Innovation over ESG Performance? The Trade-Offs of STEM Leadership in Top Sustainable Firms

Abstract

Considered as innovation-oriented, this research was conducted to examine whether STEM-educated CEOs drive better ESG performance. Using OLS regression, this research was conducted using listed companies assessed for their ESG performance on Sustainalytics in 2022 and identified as “top sustainable companies”, encompassing 1039 observations. The findings of this research reveal that STEM-educated CEOs are negatively associated with ESG performance in the top sustainable companies. Robustness analysis was also conducted to prevent endogeneity issues. This study introduces the novel idea of strategic trade-offs in ESG leadership. While STEM leaders drive innovation, their focus might lead to underinvestment in other crucial ESG aspects within already-sustainable firms. In addition, this research offers a contribution to governance and ESG research by bringing new insight on CEO selection for top ESG companies to better consider a balanced skillset beyond technological solutions.

1. Introduction

Corporate sustainability has become the central topic within the business field in recent years. However, its perceived relevance is now facing scrutiny. While prior research has emphasized the importance of prioritizing all stakeholders and a long-term perspective over just shareholder profits and short-term gains (Borralho et al., 2022; Harymawan et al., 2021; Kiron et al., 2013; Visnjic et al., 2025; Yang & Li, 2023), some now argue that practical sustainability efforts are losing their relevance. For instance, as reported by the BBC (2023), figures like former US Vice President Mike Pence and UK Prime Minister Rishi Sunak exemplify this counter-argument, viewing ESG investment as a political wedge issue. They contend it prioritizes ideological gsoals over financial returns and represents excessive governmental involvement rather than sound business practice. In stark contrast, the World Economic Forum (2023) stated that maintaining and truly incorporating ESG analysis is crucial. They argue it is key to understanding holistic risks and opportunities, drives long-term value, and is essential for supporting the transition to a greener, sustainable future. This ongoing debate highlights that, even amid questions of the relevancy issue, the advantage of integrating sustainability into a core business strategy remains significant for long-term resilience, value creation, and navigating evolving global challenges.

Despite the ongoing debate, the topic of sustainability remains in the spotlight. Many countries have responded by mandating that companies publish sustainability reports annually (Z. Li & Jia, 2022). According to KPMG (2024), over the past few years, the percentage of companies reporting their sustainability has risen to 96%, which shows an increasing acknowledgment of sustainability commitments as company leaders declare a renewed focus on sustainability goals. However, mere disclosure is not enough to guarantee ESG performance, highlighting the need for a standardized, forward-looking assessment for investors (Charlin et al., 2022; Jámbor & Zanócz, 2023; Walter, 2020). Investors currently depend on ESG scores provided by third parties to assess companies’ sustainability initiatives. Consequently, it is pertinent to investigate the underlying factors that contribute to ESG performance.

Previous research indicates that corporate governance is a significant factor influencing companies’ ESG performance (Aguilera et al., 2021; Daugaard & Ding, 2022; Justina & Lantara, 2023; Khan, 2019; Octavio et al., 2025; Zhang et al., 2023). Corporate governance, specifically the top management, is seen as the set of actors that play a role in determining company strategy, including sustainability matters. As the leader of the top management, the chief executive officer (CEO) is central to shaping the firm’s strategic priorities and their integration of sustainability into corporate strategy (Aryani et al., 2025; Jaggia & Thosar, 2021; Putra & Setiawan, 2024; Shu et al., 2024). The leadership also sets the tone for the organization’s culture and future direction (Naaman & Sun, 2022; Yao, 2023). Thus, it is essential to be able to define clear goals and keep companies on track. Therefore, the CEO’s specific capabilities and strategic mindset become critical in effectively steering the organization, especially when integrating new and complex domains like sustainability.

Drawing upon echelon theory, extensive studies have examined how corporate sustainability is influenced by CEO attributes, such as gender, age, tenure, experience, personal traits, and behavioral biases (Aabo & Cristina, 2022; E-Vahdati & Binesh, 2022; Garcia-Blandon et al., 2019; Kind et al., 2023; Kutzschbach et al., 2020; Lazareva, 2022; Liu et al., 2023). However, the study of how educational background drives corporate sustainability is still underexplored. Some research has examined the level of education, university prestige, and financial background of CEOs. Studies examining the impact of a CEO’s science, technology, engineering, and mathematics (STEM) degree remain limited; meanwhile, companies must adapt by incorporating ESG criteria enhancements across all relevant areas to secure their future. The existing literature indicates that embedding innovation development into a company’s strategy aids in effectively managing ESG initiatives (Adomako & Tran, 2022; Chen et al., 2022; Kiron et al., 2013; Truant et al., 2023; M. Wagner, 2010; J. Xu et al., 2021).

According to Siepel et al. (2021) and Cahyono et al. (2024), the innovative nature to better create a culture of innovation in a company can be found in CEOs who have a STEM degree. These leaders are more likely to engage in technological investments, pursue research-driven strategies, and lead firms toward innovation-oriented outcomes (Hsieh et al., 2022). These qualities are generally assumed to enhance firm competitiveness and innovation outcomes. However, whether these innovation-oriented CEOs also excel at advancing ESG performance remains uncertain, particularly in firms that already score highly in sustainability. In companies that already exhibit high sustainability performance, the challenge lies in maintaining their ESG performance. This might require a broader focus that extends beyond mere innovation. Thus, this research was conducted on companies with high sustainability performance.

In this context, STEM CEOs’ innovation-centric character creates a potential trade-off. While this technological focus can benefit specific environmental aspects of ESG, it might inadvertently lead to a situation where resources are disproportionately allocated towards R&D and technology adoption, potentially limiting investment into and attention on other crucial areas of ESG management. Consequently, the assumption that innovation-oriented STEM CEOs automatically translate to superior ESG performance, which is multifaceted, remains empirically uncertain. Referring to this potential trade-off, a pertinent question arises concerning the capacity of STEM-educated CEOs to contribute to the advancement of ESG performance in companies renowned for their high sustainability standards. Therefore, to address this gap, this research investigates the relationship between STEM-educated CEOs and ESG performance within firms identified as leaders in sustainability.

