Appendix G  831
Local CAPM
Some practitioners and academic researchers propose estimating the cost of 
capital for an investment opportunity in a particular country by using a local 
CAPM. The investment’s beta is then estimated versus the market portfolio 
of the country, and the market risk premium follows from the excess return 
of that same market portfolio over the local risk-free rate. The approach is 
theoretically correct if stocks are correlated to the global market portfolio only 
through the local market:6
β
β
β
j G
j L
L G
,
,
,
=
×

(G.3)
where β
β
j G
j L
j
G
,
,
=
=
beta of asset
versus global market portfolio
beta of asset
versus local market portfolio
beta of local market port
j
L
L G
β , =
folio
versus global market portfolio
L
G
This implies that any international risk factors influencing the returns of com-
panies in a given country are fully captured by the local market portfolio of 
that country. You can then indirectly estimate any asset’s global beta by mul-
tiplying its local beta by the global beta of the local market. If the local stock 
market is fully integrated and correctly priced in the global market, its ex-
pected return is:
E r
r
E r
r
L
f
L G
G
f
( )
[ (
)
]
,
=
+
−
β

(G.4)
where r
L
r
r
L
f
=
=
expected return for local market portfolio
risk-free rate
G
G
= return for global market portfolio
Combining Equations G.3 and G.4 shows that the expected return for a stock 
j estimated via the local and global CAPM should be equal as well. Following 
the global CAPM, this return is given by:
E r
r
E r
r
j
f
j G
G
f
( )
[ (
)
]
,
=
+
−
β
Substituting the asset’s global beta by the indirect beta defined previously in 
Equation G.3 leads to:
E r
r
E r
r
j
f
j L
L G
G
f
( )
[ (
)
]
,
,
=
+
×
−
β
β
6 See R. Stulz, “The Cost of Capital in Internationally Integrated Markets: The Case of Nestlé,” European 
Financial Management 1, no. 1 (1995): 11–22.

832  Appendix G
This can be rearranged to show equivalence with the local CAPM:
E r
r
E r
r
j
f
j L
L
f
( )
[ ( )
]
,
=
+
−
β
Although the assumptions may not seem very realistic at face value, there 
is evidence that the local and global CAPM generate similar results. Empirical 
research finds that the cost of capital estimated for U.S. companies with a local 
CAPM is very close to the estimate based on a global CAPM.7 For U.S. stocks, 
this may not be surprising, as the U.S. market portfolio is well diversified and 
highly correlated with the global market portfolio. But supporting evidence 
also comes from nine developed economies, including not only the United 
States but also the United Kingdom, Germany, France, and smaller economies 
such as the Netherlands and Switzerland. An analysis of beta estimates for 
companies versus a local and global market portfolio has shown that for these 
countries, the betas are typically related, as indicated by Equation G.3.8
However, the local CAPM approach, when compared with the global 
CAPM, has some practical drawbacks. First is that when you apply the local 
CAPM to investments in different countries, you should estimate the local 
market risk premium and beta for each of these countries, instead of only 
the global market risk premium, as you would do when applying the global 
CAPM. Also, with a local CAPM, you cannot make a straightforward estimate 
of a company’s beta based on the average of the estimated betas for a sample 
of industry peers (which Chapter 15 recommends to reduce the standard error 
of the company’s beta). The reason is that if the peers are in different countries, 
their local betas are not directly comparable. Finally, local risk premiums are 
typically less stable over time than their aggregate, the global risk premium. 
For example, Exhibit G.1 compares the realized premiums on local stock mar-
ket indexes with government bond returns for several countries and a globally 
diversified portfolio, using data from Dimson, Marsh, and Staunton’s analy-
sis of long-term average returns on equities and corporate and government 
bonds.9 The individual countries’ risk premiums vary considerably, depend-
ing on the time period over which they are measured, while the global pre-
mium remains almost unchanged.
Note that the risk premium differences shown in Exhibit G.1 do not mean 
that the price for risk varies across these countries. These differences are driven 
by several factors. First, levels of economic development and, therefore, profit 
7 R. Harris, F. Marston, D. Mishra, and T. O’Brien, “Ex-Ante Cost of Equity Estimates of S&P 500 Firms: 
The Choice between Domestic and Global CAPM,” Financial Management 32, no. 3 (2003): 51–66.
8 See C. Koedijk, C. Kool, P. Schotman, and M. van Dijk, “The Cost of Capital in International Financial 
Markets: Local or Global?,” Journal of International Money and Finance 21, no. 6 (2002): 905–929.
9 E. Dimson, P. Marsh, and M. Staunton, Triumph of the Optimists: 101 Years of Global Investment Returns 
(Princeton, NJ: Princeton University Press, 2002); and E. Dimson, P. Marsh, M. Staunton, and J. Wilmot, 
Credit Suisse Global Investment Returns Yearbook 2016 (London: Credit Suisse Research Institute, 2016).