466 V. Maksimovic and G. Phillips
4.6. Mergers and acquisitions, divestitures and spinoffs
4.6.1. Diversified firms and the market for assets
The early theoretical literature on internal capital markets, such asStein (1997),ex-
plicitly recognizes the importance of the size of the internal market for its efficiency.
Thus, while the importance setting the firm’s boundaries for the quality of the internal
capital market was recognized early, much of the literature takes a partial equilibrium
approach and assumes that the firm’s boundaries as given. This is potentially important
since many of the hypothesized conflicts within the firm can be solved or mitigated by
selling assets that do not fit well with the firm’s total portfolio. Thus, if the market for
firms’ assets is efficient, the magnitude of the investment distortions that can be created
by conflicts within the firm is likely to be tightly bounded. Of course, there may still be
conflicts of interest between top management and shareholders. However, top managers
have clear incentives to maximize firm value.^39
An objection to this might be that the market for corporate assets is insufficiently
liquid so that firms which attempt to readjust their portfolio by selling segments face a
capital loss. This is unlikely.Tables 2A and 2B, fromMaksimovic and Phillips (2001),
shows that there exists a large, procyclical market for segments and individual plants.
Using Census, dataMaksimovic and Phillips (2001)show that in the period 1974–
1992, 1.94% of all manufacturing plants change ownership annually in partial-firm
transactions.^40 This is comparable to the total rate at which manufacturing plants change
ownership in all-firm mergers and takeovers over this period, 1.95% annually. Similar
rates of partial firm sales occur in both growing and declining industries. The market for
divisions and plants is a market dominated by conglomerates. MP report that the sellers
operate in an average of 10 4-digit SIC industries and the buyers in an average of 8 such
industries.
MP test whether diversified firm’s decision to sell a manufacturing plant can be ex-
plained by their neoclassical model. They run a probit regression on a panel of plants
1979–1992 from the LRD, where the dependent variable, PLANT SALE, takes on the
value of 1 if the plant is sold and the value of 0 if the plant is not sold in a given year.
PLANT SALE=α+β(Industry shock)+γ(Segment TFP)
+δ(Industry shock)×(Segment TFP)
+φ(Other segments’ TFP)
+θ(Relative demand)×(Other segments’ TFP)+controls.
Consistent with the profit maximizing model, MP find that plants in productive seg-
ments are less likely to be sold(γ < 0 ), especially in industries which have experienced
(^39) For a contrary view, seeAggarwal and Samwick (2003).
(^40) SeeMaksimovic and Phillips (2001)for a detailed description of the sample. See alsoSchlingemann, Stulz
and Walking (2002)for a discussion of liquidity in the market for assets on the rate of sales.