribution.

The normal distribution with mean 10% and standard deviation 20%.

The normal curve with meanof 10% and standard deviation of 20%. A
smaller SD means that possible outcomes cluster more tightly around
the mean, while a higher SD implies more diffuse distributions. The
likelihood of realizing any particular outcome when sampling from a
normal distribution is fully determined by the number of standard
deviations that separate that outcome from the mean. Put differently,
the normal distribution is completely characterized by two parameters,
the mean and SD.

5.6 Deviations from Normality and Tail Risk

One potentially important departure from normality relates to any

asymmetry in the probability distribution of returns.
 10

CAHAPTER 5 RISK, RETURN, AND THE HISTORICAL RECORD

Just as variance is the average value of squared deviations from the

average, skew is the average cubed deviation from the mean, expressed
as a multiple of the third power of the standard deviation

When negative deviations are raised to an odd power, the result remains
negative. Therefore, negative values of skew indicate that extreme bad
outcomes are more frequent than extreme positive ones.

Normal and skewed distributions (mean = 6%, SD = 17%)

Figure shows both positively and negatively skewed distributions as well

as the symmetric normal. Notice that the negatively skewed
distributions has a “fatter” left tail than the normal with higher
probabilities of extreme negative outcomes, while the positively skewed
distribution has a fatter right tail.
 11

CAHAPTER 5 RISK, RETURN, AND THE HISTORICAL RECORD

Normal and fat tailed distributions (mean = .1, SD = .2)

Kurtosis measures the degree of fat tails. It is calculated from the

average value of deviations raised to the fourth power (expressed as a
multiple of the fourth power of the standard deviation). because the
deviations are raised to the fourth power while they are raised only to
the second power when calculating variance, kurtosis is much more
sensitive to extreme outcomes and therefor is a natural measures of tail
risk, that is, the risk of outcomes in the far tail of the probability
distribution.

Value at Risk:
VaR is the loss corresponding to a very low percentile of the entire
return distribution. Therefore, it is another name for the quantile of a
distribution. The quantile, q, of a distribution is the value below which lie
q% of the possible values.

Expected shortfall:
VaR is the most optimistic measure of bad case outcomes as it takes the
highest return (smallest loss) of all these cases. in other words
investment loss at first percentile of the return distribution, but it
ignores the magnitudes of potential losses even further out in the tail.
 12

CAHAPTER 5 RISK, RETURN, AND THE HISTORICAL RECORD

There is two names for the values, expected shortfall (ES), or conditional
tail expectation (CTE). This expectation is conditioned on being in the left
tail of the distribution. ES is the more commonly used terminology.

Lower partial standard deviation and the sortino Ratio:

The use of SD as a measure of risk when the return distribution is non
normal presents two problems: the asymmetry of the distribution
suggest we should look at negative outcomes separately and because an
alternative to a risky portfolio is a risk free investment.
A risk measure that addresses these issues is the lower partial standard
deviation (LPSD) of excess returns, which is computed like the usual
standard deviation, but using only bad returns.

Relative frequency of large, negative 3-sigma returns:

Here we concentrate on the relative frequency of large, negative returns

compared with those frequencies in a normal distribution with the same
mean and SD. Extreme returns are often called jumps. We compare the
fraction of observations with returns 3 or more SD below th mean to the
relative frequency of negative 3-sigma returns in the corresponding
normal distribution.

5.7 Historic Returns on Risky Portfolios

We can now apply the analytical tools worked out in previous sections to
look at the historic performance of several important portfolios.
Comparing the performance of Treasury bills, Treasury bonds, and a
diversified portfolio of U.S. stocks. T -bills are widely considered the least
risky of all assets. Long term treasury bonds are also certain to be repaid,
but the prices of these bonds fluctuate as a interest rate vary so they
imposing meaningful risk. Common stocks are the riskiest of the three
group of securities. Return will depend on the success of failure of the
firm.
 13

CAHAPTER 5 RISK, RETURN, AND THE HISTORICAL RECORD

T-Bills T-Bonds Stocks

Average 3.38% 5.58% 11.72%
Risk premium N/A 2.45 8.34
Standard deviation 3.12 11.59 20.05
Max 14.71 41.68 57.35
Min −0.02 −25.96 −44.04

A global view of the historical record.

U. S. investors look to improve diversification by investing
internationally. Foreign investors that traditionally used U.S. financial
markets as a safe haven to supplement home country investments also
seek internationa diversification to reduce risk.

Figure shows a century plus long history (1900-2017) of average excess

return in 20 stock markets. The mean annual excess return across these
country was 7.4% and the median was 6.6%. The U.S. is roughly in the
middle of the pack, with a historical premium of 7.7%. Similarly, the SD
of returns in the U.S. was just a shade below the median volatility in
these other countries. So the U.S. performance has been pretty much
 14

CAHAPTER 5 RISK, RETURN, AND THE HISTORICAL RECORD

consistent with international experience. The characteristics of historical

return in the U.S. also can serve as rough indication of the risk return
trade off in a wider range of countries.

5.8 Normality and Long Term Investments

Long term performance cannot be normal. If r 1 and r2 are the returns in

two periods, and each has the same normal distribution, them the sum
of returns, r1 + r2, would be normal. But the two period compound to
(1+r1)(1+r2), which is not normal. This is more than a theoretical point.
The shape of the distribution changes noticeably as the investment
horizon extends. Longnormal distribution means that the long of the
final portfolio value, In(WT), is normally distributed.

Short run versus long run risk

Result on the risk and return on investment over different time horizons
appear to offer a mitigation of investment risk in the long run: because
expected return increases with horizon in propotion to T. while standard
deviation increases in proportion to the square root of T, the expected
return of a long term risky investment becomes ever larger relative to ts
SD.

Forecast for the Long Haul

We use arithmetic averages to forecast future rates of return because
they are unbiased estimates of expected rates over equivalent holding
periods. But the arithmetic average of short term returns can be
misleading when used to forecast long term cumulative returns. This is
because sampling error in the estimate of expected return will have
asymmetric impact when compounded over long periods. Positive
sampling variation will compound to greater upward errors than negaive
variation.