St. Petersburg State University
Graduate School of Management
Master in Corporate Finance Program
IMPACT OF CORPORATE DIVERSIFICATION
ON COMPANY PERFORMANCE AND RISK: EVIDENCE FROM RUSSIA
Master’s Thesis by the 2nd year student
Concentration - Corporate Finance
Iskander Shafigullin
Research advisor:
Anna Loukianova, Associate Professor
St. Petersburg
2016
ЗАЯВЛЕНИЕ О САМОСТОЯТЕЛЬНОМ ХАРАКТЕРЕ ВЫПОЛНЕНИЯ
ВЫПУСКНОЙ КВАЛИФИКАЦИОННОЙ РАБОТЫ
Я, Шафигуллин Искандер Ильдарович, студент второго курса магистратуры
направления 080200 - «Менеджмент», заявляю, что в моей магистерской диссертации на тему
«Влияние корпоративной диверсификации на показатели и риски компании: на примере
компаний из России», представленной в службу обеспечения программ магистратуры для
последующей передачи в государственную аттестационную комиссию для публичной защиты,
не содержится элементов плагиата.
Все прямые заимствования из печатных и электронных источников, а также из
защищенных ранее выпускных квалификационных работ, кандидатских и докторских
диссертаций имеют соответствующие ссылки.
Мне известно содержание п. 9.7.1 Правил обучения по основным образовательным
программам высшего и среднего профессионального образования в СПбГУ о том, что «ВКР
выполняется индивидуально каждым студентом под руководством назначенного ему научного
руководителя», и п. 51 Устава федерального государственного бюджетного образовательного
учреждения высшего образования «Санкт-Петербургский государственный университет» о
том, что «студент подлежит отчислению из Санкт-Петербургского университета за
представление курсовой или выпускной квалификационной работы, выполненной другим лицом
(лицами)».
_______________________
26 мая 2016
2
STATEMENT ABOUT THE INDEPENDENT CHARACTER
OF THE MASTER THESIS
I, Iskander Shafigullin, second year master student, program 080200 - «Management», state
that my master thesis on the topic «Impact of corporate diversification on company performance and
risk: evidence from Russia», which is presented to the Master Office to be submitted to the Official
Defense Committee for the public defense, does not contain any elements of plagiarism.
All direct borrowings from printed and electronic sources, as well as from master theses, PhD
and doctorate theses which were defended earlier, have appropriate references.
I am aware that according to paragraph 9.7.1. of Guidelines for instruction in major
curriculum programs of higher and secondary professional education at St. Petersburg University «A
master thesis must be completed by each of the degree candidates individually under the supervision
of his or her advisor», and according to paragraph 51 of Charter of the Federal State Institution of
Higher Education Saint-Petersburg State University «a student can be expelled from St.Petersburg
University for submitting of the course or graduation qualification work developed by other person
(persons)».
_______________________
May 26, 2016
3
АННОТАЦИЯ
Автор
Шафигуллин Искандер Ильдарович
Название магистерской
Влияние корпоративной диверсификации на показатели и риски
диссертации
компании: на примере компаний из России
Факультет
Высшая Школа Менеджмента
Направление
080200 “Менеджмент” (Профиль: Корпоративные финансы)
подготовки
Год
2016
Научный руководитель
Лукьянова Анна Евгеньевна, Доцент
Описание цели, задач и
Цель
основных результатов
диверсификации на показатели и риски компании, используя
работы:
исследовать
влияние
корпоративной
пример компаний из России.
Задачи: изучить теорию по корпоративной диверсификации,
изучить существующие исследования по влиянию корпоративной
диверсификации на показатели и риски компаний, сформировать
эмпирическую модель исследования, собрать выборку для
эмпирического анализа, провести эмпирический анализ по
собранной выборке, интерпретировать результаты и предложить
рекомендации.
Результаты: изучена теория по корпоративной диверсификации,
проанализированы существующие исследования по влиянию
корпоративной диверсификации на показатели и риски компаний,
сформирована эмпирическая модель исследования, собраны
данные по 64 Российским компаниям за 2000-2015 годы формируя
выборку из 465 наблюдений, проведен регрессионный анализ на
основе
собранной
выборки,
результаты
анализа
были
интерпретированы и обсуждены, и на основе результатов были
предложены рекомендации.
Ключевые слова
Диверсификация, корпоративные результаты, корпоративный
риск, интернационализация, диверсификация по продукту,
диверсификация по географии
4
ABSTRACT
Master Student's Name
Iskander Shafigullin
Master Thesis Title
Impact of corporate diversification on company performance and risk:
evidence from Russia
Faculty
Graduate School of Management
Main field of study
080200 “Management” (specialization: Master of Corporate Finance)
Year
2016
Academic Advisor's
Anna Loukianova, Associate Professor
Name
Description of the goal,
The goal of the research is to determine the relationship between
tasks and main results
corporate diversification and company performance and risk, using
evidence from Russian companies.
In order to achieve the goal, we complete the following objectives: to
study the theoretical background on corporate diversification; to study
existing literature on corporate diversification and performance
relationship, and on corporate diversification and risk relationship; to
propose an empirical methodology of the analysis on the impact of
corporate diversification on performance and risk; to build and describe
a sample for the analysis; to conduct an empirical study on the built
sample; to interpret results and provide managerial implications based
on the findings.
Main results: theoretical background on diversification was studied;
existing literature on corporate diversification and performance
relationship, and on corporate diversification and risk relationship was
studied; empirical model for the research was created; data on 64
Russian companies for 2000-2015 time period was collected, forming
a sample of 465 observations; the sample was analyzed with regression
analysis; results of regression analysis were interpreted and discussed
managerial implications were provided.
Keywords
Diversification, corporate performance, corporate risk,
internationalization, product diversification, geographical
diversification
5
TABLE OF CONTENTS
INTRODUCTION ................................................................................................................................ 7
CHAPTER 1. LITERATURE REVIEW .............................................................................................. 9
1.1.
Theoretical background on diversification ........................................................................... 9
1.2.
Corporate diversification and performance ........................................................................ 15
1.3.
Corporate diversification and risk ....................................................................................... 21
1.4.
Summary of Chapter 1 ........................................................................................................ 25
CHAPTER 2. RESEARCH DESIGN ................................................................................................. 27
2.1.
Diversification definition .................................................................................................... 27
2.2.
Sample description .............................................................................................................. 28
2.3.
Methodology ....................................................................................................................... 30
2.4.
Variables ............................................................................................................................. 33
2.5.
Summary of Chapter 2 ........................................................................................................ 36
CHAPTER 3. RESEARCH FINDINGS ............................................................................................. 38
3.1.
Descriptive statistics ........................................................................................................... 38
3.2.
Model findings .................................................................................................................... 40
3.3.
Results ................................................................................................................................. 45
3.4.
Discussion ........................................................................................................................... 51
3.5.
Summary of Chapter 3 ........................................................................................................ 58
CONCLUSION ................................................................................................................................... 61
REFERENCES ................................................................................................................................... 63
6
INTRODUCTION
In the modern globalized world with increasing international trade and booming information
volumes one of the ways available for companies to stay ahead of the competition is diversification,
both product and geographical. Theoretically, there is a lot of benefits that are commonly attributed
to diversification (Williamson, 1979), however in practice, it is often the case that diversification
becomes a value destroying activity (Porter, 2008). While more and more companies make attempts
to diversify, it is important to understand what impact such activities can make on two of the most
important matters of any company: performance and risk.
It is usually perceived that corporate diversification opens up new opportunities in terms of
performance and therefore it is beneficial for companies. However, since 1970s academicians tried to
understand the relationship between diversification strategy and firm performance, and in turns out
that there is no one single answer; the evidence is mixed, and there are different views on the
relationship (Dey & Banerjee, 2011).
Same goes for diversification and risk: there is a common perception that diversification
reduces corporate risk, however the empirical results are mixed as well (Anderson et al., 2011). The
motivation of reduction of risk by diversification, highlighted in the literature, contributes largely to
explain the choice to integrate the notion of risk in this study.
In general, most papers study the impact of diversification on performance and risk separately,
however there is lack of empirical studies on these two issues simultaneously. Moreover, as per our
knowledge, very few studies were devoted to conducting such analysis with regards to the emerging
markets, let alone Russia. With this paper we aim to fill this gap. Also, most studies are concentrated
on cross-section analyses. Differently, this research uses a longitudinal data from a sample composed
of large Russian firms, in order to analyze the firm activity perimeter evolution effect on its level of
risk and performance in a dynamic prospect.
The research goal of the paper is to determine the relationship between corporate
diversification and company performance and risk, using evidence from Russian companies.
In order to achieve the outlined research goal, we define the following objectives:
•
To identify the theoretical background on corporate diversification;
•
To study existing literature on corporate diversification and performance relationship, and
on corporate diversification and risk relationship;
7
•
To propose an empirical methodology of the analysis on the impact of corporate
diversification on performance and risk;
•
To build and describe a sample for the analysis;
•
To conduct an empirical study on the built sample;
•
To interpret results and provide managerial implications based on the findings.
This master thesis is an empirical research, in order to achieve the goal of the study we conduct
quantitative analysis using econometric tools built in the Stata software.
The main sources of information we use for the purposes of this research are academic articles
devoted to: theoretical studies of diversification, motivation to diversify, determinants of
diversification premium or discount, the effects of corporate diversification on company performance
and risk. In order to gather data for our empirical study we use Thomson Reuters Datastream database
and annual reports available on official websites of the companies selected for the study.
In order to achieve the defined goal of the research, we structure this thesis as follows: an
introduction, three chapters that cover all objectives of the research and a conclusion. The introduction
includes goals and objectives of the research, along with the motivation and background of the study.
The first chapter covers the first two objectives as is devoted to analysing the theoretical framework
of diversification, and the impact of diversification on performance as risk.
The second chapter corresponds to the third and fourth objectives, as there we describe the
empirical research methodology, sample selection and variables calculation. In the third chapter we
cover the last two objectives, as we present the results of the econometric analysis and then discuss
these findings as well as develop managerial recommendations.
Finally, the conclusion summarizes the results of the research in accordance with the goals set.
Also, at the end of each chapter we provide a short summary in order to help a reader better catch the
main points discussed in the chapter.
8
CHAPTER 1. LITERATURE REVIEW
1.1.
Theoretical background on diversification
One of the approaches companies employ to remain competitive in business is diversification.
In the most general classification, there is product diversification, where a company gets revenue from
several industry sources; and geographical diversification, where a company gets its revenue from
several geographical locations.
Product diversification
According to Nayyar (1992), a company can be called a diversified one if it operates in more
than one business. In turn, the level of involvement in different businesses, is called diversification
strategy.
To give a definition of diversification strategy, it can be “expanding to new industries and
markets which differ from company’s main markets or product lines” (Johnson and Scholes, 2002).
According to them, there are various drivers and reason why companies choose to diversify, these are
all kinds of advantages connected to:
•
Higher market power,
•
Use of existing capacities and resources in other dimensions,
•
Better allotment of assets through internal capital markets,
•
Greater debt capacity,
•
Decreased performance variation by virtue of a portfolio of imperfectly correlated set of
business.
In other words, leveraging company resources in more than one industry creates beneficial
synergies between the sectors, and allows the firm to gain benefits connected to cost or differentiation.
Other advantage from diversification include tax and general financial benefits. Essentially, these are
the main reasons why companies diversify. It is important to note here, however, that diversification
advantages are highly dependent to institutional development: in the context of developed institutional
economies diversification strategy is more beneficial (Kock & Gulline, 2001).
As product diversification has its advantages, there are also numerous potential costs in place.
These include the information asymmetry costs that can emerge between headquarters and divisional
managers in diversified companies, where handling information becomes more expensive (Harris et
al., 1982). Jensen (1986) claims that decision makers of companies that have untapped borrowing
9
power and substantial cash at hand are more likely to chase projects and investments that may not be
highly valuable. Besides, according to Meyer et al. (1992), there is a common asymmetry between
growth and decline for divisions of a diversified company. To sum up, the potential drawbacks of
diversification strategy contain existence of abundant cash to make value-decreasing investments,
cross-subsidies that allow poor units to suck the capacities from higher-performing units, and unequal
division of incentives between HQ and unit managers.
Product diversification can be of two kinds: related and unrelated. If a firm operates in a market
that possesses a strategic fit with the core business, then it is related diversification. Otherwise, if there
is no direct strategic fit between the two, nor significant interrelationships in the value chain, it is
considered unrelated diversification (Allen & Gorgeon, 2002).
When a company decides to diversify in a related manner, the main purpose it has is to leverage
the synergies – advantages between the operations in the two related sectors. Synergies occur when
the combined outcome of the merged actions becomes bigger than the sum of the individual effects.
To put it in another way, a synergistic effect arises when one plus one add up to three. When it comes
to synergy, this is not some abstract, created connection, it is selected capabilities (financial, human,
technological) and opportunities (R&D, brand management, customer service), which can be
transferred between sectors.
Geographical diversification
Geographical diversification and its benefits and costs is quite a well-studied topic in the
academic literature. In general, academician divide internationalization by two types: multinational
companies and foreign direct investments. Theory of multinational companies involves arguments on
shared resources and experience exchange between divisions in different countries (Wernerfelt, 1984),
global arbitrage opportunities (Kogut, 1989) and synergies in business process and systems
optimization (Fayerweather, 1978). School of foreign direct investment puts that internationalizing
companies are able to reduce risk by diversifying their businesses in different locations (Lessard,
1976), and improve performance via economies of scale and resource sharing (Rieck et al., 2005).
Moreover, various authors have argued that international diversification enhances shareholder
value by exploiting firm-specific assets, by increasing operating flexibility and by satisfying investors’
preferences for holding globally diversified portfolios.
According to Morck et al. (1998), geographical diversification is especially beneficial for
companies which possess a significant information-based asset base (it can be connected with R&D
10
and advertising). This type of assets shows evidence of improving returns of scope and scale although
it can be at times challenging at times to sell. Internationalization solves this problem.
Another benefit of geographical diversification is that it enables the company to leverage
market conditions, thus adding value through operational flexibility. An international company can
choose if it wants to move manufacturing to a location with cheaper costs, or to move distribution to
a place with bigger buying potential; plus the difference is the tax system can be leveraged too.
Finally, another advantage of geographical diversification is based on invertor preferences.
Theoretically, investors favor companies which are diversified internationally and are more likely to
pay a premium for such firms, due to lower cost to have a diversified portfolio, rather than for
diversifying as a separate investor.
Similar to product diversification, while internationalization strategy has its advantages, there
are also numerous potential costs in place. One of the drawbacks is the same as with product
diversification: due to internationalization the divisions that are less profitable can potentially be
cross-subsidized inefficiently. Also, there is again information asymmetry in place with concerning
divisions and head office (Harris et al., 1982).
Diversification premium and discount
A big question in diversification research which researchers and practitioners try to answer, is
whether there is value created for a company when is diversifies across industries and locations. It is
said, that if diversification premium is there is indeed excess value creates; on the other hand,
diversification discount arises when value is destroyed.
These two concepts, of diversification premium and discount, are aligned with benefits and
costs of diversification, respectively; which are in turn explained differently by different researches.
For the costs of diversification, Jensen & Murphy (1990), for one, explain it by agency arguments;
Gomes & Livdan (2004) argue that one of the costs is company’s value maximization behavior; and
Choe & Yin (2009) explain it by inefficient investment because of rent-seeking activities. Maksimovic
& Phillips (2007) state that one of the reasons for diversification discount is self-selection of
companies that have varying financing opportunities to begin with, contradictory to the previous
explanations with inefficient internal capital markets. Their findings also are based on the more recent
empirical evidence in the research. Another factor they state that can be important for explaining the
discount is capital budgeting procedure in companies that target to solely maximize their returns,
rather than diversification itself. Moreover, according to a survey by Stein (2003), the discount arises
11
mostly due to agency issues and information asymmetry. It is worthwhile to notice, that he studied for
the most part the body of articles with the focus on internal capital markets and strong capital
allocation. Finally, according to the survey by Martin & Sayrak (2003), potential reason for the
discount may not be diversification itself, but some kinds of bias and problems with measure
management.
As for the advantages (benefits) of diversification, these are explained by economies of scope
(Teece, 1892), well-functioning internal capital market (Stein, 1997) and debt concurrence effects
(Shleifer & Vishny, 1992). Previous empirical research shows evidence of contradictory points of
view on advantages of diversification strategy (Lins & Servaes, 1999; Villalonga, 2004; Rajan et al.,
2000; and others), this topic is still in an infancy stagy and discussion is still ongoing.
Since there is open discussion, in order to proceed with discussing diversification, it will be
beneficial to understand the theoretical and methodological arguments on the topic.