This research is conducted using listed companies assessing their ESG performance on Sustainalytics over the period of 2022 and received negligible and low risk, which were identified as “top sustainable companies” encompassing 1039 observations. By selecting companies that have already demonstrated a strong commitment and effective management of ESG risks, this research aims to explore the extent to which STEM-educated CEOs can contribute to further enhancing or maintaining this high level of ESG performance. The findings show that having STEM-educated CEOs can sometimes lead to lower ESG performance in leading sustainable companies. This suggests that in this group of top sustainable companies, CEOs with a STEM background tend to have a higher ESG risk rating, according to Sustainalytics. This indicates that they might not manage ESG risks as effectively as their non-STEM counterparts within this high-performing group. Importantly, these results have been tested to ensure they remain robust and free from any endogeneity issues. This relationship between STEM-educated CEOs and ESG performance remains strong across various robustness tests. Additionally, the findings indicate that this result is pronounced in the early-adopter countries, developed nations, and companies led by female CEOs.

This research provides some contributions. First, it challenges the prevailing view that STEM-educated CEOs uniformly enhance firm performance by introducing the notion of a strategic trade-off in highly sustainable firms. Second, it extends upper-echelon theory by highlighting how the CEO’s background interacts with the ESG context. Third, it provides practical insights for boards and policymakers about the importance of aligning CEO capabilities with firm-level ESG maturity. Our findings suggest that sustaining ESG leadership may require broader managerial competencies beyond innovation-focused leadership.

The subsequent sections of this paper are structured to offer a thorough examination of the research topic as follows. Section 2 explains the relevant literature and theoretical frameworks, thereby establishing the foundation for the research hypothesis. Subsequently, Section 3 elucidates the research methodology and variable measurements, while Section 4 encompasses the empirical results, analysis, and discussions. Lastly, Section 5 concludes this study.

2. Literature Review and Hypothesis Development

2.1. STEM-Educated CEO

The CEO is the leader of the board of executives (Putra & Setiawan, 2024). CEOs sits at the apex of an organization, wielding significant influence over its strategic direction, resource allocation, and overall culture (Aryani et al., 2025; Harymawan et al., 2023; Hrazdil et al., 2021; Jaggia & Thosar, 2024; Setiawan et al., 2024; Z. Xu & Hou, 2021). In leading sustainable companies, the CEO’s role goes beyond financial metrics, including a commitment to ESG principles. These leaders often set ambitious sustainability targets, promote ESG integration across all business functions, and ensure accountability for achieving their goals. Their leadership is characterized by a long-term vision that recognizes the interconnectedness of financial success and responsible corporate citizenship (Yi et al., 2021; Zaman et al., 2020), often driving a proactive approach to risk management and pursuing innovative solutions for a sustainable future.

Recognizing that achieving strong ESG performance increasingly demands innovation (Setiawan et al., 2024; J. Xu et al., 2021), this research focuses on CEOs with a STEM educational background. A STEM-educated CEO is defined as a CEO with a degree in one or more of these disciplines. Prior research conducted by Rizki et al. (2024), Siepel et al. (2021), Alderman et al. (2022), Cahyono et al. (2024), Rodríguez-Becerra and Pernaa (2023), S&P Global (2023) reveals that CEOs with STEM backgrounds often emphasize and excel in fostering technological innovation and applying data-driven, measurable solutions.

2.2. Sustainalytics’ ESG Performance

Prior studies have assessed ESG performance primarily through company disclosures (Harymawan et al., 2021; Mohammad & Wasiuzzaman, 2021; Pulino et al., 2022; Saha et al., 2023). However, this approach has limitations, as disclosures may not accurately reflect a company’s true ESG performance and can be susceptible to greenwashing practices (Jámbor & Zanócz, 2023). To address these concerns, this study employs a performance measurement approach that utilizes third-party ESG ratings. ESG ratings offer a more condensed and analytical assessment, making them a valuable tool for evaluating companies’ ESG performance (Khandelwal et al., 2023; White, 2015). Numerous rating agencies, such as Bloomberg, MSCI, Sustainalytics, CDP, ISS, RobecoSAM, and Corporate Knights, each with their own criteria and standards, evaluate companies’ ESG performance (Charlin et al., 2022; Madison & Schiehll, 2021).

Identifying and understanding their material ESG issues, specifically the environmental, social, and governance risks they are exposed to, are crucial for companies (Eccles & Serafeim, 2013). By effectively managing these risks, companies can significantly improve their operational efficiency, effectiveness, and overall ESG performance. This targeted approach aligns well with the strengths of Sustainalytics’ ESG risk score. This rating, offered by Sustainalytics from Morningstar, was chosen as the primary ESG rating agency for this research due to its comprehensive approach (Mandas et al., 2023). Sustainalytics evaluates a company’s exposure to industry-specific ESG risks and its risk management practices. This dual approach offers insights for investors. Their ESG risk rating measures a company’s effectiveness in managing ESG risks, leading to a score that reflects its risk management capability.

Sustainalytics’ ESG risk rating goes beyond a simple numerical score (Walter, 2020). It also categorizes companies based on their risk profile into five categories, ranging from the best to the worst (Zioło et al., 2023). Companies with negligible risk (0–10) are best at managing to minimize their company’s ESG risks. Those with low risk (10–20) have relatively low exposure. The medium-risk category (20–30) indicates moderate ESG risks needing attention. Companies with high risk (30–40) face significant challenges, while severe risk (above 40) indicates very high ESG risk. Following Paramayuda et al. (2025), we classified the companies that received low or negligible risk scores as “top sustainable companies.” These companies exhibit strong commitment and competence in managing their ESG risks.

