Global standards and local authenticity: Cross-cultural leadership and employee behavioural adaptation in heritage-based hotels in Baghdad

2.1. Social Exchange Theory as the Organising Framework

SET, initially formalised by Blau (1964) and later extended in organisational contexts by Cropanzano and Mitchell (2005), holds that social behaviour is the product of an exchange process where actors weigh the costs and benefits of their relationships and respond accordingly. The theory has been applied to leader-member exchange research, with evidence that subordinates who perceive high-quality, reciprocal relationships with their leaders invest more discretionary effort in their work (Eisenberger et al., 2019). In the context of this study, SET provides the micro-level logic connecting leadership behaviour to employee adaptation. When a leader demonstrates cross-cultural competence—adapting communication style, acknowledging cultural difference openly and guiding employees through unfamiliar protocols—employees may experience this as a form of organisational support that invites reciprocal investment. That reciprocal investment can take the form of behavioural adaptation: the willingness to modify one’s communication style, service approach and professional self-presentation to meet diverse demands.

This framing matters for heritage hotels specifically because the cultural load on employees in such settings is unusually high. An employee at a hotel in a generic urban business district adapts to guests. An employee at a Tigris-facing historic property adapts to guests while simultaneously acting as an informal ambassador for a living cultural heritage. SET suggests why the quality of the leader-employee exchange in that environment may have outsized consequences: a leader who treats cultural adaptation as a purely technical skill problem (learn the software, use the English greeting) strips the exchange of its relational depth, whereas a leader who treats it as a culturally meaningful transition provides resources employees can more readily use.

2.2. Cross-Cultural Leadership: Construct Definition and Relevant Evidence

Cross-cultural leadership (CCL) in this study is understood as a behavioural construct encompassing four interrelated capacities: cultural intelligence, adaptive communication, inclusive decision-making, and what Bukhari et al. (2025) describe as cultural perspective-taking—the capacity to actively simulate the cultural assumptions of others when evaluating organisational problems. These capacities are treated as behaviours that can be trained, observed and evaluated, rather than as fixed personality traits.

The empirical record on CCL’s association with employee outcomes is broadly consistent with a positive relationship. Bukhari et al. (2025) linked employees’ cultural intelligence to constructive workplace behaviour through the mediating role of engagement in hotel settings. Cultural intelligence and well-being research in hospitality more broadly (Saito et al., 2024) situates culturally responsive management practices among the correlates of positive employee outcomes, although that review’s primary focus is employee well-being rather than CCL specifically, and it is cited here accordingly. Evidence from other emerging hospitality markets points the same way. Among 360 hospitality managers in India, cultural intelligence emerged as a foundational competency underpinning both communication effectiveness and global leadership capability, which in turn related to organisational effectiveness (Resho & Singh, 2025). Edmondson (1999) showed that positive team and leadership climates operate partly through psychological safety; when employees trust that their leader will not penalise cultural expression, they are more willing to engage in the interpersonal risk-taking that adaptive behaviour requires.

What the existing literature has not done is examine these effects in post-transition, heritage-specific contexts outside the West. Baghdad’s hospitality sector was essentially dormant for two decades of conflict. The employees currently working in its revived heritage hotels have, in many cases, no institutional memory of international service standards. Their cultural frames are deeply local, their foreign language capabilities often limited, and their professional socialisation largely domestic. The demands placed on them by newly globalised operations are therefore qualitatively different from the demands examined in most CCL studies, which assume, at minimum, some baseline of prior cross-cultural exposure.

On the basis of this reasoning, the first hypothesis is:

Hypothesis One: Cross-cultural leadership positively predicts employee behavioural adaptation in heritage-based hotels in Baghdad.