According to Wan et. al (2011), one of the most important theoretical points of view on
diversification from the research is the importance of resources, mainly due to the fact that the way
and the outcome of company performing is heavily connected with its resources and its presence in
the firm. According to Farjoun (1994), when a company shares resources close to market failure,
diversified firms tend to do better that focused ones, this is the key idea behind the Resource-based
view.
Some studies also provide another view on diversification strategy, namely through Real
Options. For example, Bernardo & Chowdhry (2002) propose an argument that companies are able to
make the value of its growth options through diversifying and expanding to different businesses.
Based on this theory, a whole new set of arguments arises, where it is possible to explain
diversification with strategic flexibility and managerial decision as per real options. According to
Bernardo & Chowdhry (2002), analyzing past diversification actions as real options, companies are
able to get better information on their capabilities and probability of future success in diversification.
Researchers tend to agree for the most part, that related diversification is beneficial for
companies. The following advantages arise: economies of scope (Nayyar, 1992), sharing resources
(Ghemawat & Khanna, 1998), synergies in the values chain (Barney, 1997).
12
1.1.1. Measures of diversification
In order to assess the effects of corporate diversification, it is essential to be able to measure it
first. In the previous researches, authors describe and employ a number of various indications to
measure company diversification. In the broadest sense, these indicators fall under two main groups,
which measure the level of diversification and diversity in a company. Diversification indicators
measure the level of company’s involvement in various sectors of economy. As for the diversity
indicators, they showcase how company’s resources and revenue streams are dispelled across these
segments.
In this paper we are mostly concerned about measures of diversification, so we do not discuss
diversity indicators in detail.
For the diversification indicators, in a broad way they can be classified by being discrete and
continuous. There are two main discrete indicators, both rather simple:
•
Diversification dummy. It is a binary measure that will equal unity of the company
operates in more than one industry and will equal zero if the company is active only in
one sector (Lins & Servaes, 2002);
•
Number of industries. This one simply measures the number of industries of company
operations (Khanna & Palepu, 2000).
It should be noted, that out of the discreet measures, diversification dummy is by far the most
frequently used one; it is used by, among others, Lins & Servaes (2002), Anderson et. al (2000),
Fleming et. al (2003), Mansi & Reeb (2002). As per whether it provides more accurate results
compared to number of segments, it is rather obvious that it provides on the issue whether
diversification in the company exists at all, rather than the degree of the diversification. For the
empirical research of diversification effects, therefore, it should be noted that this measure is only
applicable when the purpose is to learn about the outcome of being diversified. Unlike the number of
industries indicator, diversification dummy is not very efficient when studying the effects of higher
or lower diversification (Villalonga, 2004).
As for the continuous indicators of diversification, the most widely used ones in the current
body of research are:
•
Herfindahl index. This indicator is calculated by summing the squared values of revenues
in each revenue segment, and dividing it by company’s total revenues. When a company
has its revenues only from one industry, i.e. it is not diversified, its Herfindahl index is
unity. This indicator was first used with regards to diversification by Berry (1971).
13
•
Entropy index. While the Herfindahl index constitutes for the weighted average of sector’s
shares in total company revenue, these shares are taken as it is. With the entropy index,
the main idea is the same, however the shares are weighted by taking natural logarithms
of the its inverses. This indicator was first proposed to be used with regards to
diversification by Jacquemin and Berry (1979).
Both of these measures can be employed for company’s revenues or assets.
As in the previous section we discussed that diversification can be classified to related and
unrelated, it is important to note that the Herfindahl index has one major drawback when comparing
it to the entropy index: it can not account for these sub-types of total diversification. The entropy index
can be broken down to two different indices: the entropy index of related diversification and the
entropy index of unrelated diversification. These indices are perceptive to the level of how the sectors
and segments are identified. The industries can be identified broadly, at the two-digit SIC or NAICS
levels (internationally accepted standards for industry classification); or in a narrower way, at the fourdigit level of SIC or NAICS. The difference here is that in the first case (broader definition) the
diversification indicator will be a lot smaller that in the second case. The majority of authors in the
current literature is the broad, two-digit SIC or NAICS way to analyze unrelated diversification, and
the narrower, four-digit SIC or NAICS level for the related diversification (Palepu, 1985; Khanna &
Palelu, 2000; Villalonga, 2004).
Opposing to the more flexible entropy index nature, the Herfindahl index does not possess an
ability to differentiate between broad and narrow level, or related and unrelated diversification
respectively. As the entropy index can be decomposed in a way that total entropy index equals related
entropy index multiplied by unrelated entropy index, the Herfindahl index can not identify the level
of relatedness between company’s diversified revenue segments (Palepu, 1985).
While both of these indices have some limitations in place, another difference between them
is that they have different receptiveness concerning changes in diversification levels. The Herfindahl
index not only does not identify the relatedness, it also catches changes in number of industries and
its weights more slowly than the entropy index. Also, the entropy index is less receptive to changes
in the biggest sector weights, but more sensitive to changes in smaller division weights, rather than
the Herfindahl index (Gorecki, 1974).
The survey of diversification measure usage in the current literature in presented in Table 1
below. We should note that for the continuous measures, in recent years the entropy index becomes
more and more widely used.
14
Table 1. Survey of diversification measure usage in existing literature.
Diversification measure
Diversification dummy
Discrete
Number of segments
Herfindahl index
Continuous
Entropy index
1.2.
Used by (among others)
Berger and Ofek (1995)
Servaes (1996)
Anderson et al. (2000)
Fleming et al. (2000)
Lamont & Polk (2001)
Villalonga (2004)
Schmid & Indo (2009)
Lang & Stulz (1994)
Berger & Ofek (1995)
Berger & Ofek (1999)
Lamont & Polk (2001)
Schoar (2002)
Villalonga (2004)
Schmid & Indo (2009)
Lang & Stulz (1994)
Berger & Ofek (1999)
Khanna & Palepu (2000)
Schoar (2002)
Villalonga (2004)
Schmid & Walter (2009)
Elsas et al. (2010)
Khanna & Palepu (2000)
Villalonga (2004)
Lee et al. (2008)
He (2009)
Corporate diversification and performance
In this section we first review research studies on diversification and performance relationship,
and then discuss measures of performance that are frequently used in the diversification context.
1.2.1. Relationship overview in the existing literature
The relationship between corporate diversification and performance is a topic that
academicians have been conducting research on since 1970s. The impact of the former on the latter is
not a mature topic and is an ingoing subject in the world of economics, finance and corporate strategy;
it is far from being exhausted (Palich et al., 2000).
In this sub-section we conduct a survey and analyze what the findings in terms of corporate
diversification and performance relationship in the existing literature are. In general, a broad body of
papers and articles analyzed this topic in the past, however there does not seem to be one unambiguous
15
finding and conclusion to fit them all. From the logic and theories described in the previous section
we can determine that diversification has a lot of benefits for a company, things such as synergies,
economies of scope and scale, decreased risk, steeper learning curve, market power and others.
However, as we will see below in this sub-section, there is a lot of evidence on negative relationship
as well.
In general, the studies on the topic in the current literature can be categorized to the following
types:
•
“Plain” linear relationship outcome;
•
U-shaped relationship;
•
Breakdown to related and unrelated diversification;
•
Comparison of relationship between different countries (Boz et al., 2013).
The first group of studies shows a “plain” linear relationship between corporate diversification
and performance. In general, studies clearly show mixed evidence. Singh et al. (2001), Piscitello
(2004), Khanna & Palepu (2000), Schoar (2002) and others find that corporate diversification impacts
the firm’s performance positively, be it revenues, profits, or company value. For one, Anil & Narendar
(1998) find relatively strong evidence in favor that diversified companies do considerably better than
focused ones.
As for the linear negative relationship, it was found existing by Markides & Williamson
(1994), Berger & Ofek (1995), Bernardo et al. (2000), Anderson et al. (2000) and others; and we have
to notice that this body of research makes up to the majority – most of the studies report a negative
linear relationship. In other words, most of the evidence shows that with corporate diversification
comes a slowdown of downturn in company’s performance (Villalonga, 2001). That said, some
authors state that the discount occurs not due to diversification itself, but rather to some side issues.
But generally, many researches conclude that diversification brings more drawbacks rather than
benefits for companies (Lins & Servaes, 2002). As most of the existing researches report a negative
relationship between corporate diversification and performance (Villalonga, 2001), the first
hypothesis we use in our empirical study is the following:
!" : Product diversification is negatively related to firm’s performance.
Finally, studies from Montgomery (1985), He (2009) and others report lack of significant
relationship between corporate diversification and company performance. Table 2 presents a summary
of linear diversification-performance relationship in existing literature.
16
Table 2. Survey of linear diversification-performance relationship in existing literature.
Finding
Positive relationship
Negative relationship
Lack of relationship
Found by (among others)
Khanna & Palepu, 2000
Singh et al., 2001
Schoar, 2002
Piscitello, 2004
Markides & Williamson, 1994
Berger & Ofek, 1995
Bernardo et al., 2000
Anderson et al., 2000
Lins & Servaes, 2002
Gary, 2005
Montgomery, 1985
He, 2009
Moving on to the second group in the body of research, we see the inverted U-shaped
relationship. This kind of nonlinear evidence is found by Kakani (2000), Palich et al. (2000) and
others. This evidence suggests that as the level of diversification in the firm rises to some specific
level, the organizational results grow as well; but when the diversification grows further on after this
peak point, the results will deteriorate. This is explained by that there are numerous inefficiencies and
extra costs in terms of organizational management, where the size of the company is out of the
maximum level control (Grant et al., 1988), therefore the more a firm diversified after some average
point, the less can benefits of diversification play a role, and therefore the discount arises.
As for the third body of research, some of the authors went further and decided to break the
total diversification down to related and unrelated diversification and see its impact on performance.
The majority of studies in this group find that the companies that diversify in the related sectors
perform better than the ones that choose to diversify unrelatedly, as shown by authors such as
Markides & Williamson (1994) and Lubatkin & Chatterjee (1994), among others. These studies
suggest that resource-based theory and synergy views on diversification hold (Desmond, 2007). Other
studies, for example Grant et al. (1988) find no advantage for related or unrelated diversification,
stating that both of these methods enhance firm performance and are advantageous for companies to
focused companies.
However, the main base point is the superiority of related diversification for the company
performance. Palich et al. (2000) conducted an extensive survey of 55 studies with evidence on this
topic and what they found is that indeed related diversification is more advantageous comparing to
unrelated types. Doukas (2003) provided evidence that firms with related diversification experience a
17
diversification premium, due to significant positive abnormal returns when operating in related
activities.
As discussed in previous section, there is a number of advantages of related diversification for
the companies (such as economies of scope, sharing resources, value chain synergies and others); and
the empirical evidence suggests that these benefits outweigh the costs, thus we propose the following
hypothesis for our study:
!# : Related product diversification is positively related to firm’s performance.
As for the unrelated diversification, for the most part researchers tend to agree that it is
negatively related to performance. Some of the reasons are the following: sharing resources and
capabilities becomes rather challenging and more expensive, higher costs due to introduced operating
inefficiencies of unrelated business operation and lack of synergies regarding existing divisions
(Palich et al., 2000). Also, this type of diversification often makes a negative impact on company’s
operating efficiencies, because unrelated diversifiers can not leverage the benefits of economies of
scope and have inefficient internal markets (Doukas, 2003).
To sum up, unrelated diversification often leaves a negative impact on corporate performance
due to lack of synergies and focus, and decreased coordination of cohesion. Thus, we propose the
following hypothesis:
!$ : Unrelated product diversification is negatively related to firm’s performance.
Finally, a lot of studies concentrate on comparing the diversification and performance
relationship between different countries. Gullien (2000) found that diversification is more beneficial
for companies from emerging countries rather than developed one; Khanna & Palepu (1997) find
similar evidence. Speaking of reasons for that, Stijn et al. (2002) argue that the capital markets tend
to be less developed, as well as legal systems, and in general there is less information available in the
emerging markets, and these factors make diversification more beneficial in these markets.
As for the evidence on the country level, Zhang et al. (2002) studied performance of 72
Chinese firms and found that diversification impacts performance negatively. Li (2004), on the other
hand, utilizes a larger sample size of Chinese companies did not find any significant relationship.
Fleming et al. (2003) found negative relationship for Australian companies. Lins & Servaes (2002) in
their studies conducted an extensive research of a number of different countries and found that the
relationship is negative in Hong Kong, UK, Singapore, South Korea and Japan; and in such emerging
markets as Indonesia, Malaysia, Thailand and India. Speaking of India, Khanna & Palepu (2000), on
the other hand, found a positive relationship for the Indian companies in terms of diversification and
18
performance. Overall, these at times contradictory results between countries can arise due to
institutional variability, different data sources, different research methods and sample selection bias.
In general, we have to note that evidence form developing countries on this topic is still in an
infancy stage and very limited. These counties are becoming more and more significant players in the
world economy, so this is definitely an open topic interesting to study.
Overall, we see that many researchers tried to study corporate diversification and performance,
and their empirical evidence is mixed; there is no consensus on the diversification effects in the current
literature.
Moving on to geographical diversification and performance, we see that existing research
papers suggest mixed views on the impact (Hitt et al., 1997). Authors such as Tallman & Li (1996)
and others provide evidence that internationalization impacts corporate performance positively;
others, such as and Lu & Beamish (2004), find negative relationship. However, according to Rieck et
al. (2004), the majority of authors find the relationship between internationalization and performance
to be positive, thus we define our next hypothesis as the following:
!% : The relationship between geographical diversification and firm’s performance is positive.
1.2.2. Measures of corporate performance
Having reviewed papers on diversification and performance relationship in the previous subsection and having seen that they yield varying results, it is important to note that these papers use
varying diversification and performance measures. As we have reviewed diversification indicators in
the first section, this sub-section aims to go through the main performance measures existing in the
literature.
In the broadest way, performance measures can be classified in two categories, both of which
are extensively used in the literature on the diversification and performance relationship:
•
Accounting-based measures;
•
Market-based measures (Palich et al., 2000).
Accounting-based measures are generally considered to be the main indicator of firm
performance, as they are able to show the result of company operations as per account books. The
main feature of these indicators is that they measure company profitability across various dimensions,
such as assets, equity, shares, and others. Some of the accounting-based measures used in the past are:
Return on assets, Return on equity, Return on sales, Return on Investment, Operating profit, Earnings
per share, Return on capital employed and others (Al-Matari et al., 2014).
19
One of the limitations of accounting-based measures is that they incorporate only the past
performance, failing to foresee or in any way account for future prospective. Also, they recognize
future events in terms of depreciation and amortization in a limited way. Besides, these measure are
subject to be affected by differing accounting methods and practices, and the accountant’s personal
skills, as he is the one calculating the measures. This especially impacts the calculation of intangible
assets (Kapopoulos & Lazaretou, 2007).
Market-based measures, on the other hand, incorporate the forward-looking features in that the
investor expectations are included in the dimension too. The investors (the market) have certain
expectations regarding the company’s potential performance in future, which are reflected in the
market-based measures. These measures include, among others, Tobin’s Q, Price-to-earnings ratio,
Market-to-book value (Wahla et al., 2012).
Return on assets (ROA) is the most widely used accounting-based measure in the
diversification-performance literature, due to the reason that this measure is an effective indicator of
both operating and financial sides of company performance. It indicates how well the assets are used,
which is an important message for shareholders (Klapper & Love, 2002).
As for the market-based measures, Tobin’s Q is the most widely used one in the
diversification-performance literature, it became so after it was first used by Lang & Stulz in their
seminal study from 1994. It is calculated by dividing the total market value of a firm by the book
value of its assets, and it has become a common indicator to see the expected long-run performance
of a company. The numerator of the formula employs the investor-based opinion on future growth
prospective, and the denominator in the asset value depends from past organizational decisions. So
the higher the indicator is, the more evident it is that the firm has taken advantage of the investments
and proceeded to develop the company efficiently (Bozec et al., 2010).
Khanna & Palepu (2000) in their study employ both of the two most popular measures in
Return of assets and Tobin’s Q in order to track company performance. Berger & Ofek (1995) only
use the Return of assets indicator. Campa & Kedia (2002) in turn, use the assets and sales measures,
is in Return on sales. Fleming et al. (2003) and Lee et al. (2008) both use Earnings before taxes. In
other words, different researches employ different measures to track the impact of diversification on
performance, and it is highly dependent on author’s goals and preferences.
20
1.3.
Corporate diversification and risk
In this section we first review research studies on diversification and risk relationship, and then
discuss measures of risk that are frequently used in the diversification context.
1.3.1. Relationship overview in the existing literature
The general view in the finance and strategy literature and theories that corporate
diversification is associated a lower corporate risk profile. Although it is a usual perception, there is
lack of relevant evidence in the existing body of research that shows the clear relationship between
corporate diversification and corporate risk (Anderson et al., 2011). In this sub-section we review the
polarizing views on the relationship and possible reasons on why diversification might reduce and
increase corporate risk.
