1 Introduction

Stakeholder dialogue (SD) is “a process in which parties with different interests and values at stake in a particular issue work together toward mutually acceptable solutions” (Perret, 2003: 385). The idea behind SD is to go beyond communication, in favor of an ongoing structured process aimed at including the interested party in the decision-making process (Golob & Podnar, 2014). Scholars argue that engaging with stakeholders is key in the relationship between companies and society because understanding stakeholders’ claims helps to develop a more holistic and comprehensive view of the firm and its responsibilities (Perrini & Tencati, 2006), and because stakeholders exert increasing pressures to make companies consider non-economic concerns and obligations (O’Riordan & Fairbrass, 2008; Perret, 2003).

An increasing number of studies supports SD as a necessary tool to define and operationalize corporate social responsibility (CSR) (Burchell & Cook, 2006; Crane & Livesey, 2003; Golob & Podnar, 2014; Illia et al., 2017; Morsing & Schultz, 2006; Pedersen, 2006; van Huijstee & Glasbergen, 2008), whereas fewer efforts have been exerted in the investigation of whether the corporate motives to engage in CSR determine different approaches to SD. Extant research has identified different corporate motives for CSR (Brønn & Vidaver-Cohen, 2009; Santos, 2011), recognizing their fundamental role in firms’ decision to adopt responsible practices. Such motives shape the way firms govern CSR topic (Matten & Moon, 2008) and engage in institutionalized dialogue with their stakeholders (Babiak & Trendafilova, 2011).

The present study originates from these considerations and aims at further understanding the relationships between CSR and SD by exploring whether and how corporate motives for CSR affect the engagement into SD. In addition, the contribution to the understanding of the relationships between CSR and SD is enriched by considering to the role of institutional mechanisms (Campbell, 2007), referring to ethics, and the role of firm size, highlighting how the availability of resources shapes CSR and SD.

First, extant research highlighted that corporate motives for CSR are strongly influenced by institutional mechanisms (Babiak & Trendafilova, 2011; Campbell, 2007), as the differences among national institutional frameworks are related to differences in the meaning and nature of CSR (Kumar et al., 2019). The characteristics of the institutional contexts in which companies operate, especially concerning cultural aspects and ethics, shape also SD (Ferri et al., 2016). Ethics influence how stakeholders are perceived (Jamali et al., 2017), so that the difference between a ‘libertarian’ (Freeman, 1984) or ‘altruistic’ vision of stakeholders (Lutz, 2009) implies different approaches to CSR and SD. This altruistic vision poses challenges to business management and may provide an alternative to stakeholder theory. In particular, it builds on the idea that firm responsibility is neither related to the explicit or implicit contracts with stakeholders nor to the benefit derived from the relationship, but to the mere existence of the relationship itself and the interconnectedness among people and organizations (Woermann & Engelbrecht, 2019).

Second, the contribution to the understanding of CSR and SD is enriched through the lens of the resource‐based view theory (Wernerfelt, 1984). The size of the firm is another variable that influences the CSR behavior of firms and their commitment to SD, as underlined by the vast literature on this topic (Baumann-Pauly et al., 2013; Deegan & Gordon, 1996; Morsing & Perrini, 2009; Reimann et al., 2015). Extant research pointed out that large firms and SMEs have different motives to engage in CSR (Russo & Perrini, 2010) and that SMEs engage less in SD, due to the scarcity of resources (Morsing & Perrini, 2009).

Therefore, the analysis of the relationships between corporate motives for CSR and SD considering ethical pressures and firm size may offer a positive contribution to literature development.

The study was realized in Mozambique, a country representing the African context, which has been severely underrepresented for many years, and only recently has been recognized as a fertile ground for building and expanding theories (Okereke et al., 2018; Phillips, 2006). In the African context, especially in Sub-Saharan countries, the literature identifies different streams of opinion in relation to the role of CSR. Some studies highlighted critics and skepticism about the true benefits of CSR actions, as often the expectations of firms’ contributions have outweighed the benefits actually offered, and have been perceived just as utilitarian (Frynas, 2015; Jamali et al., 2020), whereas most studies assume CSR efforts of corporations are positive for local development (Amponsah-Tawiah & Dartey-Baah, 2016; Katamba & Nkiko, 2016; Painter-Morland & Dobie, 2009; Phillips, 2006).

With regard to Mozambique, this country is an interesting context for the aim of the study, as it is characterized by institutional voids and by the influence of Ubuntu ethics.

While previous contributions in this domain have focused mostly on developed countries, little attention has been paid to less developed countries, particularly to countries marred by institutional voids (Davila et al., 2018). By institutional voids, we refer to a situation where the regulatory system required to support market economies is either absent or too weak to function properly (Khanna & Palepu, 1997, 2010). Most African countries are characterized by institutional voids (Ahen & Amankwah-Amoah, 2018; Amaeshi et al., 2016), as there is a widespread lack of political stability, effective law and regulation, and control of corruption (Kaufmann & Simons-Kaufmann, 2016). Many Sub-Saharan countries suffer from poor governance, determined by inefficient bureaucracy and high levels of corruption (Hussein, 2005), inadequate investments in infrastructure, high transactional costs, scarce opportunities for economic development, a high level of informal economy and poverty (Mzembe et al., 2019). Notwithstanding the high population growth, most Sub-Saharan countries are in the lowest positions of the World Bank's (2020) “Doing Business” ranking.

Contexts with institutional voids are an interesting setting to analyze CSR, as the weakness of institutional arrangements also regards socio-environmental issues (Davila et al., 2018), and because in these contexts CSR has recently gained importance as a tool to address societal challenges (Adeleye et al., 2020; Littlewood & Holt, 2018). However, other studies have questioned the effectiveness of CSR within weak institutional contexts (Amaeshi et al., 2016), as institutional voids create a permissive environment due to a lack of regulation and enforcement, potentially incentivizing misconducts (Zhao et al., 2014).

For this study, analyzing corporate motives for CSR in such contexts is useful to highlight their effect on SD in a clearer way, observing firms’ behavior in challenging and non-enabling institutional settings (Amaeshi et al., 2016). The lack of socio-political structures, such as law enforcement systems and functional consumer redress mechanisms, and the scarce development of local markets may change the relevance of motives to engage in CSR and influence the interaction with stakeholders. Analyzing the relationship between corporate motives for CSR and SD under institutional voids is a topic of utmost relevance to advance the understanding of the relationship between CSR and SD and to better understand the dynamics driving SD (Andersen & Høvring, 2020; Guibert & Roloff, 2017).

