1 Introduction

NPOs are non-lucrative associative forms that constitute the Third Sector (Seibel 1990; Anheier 2006). NPOs occupy “a distinctive social space outside of both the market and the state” (Salamon and Anheier 1997, p. 1) to overcome the emergence of market and Government failures (Gangi and Trotta 2015). In particular, NPOs’ economic activities must be instrumental in the pursuit of social goals (Hansmann 1980; Arena 2009). NPOs’ main outcome refers not to specific products or services but to a positive developmental change both to individuals and society in general (Drucker 2004). For this reason, NPOs should be interpreted as economic entities that internalize the expectations of their stakeholders in defining and developing strategies. Stakeholders are here defined as critical influencing groups showing a strong interest toward NPOs, such as “foundation, corporation, and federated funding officials; individual donors and volunteers; government agents who contract with NPOs; and board and staff members” (Herman and Renz 1998, p. 24). Particularly, NPOs are driven in light of the interests of the “community-at-large”—interpreted as the surrounding community as a whole that becomes involved in how well NPOs fulfil their mission (Brown 2002; Low 2006; O’Mahony 2007).

Scholars have progressively witnessed an expansion of the Third Sector and, today, NPOs represent a significant societal actor—along with the Government and for-profit organizations (FPOs)—in the development of the socio-economic scenario (Hansmann 1980; Anheier 2006; Hinna and Monteduro 2016). Considering this key role of NPOs in modern economy, their governance has become a growing and sizable topic in the non-profit literature, which aims at identifying the specific governance system able to ensure efficiency, transparency, and accountability for stakeholders (Cornforth 2012). Actually, there has been a burgeoning stream of literature assessing governance in the non-profit context, stressing its complex nature (Dyl et al. 2000; Mersland 2011). However, as underlined by Ostrower and Stone (2006), governance is not clearly defined in the realm of NPOs and the concept is still scattered (see also Cornforth 2012). Considering that NPOs’ governance problems are similar to those of FPOs and that the term “corporate” in corporate governance does not refer exclusively to FPOs (Jegers 2009), it is possible to apply various definitions of corporate governance to the specific context of NPOs. In this study we specifically use the widely acknowledged definition of Tirole (2001), “corporate governance is the design of institutions that induce or force management to internalize the welfare of stakeholders” (p. 4); this concept is consistent with Anheier (2006) who underlines how the corporate governance system is necessary to distribute rights and responsibilities among different categories of stakeholders. Consistent with the pertinent literature, we refer to NPO governance effectiveness as the board structure, processes, and tasks allowing the effective functioning of the organization (Bradshaw et al. 1992; Brown 2002; Cornforth 2001, 2012).

In line with FPOs, it is possible to distinguish two categories of NPOs’ stakeholders: internal and external. While the former are board members, managers, employees and volunteers, the latter refer to donors, funding sources and beneficiaries (Petkus 2001; Liston-Heyes and Liu 2013). Among NPOs’ stakeholders, scholars have mainly focused on the board, particularly on two types of relationships: (a) board and management and (b) board (governance) effectiveness and organizational effectiveness (Ostrower and Stone 2006; Jegers 2009; Cornforth 2012). In fact, Ostrower and Stone (2006) pointed out that the main areas of NPOs’ governance research refer to board composition, the relationship between board and management, the functions and responsibilities of the board, the effectiveness of the board, and the link with organizational effectiveness. Jegers (2009) highlighted that most NPOs’ governance studies have followed the theoretical model of agency theory, focusing on the agency relationship between board and management. More recently, Cornforth (2012) pointed out that few researches have focused on critical stakeholders of NPOs, such as donors, funders, beneficiaries and volunteers. Therefore, the pertinent literature agrees that it is necessary to deepen the study on these stakeholders within NPOs’ governance, with particular reference to their interests, relationship with the board, and commitment (Jegers 2009; Cornforth 2012; Van Puyvelde et al. 2012). However, the relationship between NPOs’ governance and internal stakeholders—such as volunteers—has received scant attention, particularly with reference to the role the latter may play in allowing NPOs to implement sustainable and effective strategies (Zollo et al. 2016, 2017a).

Building on this, the present study attempts to contribute to NPOs’ governance literature by proposing and empirically testing a conceptual model that builds on an “enlightened” stakeholder theory perspective (Jensen 2001; Tricker and Tricker 2015), which takes into consideration both agency and stakeholder theories (for a review see Caers et al. 2006; Jegers 2009; Van Puyvelde et al. 2012). Focusing on the agency relationship among NPOs’ boards and internal stakeholders such as volunteers (Jegers 2009; Cornforth 2012), our aim is to unpack the underlying mechanisms between governance effectiveness and organizational effectiveness. Specifically, this paper takes an individual perspective of analysis to theorize and empirically demonstrate how volunteers’ perceptions of NPOs’ governance significantly impact on their individual commitment and identification with the NPO, resulting in higher perceptions of the organizational NPO effectiveness. The empirical results stress the significant role of both organizational identity (OI) and volunteers’ commitment as mediating variables, thus providing managerial and practical implications for board members and managers of NPOs.

