Chapter 12 Cash Flow Estimation and Risk Analysis 38112-6 WITHIN!FIRM AND BETA RISK
8
Sensitivity analysis, scenario analysis, and Monte Carlo simulation as described in
the preceding section dealt with stand-alone risk. They provide useful information
about a project’s risk; but if the project is negatively correlated with the " rm’s
other projects, it might actually stabilize the " rm’s total earnings and thus be rela-
tively safe. Similarly, if a project is negatively correlated with returns on most
stocks, it might reduce the " rm’s beta and thus be correctly evaluated with a rela-
tively low WACC. So in theory, we should be more concerned with within-" rm
and beta risk than with stand-alone risk.
Although managers recognize the importance of within-" rm and beta risk,
they generally end up dealing with these risks subjectively, or judgmentally,
rather than quantitatively. The problem is that to measure diversi" cation’s ef-
fects on risk, we need the correlation coef" cient between a project’s returns and re-
turns on the " rm’s other assets, which requires historical data that obviously does
not exist for new projects. Experienced managers generally have a “feel” for
how a project’s returns will relate to returns on the " rm’s other assets. Gener-
ally, positive correlation is expected; and if the correlation is high, stand-alone
risk will be a good proxy for within-" rm risk. Similarly, managers can make
judgmental estimates about whether a project’s returns will be high when the
economy and the stock market are strong (hence, what the project’s beta should
be). But for the most part, those estimates are subjective, not based on actual
data.
However, projects occasionally involve an entirely new product line, such as a
steel company going into iron ore mining. In such cases, the " rm may be able to
obtain betas for “pure-play” companies in the new area. For example, this steel
company might get the average beta for a group of mining companies such as Rio
Tinto and BHP, assume that its mining subsidiary has similar characteristics, and
use the average beta of the “comparables” to calculate a WACC for the mining
subsidiary. While the pure-play approach makes sense for some projects, it is actu-
ally rare. Just think about it. How would you " nd a pure-play proxy for a new in-
ventory control system, machine tool, truck, or most other projects? The answer is,
you couldn’t.
Our conclusions regarding risk analysis are as follows:
- It is very dif" cult, if not impossible, to quantitatively measure projects’ within-
" rm and beta risks. - Most projects’ returns are positively correlated with returns on the " rm’s other
assets and with returns on the stock market. This being the case, stand-alone
risk is correlated with within-" rm and market risk; so not much is lost by
focusing just on stand-alone risk.
(^8) This section is relatively technical, but it can be omitted without a loss of continuity.
SEL
F^ TEST Explain brie! y how a sensitivity analysis is done and what the analysis is
designed to show.
What is a scenario analysis, what is it designed to show, and how does it dif-
fer from a sensitivity analysis?
What is Monte Carlo simulation? How does a simulation analysis di# er from
a regular scenario analysis?