2.3. Theoretical Framework and Hypothesis Development

The relationship between STEM-educated CEOs and ESG performance can be theoretically grounded in several organizational perspectives. This research utilizes the upper-echelon perspective and posits that an organization’s strategic decisions and performance outcomes directly reflect the cognitive foundations, values, and experiences of its top management team (Hambrick & Mason, 1984). This theory suggests that these background characteristics shape a leader’s unique cognitive foundation, which in turn influences how they interpret information and perceive threats and opportunities, thereby creating a distinct approach for decision-making strategies. Previous research confirms this theory regarding how the backgrounds of top management influence their financial decision-making processes (De Silva & Banda, 2022; Minh Ha et al., 2021; Nawaz, 2022; Setiawan et al., 2024; Shao et al., 2020; L. Xu, 2024). In the context of sustainability, prior research has confirmed that the decision-making outcomes are not solely about financial outcomes but also non-financial outcomes (Kind et al., 2023; Saha et al., 2023; A. Wagner & Fischer-kreer, 2024).

Examining the traits of STEM-educated CEOs uncovers potential trade-offs related to their ESG performance. Referring to the upper-echelon theory, a STEM-educated CEO imbues the organization with a cognitive base that favors an innovation-driven strategy. In the context of sustainability, the characteristics associated with STEM education often led to innovations with a stronger environmental focus (Roos, 2015; Siepel et al., 2021; Zizka et al., 2021). However, this STEM-shaped cognitive base, while beneficial for specific aspects of sustainability, can inadvertently create a strategic trade-off, particularly within top sustainable companies. Once a company has managed its risk to a low level, achieving further improvements in ESG performance may demand more subtle and resource-heavy initiatives in the less tech-centric aspects of social and governance (Martínez-Martínez et al., 2024). This is where the upper-echelon theory-driven ‘selective perception’ becomes critical: the expertise and cognitive comfort of STEM CEOs may lead them to continue prioritizing investments and strategic focus on familiar technological solutions and innovations, even when the marginal gains in ESG would be greater from addressing social and governance factors. This strategic bias, stemming from their unique cognitive framework, might result in reduced attention to and, thus, a greater allocation of resources (both financial and human capital) toward R&D and environmental upgrades, rather than a comprehensive enhancement of all ESG pillars. As a result, this may increase the overall ESG risk rating. Based on these considerations, we developed research Hypothesis 1 as follows:

Hypothesis 1.

STEM-educated CEOs are associated with corporate ESG performance in leading sustainable companies.

3. Research Methodology

3.1. Sample Selection

The research sample consists of listed top sustainable companies according to Sustainalytics during 2022. These top ratings correspond to companies deemed “negligible risk” and “low risk” by Sustainalytics. This specific year was chosen due to being the most recent publicly available data at the time of data collection. Initially, we gathered a sample of 2045 observations. However, following a rigorous data-cleaning process to remove entries with missing values, the final sample size for analysis was reduced to 1039 observations. Details of the criteria applicable to the final sample are provided in

Table 1. Sample selection criteria.

Description	Total
Top-rated Listed companies rated by Sustainalytics in 2022	2045 observations
Less:
Missing data: STEM	520 observations
Missing data: CEOAGE	228 observations
Missing data: FSIZE	65 observations
Missing data: BSIZE	85 observations
Missing data: FAGE	30 observations
Missing data: ROA	27 observations
Missing data: LIQUIDITY	51 observations
Final Sample	1039 observations

.

Table 2. Data tabulation.

Panel A: Data Tabulation of STEM CEO across industry
Industry		non-STEM		STEM		Total
		N	%	N	%	N	%
1	Precious Metals	1	17%	5	83%	6	100%
2	Homebuilders	1	20%	4	80%	5	100%
3	Semiconductors	7	23%	24	77%	31	100%
4	Technology Hardware	43	39%	68	61%	111	100%
5	Paper and Forestry	2	40%	3	60%	5	100%
6	Software and Services	54	46%	63	54%	117	100%
7	Machinery	9	47%	10	53%	19	100%
8	Utilities	9	50%	9	50%	18	100%
9	Chemicals	7	50%	7	50%	14	100%
10	Electrical Equipment	7	50%	7	50%	14	100%
11	Construction and Engineering	2	50%	2	50%	4	100%
12	Oil and Gas Producers	1	50%	1	50%	2	100%
13	Containers and Packaging	9	53%	8	47%	17	100%
14	Traders and Distributors	9	53%	8	47%	17	100%
15	Pharmaceuticals	19	54%	16	46%	35	100%
16	Building Products	7	58%	5	42%	12	100%
17	Healthcare	29	63%	17	37%	46	100%
18	Auto Components	19	63%	11	37%	30	100%
19	Energy Services	2	67%	1	33%	3	100%
20	Real Estate	78	68%	36	32%	114	100%
21	Food Retailers	11	69%	5	31%	16	100%
22	Telecommunication Services	12	75%	4	25%	16	100%
23	Consumer Durables	33	80%	8	20%	41	100%
24	Commercial Services	57	81%	13	19%	70	100%
25	Transportation	18	82%	4	18%	22	100%
26	Media	50	83%	10	17%	60	100%
27	Textiles and Apparel	30	83%	6	17%	36	100%
28	Consumer Services	21	84%	4	16%	25	100%
29	Retailing	76	85%	13	15%	89	100%
30	Transportation Infrastructure	11	85%	2	15%	13	100%
31	Food Products	8	89%	1	11%	9	100%
32	Diversified Financials	18	90%	2	10%	20	100%
33	Automobiles	4	100%	0	0%	4	100%
34	Household Products	3	100%	0	0%	3	100%
35	Insurance	3	100%	0	0%	3	100%
36	Refiners and Pipelines	2	100%	0	0%	2	100%
37	Pharmaceuticals	3	100%	0	0%	3	100%
Total		675	64%	377	36%	1052	100%
Panel B: Data Tabulation of Risk category across continent
Continent		Risk Category
		Negligible Risk		Low Risk		Total
		N	%	N	%	N	%
America		37	8%	454	92%	491	100%
Asia		12	9%	128	91%	140	100%
Africa		2	67%	1	33%	3	100%
Europe		45	12%	334	88%	379	100%
Oceania		4	15%	22	85%	26	100%
Total		100	10%	939	90%	1039	100%

Panel A presents the sample distribution of CEO education, specifically STEM majors by each industry, showing a notable disparity in the representation of STEM and non-STEM CEOs. Of the total 1039 observations as the final sample, 675 observations (64%) are classified as non-STEM, while 377 observations (36%) fall under the STEM category. This indicates that non-STEM backgrounds remain dominant among corporate leaders. The table also highlights the varying levels of STEM-educated CEO representation across different industries. Precious Metals leads the pack with an impressive 83% of STEM CEOs, followed by Homebuilders at 80% and Semiconductors at 77%. These findings align with the expectation that industries heavily reliant on technological innovation would favor STEM CEOs’ leadership. In contrast, certain industries exhibit a lower proportion of STEM CEOs. Transportation holds such a low percentage at 18%, followed by Textiles and Apparel at 17%, and the lowest at 0% by Automobiles, Household Products, Insurance, Refiners and Pipelines, and Pharmaceuticals.