Under this perspective, Mozambique is one of the Sub-Saharan countries where institutional voids are most significant (Webster & Wood, 2005), but also a context representative of an unusual vision of stakeholders, derived from Ubuntu ethics (West, 2014). Indigenous Ubuntu ethics, deep rooted in Sub-Saharan Africa, assume that caring for other people is a fundamental approach to life and commonality (Le Roux, 2000). According to Ubuntu ethics firms have an a priori moral responsibility toward all stakeholders (Woermann & Engelbrecht, 2019), thus exerting significant ethical pressures on local firms. However, the country, after an extended period of civil war, has shifted toward a liberal economy, attracting significant investments (Sartorius et al., 2011). As the incomplete regulatory mechanism and the scarce development of the economic system (Visser, 2008) may induce firms to adapt to sub-standard business ethics (Zhao et al., 2014), the economic development of the country may potentially create a clash with the prevailing culture.

Thus, Mozambique appears as a particularly interesting setting under institutional voids, suitable to highlight the different influence of corporate motives for CSR on SD, and the effects of the Ubuntu ethical pressures on this relationship. Other studies have considered SD in Mozambique, with specific attention given to agriculture (Sánchez‐Reparaz, 2020), energy (Ahlborg & Hammar, 2014), water basin (Alba et al., 2017), marine preservation (Lucrezi et al., 2019), and biofuel sustainability (Schut et al., 2010). All these studies have validated that SD in Mozambique is an effective practice to support management agendas, thus confirming the validity of the country as a research setting for this study.

In summary, the present paper contributes to literature development along four main directions. First, it advances the knowledge around the factors that determine firm commitment to SD, by exploring how corporate motives for CSR affect the systematic and structured interaction with stakeholders. Most of extant studies consider how SD integrate CSR strategies and initiatives (Golob & Podnar, 2014; O’Riordan & Fairbrass, 2008), whereas few works have investigated the opposite relationship and no evidence exist on whether and how different CSR motives affect SD. Second, the paper addresses the paucity of research on contexts under institutional voids, common in Sub-Saharan countries—among which Mozambique—which are underrepresented in literature (Amaeshi et al., 2016). In so doing, we therefore also extend the knowledge on SD in developing countries (Davila et al., 2018). Third, the study sheds a light on the role of Ubuntu ethics, which propose a vision of stakeholders alternative to the traditional and “libertarian” vision usually considered in the academic debate. Finally, by referring to the resource‐based view theory (Wernerfelt, 1984), the paper extends the understanding of whether the availability of resources may influence directly and indirectly firms’ commitment to SD, as derived by a vast literature on CSR development.

The remainder of the article is as follows. The first section presents the theoretical background of the study, which is rooted in stakeholder theory, whereas the second discusses CSR and SD in the African context. The third section presents the Mozambican context and its adherence to Ubuntu ethics, referring it to the relationship between CSR and SD, while the fourth section deepens the understanding of the latter topic by considering the influence of firm size. The following part of the article is centered on the results of the study; thus, first the research method is explained, and then the results are presented and discussed. Finally, conclusions summarize the contributions and implications of the research and identify the limitations that draw directions for future research.

2 Theoretical background

2.1 CSR and SD

The concept of SD originates from the stakeholder theory (Freeman, 1984), which introduced the idea that interests from groups other than shareholders converge on the firms (Campbell, 2007), and organizations should consider such interests by managing an extended network of relationships that include customers, employees, suppliers, trade unions and all other interesting parts (Jamali, 2008).

In line with the academic literature, SD represents a useful tool that companies can implement to operationalize their CSR commitment (Maon et al., 2009; Pedersen, 2006). Two mainstreams of literature have addressed SD in CSR (Lane & Devin, 2018). The former considers SD useful in constructing meaningful interactions with the stakeholders in pursuit of mutual understanding (Morsing & Schultz, 2006); the latter considers SD an alternative approach to resolving conflicts around public issues between the company and its opponents (Calvano, 2008). Exchanging information between the firm and its stakeholders contributes to a better understanding of societal expectations or concerns, thus providing useful insights on a broad range of issues beyond the traditional economic ones. SD is therefore considered crucial in implementing CSR (Rinaldi et al., 2014), as it can contribute to determining the effectiveness of a company’s CSR commitment and the evaluation by the stakeholders (O’Riordan & Fairbrass, 2014).

Although extant research has recognized that the way firms govern CSR (Matten & Moon, 2008) and engage in institutionalized dialogue with their stakeholders (Babiak & Trendafilova, 2011) is related to corporate motives for CSR, previous studies have superficially explored how such motives can influence a firm’s choice to promote SD (Airike et al., 2016).

Among the various contributes that have discussed the different motives for firms engage in CSR (Brønn & Vidaver-Cohen, 2009; Santos, 2011). a well-recognized and accepted classification identifies “legitimacy”, “sustainability”, and “profitability” as the three main corporate motives for CSR (Brønn & Vidaver-Cohen, 2009). “Legitimacy” refers to a positive perception of the firm, which is obtained by meeting social expectations of external observers, and which allows a firm to be legitimate by the public eye to operate; “sustainability” stands for personal ethics and managerial values that lead the firm to operate for preserving the natural environment and promoting the development of human society; “profitability” indicates the firm’s goal to obtain positive financial results with sustainability initiatives, either by generating new revenues or by protecting existing revenues (Brønn & Vidaver-Cohen, 2009).

However, the motives for engaging in CSR often overlap, making it difficult to clearly distinguish the various components. In this regard, this study has attempted to overcome this difficulty by observing firms’ behavior under institutional voids, as these challenging and non-enabling institutional settings (Amaeshi et al., 2016) better define the boundaries of the various motivations (Zhao et al., 2014). Institutional voids here refer to a situation where the formal institutional arrangements are too weak to support market economies or either absent (Khanna & Palepu, 1997, 2010). Previous studies have superficially investigated contexts characterized by institutional voids (Davila et al., 2018), even if such contexts may be interesting for their distinctive characteristics. Under institutional voids, the incompletely formulated and poorly enforced regulatory system (Khanna & Palepu, 1997, 2010), socio-environmental issues (Davila et al., 2018), and the weak development of local markets may create a permissive environment (Zhao et al., 2014). However, extant research has recognized the positive contribution of CSR in areas with limited infrastructure and poor social welfare programs, particularly concerning poverty alleviation and sustainable development (Amaeshi & Idemudia, 2015; Idemudia, 2009). Given these peculiarities, corporate motives for CSR will play a different role in institutional void contexts, thus having a different effect on SD. It is possible to expect that legitimacy will be in some cases less relevant, as firms are less subject to stakeholders' influence (Kim et al., 2018), while sustainability motives may be prioritized by firms’ willingness to have a positive impact (Jamali, 2010). Instrumental motives, instead, may be pursued by firms able to improve their revenue by engaging in social initiatives, however the scarce development of the market (Zhao et al., 2014).

