Assessing the Impact of ESG Performance on Firm Competitiveness: A Meta-Frontier DEA and OLS Regression Analysis of Global Optical Component Firms

Abstract

The optical component industry is an industry that generates a high profit, yet it has rarely been explored on its management in the previous research. This study aims to investigate the impact of environmental (E), social (S), and governance (G) performance (ESGP) on firm competitiveness. Specifically, by using ordinal least square (OLS) regression, we analyze how overall ESGP and its components (E, S, G) impact the competitiveness including component of marketability, profitability, and innovation ability of global optical component firms, including those from Taiwan, Japan, China, and South Korea. We developed a meta-frontier data envelopment analysis (DEA) model including meta frontier, group frontier, and operational technology gap ratio to estimate firm competitiveness and analyzed 95 firm-year observations from 2016 to 2020. The meta-frontier DEA results indicate that optical component firms from Taiwan and Japan exhibit relatively higher competitiveness. Additionally, the ordinary least squares and truncated regression results reveal a significant positive association between ESGP and firm competitiveness, aligning with the resource-based view theory, where firms that performs in ESG practices obtaining unique resources and capabilities would develop a long-term competitive advantages, and also aligning with stakeholder theory, where firms that engage in ESG attract support from stakeholders, leading to a higher firm performance and competitiveness. The findings of the provides insights for the optical policymakers of the proper resource allocation to maximize firm competitiveness under the influence of ESGP.

Plain Language Summary

How ESG Performance Affects Business Success: Insights from The Global Optical Component Industry

The optical component industry plays a crucial role in modern technology and generates high profits. However, little research has explored how companies in this industry manage their business performance. This study examines how Environmental, Social, and Governance (ESG) performance affects the competitiveness of optical component firms worldwide, focusing on companies from Taiwan, Japan, China, and South Korea. To measure firm competitiveness, we use a method called meta-frontier data envelopment analysis (DEA) and support our findings with ordinary least squares (OLS) regression analysis. Competitiveness is assessed in three key areas: marketability, profitability, and innovation ability. Our study analyzes 95 firm-year data points from 2016 to 2020 to understand the link between ESG performance and business success. The results show that firms in Taiwan and Japan generally have higher competitiveness compared to their counterparts in China and South Korea. Additionally, we find that companies with strong ESG performance tend to be more competitive. This aligns with two well-known theories: 1. Resource-based view theory – Companies that invest in ESG build unique resources and long-term advantages. 2. Stakeholder theory – Companies that perform well in ESG attract greater support from investors, customers, and society, leading to better business performance. Our findings provide useful insights for business leaders and policymakers in the optical component industry, helping them allocate resources effectively to enhance firm competitiveness through ESG strategies.

Keywords

ESG firm competitiveness data envelopment analysis meta-frontier optical

Introduction

The optical component industry has always been in a competitive and profitable situation, where the profit-generating ability of an organization always draws the attention of investors. Although firms aim to expand the organization, competitiveness is the main focus of firms to keep up with competitors (D. Yang et al., 2022). In this study, the firm competitiveness of the optical components industry viewed in three dimensions, namely, innovation ability, profitability, and marketability efficiency. Companies must invest in R&D to secure competitive advantages through differentiation and intellectual property. Profitability reflects financial health and attracts investors, while marketability shows the market's response to the company's performance.

In this modern era, the importance of chips for intelligent machines is like the heart of humans; it is essential for aerospace, military industry, automatic controls, medicine, transportation, telecommunication, safety controls. Optical components, such as the lens and camera lens, are the eyes of all types of equipment to comply with chips in the aforesaid areas, allowing equipment to implement the intelligent operation. Vantage Market Research (2022) forecasted that Asia Pacific is the fastest-growing region of optical component industry, where Taiwan, Japan, South Korea, and China are the regions that are mainly covered. The optical component market was valued at USD 42.32 billion in 2021 and is projected to attain a value of USD 60.55 billion by 2028 at a compound annual growth rate of 4.9% in 2022 to 2028. Hence, the companies chosen in this study cover the top-performing companies in the optical component industry. This could help company CEOs or managers to understand their competitiveness and shortcomings compared with their rivals in the field.

In addition, the relationship between environmental (E), social (S), and governance (G) performance (ESGP) and companies’ performance have been looked into in many prior studies that focus on different regions and industries, which will be covered in Section “Literature Review.” The European Commission (2021) proposed the Carbon Border Adjustment Mechanism, where it set a series of policies to trade with EU countries to fight climate change, targeting a 55% reduction in carbon emissions by 2030 compared with 1990 levels and to become a climate-neutral continent by 2050. The carbon emissions generated in the process of manufacturing optical components would be covered if the company excels in E performance, where it monitors the impact of the company's operational process on the environment. The European Commission has initiated several reform activities to enhance the quality of corporate governance, where ESGP is one of the main focuses, causing the increasing importance of ESGP (Velte, 2017). As the inevitable trend of incorporating ESG into corporate governance, the organization's resource allocation could be affected, with the importance of understanding the impact of ESG on the company's competitiveness and performance.

However, prior academic studies do not have enough coverage of the relationship between ESGP and the competitiveness or efficiency performance of the optical component industry, where most of the literatures focused more on LED/LCD industry (Chiu et al., 2008; Z.-Y. Lee & Pai, 2011; Sun, 2014; C.-N. Wang et al., 2018; K. Wang et al., 2021).None of the aforementioned previous studies had examined firm competitiveness, which includes various aspects by combining innovation capability, profitability, and marketability efficiency, as in this study. In addition, it is rare to examine the possible impact of ESGP on the firm competitiveness of the optical components industry.

This study focuses on exploring the relationship between the ESGP and the competitiveness of nineteen top optical component manufacturing organizations from four Asian countries, namely, Taiwan, Japan, South Korea, and China, from 2016 to 2020. The study first identifies the firm competitiveness using meta-frontier data envelopment analysis (DEA) and then examines the impact of ESGP on firm competitiveness by using ordinal least square (OLS) regression model. This study fills the gap where scant research focused on solely optical component industry, incorporate R&D efficiency, profitability efficiency, and marketability efficiency as the competitiveness. The optical component industry is an innovation-intensive industry, which is suitable for exploring the competitiveness that includes the R&D efficiency of the firms. The companies from these four countries were chosen because they are the best performance companies in Asian region.

The Data Envelopment Analysis (DEA) model is a linear programming-based method used to measure the efficiency performance of decision-making units (DMUs) with multiple inputs and outputs (Charnes et al., 1978, 1997). Various kinds of DEA methods have been developed throughout the years for various functions and purposes (Cheng et al., 2020). In the present study, the meta-frontier DEA method was used to measure the meta-frontier and group frontier efficiencies and inefficiencies by including the technology gap ratio among the organizations. This method could serve as the first contribution of this study. The meta-frontier DEA method is new to the optical component industry but is used in various studies of other industries (Chiu et al., 2012, 2016; Q. Wang et al., 2013; Yu & Chen, 2020; Zhang et al., 2013). On the other hand, this study combines the innovation ability, profitability, and marketability efficiencies to identify the firm competitiveness of the optical component companies is the innovative contribution to the literature where rare prior studies had been done. Prior studies only studied one of the components but not all three. This serves as the second contribution of study.

Prior studies on the relationship between the ESGP and the competitiveness of companies from the optical component industry are few, serving as the third contribution of study. The importance of how the results show the ESGP could positively affect firm competitiveness, aligning with the stakeholders theory and resource-based view (RBV) theory, emphasized to raise awareness among stakeholders and investors, which is the fourth contribution of this study. This study will continue with literature review in Section “Literature Review,” research design in Section “Research and Data Methodology,” empirical results and analysis in Section “Empirical Results and Analysis”, and concluding remarks in Section “Empirical Results and Analysis”.

