Corporate social responsibility as a strategic resource: examining the mediating role of corporate reputation in securing competitive advantage

1 Introduction

Organizations are now judged on a broader basis than financial return. Global sustainability commitments, tighter disclosure expectations, and better-informed stakeholders have moved corporate social responsibility (CSR) from the margins of corporate philanthropy to the center of strategy formulation. The question is no longer whether firms should invest in CSR, but whether and how such investment yields something competitors cannot easily copy. That question has practical weight: CSR programs absorb managerial attention and financial resources, and boards increasingly ask what strategic return those resources generate.

Carroll (1979) was among the first to formalize CSR, and later work, such as Latapí Agudelo et al. (2019), traced its evolution into a strategic construct. Since then, a substantial body of research has linked CSR to organizational outcomes. Leclercq-Machado et al. (2022) report that CSR initiatives are often designed with competitiveness in mind, and Li et al. (2024) show that environmental, social, and governance programs have become standard vehicles for organizational development and sustainability. Dahlsrud (2008) and Aslaksen et al. (2021) document how firms have come to accept responsibility for the social and environmental conditions of the communities where they operate, while Sisodia et al. (2014) argue that durable success requires a management approach that balances all stakeholders’ interests.

Despite this accumulation of evidence, three gaps remain, and together they motivate the present study. First, the mechanism linking CSR to competitive advantage (CA) is more often asserted than modeled. A large share of prior work estimates a direct CSR–performance path. It interprets a significant coefficient as evidence of strategic value, without specifying the intervening asset through which the value is realized or holding constant the structural characteristics of the firms involved. Recent meta-analytic evidence reinforces the value of this mechanism-focused approach: pooling 98 studies, Liu et al. (2026) show that reputation is a significant channel through which CSR shapes firm performance. Yet this evidence centers on financial and non-financial performance rather than on competitive advantage as a distinct, resource-based outcome, and it draws only sparingly on GCC data. Second, stakeholder theory and the resource-based view (RBV) are routinely invoked side by side. Yet, the two are rarely integrated into a testable statement of how stakeholder approval translates into an inimitable resource position. Third, the evidence base is geographically uneven. Research on CSR in the GCC region, and in the United Arab Emirates (UAE) in particular, remains largely descriptive; Al-Jenaibi (2017) found that UAE organizations frequently endorse CSR principles without embedding them in management practice, which makes the UAE an informative setting in which to test whether declared responsibility translates into reputational and competitive returns.

The UAE’s policy environment heightens the urgency of these gaps. National sustainability and economic diversification agendas have placed measurable social and environmental obligations on firms across sectors, while the domestic market is open and highly competitive. In such a setting, organizations that cannot demonstrate a strategic return on responsible conduct face pressure to reduce it to a communication exercise, a practice that erodes rather than builds reputational capital.

Accordingly, this study examines the relationships among CSR, CR, and CA, and tests whether CR conveys the association between CSR and CA. The following research question guides the investigation: How is corporate social responsibility associated with a firm’s competitive advantage, and does corporate reputation mediate this association?

The study makes three contributions. Theoretically, it specifies and tests reputation as the conversion mechanism linking stakeholder-oriented conduct (stakeholder theory) to a resource position that is valuable, rare, and costly to imitate (RBV), moving beyond parallel citation of the two perspectives toward an integrated and falsifiable account. Empirically, it shows that reputation fully carries the association between CSR and competitive advantage. Once reputation is accounted for, the direct link between CSR and competitive advantage no longer holds, which indicates that CSR translates into advantage through the reputation it builds rather than on its own. Contextually, it supplies quantitative evidence from a GCC economy in which CSR is institutionally encouraged but unevenly implemented.

The remainder of the paper is organized as follows. Section 2 develops the theoretical background and reviews the relevant literature. Section 3 derives the hypotheses. Section 4 sets out the research design, sampling procedure, measures, and analytical strategy. Section 5 reports the descriptive, measurement, and hypothesis-testing results. Section 6 discusses the findings and their implications, and Section 7 concludes.

2 Theoretical background and literature review

2.1 Stakeholder theory

Stakeholder theory, introduced by Freeman (1984), holds that businesses have responsibilities not only to their shareholders but also to the broader set of parties that affect or are affected by the firm. The organization depends on these parties to function. Hartman et al. (2011) explain that the stakeholders with whom a company interacts are a source of value and play a significant role in shaping its CSR and the relationships it builds. Stojanović et al. (2025) argue that socially responsible practice helps an organization meet stakeholder expectations, and Al-Alawi et al. (2023) note that many organizations pursue CSR to attract employees who share their values and to strengthen ties with other constituencies.

For this study, the theoretically important implication is narrower. Stakeholder theory explains why responsible conduct generates approval: when a firm attends to the legitimate claims of employees, customers, regulators, and communities, those groups revise their evaluations of the firm upward. Stakeholder theory does not, by itself, explain why rivals should find such approval difficult to replicate, or why it should persist long enough to matter strategically. Answering that question requires a resource-based argument.

2.2 The resource-based view and CSR as a strategic intangible resource

The RBV, developed by Barney (1991), holds that firms achieve competitive advantage by deploying resources that are valuable, rare, imperfectly imitable, and non-substitutable. Such resources may be human, tangible, or intangible. Aldabbas and Gianni (2026) show that firms outperform rivals by exploiting resources already available internally, while Al-Essa et al. (2022) argue that organizational performance depends on the knowledge and capabilities held by the firm and its stakeholders and on the firm’s ability to deploy them. This strategic reading of CSR is echoed in recent work: integrating stakeholder theory with the RBV and dynamic capabilities, Zhang et al. (2026) show that CSR has moved beyond a compliance exercise to function as a strategic asset that firms configure to secure competitive advantage.

