Journal of World Business 52 (2017) 546–563

Contents lists available at ScienceDirect

Journal of World Business

journal homepage: www.elsevier.com/locate/jwb

Overcoming the liability of origin by doing no-harm: Emerging country

firms’ social irresponsibility as they go global
Davide Fiaschi, Elisa Giuliani* , Federica Nieri
Responsible Management & Development Lab, Department of Economics and Management, University of Pisa, via Ridolfi 10, 56124 Pisa, Italy

A R T I C L E I N F O A B S T R A C T

Article history:
Received 20 May 2015 As emerging (and developing) country firms internationalize, they often need to build legitimacy to
Received in revised form 1 September 2016 overcome home-country liabilities. We argue that international legitimacy is at risk if these firms do
Accepted 4 September 2016 harm in the conduct of their business, and we investigate the extent to which host countries’ speech and
Available online 27 September 2016 press freedoms influence corporate social irresponsibility (CSIR) for a sample of Multilatinas, observed
during the period 2003–2012. We do find evidence of lower CSIR among Multilatinas which have adopted
Keywords: explicit CSR policies and have higher levels of investment in countries characterized by strong speech and
Emerging country firms press freedoms.
Multilatinas
ã 2016 Elsevier Inc. All rights reserved.
Neo-institutional theory
Corporate social irresponsibility (CSIR)
Corporate social responsibility (CSR)
Host country speech and press freedom

1. Introduction (Duhigg & Barboza, 2012), and the South African AngloGold
Ashanti allegations about workers and indigenous rights violations
In 2012, the Brazilian mining company Vale, won the ‘Nobel in its home country and also Colombia, Ghana, and Tanzania3 ;
Prize of Shame’1 for its contribution to building the Belo Monte among others.
dam which will be the largest hydroelectric power plant in the Irresponsible conduct manifested through the violation of
Amazon rainforest. It is forecast that the dam will have a universal human rights4 may not be just due to simple accidental
devastating environmental impact, and will inflict forced reloca- or unintentional events. It frequently responds to firms’ rent-
tion of the local population to the proximate, large indigenous seeking and resource appropriation strategies (Giuliani, Santan-
community which constitutes an infringement of the local gelo, & Wettstein, 2016; Surroca, Tribo, & Zhara, 2013, among
population's human rights. Belo Monte is just one source of the others). For instance, violations of labor rights (e.g. child labor,
controversy currently surrounding Vale which, allegedly, is labor discrimination, union busting, etc.) and human trafficking
involved in irresponsible business conduct in other parts of Brazil, allow for efficiency gains in production; violations of local
as well as in Guatemala, Mozambique, and Peru.2 The case of Vale indigenous communities’ rights to land and to life often occur
is among the relatively well known cases of emerging country as a result of firms’ seeking access to mines or exploitation of other
firms involved in controversies over irresponsible business natural resources; violations of the right to health of local
practices or corporate social irresponsibility (CSIR). Other eminent communities can be due to poor maintenance of production
cases include the Foxconn scandal related to labor rights in China

3
For evidence on these cases, we rely on Sustainalytics (2014), as well as on other
* Corresponding author. sources (available by the authors).
4
E-mail addresses: davide.fi[email protected] (D. Fiaschi), [email protected] We conceptualize CSIR as human rights abuses, and similar to previous studies
(E. Giuliani), [email protected] (F. Nieri). (e.g. Ruggie, 2008), our reference is the 1948 Universal Declaration of Human Rights
1
The award is based on ‘Public Eye’ an online campaign organized by the Berne (UNDHR) and subsequent covenants and treaties including the International
Declaration and Greenpeace, which started as a critical counterpoint to the annual Covenant on Civil and Political Rights, and the International Covenant on Economic,
World Economic Forum (WEF) in Davos. Social, and Cultural Rights. Therefore, we define human rights as inalienable
2
Chronicles about the Belo Monte Dam are available at: http://amazonwatch.org/ fundamental rights to which a person is inherently entitled simply by virtue of
work/belo-monte-dam, last accessed 24 November 2015. For the other evidence, we being a human being. Note that our notion of CSIR as human rights abuses may not
draw on Sustainalytics’ firm-specific controversy reports (Sustainalytics, 2015) and capture irresponsible business conduct which does not generate infringement of a
analyses of reports by non governmental organizations (available upon request by person’s inherent rights – e.g. damages to wildlife, animals, or misconduct that does
the authors). not affect specific individuals.

http://dx.doi.org/10.1016/j.jwb.2016.09.001
1090-9516/ã 2016 Elsevier Inc. All rights reserved.
 D. Fiaschi et al. / Journal of World Business 52 (2017) 546–563 547

plants, or failure to invest in cleaner and more environmentally In contrast, we argue that EMNEs whose foreign investments
sustainable production processes (Giuliani & Macchi, 2014). are predominantly in countries characterized by weak speech and
Given the growing presence of emerging country firms’ in the press freedoms will be less exposed to the risk of de-legitimation
global economy (e.g. Cuervo-Cazurra & Ramamurti, 2014; Ram- because the relevant international constituencies are mostly
amurti & Singh, 2009), more and more evidence of their CSIR is unlikely to know about the firm’s irresponsible conduct. Hence,
being produced as a result of non-governmental organization we posit that managers in EMNEs whose investments are mainly in
(NGO) monitoring, and closer media scrutiny and reporting of countries characterized by weak speech and press freedom will
corporate misconduct.5 This evidence is a cause for concern, and face weaker pressures to avoid doing harm, and will be more
calls for more research to understand the conditions under which inclined to behave irresponsibly as part of their firms' rent-seeking
CSIR is more or less likely to take place. The heightened attention of strategy.
international organizations such as the United Nations (UN), the Our theoretical development is framed within the tradition of
Organisation for Economic Co-operation and Development neo-institutional theory (DiMaggio & Powell, 1983; Meyer &
(OECD), and the International Labor Organization (ILO), on the Rowan, 1977; Oliver, 1991) and its expansions in the context of
irresponsible business conduct from of firms, particularly that of MNEs (Kostova & Roth, 2002; Kostova, Roth, & Dacin, 2008; Marano
large multinational enterprises (MNEs) is a reflection of the & Kostova, 2015). We extend earlier research on the relationship
increasing evidence of such activity. between internationalization and CSIR (Strike, Gao, & Bansal,
Investigating CSIR in the context of emerging country MNEs 2006; Surroca et al., 2013) by offering a novel perspective on the
(EMNEs) is essential because these firms are notorious for country type of institutional pressures to which MNEs are exposed when
of origin liabilities due to the perceived poor institutional quality of they invest globally, and consider speech and press freedom as a
their home countries (Khanna & Palepu, 1997) which in the eyes of key construct explaining the processes of legitimacy building of
many international stakeholders translates into credibility and EMNEs suffering from the liability of origin. We defend our focus
legitimacy deficits for EMNEs (Madhok & Kayhani, 2012; Ram- on the grounds that legitimacy building processes are influenced
achandran & Pant, 2010). Hence, their involvement in a CSIR event strongly by the extent of media freedom to report on firms’
can be detrimental to their legitimacy building process in the misconduct, and that loss of legitimacy can be value-destroying for
expansion of operations to other countries (Jonsson, Greve, & firms.7
Fujiwara-Greve, 2012,Lin-Hi & Muller, 2013; Muller & Kraussl, We further enrich our argument by proposing that the
2011). This creates an incentive for EMNEs to ‘strategize’ on their relationship envisaged between internationalization and CSIR is
CSIR by adjusting it depending on the context where most of their moderated by certain firm-level characteristics. We posit that
foreign investments are undertaken. EMNEs' adoption of explicit CSR policies, understood as explicit
In this paper, we maintain that CSIR endangers EMNEs’ and voluntary corporate policies which “assume and articulate
legitimation strategies as they go global. We argue that this risk responsibility for some societal interests” (Matten & Moon, 2008;
of de-legitimation is higher in the case of EMNEs' investment in p. 409),8 is likely to be an important moderator: we expect that the
countries characterized by strong speech and press freedoms negative relationship between EMNEs’ host country speech and
(Cingranelli & Richards, 2010) where any evidence of CSIR is press freedoms and CSIR will be stronger for EMNEs with CSR
scrutinized, shared, and amplified by relevant constituencies (e.g. policies compared to other EMNEs. We justify this on the grounds
investors, suppliers, governments, consumers, etc.) Constituencies that CSR adopters are generally more closely monitored by the
in host countries characterized by speech and press freedoms have relevant constituencies than non-adopters (Ashforth & Gibbs,
access to relevant information and can act collectively to condemn 1990; Morsing & Schultz, 2006), and therefore their risk of de-
firms involved in allegations of human rights (e.g. by publicly legitimation in case of involvement in CSIR is higher, and it
discrediting the firm, organized campaigns to boycott the firms’ increases with greater investment in countries with strong speech
investments or products in the host countries, etc.) By most and press freedoms. Hence, in these cases we posit that managers
accounts, allegations of human rights abuses which receive global of EMNEs with CSR policies will hold an even tighter control over
resonance in the media, are among the “worst nightmares” of their firms’ irresponsible activities and will reduce their incidence
managers responsible for obtaining ‘social license’ for the firm's accordingly.
operations.6 This implies that managers of EMNEs whose foreign We address our research questions by focusing on a set of
investments are predominantly in countries characterized by Multilatinas from Brazil and Mexico (Casanova, 2009; Cuervo-
speech and press freedoms will try to ensure control of CSIR Cazurra, 2008) to analyze the factors that contribute to their
through initiatives that prevent harm from occurring and avoid any involvement in CSIR events. Multilatinas are an ideal setting to
risk to their firm’s legitimacy. investigate CSIR, for two reasons. First, the strong catholic roots of
many Latin American entrepreneurs have engendered a culture of
involvement of firms in societal matters with a view to solving
problems which their home country governments are failing to
5
Traditionally watchdog NGOs and the press have focused on advanced country address (e.g. poverty and other social welfare issues) (Fiaschi,
firms. Such cases as the negative impacts of Shell Oil in Nigeria (Idemudia, 2009), Giuliani, & Nieri, 2015; Griesse, 2007; Medeiros Peliano, Beghin, &
the complicity of Nike in child labor in the 1990s (Lund-Thomsen & Nadvi, 2010), the
Union Carbide (now Dow Chemical Co.) accident in Bhopal, India (Dinham &
Sarangi, 2002; Shrivastava, 1995), the recent scandal surrounding Monsanto’s
7
glyphosate spraying in Argentina (Lapegna, 2014) are emblematic examples. In this sense, other host countries’ institutional characteristics such as e.g. their
6
We quote here the words of one of the managers (whose name we keep governments’ capacity to ensure the rule of law, may not necessarily be more
confidential) involved in remedying the damage caused by the collapse of the effective in deterring firms from doing no harm – not least because ensuring justice
Samarco dam in 2015. This is one of Brazil’s worst environmental disasters, when a when international firms infringe human rights is fraught with difficulty, as
dam from an open-pit mining complex owned by Samarco burst, flooding a nearby documented by a great deal of business ethics and international law literatures
community, killing 19 and dumping illegal levels of arsenic, mercury and other (Kobrin, 2009).
8
poisoning metals into the river. Since Samarco is a joint venture between Brazil’s Note that explicit CSR policies may include a set of initiatives spanning
Vale and the Anglo-Australian BHP Billiton, the news on the event was given wide philanthropy, CSR reporting and greater accountability, adoption of principle-based
resonance not only in Brazil, but also in Australia (and elsewhere). While Samarco’s initiatives and socio-environmental certifications (Gilbert, Rasche, & Waddock,
responsibility for this event is still under investigation, this event put Samarco (and, 2011) often aligned to global socio-environmental codes of conduct and norms
indeed its two owners) under great pressure for the damage that this event has (Fiaschi et al., 2015; Marano, Tashman, & Kostova, 2016; Marquis & Qian, 2014;
created to their legitimacy, as well as to their capacity to raise investors’ money. Zheng et al., 2015).
548 D. Fiaschi et al. / Journal of World Business 52 (2017) 546–563