Literature Review

Overview of Optical Industry

The optical industry is the opto-electronic industry that covers five main sectors: optical technology, communications, components, device and equipment, and application (Chang et al., 2002). The optical technology sector consists of three sub-sectors, namely, optical storage, optical output, and optical input. The optical communications sector consists of products, such as fiber and optical cables, optical connectors, optical active components, and more. The optical component sector consists of laser diode, LED, and LCD. The optical device and equipment sector consists of optical elements and optical equipment. The optical application sector consists of optical-electronic measurement, laser medical instruments, laser machinery, and solar application.

Prior Studies in Optical Industry

Due to its various advantages, DEA has been widely applied in prior studies. Table 1 presents the prior studies used DEA models for measuring efficiency. As shown in Table 1, most of the studies have been used the traditional DEA models, such as CCR, BCC, Malmquist, Window Analysist (Broekel et al., 2018; Jens & Petra, 2009; Vladislav et al., 2016). Although traditional DEA is simple and widely used, it assumes one common frontier for all units, limiting its ability to capture technology differences. This assumption implies that all firms operate under the same production technology, regardless of regional, sectoral, or technological differences. In practice, however, firms often face heterogeneous environments—such as varying resource endowments, regulatory conditions, or technological capabilities—that shape their production possibilities.

Table 1.

DEA Models Used in Prior Studies.

Author(s)	DEA model	Units	Inputs	Outputs
K. Wang et al. (2021)	-CCR and BCC	34 LED energy enterprises in China	-Operating costs -R&D expense -General and administrative expense -Net value of fixed assets -Return on total assets ratio	- Gross revenue
C.-N. Wang et al. (2018)	-Malmquist Productivity Index	22 LED companies in Taiwan (2011–2015)	-Current assets -Fixed assets -Operating expenses	-Operating revenue -Net profit
Sun (2014)	-Window Analysis -Malmquist Productivity Index	10 LED companies in Taiwan	-Total assets -Number of employees -Operating expense -Operating cost	- Net operating revenue - Gross profit
Z.-Y. Lee and Pai (2011)	-CCR	10 TFT-LCD manufacturers: -Korea: Samsung and LG Display -Taiwan: AUO, CPT, HannStar, CMO and Innolux -Japan: SONY, Toshiba. and Sharp	-R&D expenditures -Fixed asset -Intangible asset -Capital stock -Currency	-Income -Pre-tax net profit -Return rate on stockholder’s equity -Return rate on assets
Chiu et al. (2008)	-BCC -Super SBM, -Malmquist Productivity Index	21 LED manufacturers in Taiwan (2003–2006)	-Fixed assets -Number of employees -R&D expenditure	-Net sales

Accordingly, several DEA models have been developed throughout the years to serve different functions, and they obtain different findings (Cheng et al., 2020). Among them, the two-stage meta-frontier DEA model incorporates the production process and intermediation approaches to identify the inputs, intermediates, and outputs using a two-stage network system and identify the group frontier and technology gap ratio using the meta-frontier (Chiu et al., 2016; Fei & Lin, 2016; Q. Wang et al., 2013). The two-stage meta-frontier DEA model overcomes the disadvantages of traditional models by integrating both production and intermediation processes by modeling inputs, intermediates, and outputs within a two-stage network system. It further identifies group frontiers and quantifies the technology gap ratio (TGR), offering a more comprehensive assessment of cross-group efficiency (Chiu et al., 2016; Fei & Lin, 2016; Q. Wang et al., 2013).

ESG and Competitiveness

The association between ESG performance (ESGP) and firm outcomes has received growing attention among managers and policymakers (Nirino et al., 2021). According to the resource-based view and Stakeholder theory, the strategic allocation of resources and the effective implementation of ESG activities can serve as valuable assets that provide firms with sustainable competitive advantage (Elia et al., 2021). Zheng and Lee (2025) stated that ESG components have been contributed to shape firms' outcomes across varying market contexts and time periods. ESG practices demonstrate that businesses affect not only profits but also the environment, society, and governance (Gillan et al., 2021). Specifically, (E) initiatives such as sustainable sourcing and emission reduction lower costs and appeal to eco-conscious stakeholders (C. L. Lee & Liang, 2024), while (S) engagement enhances reputation, loyalty, and long-term resilience (Abdi et al., 2022). Meanwhile, (G) is strongly linked to efficiency by promoting transparency, reducing risks, and attracting investors (Xie et al., 2019).

Prior studies suggest that ESGP may enhance corporate value creation and strengthen sustainable competitive advantage (Behl et al., 2022; Miralles-Quirós et al., 2019; Zumente & Bistrova, 2021). Firms that commit to sustainability are often found to attract a broader investor base, lower their cost of capital, and improve operational efficiency (Bridoux & Stoelhorst, 2022; Freeman et al., 2021). However, the empirical evidence on the ESGP–firm competitiveness relationship remains mixed, with some studies reporting strong positive effects, while others find weak or even indirect links (as shown in Table 2). Consequently, there is a pressing need to re-examine this relationship, particularly in underexplored industries such as the optical sector, where ESG practices and performance outcomes may follow different patterns compared to more widely studied industries like LCD/LED (Chiu et al., 2008; Z.-Y. Lee & Pai, 2011; Sun, 2014; Q. Wang et al., 2018; K. Wang et al., 2021). Based on the discussion above, the following hypothesis is developed:

H0: ESGP would significantly impact the firm competitiveness.

Table 2.

Literature Review ESG and Competitiveness.

Author(s)	Focus	Unit(s)	Findings
Zhou et al. (2022)	Marketability	167 listed firms in Shanghai and Shenzhen A-share markets (2014–2019)	-Improvement of ESG performance can improve market value performance.
Di Simone et al. (2022)	Innovation	909 innovative listed firms globally (2013–2017)	- Market-perceived innovation (R&D intensity and market-to-book ratio) positively impacts economic sustainability; the social component has the highest effect on economic sustainability when ESG components are considered. - Indirect relationship between ESG and innovation.
Tang (2022)	Innovation	1,070 firm-year sample of listed firms of A-share markets achieving the ESG rating of the SynTao Green Finance Company (2015–2020)	- ESG performance significantly increases the quantity and quality of corporate innovation, and the result is mediated by alleviating financial constraints and agency costs.
Kluza et al. (2021)	Innovation	72 articles	-Innovations affect sustainable business models in an unambiguously positive way for every country. - No standardization and categorization to include environmental, social, and governance factors in the study are the limitations of the study.
Nekhili et al. (2021)	Marketability	91 French firms listed on the SBF 120 index (excluding financial, insurance, and real estate companies) (2007–2017)	- Investors react positively to ESG performance, but employees on the board react negatively to ESG. -Hence, ESG performance and market value have a negative relationship in firms with employee directors on the board.
Broadstock et al. (2020)	Innovation	320 Japanese firms from Bloomberg (2008–2016)	-Indirect value-creation: CSR/ESG policy adoption enhances the firm’s ability to pursue innovation activities and positively affects their value creation and operational performance.
Brogi and Lagasio (2019)	Profitability	3,476 US firms (2000–2016)	- ESG is strongly related to profitability.
Dalal and Thaker (2019)	Profitability	65 Indian firms listed on the NSE 100 ESG Index database (2015–2017)	-Corporate ESG performance enhances financial performance.
Zhao et al. (2018)	Profitability	20 China’s electric power firms under four groups (2016)	- High ESG performance could positively affect financial performance.
Velte (2017)	Profitability	412 firm-year sample firms listed on the German Prime Standard (DAX30, TecDAX, MDAX) (2010–2014)	-Governance performance has the strongest impact on firms’ financial performance as compared to environmental and social performance.