Another important contribution of this study is related to assessing NPOs’ governance mechanisms in a scarcely investigated geographical area, i.e. Italy, which has an important and ancient tradition concerning the Third Sector (Evers and Laville 2004; Manetti et al. 2017; Thomas 2004; Zollo et al. 2017a). In fact, another important limitation of the NPO governance literature is that its main focus is on large NPOs located in the United States (McClusky 2002) and the UK (Mason 2010), largely ignoring the analysis of other geographical contexts—such as Italy—where the Third Sector plays an important role at the societal level.

1.1 Governance effectiveness and organizational effectiveness

The literature on NPOs’ governance stresses the relevance of the influence of governance effectiveness on organizational effectiveness (Herman and Renz 2004; Brown 2005). On the one hand, governance effectiveness is essentially interpreted as board effectiveness (Cornforth 2001; Herman and Renz 2004; Brown 2005; Hoye and Doherty 2011). Specifically, Cornforth (2001) associates board effectiveness with its functions, analyzing the relationship between the board structure and processes, and the board’s ability to conduct specific functions. Building on this, we focus on the following three responsibilities to capture volunteers’ perceptions of NPOs’ governance effectiveness, mainly because such responsibilities are particularly significant and representative functions for the purpose of the present study: (1) strategy and policy making, which refers to “setting the organisation’s mission and values, reviewing and deciding the organisation’s strategic direction, setting organisational policies”; (2) stewardship, referring to “overseeing the financial management of the organisation, ensuring the organisation has adequate financial systems and procedures, monitoring organisational performance and taking action when required”; (3) external relations and accountability, which refers to “ensuring that the organisation fulfils its legal obligations, representing the interests of stakeholders in the organisation, representing the organisation externally” (Cornforth 2001, p. 10; see also Cornforth 2012). This allows us to define the “well-functioning” of NPOs’ boards as able to significantly affect the organizational outcomes and, thus, the overall organizational effectiveness (see Brown 2005; Herman and Renz 2008; Jegers 2009).

On the other hand, organizational effectiveness can be defined as a type of performance measurement (Starbuck 2004). Many scholars have emphasized the difficulty in defining a clear and unique metric for this variable in NPOs because of the lack of the same financial assessments used by FPOs, such as shareholders’ remunerations (Kaplan 2001; Parisi 2013). However, the multifaceted nature of NPOs brings with it the development of a “multidimensional” approach—which refers to the integration of financial performance, social effectiveness, and institutional mission pursuit—through which their organizational effectiveness is analyzed (Brown 2005; Herman and Renz 2008; Hinna and Monteduro 2016). Consistently, in the present research we investigate volunteers’ perceptions of NPOs’ organizational effectiveness from two points of view: organizational responsiveness (Herman and Renz 2004) and social effectiveness (Brown 2005). While the former is fundamental in analyzing the ability to satisfy stakeholders’ expectations and pursue NPOs’ objectives, the latter expresses NPOs’ ability to achieve social goals. Specifically, organizational responsiveness has been used by Herman and Renz (2004) to indicate compliance with the institutional mission and stakeholder expectations; the dimension of social effectiveness is instead related to the quantity and quality of services and the ability to pursue social goals (Brown 2005).

1.2 Volunteers’ commitment and organizational identity in NPOs

A critical category of NPOs’ internal stakeholders is represented by volunteers whose presence is essential for the survival of many NPOs that are primarily or exclusively based on their activity in the pursuit of their institutional mission (see Hustinx 2005; Kreutzer and Jäger 2011; Manetti et al. 2015; Zollo et al. 2017a). Volunteers are defined as people who invest their free time to bring benefits to third parties, without receiving any economic reward (Musick and Wilson 2007; Wilson 2012). Jegers (2009, p. 157) highlighted that “although academic economic and managerial literature on volunteers abounds, it rarely pertains to topics dealing with the functioning of nonprofit organizations within agency or governance frameworks” (see also Sampson 2006). Hence, a focus on volunteers’ perceptions of NPOs’ governance and organizational effectiveness might provide relevant insights for the management of NPOs both at a theoretical and practical level (see Zollo et al. 2017a).

Most of the literature regarding volunteers in NPOs focuses on studying their motivations, relations with paid staff and employees, organizational support, or human resource management practices (Wellens and Jegers 2014). Particularly, the pertinent literature has widely stressed the relevance of volunteers in NPOs due to the importance of their work at the operational level (Wellens and Jegers 2014; Zollo et al. 2017a). Hence, an important critical issue in NPOs is the need to enhance volunteers’ commitment; thus avoiding their turnover intention, which may cause loss of know-how, quality reduction, and other operational difficulties (see Clary et al. 1992; Hustinx et al. 2010). According to Hustinx (2005), volunteers’ commitment may be theorized as a multidimensional construct composed of different styles of volunteering, expressed in terms of loyalty, devotion and choosiness. Loyalty refers to volunteers’ strong self-identification with the NPO; devotion is the feature of volunteers who commit themselves unconditionally following the NPO mission; finally, choosiness refers to volunteers’ availability to engage in all the NPOs’ activities regardless of their interest or willingness (see also Hustinx et al. 2010).

