A Study on ESG Evaluation Indicators Through Chemical Accident Data Analysis and Double Materiality Assessment

Abstract

This study focuses on identifying key factors that companies should prioritize to prevent chemical accidents from an ESG (Environmental, Social, and Governance) perspective. ESG provides insight into corporate sustainability by comprehensively considering both external impacts and internal risks. To achieve this, this study applied a scoring approach based on double materiality, assessing both internal and external impacts. The assessment process involved data collection, categorization of accident causes, score calculation, and prioritization of safety management items. Data from chemical accident statistics revealed that mechanical integrity, human factors, and preventive maintenance were the primary causes across all three countries. Internal impacts were evaluated by accident severity and frequency, while external impacts considered casualties and management priorities. Internal impact results showed mechanical integrity as critical in the U.S., preventive maintenance in the U.K., and human factors in Republic of Korea. For external impacts, human factors were most critical in Republic of Korea. To prioritize safety management elements for chemical accident prevention from an ESG perspective, this study categorized them into three tiers. Tier 1 represents the most critical elements requiring urgent attention, while Tier 3 includes the least critical elements. This tier classification is intended not as an absolute ranking but as a general reference for identifying overall trends in safety management priorities. Tier 1 included the U.K.’s preventive maintenance and human factors across all countries, with Republic of Korea’s human factors being the most vulnerable. Tier 2 revealed operating procedures and human factors as critical, with U.S. emergency preparedness and U.K. design highlighted. Tier 3, with impact scores below 1, was safest. These findings effectively identified safety management priorities to enhance accident prevention.

1. Introduction

1.1. Outline

Climate change risk (CCR) resulting from global warming poses a significant threat to humanity and corporate activity, causing both environmental disasters and economic instability. Ref. [1] indicated that the nationally determined contributions (NDCs) announced by each country in line with the 2015 Paris Climate Agreement project a global temperature increase of 2.6–3.1 °C by 2100, suggesting that more stringent NDCs are required to keep this increase below 2 °C. Ref. [2] noted a 98% probability that temperatures will reach record highs between 2023 and 2027 due to El Niño, with a 66% likelihood of global average temperatures surpassing a 1.5 °C increase. The Intergovernmental Panel on Climate Change predicted with high confidence in its 2021 report that a 1.5 °C increase would lead to an increase in the frequency and intensity of heavy rainfall and drought across most continents [3]. Furthermore, it was estimated (with medium confidence) that intense tropical cyclones would increase by 10%, and precipitation associated with tropical cyclones would increase by 11% (with high confidence) [3]. Ref. [4] highlighted that the financial loss from natural disasters in the United States of America (USA) reached $280 billion in 2021, with an increasing likelihood of extreme weather events. The Economist Intelligence Unit estimated potential financial losses due to CCR at approximately $4.2 trillion [5]. Ref. [6] warned that, without effective global warming mitigation policies, global actual gross domestic product (GDP) could decrease by 7.2% by 2100 due to production losses. Ref. [7] estimated that the USA’s GDP could decrease by as much as 10% due to CCR by 2100. Refs. [8,9] found that companies exposed to CCR face higher expected stock returns and loan interest rates than those without CCR concerns. Ref. [10] indicated that increased exposure to extremely hot days could reduce corporate revenue and operating profits. In response to anticipated decreasing profits, companies have adopted corporate social responsibility (CSR) policies [11]. CSR involves accountability for the economic and social impacts of corporate activity, with the goal of promoting sustainable development [12,13].

It is natural that this study focuses more on CSR and ESG than chemical accidents in the introduction. This emphasis originates from the foundational need to analyze corporate practices from an ESG perspective. Although ESG and CSR are often mistakenly considered identical concepts, they are distinct. CSR primarily emphasizes corporate accountability for social impact, while ESG extends this by integrating a company’s sustainable competitiveness and its coexistence with external environmental and social factors. However, since many existing references still predominantly use CSR rather than ESG terminology, this study frequently refers to CSR out of necessity. Nonetheless, it should be clearly stated that, unlike CSR, which often demands maximal social contributions from companies, ESG should be approached as a comprehensive framework for balancing internal sustainable competitiveness with broader external sustainability considerations. As authors, we emphasize this foundational distinction before presenting our research. The following section provides a literature review on corporate sustainability trends and ESG frameworks relevant to chemical accident prevention.

1.2. Literature Review

According to the Governance and Accountability Institute, the percentage of S&P 500 companies submitting sustainability or corporate responsibility reports surged to 86% in 2018, up from 20% in 2011 [14]. By 2021, more than 3800 institutional investors and service providers had signed the Principles for Responsible Investment, incorporating ESG and CSR into investment analysis and decision-making processes [15]. Their assets under management increased significantly, from $6.5 trillion in 2006 to $121.3 trillion in 2021 [15]. The increasing corporate focus on ESG and CSR has attracted greater investor interest, with cash flow into 300 operating mutual funds based on ESG in 2019 reaching $20.6 billion, more than quadruple the amount in 2018 [16].

Given prolonged global supply chain disruptions, such as COVID-19 in 2021 and the Russia–Ukraine War in 2023, economic nationalism has risen, especially in the USA and Europe [17,18,19]. This shift has increased demands for corporate social responsibility (CSR), influencing economies and leading to regulations like the Corporate Sustainability Reporting Directive (CSRD) [17,20]. The CSRD expands the EU’s 2014 Non-Financial Reporting Directive, requiring public small and medium-sized enterprises to report CSR activities [21]. Recent CSR studies emphasize the concept of double materiality, assessing both internal and external corporate impacts [22]. ESG financial data focus on materiality, which refers to information that influences investor decision-making [23,24]. Financial materiality affects a company’s finances, while stakeholder materiality involves broader societal impacts, such as government, employees, and communities [23]. Double materiality merges both aspects, making it vital for evaluating sustainability and corporate social responsibility in ESG assessments. It is increasingly applied in due diligence to assess direct and indirect social responsibilities [22]. Direct responsibility involves employee job safety and HR policies, while indirect responsibility covers supply chain safety and environmental protection [13,25]. With the rising importance of occupational safety under CSR, various studies on occupational health and safety (OHS) have emerged. Ref. [26] emphasized the need to integrate industrial safety into CSR using ISO 45001. Ref. [27] explored how internal CSR influences employee OHS, while Ref. [28] highlighted collaborations between external stakeholders and companies practicing CSR. Ref. [13] reviewed CSR reports from three industries (aviation, energy, and finance) and noted a gradual increase in OHS content, which reached 10% by 2011, focusing on occupational health (44%), industrial health (33%), and employee welfare (22%). Ref. [29] analyzed CSR reports from 41 airlines and identified 12 OHS indicators related to working conditions. Ref. [30] found that OHS was often categorized under human resources without separate evaluation in 58 energy companies across Norway, Denmark, and the USA. Ref. [31] reported that only 1 out of 40 insurance companies addressed an OHS indicator related to employee safety. Ref. [32] revealed low OHS scores (30–35%) in CSR reports from 19 multinational construction companies. Ref. [33] noted the frequent use of OHS indicators in the petroleum, gas, mining, electrical, construction, and chemical sectors, with 31 out of 94 companies following the GRI G4 sustainability reporting guidelines for assessing OHS using 10 criteria [32,34]. However, Ref. [35] indicated that social aspects in CSR assessments using GRI can sometimes be overestimated.