2.3. Employee Behavioural Adaptation: Dimensions and Organisational Correlates

Employee behavioural adaptation (EBA) in the hospitality context refers to the capacity and willingness of frontline workers to modify their interpersonal behaviour, communication patterns and service approach in response to the cultural, technological and organisational demands of their environment. This construct draws on the cross-cultural adjustment literature associated with Black and Stephens (1989), further developed by Shaffer et al. (2021) into a validated scale covering professional and social dimensions of adjustment. In the present study, EBA addresses three dimensions: interpersonal adaptation (adjusting communication and courtesy norms for diverse guests), professional adaptation (integrating global service protocols and digital tools into daily routines) and cultural maintenance (retaining local warmth and authenticity within standardised service delivery).

The third dimension is the most theoretically interesting, and the most difficult to train. Most behavioural adaptation research treats cultural maintenance as a separate variable—something that can be traded off for global standardisation. In heritage hotels, that trade-off is both commercially costly (local authenticity is a core product differentiator) and ethically fraught (it instrumentalises employees’ cultural identity). EBA, as conceptualised here, treats cultural maintenance not as an adaptation obstacle, but as an adaptation resource. Employees who retain their cultural confidence may be more effective adapters because they are not performing a foreign identity—they are selectively translating their own.

Ashill et al. (2005), studying service recovery performance in a public healthcare setting, found that employee behavioural flexibility was associated with lower turnover intentions and stronger service recovery outcomes. Since this evidence comes from a healthcare context rather than a hospitality context, it is cited here as suggestive cross-sector evidence, rather than as a hospitality-specific finding, and readers should weigh it accordingly. Nair and Vohra (2012), in a conceptual clarification of the work-alienation literature, argue that alienation—the psychological state that results when employees cannot reconcile their values with their institutional role—is likely most damaging in service work where the emotional labour content is high. These sources converge on the plausibility, rather than the demonstrated certainty, that EBA carries organisational consequences beyond individual performance. If adaptable employees handle cross-cultural encounters with less strain and greater flexibility, one plausible organisational expression of that capacity in a heritage hotel is the degree to which employees perceive global standards and local authenticity as combined in the service they deliver. That reasoning underpins the second hypothesis:

Hypothesis Two: Employee behavioural adaptation positively predicts employee-perceived global-local integration in heritage-based hotels in Baghdad.

2.4. Employee-Perceived Global-Local Integration and the Mediating Role of Behavioural Adaptation

Employee-perceived global-local integration (GLI) in this study refers to frontline employees’ and supervisors’ own assessment of the degree to which international operational standards and local cultural authenticity are combined in their hotel’s service delivery, rather than treated as competing imperatives. GLI is measured, analyzed, and interpreted here at the individual respondent level. It describes how integration is perceived by the people delivering the service, not an independently verified organisational-level outcome. This construction is informed by research on cultural and social experience in hotel service settings (Ogunnaike et al., 2022) and by broader work on digitalisation and organisational change management (Aruppala et al., 2026). It should be noted that neither source specifically addresses heritage tourism, and both are cited here for their general relevance to service-culture and digital-adoption dynamics, rather than as heritagespecific evidence. GLI, as perceived by employees, is not simply the coexistence of global and local elements; it captures whether employees experience service as simultaneously efficient by international benchmarks and rooted in local hospitality tradition.

The argument for behavioural adaptation as a mediating mechanism is both theoretical and empirical. Theoretically, SET implies that the association between leadership and more distal outcomes is often carried in part by proximal employee responses—attitudes, behaviours and cognitions that translate leadership input into downstream perceptions. Empirically, prior work in adjacent literatures (Edmondson, 1999; Aruppala et al., 2026) is broadly consistent with employee-level variables accounting for a substantial portion of leadership-to-outcome associations in service organisations, though that literature was not developed in heritage-hotel or cross-cultural-adaptation settings specifically.

