CORPORATE GOVERNANCE AND PERFORMANCE: A STUDY OF LISTED SUBSIDIARIES

How to cite this paper: Di Carlo


INTRODUCTION
The literature on corporate governance has sought various ways to find out how firms' financial performance can be affected, among others, by the ownership structure (e.g., concentrated vs. widely held firms; family vs. non-family firms) and board demographics, (e.g., board size, CEO duality, outsider ratio).However, the so-called "input-output studies" have given ambiguous results, stimulating scholars to present new perspectives on corporate governance research (Daily et al., 2003;Huse et al., 2011;Zattoni & Pugliese, 2019).
Many of the corporations analysed by scholars in empirical studies are parent companies since they control several subsidiaries, either wholly or partially owned (La Porta et al., 1999), whose financial data are consolidated into the parent's financial statements.
When the company under observation is the holding of a business group, normally the focus of researchers is not on its separate financial statements but instead on the consolidated of the group (Di Carlo et al., 2016).Indeed, the separate financial statements of the holding company, and those of subsidiaries, have the problem of being affected by intra-group transactions whereas the consolidated is considered as the financial statements of a group of legal entities presented as those of a single economic entity, since intragroup assets and liabilities, equity, income, expenses and cash flows relating to transactions between entities of the group must be eliminated in full.From this consideration, it seems that the consolidated financial statements as well as the reporting economic entity (i.e., the business group) are independent of external influences.
However, the consolidated financial statements used by researchers may be provided by sub-holdings that operate in wider business groups.The effects of transactions, when present, between the first level holding and the controlled sub-holding, as well as those between the latter and the other affiliates are not eliminated by the consolidation procedures.
This aspect is particularly relevant when the sub-holding is listed and the transactions with the parent and the other affiliates are carried out to serve the interest of the parent itself or that of the wider group.Indeed, the directed transactions could strongly affect the financial performance of directed listed companies.This is obviously an important area in terms of the relationship between the board of the parent and that of the sub-holding under observation.Indeed, the sub-holding board may be totally dependent from the parent or, instead, it could present an elevated degree of decision-making autonomy (Di Carlo et al., 2016).In the former case, even if a sub-holding provides the consolidated financial statements of its group as if it were a single business entity, the relevant economic entity is represented in the consolidated of the first-level parent company.
So far, the literature seems not to have adequately considered these important aspects.In fact, many scholars collected samples in countries where business groups are pervasive (e.g., Italy, France, Spain), without taking into account the dependency of listed firms from their parents (e.g., in Italy: Barontini Zattoni and Minichilli (2009)).In other terms, researchers seem to underestimate the delegation/centralization of the decision-making power by the holding when the subsidiary is a listed company since they mainly treat listed subsidiaries and groups they eventually control as independent economic entities.
Moreover, parents of listed subsidiaries may be in turn listed.For example, for the U.S. companies it has been observed that "although in most situations the parent of a controlled company is an individual or a non-public entity, in some instances controlled companies are, or may become, controlled by a public parent" (Rubin, 2006).This phenomenon is also relevant in all countries that see the presence of pyramidal business groups (Di Carlo, 2014a; La Porta et al., 1999).For instance, in Germany, the largest shareholder of Volkswagen AG is another listed company, Porsche AG; in France, the listed Louis Vuitton Moët Hennessy (LVMH) is controlled by the listed company Christian Dior SA; in Italy, the listed Saipem SpA is controlled by the listed Eni SpA.
Stemming from these considerations and focusing on listed subsidiaries, the main objective of this article is to explore how and why this phenomenon is relevant, giving some suggestions on the interpretation of the ownership structure, board demography, and the financial performances of directed listed subsidiaries.
In order to answer these research questions, supported by the literature on centralizationautonomy in business groups and on subsidiaries' boards, the article focuses the attention on the Italian-listed companies.We use the Italian context because according to the regulation introduced by the Italian Corporate Law Reform of 2003, it is possible to know if the parent company exercises the decision-making power or whether this power is delegated to the subsidiaries' insiders (directors and executives).Indeed, subsidiaries have to disclose in their correspondence and official documents, such as in their financial statements, whether the controlling parent company exercises a management activity over them (Di Carlo, 2014a).
To explore the relevance of the phenomenon we use a descriptive statistic on the sample of companies listed in the Italian Stock Exchange controlled and consolidated by other companies at the end of 2018.The analysis shows that 121 firms (74%) of Italian non-financial listed companies are consolidated by the respective controlling entities, 35 of which (21.5%) declare, according to the Italian regulation, to be directed by their parents.Thus, they are not independent economic entities and the effort to study the relationship between corporate governance variables (e.g., ownership structure and board composition) and firm performance could be strongly biased.
Our findings have several implications for academics, practitioners, and policymakers.
The remainder of this paper is structured as follows.Section 2 reviews the relevant literature.Section 3 provides the research methodology.Section 4 presents the results and discusses the findings.Finally, Section 5 is dedicated to the conclusions.Khanna & Yafeh, 2005).In this study, we adopt that of Chang and Hong (2002), who define the business group as a "gathering of formally independent firms under single common administrative and financial control, and are owned and controlled by certain families" (p.266).That definition is particularly useful for our purpose because, despite the presence of formally independent legal entities, the business group is considered as a single economic entity.

