Abstract
Tourism and hospitality firms generally recover from crises but the resources and capabilities they leverage to maintain social performance remain elusive. Embedded in the resource-based view of the firm, we argue that psychological resources, such as employee resilience, and knowledge embedded in teams, through a transactive memory system, as well as dynamic capabilities can sustain social performance during crises. We test this proposition on a sample of 350 UK tourism and hospitality firms that survived the COVID-19 pandemic. Results from partial-least square structural equation modeling and fuzzy set qualitative comparative analysis identify different combinations of employee resilience, transactive memory systems, and dynamic capabilities that can boost social performance. Theoretically, micro and meso-level organizational resources and capabilities are intertwined in shaping social performance during crises. Managerially, recommendations on how to improve employee resilience in a way that boosts team knowledge, dynamic capabilities and ultimately social performance are offered.
Keywords
Introduction
Against the backdrop of the COVID-19 crisis, the United Nations World Tourism Organization (UNWTO, 2022) highlights that long-term goals of the tourism industry should not be traded against short-term recovery (UNWTO, 2022). This implies that sustainability must remain a core focus throughout COVID-19 recovery. A critical issue for social sustainability is workers’ rights and social protection in the tourism industry (UNWTO, 2022). While the International Labor Organization (ILO) highlights that high-income countries have in large part recovered from the COVID-19 crisis but have also shown resilience to new challenges (ILO, 2023), workers in the tourism and hospitality industries still face precarious employment (ILO, 2022). The necessity for facilitating a rapid, safe, and inclusive recovery that builds toward a more sustainable tourism sector has been highlighted (ILO, 2022). Alongside, prioritizing safety and health at work is indispensable for sustainable recovery (ILO, 2022). Increasing concern about social and environmental issues has raised the importance of the sustainability performance of firms (Silva et al., 2021). In this study, we focus on social performance because of COVID-19 impacts on the working conditions of tourism and hospitality employees. While extant studies allude to the ability of tourism and hospitality firms to recover from crises and disasters (El-Said et al., 2023; Mair et al., 2016; Ritchie & Jiang, 2019), their social performance in these challenging times has not been examined. Social performance can be defined as the improvement of human values and concerns, working conditions, diversity, and corporate social responsibility (CSR) (Sudusinghe & Seuring, 2020), which should be a strategic focus of tourism and hospitality firms during recovery (ILO, 2022). To date, tourism and hospitality scholarship has largely focused on firm strategy (Uyar et al., 2023) and green management practices (Elshaer et al., 2023) as drivers of organizational social performance, omitting to consider that employees (Prayag & Dassanayake, 2023) and dynamic capabilities (Y. Jiang et al., 2019) can play a critical role in facilitating recovery.
While dynamic capabilities have received some attention in tourism and hospitality scholarship (Y. Jiang et al., 2019, 2023; Q. Jiang & McCabe, 2021; Prayag, Jiang et al., 2023), the role of teams as an enabler of crisis recovery has been neglected. Both teams (Lewis, 2003) and capabilities (Helfat & Peteraf, 2015) require knowledge as a resource to facilitate response and recovery. Improving tourism firm’s sustainability performance requires changes in strategy and structure, which is enabled by dynamic capabilities in dynamic environments (Beske, 2012). Organizations also increasingly rely on knowledge assets to provide superior performance, which requires knowledge embedded in teams, and members using their experience and expertise to solve problems and offer creative solutions (Lewis, 2003). Yet, how knowledge embedded in teams is leveraged to sustain business performance during crises has alluded tourism and hospitality scholars. In other fields, a transactive memory system (TMS) represents the repertoire of knowledge extracted from teams and embedded in organizational systems (Lewis, 2003). In a TMS, three activities are critical: knowledge specialization, credibility, and coordination (Bachrach & Mullins, 2019; Lewis, 2003). A TMS is important in uncertain and dynamic environments for two reasons: First, survival is often dependent on how organizations foster learning and knowledge creation processes. These environments also require new knowledge generated rapidly to face unpredictable events (Gligor et al., 2015). Second, the coordination of this learning and knowledge is key to managing uncertainty. Highly uncertain environments require non-programed or relational coordination (Argote & Guo, 2016), which can be facilitated by a TMS (Fernandez-Giordano et al., 2022).
As Cotta and Salvador (2020) suggest, managing disruptions is a knowledge-intensive task that requires the contribution of different domain experts within the organization. These authors suggest that resilient practices can enhance a TMS, but their focus linger on practices that enhance organizational rather than employee resilience. Individuals may act as catalysts in the development and functioning of a TMS (Mell et al., 2014) and it is, therefore, plausible that resilient employees can have knowledge that can be codified, coordinated, and retrieved from a TMS to respond and recover from crises. However, this proposition remains untested in existing management, organizational behavior, and tourism and hospitality literatures. So far, the concept of a TMS, its antecedents and outcomes have mainly eluded the tourism literature. Table 1 provides definitions of all key terms in this study. Grounded in TMS theory (Hollingshead, 1998; Lewis, 2003) and dynamic capabilities (Augier & Teece, 2009; Teece, 2018), we ask two research questions:
Operationalization of TMS and Dynamic capabilities.
By asking these two questions, this study provides a unique perspective on the shared cognitive mechanism linking employee resilience to a TMS and dynamic capabilities, in turn affecting the social performance of tourism and hospitality firms as shown in Figure 1.
Conceptual framework.
Understanding these relationships can highlight trade-offs that are possible during crises to sustain or improve social performance. The contributions of this study are four fold. First, we extend the nomological network of social performance by demonstrating that resilience, team knowledge, and dynamic capabilities are antecedent factors. So far, there are no studies linking these concepts in a singular framework within tourism and hospitality or management literatures (see Table 2). Second, existing tourism and hospitality studies (Prayag et al., 2020; Saad & Elshaer, 2020; Senbeto & Hon, 2021) on employee resilience have omitted a TMS as an important outcome, and we establish a positive link between the two. Third, while tourism and hospitality studies (Y. Jiang et al., 2023; Prayag, Jiang et al., 2023; Shrestha & L’Espoir Decosta, 2023) have exalted the virtues of dynamic capabilities in facilitating crisis recovery of tourism firms, the micro-dynamics that link a TMS to dynamic capabilities (Argote & Ren, 2012; Martin & Bachrach, 2018) have so far been ignored. Thus, we confirm the integration of team knowledge with existing knowledge and capabilities confers positive organizational benefits. Fourth, the link between a TMS and firm performance is contested, with Batra et al. (2023) demonstrating no direct link. Likewise, the focus has been on team performance (see Table 2) in existing studies (Bachrach et al., 2017; Guchait et al., 2014) rather than sustainability performance. Thus, we demonstrate that having a TMS can also influence social performance positively.
Selected Tourism and Management Studies on Employee Resilience, TMS, Dynamic Capabilities, and Social Performance.
Conceptual Framework
System-Thinking, Transactive Memory System, Resource Based View (RBV), and Dynamic Capabilities
Recognizing that the type and nature of crises determine tourism firms’ response and recovery trajectories (El-Said et al., 2023; Ritchie & Jiang, 2019), our proposed model (Figure 1) is grounded in tourism system thinking and two main theories, TMS and the resource-based view (RBV) of the firm. From a system thinking perspective, tourism as a system comprises of several sub-systems that interact and determine resilience at micro, meso, and macro levels (Hall et al., 2023; Prayag, 2023). Within organizations, it remains largely unclear how resilience works and interacts at different levels (Raetze et al., 2021). Interactions at different levels (e.g., between micro-individual and meso-team and organizational) imply that different organizational sub-systems can influence outcomes such as organizational resilience and sustainability. In the proposed model (Figure 1), we argue that micro-level processes that include employee resilience can potentially contribute to meso-level systems (e.g., TMS, dynamic capabilities) through knowledge transfer via a TMS to strengthen capabilities and organizational outcomes. Given that the resilience of a system is intertwined with its sustainability (Fiksel, 2006) and in tourism systems, including organizations, sustainability is a desirable outcome of resilience (Espiner et al., 2017), it is plausible that employee resilience can affect sustainable outcomes. From this perspective, resilient employees through their behaviors and knowledge deployment during adversity can be harnessed and when integrated in a TMS can be utilized alongside other capabilities such as dynamic capabilities to respond to adversity. Thus, strengthening micro processes related to resilience building of employees can strengthen sustainability via meso-level systems. However, given that a TMS and dynamic capabilities are organizational level systems and capabilities, the meso-level impacting the micro-level processes is also plausible but does not detract from sustainability being the overall organizational system goal. Thus, our study proposes micro-level processes (employee resilience) as antecedents of meso-level processes (TMS and dynamic capabilities) in determining one facet of sustainability- social performance. The focus on social performance is important given that employee welfare can be impacted during adversity, and if employees are resilient, this should reflect positively in meso-level metrics for sustainability, including those that evaluate social performance.