Sample distribution of risk category across the top sustainable companies is presented in Table 2 Panel B. Out of 1039 observations, 10% of the sample (100 observations) has negligible risk. On the other hand, 90% of the sample (939 observations) has low risk. From 5 continents of top company headquarters’ location, continents that have highest percentage of negligible risk include Africa, with 67% of the companies in this continent having negligible risk, followed by Oceania and Europe, with 12% and 15%, respectively. Meanwhile, for low risk, America and Asia exhibit the highest concentration, with 92% and 91% of their respective top sustainable companies falling into this category. Europe and Oceania also show a significant majority of their top sustainable companies classified as low risk, with 88% and 85%, respectively. Notably, Africa has the lowest percentage of companies in the low-risk category at 33%, which is a direct inverse of its high proportion of negligible-risk companies.

3.2. Variable Definition and Measurement

Dependent variable. In this study, we employ ESG risk rating as a dependent variable (RATING). ESG risk rating is determined using the ESG risk score developed by Sustainalytics. The higher the ESG risk rating companies have, the greater portion of their unmanaged ESG risk remains, and the lower the ESG risk score, the better the management is at handling the risks. For easier interpretation, we multiplied the score by negative 1 (−1). As Sustainalytics classifies risk rating into 5 categories based on the assessed score and the aim of this study is to analyze top-rated firms’ scores, we solely include firms that had negligible and low risk from the company assessed by Sustainalytics in 2022.

Independent variable. The independent variable in this research is STEM CEOs (STEM), measured from the educational background of CEOs. To determine whether a company’s CEO possesses a STEM background, we manually gathered data from two primary sources: company annual reports and the LinkedIn website. Borrowing the approach from Alderman et al. (2022), Hsieh et al. (2022), Kong et al. (2023), once the educational background of each CEO was ascertained, a dummy variable was created to represent STEM CEO status (Alderman et al., 2022; Hsieh et al., 2022; Kong et al., 2023). This dummy variable assigned a value of 1 to CEOs with STEM degrees and a value of 0 otherwise.

Control Variables. Various controls were added to account for CEO, auditor, firm, governance, and finance factors that could affect the relationship between STEM and RISK_SCORE. Building on previous research regarding governance and sustainability, we adopted a similar dimensional approach in selecting control variables. This includes CEO characteristics (CEOAGE) to account for relevant factors, governance (BSIZE) to address potential mechanisms impacting the association of the primary variables, firm characteristics (FSIZE; FAGE) to consider the firm’s capabilities that could affect this relationship, financial elements (ROA; LIQUIDITY) to assess the economic conditions of the firm, and the economic conditions of the countries (GDP) to regulate country conditions that may also affect this association. Furthermore, fixed effects are applied to account for unobserved characteristics related to industry categories and country differences within the sample that could influence this association. The operational variables used in this study are summarized in

Table 3. Variable definition and measurement.

Variable	Definition	Measurement	Source
Independent Variable
ESG Rating	RISK_SCORE	ESG Risk Rating Score then multiplied it with (−1)	Sustainalytics
Dependent Variable
STEM CEOs	STEM	Dummy variable of CEOs’ educational background set to 1 if CEOs obtained one or more degree from STEM major and 0 otherwise.	Annual report and Companies website
Control Variables
CEOs’ age	CEOAGE	Natural logarithm of CEOs’ AGE	Annual report and Companies website
Board Size	BSIZE	Natural logarithm of number of Board in the firm	Bloomberg
Firm age	FAGE	Natural logarithm of firm age calculated from the point of the company’s initial public offering	Bloomberg
Firm size	FSIZE	Natural logarithm of total assets	Bloomberg
Return of assets	ROA	Ratio of EBIT to total assets	Bloomberg
Liquidity ratio	LIQUIDITY	Ratio of current assets to current liabilities	Bloomberg
Gross Domestic Product	GDP	Natural logarithm of Gross Domestic Product	World Bank

.

3.3. Model Specification

This research employs a range of statistical techniques to analyze the relationship between STEM CEOs and ESG risk rating. These techniques include descriptive statistics, univariate analysis, and least squares regression analysis. Before conducting these analyses, the data undergo winsorization to address potential outliers that could distort the results. Winsorization involves capping extreme values within a specified range to minimize their influence on the overall data distribution. This ensures that the analyses are conducted on a more robust and representative dataset (Reifman & Garrett, 2010). After winsorizing the data to 1% and 99%, the regression test was conducted. The equation model used in this study is shown below:

$${SCORE}_i=\alpha +{\beta }_1{STEM}_i+{\beta }_{2\text{–}7}{CONTROLS}_i+INDUSTRYFE+COUNTRYFE+\epsilon$$

4. Findings and Discussions

4.1. Descriptive Statistics and Correlation Analysis

Table 4. Descriptive statistics.