Several researches emphasize the need for NPOs to effectively manage volunteers, for example through constant training and development of organizational identification, which positively impact on volunteers’ satisfaction and commitment (Costa et al. 2006; Cuskelly et al. 2006; Garner and Garner 2011; Manetti et al. 2015). OI has been seminally defined by Mael and Ashforth (1992, p. 103) as “a perceived oneness with an organization and the experience of the organization’s successes and failures as one’s own”. Like employees, volunteers may “choose activities consistent with the salient aspects of their identity and support institutions incorporating such identities” (Ashforth and Mael 1989, p. 25; see also Golden-Biddle and Rao 1997; Kreutzer and Jäger 2011). Hence, both volunteers’ commitment and identification with the NPO emerge as relevant organizational factors to which scholars and managers should pay attention when defining and developing NPOs’ governance and related strategies.

2 Hypotheses

Our hypothesized model (see Fig. 1) builds on an “enlightened” notion of stakeholder theory (Jensen 2001; Tricker and Tricker 2015) which integrates agency and stakeholder theories (Jegers 2009; Cornforth 2012; Van Puyvelde et al. 2012). Jensen (2001) seminally argued that “Enlightened stakeholder theory, while focusing attention on meeting the demands of all the important corporate constituencies, specifies long-term value maximization as the firm’s objective. In so doing, it solves the problems arising from the multiple objectives that accompany traditional stakeholder theory, by giving managers a clear way to think about and make the tradeoffs among corporate stakeholders” (p. 9; italics added). Once applied to the non-profit context, such a perspective allows us to consider the agency relationships concerning all the stakeholders of an NPO, particularly emphasizing the role of volunteers as crucial internal stakeholders of the organization (Van Puyvelde et al. 2012; Wellens and Jegers 2014). Thus, the enlightened stakeholder theory emphasizes the need for NPOs’ management to effectively take into consideration both external and internal stakeholders—such as volunteers—when dealing with governance and strategic issues.

Fig. 1
figure 1

Hypothesized conceptual model

Full size image

2.1 The role of volunteers’ commitment

Given that NPOs’ governance literature shows that governance effectiveness influences organizational effectiveness (Herman and Renz 2004; Brown 2005), the present study empirically assesses whether volunteers’ commitment results in being a significant underlying mechanism in such a relationship.

The previous review of NPOs’ governance literature has shown that studies in this area are focused on the agency relationship between board and management (Ostrower and Stone 2006; Jegers 2009; Cornforth 2012). According to this perspective, the literature on NPOs’ governance highlights that few researches have focused on specific stakeholders, such as volunteers, who represent critical stakeholders for the NPOs (Jegers 2009; Cornforth 2012). However, most of the literature regarding volunteers in NPOs focuses on the influence of motivations and organizational support on their commitment (Manetti et al. 2015; Zollo et al. 2017a). For the purpose of the present study, the relationship between governance effectiveness and volunteers’ commitment is interpreted according to the participation-responsiveness mechanisms characterizing the governance (Eckstein and Gurr 1975). The literature points out that a governance able to create common sense within the organization is an important tool to promote commitment (O’Reilly and Chatman 1986) and engagement (Gruman and Saks 2011). Indeed, stakeholders’ participation is influenced by the responsiveness of those who govern (e.g. board members), which results in the ability to make decisions in line with stakeholders’ (e.g. volunteers) expectations and needs. Specifically referring to the context of NPOs, if volunteers perceive an alignment between their own values and the NPO’s mission, according to the agency theory perspective (Jegers 2009; Tricker and Tricker 2015) this will lead to an enhancement of their organizational involvement, and hence commitment (Hustinx 2005; Kreutzer and Jäger 2011). In line with these considerations, we hypothesize the following:

H1a

Volunteers’ perceptions of governance effectiveness positively influence their organizational commitment.

NPOs’ organizational effectiveness has been traditionally investigated from two points of view: organizational responsiveness (Herman and Renz 2004) and the social effectiveness developed by Brown (2005). The former highlights the compliance with the institutional mission and stakeholder expectation, whereas the latter is related to the quantity and quality of services and the ability to pursue social goals. Social effectiveness and organizational responsiveness are appropriate to investigate organizational performance; actually, the economic and financial performances do not represent the main NPOs’ goals, but only represent an instrumentally necessary condition in the pursuit of social objectives and institutional mission (Brown 2005; Herman and Renz 2004). For-profit scholars have traditionally highlighted a positive relationship between the organizational commitment of employees and organizational effectiveness (Lee and Miller 1999; Abdul Rashid et al. 2003). By integrating agency theory with some insights of stakeholder theory (see Caers et al. 2006; Jegers 2009; Van Puyvelde et al. 2012), we argue that volunteers’ commitment to NPO—thanks to its sub-dimensions of loyalty, devotion, and choosiness—has a significant impact on both the fulfilment of the NPO’s institutional mission (e.g. organizational responsiveness) and the ability to accomplish social objectives (e.g. social effectiveness) (see Brown 2005; Herman and Renz 2004; Kreutzer and Jäger 2011). Therefore, the present research aims to study, in the non-profit context, whether volunteers’ commitment positively influences organizational effectiveness. Hence, we hypothesize that:

H1b

Organizational commitment positively influences volunteers’ perceptions of NPOs’ organizational effectiveness.