From a TMS theory perspective, members of collectives can function as external memory aids to one another (Wegner, 1987), suggesting that the knowledge and behaviors of individuals and teams may be useful in strengthening organizational capabilities. In the organizational context, a TMS provides members with insights into who knows what and who is the best at what within a team (Bachrach & Mullins, 2019). A central tenet of this theory is that organizational systems can be described as transactive because they depend on inter-member interaction and communication, aligning with the system thinking perspective of how sub-systems interact. This fosters deeper, more functional, and specialized collective knowledge that allows teams to get more efficient access to critical information (Lewis, 2003), that may be useful for response, mitigation, and recovery following adversity. The theory has been operationalized through a multidimensional concept, referred to as a TMS (see Table 1), which manifests through specialized knowledge, credibility perceptions, and coordination processes of the knowledge generated from teams (Lewis, 2003). In a TMS, the differentiated structure of members’ knowledge (specialization), members’ beliefs about the reliability of other members’ knowledge (credibility), and effective, orchestrated knowledge processing (coordination) interact to facilitate positive organizational outcomes (Lewis, 2003; Moreland & Myaskovsky, 2000). Thus, in group and organizational settings (Liang et al., 1995), a TMS is used to store, retrieve, and share knowledge. Each element plays a unique role in knowledge collaboration across multiple parties within and outside of the organization (Scheibe et al., 2022).
While TMS theory focuses on team knowledge and how this knowledge is embedded in organizational structures and processes (Liao et al., 2012), the RBV suggests that firm competitive advantage and performance are inextricably linked to unique bundles of resources and capabilities (Eisenhardt & Martin, 2000). The RBV argues that resources, both tangible (e.g., physical assets) and intangible (e.g., knowledge), must be rare, inimitable, and non-substitutable to provide sustained competitive advantage. However, RBV provides a static snapshot of a firm’s resources and capabilities, requiring dynamic capabilities to understand how they evolve and adapt over time (Eisenhardt & Martin, 2000). Thus, dynamic capabilities focus on how knowledge management is transformed into organizational capabilities (Eisenhardt & Martin, 2000) and is complementary to RBV. While RBV focuses on the identification and leveraging of valuable resources and capabilities, dynamic capabilities focus on the adaptation and evolution of sensing, seizing, and reconfiguration capabilities (Lin & Wu, 2014). Table 1 defines these capabilities. Given that both transactive memory systems and dynamic capabilities, at their core, delve into knowledge management structures and the benefits they confer to organizations, TMS theory has been described as a subset of dynamic capabilities (Argote & Ren, 2012). TMS theory explains the micro-dynamics of dynamic capabilities through a focus on team knowledge, which can then be embedded into practices that build capabilities (Argote & Ren, 2012). For this to happen, the knowledge and expertise embedded in teams need to be shared, capitalized, and brought to use by its members to make informed and collaborative decisions (Liao et al., 2012; Sonnentag & Volmer, 2009). To this end, individual and team capacities are critical when responding to adversity (Lengnick-Hall et al., 2011), which implies that resilience matters given that it provides the developable capacity for individuals to deal with workplace adversity (Kuntz et al., 2016). Thus, we can argue that employee resilience is an antecedent factor that strengthens both a TMS and dynamic capabilities. Yet, existing empirical evidence from the TMS literature largely supports team rather than individual adaptive capacities in enhancing performance (Marques-Quinteiro et al., 2013).
Given that knowledge-based capabilities such as dynamic capabilities (Batra et al., 2023) and a TMS (Bachrach et al., 2022) have been linked to different facets of firm performance, we argue that social performance can be one of these performance-related outcomes. Social performance incorporates organizational behavior toward its workforce and employment practices such as health and safety (Winter & Lasch, 2016). Both dynamic capabilities and a TMS are knowledge repositories that can be leveraged by tourism and hospitality firms to sustain or improve social performance during crises. Social performance highlights existing internal practices that support employees and external networks that support the organization (A. Kumar et al., 2023; Shang et al., 2020). Implicitly, the availability of knowledge-based resources such as those embedded in a TMS and dynamic capabilities should strengthen internal and external bonds, driving a firm’s social performance. The strategic management literature offers a compelling argument that a comprehensive understanding of firm-level behavior or practice requires consideration of individual level behavior and action (Abell et al., 2008; Bouguerra et al., 2024; Coleman, 1990). Aligning with this view, Prayag and his colleagues offered empirical evidence on the role of resilient employees in building resilient organizations (Prayag et al., 2020, 2024b; Prayag & Dassanayake, 2023) and other studies demonstrate the subsequent impacts on organizational performance enhancement (Bouguerra et al., 2024). Thus, empirical evidence justifies examining the role of micro-level factors such as employee resilience in influencing meso-level practices such as dynamic capabilities, transactive memory systems, and social performance.
Hypothesis Development
Next, we present the six hypotheses embedded in our theoretical model (Figure 1).
Employee Resilience and Transactive Memory System
While early conceptualizations of employee resilience viewed it as a dispositional trait, researchers now consider the concept as a state-like ability, capacity, or capability, which allows an employee to maintain normal functioning under challenging conditions and recover quickly from setbacks (Hartmann et al., 2020). This view implies that resilience is a stable state, developable in the workplace (Kuntz et al., 2017). Organizational factors such as job design, positive and supportive work climate, and employee-oriented human resource management (HRM) practices can support employee resilience (Raetze et al., 2021). While employee resilience studies are burgeoning in the tourism and hospitality literatures (Prayag et al., 2020; Saad & Elshaer, 2020; Senbeto & Hon, 2021), there is no study linking the concept to a TMS (see Table 2). This is despite several studies arguing that resilience in the workplace is multi-level and interactions exist between employee and team resilience (Hartmann et al., 2020; Raetze et al., 2021). Yet, existing studies fail to consider that employee resilience may transform into team resilience through a TMS. This is because team member interaction over time and collective teamwork patterns create knowledge that may be useful during adversity (Hartmann et al., 2020). As such adaptive, learning, and network-leveraging behaviors that signal employee resilience but also resource availability and capacity to utilize resources (Kuntz et al., 2016), would need to be identified, stored, processed, and coordinated for crisis response and recovery, implying that a TMS could potentially serve this purpose. Teams with individuals who each hold diverse knowledge domains are also more likely to combine them to improve accessibility for those who do not have such a knowledge base (Taylor & Greve, 2006), if a TMS exists. While knowledge combination is inherently difficult and occurs most easily when a team has past experience working together (Taylor & Greve, 2006), such teams and other organizational networks are repositories of knowledge, including specialized knowledge for crisis response. Resilient employees through their ability to share and leverage resources (Raetze et al., 2021) can embed this specialized knowledge in organizational systems such as a TMS. Individuals with strong knowledge are more likely to tie in with each other, more open to novel ideas, and can bring more resources to tasks (Taylor & Greve, 2006). A TMS through its coordination component would facilitate the extraction of resources and knowledge from resilient employees and make these available to others. Working together and collaborating are essential facets of employee resilience (Hartmann et al., 2020; Kuntz et al., 2017) during adversity, and these practices can improve knowledge credibility emanating from individuals and teams as they become codified and embedded in a TMS (Hollingshead, 1998). Thus, it seems that employee resilience could influence the specialization, credibility, and coordination facets of a TMS through knowledge transfer that captures resilient practices and behaviors, making them available to support others during crisis response and recovery. Thus, we propose:
Employee Resilience and Dynamic Capabilities
If employee resilience is reflected in the capacity to utilize and proactively develop personal and workplace resources (Kuntz et al., 2016) during adversity, then these resources must emanate from collaborative practices (Hollingshead, 1998). Given that resource deployment relies on the individual skills and knowledge that employees possess (Teece, 2012), it is plausible that resilient behaviors may strengthen the capacity to leverage dynamic capabilities during adversity. However, this proposition remains untested in the dynamic capabilities, tourism, and management literatures (see Table 2). Given that managers and employees initiate and shape the exploitation of market opportunities through sensing and seizing capabilities (Ambrosini et al., 2009), the interaction between individual and organizational capabilities can affect which opportunities are identified and how they are exploited. Resilient employees tend to be adaptable and learn from experiences (Kuntz et al., 2017), which are critical behaviors during crises. These capacities can enable organizations to better utilize knowledge as a resource to locate new opportunities, heightening sensing capacities. Individual elements such as characteristics, cognitions, and abilities are important building blocks for understanding collective phenomena such as routines and capabilities (Felin et al., 2012). For example, middle managers were found to actively push for opportunities for refinement and seize them when these were initiated internally by senior managers during crises (Funke et al., 2023). This is because learning, experience, resources, and routines are inputs to capabilities that reside in both individuals and organizations. Different combinations of these are required at different times (Felin et al., 2012).
