Value co-creation as a mediator between strategic planning and social enterprise performance Open Access

There is a noted focus in empirical studies on social enterprise (SE) development. This becomes very pronounced, especially within developing countries (Davies and Doherty, 2018). Others emphasise the need to pay attention to the role that processes such as strategic planning (SP) may have in SE development. Despite all this, especially on the global front, SEs continue to receive focus within the practitioner and academic press (Block et al., 2021; Cheah et al., 2019). SEs are seen as important drivers for solving poverty and employment challenges while attempting to address socio-economic inequalities through creating sustainable social wealth (Maher and Hazenberg, 2021; Goyal et al., 2017). Co-creation plays an important role as an enabler and generator of economic value for the sustainability of SEs (Dwivedi and Weerawardena, 2018). In essence, co-creation allows organisations to create value through interactions. Hence, SEs are also referred to as hybrid organisations (De Silva et al., 2020; Peng et al., 2019; Wang et al., 2015).

SEs rely heavily on dynamic capabilities for growth (Ge et al., 2019). Capabilities are defined by Dosi and Teece (1998) as firm-specific abilities related to the organisation, management, coordination or the governance of activities within the organisation. Two types of capabilities exit. Firstly, resource capabilities, which relate to how human capital components such as knowledge and skills are exhibited (Leonard-Barton, 1992). Secondly, resource utilisation capabilities, aspects related to the role of managerial systems, organisational processes and organisational capital (Lyu et al., 2019; Glaveli and Geormas, 2018; Hamel and Prahalad, 1994). These two dynamic capabilities bring in the hybrid nature of a SE (Peng et al., 2019; Hoogendoorn, 2016).

The study borrows from the resource-based view (RBV) theory of the firm (Dosi and Teece, 1998; Doyle, 2019; Sinthupundaja et al., 2020) and the theory of value co-creation (VCC) (Nadeem et al., 2020; Zhang et al., 2019; Prahalad and Ramaswamy, 2000), to understand how a SE can use internal resources and dynamic capabilities to achieve sustainable competitive advantage (Barney, 1991; Nelson, 1991; Zhao et al., 2019). The emphasis is on how SP and VCC influence social enterprise performance (SEP) (Dwivedi and Weerawardena, 2018; Islam, 2020).

Some scholars contend that co-creation occurs where organisations and society work jointly to generate value (Galvagno and Dalli, 2014; Nadeem et al., 2020). Value can consequently be co-created in three different ways:

1. the involvement of actors, and their roles in different phases of co-creation;

2. the use of sets of activities to foster transformative power; and

3. new knowledge, new relationships and new solutions, as well as increased efficiency and social acceptability of the transition processes as the outcomes of co-creation (Sillak et al., 2021).

Successful co-creation can contribute towards the creation of high performing SEs (Murthy, 2021). This study sought to address the growing need for an inquiry into the SE literature stream by looking at the role that VCC plays in SE performance (Eikebrokk et al., 2018; Torres and Augusto, 2020).

The study aims to contribute in three ways. Firstly, the study narrows focus on the role of SP and VCC as important factors influencing SE performance. Secondly, there is no consensus on how the performance of SEs should be studied (Mamabolo and Myres, 2020) and so by exploring the relationship between SP and enterprise performance, while using VCC as the mediating factor, this paper hopes to address this gap.

Thirdly, there are calls for more research, especially within developing countries paying specific attention to SE research (Degbey et al., 2021; Guidice et al., 2019). Again, many of the studies around SEs are descriptive in the form of a case study, while limited empirical studies exist (Kajiita and Murote kang’ethe, 2021, 2020). This paper provides a platform for future research geared towards understanding aspects related to SEs in developing economies. Finally, this study tested a model that narrows focus into VCC as a mediating effect between SP and SEP. This model becomes a first to be tested specifically within the context of a developing nation.

This section discusses the context in which this paper is based, which is South Africa. Firstly, the work done by predecessors in understanding SEs through different context is acknowledged (Zhang et al., 2021; Kajiita and Murote kang’ethe, 2021). There are many social challenges facing South Africa. These include inequality, poverty and a high unemployment rate (Manyaka-Boshielo, 2017). These challenges call for the need of research on SEs (Littlewood and Holt, 2018), especially in the South African context, where there is limited empirical research (Urban and Gaffurini, 2017). Furthermore, the existence of challenges related to management and planning issues as affecting SEs’ operation and impact in South Africa has been acknowledged (Ngatse-Ipangui and Dassah, 2019). Amid all this, calls for research continue to argue for the prioritisation of SE research (Mamabolo and Myres, 2020).