2.2 The role of organizational identity

According to our “enlightened” agency theory perspective (Jensen 2001), the commitment of the agent (e.g. volunteer) to pursue the interests of the principal (e.g. NPO management) may be strengthened by a high OI (Mael and Ashforth 1995; Van Puyvelde et al. 2012). Actually, a high OI promotes principal-agent relationships where volunteers commit themselves to pursue the interests of the NPO management, thus strengthening their identification with the organizational mission (Van Puyvelde et al. 2012). Therefore, NPO governance should enhance OI to facilitate volunteers’ commitment. Through the stakeholder theory lens (Van Puyvelde et al. 2012), governance seems capable of generating OI through the definition of the organization’s vision, mission and values (Cornforth 2001). In this way, an effective governance is able to create a shared value within the organization, thus increasing OI (O’Reilly and Chatman 1986). Based on the above, we hypothesize that:

H2a

Volunteers’ perceptions of governance effectiveness positively influences their OI.

At the same time, a high OI positively influences organizational performance (Dutton et al. 1994; Rupp et al. 2006). In greater detail, Voss et al. (2006) highlighted the positive correlation between OI and organizational performance in NPOs. More recently, nonprofit scholars found that managers of NPOs should invest in OI in order to improve the effectiveness of their attracting, recruiting, and retaining strategies (Costa et al. 2006; Garner and Garner 2011; Zollo et al. 2018a). Hence, if the management of an NPO is able, thanks to its values and mission, to satisfy the volunteers’ expectations, this will result in a higher organizational identification and intention to stay with the organization (Kreutzer and Jäger 2011; Zollo et al. 2017a). As a result, the following is hypothesized:

H2b

OI positively influences volunteers’ perceptions of NPOs’ organizational effectiveness.

Moreover, the identity of an organization influences the way employees, and more generally internal stakeholders, feel about the organization, thus enhancing employees’ commitment (Dutton et al. 1994; Rupp et al. 2006). Applying this to the context of NPOs, OI results as being very important for volunteers to develop a feeling of “oneness” with the organization, a type of psychological and emotional bond that NPOs’ managers have to increase in order to maintain both mutual support and a symbiotic relationship with internal stakeholders (see Rupp et al. 2006; Van Puyvelde et al. 2016). In addition to this, the literature shows that stakeholders’ commitment is influenced by organizational identification and, particularly, by the internalization of the organizational values and goals (for a review see O’Reilly and Chatman 1986; Gruman and Saks 2011; Ng and Wyrick 2011). Hence, we propose that the more OI is developed in the context of NPOs—expressed by the values, tradition, and culture of the organization (Golden-Biddle and Rao 1997; Young 2001)—the more volunteers will be willing to actively participate in the organization’s activities, thus showing higher levels of commitment. We therefore hypothesize:

H3

OI positively influences volunteers’ commitment.

Based on the above, we propose OI and volunteers’ commitment as significant mediators of the relationship between NPOs’ governance and organizational effectiveness. We therefore hypothesize:

H4

The relationship between perceived governance effectiveness and organizational effectiveness is mediated by both OI and volunteers’ commitment.

3 Research methodology

3.1 Sampling

The sampling frame is comprised of charitable organizations based in Italy, specifically focused on religious voluntary associations named Confraternite della Misericordia (“Brotherhoods of Mercy”). These charities have an ancient tradition, dating back to 1244, and a crucial role in the regional socio-health service system (Bagnoli and Megali 2011; Zollo et al. 2018a). Specifically, “the Confraternita of Misericordia was a charitable society that provided relief to the poor in an urban setting” and “represented a means through which the citizenry could participate in the social and political life of the community” (Manetti et al. 2017, p. 511). As a result, volunteers’ commitment, participation, and organizational identification in Italy are extremely high, thus representing an interesting sample for the purpose of the present study (see also Zollo et al. 2017a, 2018a).

A ten-page questionnaire incorporating 39 items on volunteering was devised early in 2016. Measures originally devised in English were translated into Italian using the translation/back-translation procedure. Initially, the questionnaire was tested in a Tuscan university; a student panel composed of 15 respondents involved in volunteering activities checked it for clarity of formulation, ease of completion, and possible ambiguity. After this pretest, no substantial changes were made to the final questionnaire, which was e-mailed or submitted directly to volunteers of 50 Italian charitable organizations (during the period March-August 2016). To decrease the social desirability bias and the related common method variance (Spector 1994), we stressed in the covering letter that volunteers’ participation was anonymous, no compensation was provided, and there were no right or wrong answers, and gave reassurance about the maximum confidentiality and independency of researchers. In addition to this, we followed Podsakoff et al.’s (2003) suggestion, by separating the questions in the survey related to the independent variables—such as governance effectiveness and OI—from those relating to the dependent variables—such as volunteers’ participation in NPOs and organizational effectiveness.