De Oliveira, 2002; Puppim de Oliveira, 2008). Second, over the last 2. Theory and hypotheses
25 years, Latin America has experienced a resurgence of social
movements and protest, possibly in reaction to military regimes 2.1. Background literature
and civil wars, and the radical neo-liberal policies to which the
region has been exposed (Silva, 2015). Some of the core strategies Neo-institutional theory suggests that to increase their chances
of these movements are focused on resisting the challenges of survival in the market firms seek legitimation within their own
imposed by global capitalism, and raising awareness of business- organizational field. They do this by adjusting to a set of external
related human rights violations, especially by MNEs and state- institutional pressures which ultimately contribute to these firms
owned firms (Stahler-Sholk & Vanden, 2011).9 For both these increased similarity (i.e. isomorphism). In a landmark paper,
reasons, we expect Multilatinas to be familiar with the need to DiMaggio and Powell (1983) suggest that firms may be subject to
prevent human rights infringements, and to be aware of the three mechanisms of isomorphic change: coercive isomorphism
accompanying risks of de-legitimation; this makes them an which refers to adaptation to formal and informal rules
interesting context for this research. (regulations, laws, codes of conduct, etc.); mimetic isomorphism
Consistent with our theoretical framework, we treat CSR and which refers to the firm's imitation of the most reputable firms, in
CSIR as separate conceptual and empirical constructs (in line with uncertain or ambiguous contexts; and to firms’ acceptance of
e.g. Chiu & Sharfman, 2016; Kang, Germann, & Grewal, 2016; Keig, professional-related normative pressures. The core idea is that
Brouthers, & Marshall 2015; Lin-Hi & Muller, 2013; Mattingly & firms adapt their behavior according to the existing institutional
Berman, 2006; Muller & Kraussl, 2011).10 We use two alternate forces, which may not necessarily make them more efficient or
concepts of CSR: (i) philanthropic initiatives, donations, and other more competitive, but grants them the legitimation needed to
strategies in favor of different types of stakeholders which we refer operate in a given environment (Meyer & Rowan, 1977).
to here as ‘social policies’, and (ii) reporting initiatives which refer Some scholars questioned the applicability of neo-institutional
to publication of CSR reports (or sections in annual reports theories as such, to the context of MNEs. Since these firms operate
referring to sustainability) that account for and communicate to in different organizational fields, they have to comply with
interested stakeholders the firm's social and environmental heterogeneous, and sometimes contradictory institutional pres-
commitment and impact (Carroll, 2008; Gilbert et al., 2011; Lu sures (Kostova & Roth, 2002), which means that legitimacy is not
& Abeysekera, in press). We understand CSIR as referring to achievable through isomorphism (Kostova et al., 2008). Such
involvement in specific controversies or events where the focal critiques have highlighted the need for a better understanding of
firm is alleged to have been involved in human rights abuses. how MNEs seek legitimation across different institutional environ-
Our study relies on an original firm-level dataset which ments (Kostova & Roth, 2002) and conceive MNEs not just as
provides information on 29 firms from Brazil and 15 firms from passive adaptive entities (Oliver, 1991) but as agents in their own
Mexico ranked by Forbes Global 2000 (2012 Edition) which are as right, which exploit institutions to their own advantage to alter the
the largest public companies in these countries. The analysis type and amount of their legitimacy (Kostova et al., 2008; Marquis
covers the period 2003–2012. Our regression analysis is based on & Raynard, 2015; Zimmerman & Zeitz, 2002).
an unbalanced panel model estimated within a random effects (RE) Legitimation is a particular problem for EMNEs. Their
model, controlling for endogeneity. We find that Multilatinas which internationalization efforts are hampered by the liability of
have adopted CSR policies display a decreasing involvement in CSIR foreignness (Hymer, 1976; Kostova & Zaheer, 1999; Zaheer, 1995)
events the more their investments are aimed at countries and, due to the backwardness and (to an extent) illegitimacy of
characterized by strong speech and press freedoms. We discuss their home countries, they also suffer from the liability of origin
our results in the context of neo-institutional theory and MNEs’ (Ramachandran & Pant, 2010) or liability of emergingness (Madhok
institutional strategies (Kostova & Roth, 2002; Kostova et al., 2008; & Kayhani, 2012), defined as: “a credibility and legitimacy deficit in
Marano & Kostova, 2015; Marquis & Raynard, 2015; Oliver, 1991), the eyes of host country stakeholders who [are] even more
and highlight some implications of our study both for CSR and CSIR circumspect due to inefficient or missing knowledge of foreign
research, and for Multilatinas’ internationalization and legitima- emerging market multinational firms, their quality and safety
tion strategies. standards, and the like” (Madhok & Kayhani, 2012; p. 31). Extant
The paper is organized as follows. Section 2 reviews the research shows that EMNEs increasingly see CSR policies a global
literature and develops the research hypotheses. Section 3 strategy to try to reduce their liability of origin (Fiaschi et al., 2015;
discusses the methodology used, and Section 4 presents the Marano et al., 2016; Zheng et al., 2015), and that the adoption of
results of the empirical analysis. Section 5 discusses the empirical these initiatives is used to demonstrate convergence towards
results and concludes the paper. accepted global (mostly Western) CSR standards, by strengthening
their moral dimension which is one of the attributes that adds to
their global reputation (Barnett, Jermier, & Lafferty, 2006;
Fombrun, 1996; Godfrey, 2005) and “generalized favorability”
(Lange, Lee, & Dai, 2011) through dialogue with, and positive
impacts on, different stakeholders.
Whereas engagement with CSR is part of many EMNEs’ global
legitimation strategies, we maintain here that EMNEs’ involve-
ment in CSIR may be accompanied by loss of reputation and
9
Examples include movements against glyphosate spraying in Argentina’s legitimacy – two different but related concepts (King and
soybean production (Lapegna, 2014); the indigenous movements in Ecuador; the
Whetten, 2008). Given that good reputation takes longer to
Zapatist movements in Mexico; the new social movements in Brazil and Bolivia
(Vanden, 2007), as well as several grassroots movements to fight against private build than to lose (Staw & Epstein, 2000; Zhang & Luo, 2013),
firms appropriation of natural resources (Kuecker, 2007), to cite but a few. doing harm can be a risky strategy for firms seeking international
10
Note that, in contrast to some of the earlier research, we do not consider CSIR legitimacy. In spite of this, very little research has investigated
and CSR as “opposite ends of a continuum” (Jones, Bowd, & Tench, 2009, p. 305). how CSIR relates to the internationalization of emerging country
Hence, we believe that failure to respect a negative duty (i.e. do no harm) cannot be
compensated by a positive duty or action in favor of the affected (or other)
firms. In the context of firm’s internationalization, moreover, only
constituencies. Therefore, CSR and CSIR remain two separate constructs requiring a few studies have examined CSIR. For instance, Strike et al.
two different measures. (2006) argue that international firms’ irresponsible behavior is
 D. Fiaschi et al. / Journal of World Business 52 (2017) 546–563 549

related to difficulties inherent to managing the increased business, may be exposed to the risk of banning, censorship and
complexity accompanying international diversification, and assassination.11 In such conditions, CSIR-related information will
suggest that “MNEs may act irresponsibly, not out of malice or not be reported or discussed publicly.
ill will, but because they have to stretch their resources and In this context, we expect that in countries with strong speech
capabilities in order to coordinate and monitor subsidiaries” and press freedoms, attributions of culpability (Lange & Washburn,
(Strike et al., 2006; p. 853). More recently, Surroca et al. (2013) 2012) for a CSIR event are more likely because individual observers
offered a somewhat more disenchanted perspective and sug- have greater access to relevant news and information. In turn,
gested that MNEs headquarters have control over CSIR and allow these attributions will have a negative influence on the firm's
their subsidiaries to act irresponsibly in a context of strong legitimacy. Thus, we argue that the more MNEs invest in countries
compliance with the institutional environment in the home with government guaranteed speech and press freedoms, the more
country and weak compliance in the host country. In so doing, likely these firms managers will exert tighter control over
Surroca and colleagues' study offers critical insights into MNEs' operations in order to minimize involvement in CSIR events12
agency role in the performance of institutional arbitrage through and reduce the risk of de-legitimation.
the relocation of CSIR practices world-wide to reduce the MNE's In contrast, MNEs whose operations are embedded mostly in
loss of reputation. This is a view that is also in line with most countries characterized by weak speech and press freedoms, will
current business ethics and international law accounts of have more leeway to enact rent seeking irresponsible business
corporate misconduct, which show that involvement of firms conduct without the risk of de-legitimation. While this logic might
in irresponsible business conduct is often not due to accidental apply to all MNEs regardless of their country of origin,
events or bad luck, but to deliberate strategic maneuvering by internationalizing EMNEs firms will have a greater incentive to
managers occupying different hierarchical positions in the MNE reduce their CSIR activity in order to overcome their liability of
(e.g. also Gond, Palazzo, & Basu, 2009; Pegg & Frynas, 2003; origin. Accordingly:
Wettstein, 2010).
Hypothesis 1. The more emerging country firms invest in
We follow this latter perspective about the ability of MNEs to
countries characterized by strong speech and press freedoms,
hold control on their own involvement in CSIR events to conjecture
the less their operations will be associated with CSIR events
that MNEs’ managers will strategically maneuver their socially
(due to their managers’ tighter control over CSIR).
irresponsible conduct depending on the speech and press free-
doms of the countries in which their operations are embedded
economically via their foreign direct investments (FDI). Next we 2.3. The moderating role of CSR in the relationship between
investigate whether the adoption of explicit CSR policies moder- internationalization and CSIR
ates the relationship between firms’ internationalization in
countries characterized by different speech and press freedom Although EMNEs may adjust their CSIR conduct depending on
stringency, and CSIR. We develop our research hypotheses as the host countries’ level of speech and press freedoms, this process
follows. may be complicated by the legitimation strategies pursued by
these firms. In parallel with attempts to control CSIR, EMNEs may
try to build global legitimacy by adopting a range of explicit CSR
2.2. Internationalization, host country speech and press freedoms, and policies (e.g. Fiaschi et al., 2015). Many emerging country firms are
CSIR familiar with the concept of social responsibility and have “deep-
rooted indigenous cultural traditions of philanthropy, business
Countries differ widely in the capacity and willingness of their ethics, and community embeddedness” (Visser, 2008; p. 481;
government to guarantee a free press and free speech (Cingranelli Matten & Moon, 2008).13 The recent increased internationalization
& Richards, 2010). Speech and press freedoms generally are of these countries’ largest industry players has further contributed
understood as people’s freedom to communicate and freedom of to the adoption of global CSR standards, and has greatly increased
expression through different media including the press, social social reporting (Fiaschi et al., 2015; Marano et al., 2016; Meyskens
media, spontaneous assemblies, etc. More specifically, free & Paul, 2010; Zheng et al., 2015).
speech means that government allows and protects the free A firm’s adoption of explicit CSR policies can play an important
circulation and open critical discussions of news and ideas, while role in moderating the relationship between host countries’ speech
a free press means the absence of government censorship on and press freedoms and CSIR: essentially, we envisage that the
media outlets. We contend that these cross-country differences negative impact of host countries’ speech and press freedoms on
are likely to have implications for whether and how news on CSIR CSIR is strengthened when firms adopt CSR. This is because CSR
is circulated, critically analyzed, and given resonance by the adopters, which are generally exposed to greater scrutiny than
relevant constituencies in various countries. Thus, reporting on non-adopters due to their ethical and social ambitions becoming
CSIR is likely to be more widespread in countries with stronger more evident and attracting critical attention from the relevant
speech and press freedoms because the media will be free to constituencies (Ashforth & Gibbs, 1990; Morsing & Schultz, 2006),
report such news – irrespective of the location of the CSIR event in
the home country or abroad – with the result that information on
CSIR events is available to the general public. Free speech means 11
Freedom House is an independent U.S. watchdog organization, founded in 1941,
also that people are free to debate openly on CSIR-related news and dedicated to the expansion of freedom and democracy around the world. See,
and contribute to a collective awareness of these issues through https://freedomhouse.org/report/freedom-press/freedom-press-2016, last
accessed 9 August 2016.
various channels such as social media, public speeches (at 12
While observing intra-firm mechanisms for avoiding human rights abuses is
universities, unions, and other relevant fora), etc. In contrast, beyond the scope of this paper, we envisage this can be done through more active
poorly policed speech and press freedoms by government can management support to detect questionable practices, through human rights
make it dangerous for journalists and other constituencies to impact assessment exercises and due diligences, as well as via goal-setting systems
circulate CSIR-related information. For instance, according to that create an environment in which misdeeds are less likely occur.
13
For instance, accounts of CSR activities in Brazil and Mexico indicate that many
Freedom House (2016), journalists reporting on land grabbing,
firms historically have taken responsibility– often via philanthropic initiatives – for
environmental degradation related to business operations – satisfying societal needs such as poverty and other social welfare issues (Logsdon,
particularly in the case of extractive industries, and corruption in Thomas, & Van Buren, 2006; Raufflet, 2008; Weyzig, 2006).
550 D. Fiaschi et al. / Journal of World Business 52 (2017) 546–563