To address some of the aforesaid gaps in the literature, this paper advances a conceptual model of VCC as a mediating effect between strategic leadership and SEP (Figure 1). This conceptual model allowed the researchers to empirically examine the contribution of SP and VCC to the SEP in a South African context.

The authors of this paper acknowledge previous work done on investigating the significance of SP on SEP (Donkor et al., 2018; Esfahani et al., 2018). The performance of SEs continues to rely heavily on internal resources and dynamic capabilities (Majama and Magang, 2017; Ge et al., 2019; Teece, 2018). Some studies have argued that through SP, SEs can leverage their performance in fulfilling their social utility goal of improving the economic and social capacities of their communities (Islam, 2020; Sabella and Eid, 2016; Porter and Kramer, 2011). Despite their contribution to a country’s economy (Mamabolo and Myres, 2020), lack of SP for some SEs affects their performance negatively (Majama and Magang, 2017). Sabella and Eid (2016) argued that effective SP, especially in a constrained environment, leads to SEs’ sustainability. In turn, this provides SE with opportunities to scale up while solving societal problems (Tykkyläinen and Ritala, 2021). Nonetheless, there seems to be limited empirical attention directed at understanding how SEs realise value through SP (Munoz and Kimmitt, 2019), or how SP and VCC influence SEP (Dwivedi and Weerawardena, 2018; Islam, 2020). This leads to the importance of investigating outcomes specific to SEP (Davies and Doherty, 2018). The relevance of activities such as SP (Tykkyläinen and Ritala, 2021) in dealing with changes both within an SE and outside the enterprise needs to be further researched. Based on the presented literature, it can be expected that:

SEs have emerged as a sustainable and innovative means to solve societal problems (Akter et al., 2020). Over the past few decades, management research has focused on three streams of SEs:

1. tangible outcomes of social entrepreneurship (Dees, 1998);

2. entrepreneurial behaviour and goals of social ventures (Dees and Economy, 2001); and

3. personality and distinctive characteristics, traits and competencies of owners or managers of SE (Sullivan Mort et al., 2003).

Little empirical attention is given to SE drivers. This is especially the case with VCC as a driver of SE (Akter et al., 2020). Co-creation is referred to as a platform for value generation through a joint, collaborative, concurrent and peer-like process (Hamidi et al., 2020; Kullak, 2021). VCC is argued to be grounded in understanding social forces, social structures and unbalanced involvement of SE actors within a given context (Sabella and Eid, 2016). There is empirical evidence that VCC has a positive impact on SE growth (Ge et al., 2019). However, understanding whether VCC has an influence on SEP still remains unclear (Elia et al., 2020; Zhang et al., 2020). Again, the effect of VCC on SEP is still under researched (Davies and Doherty, 2018; Tsai et al., 2021). Based on the presented literature, it can be expected that:

The study proposes that VCC can act as a mediator between SP and SEP. Though previous research found a positive impact of VCC on the SE growth (Qalati et al., 2020), some scholars have found that in developing countries, it is unlikely for VCC to influence SEP (Cheah et al., 2019). It is likely due to the hybrid nature, that SEs appear to have complex and conflicting demands (Liu, 2020; Nadeem, 2020).

Through VCC, SEs can leverage their performance in fulfilling both their social and financial goals (Frempong et al., 2020). It has been proven that a lack of research exists around the topic of VCC and its link to SP, more so from a South African perspective (Dwivedi and Weerawardena, 2018; Islam, 2020; Ngatse-Ipangui and Dassah, 2019). However, there is still limited research on how SP and VCC influence SEP (Dwivedi and Weerawardena, 2018; Islam, 2020), especially in a South African context (Littlewood and Holt, 2018), where SEs are faced with management and planning challenges (Ngatse-Ipangui and Dassah, 2019). Based on the presented literature, it can be expected that:

A quantitative methodological approach was adopted in this empirical study. This approach seeks to establish the possibility of a causal relationship amongst the variables in question. In calculating the proposed model, a multiple regression with partial least squares (PLS) was used. This was conducted using a structural equation model (SEM) approach. Such an approach, including the adopted techniques, allows for the building of the research models by establishing latent variables. According to Haenlein and Kaplan (2004), latent variables are those that are not observed directly but inferred from other observed variables. There is general acceptance in the literature that such an approach and technique is useful when working with small samples, estimating complex models and making predictions and explanations (Hair et al., 2019). There is an observation of the suitability of such approaches and techniques when seeking to ascertain success factor research or when exploring the source of a competitive advantage (Hair et al., 2012).

