Does the Cost of Capital Mediate the Link Between Sustainability Commitment and Innovation Intensity? Dynamic Evidence from Saudi-Listed Companies

2.1. The financial channel: Sustainability as a risk reduction tool

The first channel suggests that a company’s ESG commitment provides a tangible financial advantage, specifically by lowering its cost of capital. Several overlapping theories explain this process. Stakeholder² and legitimacy³⁵ theories state that a company that successfully manages its relationships with all its stakeholders (employees, customers, communities, regulators) and aligns its actions with societal norms enhances its license to operate. This legitimacy reduces non-financial risks, such as regulatory, reputational, and operational risks, thereby stabilising future cash flows and lowering the risk premium investors require.^36,37

Signalling theory⁵ complements this view by asserting that the disclosure of high-quality ESG information reduces information asymmetry between managers and capital providers.⁴ By signalling proactive management and a long-term vision, the company strengthens investor confidence, which makes investors willing to accept a lower return for a perceived lower level of risk.⁷ Finally, this mechanism is part of the debate raised by agency theory.¹² While the “overinvestment” hypothesis suggests that ESG spending may be an agency cost penalised by markets,¹¹ the now dominant “risk mitigation” perspective views ESG engagement as a prudent investment that protects long-term shareholder value.³⁸ Together, these theories predict that credible ESG performance is rewarded with a lower cost of capital.

A body of literature shows a negative relationship between a firm’s environmental, social, and governance (ESG) performance and its cost of capital.³⁹ The prevailing logic, derived from stakeholder, legitimacy, and signalling theories, is that ESG engagement acts as a powerful de-risking mechanism. By proactively managing non-financial risks (regulatory, reputational, climate) and improving transparency, sustainable firms reduce information asymmetry and strengthen investor confidence, thereby requiring a lower risk premium.

Recent empirical studies confirm this link, showing that high ESG scores are associated with lower costs of equity and debt.^7,10,40 Recent studies refine this conclusion by showing that financial markets integrate specific factors such as carbon risk,¹³ environmental disclosure policies 16, or voluntary green certifications¹⁷ into the assessment of the cost of capital. Although some studies highlight contextual nuances, such as market maturity or governance quality,^41,42 the general trend confirms that ESG engagement is financially rewarded. In the Saudi context, where institutional investors are paying increasing attention to ESG transparency,⁴³ we expect this mechanism to be at work as well. This leads us to our first hypothesis:

Hypothesis 1 (H1): A more substantial sustainability commitment is associated with a lower cost of capital.

2.2. Cost of capital and innovation: The financial constraints channel

The corporate finance literature clearly establishes that the cost of capital is a constraint on innovation. Innovation, as measured by R&D expenditure, is a long-term investment inherently risky and has uncertain returns. The cost of capital serves as a “hurdle rate” for such projects. A high cost of capital makes it more challenging to reach the break-even point for R&D projects, discouraging managers from undertaking them and favoring shorter-term investments.^32,33

Empirical studies confirm that financial frictions hamper R&D investment⁴⁴ and that, conversely, a lower cost of equity is associated with greater innovation efforts.⁴⁵ An interesting nuance is provided by Peia and Romelli,⁴⁶ who show that a moderate increase in the cost of capital can encourage firms to allocate their resources more efficiently by protecting the most strategic R&D projects. Nevertheless, the consensus remains that a structurally high cost of capital remains a significant obstacle to a sustained innovation effort. We therefore formulate our second hypothesis:

Hypothesis 2 (H2): A higher cost of capital is negatively related to firms’ innovation intensity.

2.3. ESG engagement and innovation: The strategic-capability channel

The second channel, strategic in nature, posits that ESG engagement directly enhances the firm’s innovation capacity, independent of any financial advantage. This link is primarily explained by the RBV and the concept of dynamic capabilities. According to the RBV, a sustainable competitive advantage stems from the possession of rare, valuable, and difficult-to-imitate resources.⁴⁷ ESG engagement specifically compels firms to develop resources, whether intangible assets like a responsible brand reputation or new organisational skills, such as mastery of eco-design or management of circular supply chains.²¹

The dynamic capabilities framework goes further by defining a firm’s ability to integrate, build, and reconfigure its competencies to respond to a rapidly changing environment.²² ESG pressures (regulatory, competitive, societal) act as “creative shocks” that force firms to develop sustainability-specific dynamic capabilities: the ability to detect opportunities and threats related to sustainable development, to seize these opportunities by mobilising resources for green innovation, and to reconfigure business models and processes to align with a low-carbon economy.⁴⁸ This process of continuous adaptation is innovation. Finally, organisational learning theories emphasise that ESG engagement fosters a culture of collaboration and knowledge sharing, creating fertile ground for responsible initiatives to transform into creative solutions.

