434 V. Maksimovic and G. Phillips
on whether the firm is a conglomerate, denoted by the indicator variablesDitwhich
takes on the value 1 if the firm is a conglomerate and 0 if it is not.
Vit=β 1 +β 2 Xit+β 3 Dit+eit, (1)whereeitis an error term.
A necessary condition for the OLS estimate of coefficientβ 3 to be unbiased is forDit
to be independent from the error termeitin equation(1). The earlier literature, such as
Lang and Stulz, implicitly assume that this condition holds and that conglomerate status
can be treated as being exogenous in the estimation. But suppose instead that the firm’s
decision to operate in more than one industry depends on a set of characteristicsWitand
a stochastic error termuit. Specifically assume thatDit=1 whenλWit+uit>0 and
Dit=0 whenλWit+uit<0. Then, the coefficient of in equation(1)will be biased
if, as seems plausible, a common determinant of both the valueVitand the decision to
become conglomerate is omitted from estimated equation(1).
Several recent empirical papers on the conglomerate discount, byCampa and Ke-
dia (2002), Graham, Lemmon and Wolf (2002), Lamont and Polk (2002)andVilla-
longa (2003, 2004a, 2004b)address these issues.Maksimovic and Phillips (2002)gives
an equilibrium justification for the endogeneity of the discount and also empirically
shows that there is a size-efficiency relationship that holds for conglomerate and single-
segment firms.
The most direct evidence on the importance of self-selection in the determination
of conglomerate discounts is provided byGraham, Lemmon and Wolf (2002).^16 They
show directly that diversification through acquisitions creates a measured discount in
thesenseofBerger and Ofek (1995)even when the diversification is value increasing.
Using a sample of 356 mergers that occurred between 1978 and 1995 and (i) which
met the Berger and Ofek criteria of inclusion in the sample of diversifiers and (ii) for
which they had data on both the bidder and the target, Graham et al. show that acquirers
register a discount computed in the sense of Berger and Ofek in a two-year window
surrounding the acquisition. However, the greater part of this discount can be explained
by the fact that the targets are selling at a discount relative to single-segment firms prior
to the merger. Thus, much of the discount associated with corporate diversification by
acquisition cannot be attributed to the costs associated with operating more diversified
firms but can be attributed to the fact that diversifying firms are on average acquiring
assets already valued at a discount relative to the industry benchmarks. To the extent
that conglomerate firms engage in more acquisition activity than single-segment firms
(as shown inMaksimovic and Phillips, 2007), it is possible that their growth pattern
might induce a discount even when it is value maximizing.
Campa and Kedia (2002)also argue that the documented discount of diversified firms
is not by itself evidence that diversification destroys value. They use three alternative
(^16) SeeChevalier (2000)for a related argument.