Stakeholder Theory posits that corporate success depends on a firm’s ability to manage relationships with a broad spectrum of internal and external stakeholders, including customers, suppliers, employees, local communities, and regulators. In the agri-food sector, environmental stewardship (e.g., sustainable farming, waste reduction) and social responsibility (e.g., fair labor practices, community engagement) foster trust and mitigate reputation risks ([5]). In emerging economies, where consumer awareness and regulatory enforcement are evolving rapidly, proactive ESG engagement can enhance social license to operate, lower capital costs, and secure market access ([7]).

The Resource-Based View argues that firms achieve sustainable competitive advantages by cultivating resources and capabilities that are valuable, rare, inimitable, and non-substitutable. Under this framework, ESG integration represents an internal capability building process. Superior environmental practices stimulate operational efficiency and technological innovation (e.g., energy-efficient processing, resource optimization), while strong governance structures improve capital allocation and operational oversight. In agricultural processing and distribution, such capabilities directly influence operational performance metrics like Asset Turnover and Return on Capital Employed (ROCE) ([8]).

Institutional Theory emphasizes the role of external institutional environments in shaping corporate behavior. Emerging markets are frequently characterized by “institutional voids”—weak regulatory enforcement, information asymmetry, and market inefficiencies. In such environments, robust corporate governance and high ESG scores act as signaling mechanisms that reduce agency costs and informational friction for international investors and trade partners ([9]). However, compliance costs associated with ESG initiatives may also impose financial burdens, creating non-linear or trade-off effects on short-term profitability metrics such as Pretax ROA.

Research Hypotheses

Based on the theoretical framework, we formulate the following hypotheses:

Hypothesis 1 (H₁): Overall ESG performance positively affects the corporate profitability (Pretax ROA) of market-leading agri-food enterprises in emerging markets. Hypothesis 2a (H_2a): The Environmental Pillar Score positively impacts operational efficiency (Asset Turnover). Hypothesis 2b (H_2b): The Social Pillar Score positively impacts corporate profitability by building stakeholder capital. Hypothesis 2c (H_2c): The Governance Pillar Score exerts a significant positive impact on firm financial performance and capital structure stability. Hypothesis 3 (H₃): Institutional quality positively moderates the relationship between ESG performance and corporate financial outcomes in emerging markets.

The empirical analysis utilizes a balanced panel dataset of 35 market-leading agri-food enterprises operating across seven key emerging economies (China, India, Brazil, Mexico, Indonesia, South Africa, and Türkiye) over the ten-year period from 2014 to 2023 (N = 35, T = 10, N × T = 350 observations). The final dataset for this study comprises 35 market-leading agri-food enterprises, selected across seven representative emerging economies (five enterprises per country). To ensure methodological consistency and sample comparability, a multi-stage purposive sampling approach was employed based on two main criteria: firm scale and data availability. First, companies were selected on the basis of their market leadership, operationalized through gross annual revenue, thereby isolating the top five dominant market players in each respective emerging market. Second, to construct a balanced dynamic panel dataset, firms were screened for data completeness within the Refinitiv EIKON database. Enterprises with significant data missingness or non-disclosed Environmental, Social, and Governance (ESG) metrics over the specified ten-year observation period were systematically excluded from the sample. Consequently, the final sample reflects the key corporate actors in the emerging agrifood market structure that exhibit sufficient non-financial reporting transparency to allow for robust econometric estimation.

Beyond the direct nexus between Environmental, Social, and Governance (ESG) performance and corporate financial performance, the econometric framework incorporates an interaction term between firm-level ESG scores and country-level institutional quality. This moderating specification addresses the theoretical premise that the financial returns of corporate sustainability practices are not uniform, but are fundamentally conditioned by the broader regulatory and macroeconomic environment in which firms operate. In emerging economies characterized by varying degrees of institutional voids—such as weaker regulatory enforcement, legal inefficiencies, or information asymmetries—the strategic value of ESG engagement is magnified, as responsible practices often act as a vital signaling mechanism to mitigate market uncertainties. Consequently, the level of institutional quality can either substitute for or complement firm-level transparency, systematically altering the marginal effect of ESG initiatives on profitability indicators such as Pretax ROA. By explicitly modeling this interaction prior to evaluating alternative mediating or direct channels, the dynamic panel specification captures cross-country heterogeneity across the seven emerging markets, ensuring a more nuanced assessment of how macro-environmental rigor shapes the financial efficacy of corporate social responsibility.

Table 1.Overview of the variables used in the article.

Firm-level financial indicators, ESG overall scores, pillar-specific scores (Environmental, Social, Governance), and ESG controversy metrics were extracted from the Refinitiv EIKON database. Country-level institutional quality indicators were collected from the World Bank’s Worldwide Governance Indicators (WGI) database (Table 1).