will be subject to even closer monitoring and requests for and as a cross-check, Sustainalytics an independent Environmen-
accountability (Zhang & Luo, 2013) if they invest in countries tal, Social and Governance (ESG) research and analysis data
with strong speech and press freedoms. In such contexts, CSR provider. We held a focus group16 to obtain recommendations on
adopters expose themselves to a heightened risk of de-legitima- data sources; BHRRC data were preferred for their more extended
tion if they behave irresponsibly,14 which strengthens managers’ longitudinal coverage compared to other sources. The BHRRC is
motivations to ensure that no harm will be caused in the conduct of considered the world’s leading independent information hub
business. This leads to the following hypothesis: providing data on the positive and negative impacts exerted by
corporations on human rights. It has offices in London and New
Hypothesis 2. For emerging country firms which have adopted
York and relies on the efforts of regional researchers based in
explicit CSR policies, the negative relationship between (i) their
Africa, Asia, Eastern Europe, and Latin America, who interact with
investments towards countries characterized strong speech and
local NGOs and gather information in the field. The BHRRC
press freedoms and (ii) CSIR, will be strengthened.
database covers the social and environmental impacts of over 5000
companies, operating in over 180 countries. Taking international
3. Methodology human rights standards as its starting point, the topics covered
include discrimination, environment, poverty and development,
3.1. Sample labor, access to medicines, health and safety, security, and trade.
BHRRC researchers collect news and reports relating to business
Our sample comprises a total of 44 firms, 29 from Brazil and 15 and human rights from the web and other sources, on a day-to-day
from Mexico, ranked by Forbes Global 2000 (2012 Edition) as the basis, paying attention to sources across the world including local
largest public companies in their respective countries. We chose to newspapers and reports produced by large and small NGOs. News,
study Brazilian and Mexican firms for two main reasons. First, most reports, and events focusing on the relations between the activities
Multilatinas' parent firms are located in one of these two countries of companies and human rights are examined, and subject to a
(Cuervo-Cazurra, 2008) – Brazil and Mexico account for a minimum criterion of credibility (therefore excluding blind attacks
significant portion of FDI outflows from the Latin American region on companies) are published on the BHRRC website. They highlight
(Brenes, Camacho, Ciravegna, & Pichardo, 2016; UNCTAD, 2014).15 the impact of business on human beings (news on the protection of
Second, both Mexico and Brazil are rent seeking economies which an endangered species but with no clear connection to an impact
often are considered to suffer from crony capitalism, endemic on human rights is generally not published).
corruption (Leahy, 2016; The Economist, 2014; see also, e.g. Dos We used this information source to search for alleged human
Santos & da Costa, 2014; Ferraz & Finan, 2011, p. 1281; Mendes & rights abuses connected to the firms in our sample. It resulted in
Junior, 2012; Oliva, 2015; Power & Taylor, 2011; Reid, 2014), over three hundred documents including news and reports
property related crimes (BenYishay & Pearlman, 2014; The providing evidence of “events” of negative human rights impacts.
Economist, 2015), and involvement in drug trafficking (Duncan, To cross-check our data, we relied on Sustainalytics as a second
2013; Morris, 2013), etc. (see e.g. Chew Sánchez, 2014 on state- data source. Sustainalytics has substantial experience in evaluating
terrorism in Mexico). For these reasons, internationalizing firms the ESG performance of publicly-traded companies (see Surroca
from Brazil and Mexico are likely to suffer from liability of origin et al., 2013 for a study using these data).17 Sustainalytics analysts
problems, similarly to other firms from emerging countries. This is draft company profiles detailing firm performance on various ESG
likely to act as a strong incentive to build legitimacy in the indicators, and firms' involvement in relevant controversies. For
international market, making these two countries an ideal setting the present study, we use Sustainalytics ‘Controversy Reports’
for this research. which provide a comprehensive understanding of companies’
Our analysis covers the period 2003–2012, and relies on 394 involvement in controversies over different types of human rights
firm-year observations. The focus on large public firms is justified violations.
by their being powerful and visible actors, whose CSIR tends to be Based on BHRRC and Sustainalytics information we identify
monitored by the press and NGOs more frequently than that of events involving the firms in our sample, related to different types
smaller companies. Large firms also possess the resources and of abuses of human rights (or CSIR events). We codified the
capacity to invest in CSR policies, and to internationalize. Our information on individual CSIR events into a dataset which
sample covers firms in the metals and mining (23%), banking (23%), includes the following items:
electricity and other utilities (14%), food and beverages (11%),
telecommunication (TLC) (7%) sectors. The remaining 22% of firms (a) a unique code for each separate CSIR event reported by our data
are from the aerospace, chemicals and pharmaceuticals, cosmetics, sources (hereafter “event”);
heavy industry, pulp and paper, real estate, and retail sectors (see (b) a firm level code associated with each event code;
Appendix A for the firms in our sample and their respective (c) a brief description of the event – e.g. “company resettled 717
industries). households without due consultation, in order to develop a
mine in Mozambique's Cateje, Moatize district”;
3.2. Variables

3.2.1. Dependent variable: CSIR 16

The focus group was organized as a semi-structured group interview process
To collect information and measure CSIR we rely on two sources and took place in 2008 at Scuola Superiore Sant’Anna (SSSA) in Pisa, Italy. Two of the
of data: the Business and Human Rights Resource Centre (BHRRC), authors of this paper acted as moderators and took notes of the conversations. The
focus group included three business and human rights experts – respectively a
practitioner in the field with extensive experience in business and human rights in
the context of Latin America; an international law specialist, and a political scientist,
14
Instances where firms adopt CSR policies and contemporaneously they are both academic scholars affiliated to SSSA. The interview lasted approximately three
involved in CSIR events are well known to neo-institutional theorists, who refer to hours and one of the topics discussed was sources of information on violations of
them as cases of symbolic or ceremonial CSR adoption (e.g. Marquis & Qian, 2014). human rights by the business sector.
17
In contrast, CSR is adopted substantively if it is not associated with CSIR. See Sustainalytics at: http://www.sustainalytics.com/, last accessed April 28,
15
According to the America Economia 2015 ranking, nearly 60% of the 100 largest 2015. Sustainalytics has its headquarters in Amsterdam, and has offices in Boston,
Multilatinas in 2015 are either Brazilian or Mexican (see http://rankings. Bucharest, Frankfurt, London, Paris, Singapore, Timisoara, and Toronto, and
americaeconomia.com/multilatinas-2015/, retrieved 6 August 2016). representatives in Bogotá, Brussels, Copenhagen, New York City, and San Francisco.
 D. Fiaschi et al. / Journal of World Business 52 (2017) 546–563 551

Table 1
Cumulative number of CSIR events by type of human rights abuses.

Brazil Mexico Examples

Abuses against local 24 7 Displacement of indigenous communities without consent, contamination of local environment, repression of the
communities protests of local communities
Abuses against workers 66 23 Anti-union practices, gender discrimination, worker fatalities, slave labor, child labor
Other abuses 1 0 Clients’ intoxication
Total: 91 30

Fig. 1. Total number of CSIR events, by year.

Source: Authors’ own elaboration based on BHRRC and Sustainalytics.

(d) year(s) in which the event took place, including for each event, event, we limited our analysis to year 2012. In addition, since data
the year in which the event is known to have started and the on emerging country internationalization through greenfield
year in which it is considered to have ceased; investments are available only from 2003 (see below), our time
(e) the year in which the event was first denounced or reported; span is 2003 to 2012.18
(f) a unique code for the document(s) reporting evidence of the Our dependent variable CSIR is the number of CSIR events in
abuse(s). These document(s) contain full news or reports of the which the firm is involved at time t. Each CSIR event captures a
abuse(s), are stored separately, and are available for consulta- different type of human right violation in which the firm has been
tion. reported as being involved in each year (e.g. if at time t, the firm is
found abusing labor rights in one of its production plants, and in
Two different coders read the BHRRC and Sustainalytics the same year, there is evidence of it violating indigenous
material and checked it for its inclusion, accuracy and complete- communities’ right to land, the CSIR value for this particular firm
ness in the dataset. Note that our data do not include events where at time t would be 2). Each single event is counted yearly as 1,
the abuse manifestly was unrelated to firm-level decision making whether it occurs in one particular year only, or extends across
or operations. For instance, cases where an employee deliberately more than a year (e.g. a firm poisoning the environment and
causes the death of other employees for personal or otherwise violating the right to health of local residents over several years). In
private motivations are not included as CSIR events. Similarly, our this case, we count this multi-year event as 1 for each year in which
dataset does not include cases of accidents related to an individual it has allegedly occurred.
employee’s lack of diligence, or to a natural disaster, unless there is Table 1 presents evidence of the cumulative number of CSIR
evidence that the accident was due to e.g. the firm’s lack of plant events for our sample firms, disaggregated by type. Fig. 1 depicts
maintenance, or other forms of complicity in the irresponsible the number of CSIR events reported in any year between 2003 and
event. Moreover, our document coding refers only to irresponsible 2012. The growing trend observed is likely due to an increase over
events which caused harm to human rights, irrespective of time of both media coverage and NGO reports of abuses. Thus, we
whether a given behavior was believed by the perpetrator to be include time dummies in the estimations. Descriptive statistics for
in line with its own duties or with profit-maximizing objectives CSIR are reported in Table 2.
(Armstrong, 1977).
Once the information was codified in the dataset, a business and
human rights expert (see footnote 16) checked the events to ensure
there were no errors or ambiguities in the events reported, and to 18
Data on emerging country internationalization through greenfield investments
check accurate coding of abuses. Data on CSIR events were are available only from 2003 onwards, which influences our sample. However,
collected for 1990–2014. However, due to an estimated time lag of growth in FDI from emerging countries is observed in the past decade, which makes
two years from a reporting of an onset or ending of a particular CSIR our observed cohort meaningful (UNCTAD, 2013).
552 D. Fiaschi et al. / Journal of World Business 52 (2017) 546–563

Table 2
Descriptive statistics.

Variable Min Max Mean (Proportion) Standard error

Dependent Variable
CSIR 0 10 0.288 0.997

Independent Variables
HCSPF 0 3 1.310 1.230
Social Policies 0 1 0.831 0.367
CSR Report 0 1 0.686 0.459
Social Policies * HCSPF 0 3 1.100 1.211
CSR Report * HCSPF 0 3 0.987 1.185

Other Variables
Size 3.090 12.110 9.730 1.446
Age 0 5.320 3.690 1.009
SOEs 0 1 0.143 0.361
ROE 0.677 8.763 0.242 0.624
Risk 0.009 2.239 0.221 0.445
Internationalization 0 24 2.680 4.616
CSR Experience Social Policies 0 23 8.410 7.005
CSR Experience CSR Report 0 16 4.280 4.050
Media Exposure 0 0.033 0.001 0.012
Industry dummy I 0 1 0.217 0.412
Industry dummy II 0 1 0.274 0.446
Industry dummy III 0 1 0.486 0.500
Brazil 0 1 0.659 0.463
Mexico 0 1 0.341 0.463
IV- Social Policies 0.500 0.966 0.831 0.136
IV- CSR Report 0.429 0.933 0.686 0.166

3.2.2. Independent variables that in our sample, firms from Mexico invest in countries
Host country speech and press freedoms (HCSPF) characterized by higher press freedoms, while Brazil’s investments
HCSPF measures the extent to which the firm’s FDI portfolio tend to be concentrated in countries with poor press freedom
includes countries with high (or low) protection of political rights records although over time the gap reduces. This difference is
to a free press and free speech and it is used to test Hypothesis 1. To explained at least in part by Mexican firms’ relatively higher
measure firms’ FDI, we rely on three complementary data sources. propensity to undertake direct investments especially M&As, in
We use FDIMarkets for data on greenfield and brownfield FDI, and signatory countries to the North America Free Trade Agreement
the Zephyr (Bureau van Dijk) and SDC Platinum (Thomson Reuters) (NAFTA) – i.e. Canada and the U.S.21 Descriptive statistics for this
databases for information on mergers and acquisitions (M&A) (we variable are presented in Table 2.
consider only majority and full stake M&A). To measure the extent
to which free speech and a free press are affected by government Corporate social responsibility
censorship including ownership of media outlets, we rely on the We consider two alternate measures of explicit CSR: Social
Cingranelli and Richards (2010) Freedom of Speech and Press Policies and CSR Reports, measured as follows:
indicator (CIRI_SP).19 In a given year a score of 1 indicates total
government censorship of the media; 2 indicates some govern-  Social Policies: refers to the firm’s adoption of CSR policies
ment censorship of the media; and 3 indicates no government conceived as discretionary activities favoring different stake-
censorship of the media.20 The index HCSPF for firm i at time t is holders, including donations and philanthropic initiatives for
defined as follows: society. We are interested in the year when the policy was
Xt XJ formally introduced at the corporate level. We retrieved this
s¼1
P  CIRI SPjs
j¼1 ijs information from direct contacts with corporations, and corpo-
HCSPF it ¼ Xt XJ i ¼ 1; :::; I; t ¼ 1; :::; T;
rate websites and CSR-specific web-pages described as “Social
P ijs
s¼1 j¼1 Corporate Responsibility”, or variations such as “Social Respon-
where Pijs is country j in which firm i invested at period s, and sibility”, “Corporate Responsibility and Sustainability”, etc. Social
CIRI SP is the value of the CIRI_SP indicator for each host country j Policies is a dummy variable that takes the value 1 if the firm has
t J undertaken any kind of social policy at time t, and 0 otherwise.
at period s; Ss¼1 Sj¼1 Pijs is the number of countries in which firm i  CSR Reports: refers to the firm’s preparation and publication on
invested up to time t. The index ranges from 0 if there are no its website of a CSR report, or an annual report with a dedicated
foreign investments, to a maximum of 3 if all investments are in CSR section. We downloaded and checked the reports. The
countries with (strong) speech and press freedoms. Fig. 2 shows variable CSR Report takes the value 1 if the firm produced a CSR
report at time t, and 0 otherwise.