Table 1. Summary of key findings from previous studies on occupational health and safety (OHS) in CSR and ESG Contexts.

Key Focus	Key Findings
Integration of Industrial Safety into CSR [26]	Emphasized the need to integrate industrial safety into CSR using ISO 45001.
Internal CSR’s Impact on Employee OHS [27]	Explored how internal CSR practices influence employee occupational health and safety.
CSR and External Stakeholder Collaboration [28]	Highlighted collaborations between companies practicing CSR and external stakeholders.
CSR Reporting Trends (Aviation, Energy, Finance) [13]	OHS content gradually increased to 10% by 2011, focusing on occupational health (44%), industrial health (33%), and employee welfare (22%).
CSR Reports in the Airline Industry [29]	Analyzed CSR reports from 41 airlines and identified 12 OHS indicators related to working conditions.
OHS Classification in Energy Sector [30]	Found that OHS was often categorized under human resources without separate evaluation in 58 energy companies across Norway, Denmark, and the USA.
OHS Indicators in Insurance Sector [31]	Reported that only 1 out of 40 insurance companies addressed an OHS indicator related to employee safety.
OHS Performance in Multinational Construction [32]	Revealed low OHS scores (30–35%) in CSR reports from 19 multinational construction companies.
GRI G4 Compliance Across Industries [33]	Noted the frequent use of OHS indicators in the petroleum, gas, mining, electrical, construction, and chemical sectors, with 31 out of 94 companies following the GRI G4 guidelines.
Limitations of GRI in CSR Evaluation [35]	Pointed out potential overestimation of social aspects in CSR assessments using GRI.

and

Table 2. Key elements and descriptions for evaluating occupational health and safety (OHS) in chemical plants.

Key Elements	Description
Chemical Plant Complexity	Numerous facilities in confined space with highly complex structures.
Hazardous Chemical Risks	Fire, explosion, and toxic gas leaks due to high temperatures and pressures.
Primary Accident Escalation	Primary incidents can lead to domino effects causing loss of life.
Vulnerability of the Chemical Industry	Higher susceptibility to human, environmental, and financial harm.
OHS Management Necessity	Management of both direct and indirect stakeholders using detailed safety indicators.
Lagging Indicators	Reflect past incidents such as accident rates and fatalities.
Leading Indicators	Reflect proactive safety measures taken to prevent accidents.

consolidate and present the summarized findings from the preceding discussions.

A chemical plant consists of numerous facilities within a confined space, creating a highly complex structure [36,37]. Hazardous chemicals stored or transported under high temperatures and pressures pose fire, explosion, and toxic gas leak risks [38]. Such primary accidents can escalate into domino effects, causing significant loss of life, environmental damage, and financial losses [39,40,41]. The chemical industry, more vulnerable to such risks than other sectors, faces heightened susceptibility to human, environmental, and financial harm [42]. To mitigate these risks, managing OHS for both direct and indirect stakeholders requires detailed safety indicators. Evaluations should cover both lagging indicators, such as accident and fatality rates, and leading indicators, which reflect proactive safety measures taken by companies.

However, current OHS evaluations focus primarily on 10 lagging indicators based on the GRI G4 guidelines, centering on injury rates due to chemical accidents, OHS management systems, and third-party disclosure verification [32]. Therefore, the leading indicators reflecting preventive safety measures have not been quantitatively assessed. Examining OHS component evaluation trends of major global and Republic of Korean ESG rating agencies reveals the challenges that companies may face when adapting to these criteria [43]. We reviewed the ‘evaluation factors’ used by major ESG rating bodies and extracted the elements related to ‘safety’ among them. These elements were organized and summarized, as presented in

Table 3. Safety assessment factors in major ESG rating bodies.

Major ESG Rating Bodies	Safety Factors in ESG Evaluation	Description
MSCI	Workplace safety program	Evaluates the effectiveness of workplace safety policies and programs to prevent accidents and injuries
	Health and safety management system	Assesses the presence and quality of certified health and safety management systems (e.g., ISO 45001)
	Accident and incident rate	Monitors and compares rates of workplace accidents and incidents relative to industry benchmarks
	Compliance with safety regulation	Ensures adherence to local and international safety regulations and standards
	Safety training and education	Measures the extent and frequency of safety training programs provided to employees
Refinitiv	OHS policy	Evaluates company OHS policies
	Lost-time injury rate	Measures the rate of work-related injuries that result in lost time
	Employee safety training program	Assesses the availability and effectiveness of safety training programs for employees
	Workplace safety certification	Checks for certifications related to workplace safety, such as ISO 45001
	Safety performance metrics	Tracks and reports various safety performance metrics, including near-misses and minor incidents
Republic of Korea Institute of Corporate Governance and Sustainability (KCGS)	Occupational safety	Assessment of workplace safety measures and accident prevention strategies
	Employee health and well-being	Evaluation of health programs and initiatives aimed at employee wellness
	Safety training and education	Effectiveness of safety training programs provided to employees
	Safety management systems	Implementation and effectiveness of systematic approaches to manage workplace safety
	Compliance with safety regulations	Adherence to local and international safety standards and regulations
Republic of Korean Ministry of Trade, Industry and Energy (K-ESG)	Industrial safety standard	Compliance with industry-specific safety standards
	Hazardous material handling	Safety measures for handling and storing hazardous materials
	Safety audits and inspection	Regular audits and inspections for safety compliance

. International ESG rating agencies tend to evaluate companies based on their safety management systems, safety programs, and adherence to safety performance indicators. However, guidance regarding specific safety programs, or the level of safety performance indicators that companies must establish and manage, is lacking. In Republic of Korea, assessments primarily focus on the ‘minimum essential safety measures’ required by law [44]. Consequently, it is challenging for companies in the chemical industry, where individual sites may have complex processes and diverse risk factors, to reflect on voluntary safety management achievements aligned with their specific realities within ESG evaluations [45].

