7: Risk, Return and the Capital Asset Pricing Model
investment will perform. For example, in Chapter 6 we saw ample evidence that investors should expect
higher returns on shares than on bonds. Intuitively, that makes sense, because shares are riskier than
bonds, and investors should expect a reward for bearing risk. However, the claim that expected returns
should be higher for shares than for bonds does not imply that shares will actually outperform bonds
every year. Rather, it means that it is more likely that shares will outperform bonds than vice versa.
In this chapter, we want to establish a link between risk and expected returns. To establish that link,
we must deal with a major challenge: expected returns are inherently unobservable. Analysts have many
techniques at their disposal to form estimates of expected returns, but it is important to remember that
the numbers produced by these models are just estimates. As a starting point, let’s see how analysts might
use historical data to make educated guesses about the future.
7-1a THE HISTORICAL APPROACH
Analysts employ at least three different methods to estimate an asset’s expected return. The first method
relies on historical data and assumes that the future and the past share much in common. Chapter 6
reported an average risk premium on US shares relative to Treasury bills of 7.5% over the 111 years
to 2010. A 2012 report by NERA Economic Consulting suggested that a premium of 7.0% would be
appropriate for Australia.^1 More recent reports suggest that this may have fallen to 6%.2, 3
In Australia, the equivalent of a Treasury bill is a short-dated instrument called a Treasury note. If a
Treasury note currently offers investors a 2% yield to maturity (YTM), then the sum of the note yield and
the historical equity risk premium (2.0% + 6.0% = 8.0%) provides one measure of the expected return
on shares.
Can we apply that logic to an individual share to estimate its expected return? Consider the case
of OrotonGroup Limited. Oroton shares have been listed on the ASX since 1987, so we can calculate
its long-run average return, just as we did for the US share market. Suppose that over many decades,
Oroton’s return has averaged 15.0%. Suppose also that over the same time period, the average return on
Treasury notes was 4.0%. Thus, Oroton shareholders have enjoyed a historical risk premium of 11.0%.
Therefore, we might estimate Oroton’s expected return as follows:
Oroton expected return = Current Treasury note rate + Oroton historical risk premium
Oroton expected return = 2% + 11% = 13%
Although simple and intuitively appealing, this approach suffers from several drawbacks. First, over
its long history, Oroton has experienced many changes, including executive turnover, technological
breakthroughs in manufacturing and increased competition from domestic and foreign rivals. This
suggests that the risks of investing in Oroton have changed dramatically over time, so the risk premium
on Oroton shares has fluctuated too. Calculating Oroton’s historical risk premium over many years
blends all these changes into a single number, and that number may or may not reflect Oroton’s current
status. Thus, the historical approach yields merely a naïve estimate of the expected return. Investors
need to know whether Oroton’s shares today are more risky, less risky or just as risky as the long-term
premium indicates.
1 The Black CAPM, Report for APA Group, Envestra, Multinet & SP AusNet, NERA Consulting Group, March 2012.
2 Australian Valuation Practices Survey 2015, KPMG, 2015.
3 Market Risk Premium used in 88 countries in 2014: a survey with 8,228 answers, Pablo Fernandez, Pablo Linares and Isabel Fdez. Acín, IESE
Business School, June 20, 2014.