CSR qualifies as a strategic resource under these criteria, but only under specific conditions; stating those conditions converts a general claim into a testable one. CSR is valuable when it reduces stakeholder-related risk and secures continued access to labor, capital, and resources. It is rare for it to exceed regulatory minimums and prevailing industry practice rather than merely match them. It is costly to imitate because it is socially complex and path dependent: an authentic responsibility profile is accumulated over years of consistent conduct and cannot be purchased or announced into existence. It is also non-substitutable because no alternative asset produces the same configuration of stakeholder trust. Branco and Rodrigues (2006) advance precisely this argument, treating CSR as a source of internal capabilities and external reputational assets, and Hart (1995) makes the parallel case for environmentally grounded capabilities. It follows that CSR’s strategic value does not lie in the expenditure itself, which competitors can match, but in the accumulated stakeholder perception it produces.

This is where reputation becomes a theoretical necessity rather than an additional variable. Fombrun and Shanley (1990) describe reputation as a signal market participants construct from observable firm conduct, and Roberts and Dowling (2002) provide evidence that superior reputation is associated with sustained superior performance and with slower erosion of that performance over time. Reputation is therefore the asset in which CSR conduct is stored: it is socially complex, causally ambiguous, and accumulated rather than acquired, which is to say it satisfies the RBV’s imitability condition in a way that CSR spending alone does not. Freeman et al. (2021), who explicitly bridge stakeholder theory and the RBV, support this reading. The integration adopted here is thus straightforward: stakeholder theory explains how responsible conduct earns approval.

2.3 Corporate social responsibility

CSR is part of organizations’ contribution to the United Nations 2030 Agenda. However, CSR initiatives may fail to deliver meaningful outcomes when they lack coordinated action across organizations and industries (Zenda et al., 2025). The construct rests on the premise that organizations have obligations to the societies in which they operate. Hartman et al. (2011) note that society expects organizations to go beyond financial objectives, and Matten and Moon (2008) characterize CSR as a commitment to ethical standards, community development, and minimizing environmental harm. Both perspectives require organizations to recognize their social and environmental footprint and act on it.

One notable trend is the growth of purpose-driven organizations. Porter and Kramer (2011) describe how firms create value for society by addressing social problems through their core business, and Ante and Sutrisno (2025) observe that consumers now evaluate organizations on their societal contribution as well as on product quality. Organizations that commit credibly to CSR therefore tend to enjoy stronger consumer support. Engaging in social responsibility has also been argued to strengthen a firm’s image and reputation and, in turn, to provide a competitive advantage in dynamic markets (Blaique et al., 2023).

2.4 Corporate reputation

CR refers to stakeholders’ aggregate perception of a company. Meirinhos et al. (2022) state that reputation is built from a company’s past interactions with the public and the impressions those interactions leave, and that it also reflects how the firm communicates information about itself. Fombrun et al. (2015) similarly treat reputation as a multidimensional evaluative judgment that can be measured and tracked over time.

Companies with weak reputations typically face lower demand and greater difficulty attracting the talent needed to operate. Yikilmaz et al. (2025) note that organizations can adopt a range of strategies to repair and strengthen their reputation, thereby reducing these difficulties. Reputation is best understood as an accumulated asset rather than a communication output: it is built slowly through consistent conduct, including CSR, and can be depleted rapidly.

2.5 Competitive advantage

Competitive advantage refers to the attributes that differentiate an organization from its rivals and enable it to outperform them. Porter (1985) explains that competitive advantage arises from the strategies and resources an organization uses to deliver superior customer value. Agustian et al. (2023) identify innovation, operational efficiency, market positioning, and research and development capability as components of competitive advantage, and Aldabbas and Oberholzer (2024) show that technological and operational systems shape market success. Together, these elements determine performance, market position, and responsiveness to environmental change, making the maintenance of competitive advantage central to long-term organizational viability.

3 Hypotheses development

3.1 CSR and competitive advantage

Competitive advantage is both an outcome of and a contributor to an effective CSR strategy. Agustian et al. (2023) argue that embedding CSR into operations improves reputation and organizational performance, both of which support firm success. From an RBV perspective, CSR initiatives constitute a resource that helps establish and maintain advantage, provided the resulting resource meets the criteria of value, rarity, inimitability, and non-substitutability.

To compete effectively, a firm must consider factors beyond economic performance. Carroll (2016) outlines four responsibilities: economic responsibility to stakeholders and society, legal responsibility to meet regulatory requirements, and ethical and philanthropic responsibilities that extend beyond what law and economics require. Empirical research supports the strategic relevance of these commitments. Al-Shammari et al. (2022) find that CSR initiatives increase competitiveness and stakeholder trust, and Famiyeh (2017) reports improved business performance among firms that implement them.

Because socially responsible conduct signals to stakeholders that a firm is trustworthy and well managed, and because such favorable evaluations tend to translate into stakeholder support that strengthens the firm’s market position, CSR is expected to be positively associated with competitive advantage.

Hypothesis 1. CSR is positively associated with competitive advantage.

3.2 CSR and corporate reputation

CSR has become an integral component of corporate strategy. Cruz et al. (2026) report that implementing CSR helps organizations strengthen their brand, differentiate their offerings, and retain customers. Reputation is an intangible asset that contributes directly to organizational success: organizations with positive reputations attract more customers and broaden their stakeholder base. Esenyel (2020) states that a positive reputation enhances competitiveness by earning stakeholder trust.

Empirical evidence for a CSR–reputation link is well established. Lin-Hi and Blumberg (2018) used data from more than 200 organizations to show that CSR practices directly strengthened reputation. Brammer et al. (2007) demonstrate that CSR initiatives improve standing within the industry and among the stakeholders who contribute to firm success. Ayada and Abuelela (2023) show how CSR communication enhances public perceptions of companies. Demeke and Ravi (2024) report that the various dimensions of CSR allow organizations to improve corporate image and secure stakeholder confidence.