To derive key prevention factors based on an ESG assessment perspective, this study attempted a safety-score evaluation formula based on the concept of double materiality and proposed prioritizing safety management components applicable to companies in the chemical industry. The safety-score formula considers both internal corporate losses and external (social) impacts of chemical accidents [46]. To evaluate the performance of the developed safety-score formula, key management factors were derived from chemical accident data from three countries (the USA, the United Kingdom (UK), and Republic of Korea). This analysis enabled the identification of specific accident causes (safety vulnerabilities), and safety management priorities were presented using a three-tier classification based on internal and external impact assessment results [47].

2. Method and Materials

This study followed a structured sequence: data collection and classification, formula development and application, and priority derivation of safety management components.

2.1. Chemical Accident Data Collection by Country

For this study, data on chemical accident occurrence patterns, causes, and casualties were collected based on the National Toxic Substance Incidents Program (NTSIP) of the US Agency for Toxic Substances and Disease Registry (ATSDR), the Hydrocarbon Release Database (HCRD) from the UK Health and Safety Executive (HSE) [48], and chemical accident data from Republic of Korea’s NICS. To focus on the chemical industry, the chemical accident data for the USA were limited to the chemical manufacturing industry, as classified by NAICS Code 325UK and Republic of Korea. Data with clearly identifiable causes and types of accidents were selected from the HCRD and NICS, respectively. The original intention of this study was to collect chemical accident data from multiple countries to provide a comprehensive comparative analysis. North America and European countries were chosen for their advanced accident safety management practices compared to Republic of Korea. However, to ensure public accessibility and avoid bias from proprietary datasets, only publicly available data were included in this study. Furthermore, among various national datasets, only those containing clear and well-documented accident causes and impact details were selected for inclusion [49]. While many countries provide chemical accident statistics, datasets with detailed causal analysis and impact severity were limited. This restriction ultimately led to the selection of data from the USA, UK, and Republic of Korea for the final analysis [50]. Regarding the time frames, the periods of data collection were determined based on the latest data availability from each source. The datasets from the USA (2010–2014), UK (2016–2021), and Republic of Korea (2014–2022) were chosen to maximize the data coverage up to the most recent records available [51]. Each data source provides records based on different timeframes and data collection standards. Since this study does not aim for direct pairwise comparisons between countries but instead analyzes general trends and key factors within each dataset, standardizing the timeframes across all countries was considered impractical and potentially limiting to the scope and quality of data available for analysis. Therefore, the most comprehensive and current datasets available for each country were included in this study. Ultimately, 1081 chemical accident cases were collected from the US NTSIP between 2010 and 2014, 531 from the UK HCRD between 2016 and 2021, and 695 from Republic of Korea’s NICS between 2014 and 2022 [52].

The NTSIP, a web-based chemical accident surveillance system designed to track the health impacts of acute toxic chemical release, was developed by the ATSDR in 2010 to replace the Hazardous Substance Emergency Events Surveillance system. The 13,532 accidents that occurred in Louisiana, North Carolina, New York, Oregon, Tennessee, Utah, and Wisconsin between 2010 and 2014 were characterized by up to 97 attributes. The HCRD, managed by the UK HSE, records hydrocarbon-release incidents within UK continental shelf offshore oil and gas facilities. It was developed following recommendations from an investigation led by Lord Cullen after the Piper Alpha disaster on 6 July 1988 [53]. HCRD data are broadly composed of equipment logs covering various leak-related components, such as equipment type, release orifice size, substances, temperature, pressure, leak duration, and leak cause [48]. The NICS, a specialized institution under Republic of Korea’s Ministry of Environment, responds to chemical accidents and terrorism by providing expert personnel, equipment, predictive range assessments, scientific response technologies, and information. Since 8 January 2014, the Chemical Safety Agency has been providing data on chemical accidents in Republic of Korea, including accident materials, causes, types, and casualty information.

2.2. Classification by Cause of Accident

Classification by cause of accident comprises two stages. The first classification is based on general contributing factors. We established five primary classification criteria for accident cause analysis: Equipment, Operational, Procedural, Design, and Other. The general contributing factors corresponding to these five categories were extracted from a prior study [54] and subsequently organized according to each classification. The resulting classification is presented in

Table 4. Combination of five primary classifications and general contributing factors.

Classification	General Contributing Factor
Equipment	Mechanical integrity
	Safeguards, controls and protection levels
Operational	Safety culture
	Personnel training
	Preventive maintenance
	Human factors
	Emergency preparedness and response
Procedural	Operating procedures
	Change management
	Pre-startup safety review
Design	Hazard awareness and identification
	Preliminary hazard analysis
	Design
	Facility siting
Other	Work-permit system
	Contractor management
	Regulations and regulatory oversight
	Natural disasters

. Using these general contributing factors, the frequency of accident occurrences by factor was used to analyze the trends in chemical accidents for each country. The second classification was based on the detailed causes contributing to the general factors, and the fundamental accident causes for each country were analyzed from the 25 detailed causes identified across the three countries.

2.3. Development and Application of Internal and External Impact Score Formulae Based on Double Materiality

The safety component scoring formula developed in this study, which is based on double materiality, is divided into internal and external impact scores. The internal impact score assesses a company’s financial materiality by evaluating the severity and frequency of chemical accidents. The external impact score assesses a company’s impact materiality by evaluating the human casualties resulting from chemical accidents and the importance of management based on the detailed causes of these accidents. The scoring criteria for the internal and external impact scores are presented in

Table 5. Conception for internal and external impact score formulae.

ESG Evaluation Perspective	Evaluation Standard
Internal impact (financial materiality)	Chemical accident risk
	Chemical accident frequency
External impact (impact materiality)	Chemical accident casualties
	Management significance based on detailed chemical accident causes

.