Recent hotel-based evidence from the Middle East region reinforces this architecture across otherwise different construct families. In Jordanian five-star hotels, work engagement mediated the association between transformational leadership and employee performance (Helalat et al., 2025). In Dubai’s four- and five-star properties, green empowerment carried much of the association between green transformational leadership and employees’ green behavioural intentions (Alshimmari & Shneikat, 2025). An integrative review of the leadership–well-being literature likewise identifies employee resilience as a recurring mediating mechanism between leadership styles and employee outcomes (Shabalala & Chitamba, 2025). The common pattern in these studies is that leadership relates to outcomes largely through proximal employeelevel states and behaviours. In the present model, EBA is treated as that proximal response, through which culturally intelligent leadership relates to employee-perceived GLI.

This reasoning leads to the third hypothesis:

Hypothesis Three: Employee behavioural adaptation mediates the relationship between cross-cultural leadership and employee-perceived global-local integration, such that the indirect effect of CCL on GLI through EBA is statistically significant.

3.1. Research Design and Sampling

This study used a quantitative, cross-sectional survey design. The target population comprised frontline employees (receptionists, concierge staff, food and beverage personnel) and mid-level supervisors in hotels formally classified as heritage or historical-theme properties within Baghdad’s administrative boundaries. That classification was based on the Iraqi Ministry of Tourism and Antiquities register, which at the time of data collection listed eleven such hotels. Nine hotels agreed to participate. Purposive sampling was used to ensure that respondents had at least six months of tenure in their current position—a threshold designed to ensure that participants had sufficient experience to evaluate leadership behaviour meaningfully and to have undergone at least one full cycle of seasonal operational adaptation.

Three design decisions deserve explicit justification. The first decision concerns the purposive sampling strategy itself. The population of interest here is not hotel employees in general but employees of a specific and rare organisational form: officially registered heritage hotels operating in Baghdad. That population is small and fully enumerable—eleven properties on the ministry register at the time of fieldwork. Probability sampling from a wider hotel population would have diluted the very characteristic the study set out to examine. Purposive selection from the official register, combined with the six-month tenure criterion, was therefore the sampling logic best aligned with the research questions.

The second decision requiring justification concerns geographic scope. Baghdad hosts the large majority of Iraq’s registered heritage properties, and concentrating fieldwork there meant the study approached a census of its target population rather than a narrow slice of a broader one. By the participating hotels’ own staffing figures, the 380 respondents cover roughly two-thirds of the frontline and supervisory workforce employed in the city’s heritage hotel sector. The findings are accordingly interpreted as statements about that bounded population, and their extension beyond it is treated as an empirical question for comparative research (Section 7).

Third, questionnaires were allocated to hotels in proportion to staffing levels, so that the largest properties would not dominate the sample. On-site administration during shift changes was chosen deliberately to reach shift workers, who are systematically underrepresented in online and mail-back surveys.

A total of 450 questionnaires were distributed across the nine participating hotels, with each hotel receiving a number proportional to its staffing level. 380 valid responses were returned, representing an 84.4% response rate. The high return rate reflects the value of the on-site administration procedure: research assistants visited each hotel during morning and afternoon shift changes, distributing and collecting questionnaires within a single visit to minimise non-response from shift-based attrition. The required sample size was determined a priori using established PLS-SEM minimum sample-size guidance for models of this complexity (three latent constructs, a maximum of six indicators per construct; Hair et al., 2022). The achieved sample of 380 comfortably exceeds the minimum benchmark. As a supplementary, retrospective diagnostic, rather than as the basis for the original sample-size decision, post-hoc power analysis (G*Power 3.1) indicated high achieved power (1 ‒ β = 0.97) at α = 0.05 for a medium effect size (f² = 0.15).

3.2. Measurement Instruments

All three constructs were measured using established scales adapted for the Iraqi hospitality context. The Cross-Cultural Leadership (CCL) scale drew primarily on the cultural intelligence leadership survey developed by Ang et al. (2007) and subsequently used in hospitality settings by Bukhari et al. (2025). The final scale comprised six items targeting behavioural manifestations of cultural intelligence rather than attitudinal or cognitive dimensions, reflecting the study’s interest in observable leadership practices.