Based on previous studies and consistent with SD as a CSR tool, the present research starts from the hypothesis that under institutional voids, a firm’s approach to SD varies depending on the different corporate motives for CSR. Hence, the following hypothesis constitutes the foundation of the analysis:

HP 1: Under institutional voids, corporate motives for CSR have a different effect on the effort of a firm in SD.

2.2 CSR, SD and ethics

Previous research highlighted that corporate motives for CSR are strongly affected by institutional mechanisms (Babiak & Trendafilova, 2011; Campbell, 2007), which influence how firms perceive and define CSR and how they consider stakeholders (Kumar et al., 2019). About SD, previous studies found that the characteristics of the institutional contexts in which companies operate may shape SD (Ferri et al., 2016). In the domain of institutional factors, ethics exert a dominant influence on how stakeholders are perceived (Jamali et al., 2017), and therefore are also able to shape SD.

Regarding ethics, the two most important approaches that can be individuate are the “libertarian” and the “solidaristic”. The “libertarian” approach, which relies on the principles of freedom, liberty, individualism, and self-responsibility, makes the argument for the traditional “stakeholder theory” (Freeman, 1984). The “solidaristic” approach, instead, relies on the principles of justice, solidarity, respect, empathy and caring (Woermann & Engelbrecht, 2019), and is mostly embedded in the Ubuntu culture. The Ubuntu ethics, indigenous to Sub-Saharan Africa, considers fundamental the good of the community, fostering generosity, compassion, and communality (Le Roux, 2000). This sense of community urges firms’ positive contribution towards the societies in which they operate and the need to take care of all stakeholders as an a priori moral responsibility (Lutz, 2009; Woermann & Engelbrecht, 2019),

This altruistic and communal vision of stakeholders (Okereke et al., 2018) is underrepresented in the academic debate (Jamali et al., 2017), as most studies refer to the traditional and ‘libertarian’ vision of stakeholders (Freeman, 1984).

Under this perspective, Sub-Saharan countries are a significant setting for analyzing the effect of the ethics on CSR and SD under institutional voids. The significant institutional voids that characterize these countries (Ahen & Amankwah-Amoah, 2018; Amaeshi et al., 2016) may allow firms to adapt to the incompletely formulated and poorly enforced regulatory system (Zhao et al., 2014), thus contradicting Ubuntu ethics.

According to previous studies, in Sub-Saharan Africa it is possible to analyze firms’ approach to CSR and SD according to ethics, distinguishing between local Sub-Saharan companies, influenced by Ubuntu ethical pressures, and foreign companies, driven by a libertarian culture (Lutz, 2009; Woermann & Engelbrecht, 2019). For local Sub-Saharan firms, as Ubuntu ethics imply a consideration of stakeholders related to moral responsibility and altruistic orientation, it is reasonable to expect that SD may be interpreted as a mean through which the moral responsibility of a firm is pursued and put into practice, thus prioritizing sustainability motives to engage in CSR. For foreign firms, mostly multinational companies (MNCs) from developed countries, the libertarian approach may induce a firm to consider scarcely relevant local stakeholders (Kim et al., 2018), thus limiting their effort in SD. Under institutional voids, the incompletely formulated and poorly enforced regulatory system, and the ethical standard conflicts between host and home countries (Donaldson & Dunfee, 1999) may allow or even incentivize foreign firms’ low efforts in CSR and SD (Zhao et al., 2014). However, exposure to institutional pressures from international markets, inter-governmental organizations, and parent companies will push foreign firms to focus on such legitimate and critical stakeholders (Jain et al., 2017; Park & Choi, 2015), thus prioritizing legitimacy and profitability motives to engage in CSR.

Starting from these considerations, the second hypothesis of the study can be defined as follows:

HP 2a: Under institutional voids, Ubuntu ethical pressures influence the effort of a firm in SD.

HP 2b: Under institutional voids, Ubuntu ethical pressures moderate the effect of corporate motives for CSR on SD.

2.3 CSR, SD and resources

Another variable that influences the CSR behavior of firms is the size of the firm, under the resource‐based view theory (Wernerfelt, 1984). The literature on the topic is vast, and many works have underlined the differences between large firms and small-medium enterprises (SMEs) when dealing with CSR-related topics. Extant research has verified that large firms are more eager to implement CSR than SMEs (da Silva Monteiro & Aibar‐Guzmán, 2010), since large firms have more resources, and subsequently can afford more easily the costs associated with CSR and dedicate more human and relational resources to the fulfillment of their socio-environmental aims (Russo & Perrini, 2010). SMEs encounter greater challenges than large firms when implementing CSR and can even perceived CSR as a burden and a threat if they cannot meet the requirements (Morsing & Perrini, 2009).

CSR in SMEs is mainly internally focused and mostly implemented through simple, easy, and cheap measures, also due to the lack of public support and the scarcity of resources available (Santos, 2011). Large firms, instead, are more oriented to externally related activities, such as reporting and stakeholder management (Baumann-Pauly et al., 2013).

Regarding motivations toward CSR, a previous contribution (Udaysankar, 2008) has verified that small and large firms have different motives to participate in CSR, even if equally strong, whereas medium-sized firms are the least motivated. Regarding SD, Morsing and Perrini (2009) suggested that SMEs depend highly on interpersonal relationships with different stakeholder groups, and therefore SD represents a fruitful opportunity. However, large firms are usually more subject to stakeholder scrutiny and, therefore, they implement SD activities as a reaction to or protection from potential critiques (Deegan & Gordon, 1996; Reimann et al., 2015).

Although the discussion is not exhaustive due to the extensive literature available, it confirms that firm size is a relevant and necessary variable to be considered in studies about CSR and SD.

Regarding the institutional voids’ context, previous studies verified that large companies, given the faulty implementation of laws (Visser, 2008), perceive external pressures lower than SMEs for the threats of litigation and punishments if non-compliance occurs (Kagan et al., 2003). Thus, weak regulatory and government systems may allow corporate misconduct to occur. However, it has also been pointed out that larger and more visible firms are more subject to stakeholders’ scrutiny, as they have more resources available to dedicate to CSR commitment in contexts characterized by institutional voids (Reimann et al., 2015).

Regarding the Sub-Saharan context, it is confirmed that most CSR initiatives have been implemented by large companies, whereas SMEs do not have the resources or mindset (Kaufmann & Simons-Kaufmann, 2016) and may perceive CSR practices as a luxury (Adeleye et al., 2020).

