Interconnectedness of trust-commitment-export performance dimensions: A model of the contingent effect of calculative commitment

Abstract This research on relationship marketing aims to revisit and reposition different foci of trust, commitment, and performance perception in the export/import relationships and explore the interconnectedness effects. We have collected self-reported survey questionnaire responses from 142 non-oil exporters in Ecuador. The data were analyzed with SmartlPls 3.0 software. We found that calculative commitment negatively moderates inter-organizational trust and affective commitment relationships. The other significant findings include the indirect relationship (mediating effect) of affective commitment to financial export performance through relationship export performance. With these novel contributions, we also identify some expected relationships- as both interpersonal and inter-organizational trust positively affects affective commitment, and relationship export performance significantly predicts financial export performance. Cross-sectional data collection and responses from one side of the export-import dyad are one of this research’s limitations. However, they are not uncommon in export marketing literature. Giving a justified position of different dimensions of trust and commitment in the export-import equation is the novelty of this scholarship. Clarifying the affective commitment and export performance relationship is another contribution of this research. Nevertheless, the dimensional views of trust and commitment re-established some known assumptions in a less researched country setting should also be considered a contribution.


Introduction
Trust and commitment are at the heart of relationship marketing (Morgan & Hunt, 1994). These sentiments are especially important for exporting firms that seek to establish, maintain, and leverage strong connections with foreign buyers over what may be considerable geographic, cultural, and institutional barriers. While there is a substantial literature base that focuses on how trust and commitment influence various performance measures (e.g., C. Bianchi & Saleh, 2020;Bloemer et al., 2013;Mahmoud et al., 2020;Styles et al., 2008), many of these studies convey that trust leads to greater commitment (e.g., T. Čater & Čater, 2010;B. Čater & Zabkar, 2009;Geyskens et al., 1996), which in turn enhances firm (export) performance (e.g., Alteren & Tudoran, 2016;Bloemer et al., 2013). However, this conceptualization may be too simplistic in that it ignores the iterative nature of these constructs in ongoing exchange relationships (Dowell et al., 2015), thus underscoring the need to extend the extant literature. Consequently, several significant gaps still remain in the existing literature.
Trust can be thought to exist in situations where one party is willing to rely on the actions of another (Morgan & Hunt, 1994). While there is a growing acceptance in the literature on trust's multidimensional nature, much of the early research on trust found in business disciplines focused on interpersonal levels of trust, i.e., sentiments held by individuals toward others and the relative strengths of the trust dimensions has not been discussed sufficiently (Dowell et al., 2015). Cropanzano et al. (2017) contend that this may represent an overly narrow perspective. More recent trust conceptualizations recognize that trust can also be directed to social units, such as workgroups, departments, organizations, and even across organizational boundaries (Schilke & Cook, 2013;Zaheer & Harris, 2006). Consequently, we witness an increasing number of studies that incorporate multiple dimensions of trust, especially interpersonal and inter-organizational, and explore the interrelationship between these dimensions and how they influence other variables (Ashnai et al., 2016;Mouzas et al., 2007Mouzas et al., , 2007Vanneste, 2016;Zaheer & Harris, 2006;Zaheer et al., 1998).
Commitment, which is construed as an enduring desire to maintain a valued relationship (Moorman et al., 1992;Morgan & Hunt, 1994), conveys the sense of an exchange partner's attachment to another and motivation for continuity (Gilliland & Bello, 2002). While many of the early studies that incorporated commitment operationalized this construct in a unidimensional fashion, more recent conceptualizations recognize that commitment is multifaceted, although varying definitions and dimensions are found in the extant literature (Geyskens et al., 1996;B. Čater & Zabkar, 2009;Kim et al., 2011;Jain et al., 2014). The various types of commitment most frequently cited in the marketing and organizational science literature are affective commitment (based on emotional attachments or social sentiments) and calculative commitment (based on rational economic calculations). Both forms of commitment reflect relatively stable attitudes and beliefs. However, the underlying motives of each and the mindsets that they evoke differ (Chang et al., 2012;Geyskens et al., 1996;Guo et al., 2016;Lee et al., 2007;Li et al., 2006). Gilliland and Bello (2002) have argued that while affective and calculative commitment may coexist, their distinct underlining motivations to maintain a relationship can have different effects on the trading partner's behavior. Hence, viewing commitment as a universal construct risks oversimplification, potential bias, and confounding effects (Hessling et al., 2018;Jain et al., 2014;Styles et al., 2008;Vanneste, 2016).
Despite the growing number of studies incorporating multiple dimensions of commitment, the influence of calculative commitment remains understudied in the context of export-import marketing despite its logical character and likely function as a rationale for parties to engage in and maintain business relationships. Nor has there been sufficient research that has addressed if there is a connection across the different types of commitment (Hessling et al., 2018).
Furthermore, export-oriented relationship marketing studies grounded on relational exchange theory (RET) have often conceptualized trust and commitment as antecedents of export performance, which has typically been expressed in terms of financial and/or non-financial performance measures (Bloemer et al., 2013;Mahmoud et al., 2020;Mysen & Tronvoll, 2020;Styles et al., 2008). However, the importance of relationship performance has not received sufficient attention. For instance, some studies have included affective commitment to examine whether it enhances performance (Ashnai et al., 2016;Bloemer et al., 2013) but do not clarify what aspect of performance (i.e., financial and/or non-financial) represents such benefits. Other recent studies have taken a narrow view of the direct link between commitment and export performance (e.g., Alteren & Tudoran, 2016;Tan & Sousa, 2015) and have overlooked the role of relational performance, which represents a significant and problematic limitation of the current export marketing literature.
We aim to contribute to the export marketing literature by addressing these aforementioned gaps. Grounded on the commitment-trust theory (Morgan & Hunt, 1994), we develop and test a novel process model that includes multiple dimensions of trust, commitment, and export performance. In sum, this study seeks to refine our understanding of the roles of different trust and commitment dimensions on export relationships and financial performance. In particular, we examine the moderating role of calculative commitment on the interpersonal and interorganizational trust links with affective commitment. The theoretical contributions of this research are to be understood in light of the Relational Exchange Theory (RET) and Commitment -trust theory, where the different foci of trust, commitment, and performance are linked. We provide empirical evidence that interpersonal trust leads to inter-organizational trust by segregating the trust conceptualization. Further, we push the boundary for international exchange relationships research by specifying the moderating influence of calculative commitment on inter-organizational trust-affective commitment relationships. As for the managerial contributions, this research could work as a recipe for exporting managers, especially those in the developing countries where we advocate the network-based exchange relationships and building interpersonal trust; while keeping in mind the critical importance of calculative commitment.
In the ensuing section, we provide our model, the theoretical underpinnings for it, and a series of hypotheses. Next, we summarize our data collection protocol, operational measures, and the methodology employed to validate our measures and empirically test our hypotheses and our findings. We conclude with a discussion of theoretical and managerial implications, limitations, and directions for future research.