Following an ethical clearance approval process, data was collected through liaison with a local government body that hosted a provincial training workshop with SEs in the Eastern Cape Province of South Africa. The municipalities that took part in the study were those in five districts in the Eastern Cape Province, notably: Buffalo City Metropolitan; Nelson Mandela Bay Metropolitan; Amathole District; Chris Hani; and Oliver Reginald Tambo District.

The aim of the workshop through which the data was collected was to bring together owners/managers of SEs in the province with the objective of capacity building in view of the socio-economic challenges faced by the province. The researchers negotiated with hosting a local government body to also distribute questionnaires related to the aim of the workshop. A total of 250 SE owners/managers were invited by the local government authority and a total of 160 attendees were present at a three-day online training session split into a minimum of 60 attendees per day. From the 160 participants, a total of 147 questionnaires were filled out, successfully yielding a response rate of 92%. The remaining 8% were rejected for analysis as these had missing data and deemed not usable.

An online Web survey was administered with permission from the local government authority. The survey was divided into four sections. The first section was the biographical section which investigated individual and organisational-specific characteristics related to the SE. The subsequent sections, all measured on a five-point Likert scale, measured the independent variable (SP), the mediator (VCC) and the dependent variable (SEP). Table 1 presents example items from these scales, their sources and the Cronbach’s alpha coefficients.

Table 2 shows that a large portion of the sample (62%) consisted of female-owned or managed SEs. Most of the owners/managers (52%) were between the ages of 41 and 50, of which 48% had a diploma/degree as their highest educational qualification. About 87% of the participants had been in operation for more than 12 years. Finally, 14% of the participants were owners, while 86% were managers of the SEs.

The measurement and structural models were tested by performing PLS and bootstrapping in Smart PLS version 2. The step entailed measuring reliability using indicator reliability, the Cronbach’s alpha coefficient and composite reliability. Table 3 indicates the factor loadings based on the scales used in the study.

An outer loading of 0.633 is acceptable while a loading of 0.700 and above is preferred (Hair et al., 2019). VCC scale is as follows: one indicator (VCC 4 with a loading of 0.295); seven SE indicators (SEP3 = 0.363, SEP5 = 0.305, SEP1 = 0.460, SEP6 = 0.496, SEP7 = 0.344, SEP8 = 0.422, and SEP9 = 0.418); and nine SP indicators (SPEI1 = 0.346, SPEI2 = 0.497, SPEI3 = 0.290, SPES1 = 0.431, SPES2 = 0.249, SPES3 = 0.450; SPVM1 = 0.434, SPVM2 = 0.494 and SPVM3 = 0.499) were deleted from the measurement model as they failed to load and explain at least 50% of the latent variables.

Furthermore, three more SEP indicators (SEP1 = 0.309, SEP2 = 0.283 and SEP4 = 0.269) were deleted from the model because their squared outer loadings were below the minimum threshold of 0.40. Once more, the enterprise performance indicator (SP5 with a variance inflation factor [VIF] value = 5.489) was deleted as this value was above the recommended threshold value of 5.000 and indicated the presence of collinearity problems. As a rule of thumb, a VIF of 5 or lower (i.e. tolerance level of 0.2 or higher) is needed to avoid the collinearity problem (Ringle et al., 2015).

Table 3 shows the results of the remaining indicators in the measurement model. Also as shown in Table 3, SP has outer loadings ranging between 0.649 and 0.807. VCC has outer loadings ranging from 0.820 and 0.885, while SEP has loadings ranging from 0.733 and 0.837. The outer loadings for all three latent variables are above the minimum threshold value of 0.633. More so, all the indicators presented in Table 3 have squared outer loadings ranging from 0.421 and 0.783, which falls within the acceptable and preferred (0.4 and above) threshold. The results in Table 3 confirm the reliability of the remaining indicators for all three variables.

Table 4 below presents example items from the three sections of the questionnaire used for this research study and the reliability scores through the Cronbach’s alpha coefficient tests. The deletion of indicators that were unreliable was made for as long as it improved the AVE, composite reliability and Cronbach’s alpha coefficients. Deletions and further computations were made four times before achieving the results. As observed from Table 4, all the scales used in the study reported Cronbach’s alpha coefficient scores of 0.7 above the recommended.

It is important to note deletion of indicators that were unreliable were made for as long as it improved the AVE, composite reliability and Cronbach’s alpha coefficients. Deletions and further computations were made four times before achieving the final results. After establishing indicator reliability, this study used Cronbach’s alpha and composite reliability to assess internal consistency (construct reliability). Table 4 reports Cronbach’s alpha coefficients of 0.863, 0.816 and 0.816 for all the three latent variables (SP, VCC and enterprise performance), respectively. All the values are above the minimum threshold value of 0.5, thus confirming that the three variables are highly reliable.