The theoretical rationale is signaling-based and grounded in stakeholder theory. Stakeholders cannot directly observe a firm’s underlying values, so they infer them from visible conduct. Sustained CSR activity is costly and therefore credible, and its credibility increases when the activity is verifiable, consistent over time, and connected to the firm’s core operations rather than peripheral to them. In digitally mediated markets, where stakeholders continuously observe and circulate evidence of corporate conduct, the informational conditions for such signaling are unusually favorable, implying that inconsistency between message and practice is quickly detected and penalized. Taken together, this suggests that firms that act responsibly, visibly, and consistently accumulate more favorable stakeholder evaluations; that is, CSR is expected to build corporate reputation.

Hypothesis 2. CSR is positively associated with corporate reputation.

3.3 Corporate reputation and competitive advantage

Reputation is widely recognized as an organizational asset with material consequences for firm success. Fombrun et al. (2015) show that reputation can be systematically measured and that strong reputations support strategic positioning and long-term viability. Porter and Kramer (2011) argue that shared-value activity allows organizations to build reputation and competitiveness simultaneously, and Du et al. (2015) find that CSR gives stakeholders opportunities to form favorable perceptions of a firm’s operations and values. Conversely, a damaged reputation constrains growth (Hartman et al., 2011).

Further evidence links reputation to competitive outcomes. Eyasu and Arefayne (2020) studied employees across branches of Lion International Bank in Ethiopia and found that CSR dimensions positively influenced elements of competitive advantage. Roberts and Dowling (2002) provide the most direct evidence for the present argument, showing that firms with superior reputations sustain superior performance longer than firms without them.

Reputation confers advantage through three simple routes. It lowers transaction costs, because partners trust a well-regarded firm more and need fewer safeguards. It creates a preference premium, because stakeholders choose the better-regarded firm when options are otherwise similar. And it offers insurance value, because built-up goodwill cushions the firm when problems arise. Rivals cannot copy these quickly, because reputation is earned through conduct over time rather than bought at a point in time. This leads to the expectation that a stronger reputation is associated with greater competitive advantage.

Hypothesis 3. Corporate reputation is positively associated with competitive advantage.

3.4 The mediating role of corporate reputation

The central purpose of this investigation is to determine the role reputation plays in the relationship between CSR and competitive advantage. Hartman et al. (2011) note that embedding responsible practices aligns with the strategic objectives of a firm’s CSR efforts. Yikilmaz et al. (2025) find that organizations with weak reputations experience lower demand and higher customer dissatisfaction than those known for quality. Budirachman et al. (2026) show that reputation is the key mechanism through which technological, organizational, and ethical capabilities translate into superior performance.

Firms seeking to meet consumer expectations and establish themselves as strong competitors are often encouraged to embed CSR into their operations and supply chains. Baron (2007) notes that the effect may vary by sector, and Velte (2026) reports that organizations that implement and disclose CSR initiatives are more successful in securing stakeholder trust.

Broader evidence supports this reasoning. In a recent meta-analysis of 98 studies, Liu et al. (2026) found that corporate reputation is a key channel through which CSR translates into stronger firm performance. Building on the previous argument, we formulate the final hypothesis:

Hypothesis 4. Corporate reputation mediates the association between CSR and competitive advantage.

Figure 1 presents the conceptual framework illustrating the direct and indirect linkages among CSR, CR, and CA. Firm age, organization size, and organization type are included as control variables in both structural equations.

4 Materials and methods

4.1 Research design

The study adopts a quantitative, non-experimental, cross-sectional survey design with a correlational and explanatory purpose. This design was selected for three reasons. First, the constructs of interest, CSR, CR, and CA, are perceptual and evaluative, and are conventionally captured through structured self-report instruments completed by informants with organization-wide visibility. Second, the objective is to estimate the strength and direction of associations among latent constructs and to test an intervening mechanism, which requires variance across a reasonably large number of organizations rather than depth within a few. Third, no manipulation of the independent variable is feasible in a field setting. The design supports inference about association and about the pattern of indirect effects; it does not license causal claims, and the findings are reported accordingly throughout.

4.2 Population, sampling frame, and data collection procedure

A survey instrument based on established scales was administered to business managers in the United Arab Emirates via electronic forms (Google Forms) in February 2025. The target population comprised managers and senior executives employed by organizations registered and operating in the UAE across the hospitality, banking, education, healthcare, government, and other service and industrial sectors. Because no consolidated public register of UAE managers exists to build an exhaustive sampling frame, and because organization-level contact information is not centrally available, probability sampling was not feasible. Therefore, convenience sampling was used, and the procedure is stated explicitly here to ensure the boundary conditions of the results are transparent.

The procedure was as follows. The research team circulated an invitation containing an information sheet and a link to the questionnaire through professional and alumni networks, professional groups on business networking platforms, and direct contact with organizations in the emirates of Fujairah, Dubai, Sharjah, and Abu Dhabi. A single eligibility criterion was applied: the respondent had to hold a managerial or executive position that gave them visibility over their organization’s responsibility practices and competitive position. No selection criteria beyond position were imposed, and respondents could decline or withdraw at any time. Participation was voluntary and anonymous, and no incentive was offered.

The questionnaire was distributed to 200 managers, of whom 150 returned valid, complete responses. This corresponds to a 75% response rate. Fifty questionnaires were excluded because they were either not returned or returned with substantial missing data. The figure of 200 refers to questionnaires distributed, and the figure of 150 to the analytical sample. All analyses reported below are based on N = 150. Data were collected at a single point in time; therefore, the design is cross-sectional.