Variable	Mean	Standard Deviation	Minimum	Median	Maximum
RISK_SCORE	15.065	3.367	4.300	15.500	20.000
STEM	0.357	0.479	0.000	0.000	1.000
CEOAGE	4.030	0.132	3.367	4.043	4.500
BSIZE	20.565	10.298	2.000	19.000	111.000
FAGE	2.812	1.000	0.000	3.045	4.804
FSIZE	14.938	2.820	6.987	15.232	21.882
ROA	0.216	1.937	−2.780	0.070	59.404
LIQUIDITY	3.408	48.992	0.000	0.677	1564.748
GDP	15.343	1.621	8.704	15.220	17.053

presents a descriptive summary of the variables used in the regression model. The dependent variable, ESG risk rating (RISK_SCORE), has a mean of 15.154. This indicates an average risk rating across the top-rated companies in our sample. However, the minimum value of 4.3 suggests that even these top companies have not entirely eliminated ESG-related risks. The maximum value is capped at 20, reflecting our selection criteria of “top-rated” companies. Turning to the independent variable, the STEM CEOs (STEM) variable has a mean of 0.336. In other words, approximately 33.6% of the top sustainable companies have CEOs with a STEM background (science, technology, engineering, or mathematics). This finding sets the stage for our regression analysis, which will investigate the potential association between a CEO’s STEM education and ESG performance, specifically within the context of these top sustainable companies. A strong focus on innovation, technological advancement, and rapid growth leads to a relatively lower emphasis on further incremental improvements in ESG, especially in companies that already have a good baseline.

Table 5. Pearson correlation.

	Variables	[1]	[2]	[3]	[4]	[5]
[1]	RISK_SCORE	1.000
[2]	STEM	−0.052 *	1.000
		(0.097)
[3]	CEOAGE	0.028	0.031	1.000
		(0.372)	(0.311)
[4]	BSIZE	−0.139 ***	−0.017	0.107 ***	1.000
		(0.000)	(0.586)	(0.001)
[5]	FAGE	0.111 ***	0.031	0.133 ***	0.124 ***	1.000
		(0.000)	(0.316)	(0.000)	(0.000)
[6]	FSIZE	0.056 *	−0.005	−0.041	−0.138 ***	0.003
		(0.072)	(0.865)	(0.190)	(0.000)	(0.928)
[7]	ROA	−0.044	−0.010	0.036	−0.011	−0.055 *
		(0.160)	(0.748)	(0.248)	(0.713)	(0.076)
[8]	LIQUIDITY	0.059 *	−0.025	−0.016	0.005	−0.016
		(0.058)	(0.412)	(0.612)	(0.883)	(0.609)
[9]	GDP	−0.101 ***	−0.066 **	0.100 ***	0.254 ***	0.065 **
		(0.001)	(0.033)	(0.001)	(0.000)	(0.035)
	Variables	[6]	[7]	[8]	[9]
[6]	FSIZE	1.000
[7]	ROA	−0.053 *	1.000
		(0.090)
[8]	LIQUIDITY	0.016	−0.004	1.000
		(0.603)	(0.898)
[9]	GDP	−0.508 ***	−0.004	0.024	1.000
		(0.000)	(0.907)	(0.443)

p-values in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.

presents a correlation matrix for the entire sample, summarizing the relationships between variables. This table allows us to assess the potential linear associations between variables before conducting further analysis. The correlation coefficient between STEM and SCORE is −0.052 with a t-value of 0.097, indicating an association between CEOs with STEM backgrounds and higher ESG risk rating scores. However, correlation does not imply causation, and further investigation is required to understand the underlying mechanisms.

Independent-samples t-test is conducted in order to compare the average value of the variables in companies led by CEOs with STEM backgrounds and non-STEM-educated CEOs. This analysis aims to investigate potential differences in the ESG performance score based on the CEO’s educational background. The results, as shown in

Table 6. Independent t-test.

Variables	MEAN		Coeff	t-Value
	STEM CEO	Non-STEM CEO
RISK_SCORE	−15.298	−14.936	−0.362 *	−1.662
CEOAGE	4.037	4.026	0.011	1.280
BSIZE	20.332	20.695	−0.363	−0.544
FAGE	2.854	2.789	0.065	1.004
FSIZE	14.918	14.949	−0.031	−0.170
ROA	0.190	0.231	−0.040	−0.322
LIQUIDITY	1.733	4.338	−2.605	−0.821
GDP	15.199	15.423	−0.224 **	−2.139

t statistics in parentheses. * p < 0.1, ** p < 0.0.

, reveal that, on average, companies with STEM CEOs have higher ESG risk scores than companies with non-STEM CEOs (−15.298). The mean difference is considered to be significant.

4.2. Baseline Regression Analysis

In examining the sample to obtain the result for the hypothesis, OLS regression is employed.

Table 7. Baseline regression analysis.

Variables	(1)	(2)	(3)
	RISK_SCORE	RISK_SCORE	RISK_SCORE
STEM	−0.362 *	−0.451 **	−0.361 *
	(−1.71)	(−2.14)	(−1.70)
CEOAGE		0.963	0.829
		(1.20)	(1.03)
BSIZE		−0.046 ***	−0.048 ***
		(−4.09)	(−4.13)
FAGE		0.438 ***	0.330 ***
		(4.49)	(3.32)
FSIZE		−0.013	−0.050
		(−0.33)	(−1.02)
ROA		−0.071 ***	−0.062 ***
		(−4.79)	(−3.85)
LIQUIDITY		0.004 ***	0.004 ***
		(6.92)	(5.20)
GDP		−0.186 **	−0.532 **
		(−2.45)	(−2.32)
Constanta	−14.936 ***	−16.024 ***	−9.524 *
	(−110.73)	(−4.69)	(−1.91)
Industry FE	No	No	Yes
Country FE	No	No	Yes
R2	0.003	0.052	0.153
Adjusted R2	0.002	0.045	0.106
N	1039	1039	1039

t statistics in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.

presents the results of the regression analysis. Column (1) presents the regression result for STEM and RISK_SCORE. The findings show a negative association between STEM and RISK_SCORE at a significance level of 10% (coeff = −0.362, t = −1.71); Column (2) shows a consistent result after the control variables are employed at a significance level of 5% (coeff = −0.451, t = −2.14); Column (3) shows the regression result after including the industry fixed effect and country fixed effect. The results show negative significance at a level of 10% (coeff = −0.361, t = −1.70).