The EBA scale was adapted from the cross-cultural adjustment literature associated with Black and Stephens (1989). The scale was then modified following the adjustment scale-development approach of Shaffer et al. (2021) to capture the three dimensions described above: interpersonal adaptation, professional adaptation and cultural maintenance. This instrument produced five items after initial piloting and factor analysis.

The Global-Local Integration (GLI) scale was developed inductively, drawing on Ogunnaike et al.’s (2022) work on cultural and social experience in hotel service, and supplemented by items informed by Aruppala et al.’s (2026) review of digitalisation and organisational change. Five items were retained after piloting. All items were rated on a five-point Likert scale (1 = Strongly Disagree to 5 = Strongly Agree).

Scale adaptation followed a structured sequence rather than ad hoc rewording. Items from the source instruments were first screened for contextual fit; wording that presupposed multinational corporate settings was rewritten to reflect concrete hotel service situations. A five-member expert panel—two hospitality management academics, two senior managers from non-participating Baghdad hotels and one specialist in Arabic-language survey methodology—then reviewed the item pool for content validity, clarity and cultural appropriateness. The expert review led to the removal of two redundant CCL items and the rewording of three EBA items whose original phrasing translated awkwardly into Iraqi Arabic. Only after this stage did translation begin.

The survey was translated from English into Arabic using the parallel back-translation method, with two bilingual translators working independently and a third reconciling discrepancies. A pilot study on 40 participants not included in the final sample confirmed item clarity and scale stability. The full wording of all sixteen retained items, together with their sources and response anchors, is reported in Appendix A.

3.3. Common Method Variance Controls

Given the single-source nature of the data, both procedural and statistical controls were applied to address the risk of Common Method Variance (CMV). Procedurally, the survey physically separated the predictor and outcome items with an unrelated filler section, respondent anonymity was assured and the questionnaire instructions stated explicitly that there were no correct or incorrect responses. Statistically, Harman’s single-factor test was applied first as a coarse screening diagnostic. The unrotated principal component analysis showed that the largest single factor explained 31.4% of total variance, below the 50% threshold at which CMV becomes a critical concern, though this test alone is not treated as conclusive.

Following Kock (2015), a full collinearity assessment was also conducted at the construct level. Each latent construct was regressed, in turn, on the remaining constructs in a saturated structural model (i.e., all three constructs interconnected rather than only the hypothesised paths). The resulting inner-model variance inflation factors (VIFs) were evaluated against the conservative 3.3 threshold. The VIF values reported in Table 4 are indicator-level (outermodel) VIFs, which assess multicollinearity among the indicators of each construct and are conceptually distinct from the construct-level full collinearity VIFs that Kock’s procedure requires; both are reported for completeness. The construct-level full collinearity VIFs, as shown in Table 1, were CCL = 1.45 (regressed on EBA and GLI), EBA = 1.82 (regressed on CCL and GLI), and GLI = 1.63 (regressed on CCL and EBA). All three values fall well below the 3.3 threshold, a pattern inconsistent with pervasive CMV. Taken together, the procedural separation of predictor and criterion items, the anonymity assurances and the two statistical diagnostics (Harman’s single-factor test and the construct-level full collinearity assessment) suggest that CMV is unlikely to fully account for the relationships reported below. However, it should be noted that a single-source, cross-sectional design cannot rule it out entirely, a point returned to in Section 7.

3.4. Analytical Approach

Data were analyzed using SmartPLS 4 (Ringle et al., 2022) following the two-stage analytical procedure recommended by Hair et al. (2022). Stage one evaluated the measurement model through confirmatory factor analysis, assessing indicator loadings, internal consistency reliability (Cronbach’s α and Composite Reliability (CR)), convergent validity (Average Variance Extracted (AVE)) and discriminant validity using both the Fornell-Larcker criterion and the Heterotrait-Monotrait (HTMT) ratio. The latter ratio was estimated with 5,000 bootstrap resamples so that confidence intervals could be inspected against the 0.85 threshold (Henseler et al., 2015). Stage two evaluated the structural model, reporting path coefficients (β), bootstrap standard deviations (STDEV), t-statistics from bootstrapping (5,000 resamples), and 95% bias-corrected confidence intervals (CI). Effect sizes (f²) and predictive relevance (Q² via blindfolding, omission distance d = 7) were also computed.