Regardless of the financial channel, a growing literature argues that ESG engagement acts as a direct catalyst for innovation. According to this perspective, often associated with the “Porter hypothesis,” environmental constraints and social expectations act as “creative shocks” that push firms to develop new products, processes, and business models, a phenomenon known as eco-innovation.¹⁸

This link is explained by the development of new organisational competencies, as suggested by the RBV and dynamic capabilities. By addressing ESG challenges, firms acquire unique know-how and strengthen their ability to adapt to a changing environment.^21,22 Recent empirical studies confirm that ESG engagement stimulates green innovation within polluting firms,²⁴ promotes the development of sustainable business models,¹⁹ and drives performance in specific sectors, such as the blue economy.²⁵ This direct effect leads us to our third hypothesis:

Hypothesis 3 (H3): A stronger sustainability commitment positively influences firms’ innovation capacity.

2.4. Channel interplay and the role of institutional pressure: The empirical gap

These two channels do not operate in isolation. Institutional theory⁴⁹ provides a unifying framework by highlighting how external pressures (coercive, normative, and mimetic) simultaneously shape firms’ financial and innovation strategies. In a context like Saudi Arabia, Vision 2030 exerts top-down institutional pressure, incentivising firms to adopt ESG practices to maintain their legitimacy and market access. This same pressure pushes them to innovate, comply with new regulations, and seize opportunities in green markets.

This integrated framework leads us to the central question of this research. If ESG engagement reduces the cost of capital (financial channel) and a lower cost of capital facilitates innovation by lowering the required rate of return for R&D projects,^32,33 it is logical to postulate that the cost of capital acts as a mediator. However, the relative strength of this indirect financial channel compared to the direct strategic channel remains an open empirical question. Our study aims precisely to separate these two mechanisms to determine whether sustainability is primarily a source of financing for innovation or an innovation imperative.

Although these pairwise links are well established, prior studies rarely test the whole chain, ESG → cost of capital → innovation in an integrated framework.^18,34 The most natural, yet under-tested, inference is that financial de-risking is a mediating mechanism: ESG lowers the cost of capital, and cheaper capital enables R&D. At the same time, recent work suggests asymmetric pricing across capital markets; equity investors often react more strongly to ESG than creditors, implying that any mediation may be uneven across cost of equity versus cost of debt.^13,39 Testing mediation in a policy-active emerging market such as Saudi Arabia is therefore essential.

Hypothesis 4 (H4): The cost of capital mediates the relationship between sustainability commitment and innovation.

Figure 1 summarises the research and the conceptual model.

Close

Figure 1: Research and conceptual model.

Export to PPT

3.1. Sample and data

Our empirical analysis examines a sample of firms listed on the Saudi Stock Exchange (Tadawul) from 2016 to 2024. We acknowledge that the Saudi Vision 2030 was officially inaugurated in April 2016. However, we have retained the full year of 2016 in our analysis for two critical theoretical and econometric reasons. First, based on signalling theory, capital markets are forward-looking; the announcement of such a significant regime shift triggers immediate anticipatory adjustments in risk premiums and corporate strategies, making 2016 a pivotal year for capturing the market’s initial reaction. Second, regarding econometric specification, the System GMM estimator is data-intensive and relies on lagged values (t-1, t-2) as internal instruments. Excluding 2016 would reduce the time series (T) to 8 years, significantly weakening the instrument set and the power of the Sargan-Hansen overidentification tests. Furthermore, including year fixed effects in our model absorbs any transitional volatility or aggregate shocks specific to the 2016 launch period, ensuring that our coefficient estimates remain unbiased. In line with standard practice, we excluded financial firms due to their unique balance sheets and regulatory frameworks. After removing observations with missing data, our final sample comprises a balanced panel of 58 non-financial firms tracked over nine years, totalling 522 firm-year observations. The sample composition, primarily consisting of basic materials (36.21%) and energy (8.62%), reflects Saudi Arabia’s historical economy. This isn’t bias but a core issue, as Vision 2030 aims to diversify and make these sectors sustainable and innovative. Financial, governance, and ESG data are sourced from the Refinitiv/London Stock Exchange Group (LSEG) database, a leading global market data source. The sample selection is shown in Table 1.