To model corporate financial performance dynamically, we specify an autoregressive model containing lagged dependent variables. Financial indicators in the agri-food sector exhibit significant persistence over time due to capital commitment cycles, long-term supply agreements, and operational momentum.

The dynamic baseline equation is formulated as:

Where:

Y i t represents the financial performance metric (ROA_it, ROCE_it, or ATO_it) for firm i in year t . Y i t − 1 is the one-period lagged financial metric capturing dynamic persistence. ESG_it denotes overall ESG performance or the vector of pillar scores (ENV_it, SOC_it, GOV_it). X i t is a vector of firm-level and country-level control variables (SIZE_it, LEV_it, INST_it, COVID_it). η i captures unobserved, time-invariant firm-specific fixed effects. λ t represents unobserved time-specific effects. ε i t is the idiosyncratic error term, ε i t ∼ IID ( 0 , σ 2 ) .

Estimating dynamic panel models using standard OLS or Fixed Effects estimators yields biased and inconsistent estimates due to Nickell bias ( Corr ( Y i t − 1 , η i ) ≠ 0 ) and endogenous regressor conditions ( Corr ( ESG i t , ε i t ) ≠ 0 ) arising from reverse causality (e.g., higher profitability providing capital for ESG investments). In the specification and execution of the System GMM estimator, the variables are systematically categorized based on their expected dynamic relationships and potential feedback effects with the error term. Specifically, the dependent variable (corporate financial performance, such as Pretax ROA) along with the core explanatory ESG metrics are treated as endogenous and predetermined variables, acknowledging the presence of potential simultaneity and unobserved firm-level heterogeneity. Consequently, these variables are instrumented using internal instruments derived from their own lagged levels and differences. Control variables, depending on their structural nature, are specified either as predetermined (weakly exogenous)—allowing for possible feedback from past shocks—or as strictly exogenous. To maintain asymptotic validity and mitigate the severe econometric risk of instrument proliferation (too many instruments), which can overfit endogenous variables and inflate the p-values of the Hansen J-test, the lag depth of the instrument matrix is strictly restricted. Specifically, the lag range for the GMM-type instruments is bounded and the instrument matrix is collapsed where necessary, ensuring that the total count of instruments remains strictly below the cross-sectional sample size (N = 35) across all estimated model specifications.

To overcome endogeneity, unobserved heterogeneity, and measurement error, we employ the System GMM estimator developed by [2] and [4] ([1]). System GMM constructs a system of two equations:

1) An equation in first-differences instrumented by lagged levels of the variables:

2) An equation in levels instrumented by lagged first-differences.

Diagnostic Tests and Specification Rules

Arellano-Bond Tests for Autocorrelation: The model requires first-order serial correlation in differences (AR(1), p < 0.05), but no second-order serial correlation (AR(2), p > 0.10). Hansen J-Test of Overidentifying Restrictions: Evaluates overall instrument validity under the null hypothesis that instruments are exogenous (p> 0.10). Instrument Count Control: To prevent instrument proliferation (“instrument proliferation bias”), the matrix of instruments is collapsed, ensuring the total instrument count remains strictly below the number of cross-sectional units (N = 35). Robustness & Non-Linearity Checks: To account for potential curvilinear relationships (e.g., U-shaped or inverted U-shaped ESG dynamics), quadratic terms ( ESG i t 2 ) and 1% - 99% Winsorized Difference GMM models are specified as stability tests ([3]).

Table 2 presents the descriptive statistics for all firm-level financial metrics, ESG scores, and macro-level controls across the 350 firm-year observations (N = 35, T = 10). Over the 2014-2023 sample period, the mean Pretax ROA stands at 7.82%, with a standard deviation of 4.15%, reflecting notable performance variance across emerging market agri-food enterprises. The average overall ESG score is 54.12, ranging from 18.40 to 88.60, which highlights heterogeneous adoption levels of sustainability practices within the sample.

Prior to dynamic estimation, Levin-Lin-Chu and Fisher-ADF panel unit root tests confirmed that all variables are stationary at levels or first differences, ruling out spurious regression risks. A baseline Hausman test (p< 0.001) rejected the Random Effects specification in favor of Fixed Effects, supporting the transition to dynamic System GMM to control for time-invariant heterogeneity and persistent financial dynamics.

Table 2.Descriptive statistics (2014-2023).

Table 3reports the main empirical results estimated via the two-step System GMM estimator with Windmeijer-corrected robust standard errors. Model 1 evaluates overall ESG performance, while Model 2 disaggregates ESG into the three individual pillar scores (ENV_it, SOC_it, GOV_it).

Table 3. System GMM dynamic panel regressions (Dependent variable: pretax ROA).

Robust standard errors are reported in parentheses in Table 3. Asterisks indicate statistical significance: ***p < 0.01, p< 0.05, *p < 0.1. Time dummies (including the 2020-2021 pandemic shock period) are included in all specifications but omitted from the presentation to conserve space.