The composite reliability values were 0.894, 0.891 and 0.879, and were all above the preferred value of 0.7, thus confirming that all the three latent variables in this study are highly reliable. Given that all the composite reliability values were high, the reliability of the three variables was further confirmed through the rho_A coefficients. The rho_A coefficients for the three variables were 0.872, 0.818 and 0.815, and were all above the recommended threshold value of 0.7. This further confirmed the high reliability of the three research variables in this study. Convergence validity was measured using average variance extracted (AVE) values. As shown in Table 4, all the latent variables had AVE values (SP = 0.548, VCC = 0.731 and enterprise performance = 0.646) above the minimum acceptable value of 0.5. This confirms the existence of convergence validity of the three latent variables.

Concerning discriminant validity, square roots of AVE values and latent variable correlations were used. The square root of AVE values of each latent variable should be greater than the correlations among the latent variables. This study followed the Fornell and Larcker (1981) method of assessing discriminant validity which compares the correlation coefficients of the latent variables against the square root of the computed AVE values.

Tables 4 and 5 present the results. As shown in Tables 4 and 5, the square root AVE values of 0.803, 0.855 and 0.740 for all the three latent variables (enterprise performance, VCC and SP, respectively) are greater than the correlation coefficients (0.708, 0.683 and 0.749, respectively) of the latent variables in this study. This proves that the questionnaire indicators of all three latent variables are weakly correlated, different and therefore confirm the presence of discriminant validity among enterprise performance, VCC and SP.

To complement the Fornell and Larcker (1981) method, the heterotrait–monotrait ratio (HTMT) was also used to assess discriminant validity in this study. Table 6 presents the results and it is clear that the HTMT values of 0.760, 0.819 and 0.826 for all the three latent variables (enterprise performance, VCC and SP, respectively) are below the stricter threshold value of 0.85. This proves that the questionnaire indicators of all three latent variables are sufficiently different and therefore confirm the presence of discriminant validity among enterprise performance, VCC and SP.

After the collinearity tests (see Tables 5 and 6), the effect size of the relationships between the three latent variables was checked through the f² values. Table 7 presents the effect size (f²) of the proposed relationships between the three variables. According to Chin et al. (1996), the model’s f² effect size shows how much an exogenous latent variable contributes to an endogenous latent variable’s R² value. Cohen (1988, pp. 410–414) provided the following guidelines of interpreting the f² values, where values of 0.02 or less are classified as small; 0.15 or less is regarded as medium; and 0.35 or less is considered to be large. As shown in Table 7, there is a medium effect size (f² = 0.197; and f² = 0.119 > 0.15) in the relationship between VCC and enterprise performance; as well between SP and enterprise performance, respectively. The relationship between SP and VCC have a larger effect size (f² = 1.277). After confirming the effect size, this study tested the significance and the nature of relationships between the variables.

Tables 8 and 9 as well as Figure 2 present the path coefficients and T-statistics values to test the nature (direction) and significance of relationships between the three variables. To test the conceptual model proposed in Figure 1, PLS-SEM testing was conducted, and the results are shown in Tables 8 and 9. The findings in Table 7 are also supported by the model shown in Figure 2.

From the results (see Table 4 and Figure 2) it is clear that SP explains 56.1% of the variance in VCC. SP and VCC explain 55.4% of the variance in enterprise performance.

The PLS-SEM results in Tables 8 and 9 as well as in Figure 2 show that all the three hypotheses in this study are positive and highly significant. All the T-statistic values were above the minimum threshold value of 2.000, and the p-values are less than 0.05, which confirms high significance. The results in Table 8 and Figure 2 report a weak positive and a statistically significant (path coefficient = 0.348, T-statistic value = 4.708 and P-value = 0.000) effect of SP on enterprise performance. SP also has a strong positive and a highly significant effect on VCC (path coefficient = 0.749, T-statistic value = 18.016 and P-value = 0.000). VCC has a weak positive and a significant effect on enterprise performance (path coefficient = 0.447, T-statistic value = 6.481 and P-value = 0.000).

VCC has a fairly weak positive and a significant indirect effect on the proposed direct relationship between SP and enterprise performance (path coefficient = 0.335, T-statistic value = 5.731 and P-value = 0.000). Because the indirect effect (i.e. path coefficient for H2 [P2] multiplied by the path coefficient for H3 [P3]) and the direct effect as shown by the path coefficient of H1 (P1) are both significant, a mediating effect was calculated as the product of the three path coefficients (P2*P3*P1) which was equal to 0.117. The mediating effect was obtained from multiplying 0.749 by 0.447 by 0.348 from the three path coefficients, respectively. Given the positive effect, this study concludes that VCC plays a significant partial complementary mediating role in the direct relationship between SP and enterprise performance.