Research and Data Methodology

Research Framework and Competitiveness Evaluation Model

The conceptual framework that contains the variables studied in this paper is shown in Figure 1. Financial statements and ESG data of the companies were obtained from the Thomson Reuters datastream database. ESGP was an independent variable and consists of three components: E, S, and G. The return rate on assets (ROA), leverage (LEV), age (AGE), and size (SIZE) are the control variables.

Figure 1.

Conceptual framework.

This study obtained the data from multiple sources. While the granted patents and cited patents are collected from the Lens, variables used for DEA measurement and regression analysis are from Thomson Reuters Eikon database. During the data collection, missing data were eliminated. The sample data for this study includes 19 firms during a period of 2016–2020, with 95 firm-year observations. The variables used for DEA model are based on previous studies which are summarized in Table 3. For the regression analysis, ESGP is served as independent variable which includes three components: Environment (E), Social (S), and Governance (G), with the measurement scores range from 0 to 100. To be specific, firms with higher ESG score are considered to perform well on the ESG activities. Control variables included in the regression consists of return on assets (ROA), leverage (LEV), firm age (AGE), and the firm size (SIZE).

Table 3.

Definitions of DEA-related Variables.

Variables	Units	Definitions	Data Source	Reference
Inputs
R&D Expenses	(Thousand TWD)	Annual expenses of a firm on R&D.	Thomson Reuters	Kweh et al. (2022)
Operating expenses	(Thousand TWD)	Cost of labor activities, administrative and business processes, which include employees’ salaries, utilities and stationery, and sales expenses.	Thomson Reuters	Sherman and Ladino (1995), Kotabe and Zhao (2002)
Plant, property, and equipment	(Thousand TWD)	Fixed assets used during the process of operating and manufacturing, which include materials, real estate, and equipment.	Thomson Reuters	Kweh et al. (2022)
Outputs
Innovation ability
Granted patent	Patent	Numbers of patents filed and published by a company on Lens Patent Database.	Lens Patent Database	Sharma and Thomas (2008), Pandit et al. (2011)
Cited patent	Patent	Number of times patents filed and published by a company are cited on Lens Patent Database.	Lens Patent Database	Sharma and Thomas (2008), Pandit et al. (2011)
Profitability
Net income	(Thousand TWD)	Company revenue from the company business activities, which include both products and services.	Thomson Reuters	Co and Chew (1997), Hashimoto and Haneda (2008)
Marketability
Market value	(Thousand TWD)	The closing price of share is multiplied by the total number of shares issued by the company of the year.	Thomson Reuters	J. Lee et al. (2019), Kweh et al. (2022)

Figure 1 illustrates a two-step framework for evaluating the impact of ESGP on firm competitiveness, incorporating meta-frontier DEA and regression analysis. The first step focuses on measuring firm competitiveness through an efficiency assessment that compares inputs and outputs across firms using the meta-frontier DEA method. The details of meta-frontier DEA models and variables are discussed below. The second step, by using OLS regression model, examines how ESGP affect firm competitiveness while controlling for various firm-specific variables. E-score measures the firm's environmental practices, such as resource use score, emission score, and innovation score. On the other hand, S score measures the aspects of workforce, human rights score, community score, and product responsibility score. G score shows the quality of corporate governance, including management score, shareholders score, and CSR strategy score. The additional truncation and bootstrap step ensure that the data is pre-processed to remove outliers or inconsistencies, thereby improving the robustness and reliability of the DEA and regression results.

Table 3 shows the meta-frontier DEA input variables (R&D expenses, operating expenses, and fixed assets) and output variables including innovation capability (granted and cited patents), profitability (net income), and marketability (market value) to evaluate the competitiveness of the optical components industry.

R&D expenditure represents a company's investment in innovation and technological advancement, which is the company’s innovation input. Operating expenses reflect the resources spent on the business operation process, where it includes the non-R&D operating costs. Property, plant and equipment records the firm's capacity to produce goods and services. Including this variable ensures that the assessment takes into account the role of capital infrastructure in supporting R&D and overall firm productivity. Granted and cited patents demonstrate the effectiveness and quality of the firm's R&D investments. Net income shows the profit generation situation of the company. Market value represents the investors' perception of the firm's future growth and profitability.

Meta-frontier Data Envelopment Analysis

Meta-frontier includes all data from different regions to investigate firm competitiveness as a whole. Group frontier divides the data according to the region of the companies to identify the technological difference that occurs among regions or groups. The operational technology gap ratio (OTGR) shows the technological gap between the meta-frontier and group frontier. Several prior studies also utilized the meta-frontier DEA in identifying the efficiency and inefficiencies of different fields (Chiu et al., 2012; Feng et al., 2021; Ma et al., 2021; Xiao et al., 2022). The meta-frontier DEA approach offers several advantages. First, it accounts for technological heterogeneity by allowing firms to be evaluated relative to both their own group frontier and the global meta-frontier, which reflects the best available technology across all groups (Q. Wang et al., 2013). This enables more meaningful cross-group comparisons than conventional DEA, which assumes all decision-making units (DMUs) operate under a common technology as the benchmark could not view the industry in whole. Group frontier considers the optical firm as a structure itself, while meta-frontier taking all the firms into account at once (Goyal et al., 2019). Second, OTGR, which measures the distance between a group frontier and the meta-frontier (O’Donnell et al., 2008), thereby identifying the extent to which firms lag behind the global best practice. Third, it improves the fairness and accuracy of efficiency measurement, as firms are not unfairly penalized for operating under less advanced technologies. Finally, the meta-frontier framework offers useful insights for policy and managerial decision-making to enhance overall competitiveness. The details and formulae of the meta-frontier, group frontier, and OTGR are elaborated in Sections “Meta-frontier and Group Frontiers of the Directional Distance Function” and “Technology Gap Ratio and Derivation of Meta-Frontier Inefficiency.”

Meta-Frontier and Group Frontiers of the Directional Distance Function

O’Donnell et al. (2008) established meta-frontier and group frontiers, where the meta-frontier includes the DMUs from all groups and the group frontiers are the DMUs for individual groups, to evaluate the influence of technological heterogeneity. Assume that a batch of N companies and these companies can be allocated into p dissimilar groups ( $G_{g}, g = 1, \dots, p$ ). The sample of the $G_{g}$ group is $N_{g}$ , where $N_{1} + N_{2} + \dots + N_{p} = N$ . All company has m inputs and s outputs. The inputs $x \in R_{+}^{m}$ were used to produce the outputs $y \in R_{+}^{s}$ . Suppose that the possibility production set of the inputs and outputs are convex. Hence, the network directional meta-distance function and the two group-specific network directional distance functions (DDF) were established as the following formula:

\begin{matrix} {D^{⇀}}^{M} (x, y; g_{x}, g_{y}) = Max {α + β : (x - α g_{x}, y + β g_{y}) \in T^{M} (x, y)} . \\ {D^{⇀}}^{G} (x, y; g_{x}, g_{y}) = Max {γ + τ : (x - γ g_{x}, y + τ g_{y}) \in T^{G} (x, y)}, G = G_{1}, \dots, G_{P} . \end{matrix}

The meta-technology and group-specific technology sets were established as the formula:

$T^{M} (x, y)$ : $x \in R_{+}^{m}$ could generate the outputs $y \in R_{+}^{s}$ throughout the production process.

$T^{G} (x, y)$ : $x \in R_{+}^{m}$ could be utilized by the firms in group $G_{g}$ of the first step to construct $y \in R_{+}^{s}$ .