Statistical power was assessed a priori using G*Power 3.1 (Faul et al., 2009) for a fixed-effects multiple linear regression testing whether R² differs from zero. With five predictors (CSR, CR, and three control variables), a medium effect size of f² = 0.15, and α = 0.05, the minimum required sample sizes are 92 observations for 80% power and 138 for 95% power. The realized sample of 150 exceeds both thresholds, so the study is adequately powered to detect medium-sized effects.

Non-response bias was assessed following the extrapolation procedure of Armstrong and Overton (1977), in which late respondents are treated as proxies for non-respondents. The sample was split into early and late respondents by submission date, and independent-samples t-tests were conducted on the study variables and demographic characteristics. No significant differences were found for CSR (t = −0.68, p = 0.501), corporate reputation (t = −1.70, p = 0.092), competitive advantage (t = −1.17, p = 0.243), firm age (t = 1.34, p = 0.182), or organization type (t = 0.19, p = 0.849). A significant difference emerged for organization size (t = 3.99, p < 0.001); however, because organization size is included as a control variable in all models, the analysis already accounts for any effect associated with this difference. The absence of differences on the focal study variables (CSR, corporate reputation, and competitive advantage) suggests that non-response bias is unlikely to threaten the validity of the findings.

The University of Fujairah (UOF/REC/January 2025) granted ethical approval for the research. The participants provided their informed consent electronically to participate in this study.

4.3 Measures

All focal constructs were measured using previously validated multi-item scales anchored on a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). A five-point scale was selected for three reasons: it matches the response format used in the source instruments (Leclercq-Machado et al., 2022; Schilke, 2014), preserving comparability of psychometric properties across studies; five points are widely reported to reduce respondent fatigue and non-differentiation relative to seven- or nine-point formats in multilingual business populations such as that of the UAE; and the format is well established in prior CSR and reputation survey research, facilitating comparison of effect magnitudes.

Independent variable: corporate social responsibility. The CSR scale was adapted from the instrument developed by Leclercq-Machado et al. (2022) and comprised six items; one (CSR5) was removed during scale purification because it loaded weakly and reduced discriminant validity, leaving five items in the final scale. The original item “This bank helps to solve social problems” was revised to “This organization helps to solve social problems” to accommodate the range of industries represented in the sample. The scale demonstrated excellent internal consistency (α = 0.916).

Mediator: corporate reputation. CR was measured with a five-item scale adapted from (Taghian et al. 2015). A sample item reads, “We believe our customers are aware that our company provides excellent value for money.” Internal consistency was excellent (α = 0.922).

Dependent variable: competitive advantage. CA was measured with a three-item scale adapted from (Schilke 2014). A sample item reads, “We have gained strategic advantages over our competitors.” Internal consistency was good (α = 0.880).

Control variables. Because the sample deliberately spans organizations of different ages, sizes, and sectors, we included three structural characteristics as covariates in every estimated model. Firm age was measured on a six-point ordinal scale (1 = less than 1 year; 2 = 1–5 years; 3 = 6–10 years; 4 = 11–15 years; 5 = 16–20 years; 6 = more than 20 years). Organization size was measured by number of employees on a five-point ordinal scale (1 = 1–5; 2 = 6–10; 3 = 11–50; 4 = 51–100; 5 = more than 100). Organization type was recorded across six sector categories (hotel, bank, educational institution, healthcare, government/public sector, other) and entered as a coded indicator. These three covariates were selected because organizational age is associated with the accumulation of reputational stock, size with the resources available for discretionary responsibility programs, and sector with both regulatory exposure and stakeholder salience.

4.4 Procedural remedies for common method variance

Because CSR, CR, and CA were reported by the same informant at a single point in time, common method variance (CMV) is a plausible threat, and the instrument design included procedural remedies following Podsakoff et al. (2003). Respondent anonymity was guaranteed and explicitly stated on the cover page, and respondents were informed that there were no right or wrong answers, which reduces evaluation apprehension and the incentive to respond in a socially desirable manner. The predictor, mediator, and outcome items were separated within the questionnaire and presented under different section headings, creating psychological separation between measures. Scale items were not presented in a fixed conceptual order that would allow respondents to infer the hypothesized model. Statistical remedies are reported in Section 5.4.

4.5 Data analysis strategy

Data were analyzed in IBM SPSS Statistics using the PROCESS macro (version 4.2; Hayes, 2018) in four stages. First, descriptive statistics and the demographic profile were examined to characterize the sample and assess the distributional properties of the variables. Second, the measurement properties of the three scales were assessed through exploratory factor analysis using principal component analysis with varimax rotation. Average variance extracted (AVE) and composite reliability (CR) were computed from the resulting loadings, and discriminant validity was assessed against the Fornell and Larcker (1981) criterion. Third, common method variance was evaluated using Harman’s single-factor test. Fourth, the hypotheses were tested using ordinary least squares regression within PROCESS Model 4, with 5,000 bootstrap resamples and percentile-based 95% confidence intervals for the indirect effect (Preacher and Hayes, 2008). Firm age, organization size, and organization type were entered as covariates in both the mediator and outcome equations. Following current reporting standards for mediation analysis (Zhao et al., 2010), total, direct, and indirect effects are reported together with unstandardized and standardized coefficients, and the completely standardized indirect effect is reported as the effect size for the mediated pathway.

A confirmatory factor analysis is presented in Section 5.3 to formally test the measurement model, including model fit indices and standardized item loadings.

5 Results

5.1 Overview of the sample and descriptive statistics

This section follows the order in which the analyses were conducted. It first describes the sample composition and response distributions, then establishes that the measures behave as intended, examines the risk of common method variance, and finally presents the correlational and mediation results that bear on the hypotheses. Presenting the evidence in this sequence allows the reader to evaluate measurement quality before evaluating the substantive claims that rest on it.

The analytical sample comprises 150 managers and senior executives from organizations of varying size, age, and sector, including hospitality, banking, education, healthcare, and government. The sample is diverse across all three structural dimensions, which motivated their inclusion as controls.