The majority of papers follow the popular opinion that there is risk reduction associated with
higher diversification. One of the arguments for why that happens is related to the real options view
of diversification discussed above: when diversifying, companies execute their growth options. In
other words, by diversifying, firms transform these growth opportunities into assets and by that reduce
the risk, as the options consist of future economics conditions, which is risky (Carlson et al., 2006).
Another argument, as proposed by Amihud & Lev (1981), consists of the assumption that by
diversification activities decision makers in companies follow the incentives to reduce risk, so they
choose to pursue investment projects that help to lower the variations in company’s revenues. Also,
there is another argument why diversification reduces the risk, and it is based on the similarity between
corporate diversification and portfolio diversification. According to Anderson et al. (2011), it is
possible to conclude that corporate diversification reduces company’s risk since corporate and
portfolio diversification are analogous since both hinge on investments in different segments or
sectors. The dominant focus in portfolio diversification is on risk management, and it is possible to
related the concept to corporate diversification as well, according to the author.
On the other hand, although it is the most popular, the opinion that corporate diversification is
associated strictly with reduction of company risk is not the only view. There are several factors
concerning why corporate diversification impacts corporate risk positively.
Consistent with the first argument of negative diversification and risk relationship above, the
first argument here also concerns real options. According to Zhang et al. (2002), it is possible that
growth options actually possess smaller amount of risk comparing to actual assets. The reason being,
that companies that have more growth options tend to have lower adjustment expenses and therefore
21
tend to keep their investments; which therefore leads to the conclusion that the counter-cyclical price
for these expensive and risky assets cause reversibility in place, more difficult to reduce than the
growth option is therefore a high risk, especially during the economic downturns with high price of
risk. The bottom line is that diversification increases risk of the company when assets have a higher
risk profile than these growth options.
The next argument concerns decision maker preference that tends to determine the risk profile.
As Hermalin & Katz (2003) state, when making decisions on whether to diversify or not, it is usually
shareholders rather than manger who decide. Since shareholders tend to be less risk-averse than
managers, as managers prefer to to secure their salaries and bonuses with safer business decisions,
shareholders are more likely to pursue riskier solutions. Therefore, when decisions on diversification
are done by shareholders, it is more likely for the risk profile of the company to increase. It is important
to note that this argument is based on the opinion that managers tend to be more risk averse than
shareholders. However, it is not a unanimous case. According to Agrawal and Mandelker (1987),
when managers have shares or options of the company, they tend to be more willing to pursue risky
decisions. Generally, there are two main outcomes of risky diversification decisions that affect the
responsible people: value of the company grows, or human capital of decision makers deteriorates. If
decision makers in managers have shares in the company, the first outcome might win and be more
important – therefore risk-increasing choices are more likely.
The final argument is based on the similarity between corporate and portfolio diversification.
It is important to note that risk reduction is the main reason for portfolio diversification, but is not
necessarily the main reason and factor for corporate diversification. There are other reasons for
corporate diversification that do not reduce risk and can even lead to risk growth, such as wellfunctioning internal capital markets, economies of scale and managerial benefits (Denis et al., 1997).
Although the issue of corporate diversification and risk relationship is not as frequently studied
in literature as the issue of corporate diversification and performance relationship, it is still an issue
worth of research and it is being studied in the current literature. For example, Anderson et al. (2011)
provide evidence that diversified firms do not tend to have lower risk, although some of them indeed
have lower risk profiles. They studied the effects through the analysis of diversifying acquisitions.
Comment & Jarrell (1995) find that there is indeed negative relationship between corporate
diversification and systematic risk, using evidence from public companies both diversified and
focused, between 1978 and 1988. Same finding goes per Chan & Steiner (2000), where diversification
negatively impacts total and market risk. On the other hand, Thomas (2002) appears to find out that
22
diversified and focused firms have similar risk profiles. Still, the majority of researchers seems to find
that diversification impacts corporate risk negatively, thus we propose the following hypothesis:
!& : Product diversification is associated with reduction of firm’s risk.
Companies bear not only market risk, but also an internal financial one, which can be measured
in financial leverage. According to Low & Chen (2004), companies with more diversification tend to
have higher leverage; and in general studies show that diversified companies tend to exhibit higher
leverage ratios.
As for internationalization, existing literature shows for the most part that geographical
diversification increases risk profiles. Fatemi (1984) finds that firms that diversify internationally have
lower risk that focused domestic companies, same evidence is reported by Borde et al. (1994); both
of these studies use companies from the US as a base and expanding internationally, primarily in
Europe. Madura & Rose (1989) find a diminishing negative effect in increased proportion of foreign
sales, Goldberg & Herflin (1995) report similar results. On the other hand, some other studies, such
as Doukas & Kan (2006), report increased risk for firms with higher degree of internationalization.
Rather interesting evidence is provided by Kwok & Reeb (2000), who analyze international
companies from 32 countries and find that companies from emerging markets experience lower risk
profiles and vice versa.
As in our study we analyze companies for an emerging market – Russia, and since the majority
of previous findings show that there is negative relationship between geographical diversification and
risk, we define our last hypothesis as the following:
!' : Geographical diversification is associated with reduction of firm’s risk.
The question of diversification and risk relationship does not get sufficient consideration in
the research, and due to not enough evidence the opinion on diversification effect on risk is not
unanimous and is still ambiguous. With this study we are looking to contribute to the existing body
of research on the topic and possible find some worthy evidence from Russian companies.
1.3.2. Measures of corporate risk
Having reviewed papers on diversification and risk relationship in the previous sub-section
and having seen that they yield varying results, it is important to note that these papers use varying
diversification and risk measures. As we have reviewed diversification indicators in the first section,
this sub-section aims to go through the main risk measures existing in the relevant literature.
23
Similar to the performance measures, we divide the risk measures to the following two
categories:
•
Accounting-based measures;
•
Market-based measures (Ecker et al., 2009).
From the accounting-based view, an important objective of the analysis of financial statements
in general and that of ratios in particular is an assessment of the risk inherent in a firm’s operations.
Two sub-types of risk exist there: credit risk and equity risk. One indicators used to forecast financial
risk measures is firm’s earnings variability. The variance of a firm’s earnings is a direct measure of
the uncertainty and therefore risk of its earnings stream. A smooth earnings stream is assumed to be
desirable by firms, their creditors, and the financial markets. To the extent that accounting earnings
mirror a firm’s economic well-being, the variance in that measure would be expected to measure a
firm’s risk (Toms, 2011).
Earnings volatility is primarily related to the underlying uncertainty of demand for the firm’s
output and thus the variability of its sales. The effect of sales variability on earnings variability is a
function of the firm’s operating and financial leverage. Earnings variability generally has a systematic
as well as unsystematic component; the systematic component is referred to as the accounting beta.
From the accounting-based view, it is recognized that there is operating and financial risk.
Operating leverage is the percentage of fixed operating costs in a firm’s overall cost structure and
financial leverage is the percentage of fixed financing costs in a firm’s overall cost structure. The
higher the percentage of fixed costs, the greater the variation in income as a result of variation in sales.
Another financial leverage used in the literature is debt and equity ratio. That said, financial leverage
can use other surrogates such times interest earned or debt and assets (Konchitchki et al., 2016).
Other frequently accounting-based measures include earnings variability, contribution margin
ratio (percentage of revenues that is available to cover a company's fixed costs, fixed expenses, and
profit), operating leverage effect ratio (how much income will change given a percentage change in
sales volume), financial leverage ratio (the amount of debt held by the business firm that they use to
finance their operations; debt creates additional business risk to the firm if income varies because debt
has to be serviced), combined leverage ratio.
As for the market-based measures of risk, two types are recognized:
•
Unsystematic risk: factors specific to the firm. Diversification eliminates unsystematic
risk.
24
•
Systematic risk: factors common across a wide spectrum of firms. The only risk measure
that remains relevant. Beta is the measure of systematic risk (Beaver et al., 1970).
Beta has become one of the most common indicators, and it is important due to the following
reasons. To construct investment portfolios with the desired risk and return characteristics, one must
know the beta of individual characteristics. Discounted cash flow valuation models require an estimate
of the firm’s expected rate of return, beta can be used to estimate that return. Management in making
capital budgeting decisions needs to know the firm’s cost of capital.
In the broadest way, beta is defined as the speed or relation to which company’s stock and
market return move together. If beta equals one, that means that the returns of the security have the
same move patter as market returns. If beta is positive, that means that the stock and market returns
move in the same direction, and the opposite holds for a negative beta (Beaver et al., 1970).
A tremendous amount of research has been conducted on the beta. Ball & Brown (1968) found
a high degree of association between the accounting beta and market beta. Lev (1973) found that the
lower the variable cost, the higher the total variance of returns and the higher the beta are. Beaver et
al. (1970) found that dividend payout, financial leverage, earnings variability, and accounting beta
have significant correlations with market beta. Fama & French (1992) found, however, that alternative
measures of risk tend to be more closely related to returns while returns are not related to the beta.
1.4.
Summary of Chapter 1
In this chapter we review the theoretical background on diversification first and then survey
the literature on corporate diversification and performance and risk relationship. We also discuss the
indicators commonly used to measure diversification, performance and risk.
In the broad way, diversification is classified to product diversification, where a company gets
revenue from several industries; and geographical diversification, where a company gets its revenue
from several geographical locations. Product diversification can be of two kinds: related and
unrelated. If a firm operates in a market that possesses a strategic fit with the core business, then it is
related diversification. Otherwise, if there is no direct strategic fit between the two, nor significant
interrelationships in the value chain, it is considered to be unrelated diversification (Allen & Gorgeon,
2002). Diversification can be measured by discrete (diversification dummy, number of industries) and
continuous (Herfindahl and entropy indices) measures; the entropy index can be decomposed to
related and unrelated entropy (Palepu, 1985).
25
Reviewing the literature, we note that most of the existing researches report a negative
relationship between corporate diversification and performance. Possible reasons for that include
agency costs, inefficient investments because of rent-seeking activities, inefficient internal capital
markets. These drawbacks, however, diminish for the related diversification: the majority of studies
in this group find that the companies that diversify in the related sectors perform better than the ones
that choose to diversify unrelatedly; these studies suggest that resource-based theory and synergy
views on diversification hold and companies are able to gain significant positive abnormal returns
when operating in related activities. As for the unrelated diversification, for the most part researchers
tend to agree that it is negatively related to performance. Finally, the majority of authors find the
relationship between internationalization and performance to be positive.
Moving on to diversification and risk relationship, the majority of researchers seem to find that
diversification impacts corporate risk negatively. The factors here include real growth options,
decision maker incentives and similarities with portfolio management. As for internationalization and
risk, existing literature shows for the most part that geographical diversification increases risk profiles;
this better applies to emerging markets rather than developed ones.
Based on theoretical review conducted in this chapter we outlined six hypotheses which will
be checked in the empirical part of this study; summary of the hypotheses is provided in Table 3.
Table 3. Summary of the hypotheses to be tested.
H
Description
H"
Product diversification is negatively related to firm’s performance
H#
Related product diversification is positively related to firm’s performance
H$
Unrelated product diversification is negatively related to firm’s performance
H%
The relationship between geographical diversification and firm’s performance is positive
H&
Product diversification is associated with reduction of firm’s risk
H'
Geographical diversification is associated with reduction of firm’s risk
26
CHAPTER 2. RESEARCH DESIGN
This chapter aims to establish the methodological basis for analysing the impact of corporate
diversification on company risk and performance. Our empirical study is a regression analysis,
therefore we describe the sample, variable calculation, and then outline the econometric models used
to conduct the analysis.
2.1.
Diversification definition
In order to proceed with the empirical analysis, it is necessary to define how exactly we
measure related and unrelated diversification. In this section we establish that, providing examples of
companies used in the study.
We employ the approach of Khanna & Palepu (2000), where the index of related
diversification is based on the various revenue sources within a company that come from the same
two-digit SIC equivalent industry, and is calculated as a function of all such instances within the
company. The index of unrelated diversification is a function of different revenue sources within the
company operating in different two-digit SIC equivalent industries. Each index is a weighted average
of the ratio of firm sales to segment sales in that two-digit category (for related diversification) or a
weighted average of the ratio of sales within a particular two-digit industry to total firm sales (for
unrelated diversification), with the weights given by the logarithm of the reciprocal of the ratio.
We provide some examples of related and unrelated diversification using companies from
Russia used in this study, the summary is provided in Table 4 below.
Table 4. Examples of related/unrelated diversification in the company data used.
Company
AFK Sistema
Irkustkenergo
Udmurtneft
Related diversification
Telecommunications
Electricity
Crude oil extraction, crude oil
distribution, oil refining
GAZ
Vehicle manufacturing, buses
manufacturing, trucks,
autocomponents, engines,
Source: company annual reports
Unrelated diversification
High technology, finance, retail, pulp
and paper, utilities, pharmaceuticals,
healthcare, railway transportation,
media, agriculture, tourism and
drilling
Metallurgy, coal, railroad services
Electricity, construction, rental
services
-
27
As a primary example we take a look at AFK Sistema, which is a major Russian group
operating in various industries. The company reports that it is active across many different segments:
telecommunications, high technology, finance, retail, pulp and paper, utilities, pharmaceuticals,
healthcare, railway transportation, media, agriculture, tourism and drilling. All of these would be
factors of unrelated diversification, except for the main sector of telecommunications.
Moving on to the next example in Irkustkenergo, the related diversification for this company
would be electric power operations, and unrelated diversification consists of divisions operating in
coal, metallurgy and railroad services.
If we consider Udmurtneft, the related diversification for it would be crude petroleum
extraction and crude petroleum distribution operations and services, and unrelated diversification
consists of electricity services, construction services, rental services.
If we consider GAZ, the related diversification segments for it would include car
manufacturing, buses manufacturing, trucks, autocomponents, and the company does not have any
unrelated diversification segments.
2.2.
Sample description
As our empirical study is based on regression analysis, we gather a sample of data for it.
Companies included in the sample had to meet two criteria. Firstly, firms had to be listed in
the Moscow Exchange in Russia. Secondly, a firm should be active in any industry other than the
financial service industry (SIC codes 6000-6999). Financial firms follow different diversification
patterns and usually have different operating strategies. Also, the debt-like liabilities of financial firms
are not strictly comparable to the debt issued by non-financial firms (Rajan & Zingales, 1995). For
these reasons, it was decided to exclude financial firms from the analysis.
The data was derived from Thomson Reuters Datastream database. The sample for the study
is a set of 64 public Russian companies that diversified over the last 15 years and operating in nonfinancial sector. The sample includes companies from 10 industries: manufacturing (11% of
companies in the sample), oil & gas (17%), construction (2%), metals & mining (16%), chemicals
(6%), real estate (2%), retail (8%), telecommunications (6%), transportation (6%), electric power
(27%). The industry classification follows the SIC standard. The summary of the sample in terms of
companies and observations is provided in Table 5.
28
Table 5. Sample composition by industry.
Industry
Electric power
Oil & Gas
Metals & Mining
Manufacturing
Retail
Chemicals
Telecommunications
Transportation
Construction
Real Estate
Total
Companies
17
11
10
7
5
4
4
4
1
1
64
%
27%
17%
16%
11%
8%
6%
6%
6%
2%
2%
100%
Observations
111
102
72
54
36
21
26
31
4
8
465
%
24%
22%
15%
12%
8%
4%
6%
7%
1%
2%
100%
All chosen companies are public and disclose all the key information which is used in the
study. As we wanted to analyze the impact in the biggest companies on the market, we have collected
data for 64 companies which have the biggest revenue in the year 2015. The data has been collected
for a time span from year 2000 to year 2015, and since different companies have not only a different
age but also have become public in a different time, the yearly data is unbalanced (yearly observation
composition is provided in Figure 1 below).
Overall we have an unbalanced panel of 465 observations.
70
60
50
40
30
20
10
0
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Figure 1. Sample composition by year.
29
2.3.
Methodology
For the purposes of this research we will use regression analysis.
As noted above, in this study we are dealing with unbalanced panel data. It is worthwhile to
note that panel data have advantage comparing to the time series of cross-sectional data. It allows to
analyze the inferences of the models more accurately, better controls for omitted variables which
allows to capture the complexities of the data incontinency (Hsiao, 2005).
The main concern with unbalanced panel data is the question why the data is unbalanced. If
observations are missing at random then this is not a problem. If the attrition of firms in the data over
time is not random, i.e. it is related to the idiosyncratic errors, then this sample selection may bias the
estimates (Wooldridge, 2012). In our sample the missing data is random, therefore it is not a problem
and we can continue.
There are several estimators that can be applied to panel data analysis, we will apply each of
them for this study and then with the help of specialized tests we will choose the most appropriate one
for each analysis.
In total we conduct 12 analyses, two per each dependent variables we have, doing it separately
for total and related and unrelated diversification. We have to do it this way because related and
unrelated and total diversification variables are correlated.
