Relation between ethical sales behaviour and switching costs: a mediation of trust in medical schemes

Selecting a medical scheme and health benefits in South Africa is a challenge encountered mainly by new members. Consumers often rely on the information presented by sales practitioners, whose low ethical standards (i.e. false promises or coercion into purchases) are perceived negatively (Chonko et al., 1996; Thomas et al., 2002), later instigating consumers’ switching behaviour. South Africa has a dual health-care system, where the private sector covers 16% of the population and 84% is covered in the public sector (M’bouaffou et al., 2022). In 2022, the South African population of medical scheme members declined from 16% to 15.8%. Groups of white (71%) and Indian/Asian (48.7%) people were significantly more likely covered by medical schemes than coloured (18.2%) and black (9.7%) individuals (Statista, 2023).

According to Moeti et al. (2023), greater efforts are needed to ensure all citizens have access to public health care, especially disadvantaged individuals like informal dwellers. Despite South Africa’s progressive constitution, which secures citizens’ rights to quality health care and protection for human rights, the quality of the health sector is rapidly declining (Republic of South Africa, 1996, p. 13). Several challenges in South Africa’s delivery of quality health care have been identified (Maphumulo and Bhengu, 2019), including imbalanced resources distribution, disaster management and poor leadership skills, slower progress in restructuring the health-care system, increased disease burden and push-and-pull dynamics, such as governmental strategies to improve the quality of health-care delivery.

Shi and Singh (2015) highlighted several areas that have been targeted for improvement in the US health-care system under the new law (i.e. revive prevention, improve care coordination, assist hospitals to improve quality and enhance quality reporting requirements). They noted that while federal assistance to primary care infrastructure is inadequate, it will help improve it and federal support should be provided for “generic” (biosimilar) versions of some biologics and insurance coverage for low-income citizens and certain vulnerable groups. Admitting similar challenges in the health-care reform in South Africa compared to the USA, Hilsenrath and Meyer (1996) noted that the role of the state in health care in the two countries differs significantly. In South Africa, the Department of Health is liable for overseeing and coordinating the provincial health services, with public facilities being the primary source of care for most of the population. Conversely, in the USA, most health-care expenditures in the public sector are funded through federal public insurance programmes (e.g. Medicare and GHAMPUS and the joint federal/state Medicaid programme), though they do not offer it. In addition, the Department of Veterans Affairs and military hospitals offer some direct health care. Katuu (2018), who analysed the three framework models that categorise South Africa’s health-care system typologies compared to Canada and the USA, reported that despite similar implementations, health-care systems in many countries are not isolated, as they function within a larger context, meaning they are not static, but always evolving, with many nuances.

In South Africa, the objective of the National Health Act 61 of 2003 (NHA) is to effectively promote and improve the national health system by promoting a spirit of cooperation and shared responsibility among public and private health-care professionals, providers and other relevant stakeholders within the context of national, provincial and district health plans (Council for Medical Schemes [CMS], 2020). The Charter of the Public and Private Health Sectors of South Africa and the Presidential Health Compact, supported the NHA in a Presidential Health Summit in October 2018, committing the public-private partnership engagement in more prolific discussions addressing the key challenges encountered in both sectors to improve the health-care delivery system for South Africans. Such a system must be coherent, quality-driven, efficient and cost-effective and optimise the use of both sectors’ resources to benefit the entire population (CMS, 2020). These discussions are progressing through regulatory interventions and will be accelerated towards the full execution of the National Health Insurance (NHI) by 2025 (CMS, 2018; Katuu, 2018; McKenzie et al., 2017). The NHI white paper details a progress towards universal health coverage entailing renovating and reconfiguring institutions for the pooling of funds and purchase of public and private health-care services to accomplish income and risk cross-subsidisation, while refining efficiency and effectiveness in purchasing personal health services (CMS, 2020; McKenzie et al., 2017). However, Persaud (1991, p. 10) asserted that the effect of any cost-containment on medical ethical behaviour should be studied, asking: “What future for ethical medical practice in the new National Health Service?”

In South Africa, a medical aid scheme coverage operates as an insurance plan offering consumers financial protection for medical treatment and related medical expenses in exchange for monthly contributions. It can either be open or closed (restricted). There are 21 open and 55 restricted schemes (Profmed, 2019) accessible to consumers in South Africa. In 2019, the ten largest open and restricted schemes with accredited managed health-care services (no transfer of risk) covered almost 8,857,408 average beneficiaries (or 99.13%), which increased by 8.38% to R4.69bn in 2019 from R4.33bn in 2018 (CMS, 2020). Anyone can join an open medical scheme, but closed schemes have a particular qualifying criteria, including direct affiliation with clients’ occupation, employer, profession or industry, in which case, a person joining an employer offering this benefit is permitted to join. For example, Profmed is a closed scheme for professionals, ranking under the top ten largest medical schemes and covers over 34,700 families. Another largest restricted medical scheme is the Government Employees Medical Scheme (GEMS), which covers over 690,000 families (Profmed, 2019).

From the R14.1bn paid to supplementary and allied health professionals in 2019, restricted health schemes spent just over R6.2bn and open medical schemes spent R7.9bn. Furthermore, the restricted schemes cost an average of R2,851 for each event in hospital, compared with open schemes’ R1,478, although the in-hospital surgical specialist visits, the use of magnetic resonance imaging scans, angiograms, bone density scans and dialysis services, are mostly higher in open medical schemes than in closed schemes (CMS, 2020). Open schemes, compared to restricted medical schemes, paid 1.39% more benefits for hospital services and 2.59% more for specialists. Moreover, unlike the restricted schemes, open schemes paid 0.94% more towards managed care arrangements (CMS, 2020). Usually, restricted schemes do not have to cover large marketing and broker costs and use that money to pay member claims (Profmed, 2019). Contrastingly, open medical schemes in South Africa that incurred broker fees, marketing and advertising expenditure and with a loyalty and wellness programme gained a 0.36% increase in the number of main members from 2018 to 2019 (2017–2018: 0.71%) (CMS, 2020), but this growth excluded younger consumers and was not linked to new members who were not covered by any scheme before, as these were people who switched from other schemes (Competition Commission South Africa, 2018).

Nevertheless, it is unclear why people switch between medical schemes. The concerns of competitive reforms on disparities, the need to effectively compare health plans or to have supportive infrastructures that provide information and facilitate choice and the requirements that dissatisfied enrollees switch from one plan to another are less known, all almost certain to be visible among advantaged populations (Axtell-Thompson, 2005). The Healthcare Consumer Survey 2016 conducted by Competition Commission South Africa (2016) probed consumers’ experiences of private health-care services. Among concerns discussed were consumers’ understanding of the legal difference between health insurance and medical schemes and their decision to join or switch medical schemes. For instance, using computer-based information tools, consumers search, evaluate and choose health-care plan designs, services, benefits and providers and accept the financial risks (Axtell-Thompson, 2005). Accessing information on some health plan features (e.g. services covered, leading service providers in the plan’s network and out-of-pocket premium) at a lower cost (e.g. if the employers release information on health plan quality), depending on what they learn, consumers may view another plan as a better quality than their current plan and switch (Abraham et al., 2006). Buyers incur switching costs or one-time costs when switching from one supplier to another (Porter, 1980, p. 10). Literature shows that perceived switching costs are positively related to trust in the service provider in various contexts (Aydin and Özer, 2005; Carter et al., 2014; Yen et al., 2011), though it is unknown how trust relates to switching costs typology between medical schemes.

Moreover, the ethical rules of the Health Professions Council of South Africa are cited as the reason for the lack of innovative models of health care and the development of alternative reimbursement models (Competition Commission South Africa, 2018). The CMS (2020) noted the deliberate and incorrect medical advice practitioners gave to members about the funding for treatment conditions listed as prescribed minimum benefits (PMBs). Members were told that medical schemes are obliged to pay the costs of treatment plans in full, despite their being non-designated service providers of those medical schemes, leaving members with huge medical bills to be paid out of their pockets. At times, the conditions were non-PMBs, but members were falsely told that they were PMBs. This practice has exposed the abuse of trust by medical advisors (CMS, 2020). As such, “We cannot trust that for-profit schemes will deliver better value for consumers given multiple information failures and adverse incentives shown to exist in the South African healthcare sector” (Competition Commission South Africa, 2018, p. 457).

Trust is defined as one party’s confidence in the exchange partner’s reliability and integrity (Morgan and Hunt, 1994, p. 23). For instance, perception is key to trust and the media reporting of misuse of public funds by commissioners could result in system-wide doubts about integrity (British Medical Association, 2017). The operational terms that Aydin and Özer (2005) used in measuring trust included reliability, ethics, service quality and cumulative process. Therefore, trust is a concept closer to customers’ perceptions of the ethical sales behaviour. In the context of this study, unethical sales behaviour is short-run salespeople’s conduct that enables them to gain at the expense of customers (Lagace et al., 1991; Román, 2003). Research shows the effect of unethical sales behaviour on trust in life insurance (Crosby and Stephens, 1987), financial services (Bejou et al., 1998), iPhone and MacBook corporate brands (Javed et al., 2019) and sportswear stores (Mansouri et al., 2022). However, it is unknown how ethical sales behaviour affects trust in medical schemes. Studying this knowledge gap will enhance recent literature. Loe et al. (2000) recommended researching the ethical ideas of ethical decision-making theory with marketing-related concepts (i.e. marketing orientation, quality and performance). Aside from the large body of marketing ethics literature produced to date, a few studies have examined ethical issues in the emerging markets of the world, which provides a definite reason to study consumers’ ethical judgements in such areas (Al-Khatib et al., 2005) and understand the effect of those judgements on trust and switching costs typology.

Considering that trust in public institutions, including medicine, is under scrutiny and that consumers no longer trust professionals with the same certainty (British Medical Association, 2017; Cosma et al., 2020), this research assesses trust as a crucial element of effective organisations in the long term (Agrawal, 2017), in which its central role has been acknowledged in medical relationships (Tang, 2011). Relationship marketing theory posits customer trust as a complete mediator (Bloemer and Odekerken-Schröder, 2002; Castaneda, 2011; Delgado-Ballester and Manuera-Alemán, 2001; Garbarino and Johnson, 1999; Han and Ryu, 2012; Morgan and Hunt, 1994; Singh and Sirdeshmukh, 2000) or a partial mediator (Ganesan, 1994; Selnes, 1998) of the buyer satisfaction–loyalty relationship. However, it is unknown whether customer trust mediates the relationship between ethical sales behaviour and switching costs typology in a medical scheme setting, which is a knowledge gap necessary to investigate in a definite view that trust positively relates to perceived switching costs (Aydin and Özer, 2005; Carter et al., 2014; Yen et al., 2011). Agrawal (2017) recommended a study that assesses trust as a mediating variable in the relationship between ethical climates and other organisational features, such as productivity, citizenship behaviour and commitment.

Moreover, Hong and Wang (2009) noted switching costs as an additional main concept in relationship marketing, which formerly advanced from the concept of transaction-specific asset proposed by Williamson (1975). In an extensive exploratory discussion of customer switching behaviour, Keaveney (1995) recommended further empirical examinations on factors, such as perceived ethics and switching costs. Consequently, this research offers a unique contribution to literature, adding the moderating effects of brand affiliation, monthly contributions and the number of dependent beneficiaries to the theoretical framework, as the consumption values are likely to be balanced against life values with the growing demand to build superior brand performance within the overall ethical business practice (Palazzo and Basu, 2007).

Accordingly, the research questions for this study are: what is the effect of ethical sales behaviour on trust and switching costs typology in medical schemes in South Africa? Does trust mediate the effect of ethical sales behaviour on switching costs typology? Do the medical scheme’s brand affiliation, monthly contributions and the number of dependent beneficiaries moderate these interactions? Therefore, the study’s objective is to examine how ethical sales behaviour and switching costs could be mediated by trust, based on ethical marketing theory (Hunt and Vitell, 1986), commitment-to-trust theory (Morgan and Hunt, 1994) and switching costs typology (Burnham et al., 2003). Unifying these theories is a key subject area of this study that often lacks empirical data. Probing these questions built on proper theories, concepts and ideas contributes to marketing literature and health-care practice, as the results can be used to enhance public policy. From this background, the theoretical framework for the study (see Figure 1) is validated via structural equation modelling (SEM) on AMOS 29 and PROCESS procedure for Statistical Package of Social Science (SPSS) release 2.041.