All columns in the regression models show consistent results of a negative relationship between STEM-educated CEOs and ESG performance. Thus, the result supports the hypothesis that there is an association between a STEM-educated CEO and ESG performance. This negative result indicates that within the context of top sustainable companies, having an innovation-oriented CEO with a STEM background is associated with a higher ESG risk rating, meaning worse performance. This finding aligns with the trade-off perspective, suggesting that the strategic focus on technological innovation and advancement, often prioritized by STEM leaders, might lead to less emphasis on other crucial aspects of ESG management in companies that have already established a strong baseline (Truant et al., 2023). This result also supports Visnjic et al. (2025), highlighting a potential inclination towards internally driven innovation, which, in this case, is fostered by STEM expertise, inadvertently creating less openness to valuable external partnerships with businesses and community stakeholders, hindering the comprehensive management of diverse ESG risks.

4.3. Robustness Test

Ensuring that the correlations between STEM-educated CEOs and ESG risk score are free from endogeneity issues that might exist is important. For endogeneity testing, first, a Wu–Hausman test is conducted. The aim is to validate the use of fixed effect as the regression model instead of using the random-effect model.

Table 8. Wu–Hausman test.

Variables	RISK_SCORE
	(1)	(2)
	Fixed-Effect	Random-Effect
STEM	−0.347	−0.451 **
	(−1.587)	(−2.108)
CEOAGE	0.957	0.963
	(1.220)	(1.225)
BSIZE	−0.042 ***	−0.046 ***
	(−4.038)	(−4.412)
FAGE	0.391 ***	0.438 ***
	(3.707)	(4.213)
FSIZE	−0.041	−0.013
	(−0.953)	(−0.310)
ROA	−0.064	−0.071
	(−1.218)	(−1.340)
LIQUIDITY	0.003	0.004 **
	(1.595)	(2.061)
GDP	−0.131 *	−0.186 **
	(−1.703)	(−2.466)
Constanta	−16.416 ***	−16.024 ***
	(−4.826)	(−4.705)
N	1039
chi2(8)	20.33
Prob>chi2	0.0092

t statistics in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.

shows the result of the Wu–Hausman test. It shows that the p-value of chi square is 0.0092. As the value is less than 0.05, the null hypothesis is rejected, which means that the use of the fixed-effect model is consistent and more reliable compared to using the random-effect model.

Second, Coarsened Exact Matching (CEM) analysis is conducted. This test is conducted by making the treatment and control groups as similar as possible in their important characteristics by temporarily simplifying those characteristics and then finding exact matches. The observations unmatched to the other group led to a reduction in sample selection.

Table 9. Coarsened Exact Matching.

Panel A: CEM Observation Sample
	Non-STEM	STEM
All	668	371
Matched	667	370
Unmatched	1	1
Panel B: CEM Regression
	(1)
Variables	RISK_SCORE
STEM	−0.358 *
	(−1.69)
CEOAGE	0.830
	(1.03)
BSIZE	−0.049 ***
	(−4.18)
FAGE	0.349 ***
	(3.50)
FSIZE	−0.050
	(−1.02)
ROA	−0.412
	(−1.49)
LIQUIDITY	−0.023 **
	(−2.22)
GDP	−0.525 **
	(−2.31)
Constanta	−9.463 *
	(−1.92)
Industry FE	Yes
Country FE	Yes
R2	0.153
Adjusted R2	0.106
N	1037

t statistics in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.

presents the result of CEM analysis.

Panel A of Table 9 shows the CEM observation sample. During this stage, the sample distribution is clustered across three groups. Consequently, 667 companies with non-STEM CEOs are matched and have similar characteristics to STEM CEOs. Meanwhile, one company remains unmatched with companies with similar characteristics that have STEM CEOs. On the other hand, 370 companies with STEM CEOs are included in the matching process, leaving one company in the unmatched condition. Subsequently, Panel B displays the regression result in the condition where the observations of CEM are matched. The results show a negative association at a significance level of 10% (coeff = −0.358, t-value = −1.69). This result is consistent with the baseline regression results, which indicate that after controlling for potential confounding factors through the matching process, the negative relationship between having a STEM-educated CEO and ESG performance in top sustainable companies remains statistically significant.

Third, to address potential self-selection bias related to the voluntary nature of CEO STEM backgrounds, we apply Heckman two-stage regression (Heckman, 1979). In the first stage, the median of STEM education in the same industry is used as an instrumental variable (Cahyono et al., 2024), under the rationale that firms tend to align with peer characteristics within their industry, even across countries. The results, as presented in

Table 10. Heckman two-stage regression.

Variables	(1)	(2)
	STEM	RISK_SCORE
IND_STEM	0.514 *
	(1.707)
STEM		−0.358 *
		(−1.694)
MILLS		3.032
		(0.500)
CEOAGE	0.381	1.632
	(1.176)	(0.892)
BSIZE	0.002	−0.045 ***
	(0.369)	(−3.237)
FAGE	0.059	0.448
	(1.332)	(1.608)
FSIZE	−0.007	−0.066
	(−0.373)	(−1.144)
ROA	−0.018	−0.102
	(−1.288)	(−1.248)
LIQUIDITY	−0.001	0.001
	(−0.716)	(0.111)
GDP	−0.015	−0.552 **
	(−0.129)	(−2.387)
Constanta	−1.901	−13.536
	(−0.818)	(−1.093)
Industry FE	Yes	Yes
Country FE	Yes	Yes
Pseudo R2	0.071
R2		0.148
N	1039	1039

t statistics in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.

, reveal a significantly positive relationship between the STEM median within the same industry and the likelihood of appointing a STEM-educated CEO, confirming the relevance of the instrument. In the second stage, the presence of a STEM-educated CEO is found to have a negative and significant relationship with corporate risk scores, consistent with the baseline model. This suggests that having a STEM CEO is associated with more measured and conservative risk-taking behavior. Overall, the robustness test reinforces the main findings and suggests that selection bias does not materially affect the observed relationship between STEM CEOs and ESG risk taking.

Lastly, we employ the Propensity Score Matching (PSM) 1-on-1 test to address potential selection bias. A 1:1 non-replacement PSM is performed with a caliper of 1%. The balance test and matching results are reported in Panel A of

Table 11. Propensity Score Matching (PSM).