Theory and hypotheses
Relationship marketing has been a dominant paradigm that has guided research in marketing channels, international business, export-import, and other business disciplines. Trust and commitment are central themes within the relationship marketing literature because of their roles in developing and maintaining long-term exchange relationships (Gounaris, 2005;Morgan & Hunt, 1994). In their seminal paper, by investigating the nature of relationship marketing, Morgan and Hunt (1994) proposed that commitment and trust are the key moderating variables and initiated the research directions further to the commitment-trust theory domain. The underpinning theory of this research is Morgan and Hunt's (1994) commitment-trust theory. We will discuss the theoretical underpinning and develop the research hypothesis in the preceding sections. Mayer et al. (1995, p. 712) define trust as "the willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the trustor, irrespective of the ability to monitor or control that other party." Within the extant literature, there has been substantial attention dwelling on whether trust is an individual or organizational property and whether this construct is best conceptualized as being unidimensional or multidimensional (Möllering & Sydow, 2018;Zaheer et al., 1998). However, a consensus now exists that trust comprises two main dimensions: interpersonal and inter-organizational (Ashnai et al., 2016;Zaheer & Harris, 2006;Zaheer et al., 1998). Zaheer et al. (1998) argue that people may trust both other individuals as well as collectives, such that interpersonal trust is defined as that "placed by the individual boundary spanner in her individual opposite member" and interorganizational trust as "the extent of trust placed in the partner organization by the members of a focal organization" (p. 142). Thus, interpersonal and inter-organizational trust differ, both empirically and conceptually (Ashnai et al., 2016;Doney & Cannon, 1997;Nielsen, 2004). Ashnai et al. (2016) portray emotions as the source of interpersonal trust. Using a four-stage process model of trust development, Schilke and Cook (2013, p. 285) argue that "the focal organization's boundary spanner is the starting point of the trust process." Interactions between boundary spanners are where valued information is exchanged, and negotiations occur, thus serving as the basis for interpersonal trust (Huang et al., 2016).

Trust and the interrelationship between its dimensions
Inter-organizational trust arises from rational and pragmatic considerations of interfirm adaptations and inter-organizational learning (i.e., has a cognition base), as well as other historical aspects of inter-organizational relationships, such as previous exchange performance, negotiations, or conflicts (Ashnai et al., 2016;Dowell et al., 2015;Geyskens et al., 1996;Nielsen, 2004).
Prior research suggests that interpersonal trust affects inter-organizational trust, but not the other way around (Ashnai et al., 2016;Vanneste, 2016), at least in the earlier stages of the relationship development (Dowell et al., 2015). The positive interpersonal sentiment that interpersonal trust represents can have a ripple effect, such that "a common understanding regarding the trustworthiness of the partner organization develops and organizational alliance routines are established reflecting' how things are done' with this partner organization (i.e., the establishment and institutionalization of organization-organization trust occurs)" (Schilke & Cook, 2013, p. 285). Consequently, we hypothesize: H1: Interpersonal trust has a positive effect on inter-organizational trust.