Most respondents represented firms operating for 1–5 years (29.3%), followed by those operating for 11–15 years (22.0%). Regarding organization size, the largest group came from organizations employing 11–50 employees (26.0%), followed by those employing 51–100 employees (22.7%). By organization type, the largest proportion came from the government and public sector (23.3%), followed by banks (18.7%) and educational institutions (14.7%). Table 1 reports the full demographic profile.

Table 1. Demographic profile of respondents (N = 150).

Mean scores were above the midpoint of the scale for all three constructs (CSR = 4.298, SD = 0.668; CR = 4.311, SD = 0.688; CA = 4.193, SD = 0.767), indicating generally favorable managerial assessments of responsibility, reputation, and competitive position. Standard deviations between 0.67 and 0.77 indicate adequate dispersion for correlational analysis. However, the elevated means suggest some restriction of range at the upper end of the scales, as discussed in the limitations.

5.2 Measurement model: reliability and convergent validity

Exploratory factor analysis using principal component analysis with varimax rotation was conducted on the thirteen items across the three scales. The rotation converged in five iterations and yielded a clean three-component solution that matched the hypothesized constructs, with all items loading above 0.50 on their intended component and no cross-loadings. The analysis produced three interpretable components that jointly explained 76.83% of the total variance; the first two components had eigenvalues above one, and the third, corresponding to the competitive advantage items, had an eigenvalue of 0.95 and was retained on theoretical grounds because it formed a clean, interpretable factor. The Kaiser–Meyer–Olkin measure of sampling adequacy was 0.926, and Bartlett’s test of sphericity was significant [χ²(78) = 1472.45, p < 0.001], confirming that the data were suitable for factor analysis.

Convergent validity and reliability were then assessed. As reported in Table 2, all standardized loadings exceed 0.50; composite reliability values range from 0.874 to 0.885 and exceed the 0.70 benchmark; and AVE values range from 0.583 to 0.719 and exceed the 0.50 benchmark recommended by Fornell and Larcker (1981) and Hair et al. (2019). Cronbach’s alpha values are similarly satisfactory. The measurement model therefore demonstrates acceptable reliability and convergent validity.

Table 2. Factor loadings, reliability, and convergent validity.

CSR, corporate social responsibility; CR, corporate reputation; CA, competitive advantage; α, Cronbach’s alpha; CR (column), composite reliability; AVE, average variance extracted. Loadings are from principal component analysis with varimax rotation.

Discriminant validity was assessed using the Fornell and Larcker (1981) criterion, which requires the square root of each construct’s AVE to exceed its correlations with the other constructs. As Table 3 shows, this condition is satisfied for all three construct pairs. The square root of each construct’s AVE (CSR = 0.763, CR = 0.769, CA = 0.848) exceeds its correlations with the other constructs, including the highest correlation, which is between CSR and CR (0.753). Discriminant validity is therefore established for all constructs. The confirmatory factor analysis reported in Section 5.3 provides a further, formal test of the distinctiveness of the three constructs.

Table 3. Discriminant validity (Fornell–Larcker criterion).

Diagonal values in bold represent the square root of AVE; off-diagonal values are inter-construct correlations.

5.3 Confirmatory factor analysis

A confirmatory factor analysis (CFA) was conducted to test the proposed three-factor measurement model and to confirm that the three constructs [corporate social responsibility (CSR), corporate reputation (CR), and competitive advantage (CA)] are distinct from one another. The model was estimated in Stata 16 using structural equation modeling with maximum-likelihood estimation.

The three-factor model fit the data well. The normed chi-square (χ²/df = 1.92) was comfortably below the recommended ceiling of 3.00, indicating that the model reproduces the observed data closely relative to its complexity. The comparative fit index (CFI = 0.957) and the Tucker–Lewis index (TLI = 0.947) both exceeded the 0.90 benchmark, showing that the model performs far better than a null model with no relationships among the items. The root mean square error of approximation (RMSEA = 0.078, 90% CI: 0.059–0.098) fell within the acceptable range, and the standardized root mean square residual (SRMR = 0.038) was well under the 0.08 threshold, indicating that the differences between the observed and model-implied correlations were very small.

To confirm that the three constructs are genuinely distinct rather than overlapping, the three-factor model was compared with two simpler alternatives. When CSR and CR were forced into a single combined factor (a two-factor model), model fit deteriorated significantly (Δχ² = 98.22, Δdf = 2, p < 0.001). When all items were forced onto one common factor (a one-factor model), the drop in fit was even larger (Δχ² = 251.87, Δdf = 3, p < 0.001). The results show that the three-factor model fit the data well and significantly better than the alternative one- and two-factor models, supporting the distinctiveness of the three constructs.

Finally, all thirteen items loaded strongly on their intended constructs. The mean standardized loadings were 0.763 for the five CSR items, 0.769 for the five CR items, and 0.848 for the three CA items—all values comfortably above the 0.50 cutoff, providing evidence of convergent validity. Table 4 summarizes the model fit indices, model comparisons, and item loadings.

Table 4. Confirmatory factor analysis: model fit and standardized loadings.

χ², chi-square; df, degrees of freedom; χ²/df, normed chi-square (chi-square divided by degrees of freedom); CSR, corporate social responsibility; CR, corporate reputation; CA, competitive advantage. Loadings are reported as the mean standardized loading per construct. ^***p < 0.001.

