The Performance Effects of Business Groups in Russia: Saul Estrin, Svetlana Poukliakova and Daniel Shapiro

Journal of Management Studies 46:3 May 2009
doi: 10.1111/j.1467-6486.2008.00820.x
The Performance Effects of Business Groups in Russia
Saul Estrin, Svetlana Poukliakova and Daniel Shapiro

London School of Economics; Simon Fraser University; Simon Fraser University
abstract This study analyses the impact of business group affiliation on firm performance
during a time when business groups are newly formed, when the economic and institutional
environment is changing, and when group survival is uncertain. Based primarily on a
transaction cost approach, we develop two hypotheses, concerning profitability and risk
sharing (redistribution) respectively. The positive profitability hypothesis proposes that
company affiliation with a business group directly and positively affects the profitability of
each affiliate. A positive direct effect emerges when each affiliate benefits from access to
group resources. The redistribution hypothesis considers the simultaneous possibility that
inter-affiliate transfers of resources through internal markets are designed to redistribute profits
among group members. We argue that variance-reducing redistribution from strong to weak
group members is linked to group survival in times of institutional change. Our empirical
approach focuses on testing these two linked hypotheses (and their alternatives) using a
relatively large, contemporary and time varying database of Russian firms. We also develop a
framework that distinguishes among the four possible empirical outcomes associated with the
hypotheses. Our results provide unambiguous support for the case where the impact of group
membership on profitability is positive and redistribution is variance-reducing. We term this
outcome Business Group Robustness, and contrast it with other possible empirical outcomes.
INTRODUCTION
Business groups (BGs) play a critical role in the governance landscape of many countries,
including most emerging economies (Khanna and Yafeh, 2007; Morck et al., 2005; Yiu
et al., 2005, 2007; Young et al., 2008). A number of theoretical perspectives have been
employed to explain their prevalence and understand their behaviour, including trans-
action cost analysis, relational analysis, political economy, and agency theory ( Yiu et al.,
2007). These theoretical approaches do not always result in common conclusions regard-
ing the nature of BGs. For example, the transaction cost approach focuses on the market
failures and institutional voids that create benefits to groups and their affiliates (Khanna
and Palepu, 1997; Peng et al., 2005). Affiliation to a business group enhances company
performance because membership allows firms to internalize market transactions and
Address for reprints: Daniel Shapiro, Simon Fraser University, Segal Graduate School of Business, 500
Granville Street, Vancouver, BC, V6C 1W6, Canada (dshapiro@sfu.ca).
© Blackwell Publishing Ltd 2009. Published by Blackwell Publishing, 9600 Garsington Road, Oxford, OX4 2DQ, UK
and 350 Main Street, Malden, MA 02148, USA.
394 S. Estrin et al.
create internal networks of value-creating relationships that minimize transaction costs
and provide access to valuable group resources and capabilities. In addition, internal
markets may permit firms to transfer financial resources so as to reduce group risk, and
ensure group survival. These advantages may be more pronounced in emerging markets,
where external markets are less efficient ( Wright et al., 2005; Young et al., 2008). Thus,
the internal markets associated with business groups reduce uncertainty and lower
transaction costs. Against this is the more negative agency theoretic view that business
groups in emerging markets suffer from multilayered agency and coordination problems
resulting not only in inefficiency but also in the expropriation of wealth from minority
shareholders (Morck et al., 2005).
Such contradictory theoretical predictions have not been resolved by the findings of
empirical work; in fact the empirical evidence on the impact of BGs on both affiliate
performance and risk sharing is ambiguous. For example, in a sample of 14 developing
economies, Khanna and Rivkin (2001) find the effect of group affiliation on firm prof-
itability to be positive in only a minority of cases. Similarly Khanna and Yafeh (2005) are
unable to identify evidence of risk sharing in most of their sample of 15 countries. Thus,
neither theory nor empirical evidence are as yet not able to resolve the issue of whether
business groups are ‘paragons or parasites’ (Khanna and Yafeh, 2007).
We extend and contribute to this literature in a number of ways. Based primarily on
a transaction cost approach, we develop two hypotheses, concerning profitability and risk
sharing (redistribution) respectively. The positive profitability hypothesis proposes that
company affiliation with a business group directly and positively affects the profitability
of each affiliate. A positive direct effect emerges when each affiliate benefits from access
to group resources. We also argue that the possibility of group failure provides strong
incentives for BGs to redistribute income to ensure group survival and so we explicitly
introduce group survival as a goal (Suhomlinova, 2006). The redistribution hypothesis
thus considers the simultaneous possibility that inter-affiliate transfers of resources
through internal markets are designed to redistribute profits among group members.
Redistribution from strong to weak affiliates reduces group risk and ensures group
survival. Although neither of these hypotheses is new to the literature, they are not
typically considered together. In developing them, we highlight the importance of defin-
ing more carefully the nature of the potential benefits (or costs) of internal markets. In
particular we distinguish benefits accruing to each affiliate from those arising from
redistribution which benefit the group as a whole. We further argue that group survival
is most relevant not only when market failures and institutional voids are present, but also
at the early stages of the BG life cycle. Thus, we argue that transaction cost theory is
particularly relevant for the understanding of how BGs form and evolve ( Yiu et al.,
2007).
We test these hypotheses on a very important emerging market, Russia. We argue that
the early Russian transition experience provides uniquely important insights into the
historic characteristics of business groups because it represents a case in which market
failures were important, institutional transition was in place, and business groups were
young and newly formed. Reforms to replace planning and state ownership with a
market economy beginning in 1991 quickly led by the mid-90s to the formation of
large, privately owned conglomerates owned by wealthy individuals (‘oligarchs’). The
© Blackwell Publishing Ltd 2009

Performance Effects of Business Groups in Russia 395
remarkable fact is that Russian business groups emerged from nothing to become the
dominant force in the economy over a 15 year period. According to Guriev and
Rachinsky (2005), the 22 largest private domestic owners in Russia controlled 42.4 per
cent of employment and 39.1 per cent of sales in 2003. This is paralleled by the
emergence of spectacular private wealth holdings; six individuals owned more than $5
billion of assets in 2003 according to Forbes.
Our empirical work focuses on the estimation of the two linked hypotheses. A unique
feature of this study is that we are simultaneously able to test both the profitability and
redistribution hypotheses. As noted above, although neither of these hypotheses is new to
the literature, they are not typically considered jointly. By estimating them together we
provide greater context for existing theory and a more nuanced approach to the study of
the strategy of business group owners in emerging and other economies (Wright et al.,
2005). In particular we are able to better distinguish among alternative theoretical
explanations for BG behaviour.
The two hypotheses are tested in a single equation framework using a specification
similar to Gedajlovic and Shapiro (2002) and Lincoln and Gerlach (2004), and are not
mutually exclusive. Thus, we provide a framework that considers four possible scenarios
associated with statistically significant outcomes. We characterize results that find both a
positive profitability effect and a redistribution effect from strong to weak firms as BG
Robustness. BG Robustness implies that business groups not only utilize internal markets
and external network resources to raise the profitability of affiliates, but that they use the
same internal markets to redistribute profits in order to reduce group risk by lowering the
variance of profitability within the group. We associate such behaviour with the need to
ensure group survival and stability and suggest that this was likely to have been the case
for Russia’s emerging BGs during our sample period (1998–2001) because of the failure
of markets and market institutions and the extreme uncertainty created during the
institutional transition from socialism to capitalism (Peng, 2003; Suhomlinova, 2006).
The framework that we develop to evaluate the empirical results also allows us to
distinguish among competing explanations regarding the nature of business groups. For
example, empirical results that indicate a negative effect of group affiliation on profit-
ability, but also show evidence of variance decreasing network redistribution, we term
BG Stability. This outcome pair implies that business groups subtract value from their
affiliates, while using internal markets to redistribute profits in a way that promotes group
stability. BG Stability may describe a situation where BGs do not create value for their
affiliates (perhaps because profit maximization is not a group goal), while at the same
time redistributing profits so as minimize group risk, as some suggest has been the case
for Japanese keiretsu (Gedajlovic and Shapiro, 2002). In addition, we consider two other
possible outcome combinations, referred to as BG Effectiveness and BG Entrenchment. We
argue that these cases, described more fully below, are consistent with a conceptualiza-
tion of business groups that either mimic markets by creating value and transferring
resources in a Pareto enhancing manner or simply destroy value, possibly through
looting.
Our study also contributes to the literature on Russian business groups. Despite the
attention paid to the role of corporate governance in Russia in recent years (Black, 2001;
Estrin, 2002), little has been done to study its business groups in a systematic manner,

perhaps because they have emerged so rapidly. For example, Khanna and Rivkin (2001)
do not include Russia in their sample, and the important recent survey by Khanna and
Yafeh (2007) does not include primary data on Russian business groups. Yiu et al. (2007)
refer to the Russian Financial and Industrial Groups of the mid-1990s, but does not
cover their evolution into business groups dominated by oligarchic owners. Such litera-
ture as there is has also tended to be highly critical of Russian business groups, often
linking them with asset stripping and looting at the time of privatization (Godoy and
Stiglitz, 2007; Goldman, 2003; Hoffman, 2003; Johnson et al., 2000). However, the few
available firm level studies suggest that they have positive effects on profitability (Black,
2001; Perotti and Gelfer, 2001) or productivity (Guriev and Rachinsky, 2005). Impor-
tantly, these empirical studies are based on relatively small samples that do not include
private firms. We contribute to this literature by employing a new, larger and more
comprehensive dataset that is likely to be more representative of the large and regionally
diverse Russian economy.
We use a large panel of nearly 1,000 firms in Russia over the period 1998–2001. The
dataset is much larger than has been previously employed to address these issues in
Russia. It is also chosen for the years immediately following the privatization process,
when the determination of which enterprises were and were not in business groups
reflected more the political processes of privatization and business group formation than
the profit maximizing choices of the business group owners. This reduces an important
source of selection bias in the sample. The dataset also contains unusually rich informa-
tion about performance and ownership for both public and private firms. Khanna and
Yafeh (2007) note that many previous studies restrict their samples to public firms, even
though many group affiliates are in fact private. Thus our sample reduces another
potential source of selection bias arising from samples that include only public firms.
Previous studies relating some aspect of corporate ownership structure to performance
have correctly stressed the issue of reverse causality (Black et al., 2004; Chang, 2003;
Gorodnichenko and Grygorenko, 2008). Of particular concern in our framework is the
possibility that a positive profitability effect of business group affiliation may reflect a
process whereby business groups tend to choose better performing affiliates, rather than
improving the performance of affiliates after they join. One important reason for focus-
ing on Russia in the period under analysis is that reverse causality may be a less serious
problem than in many other contexts because of the manner in which business groups
were formed (see Desai, 2006; Djankov and Murrell, 2002; Goldman, 2003). When
Russian business groups suddenly emerged, in the early to mid-1990s, the new owners
were not for the most part already wealthy individuals using their significant holdings to
purchase new firms and build up their groups. Rather they initially used the privatization
process to convert their effective control of Soviet conglomerates or Ministries into
ownership via a process which was politically determined. To quote The Economist (2008,
386, 8571, p. 76), ‘To date the Kremlin has treated oligarchs like rentiers rather than
owners, and no significant sale has been possible without its blessing, even when no
foreigners are involved’.
Nevertheless, endogeneity may still be an important issue in all such situations and we
control for it by using a two-stage estimation procedure in which business group mem-
bership is determined jointly with profitability. In so doing, we utilize an instrument,