The relationships among trust and commitment dimensions
In the realm of relationship marketing, the interrelated notions of trust and commitment are associated with positive collaborative benefits, such as uncertainty reduction, increased resource utilization efficiency, and value for both partners (Ashnai et al., 2016;Gounaris, 2005).
A committed relationship is where the partners are "forward-looking" and want the relationship to endure indefinitely (Morgan & Hunt, 1994). While commitment was initially conceptualized and operationalized as a unidimensional construct, more recent conceptualizations recognize that commitment is multifaceted, albeit with varying conceptualizations of what dimensions exist. Two dimensions, affective commitment and calculative commitment are most prominent in the extant literature (Geyskens et al., 1996;Meyer et al., 2004;B. Čater & Zabkar, 2009;Kim et al., 2011;Jain et al., 2014).
These dimensions imply that inter-organizational relationships can be viewed from two diametrically opposite perspectives. On the one hand, they can be seen as embedded institutions where the focus is on social sentiments, attributes, and bonds. Hence, the attention to affectional commitment. On the other hand, calculative commitment embodies a utilitarian perspective that reflects implicit or explicit cost-benefit analyses that serve as a pragmatic lens by which economic and strategic benefits of commencing and remaining in a current exchange relationship versus engaging an alternate partner are evaluated (Bansal et al., 2004;Dwyer et al., 1987;Gilliland & Bello, 2002;Lopes, 2016;Meyer & Allen, 1991;Styles et al., 2008). Hence, strong arguments exist that these two commitment dimensions are separate, leading to the call by Liu et al. (2010) for more research to investigate the link between these dimensions.
Affective commitment arises from identification, common values, attachment, involvement, and similarity. It implies a positive motivation to continue the relationship because one party likes the other and enjoys working with them, thus reflecting favorable feelings, mindsets, and attitudes toward the trading partner and "wanting to" or having a desire to continue the relationship (Bansal et al., 2004;T. Čater & Čater, 2010;Geyskens et al., 1996;Gilliland & Bello, 2002;Lariviere et al., 2014;Meyer & Herscovitch, 2001), which creates a sense of unity or attachment to the exchange partner.
Affective commitment evolves over time. When grounded on trust, it signals a desire to maintain a relationship in order to attain a cooperative, mutually beneficial future. Thus, greater affective commitment can be expected to lead to stronger relationship stability and continuity (Bloemer et al., 2013;Skarmeas et al., 2002).
Based on the principle of generalized reciprocity from social exchange theory (Gounaris, 2005), we anticipate that the emotional attachment, identification, and sense of involvement that arises from interpersonal trust also contribute to exporters' affective commitment to importers (Chang et al., 2012;Meyer & Allen, 1991). In the context of emerging markets, Park and Luo (2001) found that the interpersonal connections between the boundary spanners serve as a lubricant within exchange relationships.
According to Sharma et al. (2006), familiarity, friendship, and personal confidence built through interpersonal interaction over time between two persons or groups initiate the desire to develop and strengthen a relationship that might lead to affective commitment. Hence, in line with previous research (Bloemer & Odekerken-Schröder, 2006;Chang et al., 2012), we predict that exporters' interpersonal trust exerts a positive effect on affective commitment. Accordingly, we propose: H2: Interpersonal trust has a positive effect on affective commitment.
Since inter-organizational trust refers to the collectively held confidence in the reliability and integrity of an exchange partner, this construct also implies expectations of the benefits that accrue from the exchange relationship and the partner's ability and motivation to satisfy the buyer's specific needs. Consequently, it can lead to the presence of predetermined ways to solve business problems jointly, thus allowing the focal organization to take risks that can strengthen the relationship and make it more likely to persist (Ashnai et al., 2016;Zaheer et al., 1998).
Relationships across the different dimensions of trust and commitment rarely have been explicitly examined in export-import settings. A positive association between inter-firm trust and affective commitment has been reported in a meta-analysis performed by Delbufalo and Wilding (2012). In a recent study of UK-based buyer-supplier relationships, Ashnai et al. (2016) argued that the rational aspects of trust (i.e., inter-organizational trust) enhance commitment. Their study reported a statistically significant relationship between inter-organizational trust with a unidimensional measure of commitment, which on closer inspection actually denotes affective commitment. Thus, we expect that when inter-organizational trust exists, it too will positively influence affective commitment.