Thanks to the direct collaboration of charities’ presidents and directors, an electronic link to the survey was emailed to 1000 volunteers by the charities’ board located in ten Italian provinces. Following Cornforth (2001, pp. 220–221), our questionnaire was sent to a specific person (President or Director of the NPO) that we directly contacted to ensure his/her knowledge of the NPO’s board and ability to diffuse the survey to “dispassionate” volunteers who were highly engaged in the organization’s functioning, tasks, and governance. We felt that these volunteers’ perceptions of NPOs’ board and organizational effectiveness, as well as their perceptions of OI and commitment, were likely to be reliable and representative of the actual and real situation of the organization. Our convenience and non-probability sampling method (Etikan et al. 2016) yielded 330 usable questionnaires, implying a response rate of 33%. After discarding the incomplete questionnaires, a valid sample of 300 questionnaires was utilized in this research. Although our sampling method has inherent limitations that may affect the representativeness of our findings—which are not generalizable for the whole Italian Third Sector but instead specifically referred to the Brotherhoods of Mercy—such a method allowed us to observe specific characteristics regarding this particular NPO, which, as stressed before, represents one the most ancient and important existing NPOs (Manetti et al. 2017; Zollo et al. 2017a, 2018a).

Next, we estimated the non-response bias by conducting wave analysis (Rogelberg and Stanton 2007) through a comparison of early and late respondents (see also Zollo et al. 2017a). Volunteers who responded during the first wave (March-May) were considered early respondents, while those who responded during the second wave (June-August) were included in the late respondent group. The results of T-tests across many variables—such as age, employment and other variables included in the study—showed no significant differences between the groups, thus assuring that non-response bias was not a concern (Armstrong and Overton 1977).

The sample characteristics related to volunteers’ gender, age, education level, and years of volunteering are shown in Table 1.

Table 1 Sample characteristic
Full size table

The respondents were male (56.3%), aged 18–24 (26.7%), with a high school degree (49.3%), and had volunteered for more than one year (88.7%).

3.2 Measures

All the items used in the present study are reported in Table 2.

Table 2 Means, standard deviations (SD), and reliabilities (Cronbach’s α) of the study variables
Full size table

The three sub-dimensions of governance effectiveness—namely strategy and policy making (three items), stewardship (three items), external relations and accountability (seven items)—have been measured using the scale originally developed by Cornforth (2001). Such a scale has been specifically built to study governance and effectiveness mechanisms in the realm of NPOs.

To assess OI, the widely-used five-item scale developed by Mael and Ashforth (1995) has been adopted. Actually, such a scale has been used by many non-profit scholars (see Sargeant and Shang 2012).

The three sub-dimensions of volunteers’ commitment to NPO—namely loyalty (six items), devotion (three items), and choosiness (five items)—have been measured with the scale developed by Hustinx (2005). This instrument has been purposely built to assess volunteers’ behavioural features in NPOs.

Finally, to assess organizational effectiveness we employed the three-item scale called organizational responsiveness developed by Herman and Renz (2004) and the five-item scale called social effectiveness developed by Brown (2005). Such instruments were used in the present study because of their specific nature related to assessing NPOs’ organizational effectiveness.

Respondents rated items on a five-point Likert scale from 1 = strongly disagree to 5 = strongly agree. All reverse items were consistently recoded.

4 Analysis and results

4.1 Means, standard deviations, and scale reliabilities

In Table 2 we present the scale reliabilities and descriptive statistics of the analyzed variables. The scales used in the research present good reliability, from 0.70 (Devotion) to 0.89 (External relations and accountability), thus showing Cronbach’s alpha ≥ 0.70 as required (Hair et al. 2006). Hence, all the variables were taken into consideration in the following analysis.

4.2 Correlations

The variables’ correlations are presented in Table 3.

Table 3 Correlations of the study variables
Full size table

All the variables present high and significant Pearson r values, except for choosiness, which is significantly correlated only with external relations and accountability.