Furthermore, the meta-technology set includes the G group-specific technology set $T^{M} (x, y) = {T^{G_{P}} (x, y) \cup \dots \cup T^{G_{P}} (x, y)}$ . According to Fried et al. (2008), the researcher should select the direction vector $g = (g_{x}, g_{y})$ before assessing the DDF. A well-defined efficiency measure should satisfy unit invariance, which means that it is independent of the measurement units of input and output variables (Knox Lovell & Pastor, 1995). In the present study, the direction was considered $g = (g_{x} = x, g_{y} = y)$ (Chiu et al., 2012; Chung et al., 1997). However, if $g = (g_{x} = x, g_{y} = y)$ is determined exogenously, it could achieve unit invariance in the efficiency evaluation if the units used for the directional vector are connected to input and output variables, that is, if the observed values of inputs and outputs are used as the directional vector (Charles et al., 2016; Färe & Charles, 2018; F. Yang et al., 2018). Hence, the inefficiency measure of the company of the meta-technology and group-specific technology sets, under convex constraints, can be denoted by the following linear programs.

\begin{matrix} {D^{⇀}}^{M} = Max α_{o}^{M} + β_{o}^{M} \\ \sum_{g = 1}^{G_{P}} \sum_{j = 1}^{N_{g}} λ_{j}^{g} x_{ij}^{g} \leq x_{io}^{g} - α_{o}^{M} g_{iox}, i = 1, \dots, m, \\ \sum_{g = 1}^{G_{P}} \sum_{j = 1}^{N_{g}} λ_{j}^{g} y_{rj}^{g} \geq y_{ro}^{g} + β_{o}^{M} g_{roy}, r = 1, \dots, s, \\ \sum_{g = 1}^{G_{P}} \sum_{j = 1}^{N_{g}} λ_{j}^{g} = 1, \\ λ_{j}^{g} \geq 0, g = 1, \dots, G_{P} . \end{matrix}

(1)

\begin{matrix} {D^{⇀}}^{G} = Max γ_{o}^{g} + τ_{o}^{g}, g = 1, \dots, P \\ \sum_{j = 1}^{N_{g}} λ_{j}^{g} x_{ij}^{g} \leq x_{io}^{g} - γ_{o}^{g} g_{iox}, i = 1, \dots, m, \\ \sum_{j = 1}^{N_{g}} λ_{j}^{g} y_{rj}^{g} \geq y_{ro}^{g} + τ_{o}^{g} g_{roy}, r = 1, \dots, s, \\ \sum_{j = 1}^{N_{g}} λ_{j}^{g} = 1, \\ λ_{j}^{g} \geq 0 . \end{matrix}

(2)

where $λ_{j}^{g}$ represents the intensity variables corresponding to the processes, and $N_{1} + N_{2} + \dots + N_{p} = N$ .

Consequently, the operational efficiency in the meta-technology and group-specific technology sets are defined as ${OE}_{o}^{M} = \frac{1}{2} [(1 - α_{o}^{M}) + 1 / (1 + β_{o}^{M})]$ and have a value between 0 and 1. If ${OE}_{o}^{M}$ and ${OE}_{o}^{G}$ are equal to unity, these variables stipulate that the target firm is efficient.

Technology gap ratio and derivation of meta-frontier inefficiency

$T^{M} (x, y) = {T^{G_{P}} (x, y) \cup \dots \cup T^{G_{P}} (x, y)}$ suggests that the meta-frontier envelopes the G group-specific frontiers; therefore, the efficiency of the input computed by the meta-frontier ( ${OE}_{o}^{M}$ ) is lower than that computed by the group-specific frontiers ( ${OE}_{o}^{G}$ ). In comparison, the efficiency of the output computed by the meta-frontier ( ${OE}_{o}^{M}$ ) is higher than that computed by the group-specific frontiers ( ${OE}_{o}^{G}$ ). Hence:

{OE}_{o}^{M} (x, y) \leq {OE}_{o}^{G} (x, y) .

(3)

The ratios between the meta-frontier efficiency of the input and the group-specific frontier efficiency of the input and between the meta-frontier efficiency of the output and the group-specific frontiers efficiency of the output were connected as the OTGR. The ratios are formulated as:

OTGR (x, y) = {OE}^{M} (x, y) / {OE}^{G} (x, y) \leq 1 .

(4)

The closer the OTGR is to one, the narrower the gap of the ratio is. In other words, the efficiencies of the input and output of the group frontiers are more adjacent compared with those of the meta-frontier.

According to the OTGR, the target firm in group-specific frontiers only differentiates between the ratios of the two frontiers and could not investigate the derivation of meta-frontier inefficiency (Chiu et al., 2012, 2013). Consequently, this paper further breaks down the inefficiencies of the target firm under the meta-frontier into operational technology gap inefficiency (OTGI) and operational technical inefficiency (OTI).

Furthermore, OTGI represents the inefficiency of the target firm in group-specific frontiers. According to the technical gap between the meta-frontier and group-specific frontiers, OTGI is expressed as:

OTGI (x, y) = {OE}_{o}^{G} (x, y) \times (1 - OTGR (x, y))

(5)

Furthermore, OTI exhibits the managerial inefficiency of the group-specific frontiers of the target firm. According to the input excess in the operational or output shortfall in the production process, OTI is expressed as:

OTI (x, y) = 1 - O E^{G} (x, y) .

(6)

Thus, the operational inefficiencies examined using the meta-frontier can be expressed as:

MOTI (x, y) = OTG I^{G} (x, y) + OT I^{G} (x, y) .

(7)

Empirical Results and Analysis

Descriptive Statistics

Table 4 shows the descriptive statistics of the variables used for DEA model to examine the efficiency of optical firms. Overall, the total sample showed that most of the variables had negative percentage changes except for operating expenses, PPE, and market value. Granted patent had the highest negative percentage change of 0.13% whereas market value had the highest positive percentage change of 0.39%. Companies from the mainland had the best overall results, as all of their variables had positive percentage changes. In comparison, the companies from Korea had the worst overall results, as almost all their variables had negative percentage changes except for market value.

Table 4.

Descriptive Statistic of DEA Variables in Country or Region.

Stages	Variables	Mean					% change	p-value
Stages	Variables	2016	2017	2018	2019	2020	% change	p-value
Input	R&D expenses	9,413,148	9,687,339	9,554,181	10,042,575	9,239,924	−0.02	<.01
	Operating expenses	122,343,285	127,786,480	124,569,527	134,150,559	122,288,955	0.00	<.01
	PPE	33,212,987	32,805,762	32,445,106	36,306,147	35,266,252	0.06	<.01
Outputs	Granted patent	413	380	375	400	359	−0.13	<.01
	Cited patent	347	310	314	338	311	−0.10	<.01
	Net income	10,644,275	11,581,640	9,602,343	12,227,819	10,192,745	−0.04	<.01
	Market value	177,936,053	186,390,525	222,449,532	205,863,050	246,567,680	0.39	<.01
Taiwan
Input	R&D expenses	629,953	707,210	716,447	843,417	857,163	0.36
	Operating expenses	5,546,712	5,443,792	5,477,312	6,015,529	6,178,000	0.11
	PPE	4,776,155	4,947,787	5,196,951	6,048,372	7,002,171	0.47
Outputs	Granted patent	45	27	50	45	42	−0.08
	Cited patent	26	25	49	44	42	0.61
	Net income	10,413,412	10,979,177	10,758,227	12,664,797	12,236,250	0.18
	Market value	67,918,300	76,766,599	61,528,399	99,664,245	67,640,859	0.00
Mainland
Input	R&D expenses	3,506,950	4,279,779	6,854,600	6,612,741	6,620,247	0.89
	Operating expenses	73,889,334	90,760,905	111,894,738	146,060,008	138,657,053	0.88
	PPE	16,417,102	23,760,517	33,480,857	33,193,128	29,333,485	0.79
Outputs	Granted patent	0	0	0	0	0	0.00
	Cited patent	0	0	0	0	0	0.00
	Net income	3,333,448	6,799,517	2,953,727	7,767,794	5,606,468	0.68
	Market value	70,097,710	121,636,733	255,677,777	143,938,543	283,249,797	3.04
Korea
Input	R&D expenses	336,952	275,427	241,324	270,166	200,295	−0.41
	Operating expenses	9,172,087	9,109,538	8,634,539	11,217,869	8,406,480	−0.08
	PPE	3,579,696	4,110,470	4,334,674	4,097,086	3,166,754	−0.12
Outputs	Granted patent	5	2	5	4	3	−0.50
	Cited patent	4	1	2	4	3	−0.29
	Net income	144,768	−217,592	−155,287	−398,272	−1,055,399	−8.29
	Market value	3,654,710	3,211,105	3,020,990	1,931,600	2,005,935	−0.45
Japan
Input	R&D expenses	27,102,572	27,501,930	25,791,903	27,280,507	24,739,988	−0.09
	Operating expenses	340,022,759	348,981,832	328,341,538	340,020,107	306,880,337	−0.10
	PPE	89,404,470	84,037,447	77,628,248	88,942,709	86,617,909	−0.03
Outputs	Granted patent	1,244	1,165	1,120	1,205	1,081	−0.13
	Cited patent	1,062	949	929	1,010	930	−0.12
	Net income	18,107,343	18,709,063	14,638,016	18,083,892	13,510,591	−0.25
	Market value	436,639,343	425,992,462	493,539,768	446,400,858	548,316,298	0.26