public listing, that we suggest is also based, in this early transition period, on the highly
politicized Russian privatization experience.
The study proceeds as follows. In the next section, we provide a more complete
description of Russian Business Groups (RBGs). We then develop the profitability and
redistribution hypotheses, and introduce a framework to evaluate the various possible
empirical outcomes. The following section presents the empirical model used to jointly
test the hypotheses. This is followed by a description of our data, and a discussion of
our estimation procedures. We then present the results, followed by discussion and
conclusions.
RUSSIAN BUSINESS GROUPS (RBGs)

Business group formation in Russia began from the start of the transition in 1991 and
RBGs (then known as financial-industrial groups) were beginning to emerge by 1994
( Johnson, 1997), following the highly controversial privatization process (Boone and
Rodionov, 2002; Goldman, 2003). Thus, in contrast to BGs in other countries that
emerged slowly over time, the RBGs emerged during a process of institutional change.
For example, Indian BGs existed before the reforms of the 1990s (Kedia et al., 2006).
Unlike many emerging markets, family ownership is not yet widespread in Russia;
RBGs are usually privately owned by a single person with a controlling interest (Almeida
and Wolfenzon, 2006; Hoffman, 2003). In addition, ownership rights are typically
exercised by means of direct control of voting shares, so that pyramids are rare (Guriev
and Rachinsky, 2005). Thus, the kind of principal–principal problems between majority
and minority shareholders that might be exacerbated by business groups ( Young et al.,
2008) may be less pronounced in Russia. The high degree of individual control suggests
that RBGs will pursue goals that reflect the self-interest of the controlling oligarch. Thus,
traditional agency costs arising from diffuse ownership are unlikely to be a source of
inefficiency in Russian business groups. Partly owing to the method of privatization,
Russian business groups typically own, control, or are affiliated with banks or investment
houses, and their affiliates are often, but not exclusively, publicly traded (Filatotchev
et al., 2001). Thus, Russian business groups are firms linked by common ownership
concentrated in the hands of individual oligarchs that share commercial and financial
relationships. They also share social and political ties. Russia is a country with a strongly
developed cultural and educational system that has promoted strong social ties and
shared values among business and other groups. Moreover, Russia inherited from the
Soviet Union the remnants of networks based on reciprocal favours, known as blat
(Ledeneva, 1998), which has evolved into a quasi-market form, now referred to as svyazi
(Batjargal, 2007; Puffer and McCarthy, 2007) from which RBGs may benefit.
The rapid formation of RBGs highlights the unusual institutional context of Russia
during the 1990s and helps to address endogeneity issues in our empirical strategy.
Gorodnichenko and Grygorenko (2008) summarize the data and the Russian press
comments concerning the formation of RBGs in four stylized facts. First, the firms
chosen as business group affiliates were large but not necessarily productive. Oligarchs
primarily sought enterprises that could generate large cash flows but they showed no
regard for enterprise efficiency or profitability (see also Hoffman, 2003).[1] Second,

business groups were constructed to develop vertically integrated production chains.
Third, business groups owned and continued to own firms that consistently made losses.
Finally, that RBGs often invested in their affiliated enterprises with the objective of
improving the affiliates’ productivity, a point also made by Perotti and Gelfer (2001) and
Shleifer and Treisman (2005).
In summary, Gorodnichenko and Grygorenko (2008) argue that:
At the time of privatization, oligarchs were typically outsiders to the firms they
bought . . . Therefore, the oligarchs did not possess any material non-public informa-
tion about the enterprises and it was not required in fact. There was no need to value
the companies being bought as the prospective buyer faced no risk to overpay: state
property was sold for a nominal hugely undervalued price and a successfully closed
deal would result in a non-negative return. With the cost of a medium and a large
enterprise being immaterially different, the well-connected would-be oligarch would
more likely target the latter for acquisition as it promises higher gain in absolute
terms. . . . the typical strategy was to pick up all the biggest firms around. (p. 27)
The random nature of the process is further explained by Hoffman (2003):
Khodorkovsky was shooting in the dark. He could not figure out which factories were
potentially lucrative so he bought many . . . then hired Anderson Consulting to survey
the crazy quilt industry he had assembled and the management consultants told him
he had gathered up the equivalent of a South Korean conglomerate. (p. 205)
All of this suggests that RBG affiliation was not determined by the economic attrac-
tiveness of the affiliate and that oligarchs used their resources to enhance affiliate
performance.
HYPOTHESES
In this section we develop two (pairs of) hypotheses about the relative performance of
firms that are affiliates of BGs in Russia, which we refer to as the profitability and the
redistribution hypotheses respectively. The former pair addresses the question of
whether group membership brings profitability advantages or disadvantages to affiliates,
other things being equal. The latter addresses the question of whether RBGs use internal
markets to redistribute profits, and the consequences for profit persistence.
The Profitability Hypotheses

The Profitability Hypotheses start from the view that emerging markets are character-
ized by a governance infrastructure that does not facilitate the development of efficient
external markets. In particular, limited property rights, lack of an independent judiciary,
inefficient and corrupt governments, and the absence of a transparent regulatory frame-
work all limit the emergence of efficient markets and supporting institutions (Globerman
and Shapiro, 2003; Young et al., 2008). In these circumstances business groups may be

an efficient and rational institutional arrangement in which internal markets and infor-
mal institutions replace external markets and formal institutions as a means of allocating
resources (Kock and Guillen, 2005; Leff, 1976, 1978). Khanna and Rivkin (2001), for
example, argue that internal markets are relatively more efficient in the face of the
widespread market failures prevalent in emerging market economies, including Russia
(EBRD, 2007). Thus internal markets minimize transaction costs and allow resources to
be allocated efficiently within the group in a manner that enhances the performance of
each affiliate. Effectively, group resources spill over to affiliates via internal markets, and
the spillovers may be international in scope (Garg and Delios, 2007). In particular,
relatively efficient internal capital markets replace relatively inefficient external capital
markets, and internal governance replaces poor external contract enforcement, thus
limiting the potential for hold-up.
Blanchard and Kremer (1997) argue that the recession in Russia immediately follow-
ing reforms was caused by hold-up problems – situations where firms refuse to buy or sell
because they are worried that, after they have committed resources irreversibly, they will
be forced back into negotiations in which their share of the surplus will be reduced.
Hold-up was made more likely in Russia because the Soviet industrial structure was
highly vertically integrated as a consequence of central planning, so that each firm only
had one or a few suppliers. As we have noted, Russian business groups often emerged
from the traditional Soviet era Ministries and were often highly vertically integrated
structures. Their common ownership by a single individual may have permitted them to
avoid endemic hold up problems that characterized Russia at the time.
Efficiency gains are not the only way that BG membership may enhance the financial
performance of affiliates. Market failures and the related absence of associated market
institutions also increase the benefits from access to social and political networks (Stra-
chan, 1976). Fisman (2001) provides empirical evidence that political connections are
valuable in emerging markets. In Russia, the weakness of market allocation mechanisms
suggest that such relationships could be used to ensure access to scarce resources,
particularly financial resources, and to publicly funded infrastructure (Ledeneva, 1998).
Batjargal (2003) emphasizes the importance of social capital in Russia, and finds that
relational embeddedness (which would characterize business groups) has a direct positive
effect on firm performance. This is consistent with Ledeneva (1998) who suggests that the
benefits of blat – Russian networking – have spilled over into the transition period. Thus
network connections (svyazi) can provide firms with access to important political
resources (Batjargal, 2007; Puffer and McCarthy, 2007). This provides a second reason
why the well connected individuals controlling Russian business groups might be able to
ensure superior economic performance for their affiliates.
Finally, it has been suggested that the highly concentrated nature of ownership in
Russian business groups and their affiliates noted above may alleviate the classic
principal–agent problem created by the separation of ownership from control, thus
contributing to a profitability advantage for business groups (Boone and Rodionov,
2002). Although ownership concentration may resolve potential agency problems in
these markets, it may also allow entrenched ownership to exercise political power to
benefit their firms (Morck et al., 2005). The latter interpretation is in a sense the dark side
of the view that relational embeddedness contributes to group performance, and suggests

that performance benefits from business group affiliation may not always be efficient
from a social perspective.
However, there are also strong arguments that BG membership might reduce the
performance of affiliated firms. The finance literature contains an important ‘conglom-
erate discount’ strand which argues that because the agency problems have a number of
layers in such firms, performance will be worse than in their stand-alone competitors
(Scharfstein and Stein, 2000). To the extent that external Russian capital markets were
efficient, one might observe that RBG affiliation was detrimental to firm performance. In
the Russian transition context, where capital market efficiency was unlikely, it is also
argued that the business groups, which emerged in the chaos of liberalization and mass
privatization, were in fact formed merely to facilitate the looting of former state owned
assets by the new owners, without any concern for the performance of the affiliates
(Stiglitz, 2002). Similarly, Guriev and Rachinsky (2005) argue that if the new business
owners have gained their control via political favour rather than expertise and skill, they
might be expected to seek to maintain and extend their position by political manoeuvring
and buying continued goodwill, rather than efforts to improve company performance. In
a similar vein, Khanna and Rivkin (2001) suggest that group managers may have weak
incentives to be efficient because they are ‘secure in the embrace of the group’ (p. 51).
Thus, there are two views of BGs in an emerging market such as Russia. One stresses
the high transactions costs associated with external markets and the corresponding
benefits of internal markets and vertical integration, the political economy benefits
associated with business group networks, and the advantages associated with concen-
trated ownership. The other highlights the potential inefficiency of BG affiliates relative
to unaffiliated firms. We summarize the alternative views as follows:
Hypothesis 1a: Affiliation with an RBG has a positive influence on the profitability of
affiliated firms.
Hypothesis 1b: Affiliation with an RBG has a negative influence on the profitability of
affiliated firms.
The Redistribution Hypotheses