2.3.1. Models
Since we have panel data in this study, we will employ panel data estimators commonly used
in research. These estimators have certain advantages, as they are able to study group and individualspecific effects and time effects, and can deal with heterogeneity. These effect generally can be fixed
and random. The fixed effect model is used to analyze whether intercepts very across groups or time
periods, while with the random effect model one can analyze the variations in error variance
components (Greene, 2008). In other words, panel data models describe the individual behavior both
across time and across entities. In general, there are three types of models: the pooled ordinary least
squares model, the fixed effects model, and the random effects model and in this research we will use
all of them:
Pooled OLS model
The pooled model specifies constant coefficients, the usual assumptions for cross-sectional
analysis (Park, 2011).
)*+ = - + /*+0 1 + 2*+
30
Fixed Effects model
The Fixed Effects model allows the individual-specific effects 3* to be correlated with the
regressors; 3* is included as intercepts (Park, 2011).
)*+ = (- + 3* ) + /*+0 1 + 2*+
Random Effects model
The Random Effects model assumes that the individual-specific effects 3* are distributed
independently of the regressors; 3* is included in the error term (Park, 2011).
)*+ = - + /*+0 1 + (3* + 2*+ )
The following carcasses of models will be used in the study with adjustments based on
conducted tests of fit for the estimators described above:
Models for accounting-based performance:
678+,* = -: + -" ;<=>?@)+A+BC + -# D<=E><F=G?<FHGIF=G?< + -$ JGIE + -% K>?L=ℎ
+ -& N>?OG=FPGHG=) + 2+,*
678+,* = -: + -" ;<=>?@)QRCB+RS + -# ;<=>?@)TUQRCB+RS
+ -$ D<=E><F=G?<FHGIF=G?< + -% JGIE + -& K>?L=ℎ
+ -' N>?OG=FPGHG=) + 2+,*
67;+,* = -: + -" ;<=>?@)+A+BC + -# D<=E><F=G?<FHGIF=G?< + -$ JGIE + -% K>?L=ℎ
+ -& N>?OG=FPGHG=) + 2+,*
67;+,* = -: + -" ;<=>?@)QRCB+RS + -# ;<=>?@)TUQRCB+RS
+ -$ D<=E><F=G?<FHGIF=G?< + -% JGIE + -& K>?L=ℎ
+ -' N>?OG=FPGHG=) + 2+,*
Models for market-based performance:
V?PG<0 WX+,* = -: + -" ;<=>?@)+A+BC + -# D<=E><F=G?<FHGIF=G?< + -$ JGIE
+ -% K>?L=ℎ + -& N>?OG=FPGHG=) + 2+,*
V?PG<0 WX+,* = -: + -" ;<=>?@)QRCB+RS + -# ;<=>?@)TUQRCB+RS
+ -$ D<=E><F=G?<FHGIF=G?< + -% JGIE + -& K>?L=ℎ
+ -' N>?OG=FPGHG=) + 2+,*
N;+,* = -: + -" ;<=>?@)+A+BC + -# D<=E><F=G?<FHGIF=G?< + -$ JGIE + -% K>?L=ℎ
+ -& N>?OG=FPGHG=) + 2+,*
N;+,* = -: + -" ;<=>?@)QRCB+RS + -# ;<=>?@)TUQRCB+RS + -$ D<=E><F=G?<FHGIF=G?<
+ -% JGIE + -& K>?L=ℎ + -' N>?OG=FPGHG=) + 2+,*
31
Models for accounting-based risk:
YEZE>F[E+,* = -: + -" ;<=>?@)+A+BC + -# D<=E><F=G?<FHGIF=G?< + -$ JGIE
+ -% K>?L=ℎ + -& N>?OG=FPGHG=) + 2+,*
YEZE>F[E+,* = -: + -" ;<=>?@)QRCB+RS + -# ;<=>?@)TUQRCB+RS
+ -$ D<=E><F=G?<FHGIF=G?< + -% JGIE + -& K>?L=ℎ
+ -' N>?OG=FPGHG=) + 2+,*
Models for market-based risk:
\E=F+,* = -: + -" ;<=>?@)+A+BC + -# D<=E><F=G?<FHGIF=G?< + -$ JGIE + -% K>?L=ℎ
+ -& N>?OG=FPGHG=) + 2+,*
\E=F+,* = -: + -" ;<=>?@)QRCB+RS + -# ;<=>?@)TUQRCB+RS
+ -$ D<=E><F=G?<FHGIF=G?< + -% JGIE + -& K>?L=ℎ
+ -' N>?OG=FPGHG=) + 2+,*
2.3.2. Tests
In order to determine which estimator to use for each analysis, we will check whether there
are fixed or random effects present; for this purposes we will use F-test and Breusch and Pagan
Lagrange multiplier (LM) test respectively. If there are both effects present, we will compare them
with the help of Hausman test.
F-test
We use this test to check whether there are fixed effects present in the data. According to this
test, the null hypothesis here is that all dummy parameters except for the dropped one are all zero,
!: : ^" =. . . = ^U`" = 0 . Accordingly, the alternative hypothesis is that at least one of these
parameters is not zero (Park, 2011). If with this test we reject the null (meaning that at least one
specific intercept 3* is not equal to zero), we will conclude that there is a significant fixed effect or
significant increase in goodness-of-fit in the fixed effect model . If this is the case, the fixed effect
model is preferred for the analysis. Otherwise, we will use the Pooled OLS.
Breusch-Pagan LM Test
With this test we will check whether there are random effects present. This is a test for the
random effects model based on the OLS residual. The null hypothesis in this test is that specific
variance components are equal to zero, !: : bT# = 0. If we reject the null, we will conclude that there
is a significant random effect in the panel data, therefore using the random effect model is preferable,
as it will handle heterogeneity better than the pooled OLS (Park, 2011).
32
Hausman Test
Finally, in case if both fixed and random effects are present, we will use the Hausman Test in
order to determine which model is better to use. It tests whether there is a significant difference
between the fixed and random effects estimators by comparing fixed and random effect estimators
under the null hypothesis that individual effects are uncorrelated with any regressor in the model
(Park, 2011). If as a result we reject the null hypothesis, we will conclude that using the random effect
model is problematic. Therefore, we will to go for the fixed effect model.
2.4.
Variables
In this section we describe the variables used in this study. Table 6 provides an overview and
classification of all variables that are included in the research.
Table 6. Classification of variables used in the research.
Type
Measure of
Independent
Diversification
Based on
Product
Geographical
Accounting
Performance
Dependent
Market
Risk
Control
Accounting
Market
Variable
Total entropy index
Related entropy index
Unrelated entropy index
Degree of internationalization
ROA
ROE
Tobin's Q
PE
Leverage
Beta
Size
Growth
Profitability
Name
ent_tot
ent_rel
ent_unrel
inter
roa
roe
tob_q
pe
lev
beta
size
growth
op_mar
Calculation of independent variables
As in this study we analyze the impact of product and geographical diversification, we use
both product and geographical diversification measure variables, and these are in fact our independent
variables.
For product diversification we use Entropy indices. For the Entropy indices we use three
different measures: total entropy, related entropy and unrelated entropy, which measure total
diversification, related diversification and unrelated diversification, respectively.
33
The Entropy measure is computed as follows:
U
L* ln(1 L* ),
V?=FH ;<=>?@) G<dEe =
*i"
where L* is weight of segment i revenue. This measure considers the weighted average
importance of each segment or industry a firm is active in (Palepu, 1985).
A significant advantage of the entropy measure over other continuous measures is that its total
diversification can be decomposed into related and unrelated diversification. Henceforth, Total
diversification = Related diversification + Unrelated diversification (Palepu, 1985). Related
diversification is defined as:
U
Lj ln(1 Lj ),
6EHF=Ed ;<=>?@) G<dEe =
ji"
where Lj is weight of segment j revenues; these j segments include revenues from all
operations within the main 2-digit SIC industry group of the company (Palepu, 1985).
Unrelated diversification is defined as:
U
Ll ln(1 Ll ),
k<>EHF=Ed ;<=>?@) G<dEe =
li"
where Ll is weight of segment m revenues; these m segments include revenues from all
operations in different 2-digit SIC segments comparing to the main SIC industry of the company
(Palepu, 1985).
As for the measure of geographical diversification, we use the degree of internationalization,
which is defined as:
n?>EG[< WFHEW
V?=FH WFHEW
It is the most frequently used measure of internationalization, even though it does not allow
mE[>EEE ?O G<=E><F=G?<FHGIF=G?< =
for a separation between sales by foreign subsidiaries and sales attributed to exports from the parent
company. We also use it in our study due to data availability reasons.
Calculation of dependent variables
As in this study we analyze the impact of the above outlined measures of diversification on
firm’s performance and risk, we use various measures of performance and risk as our dependent
variables.
34
As outlined in the first chapter of this study, performance can be captured by a wide variety of
measures, both accounting and market-based, to measure operating and market-based performance
respectively. To better analyze the impact of diversification on the performance, in this study we use
two accounting measures and two market-based measures, making it four performance variables in
total. For the accounting-based measures we use ROA and ROE, which are calculated as follows:
oE= G<p?qE
V=FH FWWE=W
oE= G<p?qE
67; =
V?=FH Er3G=)
678 =
As outlined in the first chapter, ROA is the most widely used accounting-based measure in the
diversification-performance literature, due to the reason that this measure is an effective indicator of
both operating and financial sides of company performance. For this reason, we use it as our main
performance measure. We also employ ROE, due to the reason that it shows how efficient a company
is in term of equity usage.
As for market-based variables, we use Tobin’s Q and Price to earnings ration which are
calculated as follows:
V?=FH qF>sE= ZFH3EE
V?=FH FWWE=W
N>GpE @E> WℎF>E
N; =
;F><G<[W @E> WℎF>E
V?PG<0 W X =
As outlined in the first chapter, Tobin’s Q is the most widely used one in the diversificationperformance literature. So the higher the indicator is, the more evident it is that the firm has taken
advantage of the investments and proceeded to develop the company efficiently. We also use the PE
ratio in order to track investor expectation of the companies.
Again, as outlined in the first chapter of this study, risk can be measured by a variety of
measures, both accounting and market-based, to measure financial and market-based risk respectively.
To better analyze the impact of diversification on the risk, in this study we use both accounting-based
and market-based measures of risk. The accounting measure we use is leverage ratio, which is
calculated as:
YEEZE>F[E =
V?=FH mEP=
V?=FH ;r3G=)
As for market-based risk, we use the systematic risk measure of beta, which is calculated as
follows:
35
1* =
t?Z(6* , 6l )
,
uF>(6l )
where 6* represents stock returns, and 6l stands for market returns.
Calculation of control variables
We use several control variables in order to more clearly determine the effect of diversification
strategies on performance and risks by isolating other influences on firm variables. To isolate the
relationship between diversification and performance, an adequate control of other independent
variables affecting performance is required. Therefore, this study controls for firm size, growth and
profitability, the measures which have shown to affect performance in prior research (Palich et al.,
2000).
We measure firm size by taking a natural logarithm of firm’s revenues:
JGIE = ln(6EZE<3E)
For the growth control measure we use yearly percentage change in revenues:
K>?L=ℎ =
6EZU − 6EZU`"
6EZU`"
As finally for the profitability control measure we use the operating margin:
7@@E>F=G<[ qF>[G< =
2.5.
7@E>F=G<[ G<p?qE
V?=FH >EZE<3E
Summary of Chapter 2
In this chapter we establish the methodology of the research, explain how we define related
and unrelated diversification providing examples of companies studied, outline variables and their
measurement, and explain which quantitative instruments we use in the empirical study.
For calculating related and unrelated diversification for the purposes of our study, we employ
the approach of Khanna & Palepu (2000), where the index of related diversification is based on the
various revenue sources within a company that come from the same two-digit SIC equivalent industry,
and is calculated as a function of all such instances within the company. The index of unrelated
diversification is a function of different revenue sources within the company operating in different
two-digit SIC equivalent industries. Each index is a weighted average of the ratio of firm sales to
segment sales in that two-digit category (for related diversification) or a weighted average of the ratio
36
of sales within a particular two-digit industry to total firm sales (for unrelated diversification), with
the weights given by the logarithm of the reciprocal of the ratio.
In order to conduct the empirical analysis of the objectives of the study, we employ quantitative
instruments, namely regression analysis on a certain data sample. Overall, the sample for the study is
a set of 64 public Russian companies that diversified over the last 15 years and operating in nonfinancial sector. As we wanted to analyze the impact in the biggest companies on the market, we have
collected data for 64 companies which have the biggest revenue in the year 2015. The data has been
collected for a time span from year 2000 to year 2015, and since different companies have not only a
different age but also have become public in a different time, the yearly data is unbalanced. Overall
we have an unbalanced panel of 465 observations.
For the purposes of this study we use 13 variables in total: 4 independent (3 on product
diversification and 1 on geographical diversification), 6 dependent (2 on accounting-based
performance, 2 on market-based performance, and 1 for accounting and market-based risk), and 3
control variables.
There are several estimators that can be applied to panel data analysis, we will apply each of
them for this study and then with the help of specialized tests we will choose the most appropriate one
for each analysis. The estimators we employ are: pooled OLS model, fixed effects model, and random
effects model; and the tests we conduct are: F-test, Breusch and Pagan LM test and Hausman test.
In total we conduct 12 analyses, two per each dependent variables we have, doing it separately
for total and related and unrelated diversification.
37
CHAPTER 3. RESEARCH FINDINGS
3.1.
Descriptive statistics
In order to provide the reader with a comprehensive overview of the data, in this section we
provide descriptive statistics of the variables used in this study; the summary statistics is presented
in Table 7 below.
Table 7. Summary statistics.
Variable
year
roa
roe
tob_q
pe
lev
beta
ent_tot
ent_rel
ent_unrel
inter
size
growth
op_mar
Mean
2010.237
0.065
0.118
0.757
14.671
1.016
0.707
0.526
0.451
0.075
0.235
11.962
0.195
0.130
Std. Dev.
3.201
0.087
0.225
1.123
20.668
2.422
0.250
0.414
0.417
0.139
0.304
1.213
0.389
0.173
Min
2000
-0.308
-1.287
0.002
0.543
0
0.103
0
0
0
0
9.700
-0.481
-1.847
Max Observations
2015
465
0.483
465
1.536
465
7.323
465
190.245
465
35.680
465
2.110
465
1.867
465
1.867
465
1.006
465
1
465
15.536
465
5.825
465
0.592
465
Based on the summary statistics we can make several remarks. If we take a look at accounting
based measures of performance used in this study, we see that companies in the sample on average
have Return on Assets of 6.5%, Return on Equity of 11.8%. It is only normal that ROA is significantly
smaller that ROE, that shows that companies in our sample are on average in a healthy shape. The
fact that ROE is almost twice as high as ROA tells us that on average the companies in the sample
have debt more or less equal to equity. This is proved by the average Leverage variable of 1.02, which
means that companies in the sample indeed have on average equal liabilities and equity.
Now moving on to the market based performance measures: if we take a look at Tobin’s Q,
we see that on average it is 0.71, meaning that cost to replace firm’s assets is greater than firm’s stock,
which in turn means that a stock is undervalued. We also see that there are extreme cases with Tobin’s
Q of 7.32 (maximum value in the sample). As for PE ratios, we see that on average it is 14.67, which
means that on average investors in these companies are ready to pay 14 times more for $1 of earnings.
As for the risk measures, we already discussed leverage of 1.01 on average, which showed that
companies in the sample on average have equal liabilities and equity. Apart from that, in the sample
38
there are companies with zero debt (and therefore 0 leverage) and as much as 35 times debt compared
to equity (maximum in the sample). As for the beta, we see its average is 0.71 – meaning that stocks
in the sample are theoretically 29% less volatile than the market. However, in our sample we have
both low and high betas of 0.1 and 2.1 respectively, meaning that different risk profiles are
represented.
In terms of diversification measures, we see that average total diversification is 0.53, which
means that companies in the sample are moderately diversified. However, there are companies in the
sample with no diversification and highly diversified, as showed by minimum and maximum values
of 0 and 1.87 respectively. Also, as average related diversification is 0.45 and average unrelated
diversification is 0.08, it can be concluded that there is much more related diversification going on in
the sample than unrelated diversification. As for geographical diversification (or internationalization,
how it is defined in this study), we see that on average it is 0.23, meaning that companies in the sample
are more skewed to being domestic rather than international. However, there are cases of both
completely domestic and completely international, as shown by minimum and maximum values of
zero and one.
The correlation coefficients between performance, risk and diversification variables employed
in the regression model are reported in Table 8.