19
Data are from the Cingranelli-Richards Human Rights Data Project (http://
www.humanrightsdata.org). Information for coding the Freedom of Speech and
21
Press indicator is from Cingranelli and Richards (2014, pp. 27–29). For instance, over the observed period, direct investments by the Mexican firms
20
Due to the need to distinguish between firms with no foreign investments at in our sample in Canada and the U.S. account for more than half of Mexico's total
time t, and those investing in countries with low levels of press freedom, we foreign investments, while Canada and the U.S. account for only about a third of
recoded the index which originally went from 0 to 2. overall Brazilian foreign investments.
 D. Fiaschi et al. / Journal of World Business 52 (2017) 546–563 553

Fig. 2. Speech and press freedoms of host countries, disaggregated by country of origin.
Source: Authors’ own elaboration.

Descriptive statistics of these variables are presented in Table 2. dispersion measured as the total number of foreign countries in
Fig. 3 depicts the share of the firms in our sample that adopted which the firm has subsidiaries, used in earlier studies (Strike et al.,
either of the CSR policies discussed above, in any year in the period 2006). Because firms with CSR experience might have accumulated
2003 to 2012. As expected, Social Policies are more frequent among more skills to hold control on CSIR, we control for the cumulative
our sample firms. Notice also that the importance of CSR reporting number of years in which the firm had a given CSR policy up to year
increases over time, with almost all firms in the sample publishing t (CSR Experience). Moreover, since the likelihood that the firm is
a CSR report in 2012. associated with a CSIR event depends also on the extent to which it
is on the media and NGO radar, we control for Media Exposure,
Interaction which proxies for the company's visibility in the global and local
To test Hypothesis 2, we use the interaction term between media (Marquis & Qian, 2014). We use Lexis Nexis (News section)
HCSPF and the two CSR variables (Social Policies and CSR Reports). as the data source for this variable which is computed as the log of
Descriptive statistics are reported in Table 2. the ratio between the number of news items/articles mentioning
firm i at time t, and the total number of articles mentioning any of
Control variables our sample firms at time t.
We also include a set of control variables in the analysis (see We use industry dummies to control for industry-related
Table 2), to account for factors that might explain firms' specificities in connection with CSIR. We draw on Giuliani and
involvement in CSIR events, based on earlier research. Among Macchi (2014) to identify groups of industries that are relatively
the firm-level controls, we include firm size (Size) proxied by the homogeneous in terms of their firms’ risk of being involved in
log of the number of workers in each year. We control for size human rights abuses (e.g. oil extraction involves a process that
because larger firms may have more operations and activities to generates negative impacts on society; other industries can be
govern which may increase the chances of being involved in organized in ways that make firms more/less prone to involvement
irresponsible behavior (Strike et al., 2006). We control also for age in abuses of human rights). The reference group (Industry dummy I)
using the log of firm age (Age), and ownership status (SOEs) – i.e. include firms in the extractive industries (Oil, Mining and Steel),
state owned (coded 1) versus private ownership (coded 0), because the second group (Industry dummy II) includes Retail, Banking,
both characteristics might influence the propensity for substantive Chemicals and Pharmaceuticals, and the third group (Industry
socially responsible conduct (Marquis & Qian, 2014). We control for dummy III) includes Cosmetics, Pulp and Paper, Aerospace, Heavy
firm performance and risk since they are likely to influence Industry, Telecommunications (TLC), Food and Beverages, Electric-
managers’ rent-seeking strategies. To measure firm performance ity and other Utilities, Real Estate. Based on available evidence,
we use return on equity (ROE at t-1), measured as the ratio between Industry dummy II and Industry dummy III firms are expected to be
firm i's net income and equity at t-1; firm’s risk (Risk) is measured less involved in CSIR events compared to the reference group.
on the basis of ROE volatility (i.e. based on annual fluctuations in Finally, we control also for country-specificities (country
ROE around its trend value, calculated using non-parametric dummies, Brazil being our reference group) since each country
estimation). We rely on Datastream for these data. has a different history and different regulation and internal
To control for firms’ internationalization we developed a institutional arrangements which might result in different
quantitative indicator to measure the number of different valuation of human rights and ethics (Matten & Moon, 2008).
countries in which the firm invested (in the form of greenfield, Since the number of reported CSIR events may increase over time
brownfield, majority, or full stake M&A) in each year in the period due to the expected increased availability of information on their
2003–2012 (Internationalization), based on FDIMarkets, Zephyr occurrence, we include time dummies (Time dummies) in the
and SDC Platinum data. This measure is similar to international analysis.
554 D. Fiaschi et al. / Journal of World Business 52 (2017) 546–563

Fig. 3. Adoption of different types of CSR policies.

Source: Authors’ own elaboration.

3.3. Econometric methodology CSIRit = ai + a1HCSPFit + a2Social Policies (CSR Report)it + a3Sizeit +
a4Ageit+ a5SOEit + a6ROEit–1 + a7Riskit + a8Internationalizationit +
There are two critical methodological issues related to our a9Social Policies(CSR Report)Experienceit + a10Media Exposureit +
estimations. First, the presence of unobserved heterogeneity at a11Industry dummiesi + a12Home country dummiesi + a13Time dum-
firm level which our controls cannot fully account for (which miest + mit (1)
suggests use of panel data estimates).22 Second, suspected
endogeneity of Social Policies and CSR Report (and thus interaction
with HCSPF), since the factors that lead firms to adopt CSR policies CSIRit = ai + a1HCSPFit + a2Social Policies (CSR
may be the same factors that induce them to behave irresponsibly Report)it + a3Social Policies (CSR Report)it * HCSPFit + a4Sizeit + a5Ag-
(Fiaschi et al., 2015; Kotchen & Moon, 2012). Fig. 4 provides indirect eit + a6SOEit + a7ROEit–1 + a8Riskit + a9Internationalizationit + a10 So-
support for our suspect of the presence of unobserved factors cial Policies(CSR Report)Experienceit + a11Media Exposureit +
positively affecting both CSR and CSIR. a12Industry dummiesi + a13Home country dummiesi + a14Time dum-
We run tests to account for both these problems. First, we run miest + mit (2)
the endogeneity test following Wooldridge (2010, pp. 352, 358) to
Appendix B reports the results of first stage estimates. Table B1
assess whether we need to apply an instrumental variable (IV)
shows the first stage estimates taking Social Policies (Models 1–2)
approach. Second, we run a Hausman test to check whether the
and CSR Report (Models 3–4) as the dependent variables. The
unobserved heterogeneity is uncorrelated with or is dependent on
coefficients for IV-Social Policies and IV-CSR Report are positive and
the observed explanatory variables, in order to decide about use of
significant (b = 0.80, p < 0.01 in Model 1, and b = 1.03, p < 0.01 in
a random effects (RE) or a fixed effects (FE) model (Wooldridge,
Model 2; b = 0.78, p < 0.01 in Model 3 and b = 1.07, p < 0.01 in
2010, p. 328). Based on the endogeneity and Hausman tests, a RE-IV
Model 4). The marginal effect of HCSPF on adoption of a CSR policy
panel two-stage least squares (RE2SLS) seems the most appropri-
on average is positive, based on the sum of the positive direct effect
ate estimation method.
(b = 0.15 and b = 0.17 in Models 2 and 4 respectively) and a negative
The IVs for CSR (Social Policies, CSR Report) for firm i are given by
effect of the interaction, which on average, is lower than the former
the share of firms in our dataset belonging to the same country as
effect (although the effect of CSR Report is stronger). Table B2
firm i which adopted the specific CSR policy (IV-Social Policies, IV-
presents the first step estimates for the interaction between the
CSR Report). The IV for the interaction between Social Policies (or
two CSR variables and HCSPF: the instrumental variables IV Social
CSR Report) and HCSPF are built following Wooldridge (2010, p.
Policies * HCSPF and IV CSR Report * HCSPF are positive and
122), by multiplying IV-Social Policies and IV-CSR Report, for HCSPF
significant (respectively b = 0.90, p < 0.01 in Model 1, and b = 0.66,
respectively. IV-Social Policies, IV-CSR Report and their interactions
p < 0.01 in Model 2). Based on these results, we can say that our IVs
with HCSPF appear to satisfy the two conditions for a good IV:
are not weak (Wooldridge, 2010, p. 108).
conceptually, they are strongly related respectively to CSR and to
Note that the correlations are presented in Appendix C. To
the interactions between HCSPF and CSR variables, and they are
investigate the potential multicollinearity problem we computed
weakly related to CSIR by firm i.
variance inflation factors (VIFs). The maximum VIF obtained in any
We estimate the following panel models (second stage):
of the full models is 4.59 (the interaction term between HCSPF and
Social Policies); mean VIF is 1.84 in the model with CSR Reportas
explanatory variable, and 2 in the model with Social Policies as the
explanatory variable, which are below the rule-of-thumb cutoff of
10. Therefore, multicollinearity is unlikely to affect our estimates
22
Tests for the absence of unobserved heterogeneity are rejected in all our significantly. In all the estimates p-values are calculated on the
regressions (Wooldridge, 2010, p. 299).
 D. Fiaschi et al. / Journal of World Business 52 (2017) 546–563 555

Fig. 4. The relationship between CSR policies and CSIR.