Internal markets permit business groups to allocate resources among affiliated firms.
Thus, ‘headquarters’ can redistribute financial resources away from some affiliates and
redirect them to others. We approach the question of redistribution through the lens of
profit persistence.
The literature on the persistence of profitability, beginning with Mueller (1986, 1990),
measures the extent to which competition forces firm profits to converge on a competitive
norm. In this literature, greater measured persistence (at least for successful firms) is
associated with some form of monopoly power or competitive advantage. However, it has
been argued that the governance structure of firms and specifically affiliation with a
business group may affect the degree of persistence (Chacar and Vissa, 2005; Gedajlovic
and Shapiro, 2002; Lincoln et al., 1996). Firms affiliated with business groups may exhibit
less profit persistence, other things being equal, because poorly performing group members

can access valuable group resources, including capital, managerial talent or even prefer-
ential access to government favours at the expense of better performing members. Lower
profit persistence across the group may therefore be associated with a variance decreasing
redistribution effect that directs resources from stronger to weaker firms.
Such a redistribution effect arises if BG objectives include the stability or survival of
the group as a whole, which can be achieved by the sharing of costs and benefits (Chang
and Hong, 2000; Suhomlinova, 2006). Thus, a potential benefit for BG affiliates comes
from the ability of group members to reallocate resources in such a way as to support
weaker members. This approach is associated with the view that BGs arise to exploit the
advantages associated with internal markets. Thus, Khanna and Rivkin (2001) suggest
that business group members exchange resources through internal markets in such a
way that ‘capital may flow to members in distress or to members whose opportunities
outweigh their ability to generate capital themselves’ (p. 48). A similar point is made by
Gedajlovic and Shapiro (2002), who argue that within the group, short-run profits of
some firms may be sacrificed in order to allow weaker but potentially profitable firms to
survive through economic slowdowns and external shocks. Khanna and Yafeh (2005)
also suggest that redistribution is designed to share risks and occurs though shared
resources, dividends, and intra-group transfers through flexible loans and receivables.
Such mutual support also sustains risk-reducing diversification for the group.
The variance reducing redistribution argument is most closely associated with Japan,
where it is suggested that Japanese keiretsu do not necessarily maximize profits per se, but
rather focus on the survival of the group through the redistribution of profits (Lincoln
and Gerlach, 2004; Lincoln et al., 1996). There is considerable evidence in support of the
variance reducing redistribution hypothesis in Japan, where studies suggest that Japanese
business groups have lower profit rates than non-affiliated companies but they are more
stable over time (Caves and Uekusa, 1976; Hoshi and Kashyap, 2001; Lincoln et al.,
1996). More recently, Gedajlovic and Shapiro (2002) find a pronounced profit redistri-
bution effect characterized by the transferring of financial resources from more to less
profitable firms, while Lincoln and Gerlach (2004) find that group intervention enhances
the performance of weaker partners. It might be argued that Japan represents a special
case, and that redistribution is observed in that country because of the degree of
cross-ownership among firms.
We suggest, however, that variance reducing redistribution may occur in other coun-
tries, including Russia, although not necessarily for the same reasons. Drawing on an
evolutionary psychology perspective, Tooby and Cosmides (1996) consider the circum-
stances under which investment funds might be used to support those in the greatest
need. They argue that individuals in societies where there are few other individuals who
are concerned about their welfare are prone to external risks. This leads to the need to
select and identify partners who can minimize threats to survival, such as partners,
family, or other groups who have a mutual interest in supporting each other against
external threats. This is a more general argument for the variance reducing redistribu-
tion hypothesis, which essentially says that in highly volatile and uncertain circum-
stances, it pays to form groups and provide mutual support. Moreover, they suggest that
if you are a valuable resource to someone else, then that person will have a strong interest
in your survival. Thus, group survival may become a goal when conditions are uncertain,

and when partners provide protection against these uncertainties, even in the absence of
ownership links. A similar argument is provided by Boisot and Child (1996), who suggest
that clans (or groups) are more likely to arise when weak institutions limit the codification
and diffusion of information that will reduce transaction costs. Groups rely on their
affiliates to coordinate activities when markets fail and conditions are uncertain. Thus,
group survival is valuable.
Internal redistribution from stronger to weaker group members is therefore likely to be
particularly important when the long run goal of profit maximization is threatened by
short run external shocks such as those that occur in a transitional period. When markets
are immature and imperfect, protecting customers, suppliers and other group members
may be essential to long run survival. Thus, the variance reducing redistribution argu-
ment may be seen as a particular extension of the market failures approach to business
groups. In periods of institutional transition, formal rules change, creating considerable
uncertainty and increased costs as new institutions emerge (Peng, 2003). Under these
circumstances, networks or business groups based on interpersonal ties and trust arise in
order to overcome institutional voids and market failures. Maintaining this system of
relationships can under these conditions become a goal unto itself in order to ensure the
security of supply and distribution channels. As we have noted above, the transition
period in Russia was characterized by both market failures and by considerable uncer-
tainty. Under these circumstances we would argue that survival of the RBG through the
redistribution of network resources is important. Thus:
Hypothesis 2a: Redistribution within the RBG will result in the transfer of resources
from stronger to weaker firms. Thus, profits of RBG affiliates will be less persistent
than those of non-affiliates: firms with low prior profits benefit from RBG affiliation,
but firms with high prior profits have their profitability reduced.
Hypothesis 2a posits that internal redistribution results in lower profit persistence,

suggesting that group survival is an explicit goal. The alternative is that internal markets
are used to mimic external markets, and weaker group members are not subsidized.
Rather, resources flow from lower- to higher-valued affiliates in order to improve
performance (Shleifer and Treisman, 2005). To the extent that such flows create com-
petitive advantage for successful firms, one expects them to be persistently profitable. The
implication is that internal redistribution from weaker to stronger firms is associated with
relatively higher degrees of profit persistence. Thus, the alternative to Hypothesis 2a is:
Hypothesis 2b: Redistribution within the RBG will result in the transfer of resources
from weak to stronger firms. Thus, profits of RBG affiliates will be more persistent
than those of non-affiliates: firms with high prior profits benefit from RBG affiliation,
but firms with low prior profits have their profitability reduced.
Discussion of Hypotheses
Hypotheses 1 and 2 are not mutually exclusive, though previous studies have looked
mostly at either the direct performance effect, or the redistribution (risk sharing) effect,

Profitability (H1)
H1a Supported No H1b Supported
(positive RBG profit Support (negative RBG profit
effect) for effect)
Either
H1a or
H1b
BG Robustness BG Stability
BGs use internal markets BG membership lowers

H2a Supported and external networks to affiliate performance on
enhance affiliate average while at the same
profitability while at the redistributing resources to
(redistribution with same time supporting ensure group survival
lower profit group survival through (perhaps for political
persistence) redistribution from strong power). Consistent with
to weak members. Most group survival and stability
likely to occur in early as a dominant goal
stages of institutional achieved at the expense of
transition and BG profitability as in Japan,
formation. Outcomes may and with looting as profits
be privately beneficial, but are directed to money
not necessarily socially losing ‘shell’ firms
beneficial if associated without any compensating
Redistribution with market or political benefit to stronger firms.
(H2) power.
No Support for
Either H2a or
H2b
BG Effectiveness BG Entrenchment
H2b Supported BGs raise profitability of BGs are inefficient and

affiliates while at the same bureaucratic, and earn
(redistribution with time apparently persistently low returns
greater profit redistributing resources sustained by political
persistence) from weak to strong firms. influence. Might also
Higher persistence is suggest looting by the
consistent with parent which drives down
competitive advantage. profits in all affiliates and
Resources are internally persists through political
allocated from worse to influence.
better uses.
Figure 1. Summary of Hypotheses 1 and 2
but rarely both. We argue, however, that both matter, because together they can help
distinguish among alternative views of the BG. Thus, by examining the profitability and
redistribution hypotheses simultaneously, we gain a degree of interpretive freedom. Our
argument is illustrated in Figure 1, in which we focus on four possible outcomes of our
empirical test that involve statistically significant outcomes. We note that it is also
possible that none of our hypotheses could receive support.
Consider first the possibility that both Hypotheses 1a and 2a are supported by the
data. This indicates a positive profitability effect and a negative redistribution effect
(which implies that resources move from stronger to weaker firms). We term this outcome