3.2. Operationalisation of variables

3.3. Econometric strategy and model specification

Estimating a firm’s sustainability impact on innovation involves key econometric challenges. Three issues arise. First, dynamic endogeneity, or Nickell bias,⁵⁰ occurs because innovation decisions are persistent, with current R&D depending on past spending. Including the lagged dependent variable causes correlation with the error term, biasing results, especially in short panels (T=9). Second, simultaneity or reverse causality exists; lower capital costs can boost innovation, but highly innovative firms may be perceived as less risky, lowering their capital costs. Third, omitted-variable bias arises from unobservable traits such as management, culture, or tech opportunities that affect ESG, capital, and innovation; ignoring these leads to biased estimates.

To address these challenges, we use the two-step System GMM estimator for dynamic panel data, developed by Arellano and Bover⁵¹ and Blundell and Bond.⁵² It outperforms the earlier Difference GMM estimator⁵³ when variables are highly persistent, as is the case with R&D expenditures. Lagged levels become weak instruments for the first-differenced equation. Still, System GMM fixes this by combining the differenced equation (instrumented with lagged levels) with an equation in levels (instrumented with lagged differences).

To ensure our GMM’s robustness and credibility, we follow best practices from the literature. We use the efficient two-step estimator and apply Windmeijer’s⁵⁴ correction to standard errors to fix bias in small samples. To avoid ‘instrument proliferation,’ which can overfit and weaken tests, we use Roodman’s^55,56 collapse option, collapsing one instrument per lag across all periods. Following this approach, we define the three dynamic panel models below to examine our hypotheses:

We estimate the following three dynamic models to test our hypotheses:

4.1. Descriptive statistics and preliminary analysis

Table 2 presents descriptive statistics for 522 firm-year observations of 58 Saudi non-financial firms (2016–2024). ESG scores are widely spread mean 27.63, SD 16.82, ranging from 0.85 to 84.24. This shows uneven sustainability practices, with some pioneering firms and many laggards. Board gender diversity is very low; female directors average just 2.16%, and the median is 0, indicating most boards are male-dominated. The weighted average cost of capital (7.1%) varies moderately, reflecting differences in financing structures.

The correlation matrix in Table 3 shows that WACC and ESG score are negatively correlated (r =–0.080), supporting H1. Firm size strongly correlates with ESG performance (r = 0.456), implying larger firms tend to invest in sustainability. This warns against interpreting ESG effects without considering firm size. Variance inflation factors are low (mean = 1.17), indicating multicollinearity isn’t an issue.

4.2. Robustness diagnostics and endogeneity tests

Before estimating the main dynamic models, we performed diagnostics to assess the exogeneity of key regressors (WACC, ESG Score, Leverage, and Firm Size) within a fixed-effects framework. These tests serve as diagnostics and do not replace the System GMM model, which addresses dynamic endogeneity (Nickell bias). We used a manual Two-Stage Least Squares (2SLS) estimation and a control function (CF) approach, employing lagged values as instruments. We evaluated instrument relevance, finding first-stage F-statistics [Table 4] for all endogenous variables well above 10, confirming their strength. However, the high values suggest potential over-fitting from numerous instruments, highlighting the need to use the collapse option in GMM estimations to prevent instrument proliferation.

We tested exogeneity using the CF approach, including first-stage regression residuals as regressors. If variables are exogenous, these residuals should be insignificant. Table 5 shows that none are significant. It is essential to clarify that, while the Durbin-Wu-Hausman test (p=0.639) suggests no simultaneity bias between the regressors and the error term, it does not account for dynamic endogeneity. The presence of the lagged dependent variable (Ln(RD)_t-1) in our model introduces Nickell bias, rendering standard ordinary least squares (OLS) and Fixed Effects estimators inconsistent. Therefore, despite the exogeneity indicated by static tests, the use of System GMM remains econometrically necessary to control the dynamic persistence of innovation and ensure consistent estimates.