Table 8. Correlation matrix.
roa
roe
tob_q
pe
lev
beta
ent_tot
ent_rel
ent_unrel
inter
roa
roe
tob_q
pe
lev
beta
1
0.740
0.225
-0.202
-0.236
0.157
-0.063
-0.086
0.071
0.402
1
0.185
-0.136
-0.100
0.040
-0.073
-0.087
0.043
0.238
1
0.033
-0.053
-0.066
-0.194
-0.127
-0.196
-0.057
1
0.024
-0.097
-0.055
-0.035
-0.058
-0.069
1
-0.022
0.025
0.030
-0.016
-0.041
1
0.069
-0.021
0.267
0.429
ent_tot ent_rel
1
0.944
1
0.146 -0.189
0.008 -0.008
ent_unrel inter
1
0.046
1
The results show that total diversification is negatively correlated to both accounting based
and market based performance measures and positively correlated to both risk measures. As for related
diversification, we see that it repeats the pattern for all performance measures, however there is
negative correlation with beta. As for unrelated diversification, it is positively correlated to ROA and
ROE and negatively correlated to both market based performance measures. Besides, there is negative
correlation with leverage and positive correlation with beta. Finally, geographic diversification repeats
the same pattern as unrelated diversification: it is positively correlated to both accounting based
39
performance measures and negatively correlated to both market based performance measures, and
there is negative correlation with leverage and positive correlation with beta.
3.2.
Model findings
In this section we describe the output the models provide for the analysis. Here we analyze
each regression with all three models (Pooled OLS, Fixed Effects and Random Effects) and describe
preliminary results. In the next session we will conduct tests necessary to chose one most fitting
model for each regression and describe final results.
Diversification/ROA
The results of regression run for ROA using all three estimators are presented in Table 9. We
see that increased total diversification leads to decreased ROA in OLS and the relationship is not
significant in Fixed effects and Random Effects models. The same pattern is in place for related
diversification: there is negative influence on ROA in OLS and there is no significant relationship in
Fixed effects and Random Effects models. As for unrelated diversification, the relationship is only
significant in the Fixed Effects model where positive relationship in 1 percent level is evident. Finally,
internationalization relationship is significant in all three models, and we see that in all of these model
the increased internationalization leads to increased ROA.
Table 9. The effects of total and related/unrelated diversification on ROA.
roa
ent_tot
OLS
-0.0161**
FE
0.0065
RE
-0.0058
roa
OLS
FE
RE
ent_rel
-0.0162** 0.0074
-0.0071
ent_unrel
-0.0148
0.1340*** 0.0302
inter
0.0449*** 0.1355**
0.0499***
inter
0.0450*** 0.1420**
0.0503***
size
0.0056**
-0.0018
0.0039
size
0.0055**
-0.0022
0.0031
growth
0.0370*** 0.0389*** 0.0370***
growth
0.0370*** 0.0392*** 0.0373***
op_mar
0.2790*** 0.2634*** 0.2734***
op_mar
0.2789*** 0.2622*** 0.2719***
_cons
-0.0476
0.0092
-0.0356
_cons
-0.0472
0.0027
-0.0291
F-test
87.26***
44.89***
F-test
72.56***
40.45***
R2
0.4877
0.4369
0.4844
R2
0.4873
0.4144
0.4813
Note: (***), (**), (*) indicate that coefficients are significant at 1, 5 and 10 percent level respectively
Diversification/ROE
Now moving on to our next accounting-based measure of performance, ROE. The results of
regression run using all three estimators are presented in Table 10. From this table we can conclude
the following: increased total diversification leads to decreased ROE in OLS model and the
relationship is not significant in Fixed effects and Random Effects models. The same pattern is in
place for related diversification: there is negative influence on ROA in OLS and there is no significant
40
relationship in Fixed effects and Random Effects models. As for unrelated diversification, it is not
significant using any of the estimators. Finally, geographic diversification relationship is significant
only in the Random Effects model, and it exerts positive influence on ROE.
Table 10. The effects of total and related/unrelated diversification on ROE.
roe
ent_tot
OLS
-0.0452**
FE
0.0102
RE
-0.0072
roe
OLS
FE
RE
ent_rel
-0.0458** 0.0116
-0.0086
ent_unrel
-0.0355
0.2018
0.0931
inter
0.0551
0.0945
0.1395**
inter
0.0556
0.1042
0.1410**
size
0.0062
0.0010
-0.0005
size
0.0060
0.0003
-0.0021
growth
0.1016*** 0.0905*** 0.0976***
growth
0.1018*** 0.0910*** 0.0979***
op_mar
0.5205*** 0.5589*** 0.5362***
op_mar
0.5197*** 0.5571*** 0.5336***
_cons
-0.0339
-0.0116
0.0132
_cons
-0.0311
-0.0214
0.0252
F-test
30.14***
19.73***
F-test
25.06***
16.93***
R2
0.2472
0.2346
0.2353
R2
0.2472
0.2225
0.2328
Note: (***), (**), (*) indicate that coefficients are significant at 1, 5 and 10 percent level respectively
Diversification/Tobin’s Q
Proceeding to the market-based measures of performance, we start with Tobin’s Q. The results
of regression runs using all three estimators are presented in Table 11. As in cases with ROA and
ROE, increased total diversification again leads to decreased performance indicator, in this case
Tobin’s Q. The relationship is not significant using Fixed Effects and Random Effects estimators.
Related diversification exerts negative influence on Tobin’s Q too; again using OLS, as the
relationship is not significant using Fixed Effects and Random Effects estimators. Same effect goes
for unrelated diversification; negative relationship is evident. To conclude with Tobin’s Q, we see that
for internationalization there is again negative relationship with Tobin’s Q, and again it is significant
only using OLS.
Table 11. The effects of total and related/unrelated diversification on Tobin’s Q.
tob_q
ent_tot
OLS
-0.5331***
FE
0.1817
RE
0.0784
tob_q
OLS
FE
RE
ent_rel
-0.4520*** 0.1784
0.0745
ent_unrel -2.1598*** -0.2492
-0.5448*
inter
-0.5772*** 0.3131
-0.0454
inter
-0.6631*** 0.2913
-0.0589
size
0.0849**
0.1568*** 0.1381***
size
0.1321*** 0.1582*** 0.1427***
growth
0.0751
-0.0211
-0.0197
growth
0.0481
-0.0222
-0.0214
op_mar
0.9976**
0.0569
0.0954
op_mar
1.145***
0.0610
0.1068
_cons
0.0124
-1.2919
-0.9146*
_cons
-0.4620
-1.2698** -0.9208*
F-test
6.82***
2.62**
F-test
9.47***
2.48***
R2
0.0691
0.0017
0.0004
R2
0.1104
0.0000
0.0149
Note: (***), (**), (*) indicate that coefficients are significant at 1, 5 and 10 percent level respectively
41
Diversification/PE
To analyze diversification relationship with the last market-based measure in the study, namely
PE ratio, we look at the Table 12, where results of regression runs are presented. We see that increased
total diversification leads to decreased PE in Fixed Effects and Random Effects models and the
relationship is not significant in OLS. Same applies for related and unrelated diversification: both of
them exert negative influence on PE ratio, relationship is significant using both effect estimators.
Finally, internationalization relationship is not significant in any of the models, so we can not make
conclusions about the relationship.
Table 12. The effects of total and related/unrelated diversification on PE.
pe
ent_tot
OLS
-2.5491
FE
-9.1695**
RE
-4.5732*
pe
OLS
FE
RE
ent_rel
-2.5676
-9.4078**
-4.3379*
ent_unrel -2.1768
-40.3454*** -12.8484*
inter
4.8376
-17.3310
4.7278
inter
4.8573
-18.9104
4.5133
size
-3.8232*** -5.3426*** -3.9444***
size
-3.8340*** -5.2440***
-3.7626***
growth 0.4201
-1.6230
-0.6680
growth
0.4263
-1.7049
-0.7556
op_mar -10.5515*
-15.1726** -12.1480**
op_mar
-10.5853*
-14.8800**
-11.7672*
_cons
61.9019*** 89.7653*** 64.7081***
_cons
62.0105*** 91.3657*** 63.0433***
F-test
5.18***
3.55***
F-test
4.3***
4.05***
R2
0.0534
0.0337
0.0518
R2
0.0534
0.0318
0.0493
Note: (***), (**), (*) indicate that coefficients are significant at 1, 5 and 10 percent level respectively
Diversification/Leverage
Moving on to analyzing relationships between diversification and risk, we start with the
accounting-based measure of risk used in this study, namely Leverage, we look at the Table 13, where
results of regression runs are presented. Relationship between total diversification and leverage is
significant at a 10 percent level using all three estimators and there are contradictory results: with OLS
the relationship is positive and using Fixed Effects and Random Effects we get negative relationship.
For the related diversification the we see significant at 5 level percent negative relationship using
Fixed Effects model. As for unrelated diversification, results are contradictory again: using OLS
increased unrelated diversification increases leverage and using Fixed Effects increased unrelated
diversification decreases the leverage, both relationships significant at 1 percent level. Finally, the
geographical diversification exerts positive influence on leverage using OLS models; relationship is
significant at 1 percent level.
42
Table 13. The effects of total and related/unrelated diversification on Beta.
lev
ent_tot
OLS
0.0360*
FE
-0.0532*
RE
-0.0402*
lev
OLS
FE
RE
ent_rel
0.0223
-0.0550**
-0.0388
ent_unrel 0.3124*** -0.2926*** -0.1153*
inter
0.1546*** -0.2039*
0.0857
inter
0.1692*** -0.2160*
0.0906*
size
0.0882*** 0.0244**
0.0508***
size
0.0802*** 0.0251**
0.0529***
growth
-0.0140
-0.0104
-0.0101
growth
-0.0094
-0.0110
-0.0105
op_mar
0.1415**
-0.0309
0.0014
op_mar
0.1165**
-0.0286
0.0049
_cons
-0.4197*** 0.4962*** 0.0939
_cons
-0.3391*** 0.5085*** 0.0726
F-test
46.78***
2.33***
F-test
43.23***
3.36***
R2
0.3376
0.0491
0.2962
R2
0.3616
0.1046
0.2698
Note: (***), (**), (*) indicate that coefficients are significant at 1, 5 and 10 percent level respectively
Diversification/Beta
Finally, we take a look at the relationship between diversification and market-based risk
measure of the study, namely beta. Results of regression runs are presented in Table 14. Relationship
between total diversification and beta is not significant using none of the estimators, therefore we do
not make any conclusions here. Same applies for related and unrelated diversification, no significant
relationship. And the same effect for geographical diversification. Overall, we do not make any
conclusions regarding diversification-beta relationship.
Table 14. The effects of total and related/unrelated diversification on Leverage ratio.
beta
ent_tot
OLS
0.1675
FE
0.2829
RE
0.2766
beta
OLS
FE
RE
ent_rel
0.1550
0.2871
0.2795
ent_unrel 0.4191
0.8285
0.7566
inter
0.4744
0.1556
0.6777
inter
0.4877
0.1832
0.6874
size
-0.2272** -0.0701
-0.0824
size
-0.2345** -0.0718
-0.0844
growth
-0.0458
-0.1151
-0.1017
growth
-0.0417
-0.1137
-0.1006
op_mar
-1.7390** -1.0046** -1.0477***
op_mar
-1.7619** -1.0097*** -1.0522***
_cons
3.7697*** 1.8228
2.4898*
_cons
3.8431*** 1.7948
2.4761*
F-test
2.45***
2.03**
F-test
2.05***
1.8***
R2
0.026
0.0206
0.0135
R2
0.0262
0.0179
0.0120
Note: (***), (**), (*) indicate that coefficients are significant at 1, 5 and 10 percent level respectively
To sum up, we see that the results are for the most part contradictory. For this reason, we run
some tests in order to determine which estimator works best for which model.
As discussed above, for each analysis we conduct three tests in order to distinguish the most
appropriate estimator for each regression model. With the F-test we examine presence of fixed effects.
If the null hypothesis is rejected, we may conclude that there is significant fixed effect or significant
increase in the goodness of fit in the fixed effect model. The presence of random effects is checked
by Breusch-Pagan Lagrange multiplier (LM) test. If the null hypothesis is rejected, we may conclude
43
that there is significant random effect in the data, therefore random effects estimator is preferred rather
than polled OLS. Finally, to determine which effect is more relevant and significant in each of the
models, we conduct Hausman test. Under this test, if the null hypothesis is rejected, then the fixed
effects model is preferred. Otherwise, we go for random effects estimator.
The results of null hypotheses check for each of the models are provided in Table 15. As we
can see, we have varying results for different analyses. For both total diversification-ROA and
related/unrelated diversification-ROA models we see that F-test null hypothesis is accepted, meaning
that there is no fixed effect in the model, therefore pooled OLS estimator is the most appropriate one.
Same applies to both ROE models, F-test null hypotheses are accepted and we choose Pooled OLS
estimator.
Table 15. Summary of tests conducted.
Measure
ROA
ROE
TOB_Q
PE
LEV
BETA
Diversification
Total
Related/unrelated
Total
Related/unrelated
Total
Related/unrelated
Total
Related/unrelated
Total
Related/unrelated
Total
Related/unrelated
F-Test
!: accepted
!: accepted
!: accepted
!: accepted
!: rejected
!: accepted
!: accepted
!: accepted
!: rejected
!: rejected
!: rejected
!: rejected
Tests
Breusch-Pagan LM Test
!: accepted
!: accepted
!: accepted
!: accepted
!: rejected
!: accepted
!: accepted
!: accepted
!: rejected
!: rejected
!: rejected
!: rejected
Hausman Test
!: rejected
!: rejected
!: accepted
!: accepted
!: rejected
!: rejected
!: rejected
!: rejected
!: rejected
!: rejected
!: accepted
test not conclusive
Moving on to regressions with market-based measures of performance. For the total
diversification-Tobin’s Q model, we see that F-test null hypothesis is rejected, which indicates
presence of fixed effects. Using the Breusch-Pagan LM test, the null hypothesis is rejected as well,
meaning that there is significant random effect. Now, as for Hausman test, the null hypothesis is
rejected again, indicating that the fixed effect estimator is preferred, therefore we choose this one for
this particular regression. As for the related/unrelated diversification-Tobin’s Q model, F-test null
hypothesis is accepted, meaning that there is no significant fixed effect in the model and we go for the
Pooled OLS estimator. Same pattern applies for both models in diversification-PE ratio analysis, in
44
each regression there is no evidence of significant fixed effect and we go for the Pooled OLS
estimators.
Finally, we take a look at the test for diversification-risk analysis. In both total diversification
and related/unrelated diversification leverage models we reject both F-test and Breusch-Pagan LM
test null hypotheses, indicating that there are both fixed and random effects in these models. In the
total diversification-beta model we accept the null hypothesis in Hausman test, meaning that random
effect estimator is preferred. For related/unrelated diversification-beta model the Hausman test is not
conclusive, therefore we also choose the random effect estimator. Finally, for both model in the
diversification-leverage relationship we see the same pattern: there is evidence of significant fixed
effects and significant random effects, and with the help of Hausman test we see that fixed effect
estimator fits better in both of these models.
Having conducted these test, we now know what are the best fitting estimators for each model;
as it is summarized in Table 16 below.
Table 16. Summary of best fitting estimators.
Measure
ROA
ROE
TOB_Q
PE
LEV
BETA
3.3.
Diversification model
Total
Related/unrelated
Total
Related/unrelated
Total
Related/unrelated
Total
Related/unrelated
Total
Related/unrelated
Total
Related/unrelated
Best fitting estimator
Pooled OLS
Pooled OLS
Pooled OLS
Pooled OLS
Fixed Effects
Pooled OLS
Pooled OLS
Pooled OLS
Fixed Effects
Fixed Effects
Random Effects
Random Effects
Results
Now that we know which estimator is the best-fitting for each model, we can gather final
results which are outlined in this section. In the next section these results will be discussed. The
summary of the results is provided in the Table 17.
Diversification-ROA
In the diversification-ROA analysis, the best fitting model (pooled OLS) gives us the following
results. Increased total diversification leads to decreased return on assets: for one-tenth of a unit
45
increase in the total diversification index, ROA is expected to decrease by 0.16 percentage points,
holding other variables constant. Related diversification exerts negative influence on ROA: whenever
related diversification increases by one-tenth of a unit, ROA again will decrease by 0.16 percentage
points, holding other variables constant. No significant relationship between unrelated diversification
and ROA is found, so we make no inferences here.
Finally, internationalization exerts positive influence on ROA: if the geographical
diversification index increases by one-tenth of a unit, a company will see a 0.45 percentage point
increase in its ROA, holding other variables constant.
The significant results have 1 to 5 percent significance level and overall this Pooled OLS
model fits the data well at 1 percent significance level. R-squared of 0.49 says that this model accounts
for 49 percent of the total variance in return of assets of the companies in the sample.
Diversification-ROE
In the diversification-ROE analysis, the best fitting estimator is again the Pooled OLS, and
using it we obtain the following results. Increased total diversification leads to decreased return on
equity: for one-tenth of a unit increase in the total diversification index, ROE is expected to decrease
by 0.45 percentage points, holding other variables constant. Related diversification exerts negative
influence on ROE: whenever related diversification increases by one-tenth of a unit, ROE will
decrease by 0.46 percentage points, holding other variables constant.