THEORETICAL BACKGROUND
The term "economic entity" is often recalled by accounting standards, especially for what concern the consolidated financial statements (International Financial Reporting Standard (IFRS) No. 10), i.e., "the financial statements of a group in which the assets, liabilities, equity, income, expenses and cash flows of the parent and its subsidiaries are presented as those of a single economic entity" (IAS Plus, 2012), without considering the borders of the legal entities.The entity theory of consolidated financial statements focuses on the economic entity as a whole, recognizing that the parent, while not owning 100% of the assets, has effective control of the entire subsidiary (Moonitz, 1942).
To be consolidated, a legal entity must be controlled by its parent company.According to IFRS 10 (para.6), "an investor controls an investee when it is exposed, or has rights, to variable returns from its involvement with the investee and has the ability to affect those returns through its power over the investee".Moreover, "an investor controls an investee if and only if the investor has all the following: (a) power over the investee; (b) exposure, or rights, to variable returns from its involvement with the investee; and (c) the ability to use its power over the investee to affect the amount of the investor's returns' (IFRS 10, para.7).
Therefore, the accounting standard requires the parent (the investor) to consolidate a subsidiary legal entity (the investee) when it is substantially controlled, even if the parent does not exert its power.Indeed, control is the ability to use the power (to direct the relevant activities) while management is the exercise of the control power, i.e., the exercise of the decision-making power.These are the definitions of control and management that we use in this article since we want to investigate the significance of the consolidated financial statements of different levels of sub-holding companies.
The control is generally presumed by reason of the ability to elect the majority of the board of directors -de jure control.Since the parent retains voting control, it has the authority to select the subsidiary's directors.However, the concept of control also includes what is often referred to as de facto control.An example of de facto control might be a situation in which a parent holds less than 50% of the voting rights of a subsidiary but it is enough to force the latter to act in accordance with its desires.
Pursuant to IFRS 10, de jure and de facto controls are only indicators of potential substantial control, i.e., the power to govern.Indeed, there could be exceptional circumstances where it can be clearly demonstrated that possession of de jure and de facto control does not constitute substantial control (IFRS 10, para.11).
Following this reasoning, substantial control can be an important indicator to presume the direction activity by the parent.However, as seen for the indicators of substantial control, the controlling entity may demonstrate that control does not lead to a direction activity.

Separation between ownership, control and management
The most common instruments to separate ownership from control are pyramiding, cross-ownership, golden shares and dual-class equity.
While the separation between ownership and control has been widely investigated (Claessens et al., 2000;Faccio & Lang, 2002;Grossman & Hart, 1988;La Porta et al., 1999) that between control and management has received only a few attentions, especially in business groups (Di Carlo, 2014a; Di Carlo et al., 2016).Claessens et al. (2000) study the separation between control and management "by investigating whether a member of the controlling family or an employee of the controlling widely held financial institution or corporation is the CEO, chairman, honorary chairman, or vice-chairman of the company" (p.81).They find that in East Asia countries the separation of management from ownership control is rare while there is no separation between control and management.
In this article, we analyse the separation between control and management referring to the parent-subsidiary relationship and, more specifically, to the parent-listed sub-holding relationship.
Indeed, we state that in a business group, the control is separated from the management when the parent (listed or not listed) company delegates the exercise of the decision-making power to the subsidiaries (Figure 1, Case A).
It can be also stated that the delegation of the decision-making power proves that the subsidiary is the parent's agent.That delegation raises the headquarters-subsidiary agency problem, mainly studied in multinational enterprises (Du et al., 2015;Kim et al., 2005), since the interests of the subsidiaries may not always be aligned with those of the parent or that of the group as a whole.Subsidiary autonomy can be defined as the decision-making rights that are granted by the parent (Gammelgaard et al., 2012).According to Brooke (1984), autonomy refers to a situation "in which units and sub-units possess the ability to make decisions for themselves on issues which are reserved to a higher level in comparable organizations" (p.9).A subsidiary possesses high autonomy when it exercises the power to direct operational and/or strategic decisions.Low autonomy arises when the parent largely makes such decisions.Some scholars make this distinction, clarifying that autonomy "may either be delegated by headquarter or developed by the subsidiary" (Young & Tavares, 2004, p. 228), while decentralization concerns delegation.However, for this study, we use the terms decentralization (of the parent) and autonomy (of a subsidiary) synonymously, without making a distinction between the two.
As stated before, the IFRS 10 requires the parent company to consolidate its subsidiaries even if the former decides to delegate the management to the latter.
Nevertheless, despite the possibility of the parent to separate control from the management of the subsidiaries, scholars on corporate governance normally consider the group as a single economic entity with a unitary direction activity exercised by the holding company.This direction facilitates the reduction or even the removal the transaction costs (Coase, 1960;Williamson, 1985), especially in markets that have a high degree of inefficiency (Claessens et al., 2006;Goto, 1982;Khanna & Palepu, 1999;Khanna & Yafeh, 2007;Leff, 1978).
Moreover, from the literature on pyramidal groups, characterised by a high separation between ownership and control, it seems that the first level holding company manages all the companies located at different levels of the control chain (Figure 1, Case B).In this case, the subsidiary does not have the role of the parent's agent, because of the absence of the decision-making power autonomy.In other terms, the parent and subsidiary are considered as a single economic entity.
For the Italian context, Zattoni (1999) states that large Italian groups are characterised by a pyramidal structure at the head of which there is a holding company, which is the managing centre of all group activity.Indeed, it would seem that the parent company at the top of the wider group not only controls but also directs its controlled sub-holdings, to achieve, among other things, the synergy effect in the interest of the whole group.
However, in pyramidal groups, scholars normally interpret the direction activity of the parent in a negative way since the controlling party, at the top of the control chain, has an incentive to divert resources within the group through intragroup transactions (Almeida & Wolfenzon, 2006;Claessens et al., 2000;La Porta et al., 1999).The high separation between ownership and control associated with the management of the controlled subsidiaries could lead to significant private benefits for the controlling party at the expense of minority shareholders (Doidge et al., 2009;Dyck & Zingales, 2004;Fan et al., 2016;Nenova, 2000).Thus, in pyramidal groups where the risk of expropriation is particularly high, the holding tends to demonstrate that the control is separated from the management, even if that separation may be more apparent than real (Di Carlo, 2014a).