2.4. Deriving Safety Management Component Prioritization

The developed safety component scoring formula was applied to chemical accident data from the three countries. Internal and external impact scores were calculated for each of the 25 evaluation components in the three countries, categorized by general contributing factors and detailed causes and plotted on a two-dimensional coordinate plane. The detailed steps for the internal and external impact score calculation, including the formulas and scoring criteria, are explained in Section 3.3 (Score Evaluation). Specifically, the internal and external scores were calculated by applying Equations (1)–(3), where each factor was systematically evaluated to ensure consistency in scoring. Based on the distribution of the internal and external impact scores, each evaluation component was divided into three tiers. A priority ranking of safety management components was derived by comparing the internal and external impact scores according to the detailed causes within each tier.

3. Results and Discussion

3.1. Initial Classification by General Contributing Factors

The results of the initial classification by general contributing factors are presented in

Table 6. US ATSDR NTSIP data during 2010–2014.

Components	General Contributing Factor	Number of Accidents	Percentage (%)
A	Mechanical integrity	831	76.87
B	Human factors	167	15.45
C	Preventive maintenance	30	2.78
D	Natural disasters	19	1.76
E	Emergency preparedness and response	15	1.39
F	Safeguards, controls and protection levels	11	1.02
G	Operating procedures	5	0.46
H	Pre-startup safety review	3	0.28
Total		1081	100.00

,

Table 7. UK HSE HCRD data during 2016–2021.

Components	General Contributing Factor	Number of Accidents	Percentage (%)
I	Mechanical integrity	286	53.66
J	Preventive maintenance	108	20.26
K	Human factors	57	10.69
L	Design	35	6.57
M	Safeguards, controls and protection levels	20	3.75
N	Operating procedures	15	2.81
O	Work-permit system	3	0.56
P	Personnel training	3	0.56
Q	Regulations and regulatory oversight	3	0.56
R	Emergency preparedness and response	2	0.38
X	Change management	1	0.19
Total		533	100.00

and

Table 8. Republic of Korea NICS data during 2014–2022.

Components	General Contributing Factor	Number of Accident	Percentage (%)
T	Human factors	349	50.22
U	Mechanical integrity	257	36.98
V	Safeguards, controls and protection levels	54	7.77
W	Preventive maintenance	25	3.60
X	Natural disasters	8	1.15
Y	Emergency preparedness and response	2	0.29
Total		695	100.00

. In all three countries, the primary causes of accidents were associated with general contributing factors related to equipment and operations.

In the USA, the contributing factors related to equipment and operations accounted for 77.89 and 19.62% of chemical accidents, respectively. Chemical accidents due to mechanical integrity and human factors accounted for 76.87 and 15.45%, respectively, totaling 92.32% of all chemical accidents.

In the UK, the combined contributing factors of equipment and operations accounted for 57.41 and 31.89% of chemical accidents, respectively. Chemical accidents associated with mechanical integrity, preventive maintenance, and human factors accounted for 53.66, 20.26, and 10.69%, respectively, representing 84.61% of all chemical accidents.

In Republic of Korea, the contributing factors associated with equipment and operations accounted for 44.75 and 54.11%, respectively. Chemical accidents due to mechanical integrity and human factors accounted for 36.98 and 50.22%, respectively, totaling 87.2% of all chemical accidents.

There were 8 general contributing factors in the USA, as shown in Table 6, 11 in the UK, as shown in Table 7, and 6 in Republic of Korea, as shown in Table 8, with the UK having the most diverse range of contributing factors, and Republic of Korea having the least diverse range. Whereas the UK had contributing factors across all four categories—equipment, operations, procedures, and design— Republic of Korea’s factors were limited to the equipment and operational categories. Overall, mechanical factors were identified as the primary cause of chemical accidents in the US and UK, whereas in Republic of Korea, chemical accidents were more frequently caused by human factors than by mechanical accidents.

This trend reflects the types of facilities where chemical accidents occurred in each country. In the US, 92.86% of chemical accidents occurred at fixed facilities, and in the UK, chemical accident data were derived solely from offshore platform incidents, which were assumed to occur at fixed facilities (100%). By contrast, in Republic of Korea, the proportion of chemical accidents caused by transportation-related incidents was 20.43%, which is approximately triple the US transportation accident rate (7.14%). Given that more than 75% of chemical accidents in the USA were related to mechanical factors, Republic of Korea’s chemical accident profile presents a striking contrast.

3.2. Secondary Classification by Detailed Cause

Detailed causes were determined based on the specified contributing factors or accident descriptions provided in the chemical accident data from the NTSIP, HCRD, and NICS. The results of the second classification by detailed causes are presented in

Table 9. US ATSDR NTSIP data during 2010–2014.

Components	Detailed Cause
A	System/process upset and shutdown due to equipment fatigue
B	Improper filling, loading, packing by employee
C	Performance maintenance failure due to regulation disobedience by operator
D	System/process upset due to bad weather
E	System shutdown and explosion due to emergency response procedural failure by operator
F	Performance maintenance failure due to unexpected reaction
G	Process control loss and fire due to procedural deficiency
H	System shutdown due to residual chemicals

,

Table 10. UK HSE HCRD data during 2016–2021.

Components	Detailed Cause
I	Degradation of flange, valve, hose sealing, corrosion and flange/valve erosion
J	Incorrectly fitted due to employee carelessness, improper maintenance due to employee preventive maintenance regulation disobedience
K	Operational excursion due to operator error
L	Valve, flange, hose wear/fatigue due to equipment design, other impact due to employee error
M	Leaking or dropping object due to employee error
N	Deficient and procedural non-compliance by employees
O	Work-permit non-compliance by employees
P	Maloperation due to lack of training
Q	Regulatory disobedience to simplify work procedures
R	Improper emergency response for other impact
X	Improper change management due to employee carelessness

and

Table 11. Republic of Korea NICS data during 2014–2022.

Components	Detailed Cause
T	Traffic accident due to employee carelessness, leaking or dropping due to employee error
U	Component/control/monitoring device failure from fatigue
V	Deficient safety equipment in unexpected reaction/phase transition
W	Improper filling, loading, packing by employees, traffic accident due to lack of preventive maintenance
X	Unexpected reaction due to bad weather
Y	Improper emergency response in unexpected reaction/phase transition

.

In the USA, the detailed mechanical integrity cause, which accounted for 53.66% of the contributing factors, was system or process upset or shutdown owing to equipment fatigue. For human factors, which represented 15.45%, the detailed causes were improper filling, loading, or packing by employees.