BG Robustness because it suggests that business groups not only use their network of
internal markets and external relationships to the benefit of group affiliates (the profit-
ability effect), but they also use internal markets to redistribute profits to weaker firms in
order to ensure group survival. Thus, support for both Hypotheses 1a and 2a is consistent
with the transaction cost view that business groups in Russia act as substitutes for failed
external markets and missing market-supporting institutions. A positive profitability
effect combined with redistribution is consistent with the view that business groups
minimize transaction costs through internal markets so as to raise the profitability of
affiliates by facilitating the internal transfer of valuable resources, but also use the same
internal markets to redistribute profits so as to lower group risk. However, we cannot rule
out the possibility that a positive profitability effect results not from internal market
allocation, but rather from broader network benefits that may include access to politi-
cally created rents. Thus, we refer to this outcome as BG Robustness because the outcomes
are profitable for the network of firms in the BG, and contribute to its survival, thus
suggesting strength and durability. We also note that if market and political power are
the source of higher profitability, then group stability may be seen as a means to maintain
group power, and this may not be socially beneficial.
There are several other possible combinations of outcomes, with alternative implica-
tions. Here we focus on three possibilities that involve support for one or more hypo-
theses. We refer to an outcome that combines a positive profitability effect with a
redistribution effect that promotes persistence as BG Effectiveness. We use the term ‘effec-
tive’ because in this case BGs might mimic efficient markets by reallocating resources to
better uses (redistribution from lower to higher profit firms). If BGs display persistence of
profits and are more profitable, one possibility is therefore that they are operating as
efficient internal markets allocating resources to the more profitable uses, thereby
creating competitive advantage that persists. However, higher profitability may result
instead from market or political power, so, though privately efficient, this case is not
necessarily social welfare enhancing.
Not all outcomes are even privately efficient, at least in terms of profitability. We refer
to one such outcome as BG Stability. In this case, BGs reduce affiliate performance on
average while redistributing resources to ensure group survival. Unlike the case of BG
Robustness where redistribution acts to support an organization which on balance is
efficiency-enhancing (or at least generates positive profits), in this case redistribution
props up an organization which is efficiency reducing. This case we suggest is consistent
with some evidence from Japan suggesting that group insurance has become a goal that
dominates profit maximizing behaviour (Lincoln et al., 1996). However, it could also be
consistent with the possibility of looting, whereby profitable affiliates are stripped of
value-creating assets (Campos and Giovannoni, 2005; Johnson et al., 2000). If looting is
widespread, business group affiliation may be negatively related to profitability. At the
same time profit redistribution towards weak firms may exist but may be motivated by
the expropriation of wealth via the transfer of profits from subsidiary firms (Bertrand
et al., 2002; Chang, 2003). Although this is normally associated with ‘tunnelling’ through
pyramids and cross ownership (Bertrand et al., 2002), it could also occur via the transfer
of wealth to apparently unprofitable affiliates that act as conduits for wealth expropria-
tion. Thus, a result that finds both a negative profitability effect and a redistribution effect

towards weaker firms suggests that business groups subtract value from their affiliates by
using internal markets to redistribute profits in a way that likely benefits only some
members of the group.
Finally, we refer to BG Entrenchment as the case where a negative profitability effect is
combined with a redistribution effect that promotes the persistence of low returns. This
case, we suggest, is consistent with the view that BGs in emerging markets are bureau-
cratic, cumbersome firms whose existence with persistently low returns is facilitated by
political connections and influence. BGs are in this case probably not best understood as
economic organizations interested in profit maximization, but rather as political orga-
nizations seeking power. Persistence (of low profits) in this case is the result of access to
political networks and may not reflect redistribution. As noted above, Gorodnichenko
and Grygorenko (2008) suggest that BGs continue to own and support money-losing
affiliates, and Chacar and Vissa (2005) find that within Indian business groups, there is
high profit persistence among unsuccessful affiliates. However, this case may also be
consistent with looting, whereby affiliate assets profits are removed, and this persists
because of political influence.
It is possible that over time one observes no statistical effects of business group
affiliation. That is, business groups begin to focus and limit their activities, become no
more or less efficient than others firms, and rely more heavily on external markets to
obtain financial and other resources. Recent evidence suggests that this may have been
the case in India (Kedia et al., 2006). This would suggest a movement towards the centre
of Figure 1.
Specification, Estimation and Data

In this section, we propose an empirical framework to test Hypotheses 1 and 2, before
outlining the dataset and discussing appropriate estimation methods. Following a
number of previous studies (Gedajlovic and Shapiro, 2002; Lincoln and Gerlach, 2004;
Lincoln et al., 1996), we test the two hypotheses within the framework of the following
single estimating equation:
ROA jt = β0 + β1′ xjt +λ RBG jt +γ ROA jt − k + δ RBG jt ∗ROA jt − k + ε jt (1)
ε jt = α jt +η jt
where subscripts j and t represent the firm and year, respectively; subscript k is number
of time periods; ROA is return on assets (defined below); b0 is the model intercept; b1 is
a vector of estimated coefficient for control variables; xjt is a vector of independent control
variables (also defined below); RBG indicates affiliation with a business group; ejt is a
disturbance term; ajt is a firm and year individual effect; and hjt is a random error. One
advantage of this specification is that the two hypotheses can be tested in the context of
a single equation which allows us to clearly distinguish between the profitability and
redistribution effects. In addition, the model is parsimonious in explanatory variables
because the lagged ROA term captures a great deal of firm-specific information.

The direct effect of RBG affiliation on profitability is given by l. A positive significant
coefficient supports Hypothesis 1a, and a negative one supports Hypothesis 1b. The
degree to which RBG affiliation creates a redistribution benefit is estimated by d, the
coefficient on the interactive term between RBG affiliation and past profitability (Lincoln
and Gerlach, 2004). The coefficient on the lagged profitability term reflects the temporal
variation in profitability or the degree of persistence. The hypothesis that network
redistribution is facilitated by RBG affiliation (Hypothesis 2a) is supported if d is negative,
indicating that RBG affiliates have lower temporal variation (persistence) of profitability.
A positive significant coefficient is consistent with market redistribution as suggested in
Hypothesis 2b.
Equation (1) treats the RBG term as exogenous. However, there may be reverse
causality, with business groups choosing new affiliates with higher performance, rather
than generating greater profitability through their own actions (Black et al., 2004;
Chang, 2003). Although we have suggested that that this problem may not be acute in
our period in the Russian context, we nevertheless control for endogeneity by using an
instrumental variable approach and two-stage least squares (2SLS) estimation.
Thus, we estimate the following two-equation model:
ROA1jt = β10 + β11′ x1jt + λ1RBG jt + γ 1ROA jt − k + δ1RBG jt ∗ROA jt − k + ε1jt (2)
RBG2 jt = β20 + β21′ x 2 jt + β22 ROA jt + β23 ROA jt − k + ε 2 jt .
This involves estimating a version of equation (1) jointly with a second equation in
which the determinants of RBG affiliation include current and lagged firm profitability
as well as a vector of control variables (x2). Equations (2) can be estimated using 2SLS
and instrumental variables approaches.
Data
Our balanced panel of Russian firms, with RBG membership identified, is constructed
from two sources. First, we surveyed published lists of Russian oligarchs and business
groups in order to create a list of RBGs and, where possible, their affiliates. Second, we
used the 2002 Amadeus database to further define affiliates and to construct a panel of
firms containing important financial and other firm-specific information. The Amadeus
database covers all Russian firms with more than 50 employees and contains a wide
variety of enterprise specific information including ownership structure. The data indi-
cated the number of recorded subsidiaries and the names of major registered sharehold-
ers which was matched with the initial list of RBGs and their affiliates to create a
balanced panel of 977 companies for three years between 1998 and 2001. Of these, 274
firms were defined as RBG affiliates. The sample includes both public and private firms
(thus reducing the selection bias identified by Khanna and Yafeh, 2007), and is large
compared to other studies for Russia (e.g. Black, 2001; Perotti and Gelfer, 2001), in part
because it is not restricted to firms whose shares trade on the public stock exchanges.

Financial service firms were dropped from the sample since their financial ratios are not
comparable with non-financial institutions.
We used a variety of sources to construct a list of RBGs, from which to identify
the affiliates. These included the industrial–financial groups’ registry book, and other
data published by the Ministry of Statistics of the Russian Federation (e.g.
); business journals such as Vedomosti, Expert, Finance, and
Kommersant-Vlast, Forbes, and Ross Business Consulting (RBC) Daily; and listings available
from institutions that specialize in corporate governance issues in Russia such as the
Russian Institute of Directors, Goskomstat and Troijka Dialog. Information available
from companies’ official publications and websites was also used. The balanced panel
was constructed for three years: 1998, 2000 and 2001; 1999 was excluded because large
numbers of firms were temporarily absent, probably because this was the financial year
that followed the 1997 Russian financial crisis. Financial performance data in Amadeus
are based on information provided by leading ‘auditors’, usually in line with generally
accepted Western accounting principles. Financial performance data were converted
into US dollars using year-specific exchange rates.
Correlation analysis (results not presented) show that our sample characteristics are
similar to what is generally understood about RBG affiliates in Russia (Guriev and
Rachinsky, 2005): they are likely to be privately owned with highly concentrated own-
ership in the hands of few individuals. There was a positive and statistically significant
correlation between RBG affiliation and whether the firm was publicly traded. Thus, 50
per cent of publicly traded firms in the sample were RBG affiliated, but only 26 per cent
of non-affiliated firms were publicly traded. Nevertheless, it is not the case that all
publicly traded firms in our sample are RBGs. We return to this issue below.
Estimation Methods
Hypotheses 1 and 2 are jointly tested by estimating equations (2) or, for tests of specifi-
cation without considering endogeneity, equation (1). Because our sample includes
non-traded companies, we measure firm performance by ROA, calculated as income net
of taxes divided by total assets. ROA has been the most widely used performance
measure in related studies of business groups performance (e.g. Caves and Uekusa, 1976;
Khanna and Palepu, 2000a; Lincoln and Gerlach, 2004).
Control variables for the performance equation were chosen based on the previous
literature (Gedajlovic and Shapiro, 2002). The following variables constitute our control
(exogenous) variables, previously subsumed under the X-vectors in equations (1) and (2).
Firm size, measured as the log of total assets, is included to account for the potential
economies of scale and scope accruing to large firms. If present, these would produce a
positive relationship between firm size and profitability. Firm growth, measured as year
over year sales growth, is used as a control for demand conditions and product-cycle
effects. Firms in relatively fast-growing markets are expected to experience above
average profitability. Financial leverage, measured as the ratio of debt to equity, is
included as a control variable in the regression models because a firm’s capital structure
may influence its performance through a variety of mechanisms, including debt

Table I. Descriptive statistics and correlation matrix
Variable Mean St. Dev. 1 2 3 4 5 6
1. ROA 0.07 0.18