3.5. Ethical Considerations

This study did not undergo formal review by an institutional ethics committee. No such committee reviews anonymous, non-clinical, voluntary organisational survey research of this kind at the authors’ home institution (University of Mustansiriyah, Baghdad), and no applicable Iraqi regulation required formal ethical review for a study of this design. The research is administrative and organisational in nature; it involved no clinical procedures, no experimental manipulation and no vulnerable populations. Data were collected exclusively through an anonymous, self-administered questionnaire that gathered no names, contact details, identification numbers or other information capable of revealing respondents’ identities, directly or indirectly. Approvals at the respondent level were oral. Before questionnaires were distributed, research assistants explained the academic purpose of the study and assured respondents of anonymity. Each employee then retained complete freedom either to complete the questionnaire or to decline it. Participation was entirely voluntary, respondents could stop or return an incomplete questionnaire at any point without giving a reason and without any consequence, and completing and returning the questionnaire constituted the respondent’s agreement to take part. No written consent form was used, a practice consistent with the anonymous, voluntary character of the survey and with the absence of any identifying or sensitive personal data. Hotel management at each participating property granted written permission for on-site data collection before fieldwork began.

4.1. Sample Description

The sample skews moderately toward male respondents (63.2%), reflecting the still male-dominant gender composition of Baghdad’s hotel sector. The majority of respondents (43.4%) were in the 26–35 age bracket, suggesting a workforce in the early-to-mid career phase. A combined 44.7% of respondents had six or more years of hotel experience, which provides a degree of professional anchoring to the adaptation assessments. The educational distribution is notable: just over half held a bachelor’s degree (50.5%), indicating a relatively educated workforce.

This demographic profile is reported descriptively. As noted in Section 1, formal statistical comparisons of EBA or GLI across gender, age, experience or education groups were not part of the pre-specified analytic plan for this study and are not reported here. A proposal for such analyses is discussed as a direction for future research in Section 7.

4.2. Descriptive Statistics

Construct means range from 3.62 to 3.92, indicating moderately positive perceptions across all three constructs. The lowest mean (GLI3: 3.62) pertains to reconciling digital protocols with local customs, consistent with the qualitative insight from pilot interviews that digital systems imported from international chains often carry implicit cultural assumptions incompatible with local service norms. The highest mean (CCL1: 3.92) indicates that employees generally perceive their leaders as adapting communication style to cultural context, although the moderate standard deviation (0.71) indicates meaningful variation across hotels.

4.3. Measurement Model Evaluation

All indicator loadings exceeded 0.70, with the lowest loading being EBA2 (0.742) and the highest CCL1 (0.825). Cronbach’s α values ranged from 0.841 (EBA) to 0.864 (CCL), all above the conventional 0.70 threshold. CR values ranged from 0.880 to 0.895. AVE exceeded 0.50 for all three constructs—CCL (0.632), EBA (0.596), and GLI (0.624)—establishing adequate convergent validity. Indicator-level VIF values ranged from 1.61 to 1.84, well below the threshold of 3.3, indicating that multicollinearity among indicators within each construct does not pose a threat to the measurement model.

Discriminant validity was examined in two stages. Table 5 reports the Fornell-Larcker criterion: the square root of the AVE for each construct (CCL = 0.795; EBA = 0.772; GLI = 0.790) exceeds its correlation with every other construct. Because the Fornell-Larcker criterion can be insensitive under certain measurement conditions, the HTMT ratio was assessed separately and is reported in full in Table 6, together with 90% bootstrap confidence intervals from 5,000 resamples. All three HTMT values (CCL–EBA = 0.598; CCL–GLI = 0.531; EBA–GLI = 0.661) fall clearly below the conservative 0.85 threshold, and no confidence interval upper bound approaches unity (i.e., the boundary at which two constructs would become statistically indistinguishable). Discriminant validity is therefore established under both criteria.