4.3 Confirmatory factor analysis and model validity

A confirmatory factor analysis (CFA) was conducted using the maximum likelihood function of AMOS v. 22 (Arbuckle 2013; Zollo et al. 2017b) to assess the psychometric properties of the considered constructs—namely governance effectiveness, OI, commitment to the NPO, and organizational effectiveness. CFA showed that all factor loadings—the path coefficients between the indicators (manifest variables) and the latent variable—were significant (P < 0.01) (see Table 4). Next, we built a measurement model and examined the goodness-of-fit measures to verify the acceptable parsimony of the proposed four-factor model (Hair et al. 2006; Zollo et al. 2018b). Concerning absolute fit indexes, the Chi square statistics of the model are significant (χ2 = 1065.022, P < 0.01) with 565 degrees of freedom, and the relative Chi square suggests a good fit with a T-test of χ2/df = 1.884 (lower than 3 as required) (Bagozzi and Yi 1988; Bentler 1990). In addition, the goodness of fit index (GFI) (0.932) and the adjusted goodness of fit index (AGFI) (0.921) suggest a good fit, being higher than 0.9 as required (Bentler 1990). Finally, a root mean square error of approximation (RMSEA) of 0.052 suggests an acceptable model fit, less than 0.07 as required (Bentler 1990). Concerning the relative fit indexes, the most commonly used are the comparative fit index (CFI), the incremental fit index (IFI), the normed fit index (NFI), and the Tucker-Lewis index (TLI), also known as the non-normed fit index (Bagozzi and Yi 1988). Following Bentler (1990), the model indicates acceptable fit indexes (CFI = 0.984; IFI = 0.987; NFI = 0.986; TLI = 0.942), all above 0.90.

Table 4 Validity of the model
Full size table

The proposed four-factor model fits the data significantly better than a series of alternative models obtained by disaggregating the sub-dimensions of the constructs, such as: (1) a six-factor model where the three dimensions (strategy and policy making, stewardship, and external relations and accountability) of governance effectiveness were treated as unidimensional constructs (χ2/df = 3.845; GFI = 0.776; RMSEA = 0.178; CFI = 0.770); (2) an eight-factor model where the three dimensions (loyalty, devotion, and choosiness) were also treated as unidimensional constructs (χ2/df = 4.392; GFI = 0.680; RMSEA = 0.185; CFI = 0.695); (3) a nine-factor model where the two dimensions (organizational responsiveness and social effectiveness) of governance effectiveness were treated as unidimensional constructs (χ2/df = 4.995; GFI = 0.618; RMSEA = 0.198; CFI = 0.620). To further stress the theoretical validity of our proposed model, we compared it with competing models showing a different flow of theoretical hierarchy among latent variables. As expected, the model in which Organizational Effectiveness influences Governance Effectiveness via OI and Volunteers’ Commitment to the NPO performed significantly worse than our proposed model, showing unacceptable fitting indexes (χ2/df = 19.765; GFI = 0.577; RMSEA = 0.288; CFI = 0.537). Similarly, alternative models, obtained by changing the order of Volunteers’ Commitment to the NPO influences on OI, performed significantly worse both in the Governance Effectiveness-Organizational Effectiveness relationship (χ2/df = 9.368; GFI = 0.655; RMSEA = 0.190; CFI = 0.682) and the Organizational Effectiveness-Governance Effectiveness relationship (χ2/df = 11.297; GFI = 0.426; RMSEA = 0.252; CFI = 0.611). Hence, the distinctiveness and theoretical foundation of the hypothesized model were supported (Hair et al. 2006).

Table 4 shows the results of the composite reliability (CR), and convergent and discriminant validity of the model.

To assess the internal consistency of the indicators, we estimated the CR for each latent construct. CR values higher than 0.6 are required (Bagozzi and Yi 1988). As Table 4 shows, all the variables indicate acceptable CR values, ranging from 0.75 (devotion) to 0.93 (external relations and accountability). Convergent validity was assessed by the average of variance extracted (AVE), which estimates the indicators’ amount of variance accounting for the latent construct (Hair et al. 2006). AVE values higher than 0.5 indicate good convergent validity (Bagozzi and Yi 1988). All constructs show acceptable values of AVE higher than 0.5, ranging from 0.59 (devotion) to 0.78 (external relations and accountability). Moreover, to assess the model’s discriminant validity, the square root of the constructs’ AVEs must be greater than the correlations among the model constructs (Fornell and Larcker 1981). The square roots of all constructs’ AVEs in Table 4 are greater than the correlations reported in Table 3, thus indicating acceptable discriminant validity. The overall fit indexes of the model and the reliability and validity analysis suggest an acceptable model fit (see Zollo et al. 2017b, 2018b).

Finally, we assessed for the presence of common method bias (CMB) following the procedures suggested by Podsakoff et al. (2003). First, we pretested the scales to delete unclear items from the questionnaire. Second, we conducted a Harman’s one-factor test, which failed to identify a general factor that accounted for the majority of the total variance, although equal to 28.54%. Next, we used AMOS CFA to compare our proposed model with a model loading all items onto a common method factor: the “one-factor model” (Podsakoff et al. 2003; Zollo et al. 2017a). The comparison yielded a significant change in Chi square as required—the χ2 difference test with one degree of freedom was 16, much greater than 3.84 which is the critical value associated with P = 0.05. Our proposed model showed a superior fit to the data with respect to the one-factor model (Podsakoff et al. 2003). Hence, CMB is less likely to be a significant threat in the present study.