2. RBG 0.28 0.45 0.02
3. Lag ROA 0.10 0.15 0.32* 0.01
4. Log assets 8.97 1.83 -0.01 0.27* -0.02
5. Sales growth 0.43 0.72 0.12* -0.01 -0.18* -0.04
6. Leverage 0.56 0.28 0.22* -0.07* 0.24* 0.01 -0.06
7. Publicly traded 0.10 0.27 0.02* 0.15* -0.01 0.50* -0.03 0.04
Notes:
* Significant at 5% level.
Data are for three years, number of firms = 977.
servicing, investment decision and the discretion afforded oligarchs (Lincoln et al., 1996).
Alternative measures of leverage such as the ratio of debt to total assets do not change the
results.
Finally, we included indicator variables to control for state ownership (equals one if
state ownership exceeds 20 per cent), industry specific effects (30 industry level dummy
variables based on two-digit NACE codes), location (six regional indicator variables), and
year. The industry and region dummy variables together control for the potential of firms
to exercise market power, and recognize that market power varies by both industry and
location. The industry dummies also control for other industry effects. Alternative
thresholds for defining state ownership do not affect the results. Equation (1) also includes
the lagged value of ROA (Lag ROA), the RBG dummy variable defined above (RBG),
and the interaction between Lag ROA and RBG.
Our primary sample uses the three years of available data, but we also estimate the
relevant model for the two-year period, 2000–01 as a robustness test. Descriptive statis-
tics are provided in Table I. In addition to the variables discussed above, this table
includes the publicly traded dummy variable that we use later in our IV estimation of
equations 2.
As indicated in Table I, the correlation among the independent variables is in general
not very high, although most coefficients are significant. The highest correlation is found
between RBG and the log of assets, an issue addressed below. Other than that, multi-
collinearity is unlikely to be an issue for the estimation of equations (1) and (2).
Dynamic panel data models such as ours pose serious estimation challenges since the
lagged dependent variable will be correlated with the error term (Baltagi, 1995, pp.
125–6). Familiar estimation techniques (e.g. OLSQ, fixed-effects, and GLS (including
random effects)) can produce biased and inconsistent parameter estimates even if the
error terms are not serially correlated. Most proposed solutions rely on first difference
transformations of the data with some form of instrumental variable estimation (Baltagi,
1995, p. 126) or generalized method of moments (GMM) techniques (Arrellano and
Bond, 1991). The differencing means that time-invariant dummy variables such as those
for RBG, the critical term, cannot be estimated. We employ OLSQ and random-effects

(GLS) estimation for equation (1) because fixed effects estimation is difficult when the
number of companies is large and the number of time periods small (Baltagi and Kao,
2000). Moreover, the bias of the fixed effects estimator increases with the coefficient
attached to the lagged value of the dependent variable in a dynamic panel data model
( Judson and Owen, 1999) and additional problems are created when the specification
includes categorical variables. Since fixed effects may be important in capturing unob-
served heterogeneity across companies, fixed effect and random effects models were
estimated on a larger sample (two to five years) using the percentage of shares held by
the largest shareholder to replace the RBG dummy variable. We found that the differ-
ence among parameter estimates of the fixed and random effects specifications was not
great, suggesting that random effects estimates are likely to be similar to fixed effects
estimates in our sample. However, the problem raised by the dynamic nature of the
panel remains. We therefore used GMM methods to estimate a model that did not
include dummy variables, and found little difference in the relevant coefficients, when
compared to random effects. We also used the Arrellano–Bond method to estimate the
ROA equation separately for the RBG and non-RBG samples, which suggests firm
specific heterogeneity is unlikely to be a major problem. We then chose between
random effects and OLSQ estimation, on the basis of the Hausman test, which indi-
cated that the null hypothesis of no difference in the estimated coefficients is not
rejected. In our estimation of equation (1), we report OLSQ (with standard errors
corrected for heteroscedasticity) results, though (unreported) random effects estimates
are in fact very similar.
Although we have suggested that endogeneity is not likely to be a significant issue in
Russia 1998–2001 we nonetheless control for potential endogeneity using 2SLS-IV
methods to estimate equation (2). To identify an instrument for RBG membership, we
first note that our sample of affiliates contains a mix of publicly quoted and unquoted
companies. As in the selection of firms for BG membership, discussed above, the
selection of firms that are publicly quoted in Russia in this period, before IPOs had
commenced, was determined by the political process underlying privatization. Thus we
suggest that being listed in this period was exogenous to the firms’ performance.
However, as suggested by Perotti and Gelfer (2001), business groups in Russia might
have better access to external funds including issuing shares for trading on public stock
exchanges. In fact, the correlation of RBGs with a dummy variable indicating that a firm
is publicly traded is relatively high and statistically significant, though the correlation
between the dummy for publicly traded and ROA is positive but only weakly significant.
This leads us to use the publicly traded dummy variable as an instrumental variable in
the RBG equation.[2] Equations (2) are therefore estimated using instrumental variables
in a two-stage approach (Woolridge, 2003).
RESULTS
We report the estimates of equations (1) and (2) in Table II. Equation (1) is estimated in
two specifications to take account of the potential multicollinearity, noted above,
between RBG (and RBG*lag(ROA)) and fixed assets and the results are reported in the

Table II. Profitability models estimation results
ROA ROA RBG

OLSQ IV LOGIT
(robust) IV
(1) (2) (3) (4) (5) (6)
ROA 0.189 0.451

(0.56) (0.13)
Lag(ROA) 0.437** 0.434** 0.287** 0.287** -0.075 0.510
(0.00) (0.00) (0.00) (0.00) (0.86) (0.18)
Log(Assets) 0.006 0.000
(0.78) (0.97)
Leverage 0.001** 0.001** 0.001** 0.001** -0.011**
(0.00) (0.00) (0.00) (0.00) (0.00)
Sales growth 0.039** 0.040** 0.037** 0.038**
(0.00) (0.00) (0.00) (0.00)
State-owned -0.009 -0.007 -0.011 -0.011
(0.30) (0.39) (0.26) (0.25)
Publicly traded 1.307** 1.143**
(0.00) (0.00)
Industry/region/Year Y*/Y/Y* Y*/Y/Y* Y*/Y*/Y Y*/Y*/Y Y*/Y*/N Y*/Y*/N
RBG 0.023* 0.026* 0.108** 0.108**
(0.02) (0.04) (0.00) (0.00)
RBG*Lag(ROA) -0.219* -0.211** -0.106** -0.105**
(0.02) (0.00) (0.00) (0.00)
Intercept 0.020* -0.047* -1.434** -1.436** -1.071** -0.735**
(0.02) (0.02) (0.00) (0.00) (0.00) (0.00)
Pseudo R2/R2 0.206 0.204 0.210 0.211 0.152 0.164
N 1938 1938 1803 1803 1818 1818
F 17.46* 22.34* 14.72* 15.15*
LR 337.86* 349.54*
Notes:
Y indicates that the equation includes dummy variables for industry/region/year.
N indicates that the equation does not include dummy variables for industry/region/year.
Y* indicates that the included dummy variables are collectively statistically significant ( p < 0.05).
* p < 0.05; ** p < 0.01; p-values in parentheses.
first two columns of Table II; the simultaneous equation estimates of equations (2) are
reported in columns (3)–(6).
Commencing with the estimates of equation (1), we report the OLSQ estimates with
heteroscedastic standard errors; the findings are more or less identical when GMM
methods are used (not reported). We find most of the control variables are significant
with the predicted signs. Higher leverage and sales growth raise ROA in Russia. As is
consistent with other studies for Russia, state ownership is not found to reduce ROA
(Estrin, 2002). The critical coefficients are the RBG coefficient, which is positive and

0.12
0.1
0.08
0.06
0.04
ROAt
Non-OBGs
0.02
OBGs
0
0 1
−0.02
−0.04
−0.06
−0.08
ROAt-1
Figure 2. Arrellano bond regression of ROAt on ROAt-1 for RBGs and non-RBGs
statistically significant, and the coefficient on the interactive term, RBG*Lag(ROA),

which is negative and statistically significant. Estimates that omit the interactive term
(RBG*Lag(ROA)) are not presented, but indicate that the coefficient on the RBG term
remains positive and statistically significant. Thus, the results are consistent with Hypoth-
eses 1a and 2a, but not with Hypotheses 1b and 2b.
These results are not sensitive to equation specification or method of estimation. We
undertook but do not report a variety of additional specifications. For example, the
equation was estimated with various combinations of the exogenous variables and the
results with respect to the hypotheses were not affected. We also re-estimate the equa-
tions using Hausman and Taylor (1981) methods on the entire dataset and once again
replicate the findings of Table II.[3] In summary, in the single equation specification we
find robust evidence that in Russia, RBG affiliates are more profitable than other firms,
and that they also practise persistence-reducing redistribution.
We also used the Arrellano–Bond method to estimate a simple version of the model,
where ROA was separately regressed only on lagged ROA, for the RBG and non-RBG
samples. The results for Arrellano–Bond regression shown in Figure 2 again suggest that
RBG affiliates exhibit less profit persistence, in support of Hypothesis 2a. These results
are in fact very similar to those reported by Lincoln and Gerlach (2004) in their analysis
of with group and non-group affiliated companies in Japan.
Columns (3)–(6) report the results of estimating equations (2) simultaneously. The two
specifications of the instrumented IV estimates of equation (1) are in columns (3) and (4)
while the two versions of the instrumenting equation determining RBG membership
estimated using dependent variable methods are reported in columns (5) and (6). As
before, we indicate robustness of findings in the face of multicollinearity by presenting
ROA results including and excluding ln Assets (columns (3) and (4) respectively). As an
indication of robustness to alternative specifications, we report an equation that uses a
slightly different first stage equation for specification (6), including the leverage variable
on the right hand side. The first stage equation is strongly significant (LR > 335) and the