In the UK, the detailed causes of mechanical integrity, accounting for 50.22% of the contributing factors, included flange, valve, and hose-sealing degradation, and flange and valve corrosion and erosion. For preventive maintenance, which accounted for 20.26%, the detailed causes were incorrectly fitted components due to employee carelessness and improper maintenance due to noncompliance with preventive maintenance regulations. Human factors (10.69%) were attributed to operational excursions caused by operator error, whereas the design category (6.57%), involved valves, flanges, and hose wear or fatigue owing to equipment design and other impacts resulting from employee error.

In Republic of Korea, the detailed causes of human factors, which accounted for 50.22% of the contributing factors, were traffic accidents due to employee carelessness and leaks or spills due to employee error. The mechanical integrity factor, accounting for 36.98%, was attributed to component, control, or monitoring device failures due to fatigue. The safeguard, control, and protection level category, accounting for 7.77%, had a detailed cause of deficient safety equipment in unexpected reactions or phase transitions.

Following the accident types analyzed in the primary classification, the detailed causes in the USA and the UK originated primarily from fixed facilities. In both countries, the detailed causes of equipment-related general contributing factors were associated with equipment fatigue in fixed facilities, such as valves, flanges, and hoses, whereas the detailed causes of operation-related general contributing factors were linked to incomplete actions and errors by on-site employees. By contrast, detailed causes in Republic of Korea have been attributed to incidents in both fixed facilities and transportation accidents. Republic of Korea’s general equipment contributing factors included control and monitoring of device fatigue in both fixed facilities and transportation vehicles, whereas general operational contributing factors were linked to incidents caused by employee negligence and unsafe actions leading to spills and leaks.

3.3. Internal and External Impact Score Evaluation

3.3.1. Internal Impact Scoring

Risk Assessment

Risk assessment was conducted using chemical accident risk factors. The chemical accident risk factors were scaled based on the potential types of accidents that could occur during a chemical incident, such as near-misses, leaks, fires, and explosions, with a weighting of 0.5 points each, resulting in a maximum score of three points. The chemical accident risk factor values according to accident type are listed in

Table 12. Chemical accident risk factor criteria.

Chemical Accident Risk Factors	Criteria
Negligible (0.5 points)	Near-miss
Very low (1 point)	Leakage
Low (1.5 points)	Fire or explosion
Medium (2 points)	Leakage and fire
High (2.5 points)	Leakage and explosion or fire and explosion
Very high (3 points)	Leakage, fire and explosion

. The risk scores were calculated by multiplying the occurrence rates of chemical accident types for each country by the corresponding risk factor values. The formula for calculating the risk score is given in Equation (1):

$$Riskofcontributingfactors(R_i)=\sum _{i=1}^n{O}_i\times R_f$$

(1)

• $i=Contributingfactorsinchemicalaccidentdatafromeachcountry$
• $O_{r,i}=Occurrenceratioforeachchemicalaccidenttype$
• $R_f=Riskfactorofchemicalaccident$

Frequency Assessment

Frequency evaluation was performed using a frequency ratio factor based on chemical accident occurrence rates. The frequency ratio factor was determined by applying a score range based on accident occurrence rates scaled to a maximum of five points. The frequency ratio factor values are listed in

Table 13. Frequency ratio factor criteria.

Frequency Ratio Factors	Criteria
Very low (1 point)	Fifth ranking factor (0–1%)
Low (2 points)	Fourth ranking factor (1–5%)
Medium (3 points)	Third ranking factor (5–10%)
High (4 points)	Second ranking factor (10–20%)
Very high (5 points)	First ranking factor (20% and greater)

.

Internal Impact Score

The internal impact score was derived by multiplying the risk score with the frequency score. The frequency score was uniformly assessed from 1 to 5 points based on the occurrence rates of chemical accidents according to the general contributing factors. In contrast, the risk score was calculated based on the product of the risk associated with chemical accidents and the occurrence ratio of specific accident types, resulting in a significant variation in scores. Consequently, the internal impact score, derived from the product of the frequency and risk scores, reflected variations in the risk score. The results of the risk and frequency assessments, along with the internal impact scores for the 25 evaluation components across the three countries, are presented in

Table 14. Internal impact score of US general contributing factors.

Components	Risk Score	Frequency Score	Internal Impact Score
A	4.950	5	24.750
B	0.941	4	3.764
C	0.562	3	1.686
D	0.018	2	0.036
E	1.010	2	2.020
F	0.010	2	0.020
G	0.005	1	0.005
H	0.003	1	0.003

,

Table 15. Internal impact score of UK general contributing factors.

Components	Risk Score	Frequency Score	Internal Impact Score
I	1.399	5	6.995
J	4.693	4	18.772
K	1.250	4	5.000
L	1.066	3	3.198
M	0.039	2	0.078
N	0.029	2	0.058
O	0.006	1	0.006
P	0.006	1	0.006
Q	0.006	1	0.006
R	0.004	1	0.004
X	0.002	1	0.002

and

Table 16. Internal impact score of Republic of Korean general contributing factors.

Components	Risk Score	Frequency Score	Internal Impact Score
T	6.663	5	33.315
U	2.319	5	11.595
V	0.600	4	2.400
W	0.321	3	0.963
X	0.014	2	0.028
Y	0.008	1	0.008

.

In the USA, the contributing factor with the highest risk score was mechanical integrity, with a score of 4.95. Emergency preparedness and response ranked second, with a score of 1.01, followed by human factors, with a score of 0.941. The contributing factor with the highest frequency was mechanical integrity, whereas human factors and preventive maintenance ranked second and third, respectively. The internal impact score for mechanical integrity was 24.750, the highest among the factors, followed by human factors at 3.764, and emergency preparedness and response at 2.02. Preventive maintenance ranked fourth, with a score of 1.686. Other factors had internal impact scores < 0.1, indicating a minimal influence on the internal impact.

In the UK, the contributing factor with the highest risk score was preventive maintenance, with a score of 4.693. Mechanical integrity had a score of 1.399, and human factors and design had scores of 1.250 and 1.066, ranking third and fourth, respectively. The factor with the highest frequency was mechanical integrity, with human factors and preventive maintenance sharing second rank, and design in third. The internal impact score for preventive maintenance was 18.772, ranking first, whereas the mechanical integrity score was 6.995, ranking second. Human factors had an internal impact score of 5.0, ranking third, followed by design with a score of 3.198. Other factors had internal impact scores < 0.1, indicating a minimal influence on the internal impact.