Panel A: Balance test and matching results
Variables	Unmatched Matched	Mean		%bias	t-Test
		Treated	Control		t	p-Value
CEOAGE	U	4.037	4.026	8.5	1.280	0.201
	M	4.034	4.041	−5.7	−0.718	0.473
BSIZE	U	20.332	20.695	−3.6	−0.544	0.586
	M	20.696	20.556	1.4	0.172	0.863
FAGE	U	2.854	2.789	6.6	1.004	0.316
	M	2.849	2.846	0.3	0.034	0.973
FSIZE	U	14.918	14.949	−1.1	−0.170	0.865
	M	14.802	14.742	2.1	0.261	0.795
ROA	U	0.190	0.231	−2.3	−0.322	0.748
	M	0.172	0.286	−4.7	−0.600	0.548
LIQUIDITY	U	1.733	4.338	−6.0	−0.821	0.412
	M	1.752	1.352	7.4	0.937	0.349
GDP	U	15.199	15.423	−13.8	−2.139	0.033
	M	15.389	15.422	−2.1	−0.263	0.793
Panel B: Relationship of STEM CEOs and ESG risk-taking by PSM samples
Variables			RISK_SCORE
STEM			−0.462 *
			(−1.848)
CEOAGE			1.706
			(1.607)
BSIZE			−0.050 ***
			(−3.354)
FAGE			0.251 *
			(1.941)
FSIZE			−0.037
			(−0.646)
ROA			−0.067 ***
			(−4.957)
LIQUIDITY			−0.019
			(−1.218)
GDP			−0.251
			(−0.796)
Constanta			−15.093 **
			(−2.168)
Industry FE			Yes
Country FE			Yes
R2			0.167
Adjusted R2			0.110
N			644

t statistics in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.

, supported by Figure 1, showing reduced bias and the aligned propensity score. Before matching, indicated by unmatched (U), all variables show large biases between the treated and control groups, with corresponding p-values indicating significant differences. After matching, indicated by matched (M), the percentage bias for all covariates reduces substantially, with a higher p-value, which reflects the absence of significant differences between the treated group and the control group. We then re-estimate the main regression using the treatment and matched control samples, with the results presented in Panel B of Table 11. Panel B shows a positive and statistically significant relationship between STEM and RISK_SCORE at the 1% level (coefficients = −0.462, t-values = −1.848). These findings indicate that there is a negative association between the existence of STEM-educated CEOs and ESG risk taking, even after addressing potential selection bias through PSM. No evidence of sample selection bias is detected in this study, reinforcing the robustness of our hypotheses.

4.4. Additional Analysis

4.4.1. Early-Adopter Countries

Being an early adopter and having developed comprehensive sustainability policies and regulations, European countries have established a significant lead in fostering strong ESG practices among their domestic companies (Iamandi et al., 2019). This proactive regulation has likely cultivated a distinct environment where ESG considerations are deeply embedded in corporate strategy and operations. To understand whether the relationship between STEM-educated CEOs and ESG performance is consistent across different regulatory and market contexts, a sub-sample analysis comparing European and non-European companies is conducted.

Table 12. Sub-sample—European and non-European countries.

	RISK_SCORE
	(1)	(2)
	EU	Non-EU
STEM	−0.841 **	−0.086
	(−2.43)	(−0.32)
CEOAGE	−0.870	1.504
	(−0.58)	(1.58)
BSIZE	−0.075 ***	−0.039 ***
	(−3.55)	(−2.78)
FAGE	0.016	0.521 ***
	(0.09)	(4.05)
FSIZE	−0.069	−0.031
	(−0.82)	(−0.52)
ROA	−0.065 ***	−0.403
	(−3.65)	(−1.59)
LIQUIDITY	−0.022	0.004 ***
	(−1.22)	(6.28)
GDP	1.394 ***	−0.489 **
	(4.15)	(−2.09)
_cons	−31.477 ***	−12.417 **
	(−3.66)	(−2.36)
Industry FE	Yes	Yes
Country FE	Yes	Yes
r2	0.169	0.154
r2_a	0.084	0.108
N	379	660

t statistics in parentheses. ** p < 0.05, *** p < 0.01.

displays the results of sub-sampling European and non-European countries. It shows that in European countries, the relationship between STEM CEOs and ESG performance indeed reveals a negative association (coeff = −0.841, t-value = −2.43), which indicates that, within this mature and regulated ESG landscape, the presence of a STEM-educated CEO is associated with a statistically significantly worse ESG performance among top sustainable European companies. On the other hand, in non-European countries, the result shows no association between STEM-educated CEOs and ESG performance. To assess whether this observed difference in effect across groups is statistically meaningful, a cross-group coefficient comparison was performed. The result indicates a marginally significant difference at a significance level of 10%, suggesting that the regulatory and market context in European countries can play a moderating role in the relationship between STEM CEOs and firms’ ESG risk.

4.4.2. Developed and Developing Countries

The global emphasis on corporate sustainability and ESG performance is not uniformly experienced across different stages of economic development (Lozano, 2015). Having different environmental regulations, stakeholder awareness, and established institutional frameworks for corporate governance may lead to different ESG performance (Mooneeapen et al., 2022). Understanding these differences is crucial for analysis of the relationship between leadership attributes and corporate sustainability outcomes on a global scale. Thus, the sub-sample of developing and developed countries is employed.

Table 13. Sub-sample—developing and developed countries.

	RISK_SCORE
	(1)	(2)
	Developed Countries	Developing Countries
STEM	−0.414 *	0.126
	(−1.85)	(0.18)
CEOAGE	0.687	0.839
	(0.77)	(0.36)
BSIZE	−0.047 ***	−0.021
	(−3.92)	(−0.54)
FAGE	0.368 ***	0.060
	(3.55)	(0.14)
FSIZE	−0.038	−0.363
	(−0.76)	(−1.20)
ROA	−0.058 ***	−0.567
	(−3.65)	(−0.39)
LIQUIDITY	−0.024	0.038
	(−0.93)	(0.95)
GDP	−0.452 **	1.330
	(−2.02)	(1.57)
_cons	−11.080 **	−32.801 *
	(−2.11)	(−1.98)
Industry FE	Yes	Yes
Country FE	Yes	Yes
r2	0.142	0.312
r2_a	0.102	0.067
N	935	104

t statistics in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.

displays the result of the sub-sample analysis. Column (1) presents the result for developed countries. It shows a negative association between STEM-educated CEOs and ESG performance, which aligns with baseline regression. This association might happen because, in developed economies, where ESG pressures are more pronounced and sophisticated sustainability practices are already widespread, the strategic focus of STEM-educated CEOs on radical innovation and technological advancements might inadvertently lead to the relative neglect of the more nuanced and comprehensive aspects of ESG management required to further improve performance and reduce risk, as assessed by stringent third-party ratings. This prioritization could result in a higher ESG risk rating compared to leaders with a broader strategic focus on all ESG pillars. Meanwhile, for developing countries, as shown in Column (2), there is no association between STEM-educated CEOs and ESG performance. Although the relationship appears stronger in developed countries, the insignificant cross-group test implies that the observed difference does not represent a statistically meaningful divergence in the relationship of CEO STEM background on ESG risk taking.