Therefore, considering the effect of firm size on corporate motives for CSR and SD, in a context with institutional voids, it is reasonable to expect that SMEs may focus on SD for profitability motives, putting an effort only on those practices that can provide the firm with a return. Conversely, large firms, due to the exposure to higher pressures from stakeholders and to the higher level of resources (Deegan & Gordon, 1996; Reimann et al., 2015), may also consider legitimacy and sustainability as motives to engage in SD.

Therefore, the present study postulates the following hypothesis:

HP 3a: Under institutional voids, firm size positively affects the effort of a firm in SD.

HP 3b: Under institutional voids, firm size moderates the effect of corporate motives for CSR on SD.

3 The peculiarity of Mozambique as a research setting

Mozambique has been considered stable and peaceful since the 90 s, after 16 years of civil war, even if political institutions, civil society, and the economic situation are still fragile (Kaufmann & Simons-Kaufmann, 2016) and terrorism still threatens the development of the country. Economic growth, which has been driven by some vibrant industries (e.g. construction, tourism, gas and coal explorations, agriculture and mining), has implied minor consequences in terms of economic development for local society because of bureaucracy, corruption, access to finance and weak public services, including poorly managed public investments (Kaufmann & Simons-Kaufmann, 2016). The competitiveness of local SMEs is low (Abbas et al., 2012), whereas large firms, both local and foreign, operate profitably thanks to mega-projects (Castel-Branco & Ossemane, 2010).

Regarding CSR initiatives, the number of projects and activities in the country is rising, but there is still no systematic approach or coordination through public authorities or the business sector itself. The government has attempted to promote broader firms’ commitment to CSR (Kaufmann & Simons-Kaufmann, 2016), for example, by making the CSR component compulsory by law in the mining sector since 2014, but has mostly focused on addressing things from a compliance perspective (Mzembe & Meaton, 2014). The pressures from civil society, both locally and internationally, have pushed companies to implement CSR initiatives for reasons strictly related to their business activities, such as training, health of workers, and relationships with the local communities.

In this context, SD may be interpreted as a means through which CSR can be put into practice, as a more useful and efficient instrument for development, suitable for solving the main issues related to institutional voids according to stakeholders’ needs. Stakeholder participation is mentioned in the Mozambican legislation and in development programs, such as the Pilot Program for Climate Resilience (PPCR), an important part of the Climate Investment Funds managed by the World Bank to support poorer countries in the fight against climate change (Shankland & Chambote, 2011).

Looking specifically at Mozambique, the dearth of existing studies does not allow for sketching a comprehensive understanding of the orientation and relevance of SD. However, to the best of the authors’ knowledge, an interesting suggestion comes from the research by Lucrezi et al. (2019), who aimed to provide empirical evidence of the impact of SD on marine protected area management in the southern village of Ponta do Ouro. The findings confirm that SD is effective in supporting management agendas and collaborative processes of knowledge acquisition, fieldwork, and educational experiences. Another study by Schut et al. (2013) considered SD as part of policy processes on biofuel sustainability in the country, thus confirming that it is a relevant component in the local governance and managerial domains. Other studies have included SD as a variable to consider in their analyses in Mozambique, with specific attention given to agriculture, energy, and water (Alba et al., 2017; Sánchez-Reparaz, 2020). These works validate that SD is increasingly related to the managerial culture in cross-sector organizations operating in Mozambique, thus highlighting idiosyncrasies and commonalities between the Ubuntu ethical pressures and the “libertarian” vision of stakeholders.

For this study, Mozambique is an interesting research setting, as it is characterized by a clash of various issues. Mozambique is one of the Sub-Saharan countries that most represents Ubuntu ethics (West, 2014), which has been historically characterized by significant institutional voids (Webster & Wood, 2005). However, the country has recently been undergoing strong economic growth (World Bank, 2017), shifting toward a liberal economy to attract new investments (Sartorius et al., 2011). Hence, the socio-economic situation of the country may allow firms to adapt to sub-standard business ethics or limit their effort (Zhao et al., 2014), thus contradicting Ubuntu ethics toward stakeholders.

4 Data and methods

4.1 Sample

The initial phase of the research was dedicated to identifying the sample of firms operating in Mozambique. When analyzing African countries, building the sample is a challenging task, as it is usually hard to access the list of existing firms, and there are few databases, usually dedicated only to listed firms. As there is not a universally accepted list of Mozambican firms, the sample has been individuated by combining different national databases, such as the list provided by the Center for Public Integrity of Mozambique (2014) and the Mozambican Yellow Pages (2014). The geographical concentration of the sample was necessary to ensure the feasibility of the data collection by surveying the presence at firms’ headquarters. Therefore, we chose to concentrate the sample in areas reachable by public transport, such as three major capital cities of the country, particularly Maputo, Beira, and Nampula. These three cities are representative of the entire country, as Maputo is located in the southern region, Beira in the central part of the country, and Nampula in the northern region. Starting from the universe of private firms, we filtered those firms nearest to these three cities, identifying a sample of 1066 firms.

From June to September 2014, we conducted an in-person survey by visiting the headquarters of the firms participating in the study to conduct the survey, as web-based surveys are not suitable in Mozambique for the lack of adequate technological infrastructures.

We invited the top managers of the selected firms to participate in the research by letter, institutional email, and then conducting three telephone recalls. Once the firm agreed to participate in the study, we scheduled a visit to the headquarters to fill out the questionnaire. We conducted interviews following a structured questionnaire with closed questions, most of which had to be scored on a 5-point Likert scale. The in-person survey facilitated the collection of data, allowing an understanding of the activities and cultures of the companies, thus limiting the distortion derived from the so-called "phenomenon of social desirability" related to the respondents' tendencies to answer in a way that may be criticized by others.

At the end of the data collection, 235 completed questionnaires were collected, with a redemption rate of 22.04%. Although moderate, the response rate is almost as high as that usually obtained in countries in which data collection is challenging, such as Mozambique.

The companies are grouped into micro, small, medium, and large companies. The research considered the number of workers solely as a criterion to be adopted because according to the country's cultural mores, it is taboo to speak of the turnover volume or annual income of a company, and there is a greater likelihood of receiving no response. In the final sample, micro-enterprises with fewer than ten employees represent 28.94% of the 235 companies surveyed, small companies between 10 and 49 workers represent 39.15%, medium companies of 50–249 workers are 17.02%, and large companies of 250–999 lie in the range of 9.79%. The class of the largest companies, including multinationals in Mozambique, is represented by 5.11% of those surveyed. Considering that current data on the composition of Mozambican firms indicate that SMEs represent the total number of firms in a range from 78 to 99% (Valá, 2009; World Bank, 2017), we can consider our sample representative of the Mozambican economic environment.

