Ch. 8: Conglomerate Firms and Internal Capital Markets 459
by underinvesting in divisions with better growth opportunities and overinvesting in di-
visions with worse opportunities. Second, they test their model’s predictions about the
relation been distortions and the diversity of the firm’s operations.
RSZ find that diversified firms invest more in segments with good opportunities than
in segments with poor opportunities. However, conglomerates might still misallocate
investment flows relative to comparable single-segment firms. Specifically, their the-
oretical model predicts that segments with good investment opportunities and above
average resources will transfer assets to segments with poorer investment opportunities
and below average resources.^32 The purpose of the transfer is to reduce the threat that
segment with poorer investment opportunities and resources will expropriate the better
segments ex-post, thereby improving the better segments’ investment incentives.
RSZ cannot directly observe resource transfers between a diversified firm’s segments.
Instead, they have to infer those transfers for each segment by comparing the segment’s
investment to the investment of comparable single-segment firms. They attribute differ-
ences between the actual investment and the investment of comparable single-segment
firms to transfers across divisions. However, RSZ also allow for the possibility that
conglomerates may systematically over-invest relative to single-segment firms because
they have better access to capital. Thus RSZ measure of the extent to which a segment
deviates from its benchmark, the Industry-Adjusted Investment (IAI), subtracts out the
weighted average industry-adjusted investment across all the segments of a firm. Thus,
IAIjkt=
Ik
BAk
−
Ikss
BAssk
−
∑n
k= 1
wk
(
Ik
BAk
−
Ikss
BAssk
)
,
whereIkis the investment in segmentk,BAkis the book value of assets in segmentk,
is the (asset-weighted) ratio of the capital expenditures to assets of comparable single-
segment firms, andwkis the ratio of segmentk’s assets to the firm’s assets.
In the econometric model they take to data, RSZ predict that a segment’s invest-
ment depends on the magnitude of its asset-weighted investment opportunities relative
to those of the rest. In particular, their model predicts that an increase in diversity
should decrease investment in segments that have asset-weighted investment oppor-
tunities above the firm average, and increase investment in segments below the firm
average.
To test their model, RSZ divide up the segments of each diversified firm in each
year along two dimensions (above vs. below average investment opportunities, above
vs. below average resources) to obtain a 2×2 classification matrix of all the segments
in their sample. Then for each firm in each year they sum up the IAIs for the firm’s
segments that fall into each cell receive (thus, in each year each firm will have four
observations for the transfers, one for each cell, although some may be missing).
(^32) The RSZ model is discussed above in Section3.4.