Different behavioral capabilities of organizational members may influence routines and capabilities. Experiential learning in particular is a key facet of dynamic capabilities development (Felin et al., 2012). Forward-looking capacities of employees facilitate problem-solving and the identification of novel solutions, which are embedded in dynamic capabilities (Teece, 2012). Since routines and capabilities can also involve patterns of interdependent actions by employees, their relational ability, which is akin to the networking ability of employee resilience (Kuntz et al., 2016), and their ability to integrate different knowledge elements may affect the execution and outcome of reconfiguration (Felin et al., 2012). During crises, managers who ensured that reconfiguring capabilities were efficient, were also those who sensed and shaped additional complementary opportunities from the marketplace (Funke et al., 2023). Thus, it seems that capabilities of adaptability, learning, and networking that demonstrate employee resilience can strengthen the sensing, seizing, and reconfiguration facets of dynamic capabilities. Thus, we propose:
Employee Resilience and Social Performance
Existing tourism and hospitality studies (Prayag & Dassanayake, 2023; Saad & Elshaer, 2020) have evaluated the employee resilience-performance relationship (see Table 2), focusing on financial performance. In the management literature, employee resilience has been linked to improved firm performance (Santoro et al., 2021). Yet, sustainable performance remains rarely considered as an outcome of employee resilience. Fundamentally sustainability is an ethical issue for organizations, requiring socially moral people and managers that enact practices that support sustainability performance (Iqbal et al., 2020). For instance, based on social standards organizations formulate and implement policies, procedures, and practices that facilitate their sustainability agenda. Social performance emerges from corporate social responsibility activities, working conditions, and organizational practices (Alghababsheh & Gallear, 2021). In trying to improve social performance, organizations should be monitoring working hours, health and safety, and employee welfare amongst others (Huq et al., 2016; Sancha et al., 2016). For employee resilience to boost social performance, the networking ability of employees is critical. Given that resilient employees can draw from networks and external resources to navigate adversity (Kuntz et al., 2017) suggest that collaborative practices within the organization matter. Through collaborative practices, worker welfare, and a safer workplace environment can be created (Alghababsheh & Gallear, 2021; Klassen & Vereecke, 2012). Through employee adaptability, new opportunities can be created and tapped into to improve sustainability across all offerings and processes (Klassen & Vereecke, 2012). Resilient behaviors are showcased through employees embracing rather than resisting change in the workplace. Such employees are life-long learners but can also learn quickly to respond to a dynamic environment (Kuntz et al., 2016). This implies that resilient employees through learning can drive the implementation of new social standards, including safer and more inclusive work practices. In this way, resilient employees can potentially strengthen social performance. Thus, we propose:
Transactive Memory System and Dynamic Capabilities
Individuals working in teams often struggle to share knowledge internally and externally due to cross-functional silos that impede knowledge transfer (Richey et al., 2012). Also, reciprocity can be an issue in knowledge transfer externally affecting the ability to leverage knowledge during crises (Scheibe et al., 2022). Organizations with a TMS can have superior information search capability, which facilitates the efficient application of knowledge and more fluid adaption to performance contingencies (Bachrach & Mullins, 2019). In response to dynamic environments, a TMS can help teams quickly locate and selectively leverage information and expertise in real-time (Bachrach & Mullins, 2019). If a TMS can facilitate access to a greater amount of knowledge, encourage knowledge sharing, and enable team members to cultivate specialized expertise (Lewis, 2003), then dynamic capabilities can be impacted positively given that knowledge resources are integral to its functioning. In particular, Y. Ren and Argote (2011) suggest that knowledge acquisition, transfer, and absorption are behavioral outcomes of a TMS, implying that they constitute the core of an organization’s knowledge management capabilities. A TMS is a capability similar to dynamic capabilities (Argote & Ren, 2012). The ability of an organization to sense changes in the external environment and how these affect internal dynamics are related. In this way, knowledge management would be a necessary factor to develop and manage dynamic capabilities (Fernandez-Giordano et al., 2022; Oliva et al., 2019). The limited evidence on the relationship between a TMS and dynamic capabilities is from studies that are either conceptual (e.g., Argote & Ren, 2012) or qualitative (e.g., Ndlela & Tanner, 2023), but these tend to examine other types of dynamic capabilities such as agility (e.g., Fernandez-Giordano et al., 2022).
For instance, specialization of knowledge enables teams to make collective decisions that require such knowledge (Scheibe et al., 2022). In crises, this specialized knowledge enables tourism and hospitality firms to deploy dynamic capabilities (Bhaskara et al., 2023; Y. Jiang et al., 2023). Knowledge underpins the ability of organizations to leverage or modify current resources and routines to sense and seize market opportunities (Eisenhardt & Martin, 2000; Teece, 2007). However, existing tourism and hospitality literatures on dynamic capabilities (Y. Jiang et al., 2019; Mansour et al., 2019; Prayag, Jiang et al., 2023) do not explicitly acknowledge that teams are potentially the initial knowledge source that allows organizations to cope with crises and disasters. A TMS can codify both implicit and explicit knowledge in teams to facilitate sense-making (Scheibe et al., 2022). In a TMS the credibility of team members’ knowledge is essential, with the implication that collaborators need to have confidence in each other’s capacity to execute tasks that are highly specialized in nature (Scheibe et al., 2022). This implies that despite knowledge being available through organizational systems, those using this knowledge for market sensing, seizing, and reconfiguration opportunities may not always trust it. Both the autonomy of teams and a culture of trust are essential to response and recovery strategies during crises (Dahmen, 2023). Thus, it can be inferred that a TMS provides the credibility necessary for employees to leverage the knowledge embedded in the system for activating dynamic capabilities. In a TMS, coordination is related not only to disseminating knowledge but also to retrieving knowledge from within and outside of the organization and embedding it in the shared memory system (Scheibe et al., 2022). This is a core activity of market sensing, which then allows for quick resource reconfiguration during adversity. The flexibility of integrating existing knowledge from a TMS into an organization’s dynamic capabilities can facilitate reconfiguration opportunities in dynamic environments (Argote & Ren, 2012). Based on the above evidence, it is plausible that a TMS (specialization, credibility, and coordination) should have a positive influence on dynamic capabilities of sensing, seizing, and reconfiguration. Thus, we hypothesize that:
Transactive Memory Systems and Social Performance
A TMS provides the knowledge basis for superior performance (He & Hu, 2021). However, the main focus of existing studies has been on team (Bachrach et al., 2017) or firm performance (Bachrach et al., 2022) rather than sustainable performance (see Table 1). In hospitality firms, Batra et al. (2023) failed to establish the direct influence of a TMS on firm performance, while Guchait et al. (2014) demonstrated that a TMS can have a positive influence on both team performance and satisfaction. Yet, no evidence links the TMS to improved social performance of organizations. A TMS can facilitate the formation of more fully developed, completely integrated, and higher-quality strategic options (Bachrach et al., 2022) such as sustainability. A TMS should enable members to more efficiently gather information both from one another, and from external sources, and leverage this information more fluidly (Bachrach et al., 2022). In this way, trust, team spirit, and cohesiveness, which imbue social performance should be related to the credibility facet of a TMS. When organizations have a TMS, knowledge coordination should allow them to have quicker access to new work safety standards and CSR initiatives. A well-developed TMS can also provide access to deeper and more functional collective knowledge (Lewis, 2003), facilitating organizations to undertake CSR activities, matching job requirements with skills development in staff, and increasing trust between management and employees. In this way, specialized knowledge should enhance social performance. The presence of a TMS should facilitate teams to focus on aspects of the environment that sustain performance (Bachrach et al., 2022), implying that social performance can be prioritized if the competitive landscape or dynamic environment requires a shift to sustainability. Different facets of a TMS can arguably be leveraged to improve social performance but this evidence is lacking in the tourism and hospitality literatures. Thus, we propose:
Dynamic Capabilities and Social Performance
In dynamic environments, previous studies (Mohaghegh et al., 2021; Siems et al., 2021) suggest that dynamic capabilities remain indispensable for firms to change and reconfigure their resources and assets for continuous sustainability transformations. Yet, the relationship between dynamic capabilities and sustainability performance is contentious, with some studies documenting a positive relationship (Helfat & Winter, 2011), while others finding insignificant or negative results (Wilden & Gudergan, 2015). The assertion has been that the differing results depend on the level of dynamism in the environment and organizational structure (Harun et al., 2023) but also on divergent measures of dynamic capabilities and social performance. However, previous studies (Helfat & Winter, 2011; Wilden & Gudergan, 2015) have not considered the long duration of an adverse event and the multiple adjustments that organizations may require to respond. Different combinations of dynamic capabilities might be needed at different stages of the event to maintain or boost social performance. Using a different operationalization of dynamic capabilities compared to our study, Harun et al. (2023) demonstrated a positive influence on social performance when the environment is dynamic. This is because dynamic capabilities can act as a catalyst for employee development through training and knowledge sharing schemes that boost social performance (Shang et al., 2020). Dynamic capabilities through its sensing and reconfiguration capabilities can also help firms to implement social norms and meet stakeholders’ expectations through corporate social responsibility activities (G. Kumar et al., 2018). The same positive relationship has been established by A. Kumar et al. (2023) using different proxies for social performance such as gross wages and number of full-time employees. Mohaghegh et al. (2021) identify that lean management practices enable dynamic capabilities to have a positive influence on sustainable performance, including social performance, but again these authors are utilizing dynamic capabilities dimensions different from Teece (2007). On the contrary, Jain et al. (2023) could not establish a link between dynamic capabilities and social performance, while others demonstrate indirect effects via variables such as resource management capabilities (Shang et al., 2020). Most of these studies have not considered the pandemic context and the challenges of deploying limited resources and capabilities when market and tourist behaviors are constantly changing.