This study tests the switching costs typology developed by Burnham et al. (2003), comprising subcategories of procedural costs (i.e. time and effort resulting from economic risk, evaluation, learning and set-up costs), financial costs (i.e. loss of financial quantifiable resources resulting from monetary and benefit costs) and relational costs (i.e. psychological and emotional discomfort resulting from personal and brand relationship breakages and identity loss). Although this typology continues to be tested across industries (Carter et al., 2014; Deng et al., 2010; El-Manstrly, 2016; El-Manstrly et al., 2011; Ha et al., 2023; Jones et al., 2007; Matzler et al., 2015; Shen and Ahmad, 2022), no research has measured how it is affected by ethical sales behaviour, mediated by customer trust and moderated by brand affiliation, monthly contributions and the number of dependent beneficiaries in the medical scheme industry.

The next section of this paper presents the literature review, followed by the development of the study’s hypotheses. Thereafter, the research methodology is discussed, followed by the presentation and discussion of the main results. The theoretical contributions and practical implications are then explained, with the study’s limitations and future research directions concluding this study.

Ethical decisions are made in companies and the transaction situation (i.e. online or in-store) (Baker and Saren, 2010). Prior studies in business ethics have assessed the marketing exchange process and the participants in this process (Ferrell, 2004; Hunt and Vitell, 1986; Muncy and Vitell, 1992), as the marketing exchange process is the main area wherein most business ethical problems arise (Vitell and Festervand, 1987). Such ethical problems include unsafe or unhealthy practices, dishonesty, intimidating behaviour and conflicting interests (Gerrard and Cunningham, 2004). The “general theory” of marketing ethics (Hunt and Vitell, 1986, 2006) is the most widely cited and applied theory in marketing ethics, which explains that once individuals perceive the evoked set of alternatives, two kinds of evaluations will occur: deontological and teleological. This theory is applied in this study to assess how individuals’ perceived unethical sales behaviour – defined as short-run salespeople’s conduct that enables them to gain at the expense of customers (Lagace et al., 1991; Román, 2003) – affects trust and switching costs typology in medical schemes. As consumers may prefer to transact determined by trust and its actors (i.e. ethics, empathy, kinship and friendship), the social exchange relies more on unspecified, implicit duties, which depend on shared systems of meaning, belief and ethics, than on formal contracts (Nooteboom et al., 1997). Trust in firms flows from a social exchange process where customers interpret and reciprocate an organisation’s actions as the primary source of the personified organisation’s actions. This role is supported by social exchange theory, as the increments in trust increase the social embeddedness of the consumer–provider relationship, thus enhancing customers’ commitment to the relationship (Singh and Sirdeshmukh, 2000).

Behaviourally driven relationship marketing (BRM) is drawn from social exchange theory and organisational sociology and signifies “relationship marketing”. It uses political economy concepts (e.g. power and dependency) and social aspects (e.g. commitment, trust, expectations, cooperation, communication and conflict behaviour) in analysing channel relationships (Baker and Saren, 2010), which are the main elements of the commitment-trust theory of relationship marketing (Morgan and Hunt, 1994). Relationship marketing theory outcomes are explained by factors about the actors and their perspectives of the relationship. It explores key concepts, such as targeting profitable customers, using the strongest strategies for customer relationships, marketing to employees and other stakeholders and building trust as a marketing tool (Berry, 1995). As principal studies show (e.g. Crosby et al., 1990; Morgan and Hunt, 1994; Parasuraman et al., 1991), relationship marketing theory is built on trust, which is vital to form a relationship with another person or organisation. Commitment is necessary to maintain a relationship, which requires diligence (Baker and Saren, 2010). Hence, trust is essential to forming service-based relationships due to the intangibility nature of services. For instance, most services are difficult to evaluate before purchasing and experiencing them and some remain difficult to evaluate even after being performed (Berry, 1995). Services marketing literature shows that trust in service providers increases commitment to the consumer–provider relationship (Moorman et al., 1992). It specifies trust as a central relationship-building block (Berry, 1995), as one party must acquiesce to the other to relish the benefits of the relationship, and trusting alters power positions in relationships (Rosenbaum et al., 2006). Trust is a major determinant of commitment in any exchange relationship, referring to “one party’s confidence in the exchange partner’s reliability and integrity” (Morgan and Hunt, 1994, p. 23).

Furthermore, commitment-trust theory of relationship marketing (Morgan and Hunt, 1994) proposes commitment and trust as crucial mediating variables in building successful relationships in the offline context (MacMillan et al., 2005; Mukherjee and Nath, 2007). Consequently, this study measures the mediating effect of trust on the relationship between ethical sales behaviour and switching costs typology, providing rare and valuable insights in existing marketing literature and health-care practice. It also examines the moderating effects of brand affiliation, monthly contributions and the number of dependent beneficiaries in medical schemes. This supports Story and Hess (2010), who reviewed the trust-based commitment relationship model and explored how evolving committed customer relationships may change brand and customer behaviours. In search of an ethics metric by which to measure the ethical effect of commitment on brand behaviour, they reviewed several ethical frameworks.

Equally, marketing relationships are assessed by criteria of fairness and trust and refraining from opportunist behaviour. In inter-organisational relations with “high relational complexity”, the mutual interdependence ties relatively heterogeneous resources, making switching difficult. However, switching is quite easy in consumer relationships with “low relational complexity” (Baker and Saren, 2010). Kanchanapoom and Chongwatpol (2021) proposed strategies to retain prospective and profitable customers, as these customers are, on one extreme, willing to switch from one vendor to another based on the products, services and related campaigns they receive; or, on the other extreme, loyal to one vendor due to their past experiences, satisfaction or high switching costs. Switching costs are one-time costs buyers incur when switching from suppliers (Porter, 1980, p. 10). Contributions to switching costs theory include Selten (1965) and Von Weizsäcker (1984). In commitment-trust theory, switching costs foster brand trust and transfer intentions into action. Switching costs act as a deterrent against the thought of another service provider and build continuance commitment due to the desire to avoid the loss of investments in the relationship (Morgan and Hunt, 1994). As switching costs can result from investment by buyers in high-cost specialised equipment, investment in learning how to operate such equipment or the result of product specifications (Avgeropoulos and Sammut-Bonnici, 2014), these investments tie buyers to particular relationship inputs; if the relationship is terminated, these investments will be lost (Dwyer et al., 1987; Klemperer, 1995; Morgan and Hunt, 1994). The potential for such loss makes investment essential to switching costs (Burnham et al., 2003). In addition to relational governance mechanisms, Li et al. (2015) noted that the investment model extends the scope of social exchange theory. They found social exchange factors, such as transaction-specific investment and switching costs, help stimulate suppliers to involve customers.

Nevertheless, it is essential to recognise that while BRM offers theory-based tools for rigorous modelling of the relationships, it is relatively weak in its contextuality. It tries to explain the outcomes of the relationships (e.g. performance, satisfaction and relationship duration) by factors related to the actors forming the relationship or the relationship itself. It is noteworthy that organisations exist to modify market relationships by offering mutuality, thus turning them into marketing relationships (Baker and Saren, 2010). To overcome deficiencies and avoid relying on a single theory, this study unifies the explanatory power of social exchange elements, such as ethics, trust and switching costs, to explain their integration in the proposed framework for the study (see Figure 1). Hair et al. (2022, p. 13) referred to this as a “structural theory”, which specifies the relatedness of the constructs in the proposed framework for the study.  Appendix summarises the related studies in literature.

In this study, unethical sales behaviour concerns short-run salespeople’s conduct that enables them to gain at the expense of customers. Examples of such activities include lying or exaggerating about the benefits of a product/service, lying about availability, lying about the competition, selling products/services that people do not need, giving answers when the answers are unknown and implementing manipulative influence tactics or high-pressure selling techniques (Lagace et al., 1991; Román, 2003). This study adopts this definition to assess how ethical sales behaviour of medical schemes affects customer trust and switching costs typology. Ethical sales behaviour can play a key role in forming and maintaining long-term relationships with customers and create liability problems for firms’ salespeople via intentional and inadvertent statements (Román, 2003), as consumers’ subjective beliefs and ethical perceptions act as bases of attitude formation and may direct their buying behaviour to the detriment of some firms (Brunk, 2010). Therefore, marketing ethics is an organisational responsibility that is vital in maintaining long-term beneficial relationships (Ferrell, 2004).

The most significant effect of higher ethical burdens may be the cost of violating customer trust. Violating trust may carry far greater penalties than simple dissatisfaction (Story and Hess, 2010), since consumers are unlikely to remain or do further business with service providers that have violated their trust (Babin et al., 2004). The customer–firm relationship exists due to mutual expectations built on trust, good faith and fair dealing in their interaction (Ferrell, 2004). The adopted definition of unethical sales behaviour (Lagace et al., 1991; Román, 2003) reveals salespeople’s behaviour perceived by customers during the interaction. During this interaction, each business partner forms confidence that the other party will act with integrity and reliability (Morgan and Hunt, 1994; Munuera-Aleman et al., 2003). The operational definition of trust to salespeople in this study is based on that of Morgan and Hunt (1994), which views trust as a confidence in an exchange partner’s reliability and integrity. The perceptions of ethics will vary among consumer groups based on their trust levels (Al-Khatib et al., 2005), as customers’ trust is affected by low-pressure selling (Kennedy et al., 2001) and dishonesty (Beatty et al., 1996), which are parts of unethical sales behaviour. Consequently, ethical sales behaviour (e.g. fair play, honesty and full disclosure) is imperative to building customer welfare for organisations (Román and Ruiz, 2005).

Research shows that sellers’ ethical behaviour can generate customer trust (Koponen and Julkunen, 2022; Morales-Sánchez et al., 2020). Empirical data supports the idea that the greater the salesperson’s ethical behaviour as observed by customers, the greater the customers’ trust in that salesperson in the business-to-business (B2B) (Hansen and Riggle, 2009) and business-to-consumer fields (Román and Ruiz, 2005). For example, retail business ethics affect trust in the integrated social discount spaces (Diallo and Lambey-Checchin, 2017), salespeople’s ethics behaviour positively affected customers’ trust towards iPhone and MacBook corporate brands (Javed et al., 2019) and sportswear stores (Mansouri et al., 2022), and in the financial services (Bejou et al., 1998) and pharmaceutical sectors (Lagace et al., 1991). Therefore:

In an earlier study outlining a complete exploratory discussion of customer switching behaviour, Keaveney (1995) identified eight significant causes of customer switching behaviour in service industries. Keaveney and Parthasarathy (2001) noted that some of these eight major causes of customer switching behaviour could be related to dissatisfaction with the service (e.g. failed service encounters, core service failures and poor service recoveries) and extrinsic or situational factors (e.g. involuntary situations, inconvenience, competition, price and ethics). Factors including core service failure, attractiveness of alternatives, price, low satisfaction, low trust and anger and failed service quality may trigger switching intention (Chuang and Tai, 2016). Liang et al. (2013) showed that the identification of seven significant groups of customer service switching causes (i.e. ethics problems, inconvenience, core service failure, high price, service encounter failure, family/friends/group impact and competition) has major implications for service marketing and service innovation research. Keaveney (1995) recommended further empirical research into search behaviour, affections, reference prices, ethics and switching costs (i.e. psychological, financial and time), noting that the view that price, competition, ethics and involuntary factors cause customers to switch services implies a need to examine these variables in addition to the more frequently researched variables. Caruana (2003) and Lappeman et al. (2022) reviewed the classification of defectors proposed by Keaveney (1995) to examine consumers’ switching behaviour. Lappeman et al. (2022) found that the implications of turnaround time, convenience and ethics drive consumers’ decisions to switch in the South African retail banking.