Most domestic industries are equally represented in the final sample, while, with regard to the geographical distribution of companies, respondent firms are almost equally distributed among the three cities included in the sample, with Nampula representing a slightly larger share (37.45%), in comparison to Maputo (32.34%) and Beira (30.21%). Finally, 14.47% of the interviewed companies are branches of Mozambican companies, 17.45% are branches of foreign companies, whereas 68.09% of companies have no parent company.

4.2 Data collection and sample

The process of data collection included a review of the literature on CSR and business analysis in Mozambique and the context of CSR in Mozambique. We adapted the research and subsequent analysis to the local context, running a pilot study with local professors and independent reviewers, to capture the peculiarities of the Mozambican context and to avoid potential distortions.

The questionnaire was translated from English into Portuguese and then back translated into English to avoid any bias resulting. We also combined back-translation with independent reviewers (i.e., parties other than the translators) who reviewed the translated questionnaire (Chidlow et al., 2014).

The survey was conducted in person, using a questionnaire structured with closed questions. The questionnaire consisted of four sections: (1) characteristics of companies and respondents: data on the firm (industry, location, employees); (2) corporate motives for CSR; (3) organizational structure of CSR; and (4) SD.

4.3 Dependent variable

Building on the discussed theoretical developments, to investigate the relationship between corporate motives for CSR, local culture, firm size and SD, the study included as dependent variables the factor “SD effort”.

4.3.1 SD effort

SD has been defined as practices that the organization undertakes to involve stakeholders positively in organizational activities. SD practices are present in many areas of organizational activity, from public relations to customer service, and from supplier relations to human resource management (Greenwood, 2007).

The study adopts a framework developed by Jamali (2008), considering five items as variables manifesting SD: “Dialogue with Customers”, “Dialogue with Suppliers”, “Dialogue with Shareholder Minorities”, “Dialogue with Investors”, and “Dialogue with Labor Unions”. The work by Jamali considered a broader range of stakeholders, but for this study, we included only the variables deemed appropriate for the Mozambican context. The respondents were asked to rank the importance of their engagement in the dialogue with each of the mentioned stakeholders by choosing a score from 1 to 5 (where 1 = not important, 2 = a little important, 3 = sufficiently important, 4 = important, 5 = very important).

A principal components analysis was conducted to reduce the items in one factor that would represent the original data much. The final factor’s internal reliability was acceptable based on the Cronbach’s alpha of 0.784; the sampling adequacy was reliable because the Kaiser-Mayer-Olkin (KMO) coefficients were over the cut-off value of 0.767.

4.4 Independent variables

Two groups of independent variables were built, starting from the existing literature: “CSR motives” and “Firm typology”, a combination of firm size and Ubuntu ethical pressures.

As with the dependent variables, each independent variable has been explored through multiple items based on the literature. Some items in the questionnaire were excluded from the final analysis based on the reliability analysis of their internal consistency.

4.4.1 CSR motives

Firm commitment toward CSR is strongly influenced by the corporate motives that firms pursue when dealing with CSR. The study builds upon a framework adapted from Brønn and Vidaver-Cohen (2009), which identifies three main motivations for CSR: legitimacy, sustainability, and profitability. Using the previous variable, we considered four variables expressing the legitimacy motive toward CSR: “Improve our image”, “Serve long-term company interests”, “Fulfill stakeholder expectations”, and “Be recognized for moral leadership”. The respondents were asked to rank the level of importance of each of the proposed “legitimacy motives” listed above by choosing a score from 1 to 5 (where 1 = not important, 2 = little important, 3 = sufficiently important, 4 = important, 5 = very important). The legitimacy motives variable was developed by running a factor analysis of the gathered data. In line with the above-mentioned study, a principal components analysis was done to reduce the items in one factor that would faithfully represent the original data. The principal components’ analysis reduced the items into one factor, whose reliability was confirmed (Cronbach’s alpha: 0.856; KMO: 0.805).

Following Brønn and Vidaver-Cohen (2009), we considered six variables expressing the sustainability motive toward CSR: “No good reason not to”, “Concern for Society’s future”, “Personal satisfaction”, “Strengthen global networks”, “Learn from social agencies”, “Share resources with society”. The respondents were asked to rank the level of importance of each of the proposed “sustainability motives” listed above by choosing a score from 1 to 5 (where 1 = not important, 2 = little important, 3 = sufficiently important, 4 = important, 5 = very important). A factor analysis of the gathered data was conducted to develop the sustainability motives variable. The principal components analysis reduced the items in one factor with an acceptable internal reliability (Cronbach’s alpha: 0.756; KMO: 0.799).

As done for the other two motives, we considered five variables expressing the profitability motive toward CSR: “Avoid regulation”, “Solve social problems better”, “Meet shareholder demands”, “Create financial opportunity”, “Remain competitive”. The respondents were asked to rank the level of importance of each of the proposed “profitability motives” listed above by choosing a score from 1 to 5 (where 1 = not important, 2 = little important, 3 = sufficiently important, 4 = important, 5 = very important).

Like previous variables, a factor analysis with the principal components method was conducted to develop the profitability motives variable: the items were reduced into one factor, which was considered reliable (Cronbach’s alpha: 0.802, KMO: 0.799).

Table 1 shows the details of variable construction.

Table 1 Emerging motive factors
Full size table

4.4.2 Firm typology

This study posits that the effect of corporate motives for CSR on the engagement of a firm toward SD under institutional voids depends on the ethical pressures (Okereke et al., 2018) and on the size of the firm (Morsing & Perrini, 2009). For this study, we codified firms subject to Ubuntu ethical pressures looking at the firm’s country of origin (Jamali & Mirshak, 2007, 2010; Wanderley et al., 2008). As Mozambique is one of the most consistent examples of Ubuntu ethics (West, 2014), we codified firms under the influence of Ubuntu ethical pressures as “Domestic” if the headquarters were in Mozambique. Given that all foreign companies analyzed in the study have their headquarters outside Sub-Saharan Africa, we preferred simplifying the codification by using the term “Foreign” for firms not under the influence of Ubuntu ethical pressures.