Note: Each point in the figure represents the combination of CSIR (average number of CSIR event per firm)
and CSR policy (share of firms in the sample having adopted that particular CSR policy) in each year.
Source: Authors’ own elaboration.

basis of robust standard errors given the presence of serial and average, a decrease of 1.11 in the number of their alleged CSIR
cross-sectional correlation (we tested for this, see Wooldridge, events. However, adoption of CSR reporting (CSR Report) of non
2010, p. 320), and heteroskedasticity (see Wooldridge, 2010, p. internationalizing firms does not significantly affect the number of
172). All the models report statistics of the adjusted R2, F-statistics a firm's alleged CSIR events.
and the serial correlation test. All the estimates are made in R and Since the results suggest that Models 1 and 3 are poorly
performed using the routines provided by various packages but specified if some firm-level characteristics such as the adoption of
primarily plm (Croissant & Millo, 2008). CSR policies are not taken into account jointly with the variable of
interest, the following comments refer only to the results of
4. Empirical results Models 2 and 4. Among the control variables, the coefficient of
Internationalization is positive and statistically significant: inves-
4.1. Regression results ting in one additional foreign country increases the predicted
number of CSIR events by 0.10 (Model 2) or 0.14 (Model 4) which is
Table 3 shows the results of the econometric estimations. In in line with previous research (Strike et al., 2006). As expected,
Models 1 and 3 the coefficient of HCSPF is not significant (at the increased media exposure is associated to greater involvement in
usual significance level) which does not support Hypothesis 1. CSIR events, confirming the importance of controlling for this
However, when we interact HCSPF with the two CSR variables, the variable when estimating these kinds of models (Marquis & Qian,
estimates change quite radically. In Models 2 and 4 the coefficients 2014). In particular, we observe that an increase of 1% in the firm's
of the interaction term are negative at the 1% significance level. The exposure to the media, involves an expected increase of 0.62
negative sign of these coefficients supports Hypothesis 2. In (Model 2) or 0.41 (in Model 4) in the number of its CSIR events.
particular, Figs. 5 and 6 report the relationships (and their relative We observe also that a 1 unit increase in the firm’s risk implies
standard deviations) between HCSPF and the number of CSIR involvement in a smaller number of CSIR events (0.47 and 0.31
events for CSR adopters and non adopters: the patterns are in Models 2 and 4 respectively). This finding is in line with the idea
consistent with our prediction of a moderating effect of CSR in the that risky operations tend to lead firms proactively to reduce their
relationship between internationalization and CSIR. Our data risk by engaging in less irresponsible business conduct. The
suggest that for adopters of Social Policies the predicted marginal coefficient of Age is positive and significant only in Model 2,
impact on the number of CSIR events ranges from – 0.22 (when the suggesting that older firms are involved in more CSIR events: an
investment is in countries characterized by low speech and press 100% increment in firm’s age (i.e. doubling the firms’ age) implies
freedoms) to – 0.66 (when the investment is in countries an increase of 0.2 in the expected number of its CSIR events.
characterized by strong speech and press freedoms) (see Fig. 5). Finally, our results show that Mexican firms are (on average)
A similar pattern is observed for firms with CSR Reports (see Fig. 6). involved in fewer CSIR events than the reference group of Brazilian
Moreover, if firms do not adopt Social Policies but are involved in a firms: Mexican firms have an average of 0.66 (Model 2) and 0.61
process of internationalization, the predicted number of CSIR (Model 4) fewer CSIR events compared to Brazilian firms. Firms in
events increases from 0.64 (0.37) in the case of firms internation- the Industry dummy III are involved in an average number of CSIR
alizing in countries characterized by low speech and press events 0.88 (Model 2) and 0.66 (Model 4) fewer than firms in
freedoms, to 1.92 (1.11) in the case of firms internationalizing in Industry dummy I, and firms in the Industry dummy II are involved
countries characterized by strong speech and press freedoms.The in 0.49 (significant only at the 10% level, Model 2) fewer CSIR
results are similar for firms with CSR reports. Firms adopting Social events compared to firms in Industry dummy I. Therefore, firms in
Policies which have not internationalized are expected to show, on the extractive industries are more socially irresponsible than firms
556 D. Fiaschi et al. / Journal of World Business 52 (2017) 546–563

Table 3
Results of regression analyses.

Model (1) (2) Model (3) (4)

Dependent variable CSIR CSIR Dependent variable CSIR CSIR
Estimation Method RE2SLS RE2SLS Estimation Method RE2SLS RE2SLS
HCSPF 0.06 0.64*** HCSPF 0.03 0.37*
(0.06) (0.16) (0.05) (0.19)
Social Policies 2.42** 1.11* CSR Report 0.45 0.22
(1.21) (0.59) (0.83) (0.63)
Social Policies * HCSPF 0.86*** CSR Report * HCSPF 0.59**
(0.20) (0.28)
Size 0.12 0.06 Size 0.06 0.02
(0.08) (0.05) (0.08) (0.06)
Age 0.36* 0.20*** Age 0.08 0.10
(0.19) (0.07) (0.10) (0.09)
SOEs 0.05 0.11 SOEs 0.00 0.18
(0.45) (0.36) (0.34) (0.34)
ROE 0.01 0.01 ROE 0.05 0.01
(0.04) (0.03) (0.06) (0.05)
Risk 0.32 0.47*** Risk 0.36** 0.31**
(0.20) (0.17) (0.17) (0.15)
Internationalization 0.09* 0.10** Internationalization 0.11** 0.14***
(0.05) (0.05) (0.05) (0.05)
CSR Experience Social Policies 0.00 0.03*** CSR Experience CSR Report 0.01 0.18
(0.03) (0.01) (0.07) (0.07)
Media Exposure 59.64*** 61.89*** Media Exposure 40.95** 41.06**
(22.30) (17.26) (17.66) (16.85)
Industry dummy II 0.54 0.49* Industry dummy II 0.34 0.40
(0.43) (0.29) (0.26) (0.25)
Industry dummy III 0.89* 0.88** Industry dummy III 0.68** 0.66*
(0.48) (0.35) (0.33) (0.30)
Mexico 0.87** 0.66** Mexico 0.53** 0.61**
(0.36) (0.26) (0.25) (0.25)
Time dummies YES YES Time dummies YES YES

Number of observations 394 394 Number of observations 394 394

Adjusted R2 0.14 0.17 Adjusted R2 0.22 0.22
F-statistics 176.1*** 101.04 *** F-statistics 125.4*** 113.4***
(df 1;349) (df 1;349) (df 1;349) (df 1;349)
Test of Serial Correlation 0.58*** 0.71*** Test of Serial Correlation 0.56*** 0.54***
(0.04) (0.05) (0.50) (0.05)

Note: *** p-value <0.01; ** p-value <0.05; * p-value <0.1; robust standard errors in brackets. RE2SLS: Random Effect Panel Two-Stage Least Squares.

in the other two groups (Industry dummy II and Industry dummy III) 5. Discussion and conclusion
which confirms the previous evidence on the problematic
implications for society of extractive industries (Giuliani & Macchi, 5.1. Discussion
2014).
It is generally assumed that emerging country firms suffer from
4.2. Robustness checks a legitimacy deficit due to their home countries' weak institutions
and poor reputation. To earn international legitimacy, these firms
We control for robustness of the estimates by changing the adopt numerous CSR policies, a phenomenon which has attracted
industry grouping specifications, by reducing the number of considerable recent scholarly attention (Fiaschi et al., 2015; Gugler
control variables,23 and by using lagged variables for Internation- & Shi, 2009; Marano et al., 2016; Marquis & Qian, 2014; UNCTAD,
alization and HCSPF. We also ran the estimations using as a measure 2008; Zheng et al., 2015; among others). Drawing on neo-
of CSR policy the adoption of the UN Global Compact, one of the institutional theory and its extensions (Kostova & Roth, 2002;
world's largest corporate sustainability initiative, which calls firms Kostova et al., 2008; Marano & Kostova, 2015), we follow prior
to align strategies and operations with universal principles on research in arguing that emerging country firms may be adopting a
human rights, labor, environment and anti-corruption, and take dual strategy as part of their internationalization plans. They may
actions that advance societal goals.24 The results of the second adopt CSR policies to gain global legitimacy but – possibly due to
stage regressions are largely in line with our findings using Social rent seeking – they may do harm and be socially irresponsible. This
Policies and CSR Report as explanatory variables (results available kind of duality is seldom discussed in the context of EMNEs’ global
from the authors on request). legitimacy building, or in studies of internationalizing firms more
generally. In this paper we conjectured that host country speech
and press freedom is a key conceptual construct influencing
Multilatinas’ CSIR conduct, and that CSR plays a moderating role in
the relationship between Multilatinas’ host countries’ speech and
press freedoms and CSIR.
23
Following the suggestion of one of the reviewers, we dropped ROE, Risk, and CSR Our study of a sample of large public Brazilian and Mexican
Experience, on the grounds that too many explanatory variables could affect the firms shows investing in countries of varying institutional quality
robustness of our estimates. Re-running the regressions without these variables did related especially to speech and press freedoms, does not affect the
not change our key results significantly, thus, for conceptual reasons, we decided to
probability that the firm will be involved in CSIR events. However,
retain the original full model.
24
For details on the UN Global Compact, see https://www.unglobalcompact.org. we found that the characteristics of investing EMNEs influence
 D. Fiaschi et al. / Journal of World Business 52 (2017) 546–563 557

Fig. 5. Interaction effect between Social Policies and HCSPF.

Note: Dotted lines represents 95% confidence bands of the estimates.

their CSIR conduct across different host country institutional institutional pressure based on the host countries’ degree of
environments. In particular, Multilatinas that adopt explicit CSR speech and press freedoms than the coercive, mimetic, and
policies are involved in a smaller number of CSIR events the more normative behaviors identified in the literature (DiMaggio &
their investments are oriented towards countries characterized by Powell, 1983). Governments protecting and guaranteeing speech
strong speech and press freedoms. This finding is coherent with and press freedoms within their own jurisdictions do not impose
the view that managers likely adjust their rent-seeking, irrespon- rules and standards of conduct on MNEs/EMNEs – nor is there
sible business conduct on the basis of the associated de- pressure for conformity to given norms. Nonetheless, the presence
legitimation risks. Our work is original in showing that two of such freedoms can influence the legitimation building processes
conditions are required to curb CSIR: the adoption of explicit CSR of MNEs/EMNEs as stronger speech and press freedoms will
policies, and pressure from the media and other reporting agencies increase the broadcasting and critical assessment of any news of
to inform the relevant constituencies and promote critical debates irresponsible behaviors. This in turn reduces information asym-
on firms’ involvement in CSIR. This result suggests also that metries in firms’ involvement in CSIR by increasing MNEs/EMNEs
Multilatinas which adopt CSR policies may be able strategically to exposure to public scrutiny and criticism which jeopardizes their
maneuver their irresponsible conduct in order to preserve their legitimacy vis a vis important constituencies in the case of
legitimacy. In contrast, Multilatinas that do not adopt explicit CSR irresponsibility. Hence, speech and press freedoms might be
policies may be less savvy and less capable of controlling their another important manifestation of the institutional pressure that
irresponsible behavior. This may be due either to these firms’ shapes MNE/EMNEs CSIR conduct. Speech and press freedoms may
lacking the relevant skills, or to their lack of interest in using CSR also influence the cognitive pillar of the host country's institutional
and CSIR to build legitimacy. We elaborate further on the environment by shaping the cognitive structure of the host
implications of these results in the section below. country's constituencies, and determining what is legitimate in
society (Kostova & Zaheer, 1999; Scott, 1995).
5.2. Contributions However, this kind of pressure does not translate into a unique
legitimacy building strategy since it works heterogeneously across
We highlight some noteworthy contributions. First, we add to firms – in this context, it serves to curb CSIR only among CSR policy
neo-institutional theorizing on the process of MNE legitimacy adopters. There are different explanations for why CSR adopters
building with particular reference to EMNEs. We concur that are less likely to do harm when investing in countries with strong
MNEs/EMNEs are organizationally complex which makes it speech and press freedoms. One interpretation which we have
difficult to understand their adaptation to the different institu- used to develop our hypothesis is that they are aware that they are
tional environments in which they invest. Our work suggests that under scrutiny due to the adoption of CSR policies (Ashforth &
MNEs/EMNEs are not equally isomorphic with the institutional Gibbs, 1990; Morsing & Schultz, 2006; Zhang & Luo, 2013), and
environment – i.e. they respond differently to the same therefore they are more exposed to attributions of irresponsibility.
institutional pressures. This is in line with the active agency Alternatively, firms that have adopted CSR policies may also be
perspective which suggests that MNEs are able to exploit either more ethical, or more skilled at managing their operations
institutions to their own advantage, rather than homogeneously without doing harm – not least because they may be able to
conforming to them (Marquis & Raynard, 2015). Within this undertake human rights due diligence or a human rights impact
framework, our study is original in theorizing a different form of assessment. Finally, it is possible that it is in the EMNE's interest to
558 D. Fiaschi et al. / Journal of World Business 52 (2017) 546–563

Fig. 6. Interaction effect between CSR Report and HCSPF.