Note. All variables have units of thousands (TWD) except for Granted Patent and Cited Patent. K-S test shows that p-value is significant at .01 level.

The Pearson correlation coefficients of the input and output variables of the meta-frontier DEA model are shown in Table 5. The results show a remarkable positive relationship between all the input and output variables. The positive relationship between variables indicates that if an increase in inputs occurs, outputs will increase as well, which means that it achieves the assumption of DEA of the isotonicity relations. Hence, the number of samples used should be at least twice the sum of the number of input and output variables. The criterion was met for this study when seven variables and 95 firm-year observations were included. In other words, the validity of the dynamic DEA in this study was established.

Table 5.

Correlation Coefficients of DEA Variables.

	R&D expenses	Operating expenses	PPE	Granted patent	Cited patent	Net income	Market value
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
R&D expenses (1)	1.000
Operating expenses (2)	.961*	1.000
PPE (3)	.975*	.934*	1.000
Granted patent (4)	.978*	.953*	.970*	1.000
Cited patent (5)	.976*	.945*	.976*	.999*	1.000
Net income (6)	.493*	.406*	.524*	.462*	.471*	1.000
Market value (7)	.818*	.754*	.833*	.769*	.773*	.763*	1.000

Correlation is significant at .05 level.

Technology Gap Ratio Analysis from the Global Optical Component Industry

This study has two stages. The first stage utilizes the meta-frontier DEA model to identify the competitiveness of the optical component industry. The mean of efficiency (MOE) and mean of inefficiency (MOI) values of the meta-frontier and group-specific frontiers (GOE) were examined, and the empirical results are presented in Table 6. Among the 19 companies, three companies (Asia Optical, Canon, and Largan Precision) are relatively efficient (MOE and GOE are equal to 1) when the meta-technology set and group technology set were considered the evaluation basis. Therefore, these companies are competitive. Among the efficient companies, one is in Japan, and two are in Taiwan, indicating that the three efficient companies allocate their resources more efficiently to increase their efficiency.

Table 6.

Mean Efficiency and Inefficiency of Meta-frontier and Group Frontiers of Optical Component Industry.

Company name	Country/region	MOE	GOE	OTGR	OTGI	OTI	MOI
Ability Opto-Electronics Technology	Taiwan	0.911	0.911	1.000	0.000	0.089	0.089
Appro Technology	Taiwan	0.954	0.968	0.983	0.014	0.032	0.046
Asia Optical	Taiwan	1.000	1.000	1.000	0.000	0.000	0.000
Calin Technology	Taiwan	0.943	0.943	1.000	0.000	0.057	0.057
Canon	Japan	1.000	1.000	1.000	0.000	0.000	0.000
Genius Electronic Optical	Taiwan	0.728	0.778	0.937	0.050	0.222	0.272
Haesung Optics	South Korea	0.799	0.956	0.831	0.157	0.044	0.201
Hoya	Japan	0.937	1.000	0.937	0.063	0.000	0.063
Kinko Optical	Taiwan	0.930	0.934	0.996	0.004	0.066	0.070
Largan Precision	Taiwan	1.000	1.000	1.000	0.000	0.000	0.000
Newmax Technology	Taiwan	0.753	0.754	0.999	0.001	0.246	0.247
Nikon	Japan	0.917	0.943	0.969	0.026	0.057	0.083
OFilm	China	0.847	0.880	0.963	0.033	0.120	0.153
Ohara	Japan	0.600	0.986	0.608	0.386	0.014	0.400
Olympus	Japan	0.998	0.998	1.000	0.000	0.002	0.002
Q Technology	China	0.715	0.881	0.820	0.166	0.119	0.285
Ricoh	Japan	0.997	0.997	1.000	0.000	0.003	0.003
Sekonix	South Korea	0.529	0.913	0.589	0.383	0.087	0.471
Sunny Optical Technology	China	0.798	0.965	0.823	0.167	0.035	0.202

When the group technology set was utilized as the evaluation basis, only one company (Hoya from Japan) turned out efficient (GOE = 1), which indicates that it is efficient by viewing it under the group technology set, the heterogeneity of technology between the meta-frontier and group frontiers occurs in this company. Considering that only Hoya from Japan was considered efficient under the group technology set, then no technology heterogeneity exists in the companies in Taiwan, China, and South Korea. In terms of OTGI, Ohara from Japan had the highest OTGI (0.386), followed by Sekonix from South Korea (0.383). Ohara and Sekonix had the largest heterogeneity of technology between the meta-frontier and group frontiers among all the studied companies.

The results of the mean of efficiency and inefficiency of the meta-frontier and group frontiers of the country or region from 2016 to 2020 are shown in Table 7. Overall, the results show that no company was operationally efficient (MOE and GOE are equal to 1). However, the companies in Japan and Taiwan had a relatively higher efficiency performance compared with the companies in China and South Korea, meaning that they have higher competitiveness. The companies in South Korea had the lowest efficiency performance.

Table 7.

Mean Efficiency and Inefficiency of Meta-frontier and Group Frontiers of Country or Region from 2016 to 2020.

Country/region year	MOE	GOE	OTGR	OTGI	OTI	MOI
China
2016	0.711	0.910	0.785	0.199	0.090	0.289
2017	0.779	0.900	0.870	0.120	0.100	0.221
2018	0.831	0.957	0.873	0.126	0.043	0.169
2019	0.796	0.874	0.912	0.077	0.126	0.204
2020	0.817	0.903	0.904	0.087	0.097	0.183
Japan
2016	0.867	0.969	0.894	0.102	0.031	0.133
2017	0.914	0.988	0.921	0.074	0.012	0.086
2018	0.915	0.985	0.930	0.070	0.015	0.085
2019	0.917	0.998	0.919	0.081	0.002	0.083
2020	0.928	0.997	0.931	0.069	0.003	0.072
Taiwan
2016	0.878	0.889	0.986	0.011	0.111	0.122
2017	0.908	0.908	1.000	0.000	0.092	0.092
2018	0.934	0.934	1.000	0.000	0.066	0.066
2019	0.909	0.917	0.991	0.008	0.083	0.091
2020	0.884	0.908	0.970	0.024	0.092	0.116
South Korea
2016	0.573	1.000	0.573	0.427	0.000	0.427
2017	0.663	0.892	0.736	0.229	0.108	0.337
2018	0.766	1.000	0.766	0.234	0.000	0.234
2019	0.554	0.891	0.635	0.337	0.109	0.446
2020	0.764	0.889	0.839	0.125	0.111	0.236

When using the group technology set as the evaluation basis, the companies in South Korea turned out efficient (GOE = 1) in 2016 and 2018, but the companies in Japan had a higher GOE in the five-year period. Generally, the companies in South Korea also had the highest OTGI whereas the companies in Taiwan had the lowest OTGI in the 5-year period.

