Horizontal mergers and divestment dynamics in a sunset industry

• Published date: 01 November 2016
• Author: Masato Nishiwaki
• Date: 01 November 2016
• DOI: http://doi.org/10.1111/1756-2171.12161

RAND Journal of Economics

Vol.47, No. 4, Winter 2016

pp. 961–997

Horizontal mergers and divestment

dynamics in a sunset industry

Masato Nishiwaki∗

Industries with declining demand tend to be riddled with chronicexcess capital due to the presence

of a business-stealing effect and fixed costs. This article highlights the potential of mergers to

internalize this business-stealing effect and therebypromote divestment. Using the case of mergers

in the Japanese cement industry, it examines whether such merger-induced divestment improves

total welfare based on a dynamic model of divestment. The findings suggest that merged firms

indeed tended to reduce capital more actively and that, as a resultof these mergers, total welfare

improved despite a reduction in the consumer surplus.

1. Introduction

In so-called sunset industries that face declining demand, an important concern is how

firms can reduce their capital stock to remain profitable. From an industry viewpoint, eliminating

production or distribution facilities would be beneficial, because it would remove excess capital

stock and thereby result in savings of the fixed costs associated with running these facilities, such

as labor- and land-related costs.

Yet in oligopolistic industries such divestment may not take place voluntarily. The reason is

similar to the Excess Entry Theorem discussed by Mankiw and Whinston (1986) and Suzumura

and Kiyono (1987). The theorem suggests that in oligopolies, the business-stealing effect and

fixed costs result in an excessive number of entrants: entrants gain sufficient demand partly by

stealing business from incumbent firms. Although this is a gain to the entrants, it is not a gain to

the industry and, in consequence, such entry is (socially) excessive.

This implies that in oligopolies, there is a tendency for more investmentto take place than the

industry as a whole would want. Firms want to invest in capital up to the level where the marginal

revenue from the next unit of capital just equals fixed costs. However, in oligopolies, part of that

∗Waseda University; mstnishi@gmail.com.

This is a revised version of Chapter 2 of my PhD thesis. I am grateful to Hiroyuki Odagiri for his constant support and

encouragement. Discussions with Hidehiko Ichimura and TetsushiMurao have been invaluable.I have also benefited from

the comments of many individuals in the course of this study, among them Reiko Aoki, RiekoIshii, Ryo Kambayashi,

DaijiKawaguchi, Toshihiro Matsumura, Noriyuki Matsushima, KeizoMizuno, Hiroshi Ohashi, Ralph Papryzycki, Tadashi

Sekiguchi, Ayako Suzuki, and Noriyuki Yanagawa. The research has benefited from the financial support of a grant-in-

aid (the 21st Century Center of Excellence Project on the Normative Evaluation and Social Choice of Contemporary

Economic Systems) from the Ministry of Education and Science, Japan, and the JSPS Grant-in-Aid for YoungScientists

(B) no. 24730202. All errors remain, of course, my own responsibility.

C2016, The RAND Corporation. 961

962 / THE RAND JOURNAL OF ECONOMICS

marginal revenue comes from the profits of competitors. This marginal revenue represents a gain

only to the firm, not to other firms in the industry, and the total amount of capital stock as a result

will be excessive, at least from the viewpoint of producers.1

In sunset industries, scrapping of capital stock may not take place for exactly the opposite

reason: each firm is unwilling to divest because part of the business it abandons by scrapping

is captured by its competitors. In other words, every firm intends to free-ride on the reduction

of industry supply expected from someone else’s divestment. The end result is that no firm will

divest even though this would reduce fixed costs, thus prolonging the situation where there is

excess capital stock.

The picture changes if two firms merge. A merger resolves the “deadlock” partly byinter nal-

izing the business-stealing effect. In a horizontal merger between firms A and B, postmerger, A

should have less incentive to maintain the same level of capital as before the merger, because the

stealing of business from B is now internalized and brings no gain to the merged firm. Therefore,

mergers can promote divestment, providing fixed cost savings, and, as a result of that, enhance

the profitability of merged firms even if it produces less than the two premerger firms together.2

In fact, it is often observed that in industries that experience a negative demand shock, firms rush

into mergers and subsequently rationalize their capital stock.

However, a merger eliminates capital stock and (at least) one competing unit. The reduction

in the number of firms within the same industry is highly likely to be harmful to consumers.

This anticompetitive effect may offset the efficiency improvement a merger can bring. This

is a fundamental trade-off a merger causes (Williamson, 1968). Thus, the change in industry

structure and the adjustment of capital following a merger do not necessarily give rise to a higher

level of total welfare. That is, from the viewpoint of total welfare, the problem is whether the

efficiency gains from mergers, which include cost savings as a result of capital divestments as

well as traditional synergy effects like a downward shift in the marginal cost curve, are greater

than the negative effect on the consumer surplus. This implies that whether mergers with capital

rationalization improve total welfare is an empirical question that depends on the magnitudes of

the various effects. The purpose of this study is to conduct just such an empirical investigation,

using the cement industry in Japan as a case study.

