456 V. Maksimovic and G. Phillips
firms. The former have no history of successful drug development and are typically
focused on one stand-alone project, such as the development of a specific drug, whereas
the latter usually have the option of picking among several projects to develop. To the
extent that the projects are discrete, the large firms closely resemble the theoretical
model of internal capital markets inStein (1997).
Guedj and Scharfstein analyze a sample of 235 cancer drugs that entered clinical
trials in the period 1990–2002. In order to be marketed in the U.S. a drug has to undergo
three separate phases of clinical trials. In each phase more information is revealed about
the drug’s prospects. These trials are expensive, and after each phase is completed the
sponsoring firm must determine whether to proceed onto the next stage or whether to
curtail the development of the particular drug.
Guedj and Scharfstein find that standalone firms are more likely to push drugs that
have completed Phase I trials into Phase II trials. However, standalone firms also have
much worse results at Phase II. This pattern especially evident for those standalone
firms that have large cash reserves. Thus, as inStein (1997), single-product firms do not
abandon projects optimally, whereas managers of multi-project firms shift resources
in response to new information. In that light firm diversification can be viewed as a
response to an agency conflict between the managers of single-product firms and share-
holders.
Khanna and Tice (2001), Campello (2002), andGuedj and Scharfstein (2004)iden-
tify several specific advantages of internal capital markets.Lamont (1997)identifies
a potentially countervailing disadvantage: a tendency to transmit investment shocks to
the firm’s main division to unrelated projects.^29 We next look at attempts to analyze the
effect of internal capital markets on a broader scale.
4.3. Efficient internal capital markets
Stein’s (1997)model suggests that there is a positive relation between the internal mar-
ket’s efficiency and the amount of external capital a diversified firm raises. Moreover,
the efficiency of external capital markets is greater when a firm has more divisions and
when the investment opportunities across divisions are not correlated.
Peyer (2001)and alsoBillet and Mauer (2003)test predictions on how conglomerate
firms allocate firms across divisions. For diversified firms, Payer estimates the firm’s
excess external capital raised as the difference between the firm’s use of external capital
compares and an estimate of how much a matching portfolio of single-segment firms
would have used.
For each diversified firm Peyer obtains the amount of external capital used as the
difference between the external capital raised from outside investors and the external
capital returned to outside investors.
(^29) Maksimovic and Phillips (2002)also find that the operations of peripheral units of conglomerates are cut
back much more severely in recessions than their main units. It is unclear whether these cuts occur because
of a reduction of the resources available to the firm’s internal capital market or because a shock triggers off a
re-evaluation of the firm’s long-term strategy.Schnure’s (1997)results suggests that it might be the latter.