4.4 Hypotheses testing

The mediational hypotheses were tested following the procedure proposed by Hayes (2013) and using SPSS PROCESS macro (v. 2.16; see also Preacher and Hayes 2004, 2008). First, control variables were treated as “covariates” during the analysis to assess their influence on the hypothesized relationships (Hayes 2013). Interestingly, only three control variables had a significant impact: (a) respondents’ years of volunteering significantly impacted on the relationship between volunteers’ OI and commitment to NPO (+ 0.11; P < 0.05); (b) NPOs’ organizational size significantly impacted on the relationship between volunteers’ OI and organizational effectiveness (+ 0.15; P < 0.05); and (c) NPOs’ number of volunteers significantly impacted on the relationship between volunteers’ commitment to NPOs and organizational effectiveness (+ 0.22; P < 0.01).

Next, to conduct multiple mediation analysis (model 6 of PROCESS), we used the bootstrapping method (based on 5000 bootstrap samples) and computed 95% bias-corrected lower level confidence intervals (LLCIs) and upper level confidence intervals (ULCIs) around the estimates of indirect effects (see Zollo et al. 2017a). According to this method, governance effectiveness—the independent variable—should be significantly related to volunteers’ commitment (path a1) and OI (path a2), our two hypothesized mediation variables. Next, the hypothesized mediation variables—OI and volunteers’ commitment—should be significantly related to one another (path d). After controlling for the effect of independent variables, mediation variables should be significantly related to organizational effectiveness (paths b1 and b2, respectively), the dependent variable of our model. Mediation is indicated by the significance level of the indirect effect (patch c—path c′) from governance effectiveness on organizational effectiveness through OI and volunteers’ commitment, as indicated by the p-value or the LLCIs and ULCIs (Hayes 2013). In other words, governance effectiveness should have a different total (path c) rather than a direct effect (path c’) on organizational effectiveness, thus yielding an indirect effect that is different to zero (see also Preacher and Hayes 2004, 2008). Table 5 shows the results of the bootstrapping procedure.

Table 5 Results of multiple mediation analyses using a bootstrapping procedure
Full size table

Governance effectiveness was positively related to volunteers’ commitment (+ 0.35; P < 0.01), providing statistical support for H1a. Similarly, volunteers’ commitment was positively related to organizational effectiveness (+ 0.21; P < 0.01), thus supporting H1b. Governance effectiveness positively influenced OI (+ 0.66; P < 0.01), which in turn positively influenced organizational effectiveness (+ 0.15; P < 0.01). Hence, both H2a and H2b were supported by the analysis. The highest influence was between OI and volunteers’ commitment (+ 0.91; P < 0.01), thus providing statistical support for H3. Finally, concerning the relationship between governance effectiveness and organizational effectiveness, the total effect (+ 0.50; P < 0.01) significantly differed from the direct effect (+ 0.10; P < 0.01), resulting in a positive indirect effect (+ 0.40) with LLCI and ULCI (0.0995; 0.3345) that did not comprise zero as required (see Hayes 2013). Thus, both OI and volunteers’ commitment were partial mediators of the relationship between governance and organizational effectiveness, providing statistical support for H4. Figure 2 shows these multiple mediation results in our conceptual model.

Fig. 2
figure 2

Results of the bootstrapped mediation analysis. *p < 0.01. a, b In parentheses and latin we reported both the totala (0.50) and directb (0.10) effects which show the significance of the mediational hypotheses: the resulting indirect effect (total − direct = 0.40) is different than zero and thus confirms the partial mediational effects of both organizational identity and volunteers’ commitment to NPO.

Full size image

5 Discussion

The results of the empirical analysis show the role of governance effectiveness as a critical element in determining volunteers’ OI and commitment to NPOs. Specifically, the mediating effect of both OI and organizational commitment in the governance effectiveness–organizational effectiveness relationship has been supported. In this way, both theoretical and practical suggestions may be derived, from the perspective of better structuring NPOs’ governance for an effective attraction of and retention strategies for volunteers.

An important first contribution, for both literature and practice, is to point out that the perceived governance effectiveness significantly influences volunteers’ commitment. This result allows us to fill an existing literature gap and to deepen the study of a critical category of internal stakeholders—such as volunteers—within the governance mechanisms of NPOs (Jegers 2009; Cornforth 2012; Van Puyvelde et al. 2012). We provide in this way interesting insights into the governance literature on volunteers’ commitment. This relationship can be interpreted through the theoretical lens of participation and responsiveness (see Eckstein and Gurr 1975). According to this theory, the participation of stakeholders is influenced by the responsiveness of those who govern, which is understood as the ability to make decisions in line with the expectations and requirements of stakeholders. In addition to this, the empirical analysis confirms that a governance able to create shared value within the organization is an important tool to promote both commitment (O’Reilly and Chatman 1986) and engagement (Gruman and Saks 2011). Consistent with previous literature (Cornforth 2001), managers of NPOs should exploit the sub-dimensions of governance effectiveness—namely strategy and policy making, stewardship, external relations and accountability—to influence the participation of volunteers.