In evolving and adverse environments, the ability to sense the market can allow firms to identify opportunities for human resource development. Processes and structures make employees more nimble, flexible, and tuned into business needs (Garavan et al., 2016) that allow firms to swiftly respond during adversity. Effective sensing processes can improve social performance by enabling team spirit and cohesiveness in teams but also enhancing employability despite the challenges posed by external threats. These practices enable employees to better sense changing environmental conditions, initiate plans, and capitalize on opportunities (Garavan et al., 2016). In dynamic environments, new actors or stakeholders (e.g., regulatory agencies and funding agencies) can become more important, requiring firms to reconfigure resources as was the case during the pandemic. The development of stakeholder relationships through CSR activities can represent a continuous routine that requires dynamic capabilities (Garavan et al., 2016). Organizational routines that enhance collaborations may be perceived as being embedded in resource configurations (Eisenhardt & Martin, 2000) that facilitate employees to perform better. While dynamic capabilities seem to be capable of enhancing social performance, this proposition remains untested in the tourism and hospitality literatures. Thus, we propose:
Methodology
To test and validate our theoretical model, a quantitative survey-based methodology was employed.
Context of the Study
The initial lockdown in the UK on 23rd March 2020 saw flights grounded and tourism and hospitality businesses closed. Over the next few months, there were several lockdowns resulting in falling revenues, job losses, and widespread uncertainty for the industry (Ntounis et al., 2022). As noted in other studies (Price et al., 2022), there was no radical transformation of the industry by 2021, despite the UK government claiming that their Tourism Recovery Plan is “one of the UK’s great success stories” (HMCR, 2021). Another key facet of the plan was to build back better with a more productive, innovative, and resilient industry, maximizing the potential for technology and data to enhance the visitor experience and employing more UK nationals in year-round quality jobs (HMCR, 2021). At the beginning of 2023, recovery outcomes have not necessarily materialized as per the plan, with original targets reassessed given the continued challenges to international travel over the course of 2021 and 2022 and by new global crises such as high inflation and war (DCMS, 2023). Also, the resilience of the tourism sector has been tested given the continued labor shortages, which have been higher than average in the “accommodation and food services” sector (DCMS, 2023). These challenges provide the opportune time to revisit the resilience of the sector by examining employee resilience and its consequences on tourism and hospitality firms.
Survey Instrument and Control Variables
Employee resilience is a unidimensional scale consisting of nine items that were adapted from previous studies (Kuntz et al., 2016; Prayag et al., 2020). The TMS is a second-order latent construct (Fernandez-Giordano et al., 2022; Lewis, 2003; Scheibe et al., 2022) that was measured using a fifteen-item scale adapted from previous studies (Bachrach & Mullins, 2019; Lewis, 2003). Each dimension of knowledge specialization, knowledge credibility, and knowledge coordination was measured using five items. Dynamic capabilities as a construct is a second-order construct consisting of three dimensions, sensing (six items), seizing (four items) and reconfiguration (five items) adapted from previous studies (Ozanne et al., 2022; Teece, 2007). Social performance was measured using seven items adapted from previous studies (Harun et al., 2023; Mohaghegh et al., 2021; Shang et al., 2020). All items were anchored on a seven-point Likert-type scale between strongly disagree to strongly agree, consistent with previous studies (Ozanne et al., 2022; Scheibe et al., 2022). Instructions specified that items had to be evaluated about the impact of COVID-19 on the firm. Several firms and industry-level control variables such as firm size, age, and sub-tourism sector were also measured alongside individual-level control variables such as respondent age, length of work experience, and education (see Figure 1). The survey instrument was refined through minor changes in the wording of items and clearer instructions after pilot testing on 34 respondents from the same panel of participants for the main study.
Sampling Procedures and Data Collection
To build the sampling frame for the study, we approached a certified global panel provider who utilizes sign-ups from various industries through incentives to participate in surveys. Potential respondents were screened on several criteria (>18 years old, had to be in a supervisory role and above, and was working in the tourism, hospitality, or events industries). The screening criteria are consistent with previous studies examining employee resilience (Prayag et al., 2020; Saad & Elshaer, 2020; Senbeto & Hon, 2021). We had several attention checks to filter out unengaged responses (Scheibe et al., 2022). Potential participants were sent an information sheet and consent form about the survey, which when read and approved allowed them to access the survey link. The median time to complete the survey was 13 min. Data collection lasted 3 weeks (12 November to 3 December) in 2021 and resulted in 350 fully completed surveys. The respondent’s profile is presented in Table 3.
Respondent’s Profile.
Common Method Bias Test
We adopted two sets of techniques (i.e., procedural control and statistical control) to address the common method bias (CMB; F. Kock et al., 2021). Under procedural control, we deployed three sets of techniques to address CMB. First, we designed the survey instrument in consultation with academics who are experts in the hospitality, tourism, and resilience fields. Second, while designing the survey instrument, the dependent and independent variables of our proposed model were placed in different parts of the survey instrument to display the “psychological separation” of these two groups of variables (F. Kock et al., 2021). This technique helps mask the underlying relationships between these two groups of variables (F. Kock et al., 2021). Third, to reflect the appropriateness of respondents for this study, we selected managers (at different levels) of hospitality and tourism firms who are well aware of the firm’s existing practices and decision-making processes (MacKenzie & Podsakoff, 2012).
Under statistics control, we mainly deployed two techniques to evaluate CMB that are appropriate for partial-least square structural equation modeling (PLS-SEM). First, we used the unmeasured common latent factor method to evaluate CMB (Hew et al., 2018; Ozanne et al., 2022). Our result revealed that the ratio of average substantive variance (0.605) to average method variance (0.008) is 75.625. This higher number reflects a reduced risk of CMB in this study. Additionally, most of the method factor loadings are insignificant (see Table 4). Second, we adopted a collinearity assessment approach to evaluate CMB (N. Kock, 2015). We assessed inner variance inflation factor (VIF) values of all potential paths in the model (see Table 9) and our result demonstrated that all inner VIF values are less than the cut-off value of 3.3 (N. Kock, 2015). This further eliminates the risk associated with CMB.
Assessment of Common Method Bias Using Unmeasured Latent Factor Method.