The autonomy-dependence of subsidiary boards in multinational enterprises
The issue of subsidiary autonomy/dependence has been mainly investigated by the literature on multinational enterprises (MNEs) (Birkinshaw, 1997;Birkinshaw et al., 2005;Gammelgaard et al., 2012;Young & Tavares, 2004).From the literature review on MNEs conducted by Young and Tavares (2004), emerges that the attitude to centralization/ decentralization of the parent company's decisionmaking power depends, among other things, on parent company factors (e.g., culture and management style, mission and objectives, planning and control mechanism), certain subsidiary characteristics (acquisition mode of establishment) and subsidiary evolution (longer established affiliates), MNEs strategies of global or regional integration.Thus, the lack of studies on subsidiaries' autonomy in non-multinational business groups probably depends on the fact that in these groups the idea is that the parents normally centralize the decision-making power since subsidiaries do not need adaptation to the local contexts.
Also, the literature on corporate governance is mainly focused on the autonomy-dependence of subsidiary boards of MNEs.Kiel et al. (2006) explore the autonomy of decision-making power for subsidiaries that are called to deal with a global business strategy.The alternatives mostly arise "considering cost reduction pressures on the one hand and the extent of local market responsiveness on the other hand" (Kiel et al., 2006, p. 570).Thus, while in a global strategy, the subsidiary boards are completely dependent from the parent (they are rubber stamp boards), a multi-domestic strategy is often associated with substantial strategic decision-making at the subsidiary level, and then the subsidiary boards are much more independent.In case of total dependence from the parent "the subsidiary's legal board, comprised entirely of local managers, is a compliance board with no formal responsibilities outside those required under law" (Kiel et al., 2006, p. 572).
Leksell and Lindgren (1982) find three different roles for the foreign subsidiary board: external roles, internal roles, legal roles.For the first role, the function of the board is to act as a link between the foreign board subsidiary and its host environment.The second focuses on the linkages and relationships between the foreign subsidiary and the MNC as a whole.The third is just a law role since the law countries in which subsidiaries operate prescribe obligations to the board of directors, which of course must be fulfilled by the subsidiary board (e.g., conducting the annual general meeting and attesting to legal matters such as the annual corporate reporting to regulators).Indeed, some countries require that foreign corporations do business through a domestically incorporated legal entity.