5.4 Common method variance (CMV)

Because the data were collected from a single source at one point in time, it was important to assess whether common method variance (CMV) might inflate the observed relationships. Harman’s single-factor test was applied, with all thirteen items entered together into an unrotated principal component analysis. The first unrotated component accounted for 58.18% of the variance. Although this is slightly above the traditional 50% benchmark, that threshold is now widely regarded as unreliable and overly conservative. Fuller et al. (2016), using simulation evidence, show that a much higher level of CMV is required before it meaningfully biases relationships among substantive variables, and that CMV does not pose a grave threat at the reliability levels typical of multi-item measures. Harman’s test is itself widely regarded as a blunt diagnostic rather than a decisive indicator of method bias (Podsakoff et al., 2003). The study therefore does not rely on this test alone, and three further points place the result in context. First, the analysis did not yield one overwhelming factor: the retained components together explained 76.83% of the variance, and the first factor did not dominate the solution. Second, when the items were rotated, they loaded cleanly on their three intended constructs with no meaningful cross-loadings—not the pattern expected if a single method factor were driving responses. Third, discriminant validity was established for all constructs (Section 5.2), and all inter-construct correlations remained well below the 0.90 level that typically signals a common method problem (Bagozzi et al., 1991).

Taken together, these findings indicate that CMV is unlikely to account for the observed relationships. However, this remains a limitation given the single-informant, single-occasion design, and future research would benefit from temporally separated or multi-source data.

5.5 Correlation analysis

Correlation analysis revealed significant positive relationships among all three focal variables at the 0.01 level (Table 5). CSR was strongly and positively correlated with CR (r = 0.753, p < 0.01), and moderately and positively correlated with CA (r = 0.511, p < 0.01). CR was likewise positively correlated with CA (r = 0.548, p < 0.01). This pattern is consistent with the hypothesized model: the CSR–CR association is markedly stronger than the CSR–CA association, as expected if reputation stands between responsible conduct and competitive position rather than alongside it.

Table 5. Descriptive statistics, reliabilities, and correlations.

Correlation is significant at the 0.01 level (two-tailed). N = 150. CSR, corporate social responsibility; CR, corporate reputation; CA, competitive advantage.

5.6 Assumption testing

Before estimating the regression-based mediation model, we examined the standard assumptions of ordinary least squares (OLS) estimation. Because the constructs had high mean scores, the raw variables were negatively skewed; therefore, we assessed normality at the level of the regression residuals, the relevant assumption for OLS estimation. The histogram and normal probability plot of the standardized residuals indicated an approximately normal distribution centered on zero, with only one case showing a larger residual. Linearity and homoscedasticity were assessed through inspection of a scatterplot of standardized residuals against standardized predicted values, which showed a random scatter with no discernible pattern or funneling. Multicollinearity was assessed through variance inflation factors and tolerance values; both predictors returned a VIF of 2.49 and a tolerance of 0.40, well within the conventional cut-off of 5, indicating that multicollinearity is not a concern.

Most importantly for the mediation analysis, bootstrapping with 5,000 resamples was used to estimate the indirect effect. This procedure does not assume a normal sampling distribution and is robust to departures from normality. Taken together, these checks indicate that the analysis assumptions were adequately met.

5.7 Hypothesis testing

The hypotheses were tested using Hayes’ (2018) PROCESS Macro Model 4 with 5,000 bootstrap resamples and percentile-based 95% confidence intervals. Firm age, organization size, and organization type were included as covariates in both equations. Table 6 reports full regression results, and Table 7 presents the decomposition of effects.

Table 6. Regression coefficients, standard errors, and model summary information with control variables (N = 150).

B, unstandardized coefficient; SE, standard error; CI, confidence interval. Sample size = 150; resampling = 5,000 bootstrap. M, mediator; Y, dependent variable. CSR, corporate social responsibility; CR, corporate reputation; CA, competitive advantage. Control variables: firm age, organization size, organization type.

Table 7. Total, direct, and indirect effects of CSR on competitive advantage.

B, unstandardized coefficient; SE, standard error; CI, confidence interval. N = 150; 5,000 bootstrap resamples; 95% bias-corrected confidence intervals. A significant indirect effect with a non-significant direct effect indicates indirect-only (full) mediation. CSR, corporate social responsibility; CR, corporate reputation; CA, competitive advantage.

In the first equation, with CR as the outcome, the model was significant and explained 57.3% of the variance in reputation [R² = 0.573, F(4, 145) = 48.565, p < 0.001]. CSR was a strong, significant predictor of CR [B = 0.746, SE = 0.059, t = 12.557, p < 0.001, 95% CI (0.628, 0.863), β = 0.724]. None of the three control variables was significant. Hypothesis 2 is therefore supported.

In the second equation, with CA as the outcome, the model was significant and explained 33.7% of the variance in competitive advantage [R² = 0.337, F(5, 144) = 14.638, p < 0.001]. CR was a significant predictor of CA [B = 0.403, SE = 0.116, t = 3.487, p = 0.001, 95% CI (0.175, 0.632), β = 0.361], supporting Hypothesis 3. The direct effect of CSR on CA, controlling for CR, was positive but not significant [B = 0.226, SE = 0.119, t = 1.889, p = 0.061, 95% CI (−0.010, 0.462), β = 0.197]. None of the control variables reached conventional significance, although organization size approached significance (B = 0.083, p = 0.101).

The total effect model shows that CSR was significantly and positively associated with CA before reputation was introduced [B = 0.526, SE = 0.086, t = 6.134, p < 0.001, 95% CI (0.357, 0.696), β = 0.458; R² = 0.281, F(4, 145) = 14.168, p < 0.001]. Hypothesis 1 is therefore supported at the level of the total effect, with the qualification that the association does not remain significant once the mediator is included.

The indirect effect of CSR on CA through CR was positive and significant, since the bias-corrected bootstrap confidence interval excluded zero [B = 0.301, BootSE = 0.108, 95% CI (0.079, 0.507)]. The completely standardized indirect effect, which serves as the effect size for the mediated pathway, was 0.262 [BootSE = 0.100, 95% CI (0.067, 0.456)], a magnitude conventionally regarded as moderate. The indirect effect represents approximately 57.2% of the total effect. Because the indirect effect is significant while the direct effect is not, this pattern corresponds to indirect-only mediation in Zhao et al. (2010), commonly described as full mediation. Hypothesis 4 is supported.