H3:
Inter-organizational trust has a positive effect on affective commitment.
While RET highlights the importance of emotional bonds to maintaining valued business relationships, business enterprises' profit-seeking and utility maximization nature also attest to a utilitarian rationale for interfirm exchange relationships (Heide & Wathne, 2006;Hessling et al., 2018). The calculative commitment represents a constraining force that binds exchange partners together, i.e., it reflects "having to," or the perceived need to stay with the current exchange partner based on instrumental reasons, i.e., the relative availability of alternative exchange partners and the dispassionate assessment of tangible benefits versus the costs of leaving the current exchange relationships (T. Čater & Čater, 2010;Fischer & Mansell, 2009;Geyskens et al., 1996;Lariviere et al., 2014;Meyer & Allen, 1997).
In the export marketing literature, calculative commitment has seldom been explicitly addressed, unlike affective commitment (Bloemer et al., 2013). Since sufficient empirical evidence is absent from the export-import marketing literature to support whether there is a direct relationship between calculative and affective commitment, we instead rely on these two concepts' purported separability as indicated in previous research (e.g., Geyskens et al., 1996).
In inter-organizational relationships, calculative commitment refers to a structural bonding reflecting a trading partner's rational concern for instrumental gain and thus is seen as representing a negative motivation for continuing an exchange relationship (T. Čater & Čater, 2010;Geyskens et al., 1996;Liu et al., 2010). An inter-organizational relationship is expected to have minimal odds of survival based solely on calculative commitment (De Ruyter et al., 2001). Previous research suggests a strong positive association of trust with affective commitment, while a negative association with calculative commitment prevails (see, B. Čater & Zabkar, 2009 for a review). However, a more recent study of export-import relationships (Bloemer et al., 2013) contends that trust is positively linked with affective commitment only.
While many past studies have investigated trust as an antecedent to commitment dimensions, the reality is that in ongoing exchange relationships, trust and commitment are interrelated in a reciprocal cycle, meaning that they iteratively influence one another (Styles et al., 2008). But merely transposing the sequence of trust and commitment may not be sufficient. Another way of thinking about calculative commitment's interrelationship with trust is to view it as a potential moderator (Kim et al., 2011), which may mitigate the effect of the two trust dimensions on affective commitment.
Given that affective commitment is seen as a positive motivation for seeking to prolong an exchange relationship and that both trust dimensions are expected to foster affective commitment, we posit that calculative commitment and its connotation of being a negative motivation can serve to moderate the positive relationship of the trust dimensions on affective commitment. Since both trust dimensions and affective commitment are phenomena that occur after an exchange relationship commences, they implicitly reflect the perception of the moral hazard risk that the exporter faces. On the other hand, calculative commitment can be construed as an evaluation of the adverse selection hazard faced before engaging the exchange partner and one that continues explicitly or implicitly throughout the interfirm relationship (Bergen et al., 1992). Thus, calculative commitment may exacerbate the exporter's perceived vulnerability to trusting and relying on a focal importer and hinder the inter-organizational relationship (Liu et al., 2010). In sum, it may impede the positive effect of both trust dimensions on affective commitment. Hence, we propose: H4a: Calculative commitment moderates the link between interpersonal trust and affective commitment.
H4b: Calculative commitment moderates the link between inter-organizational trust and affective commitment.

Export performance and how its dimensions are influenced by trust and commitment dimensions
Export performance is another multifaceted concept, usually classified along the lines of financial and non-financial aspects (Bloemer et al., 2013;Lages, 2000). Financial export performance has typically been assessed using objective indicators such as export sales and growth, export profit levels and growth, return on investment, and achieved market share (Lages, 2000). Non-financial export performance refers to more subjective indicators of outcomes and the relationship's efficacy, such as goal achievement, satisfaction, and perceived success. It has also been operationalized as customer, employee, and shareholder's satisfaction and loyalty (Bloemer et al., 2013;Lages, 2000).
Relationship performance stems from a specific set of non-financial export performance measures that seek to assess inter-organizational relational dynamics attributed to a focal exchange partner (e.g., level of efficiency, productivity, and contribution toward achieving financial goals; Luo et al., 2015;O'Toole & Donaldson, 2002). Another definition of relationship performance refers to a particular dyad's perceived economic performance relative to expectations within a broader network of exchange relationships (Medlin, 2003), which can be instrumental to sales growth, market positions, marketing support, or qualified services for the parties involved. Others have used satisfaction as a proxy for relational performance (Lui et al., 2009;Saxton, 1997). Successful inter-organizational relationships can enhance firms' financial performance by leveraging partners' environmental scanning efforts, capabilities, and resources to compete effectively.
Prior research has found that trust fosters positive sentiments to stay in the relationship, propelling financial performance (C. C. Bianchi & Saleh, 2011). Beyond the direct effect of strong interfirm relationships on sales and profits (Palmatier et al., 2006), they encourage increased cooperation and reduced conflict, so they also can benefit from innovation efforts, expanded markets, and reduced costs (C. C. Bianchi & Saleh, 2011). Bloemer et al. (2013) note that trust and affective commitment energize this performance. According to Lee et al. (2007), affective commitment also leads to relationship performance through altruistic benevolence. Dowell et al. (2015), also found a positive effect of commitment on relationship performance.
Thus, we predict that affective commitment has a central role in determining relationship outcomes (Ashnai et al., 2016;Morgan & Hunt, 1994;Palmatier et al., 2006), which ultimately facilitates financial performance through relationship performance. Medlin (2003) suggests that relationship performance can directly lead to economic outcomes without relying on mediating indicators (e.g., level of satisfaction or cooperation). Formally, based on the internationalization process model (Johanson & Vahlne, 1990), we hypothesize that pro-social attitudes, namely interpersonal trust, inter-organizational trust, and affective commitment, are determinants of enhanced relationship performance, which in turn propels financial export performance: H5a: Interpersonal trust has a positive effect on relationship export performance.
H5b: Inter-organizational trust has a positive effect on relationship export performance.
H5c: Affective commitment has a positive effect on relationship export performance.