The legal role of the board of directors and the subsidiary directors' dilemma
As observed in the previous sub-section, scholars agree on the fact that when the parent centralizes the decision-making power the subsidiary board plays merely a passive legal role.Thus, the importance of that role seems to be underestimated by the literature on MNEs, especially for what concerns the influence of legal obligations on decision-making power.
Conversely, the legal role of subsidiary boards has been underlined by the corporate law literature.
In his article, Padfield (2004) states that "subsidiaries: separate entities with their own board of directors-that should mean something" (p.111).The importance is primarily connected to legal responsibility (Huse & Rindova, 2001) toward the parent company, creditors, minority shareholders, and government.Gouvin (1996) argues that "although subsidiaries play a significant role in our economy, surprisingly little has been written about the duties of their directors.Holding companies raise legal dilemmas for subsidiary directors that are easier to ignore than to resolve" (p.288).The subsidiary director's dilemma is well represented by Padfield (2004) when he writes: "Imagine, for example, that you are a director of a wholly-owned subsidiary.You are faced with a dilemma.You can either: serve the parent and risk being sued by creditors, regulators and/or other constituencies; serve the subsidiary and risk being sued and/or fired by the parent; or, you can quit.Where shall you look for guidance?"(p.80).
As we will discuss later, the subsidiary director's dilemma is particularly relevant in the case of non-wholly-owned subsidiaries, especially if they are listed, because of the conflict of interest that could arise from the agency problem between the directing holding company and minority shareholders of directed subsidiaries.Indeed, boards have a legal responsibility to protect owners' interests (Fama & Jensen, 1983;Jensen & Meckling, 1976).This role is important not only when the controlling shareholder (or the management) intends to extract private benefits but also when the subsidiary is managed by the parent in the interest of the group but against the interests of its shareholders.
We could expect that the outsiders of listed subsidiaries ask for a board that is autonomous from the parent mainly when the separation between ownership and control is elevated.Indeed, there are at least two perspectives through which to interpret the directing of the parent company over the subsidiaries: efficient and opportunistic.
The efficient perspective interprets the directing activity as favourable to the firm and its outsiders (Di Carlo, 2014a).The main assumption is based on transaction costs theory and assumes that the corporate hierarchy performs more efficiently than is possible through the market governance of transactions (Williamson, 1985).The opportunistic perspective is based on agency theory (Jensen & Meckling, 1976) and interprets the direction activity as potential tools of expropriation by insiders, i.e., managers, dominant shareholders, or both, to extract private benefits at the expense of the outsiders.Thus, in controlled listed subsidiaries with high separation between ownership and control the direction activity may be interpreted in a negative way since that separation is considered as a proxy to assess indirectly the degree of expropriation (Claessens et  In the case of ownership concentration, the dominant shareholder has the power and interest to contrast the managerial abuse, being the marginal benefits that come from the improved performance higher than the marginal costs of monitoring (La Porta et al., 1999).In Italy, the major shareholder generally controls the composition of the board of directors and influences the corporation's activities (Di Pietra et al., 2008).That situation gives a high probability of conducting actions that can be advantageous to the controlling shareholders but cause damage to minority ones through the extraction of private benefits of control (Type II agency problem).In Italy, the poor investor protection and the risk of expropriation by the dominant shareholder (e.g., through related party transactions) have been widely investigated by scholars (Melis, 2000;Volpin, 2002;Zingales, 2008), especially in complex group structures (Kirchmaier et al., 2009).

The interest of the whole group and the holding liabilities: Pierce the corporate veil
Normally the law treats parent companies and their subsidiaries as separate legal entities (Padfield, 2004).
Parent companies are not traditionally held liable for the debts or actions of their subsidiaries.However, under corporate law of some countries, a court may "pierce the corporate veil" of the parent if it finds an appearance of impropriety through questionable share transfers or other fraudulent means of avoiding the subsidiary's liabilities (Beaver et al., 2015;Erens et al., 2008).In other words, when the group is managed as a single economic entity, all the legal entities of a group may be treated as a single legal entity.
The Italian "Corporate Law Reform" (Legislative Decree No. 6/2003) has introduced some corporate rules for the regulation of business groups.These rules are related to the "management and coordination activity" (attività di direzione e coordinamento) exercised by the parent over its subsidiaries.
The headquarters-subsidiaries relationship is often based on the centralization of the decisionmaking power in the parent company, in order to manage the group as if it were a single economic entity and then to reach the system effect allowing subsidiaries to bring different benefits (e.g., scope and scale economies, financial synergies).
However, complying with the group's policy may have negative consequences for subsidiaries, for example, when the interest of the group conflicts with that of the single subsidiary and consequently with its stakeholders (e.g., minority shareholders and creditors).
Under Article 2497 of the Italian Civil Code, a specific liability of parent companies with directing and coordinating power (directing company) over their subsidiaries (directed legal entities) for damages incurred by the latter is expressly maintained.In particular, legal entities which, in carrying out management and coordination activities on other companies, act in the interest of their own or third parties business, in breach of the principle of fair corporate management, are directly liable: 1) with the minority shareholders of the directed and coordinated companies and 2) with the creditors of the same companies.Thus, the Italian group regulation pierces the corporate veil of the parent, treating the group as a single legal entity.
The Italian group regulation aims to discourage the activities imposed by the parent without considering the interests of the subsidiaries and their stakeholders.
This law is significant at least for two other reasons: it is based on the theory of compensatory advantages (Cariello, 2006;Denozza, 2000;Fasciani, 2007;Rossi et al., 2002), where the prejudicial impact of the parent company's decision may eventually be offset by the benefits arising from the directed companies' participation in the group structure; the subsidiaries board have the obligation to disclose if they are managed and coordinated by the parent company and the reasons that led to transactions with some particular types of related parties (e.g., transactions between two subsidiaries directed and coordinated by the parent company) (Di Carlo, 2014b).
Thus, because of the group regulation, the legal role of Italian subsidiary boards has an important impact on the decision-making power of the holding company and its subsidiaries.
Indeed, the subsidiary board members will be able to disregard the directives of the parent only if they consider that the latter acts in the interest of their own or third parties' business, in breach of the principle of correct corporate management.If board members carry out the detrimental directives of the parent they are jointly and severally liable with the parent for the damage caused to the outsider of the subsidiaries.Thus, the Italian group regulation allows the subsidiary board members to solve their dilemma (serve the parent or the subsidiary interest?).