In summary, Hypotheses 2, 3, and 4 are supported. Hypothesis 1 is supported at the level of the total effect but not at the level of the direct effect once reputation is included, and this qualification is central to the interpretation developed in the following section.

6 Discussion

This study aimed to determine how CSR relates to competitive advantage and whether corporate reputation mediates that relationship. The results are clear, and in one respect, the value of the study lies less in confirming that the three constructs are related than in identifying which pathway carries the relationship.

6.1 Interpreting the mediated pathway

The most consequential finding is the change in the CSR–CA association across model specifications. On its own, CSR is a strong predictor of competitive advantage (β = 0. 458, p < 0. 001), consistent with the pattern reported throughout the literature (Al-Shammari et al., 2022; Famiyeh, 2017). Once reputation is included, however, the association falls by roughly 57.2% and becomes statistically insignificant, while reputation itself remains a significant predictor. The natural reading of this pattern is not that CSR is strategically irrelevant, but that its relevance is indirect: what appeared in the reduced model to be a direct return on responsible conduct is, on closer specification, a return on the reputational asset that responsible conduct accumulates.

This is precisely what the integration of stakeholder theory and the RBV in Section 2.2 predicts, and it is worth explaining why. CSR expenditure is observable and describable, and therefore imitable: a competitor that decides to match a rival’s community investment or environmental program can generally do so. If CSR conferred advantage directly, that advantage would be transient, because the RBV holds that resources acquired at will do not sustain a superior position. Reputation, by contrast, is accumulated rather than acquired. It is built from a history of observed conduct; it is held in the minds of stakeholders rather than on the balance sheet, and its formation is causally ambiguous and socially complex in exactly the sense Barney (1991) specifies for inimitability. The finding that reputation, not CSR itself, is associated with competitive advantage is therefore consistent with the theory rather than an anomaly requiring separate explanation.

Three mechanisms plausibly underlie the reputation–advantage link, and each aligns with the pattern observed here. First, reputation reduces transaction costs because customers, suppliers, regulators, and employees extend trust to well- regarded firms more readily and demand fewer contractual safeguards. Second, it generates a preference premium: when offerings are comparable, stakeholders select the firm they regard more highly, a mechanism Roberts and Dowling (2002) identify as explaining the persistence of superior performance among high-reputation firms. Third, it provides insurance value, buffering the firm during adverse events. Each mechanism operates through stakeholder evaluation rather than operational capability, which explains why the effect appears at the reputational stage rather than at the CSR stage.

The strength of the CSR–CR association (β = 0.724) relative to the CR–CA association (β = 0.361) is also informative. This suggests that translating responsible conduct into reputational standing is the more reliable of the two conversions, whereas translating reputational standing into competitive position is more contingent, presumably because it depends on market structure, the firm’s visibility to customers, and the extent to which stakeholders can act on their evaluations. This asymmetry supports more optimistic readings in the literature: reputation is a necessary conduit, but not a sufficient guarantee of competitive return.

6.2 Relating the findings to the research objectives and framework

The research question asked how CSR is associated with competitive advantage and whether reputation mediates that association. The evidence can now be evaluated against each element of the framework. Hypothesis 2, derived from the stakeholder-theoretic proposition that responsible conduct earns stakeholder approval, is strongly supported and explains 57.3% of the variance in reputation, indicating that the approval mechanism operates as theorized in this context. Hypothesis 3, derived from the RBV proposition that an inimitable intangible asset confers advantage, is moderately supported. Hypothesis 4, the empirical expression of the integrated two-step argument, is supported in its strong form. Hypothesis 1 is supported only as a total effect. This qualification is theoretically informative rather than a shortfall: it indicates that the CSR–CA relationship reported in much of the prior literature may be a reduced-form representation of a mediated process, and that studies which omit the reputational pathway may attribute to CSR an autonomy it does not possess.

The null findings for the control variables also merit brief comment. Neither firm age, organization size, nor organization type reached significance in any model, suggesting that the reputational pathway operates similarly across organizations of different maturity, scale, and sector within this sample. This is a useful result for generalization within the UAE context. However, modest cell sizes for individual sectors mean this should be treated as provisional rather than as evidence that sector does not matter.

The reputational pathway identified here is also consistent with evidence on the specific stakeholder groups through which reputation is likely to operate. Iglesias et al. (2020) show that responsible conduct builds customer trust and loyalty, the demand-side expression of the preference premium described above; Bizri et al. (2021) find that CSR practice raises employee commitment and engagement, which corresponds to the labor-market route; and Koenigsmarck and Geissdoerfer (2021) document investor preference for organizations committed to responsible practice, which corresponds to the capital-market route. Aguinis and Glavas (2019) further argue that CSR effects are realized through stakeholders’ interpretation and sense-making of organizational conduct, which is why a perceptual construct such as reputation, rather than the activity itself, carries the association observed here. The absence of a significant size effect is also consistent with Cabrera-Gala and Miranda (2026), who document reputational returns to CSR among small and medium-sized enterprises. At the same time, Mhlanga (2022) cautions that organizations genuinely struggle to reconcile the competing claims of different stakeholder groups, which may explain why the reputation-to-advantage conversion is weaker and more contingent than the CSR-to-reputation conversion.

These findings are consistent with prior work. Lin-Hi and Blumberg (2018) report that organizations implementing CSR are perceived as more ethical and trustworthy than competitors; Chen et al. (2026) find greater stakeholder support for more socially responsible organizations; and Mou et al. (2026) characterize CSR as a commitment to societal well-being that shapes professional standing. Xuetong et al. (2024) report a comparable mediating role for reputation in the CSR–competitive advantage relationship, and Babu et al. (2026) document the contribution of responsible practice to business performance. The present study adds to this body of work by showing that, once structural controls are included and the full decomposition of effects is reported, the direct pathway does not survive. This is a stronger claim about the primacy of reputation than most of these studies make.