Meta-Frontier Inefficiencies from the Global Optical Component Industry

Many studies explored the efficiency performance of companies from different industries using various types of DEA models, but few studies focused on meta-frontier efficiencies (Chao et al., 2018; Chiu et al., 2016; Lu et al., 2023). The studies that explored the meta-frontier efficiencies focused on companies from different industries. The present study is the first to explore the meta-frontier inefficiencies of global companies from the optical component industry.

The MOI values of the global companies in the optical component industry are shown in Tables 6 and 7. The main reason for the results of the MOI could be due to its breakdown as the inefficiencies of the target company in GOE (OTGI) and the managerial inefficiencies of the target company in GOE (OTI). The OTI has to be equal to zero when GOE is unity. The MOI is derived primarily from OTGI, such as that in Hoya from Japan. The GOE has to be equal to MOE when OTGI is equal to zero, such as that in Olympus from Japan. These results indicate that the inefficiency might be extracted from the input and output excess or output shortfall.

According to Table 6, Sekonix from South Korea had the highest MOI (0.447) among the 19 companies, followed by Ohara from Japan (0.400). According to Table 7, the companies in South Korea had the highest MOI overall, and they had the highest MOI in 2019 (0.446). In comparison, the companies in Taiwan had the lowest MOI in 2018 (0.066).

Relationship Between ESGP and Competitiveness

The second stage examines the relationships of ESGP and its components with MOE, GOE, and OTGR using OLS and truncated regression. The OLS regression results of the relationships of ESG score and ESG components with MOE, GOE, and OTGR are shown in Table 8. ESG affected MOE positively (coefficient = 0.003, p < .01). The ESG components also positively affected MOE (E coefficient = 0.003, p < .01; S coefficient = 0.003, p < .01; G coefficient = 0.003, p < .01). By contrast, ESG had no significant relationship with GOE (coefficient = 0.001, p > .01). E and G positively affected GOE (coefficient = 0.001, p < .10; coefficient = 0.001, p < .05 respectively).

Table 8.

Regression Results of ESG and ESG Components.

Variables	MOE				GOE				OTGR
Intercept	0.722***	0.622**	0.718***	0.665**	1.071***	1.069***	1.029***	1.105***	0.626***	0.525***	0.665***	0.529**
Intercept	0.010	0.016	0.010	0.016	0.001	0.001	0.001	0.001	0.008	0.012	0.005	0.026
ESG	0.003***				0.001				0.003***
ESG	0.002				0.142				0.007
E		0.003***				0.001*				0.002***
E		0.004				0.064				0.007
S			0.003***				0.000				0.003***
S			0.002				0.563				0.002
G				0.002***				0.001**				0.002**
G				0.006				0.037				0.034
ROA	0.262***	0.266***	0.261***	0.258***	−0.010	−0.009	−0.011	−0.010	0.294***	0.296***	0.294***	0.290***
ROA	0.001	0.001	0.001	0.001	0.778	0.803	0.780	0.758	0.001	0.001	0.001	0.001
LEV	0.024	−0.043	0.005	0.045	−0.228***	−0.240***	−0.236***	−0.212***	0.260***	0.204*	0.250**	0.266***
LEV	0.854	0.692	0.976	0.689	0.006	0.001	0.005	0.005	0.022	0.069	0.024	0.026
AGE	0.000	0.000	0.000	0.001	0.001*	0.001*	0.001**	0.001**	0.000	0.000	−0.001	0.000
AGE	0.851	0.500	0.871	0.316	0.062	0.059	0.035	0.041	0.526	0.821	0.253	0.882
SIZE	0.000	0.007	0.001	0.002	−0.007	−0.007	−0.004	−0.009	0.007	0.014	0.006	0.012
SIZE	0.990	0.573	0.941	0.890	0.279	0.236	0.483	0.111	0.525	0.168	0.594	0.301

Note. ESG = ESG score; E = environmental score; S = social score; G = governance score; ROA = net income/total assets; LEV = total liability/ total asset; AGE = logged firm age; SIZE = logged total asset.

***

Statistical significance at 1% level.

Statistical significance at 5% level.

Statistical significance at 10% level.

Furthermore, ESG positively affects OTGR (coefficient = 0.003, p < .01). The components of ESG also positively affect OTGR (E coefficient = 0.002, p < .01; S coefficient = 0.003, p < .01; G coefficient = 0.002, p < .01). The results indicate that the effect of ESG and its components on MOE and OTGR were remarkable, whereas only the effects of the E and G on GOE was substantial.

Four control variables were added to the regression model. The result showed that only ROA positively affected MOE, and LEV and AGE positively affected GOE, where the effect of LEV was more substantial than the effect of AGE on average. ROA and LEV positively affected OTGR, where the effect of ROA was slightly more remarkable than the effect of LEV on average. SIZE had no considerable effects on MOE, GOE, and OTGR. The unablated variance inflation factor of all the components was lower than 5, which indicates the absence of a multicollinearity issue in this study (Khong et al., 2020; Pakravan-Charvadeh et al., 2021).

The result accepting H0 as ESGP positively impacts firm competitiveness, which is similar to prior studies (Broadstock et al., 2020; Brogi & Lagasio, 2019; Dalal & Thaker, 2019; Di Simone et al., 2022; Kluza et al., 2021; Nekhili et al., 2021; Tang, 2022; Zhao et al., 2018; Zhou et al., 2022). This result could be interpreted as firms with strong ESGP are viewed more favorably by consumers, investors, and other stakeholders. A positive reputation helps attract and retain customers, leading to higher sales and market share. Therefore, stakeholders and investors should be aware of the importance of ESGP in making decisions and investing more in ESG, as ESG engagement benefits from its involvement in firms’ financial activities and social responsibility engagement (Chouaibi et al., 2022).

Discussion

This study explored firm competitiveness in the optical component industry in East Asia, which is rare to find in prior studies. Prior studies focused on other optical-related industries, such as LED or LCD manufacturing and optical fiber manufacturing. Moreover, the relationship between ESGP and competitiveness by integrating meta-frontier DEA and OLS regression model has been rarely examined in prior studies, especially in the optical component industry.

The results from Table 6 indicate that Asia Optical, Canon, and Largan Precision are highly competitive, as they are efficient for meta-frontier and group frontier. They allocate their resources and investment efficiently to obtain the most ideal outcome and increase their competitiveness (Du et al., 2014). Companies with lower competitiveness should try to reallocate their resources in a more efficient form and spare their investment into R&D investment to obtain more competitive advantage (Xu et al., 2019)

The findings are consistent with both RBV and stakeholder theory. From the RBV perspective, ESGP can be seen as a strategic capability that enhances competitiveness when it is effectively embedded in firms’ resource allocation and operational processes. The high efficiency of Taiwanese and Japanese firms, particularly Asia Optical, Canon, and Largan Precision, suggests that these firms have developed ESG-related capabilities that are valuable, rare, and difficult to replicate. Their ability to leverage ESGP to strengthen both efficiency and competitiveness highlights the role of ESG as a dynamic capability that supports long-term advantage. However, the weak significance of environmental and governance dimensions on GOE indicates that institutional structures and existing regulatory frameworks in the region may reduce the differentiation effect of these pillars. Environmental initiatives may be shaped more by national policy frameworks than firm-specific strategies, thereby limiting their contribution to performance variation. The positive association between ESGP, MOE, and OTGR further reinforces the notion that ESG-related capabilities function as dynamic resources that enable firms to adapt to technological heterogeneity and industry demands.