The publicity of management and coordination activity within the Italian business groups
The Italian group regulation (Article 2497-bis of the Civil Code) requires that the directing activity be expressly indicated in the company's correspondence and official documents, including the notes to the financial statements and the management report of the directors.
That disclosure allows analysing the parentsubsidiary relationship in order to understand whether the controlling company also manages and coordinates its subsidiaries, regardless of the presence of directors and managers of the parent within the subsidiary board.
Figures 2 and 3 show the legal entities map of two Italian family-listed groups.The listed group controlled by De Benedetti family has a pyramidal structure (Figure 2) and operates in unrelated sectors.Cofide is the pure listed holding of De Benedetti group.CIR is a Cofide investment listed subsidiary.It is active in the energy sector, media, automotive components, healthcare, and financial investments (private equity, venture capital, non-performing loans, start-ups).
Sogefi is a CIR-listed subsidiary and one of the major international groups operating worldwide in the sector of automotive components.GEDI Gruppo Editoriale is also a CIR-listed subsidiary and it is one of the leading media groups in Italy with interests in publishing, radio, advertising, internet businesses, and television.Consolidated of Cofide group combines financial statements of the listed sub-holdings CIR, Sogefi, GEDI Gruppo Editoriale, and all of those of their subsidiaries, while that of CIR combines Sogefi and GEDI Gruppo Editoriale.Thus, the consolidated statements of the Cofide group contain the effect of the different industries and sectors where listed De Benedetti companies and their subsidiaries operate.Often researchers use the variable "industry" in their empirical study without considering this aspect.In other terms, the industry and sector are normally referred to the holding company while the financial performance represented in the consolidated is referred to the different industries in which the whole group operates.
In Figure 3, Fininvest is the non-listed holding company of the Berlusconi group and one of the world's largest communication groups that includes Mediaset, Medusa, a leading company in the cinema sector, and the publishing listed company Mondadori.Mediaset listed group is the first commercial broadcaster in Italy and one of the major media companies at the European level.Mondadori is Italy's biggest book and magazine publisher and the third-largest publisher of consumer magazines in France.Consolidated Fininvest combines the sub-consolidated of Mediaset and Mondadori and all their subsidiaries.De Benedetti and Berlusconi listed groups present different situations in terms of both the separation between ownership and control and between control and management.Indeed, all the listed sub-holdings of the Berlusconi group are not managed and coordinated by the parent company Fininvest, whereas in the De Benedetti group, all listed sub-holdings claim to be managed and coordinated by the respective parent.

RESEARCH DESIGN
This section discusses the methods, variables, and data collection employed to explore the phenomenon of the separation between control and management within business groups, giving some suggestions on the interpretation of the ownership structure, board composition and the financial performances of the directed listed subsidiaries.
Taking into consideration De Benedetti and Berlusconi groups (Figures 2 and 3), we present the results of an empirical investigation conducted on companies listed in the Italian Stock Exchange at the end of 2018.Because of a particular regulation provided for financial companies they were excluded from the initial sample.Indeed, compared to other companies, financial ones have to follow the Banca d'Italia regulation in terms of financial statements disclosure.
The initial sample for this study is composed of all the non-financial listed companies of Borsa Italiana Stock Exchange (164 firms).Table 1 shows the selection procedures.We also excluded all the 43 listed companies that are not controlled and consolidated by other companies.For instance, we did not include companies controlled directly by individuals, coalitions, and those widely held.According to the Italian group regulation, only for controlled and consolidated subsidiaries is due to the presumption of management and coordination activity of the parent (the Italian Civil Code, Article 2497-sexies).
Thus, the sample selected consists of all the non-financial companies listed on the Italian Stock Exchange, controlled and consolidated by other companies with different types of ultimate owners (e.g., state, family, coalitions).Since the selected listed companies control at least one entity they all present the sub-consolidated financial statements, and therefore they are all sub-holdings.
According to the Italian group regulation, the management and coordination activity by the parent is presumed in case of consolidation.In other terms when a parent controls a subsidiary it is presumed that the former exercises the direction activity over the latter.
The analysis is based on secondary data collected manually using the annual report of the listed companies, the Borsa Italiana (the Italian Stock Exchange), and Consob websites, all for the year ending December 31, 2018.As the analysed companies are listed, they all adopt the International Standards (IAS/IFRS) and, in particular, the same consolidation procedure (IFRS 10).
In particular, data on ownership structure were hand-collected from the Consob website.Companies' annual reports provided the necessary data to populate the information regarding the direction of activity.