1)Profitability Persistence and Overall ESG Effect

The coefficient on lagged profitability ( ROA i t − 1 ) is statistically significant and positive across all specifications (β ≈ 0.40, p < 0.01), confirming strong performance inertia in the agri-food sector. Overall ESG score exhibits a positive and statistically significant coefficient (β = 0.054, p < 0.05), lending empirical support to Hypothesis 1 (H₁). In line with Stakeholder Theory, superior ESG integration generates tangible strategic value for emerging market enterprises, outweighing initial compliance expenditures.

2)Disaggregated Pillar Dynamics

The disaggregated analysis in Model 2 yields essential insights into specific value-creation channels:

Governance Pillar (GOV_it): Exerts the strongest positive impact on corporate profitability (β = 0.062, p < 0.01), validating Hypothesis 2c (H_2c). Robust corporate governance functions as a critical risk-mitigation tool in emerging markets, reducing agency costs and information asymmetries caused by institutional voids. Environmental Pillar (ENV_it): Shows a moderate positive relationship with Pretax ROA (β = 0.038, p < 0.10), supporting Hypothesis 2a (H_2a). Additional estimations using Asset Turnover (ATO_it) as the dependent variable indicate that resource-efficient operational practices improve asset utilization efficiency over time. Social Pillar (SOC_it): Displays a positive but statistically insignificant linear effect (β = 0.021, p > 0.10), providing partial support for Hypothesis 2b (H_2b).

This study provides comprehensive empirical evidence on the dynamic relationship between Environmental, Social, and Governance (ESG) performance and corporate financial outcomes in the agri-food sector across emerging markets. Utilizing a balanced panel dataset of 35 market-leading enterprises across seven developing nations over the 2014-2023 period and employing a robust dynamic System GMM estimation framework, the empirical findings demonstrate that overall ESG performance exerts a statistically significant positive effect on corporate profitability (Pretax ROA).

Disaggregating ESG into individual pillar scores reveals distinct value-creation mechanisms:

1) Governance (GOV) serves as the primary catalyst for financial performance, acting as a vital mechanism to mitigate agency costs and institutional voids inherent in developing capital markets.

2) Environmental (ENV) capabilities contribute positively to operational efficiency and asset productivity, validating the Resource-Based View regarding eco-efficiency gains.

3) Social (SOC) initiatives yield positive but delayed financial benefits, reflecting long-term stakeholder trust-building dynamics.

Furthermore, the results confirm that country-level institutional quality positively moderates the financial returns on ESG adoption, while macro-level shocks such as global health crisis temporarily depressed profit margins across global agri-food supply chains.

Theoretical Contributions

Institutional Contextualization: Extends Stakeholder Theory and the Resource-Based View to emerging economies by demonstrating how firm-level ESG strategies interact with macro-level institutional voids. Dynamic Econometric Modeling: Demonstrates the empirical necessity of controlling for performance persistence, reverse causality, and unobserved firm heterogeneity via System GMM estimators when modeling sustainability dynamics.

Managerial Implications

Prioritization of Corporate Governance: Agri-food executives in emerging markets should prioritize internal governance mechanisms (e.g., board independence, anti-corruption policies, transparent reporting) as the immediate foundation for corporate value creation. Operationalizing Eco-Efficiency: Environmental investments should be tied directly to resource optimization (e.g., precision agriculture, energy efficiency, waste minimization) to yield observable operational cost savings (Asset Turnover). Strategic Stakeholder Engagement: Rather than viewing ESG as a compliance cost, management should leverage sustainability credentials to secure access to international trade networks and lower capital costs from ESG-oriented institutional investors.

Policy and Regulatory Implications

Standardization of ESG Disclosures: Emerging market regulatory authorities should establish standardized, transparent ESG reporting guidelines aligned with international frameworks (e.g., ISSB/GRI) to lower information asymmetry and attract foreign direct investment. Institutional Support & Incentives: Governments should offer targeted fiscal and regulatory incentives (e.g., green credit lines, tax allowances for sustainable technology adoption) to offset initial compliance costs for agri-food firms transitioning to sustainable production models.

While this study offers robust empirical insights, several limitations highlight avenues for future research:

1) Sample Scope: The sample is restricted to 35 market-leading agri-food enterprises listed in the Refinitiv EIKON database. Future research could expand the dataset to include small and medium-sized enterprises (SMEs) as reporting coverage improves across developing nations.

2) Geographic Generalizability: Although the sample spans seven major emerging markets across four continents, country-specific agricultural policies and local regulatory frameworks warrant deeper single-country or comparative regional studies.

3)Non-Linear Dynamics: Future empirical work can explore potential threshold or non-linear effects (e.g., inverted U-shaped relationships) using dynamic panel threshold regression techniques to pinpoint the exact capital expenditure levels beyond which ESG investments exhibit diminishing returns.