results are robust to the reported and other unreported changes in specification. Note, for
example, that the results from using specification (5) versus (6) as the instrumenting
equation do not influence the results with respect to the main hypotheses in columns (3)
and (4). There is also no evidence of reverse causality, in that affiliate membership of an
RBG is not associated with current or previous profitability. These results confirm the
findings of Perotti and Gelfer (2001) and Guriev and Rachinsky (2005) and are also
consistent with the anecdotal evidence discussed above that oligarchs did not seek out
profitable firms in this period (Hoffman, 2003). However, as we suggested, stock market
listing is found to be an important advantage to RBG affiliates in the raising of capital,
while being uncorrelated with other aspects of RBG affiliate performance. The fit of the
first stage equation is also lower than that of the ROA equation, and it is even lower when
industry and location dummies are excluded, which is also indirect evidence against the
reverse causality argument. Thus although RBGs have a positive and significant influ-
ence on ROA in all estimates, the reverse does not appear to be true.
The relevant coefficients in the second stage ROA equation in columns (3) and (4) are
similar in terms of sign and significance of estimated coefficients to those reported in the
single equation framework (columns (1) and (2)). In terms of the two hypotheses, we find
in both specifications that the RBG coefficient is positive and statistically significant,
supporting Hypothesis 1a while the interactive term RBG*Lag(ROA) is negative, and
statistically significant, supporting Hypothesis 2a. The findings of the 2SLS-IV estima-
tion are therefore consistent with those of the single equation OLSQ estimation results.
Thus, we find robust evidence supportive of Hypotheses 1a and 2a. However, the
positive profitability effect (Hypothesis 1a) may be the result of internal market efficiency,
market power, or access to politically created rents. While we cannot fully distinguish
among these, we favour either the internal market efficiency or the rents interpretations
because of press reports and academic studies noted above that suggest oligarchs acted
to improve the performance of their affiliates. On the other hand, the finding of what we
have termed ‘BG Robustness’ (positive profits plus redistribution), is also supportive of a
politically inspired interpretation in which RBGs use their political connections to
maintain weaker firms so as to curry favour with the political elite. We are less convinced
about the market power argument, both because we control for industry structure, and
because Guriev and Rachinsky (2005), among others, fail to find evidence that RBGs
possess market power.
The evidence in favour of Hypothesis 2a relies on our interpretation of the lagged
dependent variable (persistence) term. In order to test the robustness of our conclusions,
we approached redistribution using two alternative empirical techniques. First, we follow
Perotti and Gelfer (2001) in using the cash flow method to test the hypothesis of
redistribution of financial resources among group members. In the finance literature, a
form of redistribution is often measured by the relationship (or lack thereof) between
investment activity and cash flows, where it is suggested that group affiliated firms may
be able to overcome financial constraints to investment because of the redistribution of
funds within the group. Though not free from criticism (see, for example, Kaplan and
Zingales, 2000), this approach was followed using our data and sample. The results
(unreported) show indirect support for redistribution of profits among group members in
that RBG affiliates are not reliant on internal cash flows to finance investments.

Table III. Variability of profits model estimation results
Mean equation ARCHM 1 ARCHM 2 ARCHM 3
RBG*Lag(ROA) -0.125* -0.159

(0.03) (0.01)
Lag(ROA) 0.440* 0.482 0.482
(0.00) (0.00) (0.00)
RBG 0.035* 0.025
(0.00) (0.00)
Lag(Var) 4.717* 4.635 1.722**
(0.00) (0.00) (0.00)
Intercept -0.066 -0.073 0.022
(0.01) (0.02) (0.01)
Variance equation
Lag(Var) -1.871 -1.885 -1.891
(0.00) (0.00) (0.00)
Lag(Error2) 0.001 0.009 0.008
(0.965) (0.613) (0.655)
Intercept 0.082 0.081 0.081
(0.00) (0.00) (0.00)
Log likelihood 661 669 664
Wald c2 266.89 283.88 279.84
N 1954 1954 1954
Notes:
p-values in parentheses.
*, ** Hausman-based test of no difference among coefficients of the mean equa-
tion is rejected at 10% level and 5% level respectively.
Second, we estimated the model of firm profitability allowing for the past conditional
variances to enter the mean equation of profits using the ARCH in Mean (ARCHM)
process (Woolridge, 2003, p. 416). We estimated jointly two equations which allow us to
test the hypothesis of risk-reduction among RBG and non-RBG firms: the lagged values
of the conditional variance enter the ROA equation and the variance covariance matrix
is simultaneously estimated as a function of its own past-values and values of past squared
disturbances.
The results are reported in Table III. Model 1 was estimated excluding the RBG
variable, while Models 2 and 3 were estimated respectively with RBG and additionally
with the interaction of RBG and lagged ROA terms. Variance was estimated as a
function of its own past values, squared disturbances and the RBG term in Model 3,
allowing for autoregressive and multiplicative conditional heteroscedasticity. Hence, in
Model 3 the variance term depends on RBG affiliation, while in the other models RBG
affiliation does not directly affect the variability of profits.
In all three models, the coefficients attached to the previous period values of the
variance are positive and statistically significant while the other coefficients in the ROA
equations are similar to the previous estimates in Table II. The main differences are

between Model 3 and the other two. The results indicate that variability plays a positive
and statistically significant role in the ROA equation, with the variability significantly
reduced for RBG affiliated relative to non-RBG affiliated firms (lagged variance param-
eter estimate in Models 2 and 3). The difference between lagged variance parameter
estimates with RBG (1.722) and without RBG influence (4.635) is significant. The
influence of past variance terms on future profitability is greatly reduced for the RBG
affiliated firms. This finding is perhaps unsurprising given the result in Figure 2. These
results also show that variability of profits is less for RBG than for non-RBG affiliated
firms, as well as the estimation results of the lagged ROA equations. The results indicate
that the past conditional variance of profits is greatly reduced when the interaction of the
RBG term and the conditional variance term is added in the mean ROA equation.[4]
Thus, taken together all three models (lagged ROA, the cash flow approach and the
conditional variance approach) suggest that the internal transfers of resources within the
BG acts to reduce risk among affiliated firms.
DISCUSSION
In this paper we hypothesize that there are important advantages to business groups and
their affiliates when market failures are present and when business groups are in the
process of forming. These advantages are linked to the relative benefits of internal
markets that permit the transfer of critical resources within the group network. These
transfers provide competitive advantage for the affiliates, but also promote group survival
through internal redistribution. Thus, we find evidence in support of what we term the
BG Robustness interpretation of RBG strategy, whereby firms affiliated with RBGs are
more profitable other things being equal, and that over time profits are redistributed
from stronger to weaker group members. Importantly, these results are robust to changes
in specification and estimation method, as well when as we control for multicollinearity
and potential endogeneity.
It is important to place these results in the context of the current literature on business
groups. While there is some evidence to support the hypothesis that group affiliation is
associated with higher profitability, the aggregate evidence is mixed at best (Khanna and
Rivkin, 2001). Likewise, while there is some evidence to support the redistribution
hypothesis, the aggregate evidence is also mixed (Khanna and Yafeh, 2005). Very few
studies have specifically examined both hypotheses together, and once again the evi-
dence suggests mixed results (Khanna and Yafeh, 2005). Thus, our results suggest that
the impact of business groups in Russia may be relatively unique, and this may relate to
their specific characteristics, including individual (oligarchic) control and the absence of
pyramids, their relative youth, as well as the specific institutional characteristics of the
Russian transition economy. In this regard, our results support the view of Yiu et al.
(2007) that it might not be possible to generalize research findings to all business groups.
In this study we have been careful to argue that our hypotheses and our results are
relevant to a specific context.
The results suggest that RBGs consider the goals of the group as a whole. Thus,
weaker firms benefit from profit redistribution practices within the RBG in a way
apparently similar to that documented in the case of business group affiliated companies

in Japan (Gedajlovic and Shapiro, 2002; Lincoln and Gerlach, 2004). However, the
statistical similarity does not account for the contextual differences between RBGs in
Russia and business groups in Japan. We argue that profit redistribution in the Russian
market is a response to external shocks in the transitional period. Profit redistribution
may be intended to help weaker (but vitally important for the functioning of the whole
group) members of RBGs to survive in this volatile environment. This is quite different
from the Japanese context. At the same time, individual firm profits are not sacrificed in
Russia, whereas this appears not to be the case in Japan. Affiliated firms benefit from
their association with RBGs though higher profitability.
We cannot claim that our approach completely resolves all of the outstanding issues
relating to BGs. However, we do suggest it narrows the number of possible inferences
that one can draw. In particular, we believe that our procedure more clearly distin-
guishes looting from other outcomes. For example, Perotti and Gelfer (2001) found it
difficult to distinguish the opportunistic transfer of resources, or looting, from more
efficient internal redistribution.
There are a number of other limitations of this study. Our sample period is short, with
a number of implications for our estimation methods, findings and conclusions. BG
Robustness may have been a sound strategy for RBG owners in the early years of
transition, when the environment was volatile, external markets were inefficient and the
options to change the RBG portfolio via sales or acquisitions were limited. However, as
Kedia et al. (2006) observed for India, the environment may evolve over time in a way
that alters the balance of advantage of different strategies. Thus, our results are not time
insensitive. We note that there is a considerable difference between Russia and India
in that family business groups existed in India even before liberalization. In contrast,
business groups in Russia were formed as a consequence of liberalization. Nevertheless,
we would not dispute the possibility that even if BGs, including those from Russia,
originally achieve both higher profitability as well as network stability, the costs and
benefits of that strategy may push them towards other outcomes as markets becomes
more efficient. Similarly, Khanna and Palepu (2000b) suggest that business group advan-
tages may dissipate over time as markets become more efficient, and Peng et al. (2005)
suggest that the efficient scope of the firm may change over time as institutional relat-
edness changes. These possibilities could be tested with a longer dataset extending over
the Putin boom era.
Moreover, a larger dataset would facilitate the consistent use of superior estimation
techniques, such as GMM, and further checks for robustness. Additionally, if the dataset
were also richer, a larger set of potential valid instruments might be available to further
address the issue of endogeneity, itself more likely to be an increasing problem as
maturity of RBGs permits their owners to increase the selection of affiliated member
firms. Finally, the time gaps in our data are an important limitation. This study was
performed using balanced panels of around 1,000 companies for 1998, 2000 and 2001.
Gaps in data included 1999 and 2002. Future research should therefore examine the
influence of RBG affiliation on firm performance using broader time frames and possibly
larger and richer samples.
This study isolates a redistribution effect of RBGs on affiliate profits. More research is
needed to examine the underlying reasons and possible consequences of this