In Republic of Korea, the contributing factor with the highest risk score was human factors (6.663), followed by mechanical integrity (2.319). The contributing factors with the highest frequency were human factors and mechanical integrity, whereas safeguards, controls, and protection levels ranked second, and preventive maintenance ranked third. The internal impact score for human factors was 33.315, which was the highest among all factors, whereas mechanical integrity scored 11.595, ranking second. Safeguards, controls, and protection levels scored 2.4, ranking third.

In the USA, the risk score for mechanical integrity was 4.95, which was the highest among the three countries. By contrast, the risk score for human factors was 0.941, which was the lowest among the three countries. Consequently, the internal impact score for mechanical integrity was the highest among the three countries, whereas that for human factors was the lowest. In the UK, the risk score for preventive maintenance was 4.693, which was the highest among the three countries, indicating a significant internal impact score. Conversely, the risk score for mechanical integrity was 1.399, which was the lowest among the three countries, and resulted in the lowest internal impact score for this factor.

In Republic of Korea, the internal impacts of mechanical integrity and human factors were significant. Notably, the risk score for human factors was 6.663, the highest among the three countries. Thus, the internal impact score for human factors was also the highest. Mechanical integrity had a risk score of 2.319 and a frequency score of 5, resulting in a high internal impact score of 11.595.

3.3.2. External Impact Scoring

Casualty Assessment

Casualty assessment was conducted using a chemical accident scale factor. The chemical accident scale factor was categorized into three grades based on the extent of casualties. The values of the chemical accident scale factor corresponding to the scale of chemical accidents are presented in

Table 17. Chemical accident scale factor criteria.

Chemical Accident Scale Factors	Criteria
Minor (1 point)	Three or fewer people injured
Significant (3 points)	1–2 people killed or 4–5 people injured
Major (5 points)	Three or more people killed, or six or more people injured

. The casualty score was calculated by multiplying the frequency of chemical accidents in each country by the value of the chemical accident scale factor. The formula for calculating the casualty score is shown in Equation (2):

$$Casualtycontributingfactors(C_i)=\sum _{i=1}^n{O}_{c,i}\times C_s$$

(2)

• $O_{c,i}=Occurrenceratioforeachchemicalaccidentscale$
• $C_S=Casualtyscalefactorofchemicalaccident$

Management Significance Assessment

Management significance was assessed based on material and personnel management. The evaluation of material management involves deriving the initiating events of chemical accidents based on the detailed causes of general contributing factors and quantitatively assessing the probability of failure on demand (PFD) values of applicable mitigation measures. Personnel management was qualitatively assessed based on the time and frequency required to implement improvements related to initiating events. The required time and frequency for implementing improvements were determined based on educational hours and the work-permit regulations outlined in the Republic of Korean Occupational Safety and Health Act, Mechanical Equipment Act, guidelines from the Republic of Korea Occupational Safety and Health Agency, and safety work-permit guidelines regarding work permits and pre-work safety meetings [55,56,57,58]. Material management was scaled to a maximum of five points and personnel management was scaled to a total of three points, with 1.5 points for both time and frequency. Finally, management significance was evaluated by summing the scores for material and personnel management, resulting in a total score of eight points. The evaluation criteria for material and personnel management are presented in

Table 18. Physical management criteria.

Material Management Importance Level	Criteria
Negligible (0.5 points)	${10}^{-5}/\mathrm{y}\mathrm{r}<\mathrm{P}\mathrm{F}\mathrm{D}\le {10}^{-6}/\mathrm{y}\mathrm{r}$
Very low (1 point)	${10}^{-4}/\mathrm{y}\mathrm{r}<\mathrm{P}\mathrm{F}\mathrm{D}\le {10}^{-5}/\mathrm{y}\mathrm{r}$
Low (2 points)	${10}^{-3}/\mathrm{y}\mathrm{r}<\mathrm{P}\mathrm{F}\mathrm{D}\le {10}^{-4}/\mathrm{y}\mathrm{r}$
Medium (3 points)	${10}^{-2}/\mathrm{y}\mathrm{r}<\mathrm{P}\mathrm{F}\mathrm{D}\le {10}^{-3}/\mathrm{y}\mathrm{r}$
High (4 points)	${10}^{-1}/\mathrm{y}\mathrm{r}<\mathrm{P}\mathrm{F}\mathrm{D}\le {10}^{-2}/\mathrm{y}\mathrm{r}$
Very high (5 points)	$\mathrm{P}\mathrm{F}\mathrm{D}\le {10}^{-1}/\mathrm{y}\mathrm{r}$

and

Table 19. Personnel management criteria.

Evaluation Components	Personnel Management Importance Level	Criteria
Frequency of improvements	Low (0.5 points)	Per half year
	Medium (1 point)	Per quarter
	High (1.5 points)	Per week or per task
Required time for improvements	Low (0.5 points)	<10 h
	Medium (1 point)	10–20 h
	High (1.5 points)	20 h or more

. The management significance score calculation formula is given in Equation (3).

$$Managementsignificance(M_s)=\sum _{i=1}^n{M}_i+P_i$$

(3)

• $M_i=Materialmanagementscoreofcontributingfactor$
• $P_i=Personnelmanagementscoreofcontributingfactor$

External Impact Score

The external impact score was determined by multiplying the casualty score by the management significance score. The results of the casualty and management significance assessments for the 25 evaluation components across the three countries, along with their external impact scores, are presented in

Table 20. External impact score of all contributing factors.

Components	Country	Casualty Score	Management Significance Score	External Impact Score
A	US	4.439	3.5	15.537
B	US	2.569	7.0	17.983
C	US	0.028	6.5	0.182
D	US	0.018	6.0	0.108
E	US	1.679	5.0	8.395
F	US	0.010	6.0	0.060
G	US	0.254	5.0	1.270
H	US	0.003	5.0	0.015
I	UK	4.231	4.5	19.040
J	UK	1.289	6.0	7.734
K	UK	1.892	7.0	13.244
L	UK	0.340	4.0	1.360
M	UK	0.127	4.0	0.508
N	UK	0.569	5.0	2.845
O	UK	0.037	5.0	0.185
P	UK	0.023	7.0	0.161
Q	UK	0.017	6.0	0.102
R	UK	0.027	5.0	0.135
X	UK	0.003	6.0	0.018
T	Republic of Korea	4.751	7.0	33.257
U	Republic of Korea	2.664	4.5	11.988
V	Republic of Korea	1.324	5.0	6.620
W	Republic of Korea	0.246	7.5	1.845
X	Republic of Korea	0.012	6.0	0.072
Y	Republic of Korea	0.003	5.0	0.015

.