Note. Substantive factor loadings (Ra) are the coefficients of all items loaded together to an unmeasured latent factor. Method factor loadings (Rb) are the coefficients of relevant items to a specific construct (Williams & McGonagle, 2016). ***p<0.001; **p<0.01; *p<0.05, and NS = Not Significant
Data Analytic Approach
Data analysis was conducted in two stages. In stage 1, PLS-SEM was employed to test the overall model. PLS-SEM was chosen because it does not require to satisfy the normality assumption for further statistical analysis (Hair et al., 2022). We assessed the data distribution of lower order scales using Zskewness and Zkurtosis values (Hair, Black et al., 2019). Our result (see Table 5) revealed that Zskewness values of all items of lower order scales (except a few) are lower than −1.96 and very few Zkurtosis values are higher than +1.96. These results established the non-normal data distribution of lower order constructs. We adopted the same techniques to evaluate the data distribution of higher order scales in the later stages of our analysis and our result confirmed the non-normal data distribution of two higher order scales (i.e., transactive memory system and dynamic capabilities; see Table 7).
Reliability and Validity of Lower Order Constructs.
Deleted the item due to low factor loading (<0.50).
In stage 2, configuration analysis was used. The main objective of an asymmetric approach such as configuration analysis is to examine how diverse antecedent arrangements result in distinct levels of outcome sets (Fiss, 2011; Olya & Gavilyan, 2017), allowing us to answer RQ2. Commonly referred to as case-oriented analysis, Fuzzy set Qualitative Comparative Analysis (fsQCA) represents the latest evolution of QCA, utilizing fuzzy logic (Geremew et al., 2024; Hosany et al., 2021; Olya & Gavilyan, 2017). fsQCA, a well-recognized set-theoretic approach, is widely used to identify consistent and sufficient models for predicting outcomes of interest within smaller populations (Ragin, 1987). In the context of this research, the process of choosing causal models that lead to social performance relies on two essential criteria: consistency, which assesses the proportion of observed cases aligning with a specified pattern, and coverage, which evaluates the relative significance of various pathways to an outcome (Ragin, 1998). As opposed to symmetrical thinking, this approach uses unique configurations of casual antecedents to explain the conditions that result in a set of desired outcomes (Olya et al., 2018).
PLS-SEM Results
Evaluation of Reflective Measurement Model
All items were loaded reflectively to their respective constructs significantly (p < .05) with factor loading greater than 0.70 (except Coord2 = 0.549). We chose not to remove this particular item from the Coordination scale as it satisfied the reliability and convergent validity requirement (Hair et al., 2022). We evaluated the reliability of all lower order scales (LOS) using Rho_A and composite reliability (CR) and our result showed that all LOS have Rho_A and CR values > 0.80 which surpasses the threshold value of >0.70 (see Table 5). We evaluated the convergent validity of all LOS using average variance extracted (AVE) and all LOS have AVE higher than the threshold value of 0.50 which satisfied the convergent validity requirement (see Table 5; Hair et al., 2022).
Discriminant validity was evaluated using the Fornell and Larcker and HTMT methods that are appropriate for PLS-SEM (Hair et al., 2022). Following Fornell and Larcker’s (1981) method, the square root of AVEs of all LOS are greater than its correlation coefficients with any other LOS. Similarly, following the HTMT method, all HTMT ratios are smaller than the conservative threshold of 0.85. The results of both methods established the discriminant validity of all LOS (see Table 6).
Assessment of Discriminant Validity (Lower Order Constructs).
Note. Diagonal numbers (in bold and italic) are the square root of AVE; the Lower half of the diagonal is the Fornell–Larcker criterion and the upper half of the diagonal is the HTMT0.85 ratio. ER = employee resilience; Spec = specialization; Cred = credibility; Coord = coordination; Sens = sensing capability; Seizi = seizing capability; Reconf = reconfiguring capability; SOP = social performance.
In this study, we conceptualized both TMS and dynamic capabilities as higher order scales (HOS). We adopted the disjoint two-staged approach that is appropriate for PLS-SEM to model both HOS as reflective-reflective scales (Becker et al., 2023). In the first stage, we generated the latent variable scores (LVS) for all LOS that are related to these two HOS after satisfying the internal consistency and convergent validity requirement. In the second stage, these LVS are used as reflective indicators to model HOS. For example—LVS for specialization, credibility, and coordination are used reflectively to model TMS. We examined the internal consistency (using Rho_A and CR) and convergent validity (using AVE) of these two HOS and the result presented in Table 7 satisfied the requirements. We further tested the discriminant validity of the two HOS by combining them with two LOS (i.e., employee resilience and social performance) using the same two methods adopted previously. Our result presented in Table 8 established the discriminant validity.
Reliability and Validity of Higher Order Constructs.
Assessment of Discriminant Validity (Higher Order Constructs).
Lower order scales (LOS).
Note. Diagonal numbers (in bold and italic) are the square root of AVE; the Lower half of the diagonal is the Fornell–Larcker criterion and the upper half of the diagonal is the HTMT0.85 ratio. ER = employee resilience; TMS = transactive memory system; DC = dynamic capabilities; SOp = social performance.
Evaluation of Structural Model
Hair et al. (2022) presented a robust guideline to evaluate the structural model for PLS-SEM which is adopted in this study. Based on this guideline, we first evaluated the existence of a collinearity problem in the model using inner VIF values with a threshold value of <3.3 and all structural paths in the model satisfied this requirement. Second, we evaluated six proposed hypotheses using path coefficient (β) and bias-corrected and accelerated (BCa) confidence intervals (Hair et al., 2022). Our result (see Table 9) revealed that employee resilience has a positive effect on a TMS (β = .766, t = 24.491, Bca CI: 0.699, 0.821), dynamic capabilities (β = .282, t = 3.878, Bca CI: 0.137, 0.422), and social performance (β = .195, t = 2.549, Bca CI: 0.039, 0.336), thus supported H1, H2 and H3 respectively. A TMS is positively related to dynamic capabilities (β = .523, t = 7.860, Bca CI: 0.387, 0.650) and social performance (β = .340, t = 4.754, Bca CI: 0.204, 0.483). As a result, we got support for H4 and H5. However, we found no effect of dynamic capabilities on social performance (β = .104, t = 4.754, Bca CI: − 0.074, 0.276), consequently, H6 was not supported. Third, we evaluated the model’s explanatory power using R2 values (Hair et al., 2022). Three dependent variables in our model have R2 values of 61.2% (TMS), 58.9% (dynamic capabilities), and 36.0% (social performance) respectively, which demonstrate the substantial explanatory power of the model. We also evaluated the strength of our five supported hypotheses using the value of effect size (f2) (Hair et al., 2022). Our result showed that we have three small effects (≥0.02 to <0.15), one medium effect (≥0.15 to <0.35), and one large effect (≥0.35). Fourth, we used PLSpredict with 10-folds and 10 repetitions to evaluate the model’s out-of-sample predictive power (Hair et al., 2022; Hair, Risher et al., 2019). Our result confirmed that all indicators of dependent variables (i.e., TMS, dynamic capabilities, and social performance) have Q2 predict > 0 which demonstrated that it performed better than most naïve benchmarks. We extend our analysis by comparing the root mean square error (RMSE) or mean absolute error (MAE) value of PLS-SEM with the naïve linear model (LM) benchmark (Hair, Risher et al., 2019). We found that PLS-SEM generates smaller prediction errors for most of the indicators of dependent variables, which explains its medium prediction power.
Hypothesis Testing.
Note. Industry-level control variables: DV_HospInd = hospitality industry (Dummy variable); and DV_TourInd = tourism industry (Dummy variable). Firm-level control variables: FirmSize = firm size; and FirmAge = firm age. Individual level control variables: Age = respondent’s age; LWE = respondent’s length of work experience. Latent variables: ER = employee resilience; DC = dynamic capabilities; TMS = transactive memory system; SOp = social performance.
We evaluated the impact of eight control variables (i.e., two at the industry level, two at the firm level, and four at the individual level) on three dependent variables (i.e., TMS, dynamic capabilities, and social performance) in the PLS-SEM model (see Figure 2). Out of these twenty-four paths in the structural model, only two paths are significant. Our results presented in Table 9 showed that a TMS is negatively impacted by the hospitality industry (β = −.301, t = 2.207, Bca CI: −0.552, −0.014). This implies that the hospitality industry may not encourage managers to conduct team-oriented activities. We also found that firm size has a positive effect on a TMS (β = .089, t = 2.585, Bca CI: 0.023, 0.156), suggesting that larger firms are in a better position to promote managers to perform team-level activities and decision-making processes.
Structural model (values within the circle are R2 values).