redistribution. It would be also interesting to examine the influence of RBG affiliation on
a wider range of performance indicators and under different model specifications. The
general approach to RBGs taken here, which views them as rational economic units and
an important business phenomenon, should guide further investigations in this area.
Finally, it would be unreasonable to completely ignore the evidence of corrupt prac-
tices among Russian oligarchs that are reported in the popular press and elsewhere
(Aslund, 2006), and future studies should more carefully examine these issues. At the
same time, one should not confuse the circumstances under which Russian Business
Groups were created with the subsequent performance of the group and its affiliates.
For example, Kramer (2007) indicates that despite the possibly questionable method
by which Mikhail Prokhorov and Vladimir Potanin gained control over Norilsk (and
the subsequent circumstances under which Prokhorov was displaced), they did succeed
in making the company a successful and globally competitive mining firm. Our claim
is therefore not that RBGs were formed in a transparent and non-corrupt way; it is
that once formed, they overcame the obstacles created by market and institutional
failures. Our evidence therefore argues for a more nuanced approach towards Russian
business groups, one that recognizes both their potential liabilities and their potential
benefits.
This study has a number of important managerial implications. First, multinational
firms operating in emerging markets will already be aware that the levels of risk may well
be higher than those pertaining in developed economies, and that the relatively more
unstable political environment is likely to be a major source of such risk. Our study
suggests that because the process of formation of business groups in Russia and the
associated privatization process were primarily driven by political rather than economic
considerations, the risks arising from political rather than business sources will be par-
ticularly pronounced in Russia, even in comparison with other emerging markets. Our
research also suggests that in Russia, new large industrial groups may be forming and
dying, driven by mechanisms deriving largely from the social and political environment
rather than the profit motive. This adds a layer of complexity for Western firms in
judging potential partners and building long term business relationships. To the extent
that this phenomenon is a consequence of the transition process, it seems likely to
generalize to other transition economies, notably those in the former Soviet Union where
the same factors seem likely to drive the emergence and survival of business groups.
Finally, our research suggests that multinationals entering Russia may face particularly
intense and unexpected competition from domestic firms that at first sight seem relatively
weak. As we have established, the Russian business environment is dominated by a small
number of huge business groups, typically with rather loose horizontal connections but
strong vertical ones (H-forms in the Yiu et al. (2007) classification). We have shown that
these groups add value to the individual firms and redistribute resources in support of
companies which are struggling. Hence one can conceive of scenarios in which a
Western firm enters a Russian market, with a few relatively unsuccessful domestic
competitors, later to discover that one or more of these are actually owned by a well
resourced RBG with the objective of maintaining the group intact. Knowledge about the
ownership of domestic firms is therefore an important piece of information in guiding the
strategy of potential Western entrants.

CONCLUSION
In this study we have examined the impact of business group affiliation on the perfor-
mance of Russian firms over the period 1998–2001. The study contributes to the
literature in four ways. First, we contribute to the understanding of how existing theories
may be context-specific across both time and space. Our study therefore contributes to
a broader understanding of the role and nature of business groups in different institu-
tional environments, and in particular allows us to study a moment in time when BGs
were being formed. Our results suggest that benefits to group affiliation may be pro-
nounced when business groups are relatively young and when the institutional infra-
structure is weak and uncertain. We therefore support the view of Yiu et al. (2007) that
firms adapt to their external environments in different ways, and that these differences
should be explicitly recognized. Second, we apply an empirical methodology that
permits us to test two non-mutually exclusive hypotheses within the same estimation
framework. Specifically, we hypothesized that firms affiliated with RBGs would be either
more or less profitable (the profitability hypothesis), and/or would redistribute profit in
order to minimize risk or maximize efficiency (the redistribution hypothesis). Although
both hypotheses have been previously considered in the literature, with the exception of
Khanna and Yafeh (2005) they have not been considered together. We argue that both
are relevant in the Russian context because of the market failures and market shocks that
are a feature of the Russian economic landscape. Third, the two pairs of hypotheses
allow us to identify empirically four alternative interpretations of the strategy and
behaviour of RBGs. Thus, we gain a degree of interpretative freedom not available in
previous studies. Finally, we add to the growing number of studies that examine business
groups in specific emerging markets. However, few of these have focused on transition
economies, and in particular on the increasingly significant and distinctive Russian case.
Thus, we use a unique and relatively large dataset to analyse Russian business groups.
These data have not been used for this purpose in previous studies. We also use the
specific features of the Russian experience to produce empirical estimates that are
relatively free of some of the statistical problems associated with endogeneity.
ACKNOWLEDGMENTS
The authors thank Andrew Delios, three referees, Randolph Bruno, Maria Bytchkova, Sasha Janovskaia,
Klaus Mayer, Mario Nuti, Enrico Perotti, Leyland Pitt, Melsa Ararat, Yao Tang and Michael Carney for
useful comments and discussion. They acknowledge financial support from the Social Sciences and Humani-
ties Research Council of Canada. Any errors are their own responsibility.
NOTES
[1] Numerous press stories support this view, even currently. Thus http://www.nns.ru (accessed 29 Feb-
ruary 2008) reports: ‘Why did Abramovich buy this company? . . . the business is not very
advanced . . . the technology is old . . . what type of innovation can we expect? Overall the explanation
seems to be simpler and not at all original; it’s a big company . . . and the oligarch is attracted by big
money.’
[2] To identify the system, we did not use this variable in the ROA equation. Usually lagged values of the
same variable can also be used as instruments. However, since the ROA equation uses a lagged-

dependent variable, lagged ROA is already an instrument. The correlation between residuals saved
from the original ROA regression (equation 1) and the dummy for publicly traded was not significant,
suggesting that it is an appropriate instrument.
[3] Neither OLSQ nor random effects estimation methods are appropriate if BG affiliation is correlated
with latent individual firm effects (and also because of endogeneity problems). Thus, we employed the
Hausman–Taylor (HT) estimation method that allows for the time-invariant regressors (RBG affiliation)
to be correlated with latent individual effects (Hausman and Taylor, 1981). The HT method makes use
of time-varying explanatory variables to serve as instruments for endogenous time-invariant variables.
Unlike fixed effects models HT produces estimates of the coefficients of time-invariant variables.
[4] The ARCHM models presented here were estimated without control variables or other variables used
in the main lagged ROA equation due to computational intensity. However, we also estimated a model
using an OLS regression approximation approach, allowing the control variables to enter the mean
equation. The results did not change.
REFERENCES
Almeida, H. and Wolfenzon, D. (2006). ‘A theory of pyramidal ownership and family business groups’.
Journal of Finance, 61, 2637–80.
Arrellano, M. and Bond, E. (1991). ‘Some tests of specification for panel data’. Review of Economic Studies, 58,
277–97.
Aslund, A. (2006). ‘The hunt for Russia’s riches’. Foreign Policy, January/February, 43–8.
Baltagi, B. (1995). Econometric Analysis of Panel Data. New York: Wiley.
Baltagi, B. H. and Kao, C. (2000). ‘Nonstationary panels, cointegration in panels and dynamic panels: a
survey’. In Baltagi, B. H. (Ed.), Advanced Econometrics. New York: Elsevier Science, 15, 7–53.
Batjargal, B. (2003). ‘Social capital and entrepreneurial performance in Russia’. Organization Studies, 24,
535–56.
Batjargal, B. (2007). ‘Comparative social capital: networks of entrepreneurs and venture capitalists in China
and Russia’. Management and Organization Review, 3, 397–419.
Bertrand, M., Mehta, P. and Mullainathan, S. (2002). ‘Ferreting out tunneling: an application to Indian
business groups’. Quarterly Journal of Economics, 117, 121–48.
Black, B. (2001). ‘The corporate governance behavior and market value of Russian firms’. Emerging Markets
Review, 2, 1–23.
Black, B., Jang, H. and Kim, W. (2004). Does Corporate Governance Predict Firms’ Market Values? Evidence from Korea.
Working Paper 02-4, Stanford Law School, Stanford, CA.
Blanchard, O. and Kremer, M. (1997). ‘Disorganization’. Quarterly Journal of Economics, 112, 1091–126.
Boisot, M. and Child, J. (1996). ‘From fiefs to clans and network capitalism: explaining China’s emerging
economic order’. Administrative Science Quarterly, 41, 600–28.
Boone, P. and Rodionov, D. (2002). Rent Seeking in Russia and the CIS. Brunswick UBS, Warburg, Moscow.
Campos, N. F. and Giovannoni, F. (2005). The Determinants of Asset Stripping: Theory and Evidence from the
Transition Economies. IZA Discussion Paper 1867, November.
Caves, R. E. and Uekusa, M. (1976). Industrial Organization in Japan. Washington, DC: The Brookings
Institution.
Chacar, A. and Vissa, B. (2005). ‘Are emerging economies less efficient? Performance persistence and the
impact of business group affiliation’. Strategic Management Journal, 26, 933–46.
Chang, S. J. (2003). ‘Ownership structure, expropriation, and performance of group-affiliated companies in
Korea’. Academy of Management Journal, 46, 238–53.
Chang, S. J. and Hong, J. J. (2000). ‘Ownership economic performance of the Korean business groups:
intra-group resource sharing and internal business transaction’. Academy of Management Journal, 43,
429–48.
Desai, P. (2006). Conversations on Russia: Reform from Yeltsin to Putin. Oxford: Oxford University Press.
Djankov, S. and Murrell, P. (2002). ‘Enterprise restructuring in transition: a quantitative survey’. Journal of
Economic Literature, 40, 739–92.
EBRD (European Bank for Reconstruction and Development) (2007). Business in Transition. London:
European Bank for Reconstruction and Development.
Estrin, S. (2002). ‘Competition and corporate governance in transition’. Journal of Economic Perspectives, 16,
101–24.
Filatotchev, I., Kapelyushnikov, R. and Demina, N. (2001). ‘The effects of ownership concentration on
investment and performance in privatized firms’. Managerial and Decision Economics, 22, 299–313.