Regarding firm size, it was hard to collect specific data, as the informal economy is predominant in Mozambique and in most Sub-Saharan countries (Tchamyou & Asongu, 2017). Thus, most firms are reluctant to disclose specific data on firm revenues and on the total number of employees, because not all activities are formalized. Therefore, the pilot study highlighted that it was only possible to collect data on firm size asking respondents to indicate to which range of size they belonged among the following: “micro-enterprises” (less than ten employees); “small companies” (between 10 and 49 employees); “medium companies” (between 50 and 249 employees), and large companies (more than 250 employees). In line with previous studies (Russo & Perrini, 2010), we preferred simplifying the codification by using “SMEs” for firms with a total number of employees lower than 249, and “Large”, for firms with a total number of employees equal to or higher than 250.

Given the absence of continuous variables, due to the difficulty of collecting data in a context with institutional voids and informal economy, for this study we eventually developed the “firm typology” variable to highlight the conjunct effect of Ubuntu ethical pressures and firm size. According to this categorical variable, we identified these expressions of firm typology: “Domestic Large”, in which we included firms headquartered in Mozambique with a total number of employees higher than 249; “Domestic SMEs”, which collected firms headquartered in Mozambique with a total number of employees lower than 249; “Foreign Large”, in which we codified firms headquartered outside Mozambique with a total number of employees higher than 249. Our sample, however, did not cover the category “foreign SMEs”, because the few foreign SMEs identified were just representative offices of companies from foreign countries not directly operating within the country. We consider that the identified categories are suitable to faithfully represent the context of firms operating in Mozambique, as confirmed by previous studies in the country (Kaufmann & Simons-Kaufmann, 2016; Valá, 2009).

As existing literature indicates that in countries with institutional voids, such as Mozambique, CSR is usually less developed than in other countries and mostly performed by large companies (Jamali et al., 2017), we considered the case of “Foreign Large” firms as a basic condition to highlight peculiarities of the Mozambican context.

4.5 Control variables

Two control variables have been considered relevant to the SD effort. First, data on the industries where firms belong were collected; specifically, thinking as an essential condition the “Consumer Goods” industry, eight dummy variables were created for these sectors: Agriculture; Basic materials; Construction; Retail; Financial services; Telecommunication; Health; Services.

Also, earlier research suggests that the level of SD effort of a company on social and environmental issues depends on the presence of a CSR unit (Yuan et al., 2011), so a dummy variable was added to indicate whether the firm already has a CSR unit within its organizational form.

4.6 Analysis

Before the analysis was run, the existence of social desirability bias, common method bias, and non-respondent bias was checked. Extant research indicates that social desirability bias has become relevant in studies dealing with CSR (Roxas & Coetzer, 2012). To address this topic, we formulated questions focusing on the activities implemented by the firm, rather than on the personal judgment of the managers and provided the interviews with a 5-point Likert scale to answer. However, we cannot exclude that the top managers we interviewed have responded by looking for answers in line with the expectations of other companies or public institutions. Given this starting point, we specified in the invitation letter and the recalls that academic researchers and only anonymous researchers managed the survey and aggregated data would have been publicized.

The research method also had to account for common method bias, as the data were gathered from a single respondent through a unique questionnaire. According to the recommendations of Podsakoff et al. (2003), procedures were used to control for common method bias, such as eliminating item ambiguity, adopting previously tested questions, and reducing social desirability bias. Nevertheless, statistical control using Harmon’s single-factor test was deemed fundamental to guaranteeing the significance and reliability of the results. The results of the exploratory factor analysis show that the factor explained less than much of the total variance. This result suggests that common method bias does not represent a problem in the study. Also, the confirmatory factor analysis applied to Harman’s single-factor model showed fit indexes to be unacceptable and worse than those of the measurement model. This suggests that common method bias is unlikely.

Besides, non-respondent bias was also checked by conducting a set of t-tests to verify whether there were significant differences between respondent and nonresponding within the sample. The tests were run for the control variable in our research (Industry) and proved to be not significant, suggesting that non-respondent bias does not impact the results.

Finally, the study has revealed to be potentially affected by endogeneity, as the independent variables are broad and their effect on the dependent variable may be related to an omitted variable. However, due to the difficulties of collecting additional variables, the only feasible method to verify endogeneity was to use instrumental variable estimators (Semadeni et al., 2014), such as two-stage least squares (2SLS) and generalized method of moments (GMM). These methods did not provide the authors with significant results, as they require clearly identifying the cause of endogeneity and justifying the instrumental variable(s). It is therefore necessary to acknowledge that results and conclusions generalizability is limited as endogeneity concerns are present, even if recent studies (Hill et al., 2021) indicate that potential endogeneity concerns can still make the results suitable to advance the understanding of a topic, if the research setting is peculiar or challenging.

5 Results

Table 2 presents the means, standard deviations, and correlations of variables in our analysis. We found only a moderate correlation (0.27–0.39) between the variables related to motives to be involved in social issues (sustainability, legitimacy, and profitability). These dimensions were orthogonal, such that each motive may have its effect on the SD effort.

Table 2 Descriptive statistics
Full size table

After assessing the normal distribution of variables with Shapiro–Wilk tests, to test our hypotheses, we ran a set of hierarchical Ordinary least squares (OLS) regressions. To assess the sensitivity of the results to outliers, we performed an additional sensitivity analysis with a robust regression. The sensitivity analysis confirmed that the results were stable and in line with those of OLS regressions.

To assess multicollinearity before running regression, we analyzed the correlation matrix. Although the highest observed value in our correlation matrix is 0.48, we nevertheless conducted the additional diagnostic measures for multicollinearity suggested by Cohen et al., (2003, pp. 422–5): analysis of single tolerance/variance inflection factors (VIF) and mean VIF and assessment of the condition index. Since the VIF values of our variables range were all below 4.0 and the mean VIF is 2.34, the first analysis establishes that the model should have no serious problems with multicollinearity.

In addition, to test for heteroskedasticity, we first analyzed the plot of sample residuals against the fitted values, and no patterns where visible in that. In addition, we ran a Breusch–Pagan test, which confirms the absence of heteroskedasticity problems.

Eventually, to test our hypothesis, we relied on a hierarchical OLS linear regression analysis. We first evaluated control variables to verify the existence of a relationship between industry and CSR organizational unit on SD effort. As Table 3 shows in model A, industry and a CSR unit are significantly (positively or negatively) associated with SD effort (p < 0.05). This follows previous studies that have highlighted that SD efforts are contingent upon the industry within which firms are embedded (Jain et al., 2017).