As our data comprises a significant proportion of junior-level managers (n = 109) compared to those in more senior positions (n = 241), observed heterogeneity can account for some of the identified differences in paths coefficients. To address this issue, we conducted two sets of additional analyses. First, we conducted the Measurement Invariance of Composite Model (MICOM) test (Hair et al., 2024) between more senior versus junior managers. Results (Appendix 2) revealed no difference between the two groups. Second, we conducted a Multi-Group Analysis (MGA) between the two groups following three different techniques that is, bootstrap MGA, parametric test, and Welch-Satterthwaite test (Hair et al., 2024). Results (Appendix 3) again suggested no differences in the path coefficients of the hypotheses based on the groupings, further confirming the absence of observed heterogeneity in the dataset on managerial position.
Model’s Robustness Check
We evaluated the robustness of our structural model by performing two additional analyses that are appropriate for PLS-SEM that is, non-linearity and endogeneity test (Ai Humdan et al., 2024; Z. Jiang & Tu, 2023). We examined the non-linearity effects by testing the quadratic effects of employee resilience on a TMS, dynamic capabilities, and social performance; a TMS on dynamic capabilities and social performance; and dynamic capabilities on social performance. Our result (see Table 10) showed that all six quadratic effects are non-significant which demonstrate the absence of any non-linearity in the model.
Evaluation of Nonlinearity Effects.
Note. ER = employee resilience; TMS = transactive memory system; DC = dynamic capabilities; SOp = social performance.
We conducted the endogeneity test for PLS-SEM by following the Gaussian Copula method (Z. Jiang & Tu, 2023). We established eleven PLS-SEM models with all possible combinations of Gaussian Copulas (see Table 11). Our result presented in Table 11 demonstrated that none of these Gaussian Copulas have any significant effects which confirmed the absence of any endogeneity problem. Therefore, the outcome of these two additional analyses confirmed the robustness of our model.
Assessment of Endogeneity Using Gaussian Copula.
p < 0.001; **p < 0.01; and *p < 0.05
Fuzzy Set Qualitative Comparative Analysis (fsQCA)
Transforming Variables Into Fuzzy Sets (Calibration)
When multiple items are used to measure a variable or construct, we must compute a single value per construct, which will be utilized as input in fsQCA. The calibration of data concerning each variable was conducted using the fsQCA 4.1 software. According to Ragin’s (2008) guidance, the “AND” command within the software was employed to compute dynamic capabilities and TMS as latent constructs. Based on previous fsQCA studies (e.g., Fiss, 2011; S. Ren et al., 2023), direct data calibration method was applied, in which we set three values that define the degree of membership in the fuzzy set for each case (i.e., full-set membership, full-set non-membership, and intermediate-set; Pappas & Woodside, 2021; Ragin, 2008).
fsQCA uses calibrated measures, as data are transformed into the [0, 1] range to form fuzzy sets (Ragin, 2008). To calibrate the data, we selected the values 0.95, 0.50, and 0.05 as the three thresholds that will convert the data into the log-odds metrics where all values fall within the range of 0 to 1 (Hsu et al., 2013). In particular, concerning the seven-point Likert scales used in this research (ranging from 1 = Strongly Disagree to 7 = Strongly Agree), research suggest that the values of 6, 4, and 2 is proposed as suitable thresholds (Ordanini et al., 2014; Pappas et al., 2016).
Construction of Truth Table
After completing the calibration process, the fsQCA algorithm is utilized to generate a truth table comprising 2k rows, where k denotes the number of outcome predictor. The analysis was run in fsQCA 4.1, generating the truth table and using the Quine-McCluskey algorithm for the Boolean minimization (Ragin, 2008). The refinement of this truth table is based on two criteria: frequency and consistency. Frequency refers to the number of observations for each combination, while consistency measures the degree to which cases align with the set-theoretic relationships indicated in a solution. This algorithm’s data matrix, through standard analysis, enables the identification of complex, intermediate, and parsimonious solutions. Among these, the intermediate solution, as recommended by Ragin (2008), has been used for interpretation in this study. The frequency value assists in determining the essential configuration of antecedents, leading to the elimination of irrelevant conditions (Pappas & Woodside, 2021). The suggested frequency threshold of 1, led to the exclusion of combinations with lower frequencies from further analysis. Subsequently, a consistency cut-off of 0.8 is determined, which is above the minimum recommended value of 0.75 (Rihoux & Ragin, 2009).
FsQCA Results
Configuration for High Social Performance
To comprehend the configuration that leads to high social performance, we conducted fsQCA analysis, incorporating three key antecedent variables: employee resilience, dynamic capabilities and TMS. The calibration of data concerning each variable was conducted using the fsQCA 4.1 software. According to Ragin’s (2008) guidance, the “AND” command within the software was employed to compute dynamic capabilities and a transactive memory system as latent constructs. Raw consistency represents the reliability and adequacy of conditions (predictors) influencing the outcome of interest, while proportional reduction in consistency indicates the degree to which skewed data influenced the results (Cooper & Glaesser, 2016; Farivar & Richardson, 2020). The analysis revealed a configuration leading to a high social performance with high overall consistency (0.814) and total coverage (0.679), indicating that this configuration covered a significant portion of the sample with a high social performance. The results demonstrated that the presence of a TMS is necessary for achieving high social performance (see Table 12), confirming the results of PLS-SEM.
Results of fsQCA: Model for Predicting High Score of Social Performance.
Note. S = solution; soper = social performance; emres = employee resilience; dycap = dynamic capabilities; trmesy = transactive memory system.
Next, we included three control variables comprising of length of work experience (in years) of the employee, firm age (years in business), and firm size (number of employees) into the model to ascertain the robustness of the earlier results. The results (see Table 13) revealed that two solutions described conditions where high social performance can be achieved (coverage: 0.815, consistency: 0.678). In both solutions, a TMS is presented as a core condition, suggesting that high social performance is dependent on the presence of a TMS. Solution 1 stated that for bigger firms with higher lengths of operation in business, the absence of dynamic capabilities, and the presence of a TMS combined with a higher level of work experience result in high social performance. Alternatively, solution 2 indicated that high social performance can also be obtained when firms have a high TMS alongside employees with more years working for the company. In this solution, high social performance is achievable for large firms (number of employees is above 250) and older firms (20 years in business and more) provided they have a high TMS. Thus, the longer the business has been in operation, the more work experience employees have, and the larger the organization, seems to facilitate high social performance. Yet, in both solutions, a TMS is a necessary condition for all configurations. This implies that a TMS plays a significant role in achieving a high social performance.
Results of fsQCA: Configurations that Generate High Level of Social Performance.
Note. S = solution; soper = social performance; emres = employee resilience; dycap = dynamic capabilities; trmesy = transactive memory system; ywrfco = number of the years working for the company; ageco = company age; sizco = company size.
We have also divided the sample based on the industry level: hospitality versus tourism firms. This aims to identify whether different sectors within the industry, with their unique characteristics, can contribute to certain outcomes. The results (see Table 14) showed that for those working in tourism firms and other industries (rather than hospitality), the presence of employee resilience is essential to achieve high social performance. This can be argued as tourism encompassing a wider range of activities and environments, demands a high level of resilience from employees to manage varying conditions and unexpected challenges (Hall, 2010). Hospitality, although diverse, generally involves more predictable environments focused on service provision within established settings. In fact, this configuration highlights the importance of further investing in employee resilience in tourism firms to enhance social performance. Furthermore, the result showed the presence of a TMS is crucial, regardless of the industry type, for achieving high social performance.
Impact of Industry Type on fsQCA Results.
Note. S = solution; soper = social performance; emres = employee resilience; dycap = dynamic capabilities; trmesy = transactive memory system; Hospt = hospitality industry.
Robustness Tests for fsQCA
We also conducted robustness checks (see Table 15) to confirm the predictive validity of configurational solutions reported in Table 12. To check the robustness of the reported configurations, following Crilly (2011) and S. Ren et al. (2023), we applied the reduced consistency threshold to the acceptable threshold (0.75; Ragin, 2008). As such, the first solution appeared with a decreased overall coverage from 0.679 reported in Table 12 to 0.623, but a higher consistency of 0.842. Although all the solutions reported in Table 12 remained, the reduced thresholds as expected generated slightly different solutions which confirms the results are robust.
Results of the Reduced Threshold for the Robustness Test.
Note. S = solution; soper = social performance; emres = employee resilience; dycap = dynamic capabilities; trmesy = transactive memory system.