Moreover, brands have recognised that there are real, if not often quantifiable, costs and rewards related with un/ethical behaviour in the market (Story and Hess, 2010). For example, in the mobile phone service sector, the charge for temporarily stopping service is not for using the service, but for not using the service. This practice deviates from the social norm of charging for service and is thus considered unethical, leading to customers switching to other service providers (Liang et al., 2013). Colwell et al. (2011) showed how switching costs may limit the degree to which consumers view supplier enforcement of ethical codes of conduct. Perepelkin and Wilson (2018) showed the non-financial costs (e.g. selling consumer data) as an ethical concern and insisted these views be assessed in the community pharmacy setting. To the best of the researcher’s knowledge, the existing literature is silent on this assertion in the context of medical schemes’ ethics. Burnham et al. (2003) insisted that businesses have strong incentives to ensure good switching cost management and customer satisfaction. In addition, ethical as well as practical considerations indicate that firms should seek to increase switching costs in a way that adds value to consumers. Increasing customer switching costs and improving customer satisfaction is possible when firms help their customers better understand their products, identify the unique features they offer, offer valuable bonus points or loyalty services and engage them in more meaningful relationships. Each of these factors relates to the typology of switching costs. Therefore:

Trust is often viewed as a dyadic construct in which the behaviour of one party influences the perceptions and actions of another party (Agrawal, 2017). The Morgan and Hunt (1994) definition of trust was adopted in this study, which considers trust as one party’s confidence in the exchange partner’s reliability and integrity. The cognitive dimension of trust relates to evaluating the referent’s reliability, integrity and fairness. In contrast, the affective dimension of trust reflects a special relationship with the referent and is based on the emotional bond between parties. At a micro level, trust is often described as a dispositional attitude or state of mind (Agrawal, 2017). It is a customer’s belief that obligations will be fulfilled (Román, 2003). In the relational sales view, trust is a confident belief that salespeople can be relied upon to behave in customers’ long-term interest (Crosby et al., 1990; Román, 2003). In this view, trust is essential in preserving relationship investments by cooperating with exchange partners, resisting attractive short-term alternatives in favour of the expected long-term benefits of staying with existing partners, viewing high-risk actions as being prudent due to the belief that partners will not act opportunistically (Morgan and Hunt, 1994). Thus, trust is theorised as a cumulative process built over repeated, successful interactions (Román, 2003).

Scholars (e.g. Moorman et al., 1993; Román and Ruiz, 2005) propose that future studies on ethical behaviour and trust measure trust more validly and incorporate measures of relationship marketing theory. Relationship quality is a multidimensional concept that reflects trust, commitment and transaction-specific investment. Customers with high relationship quality feel greater betrayal and their trust in the suppliers is broken when they perceive high switching costs (Li et al., 2022). The definition of switching costs adopted in this study is from Porter (1980), which refers to switching costs as one-time costs, rather than the ongoing costs accompanying the use of a product or service after establishing a repeat purchase relationship. As customers incur one-time costs in the switch from one provider to another (Burnham et al., 2003), maintaining strong communication with customers is essential to generate trust and positive attitudes towards the brand and reduce the likelihood of customers switching service providers (Farah, 2017). As a core attribute and central construct in relationship marketing, trust affects perceived switching costs (Carter et al., 2014; Hidayat and Idrus, 2023; Kaur and Soch, 2018; Qayyum et al., 2013; Yen et al., 2011).

Nonetheless, studies on the downstream effects of diverse types of switching costs are scant (Jones et al., 2007). In literature, switching costs are viewed as a multidimensional construct. For instance, switching costs refer to the total costs incurred in leaving one service provider for another (Aydin et al., 2005; Ram and Wu, 2016), which include monetary and non-monetary costs (time and effort) (Burnham et al., 2003; Jones et al., 2000) and profit losses derived from loyalty (Han and Ryu, 2012; Xhema et al., 2018). From buyers’ perceptions, switching costs are perceived, anticipated and experienced costs consumers incur to switch a relationship from an existing seller to another (Avgeropoulos and Sammut-Bonnici, 2014; Matzler et al., 2015). While switching costs contain customers’ perception related to individual criteria alongside the monetary value, which can be objectively measured (Aydin and Özer, 2005), switching costs should not be limited to objective measurable monetary costs or “economic” costs, as they are not only economic in nature, but also involve the psychological, time and emotional effort incurred from the uncertainty of dealing with a new service provider (Chang and Chen, 2008; Dick and Basu, 1994; Edward and Sahadev, 2011; Jones et al., 2002; Yen, 2010). The relationship exchange includes an investment of unique attributes related to the effort, time and money and an obstacle if customers dissatisfied with one service provider want to switch to another (Burnham et al., 2003; Calvo-Porral and Lévy-Mangin, 2015; Chang and Chen, 2008; Jones et al., 2000, 2002; Yen et al., 2011).

Moreover, Klemperer (1987) identified three types of switching costs. Firstly, the learning costs incurred to switch to a new brand of the medical scheme after learning to use the current brand. Secondly, the transaction costs customers forfeit from switching their current service provider and acquiring a new service provider. This occurs, for instance, if two medical aid schemes offer identical medical benefits, but there are higher transaction costs in closing the current account and opening another with a competitor. Thirdly, contractual costs are the loss of repeat-purchase coupons and rewards customers have earned from their medical scheme and penalties for brand switchers. Therefore, switching cost dimensions are manifold, including lost performance, uncertainty, set-up costs and sunk costs (Jones et al., 2002).

Of special interest is the typology of switching costs proposed by Burnham et al. (2003), which groups switching costs into:

• procedural switching costs involving the loss of time and effort;

• financial switching costs involving the loss of financially quantifiable resources; and

• relational switching costs, such as psychological or emotional discomfort incurred due to the loss of identity and the breaking of bonds.

Jones et al. (2007) further classified switching costs into procedural costs, social costs (i.e. similar to relational switching costs) and lost benefit costs (i.e. similar to financial switching costs). According to Blut et al. (2014), procedural switching costs are classified as internal switching costs, while financial and relational switching costs are external switching costs. Research indicates the direct effect of procedural switching costs on customer retention, and a non-significant direct effect of financial switching costs on customer retention (Li et al., 2023). Chuah et al. (2018) noted that except for the social relationship loss, all antecedents of switching barriers (i.e. loss of material benefits, monetary switching costs and procedural switching costs) positively affect switching barriers, and procedural switching costs have the strongest effect on switching barriers.

Adopting the Burnham et al. (2003) classification of switching costs, Ha et al. (2023), Huang et al. (2021), Jones et al. (2007) and Shen and Ahmad (2022) examined “relational switching costs” as positive switching costs (foregone gains), “financial switching costs” as positive switching costs (foregone gains) and “procedural switching costs” as negative switching costs (actual losses). These scholars echoed this distinction as imperative due to the differential mediating roles of different types of commitment related to positive versus negative switching costs. They asserted that this classification is vital for industry experts, giving them more detailed guidance for applying switching costs to enhance customer loyalty.

Unlike Nagengast et al. (2014), who adopted and merged the two positive types of switching costs (i.e. financial and relational loss switching costs) as one construct, Moliner-Tena et al. (2018) suggested that it is interesting to categorise switching costs based on the underlying nature of the constraint: positive switching costs (social and lost benefits costs) generate rewards, such as benefits and value; while negative switching costs (procedural and economic costs) derive primarily from the negative sources of constraint (i.e. penalties). The positive switching barriers are regarded as “wanting to be” in a relationship (i.e. a positive reason to stay with a provider), whereas negative barriers can be described as “having to be” in a relationship (i.e. a negative reason to remain) (Jones et al., 2000, p. 269). Thus, negative switching costs are observed as opportunistic behaviour suppliers perform in a position of power (Matos et al., 2009). For instance, if consumers want to switch medical schemes, they must accept the negative consequences of cost expenditures, such as time and money and the loss of the long-term relationship (Lien et al., 2014). Lee and Kim (2022) found customers’ brand attachment (brand-self connection and brand prominence) to a hotel to be reinforced not only by positive switching barriers (relational benefits), but also by negative switching barriers.

While research confirms the effect of switching barriers on customers’ trust in the banking sector (Hidayat and Idrus, 2023), this study tested the effect of customer trust on switching costs typology classified by direction (i.e. positive and negative) (Burnham et al., 2003; Jones et al., 2007) in the medical scheme industry. Empirical evidence shows that the relationship between trust and perceived uncertainty cost is higher than the relationship between trust and other switching cost dimensions (Aydin and Özer, 2005).

Regarding the procedural switching costs, global examples of product tampering illustrate the high economic costs that result when the high level of trust based on fairness and intrinsic motivation is damaged between consumers and organisations (Choi et al., 2007). As per Moliner-Tena et al. (2018), trust is significantly affected by negative switching costs (i.e. procedural and economic switching costs) deriving mainly from negative sources of constraint (penalties for switching). These deductive attributes of freedom or penalties adversely influence customer trust. They reported that the relationship between trust and positive switching costs is stronger among older than younger consumers, and the relationship between negative switching costs and trust is weaker among older than younger consumers. They confirmed that during a financial crisis, suffering negative switching costs (penalties) causes mistrust, as younger consumers considered them not only as obstacles designed to retain them, but related this strategy with the bad banking practices that arose with the financial crisis.

From the transaction costs perspective, Teo and Yu (2005) stated that mistrust results from uncertainty and risk and tends to increase transaction costs in online purchasing behaviour. Results showed that customers’ trust in online stores is negatively related to transaction costs. Pinto et al. (2009), Teo and Yu (2005) and Yen et al. (2011) asserted that the role of trust in transaction cost theory is important because, to some extent, it reduces transaction costs. Customers who have formed trust in specific providers are less likely to switch due to the difficulties associated with establishing new trusting relationships in a context of uncertainty and risk (Carter et al., 2009). Moreover, the perceived evaluation costs (e.g. time and effort required to evaluate and select an appropriate service provider) are reduced based on trust issues (Bromiley and Harris, 2006; Huang et al., 2021). Masri et al. (2021) proved that product evaluation costs relate to trust in online vendors in Taiwan’s online shopping. Results illustrated that product evaluation costs have a positive relationship with trust towards an online vendor. In addition, trust eases the learning processes in service encounters as both parties are more open, thus enhancing perceived value (El-Manstrly, 2016). Trust increases perceived value by lowering perceived non-monetary costs, such as the time and effort needed to select a suitable service provider (Ponte et al., 2015).

Regarding the financial switching costs, Aydin and Özer (2005) indicated that the most significant information about monetary costs relates negatively to trust and customer satisfaction in the Turkish mobile phone business. They found that as users’ perceptions of trust or satisfaction increases, the perceived monetary cost of switching to a new operator decreases. In addition, Moliner-Tena et al. (2018) found that trust positively impacts positive switching costs (social and lost benefits costs), creating rewards like benefits and value. They reported that loyalty rewards that customers have to give up when ending their bonds with their service providers are positive switching costs (foregone gains) and customers enjoying these benefits may recommend these service providers to other customers.

In terms of relational loss switching costs, compulsive buyers appear to switch a lot among brands due to their variety-seeking nature and focus less on the functional benefits of brands, such as quality. Consequently, they lack the opportunity to build trust in brands (Horváth and Birgelen, 2015). Contrarily, consumers’ habitual purchases with an online hotel booking service provider build trust, increase the cost of brand relationship losses and prevent customers from switching (Huang et al., 2021). As switching costs assist firms in building personal relations (Burnham et al., 2003; Li et al., 2015), customer trust should focus on personal relationships that matter (Bromiley and Harris, 2006). El-Manstrly et al. (2011) and Maicas Lopez et al. (2006) reported that trust reduces the perception of risk and increases the feeling of wanting to remain in a relationship. It can be posited that trust promotes positive switching costs. A mutual sense is that relationship maintenance costs must be greatly reduced as customer trust increases (Pinto et al., 2009). El-Manstrly et al. (2011) confirmed that trust is correlated with relational costs. Therefore:

Trust is typically built through repeatedly keeping quality promises (Horváth and Birgelen, 2015). According to Morgan and Hunt (1994), trust is an indicator of consumers’ confidence in an exchange partner’s reliability and integrity, embodying customers’ beliefs of receiving a promised service. Trust is built when consumers believe that a company acts ethically, legally, favourably and responsibly (Pavlou and Fygenson, 2006). Within shared values, maintaining the highest level of ethics in all business transactions is the most significant issue (Mukherjee and Nath, 2007). Both ethical and practical considerations suggest that firms should seek to increase switching costs in ways that add value to consumers (Burnham et al., 2003). Colwell et al. (2011) showed the link between ethical code enforcement and continuance commitment but argued that enforcing ethical codes matters less when switching is perceived as too costly. Despite the significant advances in the research on trust, the concept of trust, as a referent in the ethical sales behaviour and switching costs typology has not been studied and is interesting to examine in medical schemes setting. For instance, this research tests the mediation effect of trust on the relationship between ethical sales behaviour and switching costs typology (Burnham et al., 2003), which is stronger when trust is higher (see Figure 1). The definition of trust operationalised in this study originates from commitment-trust theory of the relationship marketing proposed by Morgan and Hunt (1994), which views trust as an essential mediating variable in building successful relationships. Prior studies (Diallo and Lambey-Checchin, 2017; Mansouri et al., 2022) have found that customer trust mediates the impact of ethical sales behaviour on loyalty in the sportswear and retail industry. Masri et al. (2021) indicated that trust significantly mediates the relationships between monetary value, product evaluation, customer enjoyment and customer intention to purchase from an online vendor and reuse the product or service. Therefore:

The proportion of beneficiaries covered by medical schemes, expressed as a percentage of the population in the country, declined from 16.5% in 2000 to 15.08% in 2019. Restricted schemes grew by 64,172 beneficiaries and open schemes added only 9.293 beneficiaries between 2018 and 2019, which represents an insignificant growth of 0.82% in the year-on-year increase in the total number of beneficiaries covered by medical schemes. Nevertheless, open medical schemes covered the largest number of beneficiaries in 2019 (55.38%) (CMS, 2020). For the purpose of this study, a sample of 250 South African men and women aged 18–65 years was targeted from members of open and closed medical schemes living in the Gauteng province, which accounted for 40% of the total proportion of the medical scheme beneficiaries in the country (CMS, 2020). According to CMS (2020), more benefits are paid to beneficiaries in the age bands above 44 years, with beneficiaries aged 65 and older (8.91% of the population) consuming over 26% of health-care benefits and beneficiaries aged 45–64 years (23.4% of the medical schemes population) consuming 34% of health-care benefits. This translated to 33% of the medical schemes’ population consuming close to 60% of the health-care benefits provided in 2019, and the cross-subsidisation of benefits between the young and healthy and the older and sicker beneficiaries. However, this study did not survey respondents over the age of 65 due to various ethical factors that can result in adults or elderly people feeling vulnerable to abuse, neglect and ill intent from the topic being studied. The discussions prevented arousing feelings of neglect and abuse, as there was no instant prepared support for their decision to participate in the research and the study avoided tempering their independence and dignity.

The sample size of 250 medical scheme members was consistent with similar extant studies on this topic (Agrawal, 2017; Edward and Sahadev, 2011). Moreover, a sample size of 200 units has formerly been proposed as the minimum for SEM (Boomsma, 1982; Wolf et al., 2013) and many scholars have used a similar sample size to test the SEM (Hidayat and Idrus, 2023; Li et al., 2022; Morales-Sánchez et al., 2020), while some have used it to test ANOVA (Zhao et al., 2023). Even when applied to least squares estimation, a sample size should be 10 times larger than the number of variables impacting the conceptual constructs (Hair et al., 2011; Ion et al., 2021).

A non-probability judgement sampling method (Hair et al., 2017) was used to identify and select the main members who used medical schemes over the six-month period preceding the survey. The structured self-administered questionnaires, written in English, were distributed by well-trained fieldworkers face to face to respondents who were elicited in public locations near private and public health-care facilities. Research ethics (i.e. consent for participation in the study, voluntary participation, decision to terminate the participation, privacy and data safety) were adhered to in accordance with the ethics permit issued by the researchers’ university. The respondents were thanked for their participation in the survey, without rewards.

The structured self-administered questionnaires contained a cover letter written in English to brief the respondents about the study’s purpose. Screening questions filtered the sample. Section A collected the demographic data, including age, gender, monthly contributions, medical scheme brand affiliation, duration of affiliation and the number of dependent beneficiaries (see Table 1). Section B measured the latent constructs adapted from pre-validated scale items, as Churchill (1979) recommended. From the switching costs typology, procedural costs were measured with six scale items of economic risk costs, four scale items of evaluation costs, four scale items of learning costs and four scale items of set-up costs. Financial loss costs measured three scale items each of monetary loss costs and benefit loss costs. Relational loss costs measured the four scale items of personal relational loss costs and three scale items of brand relational loss costs, all adapted from Burnham et al. (2003). The ethical marketing issues were measured with nine scale items adapted from Murphy et al. (1992), and customer trust was measured with six scale items adapted from Morgan and Hunt (1994). These 46 scale items were modified to generate statements suitable to measure this study’s objectives and were recorded using a five-point Likert-type scale (1 = “strongly disagree”; 5 = “strongly agree”). Table 2 illustrates how each construct was operationalised in the study.

The questionnaire was pre-tested on three experts in the medical and academic fields to gather their subjective opinions before the main survey and improve the quality of the study. A pilot study comprised a sample (n = 50) of medical scheme members sharing similar attributes to those targeted in the initial survey. It tested the measurement errors (i.e. sentence and sequence of questions) and improved the final version of the questionnaire. Cronbach’s alpha ranged from 0.394 for evaluation costs to 0.923 for customer trust. In total, 208 completed questionnaires were returned in the primary survey, yielding an 83.2% response rate usable for data analysis. Due to a high percentage of missing data (n = 42), the study applied the data removal approach (e.g. remove rows or columns with a high proportion of missing values in the data set) as the primary method to handle and solve this error, rather than the imputation method, which substitutes reasonable guesses for missing data.

Table 1 outlines the descriptive statistics analysed using SPSS version 29. The final sample comprised 208 respondents, with more female (114; 54.8%) than male (94; 45.2%) respondents. Majority of these respondents (50; 24.0%) were aged 25–29, followed by 30–34 (43; 20.7%). Many respondents belonged to the Discovery Health Medical Scheme (115; 55.3%), followed by BankMed (42; 20.2%) and momentum (23; 11.1%). More respondents (84; 40.4%) contributed up to R3,000 monthly than those who contributed between R3,000 and R4,000 (74; 35.6%). A large number of respondents had exceeded five years’ service experience with their medical schemes (87; 41.8%), compared to those who only had one to two years of service experience (59; 28.4%). Respondents who were registered as main members (88; 42.3%) equalled those who registered two to three dependent beneficiaries (88; 42.3%), followed by those who registered four to five dependent beneficiaries (30; 14.4%).

According to MacKenzie and Podsakoff (2012), significant evidence suggests that method bias could extremely affect the reliability and validity of the scale items in a questionnaire and the covariation between latent constructs. Examples range from a series of causes, including manipulating the respondents’ capabilities, creating complex opportunities for more accurate responses, discouraging the motivation to respond accurately and making statements easier for respondents to answer without proper adherence. In minimising these issues, some procedural remedies were applied when developing, distributing and administrating the questionnaire. In addition, more sophisticated statistical remedies were applied to evaluate the extreme effects of common-method bias (CMB). During the survey design, various predictor variables were separated from the outcome variables to reduce respondents’ assumptions of their associations and anonymise them to hide their recognition deliberately. Respondents were asked to answer each statement honestly and from their own views, and were assured that there were no “correct”/“wrong” responses. For statistical evaluation, the Harman single-factor test was conducted using SPSS version 29.

The results indicated that a single factor accounts for 21.61% of the total variance, less than an acceptable threshold of 40% suggested by Babin et al. (2016). As Bagozzi et al. (1991) recommended, the inter-factor correlation matrix was also tested. The largest inter-factor correlation of 0.578 (between evaluation costs and economic risk costs; see Table 3) was below the 0.90 acceptable estimated threshold. Podsakoff et al. (2003) advised researchers to assess CMB by measuring the confirmatory factor analysis (CFA) as a more sophisticated assessment of the hypothesis that a single factor can account for all the variance among the items in the data. Tested on AMOS version 29, this diagnostic technique produced a one-factor model fit (χ² = 4815.252/df = 989; χ²/df = 4.869; comparative fit index [CFI] = 0.300; root mean square error of approximation [RMSEA] = 0.137) significantly weaker (Δχ² = 4142.651; df = 576, χ²/df = 3.24; CFI = 0.638; RMSEA = 0.082) than that of the initial hypothesised CFA model (χ² = 672.601/df = 413; χ²/df = 1.629; standardised root mean square residual [SRMR] = 0.0618; CFI = 0.938; RMSEA = 0.055). These findings strongly confirmed that CMB did not pose an extreme concern for this study.

The measurement model, or the initial hypothesised CFA model, was tested on AMOS version 29 using a maximum likelihood estimation to assess the covariance matrix and identify the estimates of various parameters (Anderson and Gerbing, 1988). Indicators of the composite reliability (CR) measured confirmed the construct reliability (Hair et al., 2017). The average variance extracted (AVE) measured confirmed the discriminant validity (Fornell and Larcker, 1981) of the interrelationships between the latent factors. The measurement model tested the possibility of outliers or negative error variances (known as Heywood cases) and the standardised factor loadings above 1.0 or below −1.0 (Hair et al., 2017). Table 2 illustrates the standardised regression weights of the CFA. All factor loadings > 0.70 showed the good internal consistency reliability of the scale items (Anderson and Gerbing, 1988; Hair et al., 2017). The factor loadings above 0.5 were retained, as they were viewed to be statistically significant (Hair et al., 2009). However, the few constructs with lower reliability scores (e.g. learning costs, set-up costs and brand relational loss costs) were deleted in the CFA1. In addition, Table 2 shows Cronbach’s alpha values ranging from 0.755 for ethical sales behaviour to 0.936 for customer trust. Cronbach’s alpha values > 0.7 and closer to 1 indicate a good internal consistency reliability of the construct measured (Hair et al., 2017).

The measurement model had good fit indices as advocated by Bentler and Bonett (1980): χ² = 672.601; df = 413; χ²/df = 1.629; goodness-of-fit index (GFI) = 0.831; adjusted GFI (AGFI) = 0.797; CFI = 0.938; Tucker–Lewis index (TLI) = 0.930; incremental fit index (IFI) = 0.938; non-normed fit index (NNFI) = 0.855; SRMR = 0.0618; and RMSEA = 0.055.

Table 3 shows the results of the discriminant validity tested using a method by Fornell and Larcker (1981). The square root of the √AVEs (bold values in Table 3) were significantly higher than the inter-correlation of coefficients, which show the discriminant validity (Henseler et al., 2015). The AVE values were more significant than the shared variance with other variables (> 0.50) (Fornell and Larcker, 1981) and the CR values were above (> 0.70 level) (Hair et al., 2017), strongly supporting the convergent validity of the measurement model tested in this study and allowing the testing of the structural model.

The latent variables’ causal structure was tested using SEM on AMOS version 29. The structural model tested the significance of the path estimates and overall model fit. The structural model had good fit indices as advised by Bentler and Bonett (1980): χ² = 802.893/df = 423; χ²/df = 1.898; GFI = 0.803; AGFI = 0.769; CFI = 0.909; TLI = 0.900; IFI = 0.96; NFI = 0.910; SRMR = 0.1129; and RMSEA = 0.066. Results showed the R² was 0.29 for trust, 0.10 for economic risk costs, 0.12 for evaluation costs, 0.10 for monetary loss costs, 0.08 for benefit loss costs and 0.22 for personal relational loss costs. Table 4 summarises the results of the hypotheses. The cut-off value of 1.96 (t >1.96) was used to determine whether the hypotheses were supported or rejected. Hypotheses with t-values smaller than 1.96 were rejected, while hypotheses with t-values exceeding 1.96 were accepted. The t-values for H1 was −5.042, which is lower than the cut-off value of 1.96. Consequently, H1 was rejected, proving ethical sales behaviour negatively affects customer trust in medical schemes in South Africa. The t-values for H2 were as follows: H2a (economic risk costs) = 1.239, H2b (evaluation costs) = 3.388, H2e (monetary loss costs) = 3.010, H2f (benefit loss costs) = 1.266 and H2g (personal relational loss costs) = 2.016. Consequently, H2b, H2e and H2g were supported, while H2a and H2f were rejected. These results proved ethical sales behaviour has a positive influence on evaluation costs, monetary loss costs and personal relational loss costs in medical schemes in South Africa. The t-values for the third hypothesis were: H3a (economic risk costs) = 3.949, H3b (evaluation costs) = 3.769, H3e (monetary loss costs) = 3.665, H3f (benefit loss costs) = 3.672 and H3g (personal relational loss costs) = 6.286. Consequently, H3a, H3b, H3e, H3f and H3g were accepted, showing trust has a positive influence on economic risk costs, evaluation costs, monetary loss costs, benefit loss costs and personal relational loss costs in medical schemes in South Africa.