As in the case with ROA, no significant relationship between unrelated diversification and
ROE is found, so we make no inferences here. Finally, there is no significant relationship between
geographical diversification and ROE.
The significant results have 5 percent significance level and overall this Pooled OLS model
fits the data well at 1 percent significance level, as shown by p-value of less than 0.0000. R-squared
of 0.25 says that this model accounts for 25 percent of the total variance in return of equity of the
companies in the sample.
Diversification-Tobin’s Q
In the diversification-Tobin’s Q analysis, the best fitting estimator for total diversification is
Fixed Effects, and for related/unrelated diversification it is the Pooled OLS. Using these estimators
we obtain the following results. If the company is more totally diversified comparing to its own
average, it has a higher Tobin’s Q: whenever total diversification index increases by one-tenth of a
unit, Tobin’s Q will increase by 0.018, holding other variables constant. Related diversification exerts
negative influence on Tobin’s Q: for one-tenth of a unit increase in the related diversification index,
46
Tobin’s Q is expected to decrease by 0.045, holding other variables constant. Unrelated diversification
exerts negative influence on Tobin’s Q as well: if the unrelated diversification index increases by onetenth of a unit, a company will see a 0.216 decrease in its Tobin’s Q, holding other variables constant.
Finally, increased geographical diversification leads to decreased Tobin’s Q: for one-tenth of a unit
increase in the internationalization index, Tobin’s Q is expected to decrease by 0.066 percentage
points, holding other variables constant.
The significant results in this output have 1 percent significance level. Overall, the Fixed
Effects model fits the data for the total diversification analysis at 5 percent significance level, and the
Pooled OLS model fits the data for the related/unrelated diversification analysis well at 1 percent
significance level, as shown by p-value of less than 0.0000. R-squared of 0.03 says that Fixed Effects
model accounts for 3 percent of the total variance in Tobin’s Q of the companies in the sample, and
R-squared of 0.11 says that OLS model accounts for 11 percent of the total variance in Tobin’s Q of
the companies in the sample.
We also see that the rho for the Fixed Effects model is around 0.83, which means that over
83% of the variation in beta is due to individual-specific terms, and the rest is due to idiosyncratic
errors. This means that the models are good, as we may not know where the variation is coming from
but it is possible to assign it to a particular company.
Diversification-PE
Moving on to the diversification-PE analysis, we see that the best fitting estimator for total
diversification is Pooled OLS, and for related/unrelated diversification it is the Fixed Effects model.
Using these estimators, we obtain the following results. No significant relationship between total
diversification and PE ratio is found, so we make no inferences here. As for related diversification,
we see that higher related diversification comparing to company’s own average exerts negative
influence on PE: for one-tenth of a unit increase in the related diversification index, PE ratio is
expected to decrease by 0.941, holding other variables constant. Same applies to unrelated
diversification, higher unrelated diversification comparing to company’s own average exerts negative
influence on PE: whenever the unrelated diversification index increases by one-tenth of a unit, PE
ratio will decrease by 4.034, holding other variables constant. Finally, no significant relationship
between internationalization and price to earnings ratio is found.
47
Table 17. Best fitting estimator results for each model.
var
divers
roa
total
model
estimator
roe
rel/unrel
total
tob_q
rel/unrel
total
pe
rel/unrel
total
lev
rel/unrel
total
beta
rel/unrel
total
rel/unrel
1
2
3
4
5
6
7
8
9
10
11
12
OLS
OLS
OLS
OLS
FE
OLS
OLS
FE
FE
FE
RE
RE
ent_tot
-0.016**
-
-0.045**
-
0.182***
-
-2.549
-
-0.053*
-
0.277
-
ent_rel
-
-0.016**
-
-0.046**
-
-0.452***
-
-9.408**
-
-0.055**
-
0.279
ent_unrel
-
-0.014
-
-0.035
-
-2.160***
-
40.345***
-
-0.293***
-
0.757
inter
0.045***
0.045***
0.055
0.056
0.313
-0.663***
4.838
-18.910
-0.204*
-0.216*
0.678
0.687
size
0.006**
0.005**
0.006
0.006
0.157***
0.132***
-3.823***
-5.244***
0.024**
0.025**
-0.082
-0.084
growth
0.037***
0.037***
0.102***
0.102***
-0.021
0.048
0.420
-1.705
-0.010
-0.011
-0.102
-0.101
op_mar
0.279***
0.279***
0.520***
0.520***
0.057
1.145***
-10.551*
-14.880**
-0.031
-0.029
-1.048***
-1.052***
_cons
-0.048
-0.047
-0.034
-0.031
-1.292
-0.462
61.902***
91.366***
0.496***
0.508***
2.490*
2.476*
F-test
87.26***
72.56***
30.14***
25.06***
2.62**
9.47***
5.18***
4.05***
2.33***
3.36***
-
-
DF
459
458
459
458
396
458
459
395
396
395
-
-
R2
0.487
0.487
0.247
0.247
-
0.110
0.053
-
-
-
-
-
R2 within
-
-
-
-
0.032
-
-
0.058
0.029
0.048
0.024
0.026
R2 between
-
-
-
-
0.001
-
-
0.043
0.004
0.059
0.066
0.053
R2 overall
-
-
-
-
0.002
-
-
0.032
0.049
0.105
0.013
0.012
sigma_u
-
-
-
-
1.064
-
-
15.488
0.255
0.270
4.306
4.344
sigma_e
-
-
-
-
0.472
-
-
17.972
0.122
0.120
0.991
0.992
rho
-
-
-
-
0.836
-
-
0.426
0.815
0.834
0.950
0.950
N
465
465
465
465
465
465
465
465
465
465
465
465
Note: (***), (**), (*) indicate that coefficients are significant at 1, 5 and 10 percent level respectively
48
The significant results in this output have a 5 percent significance level. Overall, both the
Pooled OLS for the total diversification analysis and the Fixed Effects estimator for the
related/unrelated diversification analysis fit the data well at a 5 percent significance level, as shown
by p-values of less than 0.0000. R-squared of 0.05 says that OLS model accounts for 5 percent of the
total variance in PE of the companies in the sample, and R-squared of 0.03 says that the Fixed Effects
model accounts for 3 percent of the total variance in PE of the companies in the sample.
We see that the rho for the Fixed Effects model is around 0.43, which means that over 43% of
the variation in beta is due to individual-specific terms, and the rest is due to idiosyncratic errors. This
means that the models are good, as we may not know where the variation is coming from but it is
possible to assign it to a particular company.
Diversification-Leverage
In the diversification-leverage analysis, we see that the best fitting estimator for both total and
related/unrelated diversification beta analysis is Fixed Effects. Using this estimator, we obtain the
following results. Higher total diversification comparing to company’s own average leads to lower
Leverage: whenever the total diversification index increases by one-tenth of a unit, leverage will
increase by 0.0053, holding other variables constant. Higher related diversification comparing to
company’s own average exerts negative influence on leverage: for one-tenth of a unit increase in the
related diversification index, leverage is expected to decrease by 0.0055, holding other variables
constant. Higher unrelated diversification comparing to company’s own average also exerts negative
influence on leverage: if the unrelated diversification index increases by one-tenth of a unit, a
company will see a 0.03 decrease in its leverage, holding other variables constant. Finally,
internationalization exerts negative influence on leverage too: for one-tenth of a unit increase in the
geographical diversification index, leverage is expected to decrease by 0.021, holding other variables
constant.
We see that the rho for both of these Fixed Effects models is around 0.83, which means that
over 83% of the variation in leverage is due to individual-specific terms, and the rest is due to
idiosyncratic errors. This means that the models are good, as we may not know where the variation is
coming from but it is possible to assign it to a particular company.
Diversification-Beta
Finally, moving on to the diversification-beta analysis, there is presence of random effects in
the data, so the best fitting estimator for both total and related/unrelated diversification beta analysis
is Random Effects. We see that using this estimator the results are not significant for any of the
49
analyses regarding diversification and beta: total, related, unrelated diversification and
internationalization.
We see that the rho for both of these Random Effects models is around 0.95, which means that
over 95% of the variation in beta is due to individual-specific terms, and the rest is due to idiosyncratic
errors. This means that the models are good, as we may not know where the variation is coming from
but it is possible to assign it to a particular company. However, as the coefficients are not significant,
we can not make any inferences regarding diversification and beta.
Summary of diversification influence on performance and risk measures is provided in Table
18 below.
Table 18. Summary of independent variable’s influence on dependent variables.
ROA
ROE
Tobin's Q
Total diversification
Negative Negative
Positive
Related diversification
Negative Negative Negative
Unrelated diversification
Negative
Internationalization
Positive
Negative
Note: (-) indicates that no significant relationship is found
PE
Negative
Negative
-
Leverage
Negative
Negative
Negative
Negative
Beta
-
As we can see, we can conclude the following: total diversification has negative relationship
with ROA, ROE and Leverage; positive influence on Tobin’s Q; and there is no significant evidence
concerning PE and Beta.
Related diversification exerts negative influence on ROA, ROE, Tobin’s Q, PE and Leverage;
and there is no significant evidence concerning Beta. Unrelated diversification has negative
relationship with Tobin’s Q, PE and Leverage; and we did not get any significant evidence concerning
ROA, ROE and Beta. Finally, internationalization exerts negative influence in Tobin’s Q and
Leverage; positive influence on ROA; and no evidence of significant relationship with ROE, PE and
Beta.
The summary of our hypotheses check results in presented in Table 19 below. Each of the
hypotheses we check separately for accounting based measures and market based measures of both
performance and risk, for this reason we sometimes get conflicting results. As can be seen, we accept
most our hypotheses, rejecting !" and !# for market-based measures and !$ for both accounting and
market-based measures. Also, we do not accept nor reject hypotheses !% for accounting-based
measures, and !& and !' for market-based measures due to lack of significant evidence.
50
Table 19. Summary of the hypotheses checks.
H
Description
Product diversification is negatively related to
firm’s performance
Related product diversification is positively related
H$
to firm’s performance
Unrelated product diversification is negatively
H%
related to firm’s performance
The relationship between geographical
H#
diversification and firm’s performance is positive
Product diversification is associated with reduction
H&
of firm’s risk
Geographical diversification is associated with
H'
reduction of firm’s risk
Note: (-) indicates that no significant evidence is found
H"
3.4.
Accounting
measures
Market
measures
Accepted
Rejected
Rejected
Rejected
-
Accepted
Accepted
Rejected
Accepted
-
Accepted
-
Discussion
3.4.1. Discussion of the findings
In this section we discuss and explain the received results, and also provide managerial
implications on the findings.
As can be seen from the previous section, the results of conducted analysis confirm the
existence of significant relationship between diversification and the majority of analyzed measures.
We will take a look at each analyses separately and try to explain it and see how it compares to
previous research findings.
To start with accounting-based measures of performance, total product diversification has a
negative relationship with ROA, or in other words, increased total diversification leads to decreased
return on assets. ROA can decrease due to the following reasons: (1) decreased returns, (2) increased
assets, (3) faster growing assets rather than returns. Usually in order to diversify into different product
or industry, a company needs to make capital investments, which tend to be rather big. In other words,
a diversified company has a much more asset-heavy profile compared to an undiversified corporation
(Villalonga, 2004). However, the returns the company gets do not necessarily match the old
undiversified profile, therefore a diversified firm tends to have a lower return on assets than an
undiversified firm, although its absolute return amount might very well be bigger.
Similar logic applies to diversification-ROE relationship, which is also negative in this study.
In order to diversify into different product or industry, a company needs to make capital investments
which are often financed by equity. Therefore, a more diversified company tends to have more equity
51
on its account, and this attributes to lower return on equity. There are outliers that despite the higher
equity have higher return on equity due to much higher returns comparing to other firms, but this is
not a common case.
In this study we analyze total and related/unrelated diversification impact on firm performance
separately and find that related diversification also exerts negative relationship on both ROA and
ROE. This is contradictory to our hypothesis 2 which states that related diversification is positively
related to firm performance. Usually it would be expected that related diversification allows a firm
with capabilities around a particular input to leverage that capability in more that one sector where the
same thing is relevant to performance, thus improving financial performance of the firm (Rumelt,
1982).
To compare our result to the previous research, there are varying results in previous studies:
among others, Miller (2006) found that there is positive relationship between related diversification
and accounting-based measures of performance, which they explained by technological synergies.
Anderson et al. (2000), in turn, found an opposing result - a diversification discount, which they tried
to explain through corporate governance structures. Finally, Diestre & Santalo (2013), although failed
to provide evidence of significant relationship and performance, found that there are critical
contingencies affecting this relationship.
In general, if we try to explain why there is negative relationship between related
diversification and accounting-based performance, there can be various reasons: for one, value loss
might occur due to a “new toy” (even though related) effect and rent dissipation by a company. Also,
as argued by Berger & Ofek (1995), overinvestment and cross-subsidization can contribute towards
the value loss of diversification. As a matter of fact, this value loss can be decreased by the tax benefits
of diversification. Overall, there can be numerous explanation for the phenomena, and this could
potentially be an interesting topic for future research.
As for the unrelated diversification and accounting-based measures of performance, our
empirical results failed to provide any significant relationship there, both for ROA and ROE, so we
make no conclusions here.
Moving on to market-based measures of performance, what is quite interesting is that although
total diversification has negative relationship with accounting-based measures of performance, it is
related positively with the market-based measure of performance used in this study: Tobin’s Q. As
we see from the results, increased total diversification leads to a higher Tobin’s Q. Since this indicator
is calculated by dividing total market value of the firm by its assets, there can be the following reasons
52
of the increase: (1) increased market value, (2) decreased assets, (3) faster growing market value rather
than assets. As discussed above, usually with diversification comes an increase in assets, therefore we
can make a conclusion that the most likely reason of Tobin’s Q increase is that diversified firms tend
to have higher market value rather than undiversified firms. Now, what could be a possible reason for
this?
Khanna & Palepu (2000) in their study have a similar result using evidence from Indian
companies. They suggest that diversified firms replicate functions or institutions that are missing in
the emerging market. Diversified firms are able to back up and justify the capital investment required
for diversification due to that they tend to be bigger and have to scope for it. Another factor here is
that the biggest and most diversified companies have political connections which they can leverage
for their benefit; it is a big factor in an economy where state regulation plays a big role. By these
factors Khanna & Palepu (2000) explain diversification premium in India. Although Indian and
Russian markets are different in context, they are both emerging markets, so we can apply these
explanations to our research as well.
Another factor to explain why more diversified firms tend to be higher valued is investor
expectations. Among other things, by diversified their businesses, companies want to avoid “having
all eggs in one basket”, in other words conduct better risk control in order to not be reliant one a single
market. This allows to spread the risk through several sectors of economy and not be exposed to
dangers of one sector declining. For investors, they view this as a productive risk management activity
and react positively to it, therefore company value goes up. A valuable note for a to-be-diversified
company here is that it is crucial to identify sectors where market slowdown will not coincide with
downturns in the main business if the firm.
Moving on, we also conducted analysis on total diversification and price to earning ratio,
however no signification relationship was found.
Related and unrelated diversification both have negative relationship with Tobin’s Q. One of
the reasons for that could be that relative to focused firms, diversified companies tend to invest more
and therefore have lower marginal return to capital, which in turn lowers the Q’s. The result is similar
to findings of Lang & Stultz (1994), who state that average Q for diversified firms is over one because
their market value capitalizes the contribution to shareholder wealth of the reduction in informational
asymmetries. Hence, we can conclude that diversification discount may occur due to inefficiencies
such as influence costs and agency costs.
53
As for another market-based ratio we employ in the study, PE ratio, we see evidence that, like
on Tobin’s Q, increased related and unrelated diversification exerts negative influence on the measure.
As we again follow a pattern similar to one employed in above in order to explain this, there can be
the following reasons for PE ratio decreasing with higher diversification: (1) decreased share price;
(2) increased earnings per share; (3) faster growing earnings per share compared to price per share.
As is evident from discussion above, price per share can both be increased or decreased with higher
diversification. As for increase in earnings per share, this also can be true, because because essentially
the undergoing motive for diversification is to increase the bottom line – so, when successful,
diversification attributes to higher earnings and therefore higher earnings per share.
Overall, our results for diversification discount coincide with numerous results of previous
researchers. Bernardo et al. (2000) found that diversified firms trade at a discount in the US, Fleming
et al. (2003) found the same evidence for Australia. However, it is not only the case for developed
markets: Lins & Servaes (1999, 2002) found that, besides UK, Singapore and Japan, the discount
existed in India, Thailand, Malaysia and Indonesia. The international evidence suggests that the
existence of discount could result from institutional differences between markets, data analysis
method differences, varying data sources or sample selection bias. One difference between markets
could be that investors in Russian companies do not favor diversification attempts by companies and
prefer them to stay focused.