Discussion
Theoretical Implications
This study makes a theoretical contribution in addressing how knowledge management in the form of a TMS facilitates the utilization of dynamic capabilities for enhancing social performance during crises. More importantly, the role of employee resilience as a facilitator of a TMS is demonstrated. In this way, our findings provide an incremental step in tying individual-level resilience (employee) to tacit knowledge management in the form of a TMS. This happens through a process of embedding resilient behaviors of employees into a management system that allows teams to structure their own knowledge (specialization) and codify them in management practices and processes that can then be retrieved by others and shared (Liang et al., 1995; Liao et al., 2012) in responding to crises. Thus, resilient employee behaviors can foster trust in other team member’s knowledge (credibility) and improve knowledge coordination to achieve positive outcomes (Lewis, 2003), including improved collaboration (Scheibe et al., 2022). From this perspective, our findings validate the assertion that tourism and hospitality firms require both employee resilience (Prayag & Dassanayake, 2023) and team knowledge to navigate crises. The heightened emphasis on the resilience of employees (Senbeto & Hon, 2021) and health and safety practices in the workplace during the pandemic (ILO, 2022) are potential explanatory factors to how social performance can be maintained and improved. More importantly, we demonstrate the invisible hand of knowledge management and learning that binds a TMS to dynamic capabilities, having positive consequences on social performance. Therefore, our findings validate the importance of the individual (employee resilience)-team (transactive memory system)-organization (dynamic capabilities) practices in the sustenance of social performance during crises thus answering RQ1.
From a system thinking perspective, the results demonstrate the importance of studying interactions of different sub-systems within an organization. As the results demonstrate, micro-level processes such as those enacted to enhance employee resilience can positively influence meso-level processes such as a TMS and dynamic capabilities. As such we concur with previous studies (Prayag et al., 2020; Raetze et al., 2021) on the importance of understanding different levels of interactions between systems for resilience purposes. Aligning with the strategic management literature (Abell et al., 2008; Bouguerra et al., 2024) we confirm that individual level behaviors in organizations can trigger processes that affect firm-level actions and practices such as social performance during adversity.
From a methodological perspective, by integrating PLS-SEM and fsQCA to confirm and identify necessary and sufficient conditions for the relationships identified in Figure 1, we extend studies that primarily have applied fsQCA (Geremew et al., 2024; Olya & Gavilyan, 2017) to study other tourism and hospitality issues. In particular, the integration of fsQCA and PLS-SEM enables researchers to conduct a more robust evaluation of the predictive power of their models, considering both symmetric and asymmetric viewpoints (Rasoolimanesh et al., 2021).
We demonstrate that a TMS is necessary for tourism firms but less so for hospitality firms to achieve high social performance from fsQCA results, thus, answering RQ2. We confirm that these results are particularly true for larger tourism organizations that have been in business for a while, and where employees have substantial work experience. In tandem, this implies that smaller tourism firms that are start-ups, with limited experience among staff, are more vulnerable and more likely to make trade-offs between resilience and social performance during crises.
As a developable state in the workplace, employee resilience provides the behaviors that enhance a TMS (H1 supported). This suggests that collective work patterns, networking skills, and team member interactions (Hartmann et al., 2020) are useful for managing adversity only if these practices are embedded in the organizational fabric through a TMS. Thus, the adaptive and learning capacity of employees transforms into behaviors (Kuntz et al., 2016) that can be mimicked by others for crisis response and recovery through leveraging a TMS . In this way, we become the first study that establishes a positive relationship between employee resilience and a TMS, supporting that individuals can be catalysts in the development and functioning of a TMS (Mell et al., 2014).
The collaborative practices that emanate from employee resilience can become a driving force for leveraging dynamic capabilities during crises. As the results of H2 suggest, stronger employee resilience improves dynamic capabilities, confirming that dynamic capabilities deployment relies on individual skills and knowledge (Teece, 2012). Thus, extending the dynamic capabilities literature (Ambrosini et al., 2009; Y. Jiang et al., 2023; Q. Jiang & McCabe, 2021; Prayag, Jiang et al., 2023), we broaden the antecedents of dynamic capabilities in the tourism, hospitality and management literatures by demonstrating the important role played by employee resilience in enabling dynamic capabilities. In this way, the micro-dynamics that link a TMS to dynamic capabilities (Argote & Ren, 2012; Martin & Bachrach, 2018) are unveiled. Specifically, team knowledge captured through a TMS becomes the go to resource during crises for improving sensing, seizing, and reconfiguration capabilities in tourism firms. Thus, we demonstrate that the resilience capacities of employees are integral to understanding collective phenomena such as organizational routines and capabilities, which extend previous studies (Felin et al., 2012; Funke et al., 2023) that have not taken into consideration employee resilience as a knowledge resource. The results of H1 and H2 provide strong empirical support for the individual-team knowledge nexus in organizations to support dynamic capabilities development, which complements the existing literature that often fails to explain how dynamic capabilities develops (Fernandez-Giordano et al., 2022). As such, TMS theory when combined with the RBV of the firm and dynamic capabilities can explain micro-level to meso-level interactions and processes within tourism and hospitality firms that have consequences on sustainable performance.
While employee resilience has been shown to boost the financial performance of tourism and hospitality firms (Prayag & Dassanayake, 2023; Saad & Elshaer, 2020), we extend this finding to include social performance. During crises, employees value employment practices that display moral and ethical decision-making. Resilient employees can contribute to boosting social performance, as suggested by the results of H3, through supporting CSR practices and improving working conditions for others. The networking capabilities of employees provide a resource base to improve worker welfare and build a safer working environment. Thus, we extend studies that illustrate how collaborative practices enhance social performance (Alghababsheh & Gallear, 2021; Klassen & Vereecke, 2012) by centering resilient practices as a foundational element of collaboration. Resilient employees embrace challenges stemming from crises through their ability to learn quickly and respond to those challenges (Kuntz et al., 2016) which can drive a more inclusive and safer workplace during crises.
Managing disruptions is a knowledge intensive task for firms requiring the expertise of individuals and teams (Cotta & Salvador, 2020). Often teams struggle to share knowledge (Richey et al., 2012) but this is not the case when a TMS exists within an organization. The results of H4 confirm that team knowledge sharing practices embedded in a TMS can be utilized to drive sensing, seizing, and reconfiguration capabilities during a crisis. A TMS can help teams locate information quickly and leverage expertise in real time (Bachrach & Mullins, 2019) to improve dynamic capabilities. In this way, we provide evidence that a TMS is a complementary capability to dynamic capabilities as suggested in conceptual (Argote & Ren, 2012) and qualitative studies (Ndlela & Tanner, 2023) but also uncover that dynamic capabilities of different types to the study of Fernandez-Giordano et al. (2022) are linked to a TMS. Knowledge management becomes critical to the deployment of dynamic capabilities as suggested in previous studies (Bhaskara et al., 2023; Y. Jiang et al., 2023) and we illustrate the importance of team knowledge in this process, extending the antecedents of dynamic capabilities in the tourism and hospitality literatures, which do not examine a TMS (Mansour et al., 2019; Prayag, Jiang et al., 2023).
As shown by the results of H5, a TMS enhances the social performance of tourism and hospitality firms, extending studies that focus mainly on outcomes such as financial (Bachrach et al., 2022; Batra et al., 2023) and team (Bachrach et al., 2017; Guchait et al., 2014) performance. Thus, a TMS can provide deeper knowledge that enables organizations to undertake external activities such as CSR while also improving working conditions internally. Thus, during crises maintaining an internal-external focus on business activities is critical for sustaining sustainability related performance such as social performance. This can shift the competitive landscape of tourism firms during crises and maintain an organization’s focus on sustainability. However, trade-offs are also possible during crises to allow firms to focus capabilities and resources on key critical activities. As improvements in sensing, seizing, and reconfiguration capabilities do not directly affect social performance (H6 not supported), tourism firms can focus on employee resilience and a TMS in boosting social performance. The lack of support for H6 differs from studies (Mohaghegh et al., 2021; Siems et al., 2021) suggesting that dynamic capabilities allow firms to undertake sustainability related transformations. A plausible explanation can be that during crises the focus of tourism and hospitality firms shifts to recovery through financial performance rather than social performance, which implies that dynamic capabilities are leveraged for purposes other than sustainability as suggested in previous studies (Y. Jiang et al., 2023; Prayag, Jiang et al., 2023).
Managerial Implications
Owners and managers of tourism firms can deploy several activities, processes, and strategies to boost social performance. According to the results, the foundation of improved social performance is employee resilience. Thus, creating a psychologically safe environment for employees to express their feelings and concerns during crises is critical. Enabling employees to leverage internal networks for support during crises (Kuntz et al., 2017) and rewarding collaborative behaviors can drive knowledge sharing. Use recognition and reward systems as a means to enhance behaviors that support resilience building during adversity. Owners and managers need to lead by example to motivate employees to enact behaviors that display resilience such as providing assurance, direction, and support (Catalano, 2022). For example, managers can demonstrate adaptability through showing openness to change and being flexible under pressure, while also practicing work-life balance and self-care during adversity. Managers by delegating effectively, staying calm under pressure, and communicating effectively can lead by example. The focus of resilient practices should be on learning and adaptation to uncertainty by equipping employees with the ability to work flexibly and autonomously to reinforce their sense of control.