4.4. Structural Model Evaluation and Hypothesis Testing

The structural model results support all three hypotheses. Hypothesis One is supported because CCL is positively associated with EBA (β = 0.543, STDEV = 0.036, t = 15.128, p < 0.001), with a large effect size (f² = 0.421), indicating that CCL accounts for a substantial share of the explained variance in EBA within this model (R² = 29.4%). Hypothesis Two is supported, as EBA is positively associated with GLI (β = 0.470, STDEV = 0.044, t = 10.561, p < 0.001), with a medium-to-large effect size (f2 = 0.317). The model explains 42.1% of the variance in GLI within the estimated model (R² = 0.421). This figure describes the explained variance conditional on the specified model and should not be read as an estimate of the construct’s full causal determinants.

Hypothesis Three is supported by the mediation analysis reported in Table 9. The specific indirect effect of CCL on GLI through EBA is β = 0.255 (p < 0.001, 95% bias-corrected cultural intelligence reported in Table 9), a value that equals the product of the component paths (0.543 × 0.470 = 0.255). This indirect effect is statistically significant, and the direct effect of CCL on GLI, while also significant (β = 0.183, t = 3.895, p < 0.001), is smaller than the indirect path. The Variance Accounted For (VAF = 58.2%) indicates that EBA accounts for the majority of the statistical association between CCL and GLI, a pattern consistent with partial mediation. CCL is associated with GLI both directly and indirectly through the behavioural responses it appears to elicit from employees. The Q² values for both EBA (0.168) and GLI (0.254) are above zero, indicating in-sample predictive relevance under blindfolding. This diagnostic speaks to predictive relevance within the estimated sample and does not by itself demonstrate out-of-sample predictive performance. The standardised root mean square residual (SRMR) of 0.047 is below the conventional 0.08 threshold. Consistent with guidance for PLS-SEM, this is reported as one supporting fit diagnostic, rather than as definitive evidence of overall model validity (Henseler et al., 2016).

One qualification applies to the language used throughout this section. Path coefficients estimated from cross-sectional data describe directional associations specified by theory, not demonstrated causal effects. Terms such as ‘predicts’ and ‘effect’ are used here in their statistical, model-relative sense, and the causal status of the model—what these data can and cannot establish—is addressed directly in Sections 5 and 7.

5.1. Interpreting the Association between Cross-Cultural Leadership and Behavioural Adaptation

The association between CCL and EBA (β = 0.543) is the central finding of this study and deserves careful interpretation, rather than simple endorsement. The effect size is large—f² = 0.421 sits above Cohen’s (1988) threshold for a large effect—meaning that leadership behaviour is statistically linked to a comparatively high share of variance in employee adaptation within this single-predictor model. One plausible reading of this pattern is that employees in Baghdad’s heritage hotels are unusually dependent on the quality of their immediate leadership to navigate the cultural demands of their work. Where leadership is culturally responsive, employees appear to adapt more readily. The standard deviation in CCL scores suggests meaningful variance across hotels in how consistently this support from leadership is provided.

How does this estimate sit against the wider literature? Direct comparison is complicated by differences in construct operationalisation, and the available comparator studies are not always a precise match to the present context. Bukhari et al. (2025) reported a positive association, of smaller reported magnitude, linking employees’ cultural intelligence to constructive workplace behaviour through engagement in hotels. Resho and Singh (2025) likewise found cultural intelligence operating as a foundational competency for global leadership and effective communication among hospitality managers in India. The present estimate is comparatively large next to those benchmarks. A plausible readingis contextual: Where institutional supports for cross-cultural work are thin—eg., no corporate training academies, no expatriate mentoring structures, no accumulated organisational routines—the immediate leader may become a more dominant channel through which employees acquire cross-cultural competence. This interpretation should be treated as a hypothesis generated by the present data rather than as an established finding, pending replication.