RESULTS AND DISCUSSION
In terms of direction activity, the sample can be divided into three different groups (Table 2): the first is composed of listed companies that declare to be subjected to the direction activity of the parent (hereinafter, directed subsidiaries or directed sub-holdings) (Figure 1, Case B); the second group includes companies that claim to be not directed (hereinafter, non-directed subsidiaries or nondirected sub-holdings) (Figure 1, Case A); while the third group includes those companies that do not declare whether they are directed or not (hereinafter, non-declaring subsidiaries or nondeclaring sub-holdings).However, under Italian law, companies must provide the disclosure of the direction activity by the parent only if they are subjected to this direction.Thus, companies that do not declare anything are presumably not directed.Nevertheless, the third group is distinguished from the first one, since the absence of such disclosure can be interpreted as an index of low transparency.Despite the presumption of direction activity in the case of control, 71.1% of the sample declares to be not directed by the controlling company, whereas 21.5% declares to be directed and 9.1% does not declare anything.These first results seem to contradict studies that consider the group as if it were a single economic entity because of the managing activity centralized by the controlling parent company.
The following example contains the declaration of direction activity given by Mediaset SpA controlled and consolidated by the non-listed company Fininvest SpA (Figure 3).

"Mediaset SpA defines its own strategies independently and that it has total organisational, operational and transactional autonomy, not being subject to absolutely any directional or coordinating actions by Fininvest, regarding its own business activities. Specifically, Fininvest does not issue any directives to Mediaset nor does it carry out any technical, administrative or financial support or coordination activities on behalf of Mediaset and its subsidiaries" (Mediaset annual report).
The main reasons provided by the controlled companies in their annual reports on the fact of not being managed and coordinated by the parent can be grouped into the following categories: 1) presence of a number of independent directors capable of ensuring the independence of the board of directors; 2) autonomy of the board of directors from the parent; 3) lack of strategic management and coordination activity by the parent; 4) the management of the investments in subsidiaries by the parent has only a financial nature; 5) prohibition in the company bylaw of the parent to exercise management and coordination activity over the subsidiaries; 6) absence of formal rules or procedures, which may emerge a concrete management and coordination activity; 7) a simple declaration of not to be subjected to management and coordination activity by the parent.
Thus, the level of transparency provided in the financial statements ranges from a minimum in which the subsidiary declares simply not to be directed at a level where firms indicate, point by point, the reasons that lead to preclude the exercise of the management activity by the parent since it is presumed in case of consolidation.
Twenty-six listed firms are controlled by other listed companies.Therefore, these companies belong to groups with a pyramidal structure (e.g., Sogefi in De Benedetti group, see Figure 2).It signals a potentially high level of separation between ownership and control and a higher risk of expropriation by the controlling owner.It could be the reason why only 9 of them declare the exercise of direction activity.Indeed, the absence of direction activity reduces the perceived risk of opportunistic behavior.
This information is extremely important for the selection criteria used by scholars who want to explore the relationship between corporate performance and corporate governance.Even if the controlled listed company is an independent economic entity, since the listed parent does not exercise a directing activity over it, the consolidated financial statements of the latter depend on the financial statements of the former, since they are combined and thus strongly correlated.Thus, the consolidated listed company should be excluded from the sample.For instance, looking at Figure 2 if a researcher selects the listed company Cofide for his/her sample, he/she should exclude all the other listed companies (CIR, Sogefi, and GEDI Gruppo Editoriale) consolidated in the financial statements of Cofide group, even if these controlled sub-holdings were not directed.
Foreign parents control 20 listed Italian firms, but only 2 exercises a direction activity over them.The percentage of firms controlled by foreign parent companies is higher in firms that declare to be not directed.
Among directed companies, there are some highlighting in their annual report the ways in which management activity is carried out, even if the Italian law does not request this further information.Also, in this case, we grouped the given statements as following: 1) provision of services, such as human resources, cash flow management, real estate, finance, legal and tax; 2) advise on business strategies; 3) guidelines on the appointment of directors; 4) approval of related party transactions.From the list above emerges that directed companies interpret the management activity in various ways.However, the "management and coordination activity" by the parent over its subsidiary is regulated but not defined by Italian Law.Italian case law has identified "management" as the direction of the group as a whole under a unitary operational direction and "coordination" as a specific means of implementing single management by creating links between the management of all the group entities.Coordination generally refers to collaborative actions taken to achieve a unity of effort within the organization (Lawrence & Lorsch, 1967).Thus, the management and coordination activity (referred to within this article as "direction activity") consists of giving a unitary operational direction to different companies, by applying a common financial policy and strategy and managing them as a unique entity.