H6
: Relationship performance has a positive effect on financial export performance.

Sample and data collection
Data for this study were collected from active, non-oil exporting companies operating from Ecuador. Our country choice was influenced, in part, by recent criticisms that have indicated that export-import studies have been conducted mainly in North America, Europe, and Asia, which has led to calls for more research involving firms from developing nations (Aykol & Leonidou, 2018;Samiee & Chirapanda, 2019). Furthermore, research involving Latin American firms has been deemed to be inadequate, especially studies addressing internationalization processes among Latin American firms (C. Bianchi & Saleh, 2020;Fastoso & Whitelock, 2011;Paul & Mas, 2019). Brazil, Chile, and Mexico, and to a lesser extent, Argentina and Colombia, have predominantly been the context when Latin companies have been studied (Fastoso & Whitelock, 2011;Paul & Mas, 2019).
We patterned our data collection procedures on the World Bank's data collection protocol for its Enterprise Survey 2017. A list of non-oil product exporting companies published by the Ecuadorian Ministry of Foreign Trade (Ministerio de Comercio Exterior; 2014) was the sampling frame of this study. Data were obtained using a telephone survey conducted by an external research company (an independent call center). Using a mail survey was not considered feasible because of the generally low and declining response rates to mail surveys in Latin American countries and Ecuador's low postal reliability. We also ruled out using an internet survey, given the low broadband penetration rate in the country.
Our data collection involved key informants from exporting firms; hence, while recruiting respondents, it was urged to direct the phone calls to the persons who were knowledgeable about the exports of the business; and who were instructed to respond relative to a single export venture, which helps reduce the potential for systematic or random sources of error (John & Reve, 1982;Krause et al., 2018). When asked to consider the single export venture, respondents were requested to reflect on any export-import relationship they wished.
A total of 1,330 companies were attempted to be contacted, but only 985 companies could be reached because of incorrect contact information for the others. Our final sample consisted of 142 valid surveys (corresponding to a 14.4% response rate). While this is a relatively small sample size, it is consistent with previous organizational research ( Baruch & Holtom, 2008;Krishnan & Poulose, 2016). A plausible explanation for such a low response rate is that some respondents feel uncomfortable talking about their relationship with a particular importer. From the authors, personal experiences in dealing with emerging countries' exporters usually have some reservations in talking about their export-import relationships.

Data profile
Among the organizations that provided data, 50% earned more than USD 5 million in annual sales, and 47% took in annual sales volumes of USD 100,000-5 million. The firms in our sample had an average of 191 employees. Their average exporting experience was approximately 18 years (SD = 11.38), while the average length of the focal organizational relationship was more than ten years (SD = 6.20). The most common job titles for the key informants were president (43%), chief financial officer (16%), and export manager (18%). The United States was the most targeted export destination, representing 37% of the Ecuadorian exporters in our sample, followed by Colombia (22%), the European Union (12.6%), and then Russia (9.2%), which is similar to other exporters from the region (Paul & Mas, 2019).

Measurement instruments
The operational measures of the constructs included in our model were based on established English language scales. Back translation was used to create a Spanish-language survey instrument (C. Bianchi & Saleh, 2020;Brislin, 1970). A professional translator in Ecuador first translated the original questionnaire from English into Spanish. Next, one of the co-authors, a native Spanish speaker, checked the translation and performed a back-translation to identify and resolve any disagreements. Finally, we pretested the questionnaire to check for any remaining anomalies using a group of 20 students (all with experience working for different corporations in Ecuador) pursuing a master's degree from a prominent university in Quito. The results of this pretest revealed no substantial flaws or misunderstandings in the questionnaire, allowing us to proceed to data collection.
To measure inter-organizational trust and interpersonal trust, we used ten and five items, respectively, drawn from Zaheer et al. (1998). The measure of affective commitment was operationalized with a five-item scale (Gounaris, 2005), whereas calculative commitment was measured by a three-item scale from Gilliland and Bello (2002). All items relied on 7-point Likert-type scales, ranging from 7 = strongly agree to 1 = strongly disagree. We measured financial export performance with four items from Lages (2000) and relationship export performance with four items derived from Luo et al. (2015). 1