The mediation analysis used the PROCESS procedure for SPSS release 2.041 (Hayes, 2013). The bootstrap samples for bias-corrected bootstrap confidence intervals was 1,000 with a 95% confidence level for all confidence intervals. Mediation analysis was conducted to find out how variables related to one another. A significant relationship between the independent, mediating and dependent variables (i.e. greater than t > 1.96) shows that trust mediates this relationship. However, if the t-values of the independent, mediating and dependent variables do not meet 1.96 levels, there is no mediation effect (mediation-free). Partially mediated effects were observed in this study if a mediating variable, namely trust, was significant between ethical sales behaviour and type of switching cost. Conversely, if there was no significant difference between ethical sales behaviour and type of switching cost when a mediating variable was present, complete mediation occurred. The mediation analysis results for the variables studied in this study can be found in Table 5, which concludes that three mediation hypotheses (i.e. H4b, H4e and H4g) partially mediated the effect of their independent and dependent variables.

Table 5 shows the significant positive direct and indirect effect of ethical sales behaviour of medical schemes on monetary costs (β = 0.2765; t = 2.7100; p < 0.0073; low-level confidence interval [LLCI] = 0.4776; upper-level confidence interval [ULCI] = 0.0150) [R² = 0.0150]. The bootstrap confirmed the indirect positive effect of ethical sales behaviour on monetary costs (BootLLCI = −0.1937; BootULCI = −0.0314), explaining 15.0% of the variance and thus accepting H4e. Furthermore, the results showed the significant positive direct and indirect effect of ethical sales behaviour of medical schemes on evaluation costs (β = 0.2624; t = 2.5159; p < 0.0126; LLCI = 0.4681; ULCI = 0.0146) [R² = 0.0146]. The bootstrap confirmed the indirect positive effect of ethical sales behaviour on evaluation costs (BootLLCI = 0.4681; BootULCI = 0.0146), explaining 14.6% of the variance, hence H4b was accepted. In addition, the results showed the significant positive direct and indirect effect of ethical sales behaviour of medical schemes on personal relational loss costs (β = 0.2566; t = 2.4677; p < 0.0144; LLCI = 0.4616; ULCI = 0.0013) [R² = 0.0013]. The bootstrap confirmed the indirect positive effect of ethical sales behaviour on personal relational loss costs (BootLLCI = −0.3436; BootULCI = −0.0788), explaining 01.3% of the variance and supporting H4g. Trust, directly and indirectly, mediated the positive relationships between ethical sales behaviour of medical schemes and monetary costs, evaluation costs and personal relational loss costs, thus showing partial mediation.

However, the relationship ethical sales behaviour of medical schemes had with economic risk costs (β = 0.1572; t = 1.7239; p > 0.0862) and benefit loss costs (β = 0.1216; t = 0.9691; p > 0.3337) produced t-values below 1.96 levels, and these were not statistically mediated by trust, thus rejecting H4a and H4f. Consequently, trust did not significantly mediate the impact of ethical sales behaviour of medical schemes on economic risk costs and benefit loss costs. Hypotheses that trust mediates the significant positive relationship that ethical sales behaviour has with learning costs (H4c), set-up costs (H4d) and brand relational loss costs (H4h) could not be measured, as these factors were deleted due to low-reliability scores in the CFA.

The moderating effects of medical schemes’ brand affiliation, monthly contributions and the number of dependent beneficiaries in the theoretical framework were tested on PROCESS procedure for SPSS release 2.041 (Hayes, 2013). The level of confidence for all confidence intervals in output was 95%. The results of the moderating effects and the conditional effects are shown in Table 6. Figures 2–8 show the directions of these conditional effects.

Table 6 shows a significant negative interaction effect of brand affiliation on the relationship between ethical sales behaviour and trust [R² = 0.1170] (β = −0.1306; t = −3.2026; p < 0.0016; LLCI = −0.2111; ULCI = −0.0502). Figure 2 shows that the moderation is more significant for members with high brand affiliation (β = −0.6394; t = −5.1115; p < 0.0000; LLCI = −0.8861; ULCI = −0.3928) than for those with low brand affiliation (β = −0.2475; t = −2.7194; p < 0.0071; LLCI = −0.4269; ULCI = −0.0680), explaining 11.70% of the variance. Thus, H5(1) was supported, showing that a high brand affiliation negatively affects the relationship between ethical sales behaviour and trust.

In addition, Table 6 shows a significant negative interaction effect of the brand affiliation on the relationship between trust and economic risk loss costs [R² = 0.0816] (β = −0.0635; t = −2.4024; p < 0.0172; LLCI = −0.1157; ULCI = −0.0114). Figure 3 shows that the moderation is strongly significant for members with low brand affiliation (β = 0.2991; t = 4.0446; p < 0.0001; LLCI = 0.1533; ULCI = 0.4450), but not for those with high brand affiliation (β = 0.1086; t = 1.2747; p > 0.2039; LLCI = −0.0594; ULCI = 0.2765), which explains 8.16% of the variance. Thus, H6a was accepted, showing that a low brand affiliation, but not a high brand affiliation, positively affects the relationship between trust and economic risk costs.

Furthermore, Table 6 shows a significant positive interaction effect of the monthly contributions on the relationship between trust and benefit loss costs [R² = 0.0788] (β = 0.1705; t = 1.9786; p < 0.0492; LLCI = 0.0006; ULCI = 0.3405). Figure 4 shows that the moderation is strongly significant for members with high monthly contributions (β = 0.5116; t = 4.0521; p < 0.0001; LLCI = 0.2627; ULCI = 0.7606), but not for those with low monthly contributions (β = 0.1706; t = 1.3305; p > 0.1848; LLCI = −0.0822; ULCI = 0.4233), which explains 7.88% of the variance. Thus, H7f was accepted, as it proves high monthly contributions, but not low monthly contributions, positively affect the relationship between trust and benefit loss costs.

In addition, Table 6 shows a significant negative interaction effect of the number of beneficiaries on the relationship between trust and economic risk loss costs [R² = 0.0970] (β = −0.2382; t = −2.9461; p < 0.0036; (LLCI = −0.3976; ULCI = −0.0788). As per Figure 5, the moderation for members with low number of beneficiaries is strong (β = 0.3945; t = 4.5011; p < 0.0000; LLIC = 0.2217; ULCI = 0.5673) compared to members with high number of beneficiaries (β = 0.1563; t = 2.1632; p < 0.0317; LLCI = 0.0138; ULCI = 0.2987), which explains 9.70% of the variance. Therefore, H8a was accepted, proving that having a low number of beneficiaries over a high number of beneficiaries positively impacts the relationship between trust and economic risk costs.

Furthermore, Table 6 shows a significant negative interaction effect of the number of beneficiaries on the relationship between trust and evaluation costs [R² = 0.0554] (β = −0.2228; t = −2.3806; p < 0.0182; LLCI = −0.4073; ULCI = −0.0383). Figure 6 illustrates a strongly significant moderation for members with low number of beneficiaries (β = 0.3401; t = 3.3528; p < 0.0010; LLCI = 0.1401; ULCI = 0.5401), but not for members with a high number of beneficiaries (β = 0.117; t = 1.4031; p > 0.1621; LLCI = −0.0475; ULCI = 0.2822), which explains 5.54% of the variance. Therefore, H8b was accepted, showing that having a low number of beneficiaries, but not a high number of beneficiaries, positively affects the trust and evaluation costs relationship.

Table 6 also shows a significant negative interaction effect of the number of beneficiaries on the relationship between trust and monetary loss costs [R² = 0.0694] (β = −0.2257; t = −2.4553; p < 0.0149; LLCI = −0.4069; ULCI = −0.0444). Figure 7 illustrates that the moderation is significant for members with a low number of beneficiaries (β = 0.3818; t = 3.8323; p < 0.0002; LLCI = 0.1854; ULCI = 0.5782), but not for members with a high number of beneficiaries (β = 0.1561; t = 1.9011; p > 0.0587; LLCI = −0.0058; ULCI = 0.3181), which explains 6.94% of the variance. Therefore, H8e was accepted, proving that having a low number of beneficiaries, but not a high number of beneficiaries, positively affects the relationship between trust and monetary loss costs.

Finally, Table 6 showcases a significant negative interaction effect of the number of beneficiaries on the relationship between trust and benefit loss costs [R² = 0.0844] (β = −0.2307; t = −2.0711; p < 0.0396; LLCI = −0.4504; ULCI = −0.0111). Figure 8 depicts the moderation as being strongly significant for members with a low number of beneficiaries (β = 0.5007; t = 4.1463; p < 0.0000; LLCI = 0.2626; ULCI = 0.7388), compared to members with a high number of beneficiaries (β = 0.2700; t = 2.7119; p < 0.0073; LLCI = 0.0737; ULCI = 0.4663), which explains 8.44% of the variance. Therefore, H8f was accepted, showing that having a low number of beneficiaries, over a high number of beneficiaries, positively impacts the relationship between trust and benefit loss costs.

This study assessed how ethical sales behaviour affects switching costs typology, mediated by trust and moderated by brand affiliation, monthly contributions and the number of dependent beneficiaries in medical schemes in South Africa. This research concerned the main members of South African medical schemes residing in Gauteng, which accounts for 40% of the largest proportion. This province recorded a 1.4% increase in medical scheme membership compared to other provinces, which declined (CMS, 2018). The respondents were affiliated with 13 schemes from the 15 (ten open and five restricted) schemes offering efficiency-discounted options as of 31 March 2020 (see Table 2), showing the representativeness of the medical schemes in this study.

The study’s findings provide evidence-based research linking ethical sales behaviour, customer trust and switching costs typology in the medical scheme sector. The results show that sellers’ ethical behaviour significantly and negatively affects customer trust and positively affects switching costs typology. Ethical sales behaviour also indirectly impacts switching costs typology, mediated by trust. Trust is essential to establishing long-term relationships between parties and continuing any exchange relationship (Morgan and Hunt, 1994). This study shows a negative relationship between the ethical sales behaviour and customer trust. Consistent with prior research (Chonko et al., 1996; Donoho et al., 2013; Schwepker and Schultz, 2013; Trawick et al., 1991), this study proves customers perceive the ethical behaviour of salespeople in the medical scheme sector negatively. A plausible explanation for this finding is that unethical sales practices (i.e. hard-selling tactics) negatively affect trust and practitioners of medical schemes must prohibit them. The extent to which salespeople are sales-orientated negatively influences the relationship quality, as the “hard sell” and pressure associated with such an approach could result in consumer suspicion and thus inhibit the relationship development (Bejou et al., 1998, p. 171).

This result contradicts prior research showing a positive impact of ethical sales behaviour on customer trust in other industries (Mansouri et al., 2022; Ou et al., 2015; Park et al., 2017; Singh et al., 2012; Wijaya et al., 2022). While a positive ethical climate may be desirable and positively influence trust in management, translating positive core values of salespeople into practical action is a complex process in ethical matters (Agrawal, 2017). Sellers must put transparency and honesty at the forefront of their work to improve the factors prominent to customer trust (Mansouri et al., 2022). As deceptive sales practices may be effective on a short-term basis (Santana et al., 2020), establishing customers’ trust necessitates respectful behaviour, tolerance and asking the right questions that best serve their needs. Sellers must avoid using false or unacceptably misleading sales presentations (Murphy et al., 1992), but provide customers with the best product and service advice, as proper guidance is linked to a higher level of trust (Mansouri et al., 2022; Wijaya et al., 2022).

Surprisingly, some constructs of switching costs typology (e.g. learning costs, set-up costs and brand relational loss costs) had low reliability and were deleted in the CFA1. From the origin of the scale, Burnham et al. (2003) reported that three of the four scale items of set-up costs had factor loadings of 0.6, and four scale items of learning costs were strongly correlated to set-up costs. One of the three scale items from brand relationship costs had factor loadings of 0.5. Loadings of lower-order constructs on higher-order types showed that brand relationship loss costs had a strong standardised regression value (0.87), but was weak compared to personal relationship loss costs (0.89). These scale inspections were relevant and led to these factors being suspicious.