In directed companies, the responsibility for the financial results is not entirely attributable to their boards because they also depend, in a more or less relevant way, on the board of the parent.
This leads to a number of considerations that impact the studies carried out so far on the relationship between board composition and corporate financial performance.
Suppose to study the effect that board composition (as well as board processes) of the directed sub-holding Sogefi (Figure 2) could have on the financial performance of that firm.Board members could be the same of the parent CIR, so these directors may be carrying the interest of Sogefi as well as the interest of the parent CIR.
Indeed, in a business group, there are three different interests, expressed by the parent company, subsidiaries, and the whole group.
The interest of the parent and the subsidiaries is to create value for their shareholders.The problem arises when the holding company is tempted, for example, to use related party transactions and transfer pricing in order to divert resources from the subsidiaries in which its percentage of ownership (cash flow rights) is lower, towards those where it is higher.However, the parent could use intragroup transactions and transfer pricing not for the private interest of the controlling shareholder but rather for the interest of the whole group.
As argued by Gouvin (1996), "in many situations, the board of directors of the subsidiary corporation is not free to take action that is in the best interests of the subsidiary as a corporation, but instead must do as the parent corporation demands […] Therefore, the boards of subsidiaries often engage in activities that serve only the interests of the parent -even when corporate law imposes a duty on the directors to act in the interests of other parties" (pp.289-290).
The interest of the whole group is the equilibrium point, the centre of convergence between the interest of the parent and that of the subsidiaries (Gambino, 1993).In particular, the interest of the business group is to guarantee the so-called "system effect", so that the whole (i.e., the group) is higher than the sum of the parts (i.e., the single legal entities).
Moreover, some harmful transactions ordered by the parent may be accepted in the interest of the whole group by the board of the directed company, because the liabilities of the former and that of the latter are excluded when the damage is offset by compensatory advantages.
In this regard, the Italian Corporate Governance Code (2015) provides (Section 1.C.6) that "the decisions of each director are autonomous, to the extent he/she makes his/her choices with free judgement, doing so in the interest of the issuer and the generality of the shareholders.Therefore, even when management choices have been evaluated, addressed or otherwise influenced in advance, within the limits and in compliance with the applicable provisions of law, by those exercising management and coordination activities, or by subjects participating in a syndication agreement, each director shall pass resolutions in autonomy, adopting resolutions which may, reasonably leadprimarily -to the creation of value for the generality of the shareholders in the mediumlong term".
Therefore, in directed subsidiaries, it is appropriate to distinguish the concept of board independence from that of dependence on the parent company's directives.Even a directed board must maintain in any case an independent judgment, because of its legal role.
The Consob Regulation (No. 16191) provides specific rules for listed companies subjected to management and coordination by another Italian or foreign company with shares listed on regulated markets.For example, the directors of the parent company appointed as directors in the subsidiaries that are managed and coordinated or in listed subsidiaries cannot be qualified as independent directors.
This last point is particularly important and it is justified by the fact that the interest that directors should safeguard is not only that of the subsidiary in which they carry out their role but also that of the directing parent company as well as of the whole group.These two interests could conflict with each other.
This situation is represented in Figure 4, where minority shareholders (and more in general all the outsiders) of directed listed subsidiary "B" find their agents in the managers of directing listed company "A".In this case, the direction activity raises the director's dilemma in subsidiary "B" influencing the service and monitoring roles in the subsidiary board (Huse & Rindova, 2001).Subsidiary "B" has two principals (the holding and minority shareholders) that could have conflicting interests (e.g., when the holding company "A" orders detrimental transactions for subsidiary "B" in the interest of the group, damaging minority shareholders of "B").
The principal-principal conflict may affect, among other things, the processes and task performance of subsidiary boards (Forbes & Milliken, 1999;Pettigrew, 1992).
In Figure 5, minority shareholders (and more in general all the outsiders) of the non-directed listed subsidiary "B" find their agents in the managers of that company, since the parent "A" does not exercise the direction activity.According to the Consob Regulation, directors of a non-directed subsidiary can maintain the qualification of independence even if they are directors on the board of the non-directing parent.However, if the separation between control and management is apparent and not real (Di Carlo, 2014a), also, the independence of these directors is only apparent.
The role of independent directors of directed subsidiaries must necessarily be reinterpreted within the business group since they should protect the outsiders' interests in the case that the directing parent (the agent) asks the directed subsidiary to conclude detrimental transactions in the interest of the parent company or of the major shareholder.Instead, when the harmful actions are ordered in the interest of the group (e.g., in Figure 4, holding "A" orders to sub-holding "B" to sell services to non-listed companies controlled directly by "A" at a lower price than the market price), the independent directors of the damaged subsidiaries should evaluate the compensatory advantages that their directed firms receive because of the belonging to the business group.