A TMS has important implications for human resource practices, through training programs, workshops, and technology systems to locate the knowledge. These systems and processes need to be aligned to facilitate teams to utilize knowledge for crisis response and recovery. Individual and team continuous performance must be monitored to locate knowledge and embed these into organizational practices through a TMS. There must be clarity in terms of team identification and intra-group processes, with employees knowing clearly which team(s) they belong to and the expertise held by each team. However, clarity on intra-group processes around shared group membership, goals, key performance indicators, and the psychological or affective ties within the group must be widely known to facilitate knowledge retrieval and coordination (Liao et al., 2012). TMS processes must also be coordinated around information allocation, directory updating, and retrieval of information (Liao et al., 2012). Using team building events to improve trust and reciprocity can facilitate knowledge sharing during crises. Improving cross-functional interfaces is critical for organizational goal achievement during crises given that a TMS feeds knowledge to other organizational systems and processes such as dynamic capabilities. This implies that participatory decision-making across team and management levels, job rotation, and establishing formal knowledge management structures are necessary.
The path-dependent nature of dynamic capabilities resides mainly within the key decision-makers (e.g., managers) for two potential reasons (Ambrosini & Bowman, 2009; Teece, 2007). First, managers need to closely monitor the external environment to identify any potential signal of changes in terms of shifting customer demand or the nature of competition (Ambrosini & Bowman, 2009; Harreld et al., 2007). Second, managers should act promptly to grab those opportunities by reconfiguring their internal and external resources (Ambrosini & Bowman, 2009; Harreld et al., 2007). So, to boost various elements of dynamic capabilities, managers must have access to updated information and knowledge. One potential avenue for managers to access real-time information is to invest in strengthening their technological capabilities (e.g., big data analytical capabilities and/or digital marketing capabilities) (Mikalef et al., 2019; Nguyen et al., 2022; Wamba et al., 2017). In parallel, effective deployment of such technology to boost dynamic capabilities also rests on two critical elements – first, sound technical, analytical, and knowledge base of existing employees, and second, a culture of the data-driven decision-making process (Grover et al., 2018; Mikalef et al., 2019). Therefore, it is imperative for managers not only to design appropriate training modules for employees to learn the use of those technologies for decision-making purposes within firms but also to hire people in the coming years to align with these requirements.
While information and knowledge are the backbone for effective deployment of dynamic capabilities, the acquisition of such critical information and knowledge from diverse sources demands managers to maintain and nurture strategic relationships with various internal (e.g., employees and managers of different departments) and external (e.g., key suppliers, customers, and other business partners) stakeholders (Blyler & Coff, 2003; Ozanne et al., 2022). Therefore, managers could meet with these stakeholders informally to strengthen their relationships and/or organize an annual event to formally invite all stakeholders to reinforce their bonding.
While employee resilience and a TMS can contribute directly to improving social performance, organizational practices that support social performance development must be clear. For example, leaders must be champions in social performance development through enabling human resource practices that improve employee engagement in social causes that are valued by the organization. Establishing a code of conduct as a frame of reference (see Figure 3), which provides written baseline expectations and standards of acceptable behaviors around social performance is necessary to guide internal behaviors and processes. To increase transparency companies can also demand that suppliers obtain well recognized certifications (e.g., SA8000), which focus on conducting business in a way that is fair and decent for workers and to demonstrate their adherence to the highest social standards. Auditing requirements that systematically evaluate working conditions and compliance with health and safety standards (Huq et al., 2016), can be part of social performance metrics that are assessed by external parties. Another important driver of a firm’s social performance is management commitment to ethical practices (Muller & Kolk, 2010) and government influence to comply with regulatory requirements (Weaver et al., 1999). These may require changes in strategy and structure to boost social performance.
Key practices to build employee resilience, TMS, dynamic capabilities, and social performance.
Conclusion, Limitations, and Future Research Direction
As outlined previously, the study makes several contributions to the tourism literature by showing the dynamics at play in improving social performance during crises. However, the study has certain limitations that must be acknowledged. First, we use a single informant survey approach which may not necessarily indicate that in the organizations surveyed, the practices and capabilities (employee resilience, TMS, and dynamic capabilities) assessed are in place (Scheibe et al., 2022). Second, the study implicitly assumes that all tourism firms surveyed were impacted similarly by COVID-19. It is, therefore, possible that the strengths of the relationships uncovered in this study may be due to tourism firms being more impacted by the pandemic compared to other service industries. Third, the results are not generalizable beyond the context of the UK and that of COVID-19. Fourth, there are other types of dynamic capabilities (Prayag, Jiang et al., 2023) and the results pertain only to sensing, seizing, and reconfiguring dynamic capabilities. Fifth, TMS and dynamic capabilities were modeled as second-order constructs in both PLS-SEM and fsQCA analyses, potentially diluting the complex interactions between their sub-dimensions. Future studies could examine which combinations of sensing, seizing, reconfiguring as well as which combinations of specialization, credibility, and coordination could stimulate social performance. This would help in developing pragmatic implications for owners and managers to learn how to prioritize and plan for resources to achieve higher social performance. Sixth, given the timing of the survey, there is a wide distribution of seniority of employees within the sample. More senior managers have answered the survey compared to junior ones, which may imply that they have been working in a near constant state of resilience given the challenges of the pandemic and therefore have higher resilience levels. Seventh, the study is based on a combination of firm-level and employee level concepts, which is not uncommon in tourism and strategic management literatures. Future research can do a multi-level study combining firm-level with employee-level data to further evaluate micro-meso sustainability and resilience enhancing processes and interactions within tourism and hospitality organizations.
These limitations also give rise to other future research direction. Future studies should use multiple informants from the same organization to ascertain organizational practices around employee resilience and a TMS. In particular, team identification, knowledge acquisition and sharing practices, and social capital in teams can be included in future studies to understand how employee resilience further supports the development of a TMS. Within tourism and hospitality studies, leadership as a driver of both employee resilience and a TMS would be a fruitful area of future research but also how team size, team performance, and team resilience can confer positive organizational benefits. Within tourism studies, assessing the behavioral, performance, and affective outcomes of a TMS (Y. Ren & Argote, 2011) remain in their infancy. Another area of future research is the outcomes of a TMS such as new product development, creative performance in teams, team cohesion, and other types of dynamic capabilities such as replicating, integrating, and assimilating, amongst others (Prayag, Jiang et al., 2023). Another important outcome of dynamic capabilities is organizational resilience (Y. Jiang et al., 2019; Ozanne et al., 2022; Prayag, Jiang et al., 2023) and so far the tourism and hospitality literatures have not comprehensively evaluated this relationship and whether a TMS could strengthen both dynamic capabilities and organizational resilience. A future study can reassess with the sampled employees if they still work in the tourism and hospitality industry and if so, assess again their resilience in relation to other crises including the labor shortage issues stemming from Brexit and the cost-of-living crisis.
Footnotes
Appendix
Multi-Group Analysis (MGA) Result.
| Original path-coefficients | Differences | Bootstrap MGA (two-tailed p-value) | Parametric test (p-value) | Welch-Satterthwaite test (p-value) | ||
|---|---|---|---|---|---|---|
|
|
All positions |
Junior positions |
||||
| Employee resilience → Transactive memory system | 0.741 *** | 0.837 *** | −0.096 | 0.090 | 0.119 | 0.081 |
| Employee resilience → Dynamic capabilities | 0.276 ** | 0.269 * | 0.006 | 0.999 | 0.967 | 0.967 |
| Employee resilience → Social performance | 0.202 ** | 0.173 * | 0.029 | 0.913 | 0.855 | 0.876 |
| Transactive memory system → Dynamic capabilities | 0.525 *** | 0.543 *** | −0.018 | 0.915 | 0.903 | 0.897 |
| Transactive memory system → Social performance | 0.273 ** | 0.543 ** | −0.269 | 0.125 | 0.098 | 0.136 |
| Dynamic capabilities → Social performance | 0.150 | −0.139 | 0.289 | 0.107 | 0.086 | 0.091 |
p < .05. **p < .01. ***p < .001.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