This interpretation is broadly consistent with Edmondson’s (1999) evidence on psychological safety as a mechanism relevant to team learning behaviour. Employees who do not feel that their leader will protect their cultural identity when they stumble in a cross-cultural interaction may be less willing to take the interpersonal risks that adaptation requires. In Baghdad, where many hotel employees have little prior experience with international guests and where the institutional infrastructure for cross-cultural training is thin, the leader may function less as a performance evaluator and more as a primary source of cultural interpretation. What Bukhari et al. (2025) term ‘cultural perspective-taking’ becomes, in this reading, a concrete managerial practice by helping employees hold two cultural frames simultaneously without collapsing one into the other.

5.2. Behavioural Adaptation as a Candidate Mechanism

The mediation finding—VAF = 58.2%—invites a reconsideration of how practitioners understand EBA. A common assumption in HR development is that adaptation is an individual trait that can be screened for at hiring or developed through generic diversity-training workshops. The present data are more consistent with a picture in which behavioural adaptation is substantially related to the organisational environment the leader creates, although the cross-sectional design cannot establish this as a causal claim. If this pattern holds under future longitudinal testing, it would suggest that training employees in cross-cultural sensitivity, without simultaneously developing their leaders’ cross-cultural competence, may produce limited returns. Employees who lack leadership scaffolding to practice adaptive behaviours may then revert to whichever default serves survival in the short run—rigid adherence to global protocols, or informal retreat into local-only interaction patterns.

This finding is broadly consistent with Ashill et al.’s (2005) evidence on behavioural flexibility and service outcomes —while noting again that their evidence is drawn from a healthcare rather than a hospitality setting. Furthermore, this finding identifies CCL as a plausible antecedent condition of that flexibility in the present sample—consistent with, but not proof of, a causal relationship. It also resonates with Edmondson and Bransby’s (2023) observation that psychological safety is most consequential when work requires interpersonal risk-taking. Service delivery in a heritage hotel, where every guest interaction is potentially both a commercial transaction and a cultural encounter, plausibly fits this risk-taking situation.

5.3. Partial Mediation and the Dual Channel of Leadership Influence

The partial mediation result—a significant indirect effect alongside a significant, smaller direct effect— deserves attention in its own right. The direct path from CCL to GLI (β = 0.183) indicates that CCL is associated with employee-perceived GLI through channels that do not appear to require behavioural adaptation as the mechanism. Two interpretations are plausible. First, culturally intelligent leaders may directly shape organisational-level practices (scheduling, service script design and physical space use) in ways that independently relate to employees’ perception of integration, without the association passing through individually observed behaviour. Second, the direct effect may capture a collective-level process—the way leadership behaviour establishes a perceived cultural climate— that is conceptually distinct from EBA. Because GLI is measured here at the individual, employee-perceived level and the study did not collect the within-group agreement or intraclass correlation statistics that would justify aggregating it to the hotel level, neither interpretation can be adjudicated with the present data. Both interpretations point toward the need for multilevel research designs, ideally with a larger number of hotel-level clusters than the nine available here, that can separate individual-level adaptation effects from organisational-level climate effects.

5.4. Consistencies and Divergences with Prior Research

Set against the existing body of work, the findings show two broad consistencies and one instructive divergence. Consistent with the cross-cultural adjustment literature, associated with Black and Stephens (1989) and its later development by Shaffer et al. (2021), adaptation here behaves as a multidimensional capacity that appears responsive to relational support, rather than as a fixed disposition. Consistent with the general pattern reported by Ashill et al. (2005), adaptive employees appear organisationally consequential in this sample. Adaptation carried the majority of the statistical association between leadership and integration, which is what one would expect if adaptability related to concrete service behaviour, rather than remaining a private attitude.