CONCLUSION
In Italy, more than half of the listed companies are controlled by other entities, some of which are listed.These companies are, in turn, sub-holdings since they control other subsidiaries.Thus, researchers who focus the attention on the Italian stock market, or on other markets with the same characteristics, should be aware that they mainly collect data from sub-consolidated financial statements.The phenomenon is particularly relevant considering that some of those subsidiaries declare to be managed and coordinated by their parent companies.Thus, the number of independent economic entities listed in the Italian stock market does not coincide with the number of listed companies, being lower.
This article has several implications for academics, practitioners, and policy-makers.
Empirical studies that do not take into account the above aspects could be heavily biased.The bias can affect the interpretation of both dependent and independent variables (Judge, 2008).We refer primarily to the: accounting-based indicators (e.g., earnings per share, return on assets, return on equity); market-based indicators (e.g., market price, price to earnings); corporate governance variables (board demographics and board processes).
With regard to studies on board, it is not adequate to analyse the effect that board composition and processes may have had on the financial performance or on the stock prices of directed (not independent) companies, without taking into account the consequences generated by the management and coordination activity of the parent.Indeed, the responsibility for the financial results of directed companies might be only partially attributable to their boards, since they also hold the interests of the directing parent company and that of the group.
Sometimes, the links between directing and directed companies are so close (e.g., they operate in the same business, the majority of their boards are composed of the same directors) that in order to study the relationships between corporate governance and financial performance it is appropriate to analyse the consolidated financial statements of Group boundaries (Listed "A" and "B" are managed as a single economic entity)

Subsidiary board of directors
the whole group as well as the board of the first level consolidating parent.Moreover, even if some listed subsidiaries are independent from their listed parents, the consolidated financial statements of the latter depend on the financial statements of the former, since they are combined and thus strongly correlated.Thus, it seems necessary to exclude the consolidated financial statements of the independent listed companies, when selecting the financial data of their listed parents.
In addition, financial indicators may be strongly affected by transactions with related parties guided by the directing parent company, while the board of directors of the directed companies may face severe constraints from the board of the parent since the latter holds the interest of the whole group.
As provided by Italian law, directors of the directed company who sit as directors in the directing company may not be considered as independent, because they carry the interest of both companies.The specificity of the Italian regulation must be taken into consideration when comparing the Italian corporate governance system with that of other countries.
The results of our analysis have shown that Italian companies interpreted the direction activity in different ways.Some directed companies do not mention the type of direction exercised by the parent.In addition, non-directed companies often do not specify the reason why even if controlled they are not directed.Nevertheless, the Italian regulation does not require this disclosure.However, the disclosure of the management and coordination activity has been of extreme importance in the analysis of the performance and responsibilities of persons called to govern the companies subject to such direction activity, since the outsiders of directed companies find their agents in the directing parent company (see Figure 4).
It would be appropriate that regulators require a disclosure on the eventual directing activity of the parent.For example, it should be disclosed what type of directing activity is exercised and what are its effects on governance (e.g., the dependence of the subsidiary's board) and the financial performance of directed companies.In particular, the objective of the regulation should be to ensure that the entity's financial statements contain the disclosures necessary to draw attention to the possibility that its financial position and profit or loss may have been affected by the existence of direction activity by the parent.In other terms, it is important having the information to evaluate the degree of autonomy of the subsidiary boards and to understand how the controlling company directs its subsidiaries.The separation between control and direction, together with the separation between ownership and control, can be seen as a further indicator of the degree of expropriation.For instance, two pyramidal groups with the same separation between ownership and control show a different degree of expropriation depending on the direction activity by the parent.
The code of corporate governance of directed subsidiaries should provide clear evidence on how boards manage the conflict of interest with their minority shareholders and creditors since they are directed also to serve the interest of the parent and of the group.
The main limitation of this study derives from the use of the publicity given by the controlled subsidiaries relatively to the fact of being managed and coordinated.Thus, it exploits the Italian legislation requiring the subsidiaries to declare their autonomy or dependence from the parent.
However, it is plausible that some companies have found convenient to deny the subjection to the directing activity (e.g., to avoid the responsibility of the parent in case of damage to the outsiders of the directed company).In other words, for some companies, we could expect that the separation between control and direction may be just apparent, and thus the number of independent economic listed entities may be lower than we have considered in our analysis.
It would be worthwhile to investigate further the usefulness of the listed sub-consolidated financial statement in corporate governance research, analysing how the management activity by the parent and the legal responsibility of the subsidiary boards could influence the effectiveness of various corporate governance and disclosure variables.
It would also be interesting to investigate how different types of ultimate owners (e.g., State, individuals, banks) influence the direction activity of the parent.

Figure 1 .
Figure 1.Control without and with direction activity of the parent

Figure 2 .
Figure 2. De Benedetti business group company Ownership link Management and coordination link Boundaries of the consolidated financial statements 53.8% + 5%* 100%

Figure 4 .
Figure 4. Agents and principals in directed subsidiaries Group boundaries (Listed "A" and "B" are managed as a single economic entity)

Table 1 .
Sample size and selection procedures

Table 2 .
Characteristic of the controlling and consolidating company