Table 3 Regression of motives and types of organisation on SD effort
Full size table

As shown in model B of Table 3, there was a significant positive relation between profitability motive and the effort in SD (β = 0.21, p < 0.01). Next, in model C, firm typology is included, considering the case of “Large Foreign” as the basic model. Results indicate that Ubuntu ethical pressures have a significant effect on SD only for small domestic firms (β =  − 0.62, p < 0.05), but the effect is negative. Third, the direct impact of Ubuntu ethical pressures on SD is negative and significant (β =  − 0.26, p < 0.05) for small domestic firms also in model D, whereas in this intermediate model corporate motives appear to have different effects (to what previously observed). Finally, the complete model shown in the Table (model E) confirms that corporate motives have different and significant effects on SD. SD is positively influenced by sustainability (β = 0.43, p < 0.05) and profitability (β = 0.71, p < 0.05) motives, while legitimacy motives (β =  − 0.52, p < 0.05) have a negative effect. Hypothesis 1, which predicted that corporate motive to engage in CSR would affect SD effort, is confirmed. Besides, the effect of Ubuntu ethical pressures, represented by the two “domestic” firm typologies, is not consistent nor significant, thus rejecting hypotheses 2a and 2b. Simultaneously, the effect of firm size on SD is confirmed, as SMEs consistently engage less in SD (β =  − 0.42, p < 0.05), so this result confirms hypothesis 3a. Finally, the interaction between SMEs and sustainability motives (β =  − 0.54, p < 0.01), and between SMEs and legitimacy motives (β = 0.55, p < 0.01) are significant, partially confirming that firm size influences the effect of corporate motives on SD, as predicted in hypothesis 3b.

To better interpret the significant interactions, we plotted the results. Figure 1 depicts a strongly positive relationship between sustainability motives and SD effort for large firms, whereas the relationship is negative in smaller firms. In Fig. 2, the relationship between legitimacy motives and SD is consistently negative for large companies, while it is positive for SMEs. Finally, as shown in Fig. 3, the profitability motives have a positive relationship with SD effort in all the organizations we considered. These results graphically support hypothesis 3b, indicating that the effect of legitimacy and sustainability motives hinge greatly depending on firms’ size, whereas Ubuntu ethical pressures are not relevant.

Fig. 1
figure 1

The moderating effect of sustainable motivation

Full size image
Fig. 2
figure 2

The moderating effect of legitimacy motives

Full size image
Fig. 3
figure 3

The moderating effect of profitability motives

Full size image

6 Discussion

The present study represents an original work in the literature for at least two significant reasons. First, it explores how corporate motives for CSR drive SD, thus proposing one of the few quantitative studies on SD and advancing the understanding of its dynamics. Second, it focuses on Mozambique, one of the Sub-Saharan countries that most represents a context with institutional voids (West, 2014), and a context where Ubuntu ethical pressures are contrasted by the socio-economic situation of the country.

The findings from the analysis of the 235 Mozambican firms provide interesting insights into the relationships between CSR motivations and SD. First, the study has highlighted that, in a context where institutional voids allow firms to adapt to sub-standard business ethics or limit their effort (Zhao et al., 2014), corporate motives for CSR commitment exert different effects on SD. Hypothesis 1 is thus supported. More in detail, among the three CSR motivations considered in the study—namely sustainability, legitimacy, and profitability—the third one has consistently affected SD. In this sense, the overall sample seems to engage in SD to pursue instrumental purposes in their CSR commitment. It is also interesting to note the negative effect that CSR motivations related to legitimacy have on SD. According to these results, the more firms are committed to CSR to gain legitimacy, the lower the effort in constructing a dialogue with stakeholders. This finding is opposite to most of the previous studies, where the search for legitimacy was identified as one of the primary drivers of SD. A possible explanation could be that, under institutional voids, firms prefer implementing initiatives with positive impact on performances rather than pursuing a better perception of the firm through responding to stakeholders’ expectations. Finally, sustainability motives appear to exert a positive influence on SD, even if not consistent across intermediate models. In summary, the findings support the idea corporate motives for CSR shape SD as a tool to contribute to firm purposes.

The second hypothesis was not confirmed, as the data did not evidence significant and consistent effects of Ubuntu ethical pressures. The stream of literature that looks at SD under institutional voids to demonstrate moral duty, as in Ubuntu ethics, is not reflected by the results of the study. The wider consideration of stakeholders pursued by Ubuntu ethics does not determine per se an engagement in SD, nor a peculiar effect of CSR motives on the latter. Notwithstanding the a priori moral responsibility toward all stakeholders, the Ubuntu vision of stakeholder does not push all local firms to engage in SD, whereas foreign companies, expressing a libertarian vision of stakeholders, may engage in SD to catalyze attention and support only from selected stakeholders, mostly international.

Finally, the study verified the third hypothesis on the role of firm size on the relationship between corporate motives for CSR and SD. Here, the statistical results support what was postulated in the theoretical section, highlighting that large and small firms have a different pattern of behaviors. Large firms in the sample present a similar commitment to SD, whereas smaller firms differ significantly. As hypothesized, large firms have more resources to dedicate to CSR and SD implementation, to define and operationalize their CSR commitment. Large firms may also have more extensive knowledge and long experience with CSR and SD practices, thus increasing the probability of success of their efforts. Smaller firms, instead, seem less sensitive to CSR motivations and exert lower efforts in SD. This result is in line with existing literature, which suggests that the scarcity of resources that characterizes smaller firms usually limits their commitment to CSR and SD. The investments to sustain the extended network of relationships that SD requires may prevent SMEs from engaging in complex SD activities, especially under institutional voids, where smaller firms may prefer reducing the risk of wasting vital and scarce resources in activities that may obtain limited results. As the data gathering demonstrated, implementing efficient systems to collect information and maintain constant contact with external stakeholders requires considerable energy and resources, especially when rural areas must be reached. However, it is not actually possible to claim that local SMEs have no interest in CSR issues, but only recognize that there is not an explicit manifestation of this interest within all business activities (Painter-Morland & Dobie, 2009). Besides, the analysis of the interaction between corporate motives for CSR and firm typology indicates that for SMEs sustainability motives are less relevant for stimulating SD. As already pointed out, the activities of SMEs under institutional voids are considered suitable to embody social value per se, given the positive impact on the economic development of the local community. Thus, for smaller enterprises engaging stakeholders is less relevant to perform their social duty, whereas it becomes important when it is necessary to achieve legitimacy of their operations. In fact, results indicate the legitimacy motives are positively related to SD for SMEs, differently from large firms. SMEs may be interested in establishing legitimacy in local markets, satisfying the expectations from local communities, customers, and public institutions, as the pressures of these stakeholders are critical for such firms. Conversely, large firms are less vulnerable to local stakeholders' influence, so that signaling their legitimacy by CSR engagement may be irrelevant, or at least less critical (Kim et al., 2018).