In the USA, the contributing factor with the highest casualty score was mechanical integrity, which recorded 4.439 points, followed by human factors at 2.569 points, and emergency preparedness and response at 1.679 points, with the latter two ranking second and third, respectively. The contributing factor with the highest management significance was human factors, whereas preventive maintenance ranked second. Natural disasters, safeguards, controls, and protection levels ranked third. The external impact score was the highest for human factors at 17.983 points, followed by mechanical integrity at 15.537 points, emergency preparedness and response at 8.395 points, and operating procedures at 1.270 points.

In the UK, the contributing factor with the highest casualty score was mechanical integrity, with 4.231 points, followed by human factors at 1.892 points and preventive maintenance at 1.289 points, ranking second and third, respectively. The contributing factors with the highest management significance were human factors and personnel training, whereas preventive maintenance, regulations and regulatory oversight, and change management ranked second. The external impact score was the highest for mechanical integrity at 19.040 points, followed by human factors at 13.244 points, preventive maintenance at 7.734 points, operating procedures at 2.845 points, and design at 1.360 points.

In Republic of Korea, the contributing factor with the highest casualty score was human factors (4.751 points), followed by mechanical integrity (2.664 points) and safeguards, controls, and protection levels (1.324 points), ranking second and third, respectively. The contributing factor with the highest management significance was preventive maintenance, followed by human factors and natural disasters. The external impact score for human factors was the highest (33.257 points), followed by mechanical integrity (11.988 points), and safeguards, controls, and protection levels (6.620 points), with preventive maintenance (1.845 points) ranking fourth.

Generally, the casualty scores for the three countries were high for both mechanical integrity and human factors. Management significance scores were notably high for general contributing factors that required a significant level of personnel management, such as human factors and preventive maintenance, whereas they were lower for general contributing factors that demanded relatively less human management, such as mechanical integrity and design. In the USA, emergency preparedness and response recorded the highest external impact score of 8.395 points among the three countries, whereas, in the UK, five general contributing factors exceeded the external impact score of 1 point. In Republic of Korea, the external impact score for human factors was 33.257 points, which was the highest among the three countries, and the external impact score for safeguards, controls, and protection levels was the highest at 6.620 points.

3.4. ESG Assessment Based on Internal/External Impacts

The ESG assessment was conducted by deriving vulnerabilities based on the internal and external impacts of the general contributing factors. The results of the ESG assessment of the 25 evaluation components across the three countries are shown in Figure 1. The statistical analysis revealed a moderate positive correlation (r = 0.47) between internal and external impact scores across the components. This indicates that components with higher internal impact scores tend to exhibit somewhat higher external impact scores as well. However, the relationship is not strictly linear, suggesting variability in how internal factors influence external outcomes. This moderate correlation suggests that while internal impact factors may contribute to external impact outcomes, they do not fully determine them. Factors such as management strategies, operational environments, and industry-specific characteristics may influence external impacts independently. Further analysis, including variance decomposition and regression modeling, may be necessary to isolate the influence of each factor more clearly. Subsequently, a tiered classification was performed to systematically evaluate components vulnerable to internal and external impacts. Accordingly, evaluation components with an internal or external impact score exceeding 10 points were classified as Tier 1, those exceeding 1 point were classified as Tier 2, and components with scores less than 1 point were classified as Tier 3.

The evaluation components classified as Tier 1 are depicted in Figure 2, and the internal and external impact scores for each evaluation component are listed in

Table 21. Tier 1 internal and external impact scores.

Components	Internal Impact Score	External Impact Score
A	24.750	15.537
B	3.764	17.983
I	6.995	19.04
J	18.772	7.734
K	5.000	13.244
T	33.315	33.257
U	11.595	11.988

. Tier 1 consists of evaluation components with either an internal or external impact score of 10 points or higher, indicating a high level of vulnerability to both forms of impact.

The evaluation components in Tier 1 included preventive maintenance in the UK and mechanical integrity and human factors in all three countries. According to a detailed causal analysis, mechanical integrity in all three countries was commonly attributed to equipment fatigue such as flange, valve, and hose-sealing issues, resulting in significant vulnerabilities to both internal and external impacts. In the USA, the internal impact score for mechanical integrity was recorded at 24.75 points, making it the most vulnerable among the three countries. Conversely, mechanical integrity in the UK had an internal impact score of 6.995 points, indicating a relatively lower vulnerability in terms of internal impact, but an external impact score of 19.04 points, which was the highest among the three countries. The human factors identified across the three countries were predominantly linked to operator error. In Republic of Korea, incidents such as traffic accidents and leaks due to operator error and negligence resulted in the highest internal and external impact scores among Tier 1 evaluation components. This indicates that chemical incidents associated with human factors occurred at a higher frequency in Republic of Korea than in the USA and UK, suggesting inadequate casualty management. Preventive maintenance in the UK was driven by operator negligence and regulatory non-compliance, with an internal impact score of 18.772 points and an external impact score of 7.734 points, marking the highest vulnerability in preventive maintenance among the three countries.

The evaluation components classified as Tier 2 are illustrated in Figure 3, and their respective internal and external impact scores are detailed in

Table 22. Tier 2 internal and external impact scores.

Components	Internal Impact Score	External Impact Score
C	1.686	0.182
E	2.020	8.395
G	0.005	1.270
L	3.198	1.360
N	0.058	2.845
V	2.400	6.620
W	0.963	1.845

. Tier 2 includes evaluation components with an internal or external impact score of one point or higher, indicating a moderate vulnerability to both forms of impact.