Fisman, R. (2001). ‘Estimating the value of political connections’. American Economic Review, 91, 1095–102.
Garg, M. and Delios, A. (2007). ‘Survival of the foreign subsidiaries of TMNCs: the influence of business
group affiliation’. Journal of International Management, 3, 278–95.
Gedajlovic, E. and Shapiro, D. M. (2002). ‘Ownership structure and firm profitability in Japan’. Academy of
Management Journal, 45, 565–75.
Globerman, S. and Shapiro, D. M. (2003). ‘Governance infrastructure and U.S. foreign direct investment’.
Journal of International Business Studies, 34, 19–39.
Godoy, S. and Stiglitz, J. (2007). ‘Growth, initial conditions, law and speed of privatization in transition
countries: 11 years later’. In Estrin, S., Kolodko, G. W. and Uvalic, M. (Eds), Transition and Beyond.
Basingstoke and New York: Palgrave Macmillan, 89–117.
Goldman, M. (2003). The Privatization of Russia: Russian Reform Goes Awry. London: Routledge.
Gorodnichenko, Y. and Grygorenko, Y. (2008). ‘Are oligarchs productive?’. Journal of Comparative Economics,
35, 17–42.
Guriev, S. and Rachinsky, A. (2005). ‘The role of oligarchs in Russian capitalism’. Journal of Economic
Perspectives, 19, 131–50.
Hausman, J. and Taylor, W. (1981). ‘Panel data and unobserved individual effects’. Econometrica, 49,
1377–98.
Hoffman, D. (2003). The Oligarchs: Wealth and Power in the New Russia. New York: Public Affairs.
Hoshi, T. and Kashyap, A. (2001). Corporate Financing and Governance in Japan: The Road to the Future. Cambridge,
MA: MIT Press.
Johnson, J. (1997). ‘Understanding Russia’s emerging financial-industrial groups’. Post-Soviet Affairs, 13,
333–65.
Johnson, S., La Porta, R., Lopes-de-Silanes, F. and Shleifer, A. (2000). ‘Tunneling’. American Economic Review,
90, 22–7.
Judson, R. A. and Owen, A. L. (1999). ‘Estimating dynamic panel data models: a guide for macroecono-
mists’. Economics Letters, 65, 9–15.
Kaplan, S. N. and Zingales, L. (2000). ‘Investment-cash flow sensitivities are not valid measures of financing
constraints’. Quarterly Journal of Economics, 115, 707–12.
Kedia, B. L., Mukherjee, D. and Lahiri, S. (2006). ‘Indian business groups: evolution and transformation’.
Asia Pacific Journal of Management, 23, 559–77.
Khanna, T. and Palepu, K. (1997). ‘Why focused strategies may be wrong for emerging markets’. Harvard
Business Review, 75, 3–10.
Khanna, T. and Palepu, K. (2000a). ‘Is group affiliation profitable in emerging markets? An analysis of
diversified Indian business groups’. Journal of Finance, IV, 867–91.
Khanna, T. and Palepu, K. (2000b). ‘The future of business groups in emerging markets: long-run evidence
from Chile’. Academy of Management Journal, 43, 268–86.
Khanna, T. and Rivkin, J. W. (2001). ‘Estimating the performance effects of business groups in emerging
markets’. Strategic Management Journal, 22, 45–74.
Khanna, T. and Yafeh, Y. (2005). ‘Business groups and risk sharing around the world’. Journal of Business, 78,
301–40.
Khanna, T. and Yafeh, Y. (2007). ‘Business groups in emerging markets: paragons or parasites’. Journal of
Economic Literature, XLV, 331–72.
Kock, C. and Guillen, M. (2005). ‘Strategy and structure in developing countries: business groups as an
evolutionary response to opportunities for unrelated diversification’. Industrial and Corporate Change, 10,
77–113.
Kramer, A. E. (2007). ‘The Kremlin flexes, and a tycoon reels’. New York Times, 8 July.
Ledeneva, A. V. (1998). Russia’s Economy of Favours: Blat, Networking and Informal Exchange. Cambridge:
Cambridge University Press.
Leff, N. (1976). ‘Capital markets in the less developed countries: the group principle’. In McKinnon, R. (Ed.),
Money and Finance in Economic Growth and Development. New York: Marcel Dekker, 97–122.
Leff, N. (1978). ‘Industrial organization and entrepreneurship in developing countries: the economic groups’.
Economic Development and Cultural Change, 26, 661–75.
Lincoln, J. R. and Gerlach, M. L. (2004). Japan’s Network Economy. Cambridge: Cambridge University Press.
Lincoln, J. R., Gerlach, M. L. and Ahmadjian, C. L. (1996). ‘Keiretsu networks and corporate performance
in Japan’. American Sociological Review, 61, 67–88.
Morck, R., Wolfenzon, D. and Yeung, B. (2005). ‘Corporate governance, economic entrenchment and
growth’. Journal of Economic Literature, 43, 657–722.
Mueller, D. C. (1986). Profits in the Long Run. Cambridge: Cambridge University Press.

Mueller, D. C. (1990). ‘The dynamics of company profits: profits and the process of competition’. In Mueller,
D. (Ed.), The Dynamics of Company Profits: An International Comparison. Cambridge: Cambridge University
Press, 1–15.
Peng, M. (2003). ‘Institutional transitions and strategic choices’. Academy of Management Review, 28, 275–96.
Peng, M., Lee, S. K. and Wang, D. (2005). ‘What determines the scope of the firm over time? A focus on
institutional relatedness’. Academy of Management Review, 30, 622–33.
Perotti, E. and Gelfer, S. (2001). ‘Red barons or robber barons? Governance and investment in Russian
financial-industrial groups’. European Economic Review, 45, 1601–17.
Puffer, S. M. and McCarthy, D. J. (2007). ‘Can Russia’s state-managed network capitalism be competitive?
Institutional pull versus institutional push’. Journal of World Business, 42, 1–13.
Scharfstein, D. and Stein, J. (2000). ‘The dark side of internal capital markets: divisional rent-seeking and
inefficient investment’. Journal of Finance, 55, 2537–64.
Shleifer, A. and Treisman, D. (2005). ‘A normal country: Russia after communism’. Journal of Economic
Perspectives, 19, 151–74.
Stiglitz, J. (2002). Globalization and its Discontents. New York: W.W. Norton.
Strachan, H. (1976). Family and other Business Groups in Economic Development: The Case of Nicaragua. New York:
Praeger.
Suhomlinova, O. (2006). ‘Toward a model of organizational co-evolution in transition economies’. Journal of
Management Studies, 46, 1537–58.
Tooby, J. and Cosmides, L. (1996). ‘Friendship and the banker’s paradox: other pathways to the evolution
of adaptations for altruism’. Proceedings of the British Academy, 88, 119–43.
Woolridge, J. M. (2003). Introductory Econometrics, 2nd edition. Mason, OH: South-Western Thomson.
Wright, M., Filatotchev, I., Hoskisson, R. E. and Peng, M. (2005). ‘Strategic research in emerging markets:
challenging the conventional wisdom’. Journal of Management Studies, 42, 1–33.
Yiu, D., Bruton, G. D. and Lu, Y. (2005). ‘Understanding business group performance in an emerging
economy: acquiring resources and capabilities in order to prosper’. Journal of Management Studies, 42,
183–206.
Yiu, D., Lu, Y., Bruton, G. D. and Hoskisson, R. E. (2007). ‘Business groups: an integrated model to focus
future research’. Journal of Management Studies, 44, 1551–79.
Young, M. N., Peng, M., Ahlstrom, D., Bruton, G. D. and Jiang, Y. (2008). ‘Corporate governance in
emerging economies: a review of the principal-principal perspective’. Journal of Management Studies, 45,
196–220.

The Performance Effects of Business Groups in Russia: Saul Estrin, Svetlana Poukliakova and Daniel Shapiro

Uploaded by

Copyright:

Available Formats

The Performance Effects of Business Groups in Russia: Saul Estrin, Svetlana Poukliakova and Daniel Shapiro

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Performance Effects of Business Groups in Russia: Saul Estrin, Svetlana Poukliakova and Daniel Shapiro

Uploaded by

Copyright:

Available Formats

Journal of Management Studies 46:3 May 2009

The Performance Effects of Business Groups in Russia

Saul Estrin, Svetlana Poukliakova and Daniel Shapiro

© Blackwell Publishing Ltd 2009

© Blackwell Publishing Ltd 2009

© Blackwell Publishing Ltd 2009

RUSSIAN BUSINESS GROUPS (RBGs)

© Blackwell Publishing Ltd 2009

The random nature of the process is further explained by Hoffman (2003):

The Profitability Hypotheses

© Blackwell Publishing Ltd 2009

© Blackwell Publishing Ltd 2009

The Redistribution Hypotheses

© Blackwell Publishing Ltd 2009

© Blackwell Publishing Ltd 2009

Hypothesis 2a posits that internal redistribution results in lower profit persistence,

© Blackwell Publishing Ltd 2009

BGs use internal markets BG membership lowers

H2b Supported BGs raise profitability of BGs are inefficient and

Figure 1. Summary of Hypotheses 1 and 2

© Blackwell Publishing Ltd 2009

© Blackwell Publishing Ltd 2009

Specification, Estimation and Data

ROA jt = β0 + β1′ xjt +λ RBG jt +γ ROA jt − k + δ RBG jt ∗ROA jt − k + ε jt (1)

© Blackwell Publishing Ltd 2009

RBG2 jt = β20 + β21′ x 2 jt + β22 ROA jt + β23 ROA jt − k + ε 2 jt .

© Blackwell Publishing Ltd 2009

© Blackwell Publishing Ltd 2009

Variable Mean St. Dev. 1 2 3 4 5 6

1. ROA 0.07 0.18

© Blackwell Publishing Ltd 2009

© Blackwell Publishing Ltd 2009

ROA ROA RBG

(1) (2) (3) (4) (5) (6)

ROA 0.189 0.451

© Blackwell Publishing Ltd 2009

statistically significant, and the coefficient on the interactive term, RBG*Lag(ROA),

© Blackwell Publishing Ltd 2009

© Blackwell Publishing Ltd 2009

Mean equation ARCHM 1 ARCHM 2 ARCHM 3

RBG*Lag(ROA) -0.125* -0.159

© Blackwell Publishing Ltd 2009

© Blackwell Publishing Ltd 2009

© Blackwell Publishing Ltd 2009

© Blackwell Publishing Ltd 2009

© Blackwell Publishing Ltd 2009

© Blackwell Publishing Ltd 2009

© Blackwell Publishing Ltd 2009

© Blackwell Publishing Ltd 2009

You might also like

RBGLag(ROA) -0.125 -0.159