Analysis and results
Partial Least Squares (PLS) is a confirmatory second-generation multivariate analysis technique that allows the examination of both latent and manifest variables simultaneously (Ahamed & Skallerud, 2013;Johnston et al., 2004). Besides the more popular covariance-based SEM (CB-SEM) approaches, PLS-SEM has also gained marketing and management research attention as an alternative (Hair et al., 2011). As PLS allows the researchers to investigate models at a higher level of abstraction, it is helpful for predictive purposes or exploratory research in situations where the theory is still developing further; it is capable of dealing with smaller sample sizes, and data normality is not required (Sarstedt et al., 2022). In view of these advantages, the hypotheses expressed in our model were tested in a structural equation model using SmartPLS 3.0 .

Analyses for biases
We first checked for nonresponse bias by applying the wave analysis guidelines suggested by Armstrong and Overton (1977) and tested for any significant differences between early (first 75 percent returned) versus late (last 25 percent) responses on demographic and substantive variables. The test of the homogeneity of variance (Levene statistic) in SPSS indicated no significant differences at the 0.05 level, so we concluded that nonresponse bias was not a concern. Although SmartPLS does not produce direct statistics for assessing common method bias, previous studies have advised using a full collinearity assessment (Kock, 2015). We found that the variance inflation factor (VIF) scores were below the tolerance level of 5.0 recommended by Kock (2015), which suggested that common method bias was not present. We also performed the measured latent marker variable (MLMV) approach to detect the Common Method Bias (CMB) as suggested by (Chin et al., 2013); where we followed the construct level correction (CLC) approach by comparing the path coefficients and variance explained (R 2 ) differences with and without measured marked variable. The PLS results revealed that the differences in β coefficients and R 2 were not greater than 10%. Therefore, the common method bias was not a problem in the present study. Finally, to attenuate socially desirable response biases, we informed the key informants that their responses were voluntary and guaranteed anonymity.

Reliability and validity
Before conducting our hypothesis tests, we ran a measurement model to evaluate the factor loading of the items of the corresponding constructs. Any item with a factor loading below 0.60 was deleted, and the purified scale was used for further analysis (Hair et al., 2014). We then assessed the reliability and validity of the measurement items in two steps. First, to determine the reliability and internal consistency of the constructs, we assessed composite reliability (CR) and Cronbach's alpha, using a threshold as displayed in Table 1, both reliability values for all substantive variables exceeded the threshold value of 0.70 recommended by Hair et al. (2014). Further evidence of reliability was demonstrated by the rho_A coefficient, a reliability measure for partial least squares. The rho_A coefficients ranged from 0.88 to 0.98, which are greater than 0.70 threshold recommended by Henseler et al. (2016).
To assess item multicollinearity, we calculated outer VIF; none of the items were found with a higher VIF score than the threshold of five.
As a check for convergent validity, each scale's average variance extracted (AVE) scores were determined. As listed in Table 1, all AVE scores are higher than 0.50, as Wong (2013) recommended. Therefore, the scales used in this research were all considered to achieve convergent validity.
To check for discriminant validity, we applied the well-known Fornell-Larcker criterion (i.e., the square root of AVE for each latent variable is higher than other correlation values among any other construct) (see , Table 1) and the Heterotrait-Monotrait (HTMT) ratio (Hair et al., 2014;Henseler et al., 2015). Table 1 shows that all the HTMT ratios are below the recommended threshold level of less than 0.85 (Henseler et al., 2015).

Assessment of structural model
To detect whether construct multicollinearity was present, we examined the VIF of each variable (Lowry & Gaskin, 2014). As shown in Table 2, all VIFs are below 5, as Wong (2013) recommended. Thus, we concluded that none of the latent variables suffer from multicollinearity.  To assess the overall fit of our model, we examined several goodness-of-fit indicators recommended when partial least squares (PLS) are used, as displayed in Table 3.

Hypotheses tests
To test the path coefficients in the structural model that underscore our hypotheses, we conducted bootstrapping in SmartPLS-3, a resampling technique that estimates the standard error without relying on distributional assumptions (Hair et al., 2014). Figure 1 displays the path coefficients and their reported significance. Table 4 reports the results of the hypothesis tests.
Though we have not specified any mediation hypotheses, we reviewed the specific indirect effects from the bootstrapped results to determine whether the upper limit (97.5%) and the lower limit (2.50%) of the bias-corrected confidence intervals included a zero (Hayes & Scharkow, 2013). Our results also indicate that inter-organizational trust partially mediates the effect of interpersonal trust on affective commitment since the confidence intervals do not include a zero. Similarly, the effect of interpersonal trust on relational performance is partially mediated by inter-organizational trust and affective commitment. On the other hand, the effect of interorganizational trust on relational performance was fully mediated by affective commitment. Moreover, the effects of interpersonal trust, inter-organizational trust, and affective commitment on financial performance are fully mediated by relational performance. 2