7 Conclusion

The study has explored whether and how different motives to CSR commitment determine different approaches to SD, considering a large sample of 235 local and foreign firms operating in Mozambique. As previously discussed, the country has represented an interesting setting for the considered topic, since it is characterized by institutional voids thus allowing the investigation of firms’ behavior in non-enabling context, and by Ubuntu ethics, which roots firm responsibility in the value of caring of others and commonality as an a-priori moral duty.

Despite a very peculiar research setting, which does not allow generalization of the results, the study also aims to offer some insights on SD. First, the analysis showed that profitability is the most consistent corporate motive for CSR suitable to drive commitment to SD, whereas legitimacy under institutional voids may have a negative effect. Previous literature mainly focused on the investigation of how SD influences CSR strategy, but limited attention has been paid to the analysis of whether corporate motives to engage in CSR determine different approaches to SD. Thus, the first result of the study indicates that corporate motives for CSR are a key antecedent of how firms operationalize SD. Second, the study built on extant literature and verified whether Ubuntu ethical pressures and firm size exert an influence on SD. In line with extant works, firm size is a significant factor that determines firm commitment to SD, with large firms being more dedicated to this tool. Interestingly, Ubuntu ethical pressures were not found to be significant, contradicting previous research on this peculiar vision of a firm’s stakeholders and opening up possible future investigations.

In summary, this study offers significant contributions to literature development. First, it advances the knowledge around SD, by exploring how corporate motives for CSR affect the systematic and structured interaction with stakeholders, and the concurring effect of ethics and resources. Besides, most of the extant works on SD are theoretical, whereas quantitative articles are very limited (Ferri et al., 2016; Habisch et al., 2011). Hence, the study presented here proposes a new approach to the topic, discussing the results of the statistical analysis of a broad database derived from 235 firms operating in Mozambique, thus stimulating more quantitative research in the field. A second contribution is related to the setting of the study, namely Mozambique. As international literature on SD is still scarce, even in developed countries, the present work aims to fill this gap by adding to the understanding of the dynamics behind SD in a context affected by institutional voids. To the best of the authors’ knowledge, this is the first quantitative study to explore firms’ CSR behaviors and SD in Mozambique. Most of the studies have been managed in more advanced sub-Saharan African countries such as Kenya (Muthuri & Gilbert, 2011), Ghana (Amponsah-Tawiah & Dartey-Baah, 2016), Nigeria (Ite, 2004), Uganda (Katamba & Nkiko, 2016), whereas poorer countries have been overlooked, mainly due to the difficulties related to data availability and collection. The present study aimed at filling this gap to contribute to furthering the understanding of emerging phenomena that will drive corporate behavior and societal progress in backward contexts. Third, the study sheds a light on the role of ethics in determining a firm’s vision of stakeholders. The study takes into consideration Ubuntu ethics, which propose a vision of stakeholders alternative to the traditional and “libertarian” vision usually considered in the academic debate. In addition, the Mozambican context allowed the authors to consider the effect of Ubuntu ethics, which imply firms’ moral responsibility toward all stakeholders, in vibrant economic system, which stimulates business ventures, thus potentially creating a clash with the prevailing culture. Finally, the paper extends the understanding of CSR and SD indicating that the availability of resources influences directly and indirectly firms’ commitment to SD.

While we identify value in the findings, we also acknowledge several limitations related to the current study. First, the results of the study may be affected by endogeneity issues, due to the nature of independent variables collected in informal economies and the reluctance to disclose specific data on firm size. Thus, the measurement of some variables and the dynamism of some results may signal an omitted variable (Semadeni et al., 2014). Although the authors’ effort in managing this potential issue by making use of instrumental variables estimators (namely 2SLS and GMM models), the results were not consistent. It is therefore necessary to acknowledge that results and conclusions are contingent on these limitations. However, given that endogeneity concerns are present in virtually all social science research (Hill et al., 2021), the peculiarity of the research setting, and the novelty of the study still make the results suitable to advance understanding of CSR and SD under institutional voids. Future studies using different designs may shed light and potentially address this endogeneity threat if local conditions in Mozambique are adequate. Given the conditions under which the study has been realized, we consider this potential threat not enough relevant to prevent using data from such a peculiar and interesting context (Hill et al., 2021), as they may advance knowledge in the field and have implications for future research.

Second, we recognize that the non-response bias is a concern in survey research, and our study was not an exception, as approximately 22 per cent of the firms we contacted did not respond. However, the rate of response we had is suitable for countries in which data collection is challenging, such as Mozambique.

A third limitation of the research design was the existing restricted literature on practical measurement of SD. The lack of reliable scales pushes us to measure the firms’ effort in SD by a factor analysis that did not be tested in earlier studies. We adopted all the methodological steps to guarantee the reliability of our scale, and, because of that, our study contributes to the development of measurements for the operationalization of SD. Additionally, our hypotheses were only supported based on self-report measures. This raises the possibility that some of our obtained associations could be due to common method variance. Unfortunately, the problem of method variance cannot be ruled out in our data, but, as explained in the methodology section, we adopted the necessary procedure.

A fourth limitation deals with the focus of the study on a single country, namely Mozambique. This could raise concerns about the absolute generalizability of our results for Africa. Because our research questions were about the motivations and behaviors of foreign and domestic firms, we felt it was appropriate to constrain the sample to a single country and to pursue a high response rate from this select group. Overall, this paper should be a basis for further investigations in other African countries to study the relationship between CSR motivations and SD practices.

The results may have some implications for future research. They show that research on the motivation of CSR and practices is still far from maturity, and it holds substantial promise for further investigation. We focused our study on the effect of motivation on SD effort, which is one of the fundamental CSR practices, but we did not consider other relevant CSR policies and practices. We hope the findings from our analysis can motivate future research to cover a broader range of CSR practices, such as social accountability, strategic philanthropy, adoption of social certification, and social initiatives toward employees.

Our results suggest that future research must also examine additional moderating effects in the relationship between corporate motives for CSR and SD. For instance, theoretical literature on motivation indicates that also the personal features of managers often play a relevant role in developing CSR practices. Future research, for example, could assess if and how managerial attitudes, social-cognitive perspectives, and leadership styles would affect the relationship between motivations and practices.

Also, future research is needed to examine the context dependence of results and examine variation in the shape of the African country as the criterion variable. By including firms operating in more countries, a prominent field of future research could assess the impact of institutional (coercive, normative, and mimetic) pressures or domestic, national culture on the relationship between motivation and CSR practices. Finally, future research may minimize endogeneity threats through design, avoiding ex-post interventions that may be difficult to realize.