Firstly, when evaluating the procedural switching costs, the effect of ethical sales behaviour on economic risk costs was not statistically significant. It is surprising that ethical sales behaviour does not have a significant positive effect on economic risk costs. A plausible explanation for this finding is that ethical sales behaviour does not predict the economic risk loss of switching medical schemes (e.g. accepting financial, performance and convenience risks when switching providers) (Carter et al., 2009). This view complements the view of Furrow (1988) that health-care costs are treated as primarily an economic problem when they are not. To be solved, they have to be treated as an ethical problem. Instead, ethical sales behaviour had a positive significant effect on evaluation costs. It shows that perceptions of evaluation costs derived primarily from the negative sources of constraint (i.e. evaluating the penalties for switching providers) are positively determined by ethical sales behaviour. These non-financial costs may be an ethical concern (Perepelkin and Wilson, 2018). In support of the literature (Keaveney, 1995; Keaveney and Parthasarathy, 2001; Lappeman et al., 2022; Liang et al., 2013), this study shows that the ethical problem is one of the causes of customer service switching, which has key implications for service marketing and service innovation research. Switching costs may limit the extent to which consumers acknowledge providers’ enforcement of ethical codes of conduct (Colwell et al., 2011).

Secondly, when evaluating the financial switching costs, the results proved ethical sales behaviour had a significant positive effect on monetary loss costs. It shows that the perceptions of monetary loss costs (i.e. the loss of financial quantifiable resources resulting from monetary costs), which are positive switching costs (foregone gains), are positively determined by ethical sales behaviour. Based on the confirmation of the association between ethical sales behaviour and monetary loss costs, medical schemes should admit that there are real, if not often quantifiable, costs and rewards associated with ethical sales behaviour in the market (Story and Hess, 2010). However, ethical sales behaviour had an insignificant effect on benefit loss costs. It shows that the benefit loss costs (i.e. positive switching costs) that generate rewards and value customers lose when switching medical scheme providers after accumulating points are not determined by ethical sales behaviour. A possible reason for this finding is that losing health-care plan benefits, premiums and discounted services, is not linked to ethical sales behaviour in medical schemes.

Thirdly, when evaluating the relational loss switching costs, ethical sales behaviour of medical schemes significantly and positively affected personal relational loss costs. This result agrees with prior research suggesting that customers experience a loss of comfort when breaking bonds with familiar service providers (Carter et al., 2009). A plausible explanation for this result is that customers experience a loss of psychological and emotional discomfort when breaking bonds with familiar medical scheme providers due to the ethical sales behaviour.

Fourthly, as for the role of customer trust in relation to procedural switching costs, the results showing a positive significant effect of trust in medical schemes on economic risk costs and evaluation costs were not expected, as in the switching costs typology, the “procedural” costs are negative penalties and actual losses customers incur for switching services providers (Burnham et al., 2003; Jones et al., 2007). A plausible explanation for this result is that the higher economic costs result when the high level of trust based on intrinsic motivation and fairness diminishes between consumers and organisations (Choi et al., 2007). This result reverses the study by Moliner-Tena et al. (2018), showing that trust is significantly influenced by negative switching costs (i.e. procedural and economic switching costs), deriving mainly from negative sources of constraint (penalties). They reported that reductive elements of freedom or penalties adversely affect consumer trust. Nonetheless, these results are consistent with prior research findings (Pinto et al., 2009; Teo and Yu, 2005; Yen et al., 2011), showing that the role of trust in transaction cost theory is necessary because, to some extent, it reduces transaction costs. In the transaction cost economics perspective, Teo and Yu (2005) stated that mistrust results from uncertainty and risk and tends to increase transaction costs in online purchasing behaviour, showing that trust in online stores is negatively related to transaction costs.

Furthermore, the results showed that customer trust in medical schemes significantly and positively affects the evaluation costs incurred in searching for and evaluating information about service providers and the time and effort needed in acquiring the skills required to use new service providers. This can be explained by the fact that the extent to which consumers trust medical schemes predicts their decision to switch due to the searched information about and evaluated health-care benefit options (Abraham et al., 2006). This result supports a study by Masri et al. (2021), showing that product evaluation costs in online shopping positively relate to trust in an e-vendor in Taiwan.

Fifthly, when evaluating the financial switching costs, customer trust in medical schemes was positively related to monetary loss costs (i.e. one-time financial outlays customers incur for switching providers other than those used to purchase the new product) (Burnham et al., 2003; Carter et al., 2009). This result differs from Aydin and Özer (2005), who reported that monetary costs relate negatively to the trust and satisfaction of customers in the Turkish mobile phone industry. Concerning their finding, as customers’ trust or satisfaction increases, the perceived monetary cost of switching to a new operator decreases. Alternatively, customer trust in medical schemes decreases due to higher acceptance of the one-time financial outlays incurred when switching service providers.

Moreover, customer trust in medical schemes significantly and positively relates to benefit loss costs, which occur when customers who switch to other providers lose accumulated points, benefits or discounts (Burnham et al., 2003; Carter et al., 2009). This result deviates from Koponen and Julkunen (2022), who showed that relational benefits relate to customer trust, commitment, social bonding, conflict management and communication. Nonetheless, this finding is consistent with a study by Nilssen (1992), which showed the effect of customer trust on feelings of losing accumulated points, benefits or discounts by switching and the total discounted future benefits from being disloyal, with only transaction costs incurred at a switch rather than loyal and with net switching costs. Moliner-Tena et al. (2018) confirmed that trust positively affects positive switching costs (i.e. lost benefits costs), such as loyalty rewards.

Finally, when evaluating the relational loss switching costs, customer trust in medical schemes was significantly and positively related to personal relational loss costs, such as the loss of comfort experience due to breaking relationships with familiar service providers (Carter et al., 2009). Huifeng and Ha (2020) concurred, showing that higher switching costs facilitate relationship termination relatively if time and personal loss are incurred. This could be because as trust increases, the costs related to maintaining the relationship with a partner should reduce significantly (Pinto et al., 2009). Hence, trust should focus on personal relationships in which it matters (Bromiley and Harris, 2006). Customers switch service providers when after-sales interactions are not provided, resulting in a lack of relationships (Kanchanapoom and Chongwatpol, 2021).

In further answering the research questions, the study offers empirical evidence of the direct and indirect mediating effects of customer trust in the theoretical framework of this study. The findings show that the significant relationships between ethical sales behaviour and switching costs typology (e.g. monetary loss costs, evaluation costs and personal relational loss costs) are mediated by trust. These results provide more evidence supporting the findings in earlier studies regarding the mediating effect of customer trust, either totally or in part (Diallo and Lambey-Checchin, 2017; Mansouri et al., 2022), which found that customer trust mediates the influence of ethical sales behaviour and loyalty in the retail industry and sportswear brands. In addition, Masri et al. (2021) showed that customer trust significantly mediated the relationships between monetary value, product evaluation, customer enjoyment and customer intention to buy from an online vendor and reuse the product or service. These results support relationship marketing theory, which views trust as a key mediating factor in relational exchanges (Garbarino and Johnson, 1999; Han and Ryu, 2012; Morgan and Hunt, 1994). Conversely, although trust mediates the relationships between ethical sales behaviour and switching costs, there is no valid evidence about the mediating effect of trust on the relationships between ethical sales behaviour and economic risk/benefit loss costs. Lin et al. (2021) indicated that green trust mediates the effect of the comparative economic value on the disruptive green product switching intention. A possible reason for these results is that customer trust in a medical scheme reinforces the significant positive influences of ethical sales behaviour on monetary costs, evaluation costs and personal relational loss costs, as opposed to economic risk costs and benefit loss costs.

In addition, the study’s findings underscore the differential moderating effects of the number of beneficiaries, monthly contributions and brand affiliation on the interrelationships between ethical sales behaviour, trust and switching costs typology. Customer trust and ethical sales behaviour continue to be key determinants of perceived switching costs and their effects are strengthened or weakened based on the observations of specific directions of the number of beneficiaries, monthly contributions and brand affiliation. Firstly, the number of beneficiaries plays a more specific moderating role (moderates the trust-economic risk costs, trust-evaluation costs, trust-benefit loss costs and trust-monetary loss costs relationships). Secondly, monthly contributions play a more specific moderating role (moderates the relationship between trust and benefit loss costs). Thirdly, brand affiliation plays a more general moderating role (moderates the ethical sales behaviour-trust and trust-economic risk costs relationships).

Regarding brand affiliation, the results indicate that brand affiliation negatively moderates the relationship between ethical sales behaviour and customer trust. The moderation is more significant for members with high brand affiliation than for members with low brand affiliation. Therefore, the relationship between ethical sales behaviour and trust is stronger for customers with high brand affiliation than low brand affiliation. Moreover, the results show that brand affiliation negatively moderates the relationship between trust and economic risk costs. The moderation is significant for members with low brand affiliation and not high brand affiliation. Thus, the relationship between trust and economic risk costs is significant for members with low brand affiliation, but is not significant for those with high brand affiliation.

In terms of monthly contributions, the results show that monthly contributions negatively moderate the relationship between trust and benefit loss costs. The moderation is significant for members with high monthly contributions, but not significant for those with low monthly contributions. Thus, the relationship between trust and benefit loss costs is significant when members have high monthly contributions, but not significant for low monthly contributions.

Concerning the number of beneficiaries, the show that the number of beneficiaries negatively moderates the relationship between trust and economic risk costs. The moderation is strongly significant for members with a low number of beneficiaries compared to a high number of beneficiaries. Therefore, the relationship between trust and economic risk costs is stronger for customers with a low number of beneficiaries than a high number of beneficiaries. Furthermore, the findings indicate that the number of beneficiaries negatively moderates the relationship between trust and evaluation costs. The moderation is significant for members with a low number of beneficiaries, but is not significant for those with a high number of beneficiaries. Therefore, the relationship between trust and evaluation costs is significant for customers with a low number of beneficiaries, but not significant for those with a high number of beneficiaries. Moreover, the results prove that the number of beneficiaries negatively moderates the relationship between trust and monetary loss costs. The moderation is significant for members with a low number of beneficiaries, but not for members with a high number of beneficiaries. In other words, the relationship between trust and monetary loss costs is significant when members have a low number of beneficiaries, but is not significant for members with a high number of beneficiaries. The findings also show that the number of beneficiaries negatively moderates the relationship between trust and benefit loss costs. The moderation is more significant for members with a low number of beneficiaries than for members with a high number of beneficiaries. Hence, the relationship between trust and benefit loss costs is stronger when members have a low number of beneficiaries.

Finally, none of the moderating variables (brand affiliation, monthly contributions and the number of beneficiaries) significantly affect the relationships between ethical sales behaviour and personal relational loss costs and trust and personal relational loss costs. A possible reason for these results is that number of beneficiaries, monthly contributions and brand affiliation do not strengthen the influence of ethical sales behaviour and trust on personal relational loss costs relationships.

There are some significant techniques in which this paper extensively expands the theoretical understanding of switching costs typology. The study undertakes the view that the relationships between ethical sales behaviour, trust and switching costs are complex and enlarge some additional components to the body of prior research in the area. The switching costs typology in this study was adapted from the work by Burnham et al. (2003), which classified switching costs into three distinct dimensions: procedural, financial and relational switching costs. This diverges from research that measured switching costs as a unidimensional construct of some diverse elements for approving its connection with other constructs. This study advocates that a more complex perspective of classifying switching costs by type and direction provides a more sophisticated understanding of the role of switching cost frameworks. The findings empirically confirm the importance of differentiating between positive and negative switching costs. Moreover, the study specifies that a unidimensional assessment may be too simplistic, thereby potentially confusing imperative theoretical and managerial implications required in the industry.

Firstly, this study redirects the academic interest away from a focus on probing the interrelationships between ethical sales behaviours, trust and switching costs towards understanding the boundary conditions of these interrelationships. Despite the fact that prior studies in numerous areas, including hairdresser service providers, the banking industry and digital scholarship settings, have confirmed the favourable relationships between ethics, trust and switching costs (El-Manstrly et al., 2011; Kaur et al., 2012; Lappeman et al., 2022; Moliner-Tena et al., 2018; Mutula, 2011), very limited research exists in the health-care context, particularly in medical schemes. Therefore, the current understanding of switching costs and their specific determinants in medical scheme health care is extended and improved by the results of this study. The study’s findings with regard to medical schemes in health care have similar results to extant research in different settings (Carter et al., 2009; Chonko et al., 1996; Donoho et al., 2013; Koponen and Julkunen, 2022; Masri et al., 2021; Moliner-Tena et al., 2018; Schwepker and Schultz, 2013; Trawick et al., 1991), which enhances the scope of marketing literature in the several subsequent approaches: ethical sales behaviour and trust are negatively related; ethical sales behaviour and switching costs are positively related; trust and switching costs are positively related; trust significantly mediates the relationship between ethical sales behaviour and switching costs; and the interactions of these variables are moderated by the number of beneficiaries, monthly contributions and brand affiliation. Consequently, the study expands knowledge on the applications of ethical sales behaviour in relation to trust and switching costs typology in health-care marketing literature, which is an exclusive contribution.