According to a detailed causal analysis of the Tier 2 evaluation components, factors other than design in the UK and safeguards, controls, and protection levels in Republic of Korea were primarily caused by work procedures or human elements. Consequently, external impact scores, evaluated based on management significance, were generally higher. The internal impact scores for emergency preparedness and response in the USA and safeguards, controls, and protection levels in Republic of Korea were 8.395 points and 6.62 points, respectively, making them the most vulnerable among the Tier 2 evaluation components in terms of external impact. By contrast, design in the UK recorded an internal impact score of 3.198 points, demonstrating the highest vulnerability among the Tier 2 components in terms of internal impact. Preventive maintenance in the USA was vulnerable to internal impacts, whereas Republic of Korea’s preventive maintenance exhibited vulnerability to external impacts. The detailed causal factors for these evaluation components were linked to operator disobedience concerning regulations; improper filling, loading, and packing by employees; and traffic accidents resulting from a lack of preventive maintenance, highlighting the significant influence of human factors on management significance scores. However, Republic of Korea’s high casualty score indicated greater vulnerability to external impacts. This implies that, whereas casualty management from chemical incidents in the USA was effectively executed, Republic of Korea faced challenges in this area, likely because of the frequent occurrence of chemical incidents involving transportation vehicles and inadequate emergency response measures.

The evaluation components classified as Tier 3 are depicted in Figure 4, and their internal and external impact scores are detailed in

Table 23. Tier 3 internal and external impact scores.

Components	Internal Impact Score	External Impact Score
D	0.036	0.108
F	0.020	0.060
H	0.003	0.015
M	0.078	0.508
O	0.006	0.185
P	0.006	0.161
Q	0.006	0.102
R	0.004	0.135
S	0.002	0.018
X	0.028	0.072
Y	0.008	0.015

. Tier 3 consisted of evaluation components with both internal and external impact scores less than one point, indicating that these components were minimally affected by internal and external impacts.

4. Conclusions

This study developed a safety score evaluation formula to derive safety management components that companies can proactively implement to prevent chemical incidents based on double materiality, which considers both internal and external impacts. Subsequently, to assess the performance of the developed formula, chemical incident statistics from three countries (the US, the UK, and Republic of Korea) were selected as sample data for ESG evaluation.

• According to the first classification results based on the general contributing factors presented in [54], chemical incidents frequently occur with general contributing factors related to equipment and operation. In the USA, the incidence rates were 76.87% for mechanical integrity and 15.45% for human factors, whereas in the UK, the rates were 53.66% for mechanical integrity and 20.26% for preventive maintenance. In contrast, Republic of Korea showed a different distribution, with human factors accounting for 50.22% and mechanical integrity accounting for 36.98%. The UK had the most diverse set of incident contributing factors, whereas Republic of Korea had the least.
• The detailed causes of the chemical incidents for each country were determined based on the types of chemical incidents occurring in the US, the UK, and Republic of Korea. In the US and the UK, detailed causes originated from fixed facilities, whereas in Republic of Korea, they were derived from both fixed facilities and transportation vehicle accidents. Commonly, detailed causes related to equipment-type general contributing factors were equipment fatigue and degradation, such as valve, flange, and hose-sealing issues. The common detailed causes for operational-type general contributing factors were incomplete actions and employee error within the facility. In Republic of Korea, traffic accidents and incidents of leakage and dropping were caused by operator negligence and unsafe behavior.
• The evaluation indicators developed in this study were established to consider double materiality in relation to chemical incidents based on internal and external impacts. Internal impacts were evaluated based on the hazard and frequency of chemical incidents, whereas the external impacts were assessed based on the casualties resulting from chemical incidents and management significance.
• Based on the previously derived internal and external impact scores, the general contributing factors for each country were classified into three tiers. The ESG evaluation results revealed common vulnerabilities in the internal and external impacts of Tier 1 general contributing factors, namely human factors and mechanical integrity, as well as preventive maintenance (Tiers 1–2) across the three countries. Additionally, Tier 2’s general contributing factors allowed for the identification of trends in internal and external vulnerabilities related to chemical incidents in each country. This indicates that, to improve the double materiality assessment scores for chemical incidents, safety management should prioritize general contributing factors, including mechanical integrity, human factors, and preventive maintenance.

Through performance evaluation, vulnerability scores based on internal and external impacts were determined for the evaluation components. Subsequently, by visualizing these two scores in a two-dimensional coordinate system, the vulnerabilities among the evaluation components were compared and individual safety lead indicators for each component were presented. This demonstrates that the safety score evaluation formula developed based on double materiality can yield the expected outcomes when applied to actual sample data.

However, owing to sample size limitations, this study did not base its research on the statistical number of chemical incidents by industry; instead, it selected chemical incident data from the US, the UK, and Republic of Korea as sample data. Because the incident data from each country were not collected in the same manner and were recorded in various formats, issues relating to data quality and non-standardization could arise. To address potential issues related to data quality and non-standardization, this study applied the following measures. Data Consistency Check: Only datasets with clearly documented accident causes and consequence details were selected to ensure uniformity across the countries. Filtering and Cleaning: Data with ambiguous or missing cause classifications were excluded from the analysis. Standardization Approach: While data periods varied between countries, the analysis was designed not for direct numerical comparison between datasets but for trend identification and key factor analysis within each dataset. Validation Through Cross-Referencing: Accident causes and classifications were cross-checked with multiple reliable data sources where available, ensuring consistency in classification. Nevertheless, a second verification, using the same method as the performance evaluation conducted in this study, may be required. Additionally, future research should focus on conducting ESG safety evaluations by applying these evaluation components to actual companies.

This study aims to inspire future efforts where accident prevention and safety investments are strategically implemented from an ESG (Environmental, Social, and Governance) perspective, emphasizing the importance of considering both internal impact and external impact for the establishment of sustainable corporate strategies. Internal impact reflects the risks and vulnerabilities within the organization, such as equipment reliability, process safety, and employee-related risks, while external impact accounts for the broader consequences affecting stakeholders, including environmental harm, public health, and community safety. By integrating and analyzing various datasets, this research compared the relative priority of safety management factors for accident prevention across multiple components. It is anticipated that this comparative approach can evolve into more extensive empirical studies, where these prioritization frameworks can be tested and validated within real industrial environments and corporate safety management systems. Such practical applications would offer valuable insights which could be useful for refining safety strategies tailored to the operational realities of different industries. Ultimately, promoting safety investments from an ESG perspective should extend beyond mere regulatory compliance, encouraging a more proactive approach where corporate safety performance is continuously enhanced. By balancing both internal risks and external impacts, organizations can achieve optimal decision-making that not only strengthens workplace safety but also fosters a broader culture of sustainability. This integrated approach can serve as a progressive management strategy, positioning safety as a core element of long-term corporate success and stakeholder trust.