From the perspective of Stakeholder Theory, the consistently positive role of ESGP in shaping efficiency and competitiveness underscores the value of aligning with stakeholder expectations. The higher GOE or OTGR observed in firms such as Hoya (Japan), Ohara (Japan) and Sekonix (South Korea) reflect disparities in R&D intensity and innovation ecosystems across the region. Countries with stronger R&D support and industrial policy, such as Taiwan and Japan, appear better positioned to translate ESG practices into technological competitiveness, whereas firms in South Korea and China face greater challenges in bridging the technology frontier. These results highlight that ESGP contributes not only to firm-level resource advantages but also to stakeholder trust and legitimacy, with regional policy environments and innovation capacity acting as critical boundary conditions. Overall, the findings suggest that ESGP operates as both a strategic resource and a stakeholder engagement mechanism, while policy and R&D disparities mediate its effectiveness in addressing technological heterogeneity.

The results of the meta-frontier DEA model in DDF could exhibit the meta-frontier and group frontier among the companies and show the inefficiencies of the companies. Some of the companies that are included in the samples were removed due to data unavailability. The regression results of ESGP and GOE were unremarkable possibly because of the small number of samples in each group. Future research could try using a similar method with a larger number of samples in each group so that the method could be more accurate.

Conclusion

This study explores the relationship between the ESGP and the competitiveness of top-performing optical component manufacturing organizations from four Asian countries, namely, Taiwan, Japan, South Korea, and China, from 2016 to 2020. The results show that the optical component companies from Taiwan and Japan have better results, which means that they have higher overall competitiveness. Hence, the companies from China and South Korea are suggested to refer to the companies from Taiwan and Japan to improve their competitiveness. ESGP has a remarkably positive relationship with MOE and OTGR and a less remarkable relationship with GOE.

This study utilized meta-frontier DEA to evaluate the efficiency of optical component companies, where firm competitiveness was explored, which has never been done in prior studies. A meta-frontier DEA model was developed to evaluate the DMUs from all groups and individual groups from group frontiers to identify the effects of technological heterogeneity. Traditional DEA could not identify group frontier and technological heterogeneity as compared to meta-frontier DEA. This study also compares samples from different countries and years of observation.

Even though some prior studies investigated the relationship between ESGP and efficiency performance using DEA models, they focused on the LED manufacturing industry instead of the optical component industry. This study provides a meta-frontier analysis to examine the operational management efficiencies and inefficiencies of optical component companies. It also explores the impact of ESGP toward the competitiveness of firms in the optical component industry. As shown in the Appendix, robust tests were conducted using truncated regression and bootstrap truncated regression.

Future research could dive into the role digital technology (i.e. artificial intelligence, big data analytics, and IoT) in improving the competitiveness and efficiency of the optical components industry. Production processes, enable predictive maintenance, and foster innovation in product development could be improved. In addition, the use of digital tools for ESGP monitoring and reporting can help companies achieve the stakeholder expectations and regulatory requirements.

In this study, some limitations were identified. First, top-performing companies from the optical component industry were included, but some companies were removed because of limited data availability. Second, the main advantage of utilizing the meta-frontier DEA model, which is the consideration of undesirable outputs, was not included in this study. For instance, the undesirable variable such as CO₂ Emissions and energy use could be added to include the sustainability aspect of the company. Third, the study only included a one-stage DEA model. More aspects of efficiency and the causal relationship of the efficiency performance could be demonstrated if a two-stage DEA model is used.

Managerial Implication

This study highlights the critical need for optical component decision to strategically allocate their resource especially on R&D investments. During the process of engaging in ESGP initiatives, managers should prioritize innovations following the industry trends to enhance their competitive advantage. Optical firms in China and South Korea can improve their competitiveness by learning from the ESG strategies of Taiwanese and Japanese firms. Companies should incorporate ESG practices into business decisions and daily operations is essential for reducing technology gaps and improving efficiency. By identifying inefficiencies which could assist in optimizing their resource use, companies can achieve greater operational efficiency and help in long-term growth. High-performing companies can serve as benchmarks for best practices, helping others refine their strategies.

Practical and Social Implication

This study provides practical insights for companies seeking to improve their competitiveness in the optical components industry. The use of meta-frontier DEA and OLS regression provides a robust framework for evaluating performance and identifying areas for improvement. Companies can use these methods to assess their relative efficiency, reallocate resources, and enhance investment strategies. By integrating ESGP into their operational strategy, companies can align financial goals with environmental and social performance to ensure sustainable growth in a highly competitive marketplace. The findings highlight the need for platforms that support cross-country knowledge sharing, training programs, and resource planning that connect ESG adoption with technological progress. Governments and regulators play a key role to promote ESG integration in helping firms to enhance competitiveness with long-term social and environmental goals.

Optical component companies that engage in ESG could contributes to broader societal goals such as sustainability and climate change mitigation. Stronger ESG performance in the optical component industry supports sustainable growth, builds trust among stakeholders, and improves regional competitiveness, even potentially benefits for other manufacturing sectors in Asia. By aligning business practices with stakeholder expectations, companies can foster trust and cooperation with communities, regulators and consumers. This approach not only enhances brand reputation, but also promotes a more sustainable industry ecosystem.

Footnotes

Appendix

Table A3.

Bootstrap Truncated Regression Results of ESG and ESG Components.

Variables	MOE				GOE				OTGR
Intercept	0.722***	0.622***	0.718***	0.665***	1.071***	1.069***	1.029***	1.105***	0.626***	0.525***	0.665***	0.529***
ESG	0.003***				0.001*				0.003***
E		0.003***				0.001**				0.002***
S			0.003***				0.000				0.003***
G				0.002***				0.001***				0.002**
ROA	0.262***	0.266***	0.261***	0.258***	−0.010	−0.009	−0.011	−0.010	0.294***	0.296***	0.294***	0.290***
LEV	0.024	−0.043	0.005	0.045	−0.228***	−0.240***	−0.236***	−0.212***	0.260***	0.204**	0.250***	0.266***
AGE	0.000	0.000	−0.000	0.001	0.001**	0.001**	0.001***	0.001**	−0.000	−0.000	−0.001	0.000
SIZE	0.000	0.007	0.001	0.002	−0.007	−0.007	−0.004	−0.009*	0.007	0.014*	0.006	0.012

Note. 2,000 bootstrap replications are used. ESG = ESG score; E = environmental score; S = social score; G = governance score; ROA = net income/total assets; LEV = total liability/total asset; AGE = logged firm age; SIZE = logged total asset.

***

Statistical significance at 1% level.

Statistical significance at 5% level.

Statistical significance at 10% level.