It should be elaborated on the fact that Tobin’s Q produced partially contradicting results
compared to ROA. This puzzle might be explained by the fact that ROA captures realized
performance, whereas Tobin’s Q is a more future oriented measure, reflecting investors’ expectations
on the (long-term) future of the firm. Henceforth, in the light of the conclusions it is importance to
notice that investors seem to relate superior performance effects to unrelated diversification which is
in turn beneficial for the value of the firm. However, it should be noted that these expectations are
speculative in nature, and therefore do not have to mean that actual performance equals expected
performance. The fact that ROA provides best insight in real, achieved performance underscores the
conclusions drawn in this section.
Now, having discussed the results for product diversification and performance relationships,
we move to geographical diversification evidence. We see that increased internationalization has
positive effect on accounting based measure of performance, ROA, and negative effect on marketbased measure of performance, Tobin’s Q. As for the case with ROA, it seems that for the companies
in the sample going international is beneficial and they are able to increase the returns without a
54
dramatic increase in assets. This shows that there is demand for Russian companies’ products abroad,
and that these products are of high enough quality to be sold to foreign consumers.
The findings here are consistent with the results of Contractor et al. (2007), who found a
positive relationship for another emerging market – India; and more importantly, the results are
consistent with findings of Shcherbakov (2012), who used evidence from Russian companies, similar
to our study. Shcherbakov found that geographical diversification and accounting-based performance
have positive relationship which follows a U-shaped curve, meaning that in the initial stage of
internationalization the performance declines and in deeper stages of internationalization performance
improves. We do not have such detailed evidence in this study, however in general our findings
coincide.
As for our findings regarding Tobin’s Q, they are the opposite to ROA. In other words, Tobin’s
Q tends to decrease with higher geographical diversification. Since this indicator is calculated by
dividing firm’s market value by its assets, we can make a conclusion that in more internationalized
companies total assets tend to grow faster than the market value, if there is any growth at all.
Moving on to diversification-risk relationship, we did not achieve any significant relationship
between any type of diversification and our market-based risk measure expressed in beta.
Diversification-beta relationship is not very frequently studied in the literature, one example of prior
evidence could be the research by Montgomery & Singh (1984), where it was found that betas for
unrelated diversifiers were significantly higher than those of other firms. As for the international
diversification and beta relationship, Kwok & Reeb (2000) found that internationalization reduces
systematic risk. Our findings fail to provide any significant evidence, so we make no inferences here.
As for our accounting-based risk, namely leverage, we see that for all types of diversification
studied, there is evidence of negative relationship, where the increased diversification comes with
decreased leverage, thus confirming our hypothesis. Leverage ratio can decrease due to the following
reasons: (1) decreased debt amount, (2) increased equity amount, (3) faster growing equity rather than
debt amount. Based on our results, we conclude that Russian companies prefer to finance their
diversification efforts with equity rather than debt. Another interesting inference we can make here is
that Russian companies prefer to use equity financing for unrelated diversification activities, and debt
financing in a bigger extent for related diversification efforts.
Our results contrast to some of the previous studies, for example Barton & Gordon (1988),
where is was found that firms developing a strategy of unrelated diversification have the highest debt
ratio; or Kochhar & Hitt (1998), which also explored the linkage between the characteristics of a
55
firm’s diversification strategy and its capital structure. According to their findings, equity financing
is preferred for related diversification and debt financing for unrelated diversification. Our results are
partially consistent with La Rocca et al. (2009), where they found that a related-diversification strategy
has a negative influence on leverage; however, unrelated diversification, has a positive effect on debt.
As for our results of internationalization-leverage analysis, they are consistent with those of
the majority of previous research: a big body of literature shows that corporate leverage is negatively
related to internationalization. For example, Kwok & Reeb (2000) found that there is a debt reduction
associated with internationalization, as did Low & Chen (2004) and others. It appears that with
internationalization activities get higher earnings for the companies, which then use these increased
earnings for debt repayment. All in all, internationalization benefits outweigh the costs for Russian
companies.
3.4.2. Managerial implications
Our findings indicate a number of managerial implications for both decision makers in
companies and investors.
First and foremost, both our theoretical survey and empirical analysis suggest that managers
should be aware that diversification can cause both positive and negative effects. Although
theoretically diversification can result in a premium, we see that in the Russian context product
diversification almost always yields value losses. In order to be able to predict potential outcomes of
diversification, managers should be informed of what creates and destroys value when diversifying
across industries. A risky move can be to diversify across segments where the the initial position of a
firm is not very well established, and moreover, even though a firm has a well established starting
point but the industry potential is not as high. On the other hand, when diversification is carried into
the segments where existing resources and capabilities can be used and leveraged, the potential of
positive value created is high. Therefore, even when a firm is not diversified yet, but potentially thinks
of doing so in the future, managers should keep in mind the importance of developing capabilities and
resources in a way that can potentially bring value across other segments in future.
As can be inferred from our research, diversification is not a safe route and managers should
be aware of that. Diversification targets should be picked really carefully, because of the results will
be highly dependent on a number of factors. Lubatkin & Chatterjee (1994) put it the following way:
the companies should “diversify in such a manner that all of its eggs are in similar baskets - not in the
same basket or in different baskets.” Managers need to make sure that their companies have all the
56
necessary skills, capabilities and resources to operate across all these “similar baskets”. If it is not the
case, then the companies are better off by operating across different baskets all along.
In general, current business landscape becomes increasingly more focused of having core
capabilities. It seems that for a large number of companies that have expertise only in their core
sectors, it is better to focus on this core sector to achieve better and more stable results, unless they
are confident that they have the right capabilities that can be leveraged in different sectors and
industries.
An important implication for decision makers in Russian companies concerns geographical
diversification. Our findings imply that going international is beneficial for both performance and risk
profile of the company. There is demand for Russian companies’ products abroad, and that these
products are of high enough quality to be sold to foreign consumers, therefore managers should pay
more attention to export possibilities. This is especially relevant during the current economic situation,
where domestic currency in Russia is rather cheap.
3.4.3. Limitations and directions for further research
There are several limitations associated with this study; these limitations, in turn, create
suggestions for further research. First limitation is the general nature of our research. In this study we
analyze companies from several industries across one emerging market country. Due to different
particularities of market conditions in different countries (developing and emerging), and due to
different particularities of different industries, a more specific approach would be beneficial. Future
research could conduct the similar analysis across various markets and industries and then compare
the results; this would allow to make industry and market specific conclusions and implications.
Another limitation of this study is a rather limited amount of performance, risk and control
variables. Future research could include and test more dependent variables not included in this study,
and also a bigger number of control variables as well. Examples of potential performance measures
could include Return on capital employed, Operating profit and others; examples of potential risk
measures to include could be credit ratings or different types of operating risk. Including more of
dependent variables would enable to track diversification impact on different parts of corporate
performance and risk and give more comprehensive results. Including more control variables, in turn,
would allow to improve the results of the empirical analysis by better controlling for more aspect of
company operations.
57
In a broad sense, as a process, companies can conduct diversification via a number of activities
including M&A, greenfield, joint ventures. An interesting topic for future research would be to
compare and contrast how diversification activities conducted via these methods differ from each
other in terms of their impact of company performance and risk. The results of such a study would be
beneficial for managers responsible for diversification decisions in companies, as it would allow them
to better understand potential costs and benefits associated with each diversification mode.
Finally, we discussed in this study how important initial capabilities and resources a company
possesses are before it decides to diversify. In this regard, a very interesting topic for future research
would be to analyze how companies could quantify these initial capabilities and resources, and
compare them. These quantified assessments could be in a form of scorecards or more complex
models, and would be tremendously useful in helping predict diversification results. At the end of the
day, whenever a company decides to diversify, it brings itself a tremendous amount of uncertainty,
and any tools to help reduce this uncertainty would be extremely useful.
3.5.
Summary of Chapter 3
In this chapter we outline how we conduct the empirical analysis, provide results and then
discuss the findings.
Based on the summary statistics we make several inferences about the data: companies in our
sample are on average in a healthy shape; their stocks on average are undervalued; they on average
have equal liabilities and equity; they are moderately diversified; there is much more related
diversification going on in the sample than unrelated diversification; the companies in the sample are
more skewed to being domestic rather than international.
With the help of the three tests we determine which estimator is more appropriate for each of
our 12 models. We see that all three estimators (pooled OLS, fixed effects, random effects) appear as
best-fitting several times, therefore we use all three of them selectively. Conducting the appropriate
analyses, we get the results; we receive significant result for the majority of relationships, with most
of them being negative.
After getting the results and checking the hypotheses, we discuss and explain the received
findings. A diversified firm tends to have a lower return on assets than an undiversified firm due to
the higher asset profile and the problem that the returns the company gets do not necessarily match
the old undiversified profile, although its absolute return amount might very well be bigger.
58
Diversification-ROE relationship is also negative. In order to diversify into different product or
industry, a company needs to make capital investments which are often financed by equity.
We find a negative relationship between related diversification and accounting-based
performance, and the reasons could include: “new toy” effect, rent dissipation, overinvestment and
cross-subsidization. As a matter of fact, this value loss can be decreased by the tax benefits of
diversification. Overall, there can be numerous explanation for the phenomena, and this could
potentially be an interesting topic for future research.
One of the few positive relationships we find is the one between total diversification and
Tobin’s Q, for which we infer that diversified firms tend to have higher market value rather than
undiversified firms. Possible reasons include: ability to back up and justify the capital investment
required for diversification, political connections in an emerging market, and investor expectations,
who see corporate diversification as a productive risk management activity. However, it seems that it
is different for the Russian market, investors in Russian companies do not favor diversification
attempts by companies and prefer them to stay focused.
Another result is that increased internationalization has positive effect on accounting based
measure of performance, ROA, and negative effect on market-based measure of performance, Tobin’s
Q. As for the case with ROA, it seems that for the companies in the sample going international is
beneficial and they are able to increase the returns without a dramatic increase in assets. This shows
that there is demand for Russian companies’ products abroad, and that these products are of high
enough quality to be sold to foreign consumers. As for our Tobin’s Q, it tends to decrease with higher
geographical diversification. Since this indicator is calculated by dividing firm’s market value by its
assets, we can make a conclusion that in more internationalized companies total assets tend to grow
faster than the market value, if there is any growth at all.
As for our accounting-based risk, namely leverage, we see that for all types of diversification
studied, there is evidence of negative relationship. Based on our results, we conclude that Russian
companies prefer to finance their diversification efforts with equity rather than debt. Another
interesting inference we can make here is that Russian companies prefer to use equity financing for
unrelated diversification activities, and debt financing in a bigger extent for related diversification
efforts. As for our results of internationalization-leverage analysis, there is a negative relationship. It
appears that with internationalization activities get higher earnings for the companies, which then use
these increased earnings for debt repayment. All in all, internationalization benefits outweigh the costs
for Russian companies.
59
There is a number of managerial implications that can be extracted from the study. Managers
should be aware that diversification can cause both positive and negative effects and should keep in
mind the importance of developing capabilities and resources in a way that can potentially bring value
across other segments in future. Resource and capability management is crucial, if it is not managed
properly for diversification, then it is better to not diversify at all. Also, it seems that for a large number
of companies that have expertise only in their core sectors, it is better to focus on this core sector to
achieve better and more stable results. An important implication for decision makers in Russian
companies concerns geographical diversification. Our findings imply that going international is
beneficial for both performance and risk profile of the company. There is demand for Russian
companies’ products abroad, and that these products are of high enough quality to be sold to foreign
consumers, therefore managers should pay more attention to export possibilities. This is especially
relevant during the current economic situation.
Finally, we also discuss limitations and suggestions for further research. The suggestions
include: including more markets and industries to the analysis and then comparing the results;
including more variables; comparing results of diversification via different modes; and analyzing how
companies could quantify the initial capabilities and resources, and compare them with regards to
diversification.
60
CONCLUSION
This thesis was devoted to studying corporate diversification and how if affects companies and
their businesses. The research goal of the paper was to determine the relationship between corporate
diversification and company performance and risk, using evidence from Russian companies. We
accomplished this goal, and achieved all of the research objectives stated.
As a first step of the study, we investigated the theoretical framework of diversification and
outlined commonly used indicators to measure diversification. We also reviewed the existing
literature on diversification performance relationship and diversification risk relationship, which
allowed us to make preliminary conclusions and define the hypotheses to be tested. As a second part
of the study, we conducted empirical analysis which allowed us to determine the impact of
diversification on risk and performance. As we measured company performance and risk by using
both accounting and market based indicators, we managed to test our hypotheses separately for these
types of measures.
We found that a diversified firm tends perform worse than an undiversified firm in terms of
ROA and ROE, due to the higher asset and equity profile and the problem that the returns a company
gets do not necessarily match the old undiversified profile, although its absolute return amount might
very well be bigger. On the other hand, when analyzing the relationship between total diversification
and Tobin’s Q, we found that diversified firms tend to have higher market value rather than
undiversified firms; investors see corporate diversification as a productive risk management activity.
Also, as per our findings, both related and unrelated diversification yield a decrease in performance.
Overall, there can be numerous explanation for the phenomena, and this could potentially be an
interesting topic for future research.
As for internationalization, we found that it has a positive effect on an accounting based
measure of performance, ROA, and negative effect on a market-based measure of performance,
Tobin’s Q. We inferred that when the companies in the sample go international, they are able to
increase the returns without a dramatic increase in assets. We also concluded that in more
internationalized companies total assets tend to grow faster than the market value.
As for the diversification and risk relationship, we found it to be negative for both accounting
and market-based measures. Also, we concluded that Russian companies prefer to finance their
diversification efforts with equity rather than debt. Another inference we made is that Russian
companies prefer to use equity financing for unrelated diversification activities, and debt financing in
61
a bigger extent for related diversification efforts. As for our results of internationalization-risk
analysis, we found a negative relationship. We inferred that internationalization activities get higher
earnings for the companies, which then use these increased earnings for debt repayment. All in all,
internationalization benefits outweigh the costs for Russian companies.
Based on our findings, we developed a set of managerial implications. Managers should be
aware that diversification can cause both positive and negative effects and should keep in mind the
importance of developing capabilities and resources in a way that can potentially bring value across
new segments in future. Resource and capability management is crucial; if it is not managed properly
for diversification, then it is better to not diversify at all. An important implication for decision makers
in Russian companies concerns geographical diversification: our findings imply that going
international is beneficial for both performance and risk profile of the company. Also, it seems that
there is demand for Russian companies’ products abroad, therefore managers should pay more
attention to export possibilities. This is especially relevant during the current economic situation.
In order to conduct a thorough analysis, we used 116 references; and the contribution of this
study is the coherent investigation of diversification relationship with performance and risk. However,
there is clearly a scope for future research: besides of including more markets and industries to the
analysis and then comparing the results, what could be valuable is to compare results of diversification
via different modes such as M&A, greenfield or joint venture. Also, analyzing how companies could
quantify the initial capabilities and resources, and comparing them with regards to diversification
could also be a promising topic to study.
62
REFERENCES
1.
Agrawal, A., and G. Mandelker. 1987. Managerial incentives and corporate investment and
financing decisions. Journal of Finance 42: 823-837.
2.
Al-Matari E. M., A. K. Al-Swidi, and F. Fadzil. 2014. The measurements of firm performance’s
dimensions. Asian Journal of Finance & Accounting 6 (1): 24-49.
3.
Allen, D., and A. Gorgeon. 2002. Diversification Strategy. Madrid, Spain: IE Publishing
Department.
4.
Amihud, Y., and B. Lev. 1981. Risk reduction as a managerial motive for conglomerate mergers.
The Rand Journal of Economics 12: 605-618.
5.
Anderson, C. W., M. Fedenia, M. Hirschey, and H. Skiba. 2011. Cultural influences on home
bias and international diversification by institutional investors. Journal of Banking & Finance
35 (4): 916-934.
6.
Anderson, R. C., T. W. Bates, J. M. Bizjak, and M. L. Lemmon. 2000. Corporate governance
and firm diversification. Financial Management 29 (1): 5-22.
7.
Anil M. P., and V. Narendar. 1998. Diversification and firm performance: an empirical
evaluation. Journal of Financial and Strategic Decisions 11 (2): 67-82.
8.
Ball, R., and P. Brown. 1968. An empirical evaluation of accounting income numbers. Journal
of Accounting Research 6 (2): 159-178.
9.
Barney, J. B. 1997. Gaining and sustaining competitive advantage. Reading, MA: AddisonWesley.
10.
Barton, S., and P. Gordon. 1988. Corporate strategy and capital structure. Strategic Management
Journal, 9: 623-632.
11.
Beaver, W., P. Kettler, and M. Scholes. 1970. The association between market-determined and
accounting-determined risk measures. The Accounting Review 45 (4): 654-681.
12.