6.3 Theoretical implications

The study makes three theoretical contributions. First, it empirically tests a mechanism that is often assumed. By specifying reputation as the conversion point between stakeholder approval and resource-based advantage, and by finding indirect-only mediation, it provides evidence that the two theoretical traditions are sequentially rather than merely jointly relevant: stakeholder theory governs the first link and the RBV governs the second. Second, it demonstrates the analytical cost of omitting the mediator. The same data yield a significant direct CSR–CA association in a reduced model and a non-significant one in the full model, suggesting that some direct effects reported in this literature may be attributable to model under-specification rather than to a genuine autonomous pathway. Third, it clarifies the boundary of the RBV argument for CSR: the resource that satisfies the inimitability condition is reputational standing, not responsibility activity, and theoretical claims about CSR as a strategic resource should be stated at that level of precision.

6.4 Practical implications

The practical implications follow directly from the estimated model and are deliberately limited to what the data support.

Because the association between CSR and competitive advantage runs through reputation rather than around it, managers should treat reputational standing as the intermediate outcome against which they evaluate investments in responsibility. Programs that are unlikely to become known to or valued by stakeholders should not be expected to yield a competitive return, however worthwhile they may be on other grounds.

Because CSR explained 57.3% of the variance in reputation while accounting for a more modest share of the variance in competitive advantage, managers should also be realistic about the size and reliability of the second conversion. Building reputation through responsible conduct is the more dependable step; converting reputation into competitive position depends on market conditions that individual firms control only partially.

Because the effect operates through stakeholder perception, consistency between conduct and communication is essential. Reputation’s mediating role implies that any divergence between what an organization does and what it claims damages the mechanism itself, not merely the message.

Finally, because firm age, size, and sector were not significantly associated with either reputation or competitive advantage in this sample, the reputational pathway does not appear to be the preserve of large or long-established organizations. Smaller and younger firms in the UAE have no structural reason, on this evidence, to regard reputation-building through responsible conduct as out of reach.

6.5 Limitations and future research

Several limitations should be noted. First, the study was conducted exclusively within the UAE, limiting generalizability to other institutional settings. Second, the sample comprised 150 respondents, which, while adequately powered for the estimated models, limits the ability to examine sector-level differences. Third, the design is cross-sectional, so the estimates describe associations at a single point in time; reputation is by definition an accumulated asset, and a cross-sectional design cannot capture accumulation. Longitudinal or panel designs would better test the temporal ordering implied by the theory.

A further limitation is reliance on self-reported data from managers, which may reflect their assessments of their organizations rather than objective conditions. Although procedural remedies were applied and statistical diagnostics were conducted, the first unrotated component in Harman’s test accounted for slightly more than 50% of the variance, so CMV cannot be ruled out entirely. Future studies should collect predictor and outcome measures from different sources, for example by pairing managerial reports of CSR with customer or third-party assessments of reputation and with archival indicators of competitive position, or by temporally separating measurement occasions.

Two further limitations warrant explicit mention. Although discriminant validity between CSR and CR was established after scale purification, the two constructs remain strongly correlated, consistent with the theoretical claim that reputation stores responsible conduct. The confirmatory analysis reported in Section 5.3 addresses this directly, and future work might adopt instruments designed to sharpen the empirical separation between responsible conduct and its perceptual consequences. Organization type was entered as a coded indicator rather than as a full set of dummy variables. Although it was not significant, a specification using dummy coding with a larger sample would provide a stronger test of sector effects. Finally, mean scores on all three constructs were high, suggesting some restriction of range that may attenuate the observed associations.

Future research may also examine how technological capability interacts with reputational standing to shape competitive position, and test whether the indirect-only pattern observed here replicates in institutional settings with stricter CSR regulation or more intense stakeholder scrutiny. A further extension would trace the model beyond the focal firm into the supply chain, where responsibility commitments are increasingly assessed and where conditions for sustainable production have been shown to shift rapidly under external shocks (Sarkis et al., 2020).

7 Conclusion

This study examined the association between corporate social responsibility and competitive advantage and tested whether corporate reputation mediates that association among organizations operating in the United Arab Emirates. The evidence indicates that CSR is positively associated with reputation, that reputation is positively associated with competitive advantage, and that the association between CSR and competitive advantage operates through reputation rather than independently. Once reputation was accounted for, the direct association between CSR and competitive advantage was no longer statistically significant. In contrast, the indirect association through reputation remained significant and accounted for approximately 57.2% of the total effect.

Viewed through the combined lens of stakeholder theory and the resource-based view, this pattern suggests that responsible conduct earns stakeholder approval, and that approval, once accumulated into reputational standing, becomes the asset competitors find difficult to replicate. The practical corollary is that organizations should treat reputation as the intermediate outcome through which responsibility investment is expected to pay off, and should recognize that consistency between conduct and communication sustains the mechanism.

These conclusions should be read with appropriate caution. The design is cross-sectional, and the measures are self-reported, so the findings establish patterns of association rather than causal effects. The single-country sample limits generalization. Within these boundaries, the study offers evidence that the strategic value of corporate social responsibility is realized through reputation. Models that omit this pathway are likely to misattribute the location of that value.

Generative AI statement

The author(s) declared that Generative AI was used in the creation of this manuscript. Grammarly Inc.{v1.2.294.1953} was used for basic language editing, limited to grammar and spelling corrections. No generative AI tools were used to create, draft, or generate any content in this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Pardeep Kumar, Chandigarh University, India

Reviewed by: Agus Subhan Prasetyo, Texas A and M University, United States

Niyom Suwandej, Suan Sunandha Rajabhat University, Thailand