1 Risk-and-Return-Basics
1 Risk-and-Return-Basics
1 Risk-and-Return-Basics
Example:
Buy a stock on Jan. 1 for $100
Sell the stock on April 1 for $102
What is the annualized return?
Annualized return assumes we had the same return for an entire year.
Quarterly Return =
ln (102
100 ) = 1.98% if we assume continuous compounding
Notice that in each case, the PV of 100 becomes a FV of 108.24 at the end of the year.
Another Example:
Suppose you buy a non-dividend-paying stock for $100, hold it for 5 years and then find
that it is worth $155. Over that time period you have a simple return of 55%. Your
annualized return has been:
ln (155
100 ) = 43.8%/5 = 8.77% if we assume continuous compounding
5
1
Calculating Average Annual Returns and Standard Deviations from HISTORIC DATA
= [Π (1+Rt)]1/T – 1
Geometric Average Return – The compounded annual return earned by an investor who
bought the security and held it for ‘T’ years. This is equivalent to earning this return each
year and reinvesting the earnings at the end of each year.
Example: Let’s take the annual returns for a hypothetical stock which you bought at the
beginning of 2001:
2001 5%
2002 9%
2003 -12%
2004 20%
Arithmetic Mean
The Arithmetic Mean is approximately equal to the Geometric Mean plus one-half the
variance. This means that the larger the variance of the data, the greater will be the
difference between the arithmetic and geometric means.
If we want to calculate the Variance and Standard Deviation, we must use the arithmetic
mean.
2 = variance = Average squared difference between actual return and average return.
Always uses arithmetic average return
2
(∑ ( )
T
1
σ 2= Rt − R̄ )2
T t =1
σ =√ .013225
= .115 = 11.5%
Note though that if these 4 years are only a sample of the annual returns for this stock
and we wish to estimate the variance and standard deviation of the population (we can’t
know them for sure without all the data), we must divide by T-1 instead of T. In this
example, that would give us a sample variance of .0176 and a sample standard deviation
of 13.279%.
Here are the 1926 – 2011 for large-cap stocks, Corporate Bonds, and Treasury Bills.
1926 – 2011
Investment Arith. Mean Return Geom. Mean Return Standard Deviation
Stocks (large cap) 11.8% 9.8% 20.3%
Treasury Bonds 6.1 5.7 9.8
Treasury Bills 3.6 3.6 3.1
Note: R̄
Basic Principle of Finance (one of them): There is an expected reward for bearing risk.
We use U.S. Treasury yields as our proxy for the risk-free rate.
Any return greater than a treasury’s return is a reward for bearing risk – We call it
a risk-premium.
The risk premium you come up with is determined, in part, by how you measure it.
3
Do we use 1926-2011, or some other time period?
Arithmetic mean or geometric mean returns?
Treasury Bond or Treasury Bill yields as the risk-free rate?
We will discuss these issues later in the course.
The key is – what do you expect the risk premium to be in the future?
The generally accepted risk premium for the S&P 500 is usually between 2% and 8%.
For uniformity, we will often use 5.7% in our examples.
If we assume a normal distribution for the population (note that these 82 years are a
sampling from that population), then we can predict the probability of a single year’s
return for each security type.
Note: When we look at this year’s probabilities, we don’t look at the historic treasury
yield as the risk-free rate. We look at today’s treasury yield. Let’s say that it’s
approximately 3%.
So, using this method, the most likely return for stocks this year is:
And since about 2/3 of the probability distribution lies within one standard deviation of
the expected value, we can say that there is a 2/3 unconditional probability that the S&P
500 will post a return between -11.6% and 29.0% this coming year.
Suppose you wanted to know what is the unconditional probability that stocks will return
more than 25% next year.
If we assume that returns are normally distributed (they aren’t, but it is close enough for
this sort of an exercise), we want to find the area of the probability distribution to the
right of 25% when the mean is 8.7% and the standard deviation is 20.3%. As you
remember from statistics, this involves calculating z-values and perhaps using tables to
find the area under the curve. Fortunately, Excel does all this for us. If we go to
NORMDIST in the Excel functions, it asks us for the value of X – which is the point we
are looking for (in this case, it is .25). Then, we enter the mean of the distribution (.087)
and the standard deviation (.203). Finally, under Cumulative, enter True.
The result will be the area under the normal curve to the left of the point in question (in
our example, the probability you will get a return that is less than 25% is .789 – or about
79%). To find the area to the right, just subtract that value from 1.0. So there is about a
21% unconditional probability that stocks will earn a return greater than 25% next year.