Among the governance functions that can influence the commitment of volunteers, the following aspects emerged: first, the NPO mission, goals and values definition. These results confirm the pertinent literature (Zollo et al. 2017a, 2018a), arguing that a clear definition of both moral and religious values characterizing the Brotherhood of Mercy’s mission results as fundamental for fostering, on the one hand, NPOs’ attracting and retaining strategies and, on the other hand, a transparent and interconnected embeddedness with the local community (see also Manetti et al. 2017). Second, the development of accountability practices allows the constant monitoring and meeting of stakeholders’ interests and expectations. In line with the pertinent literature (Manetti et al. 2015), the Brotherhood of Mercy’s ability to be accountable, not only represents an important instrument of financial control, but also enables the improvement of volunteers’ commitment.

These results may also prove beneficial to NPOs’ practitioners and policy makers. In fact, to strengthen the positive effect of governance effectiveness on volunteers’ commitment, managers should effectively focus on the considered governance mechanisms to enhance volunteers’ loyalty, devotion, and choosiness, thus deterring their turnover intention, which actually represents a key factor for the long-term survival of NPOs (see Zollo et al. 2017a). Given that our results mimic the empirical evidences of FPOs’ governance mechanisms (Gruman and Saks 2011; Lee and Miller 1999), we suggest NPOs’ management should provide a clear definition of organizational goals, the creation of a challenging work environment and a good work atmosphere, and manage the frustration of volunteers coping with difficulties, in order to increase their organizational commitment.

A further significant contribution refers to the crucial role of OI, which was shown to be a strong predictor of volunteers’ commitment and was highly influenced by governance effectiveness. As a result, NPO board members should act on OI to influence and encourage the participation of volunteers, for example by the internalization of the organizational values and goals (Gruman and Saks 2011; Ng and Wyrick 2011). Consistently with recent research (Manetti et al. 2017; Zollo et al. 2018a), volunteers’ identification with the moral values of the Brotherhood of Mercy is crucial for its long-term sustainability and ability to retain committed volunteers. In this light, it is important to point out that OI was positively and strongly correlated with devotion and loyalty—two of the sub-dimensions of volunteers’ commitment—thus offering significant practical suggestions to NPOs’ managers for reducing volunteers’ turnover. Instead, as far as concerns the sub-dimension of choosiness, our empirical analysis showed that it was not significantly correlated with any of the other constructs assessed in the present study. This seems to highlight the spontaneous nature of volunteering and the volunteers’ difficulty in accepting obligations and impositions (Hustinx 2005). Therefore, we suggest NPOs’ management should enhance volunteers’ availability in order to act on their consensus and identity. This allows volunteers to perceive the required tasks not as imposed duties but as activities related to the shared value and interest of the organization. Moreover, this result stresses how volunteers’ choosiness—here interpreted as their availability to engage in all NPOs’ activities, and accepting every kind of required task (Hustinx et al. 2010)—has to be enhanced by NPOs’ policy makers, who otherwise risk creating an excessively selective behaviour, which may be detrimental in terms of NPOs’ effectiveness. In line with recent empirical evidence (Zollo et al. 2017a), a possible solution for increasing volunteers’ willingness to engage in NPOs’ activities could be to implement a reciprocal and relational organizational behaviour. Actually, given the importance of volunteers’ OI and commitment, NPO governance should be able to encourage a “gratuitous reciprocal exchange of trust, commitment, and leniency between volunteers and management” (Zollo et al. 2017a, p. 524). This is also in line with recent research on relational signalling in both FPOs and NPOs (Lindenberg and Foss 2011), according to which, if managers address employees’ (volunteers’) motivations, and appropriately reward their efforts, this will be beneficial to the organizational identification, which will in turn positively impact on the perceived organizational effectiveness. Overall, we suggest NPOs’ managers should promote organizational effectiveness through the development of volunteers’ commitment, especially by fostering OI. Actually, if the governance of Brotherhood of Mercy is able to effectively foster volunteers’ signals of commitment—i.e., loyalty, devotion, and choosiness—this will incentivize and reward their volunteering activities (Manetti et al. 2015).

6 Limitations and future research

Our study suffers from limitations that offer interesting avenues for future research in this area. The first limitation concerns the sampling procedure because we used an Italian convenience sample to collect data; it could be interesting to empirically test our proposed conceptual model in different geographical contexts, such as the United States, where the volunteering phenomenon is increasingly expanding (Zollo et al. 2017a). Moreover, our framework could be replicated in Eastern collectivistic countries, which are sociologically different to Italy or the US, such as Japan or China. Next, we only took into consideration two of the many mediating variables explaining the relationship between governance and organizational effectiveness in the context of NPOs. Future studies could undertake a functional approach (Clary et al. 1992) by better investigating the motivations—both altruistic and egoistic—that lead volunteers to positively respond to NPO governance. Actually, knowing internal stakeholders’ main work motivation represents an important strategic lever for managers and governance in general (Lindenberg and Foss 2011). Finally, it would also be interesting to empirically test our conceptual framework in the for-profit sector. For example, governance scholars may better unpack the linkages between governance and organizational effectiveness by focusing on employees’ OI and organizational commitment, which represent important mechanisms to improve an organization’s performance.