Discussion and implications
Our results convey that greater understanding is achieved by taking a dimensional view when examining the interrelationships among trust, commitment, and performance dimensions. This research provides empirical evidence that largely supports our conceptual model of the drivers of exporter financial performance (i.e., interpersonal trust, inter-organizational trust, affective commitment, and relational performance) and the process by which these variables enhance financial performance. Unlike many export relationship marketing studies, this research highlights the role of calculative commitment in exporter-importer relationship dynamics and provides empirical evidence that calculative commitment moderates the influence of inter-organizational trust on affective commitment, but it did not moderate the effect of interpersonal trust on affective commitment. Towards the end, this paper agrees with Kemp et al. (2018); that "human decision-making often includes a complex cadre of emotions and rationalizations" (p. 19). In the following paragraphs, we review the results of the hypothesized relationships along with theoretical and managerial implications derived from them.

Direct relationships
Our findings confirmed H1, which predicted a positive association between interpersonal trust and interorganizational trust and gave further credence that interpersonal and inter-organizational trust are interrelated, but empirically and conceptually distinct (Ashnai et al., 2016;Doney & Cannon, 1997;Nielsen, 2004), and that interpersonal trust causes inter-organizational trust and not the other way around (Ashnai et al., 2016;Vanneste, 2016). Our results are also consistent with the process model of trust development conceptualized by Schilke and Cook (2013), which depicts an organization's boundary spanner is the starting point of the trust process. Repeated and cordial interactions between boundary spanners provide the context for interpersonal trust to develop and facilitate a common understanding of the trustworthiness of the partner organization (Huang et al., 2016;Schilke & Cook, 2013).
The significant positive association of interpersonal and inter-organizational trust with affective commitment (H2 and H3) indicates that both dimensions of trust are instrumental in fostering affective commitment between an exporter and an importer. Thus, both trust dimensions can be thought of being facilitators to this positive motivator for the continuity of the exchange relationship and reemphasize the proposition that "personal chemistry" (Dowell et al., 2015) between the trading managers' is a decisive factor for the fate of the E-I relationship. These findings are consistent with the empirical evidence from earlier studies that examined trust's influence on commitment dimensions, (e.g., T. Čater & Čater, 2010;B. Čater & Zabkar, 2009;Geyskens et al., 1996;Gilliland & Bello, 2002). Our findings are distinguished because we explicitly included both trust dimensions rather than the unidimensional trust measures common to earlier studies. These results provide additional support for Zaheer et al. (1998) conceptualization that the two trust dimensions are distinct though

Figure 1. Research model with path coefficients (β) and p-values in parentheses.
Notes: Significant paths are displayed as solid lines; nonsignificant (NS) paths are in italics and shown as dotted lines. R 2 refers to the R-square of endogenous variables.
interrelated. Further, we found that interpersonal trust, directly and indirectly, influences affective commitment (i.e., its effect is partially mediated by inter-organizational trust).
We found that interpersonal trust and affective commitment are positively linked with the export relationship performance, supporting H5a and H5c. These findings, i.e., that the positive sentiments that arise from interpersonal relationships between boundary spanners (interpersonal trust) and the social motivation for continuity (affective commitment) increase relationship performance, which is aligned with conceptualizations that economic actions are embedded in interpersonal ties and relations (Granovetter, 1985;Huang et al., 2016;Peng & Luo, 2000). It also attests to a foundational premise of relationship marketing theory that positive interpersonal sentiments between managers can enhance performance (C. C. Bianchi & Saleh, 2011;Skarmeas et al., 2002).
However, we found no statistically significant direct relationship between inter-organizational trust and relationship performance (H5b). This finding contends with some of the previous research where it is argued that cognitive trust maximizes the relational benefits (Mayer et al., 1995). However, it seems that this result, coupled with support for H3 and H5c, suggests that interorganizational trust still facilitates relational performance, but indirectly, i.e., its effect is mediated by affective commitment. Implying that an emotional/affectional concept mediates the effect of cognition-based trust, resulting in relational benefits (Dowell et al., 2015).
As expected, relationship performance positively influences financial performance (H6), which is indicative that relationship performance channels the influence of affective commitment and its trust dimension facilitators to financial performance. This finding also supports the premise that pro-social forces are complementary to economic forces (Lee et al., 2007;Luo et al., 2015). Collectively, interpersonal trust, inter-organizational trust, and affective commitment can be viewed as energizing forces (Vanneste, 2016) that stimulate relational performance and financial performance. Each represents a "form of motivation that binds individuals or organizations to actions and decisions that have relatively long-term implications" (Bloemer et al., 2013, p. 364).