Note: Dotted lines represents 95% confidence bands of the estimates.

maintain internal legitimacy within its organization (Kostova & rights – i.e. neatly defining the responsibility of international
Zaheer, 1999) since information on irresponsible behavior from firms to do no-harm. Multilatinas, similar to other EMNEs, are
sources outside the firm, could infiltrate the firm's divisions and learning fast, and research shows that they are aligning quickly to
subsidiaries resulting in harmful intra-organizational tensions. international CSR practices and standards. However, it remains to
These various interpretations call for more research to investigate be seen whether their adherence to this new agenda implies that
the internal and external pressures that influence processes of they will be less likely to do harm in the future, and whether this
legitimacy building by doing no harm. extends to all the countries in which they have investments. We
This leads to our second contribution to the international think that this remains a major untapped challenge for Multilatinas
business and management literatures on CSIR and related as well as for many MNEs around the world.
constructs. We noted that compared to CSR, CSIR has been Our research contributes to scholarship as it offers also a
somewhat overlooked by international business and manage- different perspective on Multilatinas’ internationalization. Earlier
ment scholars (Giuliani et al., 2016; Whiteman & Cooper, 2016). research on the internationalization of Latin American firms
Earlier research shows that internationalization can increase focuses either on the motivations and factors affecting their
instances of MNEs' involvement in irresponsible business conduct internalization strategies and entry modes (e.g. Ciravegna, Lopez, &
(Strike et al., 2006), and that MNE headquarters transfer Kundu, 2014; Cuervo-Cazurra, 2008; Gonzalez-Perez & Velez-
irresponsible practices to subsidiaries located in relatively weak Ocampo, 2014; Losada-Otálora & Casanova, 2014; Santiso, 2008),
institutional environments (Surroca et al., 2013). Extant research on the barriers to internationalization (e.g. Ciravegna, Lopez, &
suggests also that CSR and CSIR tend to coexist (Idemudia, 2009; Kundu, 2016), or on the factors contributing to their achievement
Strike et al., 2006; among others), or that CSIR is a predictor of CSR of a competitive edge in the home and the host countries (e.g.
(Fiaschi et al., 2015; Kotchen & Moon, 2012; Muller & Kraussl, Cuervo-Cazurra & Genc, 2008; Tavares, 2007). We add to this
2011). However, few studies try to predict CSIR or assess the literature by providing evidence on Multilatinas’ strategies for
effectiveness of existing CSR policies for curbing CSIR. We found meeting the expectations of the institutional environment where
that the adoption of CSR policies is not sufficient to reduce they invest, and on improvements over time in the institutional
incidences of CSIR, since CSR adoption is not effective when firms quality of their host countries, observed through the growing share
invest in countries with lower speech and press freedoms. This of FDI in countries characterized by strong speech and press
calls for caution in attributing value to voluntary and explicit CSR freedoms. This might reflect the increased ability of Multilatinas to
policies, which in turn points to the normative implications of operate in more demanding institutional contexts, and suggest
CSR research and practice. The 2015 UN Sustainable Development that some Multilatinas may have the capabilities to control their
Goals and the 2011 UN Guiding Principles on Business and Human socially irresponsible conduct when they invest in such contexts.
Rights are two of the most recent initiatives promoting the Multilatinas are an ideal context to investigate the proposed
human rights agenda in the business sector. They reflect a broader research questions because of their long-standing involvement in
international agenda to ensure business firms' respect of human resolving societal issues ignored by the home governments
rights in their operations, and are establishing a new frontier to (Fiaschi et al., 2015), and because of their intense engagement
CSR practice based on respect and promotion of universal human in both corporate giving and other CSR policies (Araya, 2006;
 D. Fiaschi et al. / Journal of World Business 52 (2017) 546–563 559

Griesse, 2007; Medeiros Peliano et al., 2002; Puppim de Oliveira, research uses more refined measures of CSR. For instance, the
2008; Vives, 2012). Having mature CSR-related skills may provide development of measures of the quality of CSR policies based on
these firms with the capabilities needed also to avoid doing harm – content analysis of firms’ CSR reports, could be one way to
for instance, vis a vis other geographical contexts, Multilatinas’ advance this research area.
managers may be more familiar with concepts such as human Third, in the measurement of CSIR we used information on
rights, poverty reduction, etc. However, these are qualities which alleged human rights violations, which may not finally be judged
remain hidden unless these firms’ investments are in institution- as such by a domestic court since only a small minority of human
ally more strict contexts where strong speech and press freedoms rights violations result in lawsuits and receive a final judicial
put their legitimation at risk. This result calls for more research to decision. This is due to many constituencies lacking access to
better understand CSIR in the Latin American context, and in turn justice, and friendly settlements (Kobrin, 2009). Another reason
to achieve a better understanding of how Multilatinas differ from for not using final judicial decision as a measure of CSIR is that
purely-domestic firms, or from foreign firms investing in Latin there is wide cross-country variety in how human rights’ treaties
America. Some of the qualitative research on Latin America reports are incorporated into national legal systems.25 Thus, relying only
gross violations of human rights in the context of the forestry on violations attached to a final judicial decision would
(Whiteman & Cooper, 2016), extractive (e.g. Dougherty & Olsen, underestimate our variable of interest. Moreover, we acknowl-
2014), agro-food (Lapegna, 2014), and many other industries (see edge that human rights violations are likely to be less intensively
Giuliani & Macchi, 2014). This body of work provides very detailed broadcast in contexts of weak political and civil rights, limited
accounts about the contexts that engender irresponsibility, and press freedom, and repression of local communities, NGOs, and
about the reactions of local constituencies – often indigenous other civil society actors. Thus, the present work and other similar
communities – to being abused. Further research could investigate works (e.g. Marquis & Qian, 2014; Strike et al., 2006; Surroca et al.,
this on a larger scale, and delve into the different outcomes 2013; among others), relies on evidence of social conduct or
generated by different types of firms and their strategies. misconduct being reported (by the press, NGOs, activists,
Finally, our work should be useful to practitioners particularly but governments, residents, etc.), which means that there is a
not exclusively, MNE/EMNE managers. One of the biggest challenges discrepancy between actual human rights violations, and
current managers face is achieving a ‘social license to operate’. Most observed or reported violations. However, our study was not
efforts have been based on creating favorable relationships with both aimed at explaining the intensity of CSIR as such but rather the
local and global relevant stakeholders, via sets of initiatives which inter-firm differences in CSIR conduct. In this context, there might
can be considered as CSR related e.g. community support, be a problem if we have a reason to believe that some firms are
stakeholder management, adoption of principle-based initiatives, observed more closely (and therefore their CSIR events receive more
etc. However, social expectations change constantly, and in the reporting attention) than others. To account for this, we controlled
contemporary international scenario respect for human rights is for media exposure – as in Marquis and Qian (2014). Nevertheless,
prioritized by several international organizations, as discussed in the we believe that the measurement of CSIR and human rights
paper. Hence, failure to respect human rights can no longer be offset violations, and the elaboration and use of measures that contribute
by CSR policies, and there will be demands for rigorous due diligence to understanding corporate wrongdoing, deserve more careful
in relation to human rights before firms can establish business consideration.
activities in given territories. There will be a requirement for firms to This study does also not allow us to investigate in more depth
address such questions as: Is this new investment likely to harm the some of the motivations for our observed empirical results, and
right to health of local residents? For how long? How seriously? Has leaves them open to speculation. The existing international
the acquisition of this new piece of land been accompanied by business research shows that MNEs are characterized by hetero-
consultation with local indigenous communities? Is there a clear geneous governance models and other characteristics which
understanding about indigenous people's requirements in relation condition the extent to which headquarters are able to control
to respect for their right to land? It is likely that response to these and the operations of their subsidiaries and influence the capacity to
similar questions will become mandatory. The soon this is accepted generate value (Ambos & Schlegelmilch, 2007). So far very little
by MNE/EMNE managers, the more this will reinforce internation- research has investigated how the heterogeneity of different intra-
alizing firms' legitimacy building processes, and the better it will be MNE characteristics such as headquarters-subsidiary relation-
for society as a whole. ships, mode of intra-MNE global governance, degree of subsidiary
autonomy in decision making, and local managers’ stewardship,
5.3. Limitations and further research condition the irresponsible conduct of MNEs from both advanced
and emerging countries. These are issues whose investigation
This paper has some limitations and the results should be would contribute to a better understanding of the firm-level
interpreted with some caution. First, our study is based on a drivers of CSIR.
restricted sample of large public companies in Brazil and Finally, we focused on how host country speech and press
Mexico which are not representative of the population of freedoms influence MNEs’ strategic choices with reference to CSIR;
Brazilian and Mexican firms. However, our sample firms are it would be interesting also to investigate how MNEs influence the
very visible and very resourceful which makes them convenient speech and press freedoms in host countries to obtain more
for this kind of analysis. Focus on these two countries also advantage from their investments. We hope that future endeavors
mean that our results apply to our sample of the largest will explore this topic.
Brazilian and Mexican public companies, not to smaller
Multilatinas or to purely domestic firms. Acknowledgements
Second, the adoption of CSR policies as measured in this paper
provides very little information on the magnitude and quality of The authors are especially grateful to Alvaro Cuervo-Cazurra
these companies' CSR investments. Unfortunately, data on CSR and Maria Alejandra Gonzalez-Perez for providing extensive
investments are not publicly available. To address this limitation
and to check the robustness of our results, we used adoption of
the UN Global Compact as another CSR measure, and found 25
For an overview, see http://indicators.ohchr.org/, last accessed November 13,
consistent results. However, we would recommend that future 2015.
560 D. Fiaschi et al. / Journal of World Business 52 (2017) 546–563

feedback on earlier drafts of this manuscript, and to two Appendix B. First step estimations
anonymous reviewers for very constructive comments. Thanks
go also to Luciano Ciravegna, Ruth Aguilera, the reviewers of the Table B1 First step regressions, CSR
Academy of International Business and Academy of Management
2016 Annual Meetings, and to the participants of the JWB Paper Model (1) (2) Model (3) (4)
Development Workshop held in San Paolo in February 2016. We Dependent Social Social Dependent CSR CSR
feel indebted to Chiara Macchi, Luisa Nardi and Emanuele variable Policies Policies variable Report Report
Estimation Method RE2SLS RE2SLS Estimation Method RE2SLS RE2SLS
Sommario for their insights on human rights and business.
Financial support from the University of Pisa (PRA-2015-082) is IV Social Policies 0.80* 1.03* IV CSR Report 0.78* 1.07*
(0.31) (0.32) (0.28) (0.29)
gratefully acknowledged. All disclaimers apply. HCSPF 0.00 0.15* HCSPF 0.016 0.17
(0.02) (0.09) (0.02) (0.07)
Appendix A. List of firms included in the study with country and IV Social Policies * 0.18* IV CSR Report * 0.23
industry specifications HCSPF (0.10) HCSPF (0.09)
Size 0.00 0.02 Size 0.06* 0.04*
(0.02) (0.02) (0.02) (0.02)
Firm Country Industry Age 0.07 0.04 Age 0.09 0.09
ALFA Mexico Metals and Mining (Hld) (0.05) (0.04) (0.04) (0.04)
America Movil Mexico TLC SOEs 0.02 0.04 SOEs 0.00 0.00
Banco Bradesco Brazil Banking (0.11) (0.11) (0.13) (0.13)
Banco do Brasil Brazil Banking ROE 0.01 0.01 ROE 0.05 0.04
Banrisul Brazil Banking (0.01) (0.01) (0.05) (0.05)
Bradespar Brazil Banking Risk 0.01 0.07 Risk 0.01 0.01
Brasil Foods Brazil Food and Beverages (0.04) (0.05) (0.14) (0.15)
Braskem Brazil Chemicals and Pharma Internationalization 0.01 0.00 Internationalization 0.01 0.00
Cemex Mexico Heavy Industry (0.00) (0.00) (0.01) (0.01)
Cemig Brazil Electricity and other Utilities CSR Experience 0.01 0.01* CSR Experience CSR 0.07* 0.07*
Cielo Brazil Banking Social Policies (0.01) (0.01) Report (0.01) (0.01)
Companhia Siderurgica Nacional Brazil Metals and Mining Media Exposure 8.28 8.80 Media Exposure 6.46 4.82
Copel Brazil Electricity and other Utilities (3.39) (3.41) (4.61) (4.53)
Cosan Industria e Comercio Brazil Electricity and other Utilities (Hld) Industry dummy II 0.14 0.10 Industry dummy II 0.14 0.11
CPFL Energia Brazil Electricity and other Utilities (0.09) (0.10) (0.12) (0.13)
El Puerto de Liverpool Mexico Retail Industry dummy III 0.15 0.13 Industry dummy III 0.19 0.12
Eletrobras Brazil Electricity and other Utilities (0.01) (0.09) (0.12) (0.13)
Embraer Brazil Aerospace Mexico 0.07 0.05 Mexico 0.04 0.07
Femsa Mexico Food and Beverages (0.08) (0.08) (0.11) (0.11)
Fresnillo Mexico Metals and Mining Time dummies YES YES Time dummies YES YES
GF Norte Mexico Banking
Grupo Bimbo Mexico Food and Beverages Number of 394 394 Number of 394 394
Grupo Carso Mexico Retail (Hld) observations observations
Grupo Elektra Mexico Retail Adjusted R2 0.62 0.63 Adjusted R2 0.55 0.55
Grupo Inbursa Mexico Banking F-statistics 29.72* 30.19* F-statistics 23.38* 22.15*
Grupo Mexico Mexico Metals and Mining (df 22; (df 23; (df 22; (df 23;
Grupo Pão de Açúcar Brazil Food and Beverages 372) 371) 372) 371)
Grupo Televisa Mexico TLC
Industrias Penoles Mexico Metals and Mining Note:*** p-value < 0.01; ** p-value < 0.05; * p-value < 0.1; robust
Itau Unibanco Holding Brazil Banking
standard errors in brackets. RE2SLS: Random Effect Panel Two-
Itausa – Investimentos Itau Brazil Banking (Hld)
JBS Brazil Food and Beverages Stage Least Squares.
Lojas Americanas Brazil Retail
Metalurgica Gerdau Brazil Metals and Mining
Minera Frisco Mexico Metals and Mining
Natura Cosmeticos Brazil Cosmetics
OGX Petroleo e Gas Participacoes Brazil Metals and Mining
PDG Realty Brazil Real Estate
Sabesp Brazil Electricity and other Utilities
Sul America Brazil Banking
Suzano Papel e Celulose Brazil Pulp and Paper
Tele Norte Leste Brazil TLC
Usiminas Brazil Metals and Mining
Vale Brazil Metals and Mining
 Stage Least Squares.
standard errors in brackets. RE2SLS: Random Effect Panel Two-