Table 10 provides a summary of the hypotheses testing done.

The study ascertains the mediation of VCC to the relationship between SP and SEP. The study drew on the theoretical perspectives of RBV to explore the relationship between SP and SEP. Specifically, this study developed a model that illustrates the role that VCC and SP play in influencing SEP (Islam, 2020). The findings prioritise the necessity of improving how SEs approach their planning as this has implications for SEP, especially in a South African context (Mamabolo and Myres, 2020). Given the challenges concerning management issues within SEs (Ngatse-Ipangui and Dassah, 2019), the ensuing model to this study offers useful insight into how SEs can be better run.

VCC is brought about by understanding social forces, social structures and unbalanced involvement of SE actors, which leads to high performing SEs (Murthy, 2021). In this study, the researchers found that SP has a strong positive and a highly significant effect on VCC. At the same time, VCC has a weak positive and a significant effect on enterprise performance (see Tables 8 and 9 and Figure 2). To this end, the results complement the recent work that demonstrates how VCC positively impacts SE growth (Ge et al., 2019). However, whereas Elia et al. (2020) and Zhang et al. (2020) focus on understanding the influence of VCC on SE performance, this study examined the role that VCC plays as a mediator between SP and SE performance. This positions the importance of not only SP but also value creation as dynamic capabilities.

Evidence shows that SE performance relies heavily on internal resources and dynamic capabilities (Frempong et al., 2020). Table 7 shows a medium effect size in the relationship between VCC and enterprise performance, as well as between SP and enterprise performance, respectively. Based on the results, the relationship between SP and VCC was found to have a larger effect size. Because the indirect effect shown by both H2 and H3, and the direct effect shown by H1 are both significant, the results of this study show a fairly weak positive and a significant indirect effect of VCC on the proposed direct relationship between SP and enterprise performance. Given the positive effect, this study concludes that VCC plays a significant partial complementary mediating role in the direct relationship between SP and enterprise performance. This study adds to the body of knowledge (Tsai et al., 2021) by addressing the effect of VCC on SEP. Based on the presented literature, it can be expected that this study has the following three contributions to the body of knowledge.

The first contribution is to the role of SP as an important organisational activity, as well as a dynamic capability in the performance of SEs (Kedmenec and Strašek, 2017; Sinthupundaja et al., 2020). This finding confirms previous findings on the role of capabilities within organisations (Elia et al., 2020; Tucker and Croom, 2021). The study inimitably finds that enacting a SP regime and improving VCC can potentially generate opportunity for social entrepreneurs. This appears to be a contribution with consequences and potential implications not only theoretically in understanding SEs better but also in informing practitioner efforts.

The study becomes one of the first within the African context to pay attention to the role that SP and VCC have in SEP. Such a focus has implications for the work done by managers and practitioners within the SEs context, that is, to take into account the role that strategy and VCC play in influencing better SEP (Eikebrokk et al., 2018; Torres and Augusto, 2020). Some implications can be drawn for research and policy. SP should be considered important to encourage SEP. This study highlights and prioritises network capabilities within SE as critical in informing SEP.

The study had two limitations. Data collection was the first limitation. This is because the data collection process happened at the height of the COVID-19 pandemic, and this affected the number of SEs that could participate in the study. This in turn affected the sample size. The second and last limitation pertaining to this study is that the study was based only on SEs within the Eastern Cape Province of South Africa. There may be a need to be expansive not only in terms of SE activity but also location.

Future studies can prioritise more research into SEP against other factors. The focus in this study was on management practices and dynamic capabilities but the breadth of extension can be widened. This may include exploring the role of individual characteristics exhibited by the social entrepreneur into how this influences SEP. There may also be an opportunity to explore the interaction that potentially exists between these individual characteristics as well as dynamic capabilities in informing SEP. This interaction appears important considering the role of the environment in informing SEP (Tucker and Croom, 2021). Future research may also look at how ownership influences the role of SP and SEP.

Future research can also consider incorporating a qualitative approach to understanding how management practices and dynamic capabilities are important within SEs. Furthermore, future research can test management practices and dynamic capabilities against other indicators of performance. This study acknowledges the efforts by Urban and Gaffurini (2017) in advancing this cause and the study encourages other researchers in the African context to follow suit. This research narrowed focus on SEP; other indicators that can be used can include sustainability outcomes. These outcomes have been prioritised as important for the future of SEs (Mendez-Picazo et al., 2020).