Table 6 compares the coefficients from a manual 2SLS estimation with a standard fixed-effects OLS model. The coefficients and p-values are nearly identical across both models, confirming that, within this static framework, instrumenting the variables does not significantly alter the results.

At first glance, these results seem to challenge the need for System GMM. However, this contradiction arises from the limitations of static tests. Static tests exclude the dynamic component necessary to address Nickell bias arising from the lagged dependent variable, which is crucial for capturing innovation’s persistence. Not detecting endogeneity in a static model doesn’t mean it’s absent in the correct dynamic model. These diagnostics support our approach by demonstrating careful consideration of assumptions. System GMM remains the most robust method for our research.

4.3. Dynamic panel estimation results

Mediation analysis: SEM approach

To test if the cost of capital mediates the ESG-innovation link (H4), a structural equation model with 2,000 bootstrap replications was used. The indirect path (ESG → WACC → ln(RD)) is negative (–0.00013) and not significant (p = 0.526). The direct ESG-innovation effect is positive (0.0119, p < 0.001). Since the total effect is nearly identical to the direct effect, mediation is rejected. Sustainability improves financing but doesn’t significantly influence innovation through this channel [see Table 10, Figures 2-4].

Table 10: Standardised SEM coefficients for mediation via WACC.

Path	Coefficient	z-stat	p-value
ESG → WACC	–0.0624	–1.57	0.117
WACC → ln_rd	0.0302	0.73	0.465
ESG → ln_rd (direct)	0.1641***	4.79	0.001
Indirect effect (ESG → WACC → ln_rd)	–0.00014	–0.63	0.526
Total effect	0.0118***	4.63	0.001

Standardised coefficients from SEM with bootstrap standard errors (2 000 replications). ***p < 0.01. The indirect effect is not significant, confirming the absence of mediation. WACC: Weighted average cost of capital, ESG: Environmental, social, and governance.

Notes : Standardised coefficients from SEM with bootstrap standard errors (2000 replications). The indirect effect is not significant, confirming the absence of mediation.

Close

Figure 2: Structural equation modelling (SEM) mediation path. ESG: Environmental, social, and governance, WACC: Weighted average cost of capital. Notes: The diagram illustrates the standardised coefficients from the mediation analysis ***p<0.01, *p<0.1. The path from ESG Score to Innovation via Cost of Capital represents the indirect effect, which was found to be statistically insignificant.*

Export to PPT

Close

Figure 3: Standardised effect on R&D expenditures. ***p<0.01, **p<0.05.

Export to PPT

Close

Figure 4: Standardised effect on cost of capital. WACC: Weighted average cost of capital. ***p<0.01, **p<0.05, *p<0.1.

Export to PPT

5.1. ESG engagement and the cost of capital: Confirmation of a risk reduction channel

Our results support hypothesis H1, showing that high ESG scores significantly reduce Saudi firms’ cost of capital. A one-standard-deviation increase in ESG lowers WACC by about 8.4 basis points, indicating that investors value ESG even in emerging markets. This rewards sustainable firms with better financing, consistent with international studies.^6,8,9

This relationship is driven by a risk-reduction mechanism supported by theoretical frameworks. Consistent with signalling theory, transparent ESG performance signals managerial quality and long-term vision, reducing information asymmetry in an often high-uncertainty market. Studies in China show that environmental disclosure, voluntary or mandatory, lowers equity capital costs.^16,17 Legitimacy and institutional theories explain that, under Vision 2030, a high ESG score aligns with national goals, enhances legitimacy, and reduces the perceived risks of sanctions or contract losses, especially in the materials and energy sectors in our sample.

Our findings support the ESG risk mitigation view over agency theory’s overinvestment hypothesis. Saudi investors see ESG spending as prudent management of climate, regulatory, and reputational risks, as well as to stabilise cash flows and lower required returns. Robustness tests show the cost of equity drives this benefit, while debt remains unaffected by ESG performance. This indicates a maturity gap: shareholders incorporate ESG risks, but creditors focus on traditional metrics, possibly reassured by state support. Improving ESG data quality and assurance could bridge this gap, aided by higher audit standards.