Table B2 First step regressions, interaction CSR and HCSPF

F-statistics
Adjusted R2

Number of

Time dummies

Mexico

Industry dummy III

Industry dummy II

Media Exposure

CSR Experience Social

Internationalization

Risk

ROE

SOEs

Age

Size

IV Social Policies *

HCSPF

IV Social Policies
Estimation Method

Dependent variable
Model
Note: ***p-value < 0.01; **p-value < 0.05; *p-value < 0.1; robust

observations

Policies

HCSPF
(df 23; 371)
54.65*
0.76

394

YES
(0.18)
0.17
(0.20)
0.27
(0.16)
0.03
(7.38)
18.50
(0.01)
0.02
(0.01)
0.01
(0.09)
0.07
(0.03)
0.03
(0.01)
0.08
(0.06)
0.00
(0.04)
0.01
(0.22)
0.90*
(0.20)
0.07
(0.45)
0.03
RE2SLS
*HCSPF
Social Policies
(1)
F-statistics
Adjusted R2
observations
Number of

Time dummies

Mexico

Industry dummy III

Industry dummy II

Media Exposure
Report
CSR Experience CSR

Internationalization

Risk

ROE

SOEs

Age

Size
HCSPF
IV CSR Report *

HCSPF

IV CSR Report
Estimation Method

Dependent variable
Model

D. Fiaschi et al. / Journal of World Business 52 (2017) 546–563

(df 23; 371)
38.42*
0.69

394

YES
(0.21)
0.04
(0.23)
0.29
(0.24)
0.07
(9.09)
12.50
(0.02)
0.09*
(0.01)
0.02
(0.13)
0.11
(0.03)
0.00
(0.16)
0.16
(0.07)
0.05
(0.04)
0.03
(0.18)
0.66*
(0.15)
0.26*
(0.56)
0.01
RE2SLS
*HCSPF
CSR Report
(2)
Appendix C. Correlation table

Variables 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
1 CSIR 1
2 HCSPF 0.135 1
3 Social Policies 0.099 0.059 1
4 CSR Report 0.134 0.171 0.502 1
5 Social Policies*HCSPF 0.164 0.85 0.406 0.328 1
6 CSR Report*HCSPF 0.155 0.76 0.275 0.558 0.82 1
7 Size 0.136 0.449 0.019 0.129 0.374 0.333 1
8 Age 0.135 0.084 0.183 0.119 0.145 0.135 0.308 1
9 SOEs 0.019 0.102 0.075 0.086 0.04 0.001 0.01 0.197 1
10 ROE 0.028 0.035 0.072 0.006 0.029 0.02 0.157 0.04 0.052 1
11 Risk 0.081 0.119 0.15 0.087 0.08 0.055 0.173 0.088 0.102 0.512 1
12 Internationalization 0.463 0.437 0.038 0.187 0.369 0.417 0.409 0.107 0.067 0.024 0.058 1
13 CSR Experience Social Policies 0.191 0.101 0.545 0.413 0.291 0.278 0.201 0.383 0.237 0.055 0.073 0.189 1
14 CSR Experience CSR Report 0.079 0.232 0.384 0.712 0.359 0.511 0.146 0.204 0.077 0.068 0.065 0.344 0.527 1
15 Media Exposure 0.101 0.174 0.093 0.145 0.188 0.208 0.018 0.155 0.082 0.038 0.119 0.255 0.019 0.166 1
16 Industry dummy I 0.19 0.025 0.098 0.03 0.038 0.032 0.134 0.058 0.215 0.024 0.014 0.005 0.065 0.102 0.013 1
17 Industry dummy II 0.119 0.101 0.021 0.002 0.015 0.004 0.122 0.051 0.098 0.124 0.26 0.112 0.02 0.007 0.12 0.323 1
18 Industry dummy III 0.042 0.047 0.083 0.083 0.054 0.001 0.171 0.014 0.284 0.126 0.226 0.111 0.002 0.126 0.11 0.511 0.597 1
19 Mexico 0.059 0.279 0.207 0.177 0.194 0.111 0.206 0.008 0.296 0.05 0.092 0.174 0.307 0.091 0.107 0.165 0.109 0.104 1
20 Brazil 0.059 0.279 0.207 0.177 0.194 0.111 0.206 0.008 0.296 0.05 0.092 0.174 0.307 0.091 0.107 0.165 0.109 0.104 1 1

561
562 D. Fiaschi et al. / Journal of World Business 52 (2017) 546–563

References Gond, J.-P., Palazzo, G., & Basu, K. (2009). Reconsidering instrumental corporate
social responsibility through the Mafia metaphor. Business Ethics Quarterly, 19
Ambos, B., & Schlegelmilch, B. B. (2007). Innovation and control in the multinational (1), 57–85.
firm: A comparison of political and contingency approaches. Strategic Gonzalez-Perez, A. M., & Velez-Ocampo, J. F. (2014). Targeting one’s own region:
Management Journal, 28(5), 473–486. Internationalization trends of Colombian multinational companies. European
Araya, M. (2006). Exploring terra incognita non-financial reporting in corporate Business Review, 26(6), 531–551.
Latin America. Journal of Corporate Citizenship, 21, 25–38. Griesse, M. A. S. (2007). The geographic, political, and economic context for
Armstrong, J. S. (1977). Social irresponsibility in management. Journal of Business corporate social responsibility in Brazil. Journal of Business Ethics, 73, 21–37.
Research, 5(3), 185–213. Gugler, P., & Shi, J. Y. J. (2009). Corporate social responsibility for developing country
Ashforth, B. E., & Gibbs, B. W. (1990). The double-edge of organizational multinational corporations: Lost war in pertaining global competitiveness?
legitimation. Organization Science, 1(2), 177–194. Journal of Business Ethics, 87(1), 3–24.
Barnett, M. L., Jermier, J. M., & Lafferty, B. A. (2006). Corporate reputation: The Hymer, S. (1960/1976). The international operations of national firms: A study of direct
definitional landscape. Corporate Reputation Review, 9(1), 26–38. investment. Cambridge, MA: MIT Press.
BenYishay, A., & Pearlman, S. (2014). Crime and microenterprise growth: Evidence Idemudia, U. (2009). Oil extraction and poverty reduction in the Niger Delta: A
from Mexico. World Development, 56, 139–152. critical examination of partnership initiatives. Journal of Business Ethics, 90, 91–
Brenes, E. R., Camacho, A. R., Ciravegna, L., & Pichardo, C. A. (2016). Strategy and 116.
innovation in emerging economies after the end of the commodity boom— Jones, B., Bowd, R., & Tench, R. (2009). Corporate irresponsibility and corporate
Insights from Latin America. Journal of Business Research, 69, 4363–4367. social responsibility: Competing realities. Social Responsibility Journal, 5(3),
Carroll, A. B. (2008). A history of corporate social responsibility: Concepts and 300–310.
practices. In A. Crane, A. McWilliams, D. Matten, J. Moon, & D. Siegel (Eds.), The Jonsson, S., Greve, H. R., & Fujiwara-Greve, T. (2012). Undeserved loss: The spread of
Oxford handbook of corporate social responsibility (pp. 19–46).Oxford: Oxford legitimacy loss to innocent organizations in response to reported corporate
University Press. deviance. Administrative Science Quarterly, 54, 195–228.
Casanova, L. (2009). Global latinas: Latin america’s emerging multinationals. Palgrave Kang, C., Germann, F., & Grewal, R. (2016). Washing away your sins? Corporate social
Macmillan. responsibility, corporate social irresponsibility, and firm performance. Journal of
Chew Sánchez, M. I. (2014). Paramilitarism and state-terrorism in Mexico as a case Marketing, 80(March), 59–79.
study of shrinking functions of the neoliberal state. Perspectives on Global Keig, D. L., Brouthers, L. E., & Marshall, V. B. (2015). Formal and informal corruption
Development and Technology, 13(1–2), 176–196. environments and multinational enterprise social irresponsibility? Journal of
Chiu, S. C., & Sharfman, M. (2016). Corporate social irresponsibility and executive Management Studies, 52(1), 89–116.
succession: An empirical examination. Journal of Business Ethics. http://dx.doi. Khanna, T., & Palepu, K. (1997). Why focused strategies may be wrong for emerging
org/10.1007/s10551-016-3089-7. markets. Harvard Business Review, 75, 41–54.
Cingranelli, D. L., & Richards, D. L. (2010). The Cingranelli and Richards (CIRI) human King, B. G., & Whetten, D. A. (2008). Rethinking the relationship between reputation
rights data project. Human Rights Quarterly, 32(2), 401–424. and legitimacy: A social actor conceptualization. Corporate Reputation Review,
Cingranelli, D. L., & Richards, D. L. (2014). The Cingranelli and Richards (CIRI) human 11, 192–207.
rights data project coding manual version 5.20.14. Kobrin, S. J. (2009). Private political authority and public responsibility-
Ciravegna, L., Lopez, L., & Kundu, S. (2014). Country of origin and network effects on Transnational politics, transnational firms, and human rights. Business Ethics
internationalization: A comparative study of SMEs from an emerging and Quarterly, 19(3), 349–374.
developed economy. Journal of Business Research, 67(5), 916–923. Kostova, T., & Roth, K. (2002). Adoption of an organizational practice by subsidiaries
Ciravegna, L., Lopez, L., & Kundu, S. (2016). The internationalization of Latin of multinational corporations: Institutional and relational effects. Academy of
American enterprises  Empirical and theoretical perspectives. Journal of Management Journal, 45(1), 215–233.
Business Research, 69(6), 1957–1962. Kostova, T., & Zaheer, S. (1999). Organizational legitimacy under conditions of
Croissant, Y., & Millo, G. (2008). Panel data econometrics in R: The plm package. complexity: The case of the multinational enterprise. Academy of Management
Journal of Statistical Software, 27(2), 1–43. Review, 24(1), 61–64.
Cuervo-Cazurra, A., & Genc, M. (2008). Transforming disadvantages into Kostova, T., Roth, K., & Dacin, M. T. (2008). Institutional theory in the study of
advantages: Developing-country MNEs in the least developed countries. Journal multinational corporations: A critique and new directions. Academy of
of International Business Studies, 39(6), 957–979. Management Review, 33(4), 994–1006.
Cuervo-Cazurra, A., & Ramamurti, R. (2014). Understanding multinationals from Kotchen, M., & Moon, J. (2012). Corporate social responsibility for irresponsibility.
emerging markets. Cambridge: Cambridge University Press. The BE Journal of Economic Analysis & Policy, 12(1), 1–22.
Cuervo-Cazurra, A. (2008). The multinationalization of developing country MNEs: Kuecker, G. D. (2007). Fighting for the forests: Grassroots resistance to mining in
The case of Multilatinas. Journal of International Management, 14, 138–154. Northern Ecuador. Latin American Perspectives, 34(2), 94–107.
DiMaggio, P. J., & Powell, W. W. (1983). The iron cage revisited: Institutional Lange, D., & Washburn, N. T. (2012). Understanding attributions of corporate social
isomorphism and collective rationality in organizational fields. American irresponsibility. Academy of Management Review, 37(2), 300–326.
Sociological Review, 48, 147–160. Lange, D., Lee, M. L., & Dai, Y. (2011). Organizational reputation: A review. Journal of
Dinham, B., & Sarangi, S. (2002). The Bhopal gas tragedy 1984 to the evasion of Management, 37, 153–184.
corporate responsibility. Environment and Urbanization, 14(1), 89–99. Lapegna, P. (2014). Global ethnography and genetically modified crops in Argentina:
Dos Santos, L. A., & da Costa, P. M. T. (2014). The contribution of lobby regulation On adoptions, resistances, and adaptations. Journal of Contemporary
initiatives in addressing political corruption in Latin America. Journal of Public Ethnography, 43(2), 202–227.
Affairs, 14(3–4), 379–391. Leahy, J. (2016). What is the Petrobras scandal that is engulfing Brazil? Financial
Dougherty, M. L., & Olsen, T. D. (2014). Taking terrain literally: Grounding local times americas. http://www.ft.com/cms/s/2/6e8b0e28-f728-11e5-803c-
adaptation to corporate social responsibility in the extractive industries. Journal d27c7117d132.html#axzz4GZXrSO9d, Retrieved 6 August 2016.
of Business Ethics, 119, 423–434. Lin-Hi, N., & Muller, K. (2013). The CSR bottom line: Preventing corporate social
Duhigg, C., & Barboza, D. (2012). In China, human costs are built into an Ipad. The irresponsibility. Journal of Business Research, 66, 1928–1936.
new York times. January 25th, http://www.nytimes.com/2012/01/26/business/ Logsdon, J. M., Thomas, D. E., & Van Buren, H. J. III (2006). Corporate social
ieconomy-apples-ipad-and-the-human-costs-for-workers-in-china.html?_r=0, responsibility in large Mexican firms. Journal of Corporate Citizenship, 21, 51–60.
Retrieved 22 August 2016. Losada-Otálora, M., & Casanova, L. (2014). Internationalization of emerging
Duncan, G. (2013). Drug trafficking and political power: Oligopolies of coercion in multinationals: The latin american case. European Business Review, 26(6), 588–
Colombia and Mexico. Latin American Perspectives, 41(2), 18–42. 602.
Ferraz, C., & Finan, F. (2011). Electoral accountability and corruption: Evidence from Lu, Y., & Abeysekera, I. (in press). What do stakeholders care about? Investigating
the audits of local governments. American Economic Review, 101(4), 1274–1311. corporate social and environmental disclosure in China. Journal of Business
Fiaschi, D., Giuliani, E., & Nieri, F. (2015). BRIC companies seeking legitimacy Ethics, forthcoming.
through corporate social responsibility. UNCTAD Transnational Corporations, 22 Lund-Thomsen, P., & Nadvi, K. (2010). Clusters, chains and compliance: Corporate
(3), 5–42. social responsibility and governance in football manufacturing in South Asia.
Fombrun, C. (1996). Reputation: Realizing value from the corporate image. Cambridge, Journal of Business Ethics, 93, 201–222.
MA: Harvard University Press. Madhok, A., & Kayhani, M. (2012). Acquisitions as entrepreneurship: Asymmetries,
Gilbert, D. U., Rasche, A., & Waddock, S. (2011). Accountability in a global economy: opportunities, and the internationalization of multinationals from emerging
The emergence of international accountability standards. Business Ethics economies. Global Strategy Journal, 2(1), 26–40.
Quarterly, 21(1), 23–44. Marano, V., & Kostova, T. (2015). Unpacking the institutional complexity in adoption
Giuliani, E., & Macchi, C. (2014). Multinational corporations' economic and human of CSR practices in multinational enterprises. Journal of Management Studies.
rights impacts on developing countries: A review and research agenda. http://dx.doi.org/10.1111/joms.12124.
Cambridge Journal of Economics, 38(2), 479–517. Marano, V., Tashman, P., & Kostova, T. (2016). Escaping the iron cage: Liabilities of
Giuliani, E., Santangelo, G. D., & Wettstein, F. (2016). Human rights and international origin and CSR reporting of emerging market multinational enterprises. Journal
business research: A call for studying emerging market multinationals. of International Business Studies. http://dx.doi.org/10.1057/jibs.2016.17.
Management and Organization Review, 12(3), 631–637. Marquis, C., & Qian, C. (2014). Corporate social responsibility reporting in China:
Godfrey, P. C. (2005). The relationship between corporate philanthropy and Symbol or substance? Organization Science, 25(1), 127–148.
shareholder wealth: A risk management perspective. Academy of Management Marquis, C., & Raynard, M. (2015). Institutional strategies in emerging markets.
Review, 30(4), 777–798. Academy of Management Annals, 9(1), 291–335.
 D. Fiaschi et al. / Journal of World Business 52 (2017) 546–563 563