4.4.3. CEO’s Gender

Due to evolutionary and societal factors, women tend to be more empathetic and nurturing, which could lead to better ESG scores when they are in leadership positions (Aabo & Cristina, 2022). By referring to the result of this research, further analysis is conducted. These traits could translate into more effective leadership, particularly in the context of ESG principles. To further explore this potential connection, researchers investigated the influence of STEM CEOs’ gender on company ESG performance. The main analysis is expanded to consider whether and to what extent the gender of STEM CEOs influences a company’s ESG performance, as measured by the rating score from Sustainalytics.

Table 14. Sub-sample—CEO gender.

Panel A: Distribution of STEM-educated CEO by gender
Gender	Non-STEM CEO		STEM CEO		Total
	N	%	N	%	N	%
Female	68	79%	18	21%	86	100%
Male	600	63%	353	37%	953	100%
Total	668	64%	371	36%	1039	100%
Panel B: Sub-sample regression results
Variables		RISK_SCORE
		(1)			(2)
		Female CEO			Male CEO
STEM		−1.542 *			−0.295
		(−1.91)			(−1.35)
CEOAGE		1.809			0.840
		(0.53)			(1.01)
BSIZE		−0.018			−0.049 ***
		(−0.34)			(−4.11)
FAGE		0.264			0.356 ***
		(0.67)			(3.46)
FSIZE		0.161			−0.067
		(0.75)			(−1.31)
ROA		−0.310			−0.065 ***
		(−0.70)			(−3.91)
LIQUIDITY		−0.033			0.004 ***
		(−0.78)			(4.97)
GDP		−1.011			−0.505 **
		(−1.25)			(−2.18)
Constanta		−11.385			−9.620 *
		(−0.68)			(−1.89)
Industry FE		Yes			Yes
Country FE		Yes			Yes
R2		0.410			0.166
Adjusted R2		0.088			0.115
N		86			953

t statistics in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.

displays the results of an analysis aimed at incorporating this potential connection. Panel A presents the distribution of STEM-educated CEOs by gender within the sample. Among the sample, there are 86 female CEOs, while the majority are male CEOs, totaling 953 people. Regarding their major, 21% of female CEOs (18 people) have a STEM education. Compared to this, 37% of all male CEOs (353 people) have a STEM education. This supports the notion of female scarcity in STEM. Panel B displays the results of regression analysis by splitting the sample based on gender. Female STEM CEOs have a negative relationship with ESG risk rating score at a level of 10% (coeff = −1.542, t = −1.912). Meanwhile, male STEM CEOs do not show a significant relationship with the ESG risk rating score. These results may arise from the challenges female STEM CEOs encounter, who, due to their underrepresented status in both STEM fields and top leadership, offer unique perspectives that can lead to a different approach to ESG management. To evaluate whether the observed differences in effect across regions are statistically significant, a cross-group coefficient comparison was conducted. The result shows a significant difference at the 5% significance level.

5. Conclusions

This study investigated the association between STEM-educated CEOs and ESG performance within a sample of top sustainable companies. Our findings, derived from OLS regression, reveal a negative association between having a STEM-educated CEO and ESG performance, which indicates that, contrary to initial assumptions rooted in the innovation-oriented nature often attributed to STEM leadership, companies with STEM-educated CEOs within this high-performing sustainability cohort tend to exhibit worse ESG performance compared to their non-STEM counterparts. The robustness tests conducted support the primary findings and hypothesis.

This research offers several implications through both theoretical and practical contributions. It enhances the expanding body of literature by examining the influence of CEO characteristics on corporate sustainability. By revealing a negative association between STEM-educated CEOs and ESG performance in leading sustainable companies, this study extends the upper-echelon theory by introducing the concept of strategic trade-offs at the leadership level as a critical factor influencing ESG performance, thereby moving beyond a mere focus on cognitive abilities. This study underscores the necessity of considering the contextual fit between leader traits and firm ESG maturity. In top sustainable firms, maintaining ESG excellence may necessitate more inclusive and integrative leadership, extending beyond individual-level innovation orientation.

In regard to practical contributions, this research serves as a valuable reference point for companies when contemplating their strategic priorities in CEO selection. It highlights that an exclusive focus on innovation-oriented leaders may not suffice for sustaining superior ESG performance. Furthermore, our findings suggest that boards should consider a balanced leadership profile beyond mere STEM expertise. Companies led by STEM CEOs should cultivate diverse leadership teams and implement comprehensive ESG monitoring. For policymakers, this underscores the necessity of incentivizing thorough ESG performance and fostering cross-disciplinary expertise in corporate leadership, rather than concentrating solely on technological innovation to achieve sustainability.

We acknowledge the limitations present within our study, particularly concerning the employment of Sustainalytics data as a proxy for ESG performance, which could solely be conducted for the year 2022 due to restricted public accessibility and the lack of available historical databases. Another limitation arises from the geographic distribution of our sample, which shows a high concentration of companies from Asia and Europe. While this reflects the available data on top sustainable companies within our chosen timeframe, we acknowledge that this regional bias may cause generalizability issues. Further research could expand the dataset by engaging alternative third-party ESG performance evaluators and considering the incorporation of personal profile characteristics of CEOs, such as their tenure and duality, to provide a more comprehensive understanding of the influence of leadership on ESG-related activities.