Secondly, trust is examined in its relationship with switching costs (Koponen and Julkunen, 2022; Masri et al., 2021; Moliner-Tena et al., 2018). Limited studies emphasise the different types of switching costs in that relationship. In this research, switching costs are conceptualised in three forms, including procedural, financial and relational switching costs in medical schemes in the health-care sector. The three conceptualised types of switching costs are very important, as attested by Ha et al. (2023), Huang et al. (2021), Jones et al. (2007) and Shen and Ahmad (2022), who stated that this distinction is necessary due to the differential mediating roles of different types of commitment related to positive versus negative switching costs. These scholars agree that this classification is crucial for industry experts, giving them more detailed guidance for applying switching costs to enhance customer loyalty. As such, this classification enriches theoretical gaps by linking ethical sales behaviour, trust and procedural, financial and relational switching costs in medical schemes in the health-care sector. Probing a three-dimensional concept compared to a unified concept, the conceptualised switching costs could provide a more nuanced understanding of the relationships between switching costs, ethical sales behaviour and trust. Thus, this study contributes to the systematic evaluation of current literature by highlighting the different switching costs as an appropriate and effective solution for firms to create a barrier in ways that add value and prevent customers from switching to other firms. The results reveal that, compared to the other two types of switching costs, procedural costs (e.g. evaluation costs) are largely impacted by ethical sales behaviour, while relationship loss costs (e.g. personal relational loss costs) are largely impacted by customer trust in the medical schemes context. Chuah et al. (2018) showed that procedural switching costs had the strongest effect on switching barriers among the mobile data service providers.

Thirdly, although earlier studies treated trust as a mediator in the relationship between ethical sales behaviour and loyalty (Diallo and Lambey-Checchin, 2017; Mansouri et al., 2022; Masri et al., 2021), no extant research has measured the mediating role of trust on the ethical sales behaviour-switching costs typology relationship. Therefore, this study contributes a unique perspective or idea that extensively advances the understanding of the mediating role of trust in the linkages between ethical sales behaviour and the three dimensions of switching costs (i.e. procedural, financial and relational) in the medical schemes. Ha et al. (2023) confirmed the three-dimensional switching costs (procedural, financial and relational) as mediators of the customer satisfaction-loyalty and service value-customer loyalty relationships in the private health-care sector. Consequently, this paper is the first to study trust as a mediator in the relationships between ethical sales behaviour and the three-dimensional switching costs (procedural, financial and relational switching costs) in the medical scheme sector. This valued conceptualisation enriches relationship marketing theory, which considers trust a key mediating factor in relational exchanges (Garbarino and Johnson, 1999; Han and Ryu, 2012; Morgan and Hunt, 1994). The empirical results in this study significantly validate the mediating role of trust in the effect of ethical sales behaviour on procedural switching costs (e.g. evaluation costs), financial switching costs (e.g. monetary loss costs) and relational switching costs (e.g. personal relational loss costs) in the medical schemes industry.

Moreover, ethical sales behaviour, trust and switching costs are three of the most recognised elements in relationship marketing literature, and this research is one of the few, if not the first empirical study to unify them for the purpose of not only confirming their interlinks, but also understanding the boundary conditions of these interrelationships in the context of medical schemes in an emerging market like South Africa. In a study testing the effect of ethical code enforcement on inter-organisational relationships, Colwell et al. (2011) reported that higher switching costs weakened the relationship between supplier-enforced ethical codes of conduct and continuance commitment. Instead, the current study confirms that higher customer trust strengthens the relationships between ethical sales behaviour and personal relational loss costs, evaluation costs and monetary loss costs in the medical scheme sector.

However, this study’s findings also support certain well-accepted conclusions from earlier research in the relationship marketing area (Carter et al., 2009; Chonko et al., 1996; Donoho et al., 2013; Koponen and Julkunen, 2022; Masri et al., 2021; Moliner-Tena et al., 2018). This is proven through ethical sales behaviour being an antecedent of trust and switching costs; ethical sales behaviour and trust being antecedents of switching costs; trust partially mediating the relationships between ethical sales behaviour and switching costs typology comprising procedural, financial and relational switching costs.

Finally, the inclusion of moderating variables in the theoretical framework provides a more detailed evaluation of how two types of switching costs, procedural and financial switching costs, are not only influenced by trust and ethical sales behaviour, but also by medical scheme industry-related services, such as brand affiliation, monthly contributions and the number of beneficiaries members have registered on their health-care plans. This sophisticated consideration is necessary to explain, for instance, why specific segments of customers perceive ethical sales behaviour and trust specific to medical scheme service providers with specific types of switching costs. Thus, this idea to test the simultaneous moderating effects of these variables on the interrelationships between ethical sales behaviour, trust and switching costs classified by type and direction is likely to provide not only a complete understanding, but also add to current inconclusive results. This study is unique in showing a simultaneous understanding of three different types and two distinct directions of switching costs in the medical scheme industry.

In addition to its scholarly contributions, this study has numerous important managerial implications. The results suggest that managers can revise and refine the ethical sales behaviours, trust and switching costs strategies, but the effectiveness of these strategies differs. Considering the specific types of perceived switching costs, managers may need to first establish trust followed by ethical sales behaviour as the most effective switching costs strategy in the medical scheme sector. Managers must invest in ethical sales behaviour practices as their sales tactics seem to negatively affect customer trust. For example, practitioners should avoid unethical sales behaviour, such as false advertising, price conspiracies, pressure selling and discriminatory pricing, as these practices are within the public view and subject to an excessive societal scrutiny (Du Plessis et al., 2010).

Likewise, in building customer trust, service managers with high employee contact can offer customised services needed to attract attention in enhancing relational (personal relational loss costs), procedural (economic risk costs and evaluation costs) and financial (monetary costs and benefit loss costs) as perceptions of these switching costs are impacted by consumer trust, and also restructure the procedural costs (evaluation costs), financial costs (monetary loss costs) and relational costs (personal relational loss costs), which seem to be affected by ethical sales behaviour. The conventional wisdom among many service managers is that the higher the switching costs are, the higher customer loyalty and profitability will be. Ironically, both ethical and practical considerations require managers to plan on building trust as a successful strategy to increase personal relational loss costs, economic risk costs, evaluation costs, monetary costs and benefit loss costs in ways that add value (Burnham et al., 2003). Managers should devote resources and competencies towards dynamic capabilities of providing service efficiency, well-explained services, discounted services, competitive prices and easy-to-use services to encourage their customers to trust their service providers. This is relevant considering that increases in switching costs by the affective dimensions are a form of avoiding customers’ negative word-of-mouth generated in a lock-in situation (Jones et al., 2007). Customers’ desire to avoid losing the financial (monetary and benefits loss costs) constraints is less likely to be affected by their inclination to trust service providers.

Furthermore, managers in medical schemes should concentrate on building positive ethical sales behaviour to increase procedural (evaluation costs), financial (monetary costs) and relational (personal relational loss costs) switching costs in ways that add value to customers. Proficient managers can operate this by offering co-created value, convenience and detailed-information services to stimulate customers’ perceptions of procedural switching costs (i.e. the time and effort in evaluating information about the health-care services and options). They may activate positive ethical sales behaviour by offering customers the monetary added value (i.e. economic savings, discounts and rewards for staying loyal). It is also necessary to offer sales training to service personnel with a focus on improving ethical sales tactics as a key to secure the employees–customer relationships (i.e. personal relational loss costs). Apparently, customers’ desire to avoid losing personnel relational bonds is less likely to be influenced by ethical behaviour of salespeople in medical schemes. Surprisingly, customers’ desire to avoid losing the non-monetary benefits (i.e. types of special treatment and customised service, and cognitive and emotional benefits not only aimed at attracting or enhancing the perceived intrinsic value, privileges and confidence benefits, but also removing customers’ anxiety in the service exchange) is not triggered by ethical sales behaviour.

In addition, practitioners who can improve ethical sales behaviour practices will be in a better position to generate customer trust as a key mediator influencing personal relational loss costs, monetary loss costs and evaluation costs in the medical schemes. This practice is in sync with relationship marketing theory, which considers trust a key mediating factor in relational exchanges (Garbarino and Johnson, 1999; Han and Ryu, 2012; Morgan and Hunt, 1994). Resultantly, generating higher levels of trust could help strengthen ethical sales behaviour in building long-term personal relationships with customers, openly discussing with them the most suitable health-care plans (discounted offers, premiums) and enabling customers to evaluate information that the salespeople present to them. Ethically and legally, a fiduciary relationship must be established based on trust (British Medical Association, 2017). As such, salespeople’s relational selling behaviour (e.g. staying in touch with clients; personalising the relationship by opening hearts to clients and sending cards and gifts; and demonstrating a cooperative, responsive service attitude) improves the relationship quality (i.e. clients’ trust in, and satisfaction, with the salespeople) in the life insurance service sector (Crosby and Stephens, 1987). Due to research centring on companies’ relationships with the practice of gift-giving to physicians by medical representatives, Marmat et al. (2020) noted that it appears that medical companies’ ethical or unethical behaviour is limited to these factors. Therefore, as ethics is a set of principles people use to decide what is right or wrong (Ferrell, 2004; Handa et al., 2014), medical scheme practitioners could establish trust in society and affect quality of life by avoiding misleading sales presentations or offering customers gifts in exchange for preferential treatment, or presenting wrong information about competitor firms, or accepting customer favours in exchange for preferential treatment (Murphy et al., 1992). Managers must ensure the value of such gifts rely on a higher level of trust between receivers and givers, as the actual economic or business value of the gift may be intangible or negligible (Choi et al., 2007).

Moreover, if the organisational climate conveys a high level of trust, employees are likely to reciprocate with high levels of trust in management (Agrawal, 2017). Practitioners need to be aware that an unethical environment and an organisational culture that does not value ethics can steadily erode the trust organisational members have (Agrawal, 2017; Mechanic, 1996). This study supports the idea that unethical sales behaviour has a significant negative effect on customer trust (Bejou et al., 1998; Chonko et al., 1996; Donoho et al., 2013). Hence, it is vital to abide by the American Marketing Association’s code of ethics regarding promotions (marketing communication), stated as follows: avoid deception in communications about products and services offered; avoid misleading and false advertising; discard the misleading, manipulative and high-pressure sales tactics; and avoid deception or manipulation in sales promotions (Du Plessis et al., 2010). It means that practitioners must avoid corporate practices that rob customers of self-esteem or justice, as these may not only be illegal, but can destroy trust and, consequently, the potential for building relationships (Berry, 1995). In this view, relationship marketers must be prepared to subject every policy and strategy to a fairness test. They must be willing to ask: “Is it legal?” and “Is it right?” (Berry, 1995, p. 243; Berry and Parasuraman, 1991, p. 145), and search for the meaning of “what is good” in a medical aid context or an answer to the question “What should I do?”, which defines ethics (Ion et al., 2021). As businesses lack a clear understanding of their consumers’ ethical beliefs (Brunk, 2010), the key is that all health facilities, services and goods must be sensitive to culture and meet the acceptable ethical standards, and health-care providers must acquire the skills needed (Competition Commission South Africa, 2018).

Trusting relationships are built over time as consumers gain evidence that their trust is well-placed (Fritz and Holton, 2019). In practice, practitioners must behave ethically and provide realistic expectations about the core medical scheme services (e.g. not exaggerate the gains to be made from an investment fund) and should be less likely to push customers into buying services they do not need (Román, 2003). Instead, they should strive to build perceived benevolence and honesty, and be competent enough to engage in proper behaviour about the relationship, ensuring customers rely on the service provided and show commitment towards the relationship (Morgan and Hunt, 1994). Salespeople’s ethical behaviour is imperative to establishing and maintaining relationship quality (Ou et al., 2015). Therefore, marketers should understand the ethical mindsets of customers, suppliers, business partners, distributors and staff (Al-Khatib et al., 2005). For instance, the ethical sales behaviour of contact employees (e.g. salespeople) can be used as a differentiating tool to achieve higher revenue growth, improved market share and competitive advantage (Mulki and Jaramillo, 2011; Román, 2003).