Matten, D., & Moon, J. (2008). Implicit and explicit CSR: A conceptual framework for Strike, V., Gao, J., & Bansal, P. (2006). Being good while being bad: Social
a comparative understanding of corporate social responsibility. Academy of responsibility and the international diversification of US firms. Journal of
Management Review, 33, 404–424. International Business Studies, 37(6), 850–862.
Mattingly, J., & Berman, S. (2006). Measurement of corporate social action: Surroca, J., Tribò, J. A., & Zahra, J. A. (2013). Stakeholder pressure on MNEs and the
Discovering taxonomy in the Kinder Lydenburg Domini ratings data. Business & transfer of socially irresponsible practices to subsidiaries. Academy of
Society, 45(1), 20–44. Management Journal, 56(2), 549–572.
Medeiros Peliano, A.-M. T., Beghin, N., & De Oliveira, N. V. (2002). Philanthropy and Sustainalytics (2014). AngloGold Ashanti Ltd. Controversy Report. Last Accessed
equity: The case of Brazil, the global equity. Cambridge, MA: Harvard University 08.11.15.
Press. Sustainalytics (2015). Vale S.A. Controversy Report. Last Accessed 24.11.15.
Mendes, C. M., & Junior, S. P. (2012). Deforestation, economic growth and Tavares, M. (2007). Outward FDI and the competitiveness of Latin American firms. In
corruption: A nonparametric analysis on the case of Amazon forest. Applied R. Grosse, & L. F. Mesquita (Eds.), Can latin american firms compete?Oxford:
Economics Letters, 19(13), 1285–1291. Oxford University Press.
Meyer, J., & Rowan, B. (1977). Institutionalized organizations: Formal structure as The Economist (2014). Our crony-capitalism index: Planet plutocrat. . . Retrieved 20
myth and ceremony. American Journal of Sociology, 83, 340–363. March 2014 http://www.economist.com/news/international/21599041-
Meyskens, M., & Paul, K. (2010). The evolution of corporate social reporting practices countries-where-politically-connected-businessmen-are-most-likely-prosper-
in Mexico. Journal of Business Ethics, 91, 211–227. planet.
Morris, S. D. (2013). Drug trafficking, corruption, and violence in Mexico: Mapping The Economist (2015). Corruption in Brazil: The big oily. the petrobras scandal
the linkages. Trends in Organized Crime, 16(2), 195–220. explained. . . Retrieved 20 March 2014 http://www.economist.com/news/
Morsing, M., & Schultz, M. (2006). Corporate social responsibility communication: americas/21637437-petrobras-scandal-explained-big-oily?
Stakeholder information, response and involvement strategies. Business Ethics: zid=298&ah=0bc99f9da8f185b2964b6cef412227be.
A European Review, 15, 323–338. UNCTAD (2008). 2008 Review of the corporate responsibility performance of large
Muller, A., & Kraussl, R. (2011). Doing good deeds in times of need: A strategic emerging market enterprises. TD/B/C.II/ISAR/CRP.3, http://unctad.org/en/docs/
perspective on corporate disaster donations. Strategic Management Journal, 32 c2isarcrp3_en.pdf, Retrieved 19 May 2015.
(9), 911–929. UNCTAD (2013). World investment report 2013: Global value chains: Investment and
Oliva, P. (2015). Environmental regulations and corruption: Automobile emissions trade for development. United Nation: New York and Geneva. http://unctad.org/
in Mexico city. Journal of Political Economy, 123(3), 686–724. en/PublicationsLibrary/wir2013_en.pdf, Retrieved 19 May 2015.
Oliver, C. (1991). Strategic responses to institutional processes. Academy of UNCTAD (2014). World investment report 2014: Investing in the SDGs: An action plan.
Management Review, 16(1), 145–179. United Nation: New York and Geneva. http://unctad.org/en/
Pegg, S., & Frynas, J. G. (2003). Transnational corporations and human rights. New PublicationsLibrary/wir2014_en.pdf, Retrieved 19 May 2015.
York: Palgrave MacMillan. Vanden, H. E. (2007). Social movements, hegemony, and new forms of resistance:
Power, T. J., & Taylor, M. M. (2011). Corruption and democracy in Brazil: The struggle for The new politics of social movements in Latin America. Latin American
accountability. Notre Dame, IN: University of Notre Dame Press. Perspectives, 34(2), 17–30.
Puppim de Oliveira, J. (2008). Corporate citizenship in Latin America: New Visser, W. (2008). Corporate social responsibility in developing countries. In A.
challenges for business. Journal of Corporate Citizenship, 27, 17–20. Crane, A. McWilliams, D. Matten, J. Moon, & D. Siegel (Eds.), The Oxford handbook
Ramachandran, J., & Pant, A. (2010). The liabilities of origin: An emerging economy of corporate social responsibility in developing countriesOxford: Oxford University
perspective on the cost of doing business abroad. In T. Devinney, T. Pedersen, & L. Press.
Tihanyi (Eds.), The past, present, and future of international business and Vives, A. (2012). Is socially responsible investment possible in Latin America?
management (pp. 231–265).Bingley: Emerald. Journal of Corporate Citizenship, 48, 59–74.
Ramamurti, R., & Singh, J. V. (2009). Emerging multinationals in emerging markets. Wettstein, F. (2010). The duty to protect: Corporate complicity, political
Cambridge: Cambridge University Press. responsibility, and human rights advocacy. Journal of Business Ethics, 96(1), 33–
Raufflet, E. (2008). Creating the context for corporate responsibility. Journal of 47.
Corporate Citizenship, 30, 95–106. Weyzig, F. (2006). Local and global dimensions of corporate social responsibility in
Reid, M. (2014). Brazil: The troubled rise of a global power. Yale University Press. Mexico. Journal of Corporate Citizenship, 24, 69–81.
Ruggie, J. (2008). Protect, respect and remedy: A framework for business and human Whiteman, G., & Cooper, W. H. (2016). Decoupling rape. Academy of Management
rights report of the special representative of the secretary-general on the issue of Discoveries, 2(2), 115–154.
human rights and transnational corporations and other business enterprises. UN Wooldridge, J. M. (2010). Econometric analysis of cross section and panel data.
Doc A/HRC/8/5 7 April 2008. Harvard, MA: MIT Press.
Santiso, J. (2008). The emergence of Latin multinationals. Cepal Review7–30. Zaheer, S. (1995). Overcoming the liability of foreignness. Academy of Management
Scott, R. (1995). Institutions and organizations. Thousand Oaks, CA: Sage. Journal, 38(2), 341–363.
Shrivastava, P. (1995). Industrial/environmental crises and corporate social Zhang, J., & Luo, X. R. (2013). Dared to care: Organizational vulnerability
responsibility. The Journal of Socio-Economics, 24(1), 211–227. institutional logics, and MNCs’ social responsiveness in emerging markets.
Silva, E. (2015). Social movements, protest, and policy. European Review of Latin Organization Science, 24(6), 1742–1764.
American and Caribbean Studies, 100, 27–39. Zheng, Q., Luo, Y., & Maksimov, V. (2015). Achieving legitimacy through corporate
Stahler-Sholk, R., & Vanden, H. E. (2011). Introduction: A second look at latin social responsibility: The case of emerging economy firms. Journal of World
american social movements: Globalizing resistance to the neoliberal paradigm. Business, 50(3), 389–403.
Latin American Perspectives, 38(1), 5–13. Zimmerman, M. A., & Zeitz, G. J. (2002). Beyond survival: Achieving new venture
Staw, B. M., & Epstein, L. D. (2000). What bandwagons bring: Effects of popular growth by building legitimacy. Academy of Management Review, 27(3), 414–431.
management techniques on corporate performance, reputation, and CEO pay.
Administrative Science Quarterly, 45(3), 523–556.