The mediation architecture observed here also mirrors recent hospitality evidence from the wider region. Work engagement carried the association between transformational leadership and employee performance in Jordanian five-star hotels (Helalat et al., 2025), and green empowerment played the equivalent role for employees’ proenvironmental intentions in Dubai’s four- and five-star properties (Alshimmari & Shneikat, 2025). That the same proximal-mechanism pattern emerges in Baghdad, with behavioural adaptation as the carrier, suggests the underlying architecture travels across construct families in the region’s hotel sector, even as the content of the mediator changes. This reading is also consistent with review-level evidence that employee-level psychological resources recur as mediating mechanisms between leadership and employee outcomes (Shabalala & Chitamba, 2025).

The divergence concerns cultural maintenance. Much of the adjustment literature implicitly treats retention of local cultural practice as friction to be overcome on the road to global competence. In these data, the cultural maintenance item (EBA5) loaded strongly on the same factor as protocol integration (EBA4: 0.756; EBA5: 0.773), and the construct as a whole predicted integration out-comes within the model. In heritage settings, retention and adaptation look like complements rather than substitutes in this sample. This reading is broadly consistent with Nair and Vohra’s (2012) conceptual account of alienation, under which employees compelled to abandon identity-relevant practices may disengage, whereas those permitted to translate them may invest more. Whether the complementarity holds outside heritage contexts— where authenticity is not a product attribute and cultural maintenance carries no commercial premium—is an open empirical question.

5.5. Baghdad as a Theoretically Productive Site

Most of the intellectual machinery in CCL research was built in North American, Northern European, and East Asian contexts. Testing it in Baghdad is not simply a matter of adding geographic diversity to a literature that could always use more. Baghdad represents a qualitatively different kind of research site: a city with a historically dense cultural heritage, a recent institutional trauma from decades of conflict and a hospitality sector in active reconstruction. The fact that the SET-based model performs as expected in this context—statistically significant paths in the hypothesised directions, meaningful explained variance, acceptable measurement properties—suggests that the core theoretical logic is not culturally parochial. What appears to differ is the specific cultural content of the exchange, the depth of the identity stakes involved and the thinness of the institutional support infrastructure. Although, this research remains one study in one city and should be read with appropriate caution.

These contextual differences do not make Baghdad an anomaly; they make it a useful stress test. If CCL is associated with behavioural adaptation in an environment where cultural pressures are high, institutional supports are limited and employees have little prior cross-cultural socialisation, the case for CCL’s practical relevance is strengthened relative to what studies conducted in well-resourced institutional environments alone would suggest.

6.1. Theoretical Contributions

This study offers three contributions to the literature. First, it extends SET into the domain of cultural heritage management. SET has been applied extensively in standard commercial service contexts; its extension into heritage institutions, where the stakes of cultural authenticity are woven into the institution’s raison d’être, opens a new application domain. The exchange logic that appears to drive EBA in this context is richer than the task-performance reciprocity typically modeled in SET research. It involves cultural identity, institutional memory and the negotiated legitimacy of an institution that derives its competitive advantage from being unlike other hotels.

Second, the study introduces behavioural adaptation as a candidate mechanism—rather than simply a dependent variable—in the leadership-to-integration pathway. This reframing has implications for how organisational behaviour researchers might think about the organisational correlates of CCL. The pattern observed here is consistent with adaptation functioning as a central pathway through which leadership relates to the perceived integration outcomes examined in this model. This is a proposition the present cross-sectional design can support statistically, but cannot establish causally, and which further longitudinal work should now test.

Third, the study provides empirical evidence from a non-Western, post-transition context that is comparatively underrepresented in the hospitality management literature. Baghdad extends the range of settings in which the core SET-based logic has been examined, while also pointing toward the contextual modifications that posttransition environments may require of the theory.

6.2. Practical Recommendations