The first companies in an industry or first movers can gain advantages, such as enjoying brand and customer loyalty, combined with switching costs. This is helpful because once customers become used to products or producers, they show a (natural) reluctance to switch (Competition Commission South Africa, 2018). Marketplace reality shows that marketing managers inflict switching costs on customers to inhibit them from defecting to new suppliers (Yang and Peterson, 2004). If a firm can develop switching costs at the same time that it differentiates, it will boost its ability to offer superior performance and build a sustainable competitive advantage (Porter, 1980). Advantages of switching costs to a firm include customers’ satisfaction and lower price sensitivity and customers perceive functionally homogeneous brands as differentiated heterogeneous brands (Aydin and Özer, 2005). The strategic implications of switching costs are that they create imperfections in competitive markets by raising the barriers of change in terms of consumers’ ability to switch suppliers and in terms of the rate of adoption of new technologies (Avgeropoulos and Sammut-Bonnici, 2014).

In a globalised economy, how ethics is lived in businesses based on these countries is essential for many successful practitioners, policymakers and academicians (Abraham et al., 2006). However, in South Africa, academics have shown little leadership in driving evidence-based best practices in the private health-care sector (Competition Commission South Africa, 2018). Consequently, this research study alerts academics in marketing ethics to develop successful strategies aimed at this vulnerable social market, and to teach or discuss them with students and colleagues to raise awareness of the quality of the products they offer, the level of fairness in pricing products and the most viable distribution channels needed in such an emerging market (Baker and Saren, 2010). Therefore, managers in medical schemes should establish trust by refining the perception of ethical sales behaviour, especially among members with high brand affiliation. Increasing economic risk costs and benefit loss costs rely on the trust of members with a low number of dependent beneficiaries. Increasing benefit loss costs rely on the trust of members with high monthly contributions.

This study offers valuable knowledge to the professional practice of health-care managers, especially in the South African medical scheme sector. Offering special treatment benefits has become part of relationship marketing programmes due to the expectation of positive financial returns. As noted earlier, managers may operate this ethically by increasing switching costs and offering added value, such as types of special treatment, economic savings, customised service and cognitive and emotional benefits (Hennig-Thurau et al., 2002). The net effect of switching efforts relies on the strength of switching costs relative to the equivalent benefits (Yang and Peterson, 2004). Customers primarily evaluate “the greatest good for the greatest number” by performing a social cost-benefit analysis, and all possible benefits and costs of the assessed act are listed and summarised as the net of all benefits minus all costs. If the net result is positive, the act is morally acceptable; if the net result is negative, the act is not acceptable (Takala and Uusitalo, 1996, p. 52).

According to Mukherjee and McGinnis (2007), any health-care reform strategy aims to reduce costs. They noted that in an efficient system, costs are reduced while patient care is improved and administrative burdens are reduced. This is relevant in the medical scheme market, which is characterised by high and increasing costs of health care and medical scheme cover, highly concentrated funders’ and facilities’ markets, disempowered and uninformed buyers, absent value-based purchases, ineffective limits on rising volumes of care, practitioners who are subjected to very little regulation and failures of accountability at various levels (Competition Commission South Africa, 2018). For example, formally launched on 1 October 2019 by South African President Cyril Ramaphosa, the Health Sector Anti-Corruption Forum pointed that in the health sector, both public and private health-care markets are vulnerable to fraud and corruption due to large and varied numbers of transactions on goods and services, poor governance, bribery, over-pricing, fraudulent orders, tender irregularities, fiscal dumping by government departments through non-governmental organisations, transfer of liabilities to the state and bogus and fraudulent qualifications (CMS, 2020). The CMS is worried about the growing market concentration of administrators and managed care providers that has occurred in the past five years. The concerns centre on market failures that could result in market dominance and anti-competitive behaviour by these entities, which may be at odds with beneficiary interests and welfare. Again, at the provincial level, there are no direct positive associations between patient loads or density ratios with the use of general practitioners in health-care services (CMS, 2020).

For policy direction, the Joint Commission, committed to improving quality and safety in health-care organisations, has extensively offered an ethics framework entailing “marketing and advertising plans” (Schenker et al., 2014, p. 37). The ethical sales behaviour of salespeople in the medical scheme industry could play a key role in eliminating unethical practices, build customer trust and minimise switching behaviour. This approach contributes insights to standardising options and consolidating medical schemes in South Africa, which offer stable planning for health-care reform strategy, such as a full legislature of the NHI by 2025 (CMS, 2018).

Finally, the effectiveness of communication in building trust can improve organisational structure, especially in organisations typically serving customers through different salespeople at each service contact (Berry, 1995). Communication must lead to trust and trust builds the relationship commitment (Morgan and Hunt, 1994). Consequently, good communication increases the chance that patients will disclose intimate information and stigmatised conditions, and further improve cooperation in treatment and adherence to medical advice, and become open to suggestions about adopting health-promoting behaviour, which are essential goals to the emerging health-care agenda (Mechanic, 1996). In leveraging this, managers can invest in loyalty programmes offering special rewards to members for frequent medical visits or provide discounted visits to key health-care services, dental services and optometrist services and offer wellness programmes that enhance a direct medical benefit, such as free medical screening, HIV programmes and counselling to main members and beneficiaries with a lengthy relationship. For value co-creation, the wellness programme may be combined with the loyalty programme (Competition Commission South Africa, 2018).

Literature is insufficient in addressing the need for the theoretical framework that explains the role of ethical sales behaviour in affecting the switching costs typology, mediated by trust and moderated by brand affiliation, monthly contributions and the number of dependent beneficiaries in the medical schemes sector. This knowledge gap prompted a study examining how ethical sales behaviour impacts trust as a mediator of switching costs typology in medical schemes. As no research has examined this role in literature, validating it remained interesting. The study shows the direct impact of the findings, which may be necessary for health-care marketing theory and relevant stakeholders, policymakers and practitioners. The results affect the target community or study area. The study contributes to general theory of marketing ethics (Hunt and Vitell, 1986; Hunt and Vitell, 2006), commitment-trust theory of relationship marketing (Morgan and Hunt, 1994), and switching costs typology (Burnham et al., 2003). The unified explanatory power of these relationship exchange theories enriches the theoretical framework and unveils new insights unifying ethical sales behaviour, trust and switching costs typology explanations in the health-care marketing field. The findings adhere to prior research (Aydin and Özer, 2005; Carter et al., 2014; Yen et al., 2011) recommending studies test the relationship between trust and switching costs in several contexts. Moreover, Agrawal (2017) proposed that future research evaluates trust in management as a mediating or moderating variable in the link between ethical climates and other organisational variables, including commitment, citizenship behaviour or productivity. Other scholars, Gundlach and Murphy (1993) even recommended the research, asking: what ethical values are vital in establishing trust, commitment and solidarity of association in exchange relationships? Loe et al. (2000) recommended a study that unifies the ethical constructs suggested by ethical decision-making theory with marketing-related variables, namely, marketing orientation, quality and performance, as appropriate in reducing the gap. Thus, the novel contributions of this study are shown in the mediation of customer trust on the relationship between ethical sales behaviour and switching costs typology, moderated by brand association, monthly contributions and the number of beneficiaries in the medical scheme. By examining these interactions, this study unveils the role of the dimensions of these theories in literature and practice.

Considering the relevance of ethical sales behaviour as the primary marketing variable in the marketing strategy formulation and success (Al-Khatib et al., 2005), this study illustrates that unethical sales practices negatively affect customer trust, despite the positive effect trust has on personal relational loss costs, economic risk costs, evaluation cost, monetary costs and benefit loss costs in the context of medical schemes in South Africa. A negative effect of ethical sales behaviour on trust differs from previous studies, showing the positive relationship between salespeople’s ethical behaviour and customer trust in service literature (Mansouri et al., 2022; Ou et al., 2015; Singh et al., 2012; Wijaya et al., 2022). Paradoxically, customer trust was expected to affect economic risk costs and evaluation costs negatively, as Burnham et al. (2003) classified these “procedural switching costs” as negative switching costs (actual losses). For instance, customers incur an economic loss if a contract is ended or they lose investments (i.e. reward points, discounts and rewards) that may have been painstakingly acquired.

Implications of trust as a mediator of the influence of ethical sales behaviour on personal relational loss costs, evaluation costs and monetary costs are crucial. The valuable insights for managers suggest how to initiate practices that increase ethical sales behaviour, trust and switching costs typology by adding value in the medical scheme industry. This is important for members with high brand affiliation, high monthly contributions and a low number of beneficiaries.

The study’s findings contribute valuable knowledge to the health-care system transitioning to the NHI by 2025 in South Africa. A unique contribution of this study is in showing the mediation effect of trust on the relationships between ethical sales behaviour of medical schemes and personal relational loss costs, monetary loss costs and evaluation costs. Furthermore, the findings show the moderating effect of brand affiliation on the relationship between ethical sales behaviour and customer trust in the medical schemes setting. Concerning procedural switching costs, the relationship between trust and economic risk costs is moderated by low brand affiliation and a low number of beneficiaries. Moreover, a low number of beneficiaries moderates the relationship between trust and evaluation costs. For financial switching costs, the relationship between trust and benefit loss costs is moderated by low monthly contributions and a low number of beneficiaries. In addition, a low number of beneficiaries moderates the relationship between trust and monetary loss costs. For relationship loss switching costs, none of the moderating variables affect the relationship between ethical sales behaviour and personal relational loss costs and the relationship between trust and personal relational loss costs.

This study validates a direct negative effect of unethical sales behaviour of medical schemes on trust. Trust mediates the impact of ethical sales behaviour on personal relational loss costs, monetary loss and evaluation costs. Similar to the origin of the scale by Burnham et al. (2003), a few facets of the switching costs typology (learning, set-up and brand relational loss costs) had low reliability scores, which require further investigation. For example, it suggests that these are not key concerns to customers in medical schemes as they were deleted due to a low reliability score. The moderating effect of the duration of affiliation was also not considered, as the topic primarily focused on customer switching behaviour between medical schemes than retention. This may limit the understanding of medical scheme practitioners who need to understand the importance of these factors, as consumers seem to value them less and cannot comprehend their significance when evaluating kinds of switching costs between medical schemes. This study assesses perceptions of beneficiaries in open and restricted medical schemes, as opposed to M’bouaffou et al. (2022) who focused on open-scheme principal members.

However, the study’s findings cannot be generalised to the broader beneficiaries of medical schemes in South Africa, as the study included only principal members residing in Gauteng, despite accounting for 40% of the medical scheme members in the country, which is considered the highest percentage of beneficiaries. Estimating a sample size using non-probability sampling could be irrelevant, mainly if convenience sampling is used, as it is likely to generate non-generalisable results, which preclude statistical inference to the larger population (Althubaiti, 2023). Therefore, this study applied a non-probability judgement sampling method. Unfortunately, this research did not include respondents over the age of 65 years due to various ethical factors that can result in older adults or elderly people feeling vulnerable to abuse, neglect and ill intent as a result of the topic being studied. Discussion prevented arousing feelings of neglect and abuse; as there was no instant prepared support for their decision-making to participate in the survey, and the study avoided tempering their independence and dignity.

Future research can undertake a longitudinal study to explore related health-care services and better understand the role of NHI in the South African environment, as customers perceive the ethical sales behaviour of medical scheme salespeople negatively. This research only focuses on buyers’ perceived switching costs and excludes sellers’ perceptions. Future studies can examine this concern. Moreover, probing the role of switching costs within a network is essential because, in particular network industries, only one standard exists from the outset (Hess and Ricart, 2003). For instance, pharmacists are also considered ethical decision-makers (Perepelkin and Wilson, 2018). In addition, future studies could explore the health-care services and outcomes associated with caesarean deliveries in the medical schemes’ population. Factors for investigation may include, but are not be limited to, gestational age at the time of caesarean delivery, post-delivery hospital admission of baby or mother, birth weight and maternal clinical comorbidities (CMS, 2020).