5.2. The cost of capital as a constraint to innovation

Consistent with hypothesis H2, our analysis confirms that the cost of capital is a significant constraint on Saudi firms’ innovation. The negative, significant coefficient of WACC on R&D expenditures (β = –5.894, p<0.10) indicates that a 1% increase in WACC is associated with a 5.9% decrease in R&D spending. This aligns with corporate finance principles: a higher cost of capital, as a “hurdle rate,” makes risky, long-term R&D projects less likely to have a positive NPV. This supports literature showing financial constraints hinder innovation.^32,33,44.

Our robustness tests, which decompose the cost of capital, confirm and refine this conclusion. The effect is more substantial for the cost of debt, indicating that creditor constraints emphasising cash flow security heavily influence R&D investments in high-uncertainty environments. Interestingly, we observe a weak positive relationship between the cost of equity and innovation. While this appears to contrast with the ‘hurdle rate’ logic of Hypothesis 2, it is consistent with the risk-return trade-off fundamental to equity markets. Innovation is inherently risky; firms with high R&D intensity often exhibit more volatile future cash flows. Consequently, equity investors, as residual claimants, demand a higher risk premium for these innovative firms. This suggests that while debt markets (WACC) act as a constraint on innovation, the equity market reflects the higher risk profile associated with aggressive R&D strategies, as discussed by Zheng et al.⁴⁵.

Despite this anomaly, the negative impact of WACC and debt costs clearly shows that tighter financing conditions hamper Saudi firms’ innovation. Peia and Romelli⁴⁶ agree, noting that moderate financial tightening might paradoxically boost innovation efficiency by encouraging firms to focus on strategic projects.

5.3. The strategic dividend of sustainability: A direct catalyst for innovation

Our study confirms a strong positive link between a firm’s sustainability commitment and its innovation capacity, supporting hypothesis H3. The ESG score positively impacts R&D spending (0.0117, p = 0.045). A one-standard-deviation increase in ESG correlates with about a 11.7% rise in R&D. For Saudi firms, ESG engagement drives strategic renewal and capability development, not just reporting or compliance.

This “strategic dividend” of sustainability is explained by resource- and capability-based theories. The resource based view says that sustainability pushes firms to develop unique resources, such as eco-design and circular-economy management, thereby aiding innovation. The dynamic capabilities framework adds that pressures such as Vision 2030 and the global shift to sustainability prompt firms to identify trends, reallocate resources, and transform business models. This ongoing adaptation and learning drive innovation.

Our results validate the “Porter hypothesis” in a non-Western, resource-dependent economy. They support literature showing that sustainability spurs new business models in sectors such as mining,⁶⁰ encourages sustainable innovation along global value chains,²³ and drives the blue economy.²⁵ Our findings confirm that green innovation is vital for sustainable development in high-polluting firms, which are dominant in our sample and are consistent with other industrialising economies.^45,61

5.4. The absence of mediation: The primacy of the strategic channel over the financial channel

The study’s key contribution is rejecting the mediation hypothesis (H4). While ESG reduces the cost of capital (H1) and facilitates innovation (H2), the indirect path is not statistically significant (p=0.526). The direct ESG-innovation effect remains large and significant (p<0.001). Our robustness tests confirm this, leading to the conclusion that, during Vision 2030 in Saudi Arabia, innovation driven by sustainability pressures is a stronger motivator than financial relief. Several factors rooted in the Saudi context explain the dominance of this strategic channel. First, institutional pressures, notably Vision 2030’s top-down push, create a crucial need for large firms to diversify and decarbonise. This coercive and normative pressure makes innovation vital for long-term survival and legitimacy. Firms innovate because their environment and license to operate require it, not just because capital is cheaper.

Second, statistical significance differs from economic relevance. An 8.4 basis-point reduction in WACC is a real financial benefit but may not strongly influence investment in large, transformative R&D projects. The strategic long-term benefits, such as ensuring future viability and political legitimacy, likely outweigh the marginal savings. Third, the firm’s financing, often from internal cash flows, may lessen the impact of external capital cost changes, like those reflected in the WACC. This challenges the assumption that ESG value is mainly financial. It shows that the strategic channel, based on developing internal capabilities, is dominant. Financial constraints matter, but are not the main pathway for ESG’s influence on innovation here. This suggests the need for more nuanced models considering parallel pathways.