Berger, P. G., and E. Ofek, 1999. Causes and effects of corporate refocusing programs. The
Review of Financial Studies 12 (2): 311-345.
13.
Berger, P. G., and E. Ofek. 1995. Diversification's effect on firm value. Journal of Financial
Economics 37 (1): 39-65.
14.
Bernardo A. E., and B, Chowdhry. 2002. Resources, real options, and corporate strategy.
Journal of Financial Economics 63: 211–234.
15.
Bernardo, A., B. Chowdhry, D. Palia, and E. Sernova. 2000. Real Options and the diversification
63
discount. Working Paper, The Anderson School, University of California in Los Angeles.
16.
Berry, C. H. 1971. Corporate growth and diversification. The Journal of Law & Economics 14
(2): 371-383.
17.
Borde, S. F., K. Chambliss, and J. Madura. 1994. Explaining variation in risk across insurance
companies. Journal of Financial Services Research 8 (3): 177-191.
18.
Boz, I., I. Yigit, and I. Anil. 2013. The relationship between diversification strategy and
organizational performance: A research intended for comparing Belgium and Turkey. Procedia
- Social and Behavioral Sciences 99: 997-1006.
19.
Bozec, Y., R. Bozec, and M. Dia. 2010. Overall governance, firm value and deviation from one
share: One vote principle. International Journal of Managerial Finance, 6(4): 305-328.
20.
Campa, J.M., and S. Kedia. 2002. Explaining the diversification discount. Journal of Finance
57 (4): 1731-1762.
21.
Carlson, M., A. Fisher, and R. Giammarino. 2006. Corporate investment and asset price
dynamics: Implications for SEO event studies and long-run performance. Journal of Finance
61: 1009-1034.
22.
Chan, K., and T. Steiner. 2000. An agency analysis of firm diversification: The consequences
of discretionary cash and managerial risk considerations. Review of Quantitative Finance and
Accounting 14: 247-260.
23.
Choe, C., and X. Yin. 2009. Diversification discount, information rents, and internal capital
markets, Quarterly Review of Economics and Finance 49: 178-196.
24.
Comment, R., and G. Jarrell. 1995. Corporate focus and stock returns. Journal of Financial
Economics 37: 67-87.
25.
Contractor F. J., V. Kumar, and S. K. Kundu. 2007. Nature of the relationship between
international expansion and performance: The case of emerging market firms. Journal of World
Business 42: 401–417.
26.
Denis, D., D. Denis, and A. Sarin. 1997. Agency problems, equity ownership and corporate
diversification. Journal of Finance 52: 135-160.
27.
Desmond, W.N. 2007. A modern resource based approach to unrelated diversification. Journal
of Management Studies 44: 1481-1502.
28.
Dey, T., and R. Banerjee. 2011. Can corporate diversification promote firm value? A survey.
MPRA Paper, University Library of Munich, Germany.
29.
Diestre, L., and J. Santalo. 2013. The related diversification discount: When does related
64
diversification destroy value? Academy of Management Annual Meeting Proceedings 2013 (1):
12298-12298.
30.
Doukas, J. A. 2003. Foreign direct investment, diversification and firm performance. Journal of
International Business Studies 34: 153-172.
31.
Doukas, J., and O. Kan. 2006. Does global diversification destroy firm value? Journal of
International Business Studies 37: 352-371.
32.
Ecker, F., J. Francis, P. Olsson, and K. Schipper. 2009. A comparison of market-based and
accounting-based descriptions of business risk. Working Paper, Fuqua School of Business,
Duke University.
33.
Elsas, R., A. Hackethal, and M. Holzhauser. 2010. The anatomy of bank diversification. Journal
of Banking & Finance 34 (6): 1274-1287.
34.
Fama, E., and K. French. 1992. The Cross-Section of expected stock returns. The Journal of
Finance 47: 427–465.
35.
Farjoun, M. 1994. Beyond industry boundaries: Human expertise, diversification, and resource
related industry groups. Organization Science 5 (2): 185–199.
36.
Fatemi A. M. 1984. Shareholder benefits from corporate international diversification. Journal
of Finance 39: 1325-1344.
37.
Fayerweather, J. 1978. International business strategy and administration. Cambridge, MA:
Ballinger Publishing Co.
38.
Fleming, G., B. Oliver and S. Skourakis. 2003. The valuation discount of multi-segment firms
in Australia. Accounting and Finance, 43, 167–185.
39.
Ghemawat, P., and T. Khanna. 1998. The nature of diversified business groups: A research
design and two case studies. Journal of Industrial Economics 46 (1): 35-61.
40.
Goldberg, S., and F. Heflin. 1995. The association between the level of international operations
and risk. Journal of International Financial Management and Accounting 6 (1): 1-25.
41.
Gomes, J., and D. Livdan. 2004, Optimal diversification: Reconciling theory and evidence. The
Journal of Finance 59: 507–535.
42.
Gorecki, P. K. 1974. The measurement of enterprise diversification. Review of Economics and
Statistics 56: 399-401.
43.
Grant R. M., A. P. Jammine, and H. Thomas. 1988. Diversity, diversification, and profitability
among British manufacturing companies, 1972-1984. The Academy of Management Journal 31
(4): 771-801.
65
44.
Greene, W. 2008. Econometric Analysis. 6th ed. Upper Saddle River, NJ: Prentice Hall.
45.
Guillen, M. F. 2000. Business groups in emerging economies: A resource-based view. Academy
of Management Journal 43 (3): 362-380.
46.
Harris M., C. Kriebel, and A. Raviv. 1982. Asymmetric Information, Incentives and Intrafirm
Resource Allocation. Management Science 28 (6): 604-620.
47.
He, X. 2009. Corporate diversification and firm value: Evidence from post-1997 data.
International Review of Finance 9 (4): 359-385.
48.
Hermalin, B., and M. Katz. 2003. Corporate diversification and agency costs. Working Paper,
University of California at Berkeley.
49.
Hitt, M. A., R. E. Hoskisson, and H. Kim. 1997. International diversification: effects of
innovation and firm performance in product-diversified firms. Academy of Management Journal
40: 767-798.
50.
Hsiao, C. 2005. Why Panel Data? Singapore Economic Review 50 (2): 1-12.
51.
Jacquemin, A. P., and C. Berry. 1979. Entropy measure of diversification and corporate
growth. Journal of Industrial Economics 27 (4): 359-69.
52.
Jensen, M. C. 1986. Agency cost of free cash flow, corporate finance, and takeovers. American
Economic Review 76 (2): 323-329.
53.
Jensen, M., and K. Murphy. 1990. Performance pay and top-management incentives. The
Journal of Political Economy 98 (2): 225-264.
54.
Johnson, G., and K. Scholes. 2002. Exploring Corporate Strategy: Text and Cases. New York,
NY: Financial Times Prentice Hall.
55.
Kakani, R.K. 2000. Financial performance and diversification strategy of Indian business
groups. PhD diss, Indian Institute of Management, Calcutta.
56.
Kapopoulos, P., and S. Lazaretou. 2007. Corporate ownership structure and firm performance:
evidence from Greek firms. Corporate Governance 15 (2): 144–159.
57.
Khanna, T., and K. Palepu. 1997. Why focused strategies would be wrong for emerging markets.
Harvard Business Review 75: 41-54.
58.
Khanna, T., and K. Palepu. 2000. Is group affiliation profitable in emerging markets? An
analysis of diversified Indian business groups. The Journal of Finance 55 (2): 867-891.
59.
Klapper, L., and I. Love. 2002. Corporate governance, investor protection, and performance in
emerging markets. Working Paper, World Bank.
60.
Kochhar R., and M. Hitt. 1998. Linking corporate strategy to capital structure: diversification
66
strategy, type and source of financing. Strategic Management Journal, 19: 601-610.
61.
Kock C., and M. F. Guillen. 2001. Strategy and structure in developing countries: Business
groups as an evolutionary response to opportunities for unrelated diversification. Industrial and
Corporate Change 10 (1): 77-113.
62.
Kogut, B. 1989. Research notes and communications: A note on global strategies. Strategic
Management Journal 10: 383-389.
63.
Konchitchki, Y., Y. Luo, M. Ma, and F. Wu. 2016. Accounting-based downside risk, cost of
capital, and the macroeconomy. Review of Accounting Studies 21 (1): 1–36.
64.
Kwok, C., and D. Reeb. 2000. Internationalization and firm risk: An upstream-downstream
hypothesis. Journal of International Business Studies 31: 611-629.
65.
La Rocca, M., T. La Rocca, D. Gerace, and C. Smark. 2009. Effect of diversification on capital
structure. Accounting & Finance 49: 799-826.
66.
Lamont, O., and C. Polk. 2002. Does diversification destroy value? Evidence from the industry
shocks. Journal of Financial Economics 63 (1): 51-77.
67.
Lang L., and R. Stulz. 1994. Tobin’s q, corporate diversification, and firm performance. Journal
of Political Economy 102 (6): 1248–1280.
68.
Lessard, D. 1976. World, country, and industry relationships in equity returns: Implications for
risk reduction through international diversification. Financial Analysts Journal 32 (1): 32-38.
69.
Lev, B. 1974. On the association between operating leverage and risk. The Journal of Financial
and Quantitative Analysis 9 (4): 627-641.
70.
Li, X. 2004. Diversification and firm performance - the case of China. Chinese Business Review
3 (5): 1-8.
71.
Lins, K. V. and H. Servaes. 2002. Is corporate diversification beneficial in emerging markets?
Financial Management 31 (2): 5-31.
72.
Lins, K., and H. Servaes. 1999. International evidence on the value of corporate diversification,
The Journal of Finance 54 (6): 2215-2239.
73.
Low, P. Y., and K. H. Chen. 2004. Diversification and capital structure: Some international
evidence. Review of Quantitative Finance and Accounting 23 (1): 55-71.
74.
Lu, J., and P. Beamish. 2004. International diversification and firm performance: The S-curve
hypothesis. The Academy of Management Journal 47 (4): 598-609.
75.
Lubatkin, M., and S. Chatterjee. 1994. Extending modern portfolio theory into the domain of
corporate diversification: Does it apply? Academy of Management Journal 37: 109–136.
67
76.
Madura J., and L. Rose. 1989. Impact of international sales degree and diversity on corporate
risk. The International Trade Journal 3: 261-276.
77.
Maksimovic V., and G. M. Phillips. 2007. Conglomerate firms and internal capital markets. In
Handbook of Corporate Finance, ed. B. Espen Eckbo, 423-477. Amsterdam: North-Holland.
78.
Mansi, S. A., and D. M. Reeb. 2002. Corporate diversification: What gets discounted? The
Journal of Finance 57 (5): 2167-2183.
79.
Markides, X. C., and P. J. Williamson. 1994. Related diversification, core competences and
corporate performance. Strategic Management Journal 15: 149–165.
80.
Martin, J. D., and A. Sayrak. 2003. Corporate diversification and shareholder value: A survey
of recent literature, Journal of Corporate Finance 9 (1): 37-57.
81.
Meyer, M., P. Milgrom, and J. Roberts. 1992. Organizational prospects, influence costs and
ownership changes. Journal of Economics & Management Strategy, 1: 9–35.
82.
Miller, D. J. 2006. Technological diversity, related diversification and firm performance.
Strategic Management Journal, 27: 601–619.
83.
Montgomery, C. 1985. Product-market diversification and market power. Academy of
Management Journal 28 (4): 789-798
84.
Montgomery, C., and H. Singh. 1984. Diversification strategy and systematic risk. Strategic
Management Journal 5: 181-191.
85.
Morck, R., D. Stangeland, and B. Yeung. 1998. Inherited wealth, corporate control, and
economic growth: The Canadian disease? NBER Working Paper 6814, National Bureau of
Economics Research.
86.
Nayyar, P. R. 1992. On the measurement of corporate diversification strategy: Evidence from
large U.S. service firms. Strategic Management Journal 13: 219–235.
87.
Palepu, K. 1985. Diversification strategy, profit performance and the entropy measure. Strategic
Management Journal 6: 239–255.
88.
Palich, L. E., L. B. Cardinal, and C. C. Miller. 2000. Curvilinearity in the diversification–
performance linkage: An examination of over three decades. Strategic Management Journal 21:
155–174.
89.
Park, H. M. 2011. Practical guides to panel data modeling: A step-by-step analysis using Stata.
Working Paper, Graduate School of International Relations, International University of Japan.
90.
Piscitello, L. 2004. Corporate diversification, coherence and economic performance. Industrial
and Corporate Change 13: 757–787.
68
91.
Porter, M. 2008. On Competition. 2nd ed. Cambridge, MA: Harvard Business School Press.
92.
Rajan R., and L. Zingales. 1995. What do we know about capital structure? Some evidence from
international data. Journal of Finance 50: 1421-1460.
93.
Rajan, R., H. Servaes, and L. Zingales. 2000. The cost of diversity: The diversification discount
and inefficient investment. Journal of Finance 55: 35-80.
94.
Rieck, O., J. Cheah, A. Lau, and S. Lee. 2004. The Relationship between degree of
internationalization and firm performance in the telecommunications industry. Paper presented
at the ITS Europe 16th European regional conference proceedings, Porto, Portugal.
95.
Rumelt, R. P. 1982. Diversification strategy and profitability. Strategic Management Journal
3: 359–369.
96.
Schmid, M., and W. Ingo. 2009. Do financial conglomerates create or destroy economic value?
Journal of Financial Intermediation 18 (2): 193-216.
97.
Schoar, A. 2002. Effects of corporate diversification on productivity. The Journal of Finance
57: 2379–2403.
98.
Servaes, H. 1996. The value of diversification during the conglomerate merger wave. The
Journal of Finance 51 (4): 1201-1225.
99.
Shcherbakov D. 2012. Povyshaet li mezhdunarodnaya diversifikatsiya biznesa operatsionnuyu
effektivnost'? Opyt rossiyskikh kompaniy [Does corporate internationalization contribute to
companies’ operational efficiency? Evidence form Russian companies] Journal of Corporate
Finance, 4 (24): 58-64.
100. Shleifer, A., and R. W. Vishny. 1992. Liquidation values and debt capacity: A market
equilibrium approach. Journal of Finance 47 (4): 1343-1366.
101. Singh, N. P., R. Kumar, and R.P. Singh. 2006. Diversification of Indian agriculture:
Composition, determinants and trade implications. Agricultural Economics Research Review
19: 23-36.
102. Stein, J. C. 1997. Internal capital markets and the competition for corporate resources. Journal
of Finance 52 (1): 111-133.
103. Stein, J. C. 2003. Agency, information and corporate investment. In Handbook of the Economics
of Finance (vol. 1A), eds. G. M. Constantinides, M. Harris, and R. M. Stulz, 111-165.
Amsterdam: Elsevier NH.
104. Stijn, C., S. Djankov, J. P. Fan, and L. Lang. 2002. When does corporate diversification matter
to productivity and performance? Evidence from East Asia. Pacific-Basin Finance Journal 11
69
(3): 365-392.
105. Tallman, S., and J. Li. 1996. Effects of international diversity and product diversity on the
performance of multinational firms. Academy of Management Journal 39: 179-196.
106. Teece, D. J. 1980. Economies of scope and the scope of the enterprise. Journal of Economic
Behavior and Organization 1: 223–247.
107. Thomas, S. 2002. Firm diversification and asymmetric information: Evidence from analyst
forecasts and earnings announcements. Journal of Financial Economics 64: 373-396.
108. Toms S. 2011. Accounting, regulation and profitability: The case of PFI hospital refinancing.
Critical Perspectives on Accounting 22 (7): 668-681.
109. Villalonga, B. 2001. Does diversification cause the ‘diversification discount’? Working Paper,
Harvard University.
110. Villalonga, B. 2004. Diversification discount or premium? New evidence from the business
information tracking series. The Journal of Finance 59 (2): 479-506.
111. Wahla, K., S. Shah, and Z. Hussain. 2012. Impact of ownership structure on firm performance
evidence from non-financial listed companies at Karachi Stock Exchange. International
Research Journal of Finance and Economics 84: 6–13.
112. Wan, W. P., R. E. Hoskisson, J. C. Short, and D. W. Yiu. 2011. Resource-based theory and
corporate diversification: accomplishments and opportunities. Journal of Management 37:
1335-1368.
113. Wernerfelt, B. 1984. A resource-based view of the firm. Strategic Management Journal 5: 171–
180.
114. Williamson, O. 1979. Transaction-cost economics: the governance of contractual relations.
Journal of Law and Economics 22 (2): 233-261.
115. Wooldridge, J. 2012. Introductory Econometrics: A Modern Approach. 5th ed. South-Western
Cengage Learning.
116. Zhang, W., F. Yuan, and Y. Chen. 2002. An empirical study on the relationship between
diversification and economic performance of listed corporations of China. Journal of
Chongqing University 25 (11).
70
Отзывы:
Авторизуйтесь, чтобы оставить отзыв