The Japanese cement industry provides a good case study for examining the welfare effect

of mergers in a period of industry decline. Japan’s cement industry can be regarded as a sunset

industry in the sense that it has faced a prolonged downward trend in demand. Following the

bursting of the bubble economy Japan experienced in the 1980s, public and private investment

in construction, which is a good indicator of cement demand, decreased substantially during the

1990s and in recent years, has settled at about the same level as that seen 30 years ago. As demand

shrank, the industry was forced to contract in size and to become more efficient to survive in

such severe circumstances. Around the midpoint of this phase of decline in the mid-1990s, four

mergers and one acquisition took place. Following the mergers, physical capital in the industry,

for example, production plants and distribution centers, contracted along with the decline in

demand. Whether this consolidation-induced contraction enhanced efficiency in the industry and

improved welfare is the main point of interest in this study.

To evaluate the welfare effect of horizontal mergers, a theoretical model to capture the

industry dynamics—namely,the downward trend in demand and divestmentof cement distribution

centers—is constructed, building on the Markov-Perfect Equilibrium framework of Ericson and

Pakes (1995). The underlying parameters of the model governing divestment dynamics are

estimated using the recently developed two-step estimator of Bajari, Benkard, and Levin (2007;

1Okuno-Fujiwara and Suzumura (1993) show that in oligopolies of symmetric firms, cost-reducing investment

such as Reseach and Development (R&D) investment becomes excessivenot only in terms of the producer sur plus but

also in terms of total welfare.

2Salant, Switzer, and Reynolds (1983) point out that even if the merged firm loses markets share, it can still save

costs by shutting down a plant. Therefore, in the case of high fixed costs, the mergercan be profitable.

C

The RAND Corporation 2016.

NISHIWAKI / 963

hereafter BBL). With the parameter estimates thus obtained, a counterfactual experiment is then

conducted to evaluate the welfare effect of the mergers.

The experimental exercise shows that the mergers improved total welfare, with most part

of the welfare improvement coming from the gains in the producer surplus. The marginal cost

functions of the merged firms shifted downward and the synergy effect stemming from this

downward shift exceeded the loss in the consumer surplus. In addition to this efficiency gain in

the traditional sense, merged firms tended to be more active in scrapping their distribution centers,

and these divestments as a result of the mergers led to additional increases in cement firms’ profits.

Specifically, fixed costs savings and the realization of the sell-off values of distribution centers

accounted for a nonnegligible part of the total welfare improvement.

The novelty of this study is that it examines the effect of mergers on the forward-looking

behavior of firms in the context of a declining industry and evaluates the welfare impact of such

mergers in a dynamic environment.This study thus contributes to both the literature on the analysis

of declining industries and that of mergers. Regarding the first of these two strands of literature,

there have in fact been relative fewstudies on declining industries since the early theoretical work

by Ghemawat and Nalebuff (1985, 1990), Fudenberg and Tirole (1986), and Whinston (1988),

despite the fact that almost all developed nations have declining sectors and how to promote

capacity reduction in such sectors is a pressing policy issue. Against this background, the present

analysis is one of a handful studies on declining industries and will providenew empirical findings

on industry contraction, with a particular emphasis on the role of mergers.

The second strand of literature the present study relates to is that on mergers. The importance

of modelling how a merger affects firms’ incentives for investment, divestment, entry, and exit

was first highlighted by Stigler (1968). Since then, a number of studies, including Berry and

Pakes (1993), Gowrisankaran (1999), and Pesendorfer (2005), have proposed theoretical models

addressing the effect of mergers. However, despite the blossoming of theoretical and computa-

tional work, only very few empirical studies on mergers from a dynamic perspective have been

conducted. To the best of my knowledge, the present study is one of only a very few attempts to

examine empirically the welfare implications of mergers by employing a fully dynamic model.

The remainder of the study is organized as follows. Section 2 reviews related empirical

studies on declining industries and mergers. Section 3 provides a brief overview of the Japanese

cement industry, and Section 4 explains the data used in this study. Next, Section 5 presents a

theoretical model describing competition in the cement industry. It allows for capital divestment

as well as traditional quantity-setting competition, building on the Markov-Perfect Equilibrium

framework of Ericson and Pakes (1995). Section 6 then presents the empirical procedure. The

structural parameters of the model are estimated using the econometric method recently developed

by BBL (2007). Section 7 provides the estimation results. This is followed, in Section 8, by a

simulation experiment to evaluate the effect of the mergers in the Japanese cement industry on

total welfare. Finally, Section 9 concludes.

2. Related studies

The current study is closely related to various previous empirical studies on declining indus-

tries and mergers. There are only a few empirical studies to date that have focused on declining

industries. These can be classified into two categories according to the econometric methodology

they employ: studies using structural econometric models, and studies using descriptive econo-

metric models. An example of studies using structural econometric models is Takahashi (2015).

He generalizes Fudenberg and Tirole’s (1986) model to analyze the exit decision in oligopolies

and develops a novel approach for estimating the exit game. Focusing on the US movie theater

industry, the main interest in his study is in the timing of exit and, in particular, how strate-

gic interactions among movie theaters delayed the exit date. He finds that strategic interactions

substantially delayed exit dates compared to the dates that would maximize the industry’s total

profit.

C

The RAND Corporation 2016.