Moderating hypotheses
While calculative commitment is an underlying motive for establishing an exchange relationship at the outset, the results of comparative assessments that continue thereafter mean that this dimension can represent a negative motivation to maintain an exchange relationship (T. Čater & Čater, 2010; Geyskens  , 1996;Liu et al., 2010). We found that calculative commitment had no moderating effect relative to the interpersonal trust-affective commitment relationship (H4a), whereas calculative commitment negatively moderated the effect of inter-organizational trust on affective commitment (H4b).
Given the emotional and social basis of interpersonal trust versus the rational and pragmatic considerations of inter-organizational trust (Ashnai et al., 2016;Geyskens et al., 1996;Nielsen, 2004), the non-significant moderating effect of calculative commitment on interpersonal trust may be a reflection of a lack of alignment between emotional sentiments versus asocial, rational assessments. Oliveira and Lumineau (2019) argue that calculative commitment helps organizations stay focused and avoid adverse effects stemming from an excessive trust, emotional attachments, and other social bonds. This underscores the notions of gesellschaft versus gemeinschaft rationales for the continuity of exchange relationships. The former, represented by calculative commitment, focuses on rational assessments and task-oriented actions that lead to economic gains or achievement of instrumental goals that do not consider the emotion or sentiments of the partner. The latter, relating to affective commitment, is more concerned with emotional content and recognizing, valuing, and preserving an existing relationship (Gilliland & Bello, 2002).
Reflecting the gesellschaft rationale for calculative commitment suggests that this commitment dimension negatively regulates the effect of inter-organizational trust on affective commitment as the organization considers pragmatic aspects, such as losing an existing importer or the possible loss of any dedicated one investments. In line with previous research, this negative moderating effect implies that despite indicating an attachment, calculative commitment does not guarantee any pro-social attitudes or behaviors between the exporting-importing firms. In other words, the linkage of two pro-social constructs (i.e., inter-organizational trust and affective commitment) appears to be regulated by pragmatism and the cognition-driven utilitarian evaluation of the exchange partner (Gilliland & Bello, 2002).
We do not find any significant moderating effect of calculative commitment on the interpersonal trust-affective commitment link, though (H4b). We posit that the social foundations of interpersonal trust and affective commitment might become so strong that calculative commitment has no room to exert a further effect. Further, this might also be linked with the boundaryspanning attributes of top executives of the exporting organizations who play a bridging role that includes information processing (i.e., decoding, filtering, and translating the information) and problem-solving. From the boundary spanners' perspectives, calculative commitment might be perceived as having a weak influence on the interpersonal ties between the exporting-importing managers (Huang et al., 2016).

Managerial implications
This research also offers implications for managers and policymakers, especially in less developed countries. In general, choosing ongoing exchange relationships with importers rather than opting to forward integration can be an effective strategy to penetrate foreign markets and achieve performance goals. A network-based approach through exporter-importer dyads can help firms grow by granting them market access and additional information, knowledge, and resources (Riddle & Gillespie, 2003). In developing countries like Ecuador, formal business network structures are apt to be less wellestablished, so firms tend to rely on informal managerial ties or interpersonal relationships to foster organizational performance (Peng & Luo, 2000). Thus, when venturing beyond home markets, enacting strategies that embody building and maintaining solid relationships with importer exchange partners is not only an extension of exporters' domestic approach but appears to be a more pronounced need. Our findings underscore that fostering a positive interfirm social climate through both trust dimensions and affective commitment creates favorable relationship dynamics that further facilitate financial performance. However, exporters should not focus solely on social aspects but must balance them against pragmatic considerations, where calculative commitment comes into play.

Limitations and future research directions
We would be remiss if we did not acknowledge the several limitations of this study and directions that future research could address. We collected data from exporting members of export-import dyads and thus lack insights from importers' perspectives. Future studies could comprise research designs to gather data on sentiments and behaviors from both exporters and importers.
Another limitation is that we only collected data from a single country, which constrains the extent to which we can generalize from the findings of this study to other developing nations; still, another limitation is that our study employed a cross-sectional design. Thus, while we present a plausible and logical sequence to the variables, we cannot establish causality with certainty. Future research could employ longitudinal data to support time-series analyses that provide more insights.
Future research could continue to explore the contingent effect of calculative commitment. What remains unknown is whether there might be an ideal level of calculative commitment or a tipping point above which an excessive degree of calculative commitment could convey a wrong signal that might overwhelm the effect of inter-organizational trust on affective commitment. This might be investigated using novel techniques like the response surface approach (Kim & Hsieh, 2003). Another direction of future research could be to take a relationship lifecycle approach and test the interaction of these relational and cognitive variables at different stages of the E-I relationship (Dowell et al., 2015;Shen et al., 2020).
Another direction for research could be to investigate whether other moderator relationships exist. For example, is there an interactive effect between the two commitment dimensions relative to performance outcomes?
Nevertheless, this study contributes to the export relationship marketing literature by providing a novel refinement to the trust-commitment-performance framework. We have offered a more nuanced portrayal of this chain by taking a multidimensional perspective of each of these constructs and testing for contingent effects.