ORCID iDs

Tzu Kuan Lin

Fang Chen Kao

Author Contributions

Tzu Kuan Lin: Formal Analysis, Writing – Original Draft, Review & Editing, Data Curation, Resources. Fang Chen Kao: Supervision, Software, Methodology

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

References

Abdi

Càmara-Turull

(2022). Exploring the impact of sustainability (esg) disclosure on firm value and financial performance (fp) in airline industry: The moderating role of size and age. Environment, Development and Sustainability, 24(4), 5052–5079. https://doi.org/10.1007/s10668-021-01649-w

Behl

Kumari

P. S. R.

Makhija

Sharma

(2022). Exploring the relationship of esg score and firm value using cross-lagged panel analyses: Case of the indian energy sector. Annals of Operations Research, 313(1), 231–256. https://doi.org/10.1007/s10479-021-04189-8

Bridoux

Stoelhorst

(2022). Stakeholder theory, strategy, and organization: Past, present, and future. Strategic Organization, 20(4), 797–809. https://doi.org/10.1177/14761270221127628

Broadstock

D. C.

Matousek

Meyer

Tzeremes

N. G.

(2020). Does corporate social responsibility impact firms’ innovation capacity? The indirect link between environmental & social governance implementation and innovation performance. Journal of Business Research, 119, 99–110. https://doi.org/10.1016/j.jbusres.2019.07.014

Broekel

Rogge

Brenner

(2018). The innovation efficiency of german regions: A shared-input dea approach. Review of Regional Research, 38(1), 77–109. https://doi.org/10.1007/s10037-017-0112-0

Brogi

Lagasio

(2019). Environmental, social, and governance and company profitability: Are financial intermediaries different? Corporate Social Responsibility and Environmental Management, 26(3), 576–587. https://doi.org/10.1002/csr.1704

Chang

P. C.

Tsou

N. T.

Yuan

B. J. C.

Huang

C. C.

(2002). Development trends in Taiwan’s opto-electronics industry. Technovation, 22(3), 161–173. https://doi.org/10.1016/S0166-4972(01)00008-6

Chao

C. M.

M. M.

Hsiung

N. H.

Chen

L. H.

(2018). Profitability efficiency, marketability efficiency and technology gaps in taiwan’s banking industry: Meta-frontier network data envelopment analysis. Applied Economics, 50(3), 233–250. https://doi.org/10.1080/00036846.2017.1316827

Charles

Färe

Grosskopf

(2016). A translation invariant pure dea model. European Journal of Operational Research, 249(1), 390–392. https://doi.org/10.1016/j.ejor.2015.09.037

10.

Charnes

Cooper

W. W.

Rhodes

(1978). Measuring the efficiency of decision making units. European Journal of Operational Research, 2(6), 429–444.

11.

Charnes

Cooper

Lewin

A. Y.

Seiford

L. M.

(1997). Data envelopment analysis theory, methodology and applications. Journal of the Operational Research Society, 48(3), 332–333. https://doi.org/10.1057/palgrave.jors.2600342

12.

Cheng

Jin

Jiang

Hua

(2020). Efficiency assessment of rural domestic sewage treatment facilities by a slacked-based dea model. Journal of Cleaner Production, 267, Article 122111. https://doi.org/10.1016/j.jclepro.2020.122111

13.

Chiu

C. R.

Chiu

Y. H.

Chen

Y. C.

Fang

C. L.

(2016). Exploring the source of metafrontier inefficiency for various bank types in the two-stage network system with undesirable output. Pacific-Basin Finance Journal, 36, 1–13. https://doi.org/10.1016/j.pacfin.2015.11.003

14.

Chiu

C. R.

Liou

J. L.

P. I.

Fang

C. L.

(2012). Decomposition of the environmental inefficiency of the meta-frontier with undesirable output. Energy Economics, 34(5), 1392–1399. https://doi.org/10.1016/j.eneco.2012.06.003

15.

Chiu

C. R.

K. H.

Tsang

S. S.

Chen

Y. F.

(2013). Decomposition of meta-frontier inefficiency in the two-stage network directional distance function with quasi-fixed inputs. International Transactions in Operational Research, 20(4), 595–611. https://doi.org/10.1111/itor.12008

16.

Chiu

Y. H.

M. F.

Cheng

C. H.

(2008). The influence of exchange rate gains or losses on the operational efficiency of the taiwanese led industry: An application of dea. Journal of Statistics and Management Systems, 11(4), 787–804. https://doi.org/10.1080/09720510.2008.10701343

17.

Chouaibi

Rossi

(2022). Esg and corporate financial performance: The mediating role of green innovation: Uk common law versus germany civil law. EuroMed Journal of Business, 17(1), 46–71. https://doi.org/10.1108/EMJB-09-2020-0101

18.

Chung

Y. H.

Färe

Grosskopf

(1997). Productivity and undesirable outputs: A directional distance function approach. Journal of Environmental Management, 51(3), 229–240. https://doi.org/10.1006/jema.1997.0146

19.

H. C.

Chew

K. S.

(1997). Performance and R&D expenditures in american and japanese manufacturing firms. International Journal of Production Research, 35(12), 3333–3348. https://doi.org/10.1080/002075497194101

20.

Dalal

K. K.

Thaker

(2019). Esg and corporate financial performance: A panel study of indian companies. IUP Journal of Corporate Governance, 18(1), 44–59. https://search-ebscohost-com-443.web.bisu.edu.cn/login.aspx?direct=true&db=buh&AN=137221144&site=ehost-live

21.

Di Simone

Petracci

Piva

. (2022). Economic sustainability, innovation, and the esg factors: An empirical investigation. Sustainability, 14(4), 2270. https://www-mdpi-com-s.web.bisu.edu.cn/2071-1050/14/4/2270

22.

Cook

W. D.

Liang

Zhu

(2014). Fixed cost and resource allocation based on dea cross-efficiency. European Journal of Operational Research, 235(1), 206–214. https://doi.org/10.1016/j.ejor.2013.10.002

23.

Elias

A. K.

Omama

K. F.

Cornelia

D. C.

Adrian

M. A.

Imola

Ana

(2024). Green financing and sustainability reporting among SMEs: The role of pro-environmental behavior and digitization. Journal of Cleaner Production, 478, 143939. https://doi.org/10.1016/j.jclepro.2024.143939

24.

European Commission. (2021, July 14). Carbon border adjustment mechanism: Questions and answers.Retrieved September 27, 2022, from https://ec.europa.eu/commission/presscorner/detail/en/qanda_21_3661

25.

Färe

Charles

(2018). A directional distance function approach to void the non-archimedean in dea. Journal of the Operational Research Society, 69(5), 772–775. https://doi.org/10.1057/s41274-017-0264-x

26.

Fei

Lin

(2016). Energy efficiency and production technology heterogeneity in China’s agricultural sector: A meta-frontier approach. Technological Forecasting and Social Change, 109, 25–34. https://doi.org/10.1016/j.techfore.2016.05.012

27.

Feng

Zhang

Chiu

Y.-h.

Chang

T.-H.

(2021). Innovation efficiency and the impact of the institutional quality: A cross-country analysis using the two-stage meta-frontier dynamic network DEA model. Scientometrics, 126(4), 3091–3129. https://doi.org/10.1007/s11192-020-03829-3

28.

Freeman

R. E.

Phillips

R. A.

Sisodia

R. S.

(2020). Tensions in Stakeholder Theory. Business & Society, 59, 213–231. https://doi.org/10.1177/0007650318773750

29.

Fried

H. O.

Lovell

C. A. K.

Schmidt

S. S.

(2008). The measurement of productive efficiency and productivity change. Oxford University Press. https://doi.org/10.1093/acprof:oso/9780195183528.001.0001

30.

Gillan

S. L.

Koch

Starks

L. T.

(2021). Firms and social responsibility: A review of esg and csr research in corporate finance. Journal of Corporate Finance, 66, Article 101889. https://doi.org/10.1016/j.jcorpfin.2021.101889

31.

Goyal

Singh

Aggarwal

(2019). Efficiency and technology gaps in Indian banking sector: Application of meta-frontier directional distance function DEA approach. The Journal of Finance and Data Science, 5(3), 156–172. https://doi.org/10.1016/j.jfds.2018.08.002

32.

Hashimoto

Haneda

(2008). Measuring the change in r&d efficiency of the japanese pharmaceutical industry. Research Policy, 37(10), 1829–1836. https://doi.org/10.1016/j.respol.2008.08.004

33.

Jens

S.-E.

Petra

(2009). Research efficiency in manufacturing: An application of dea at the industry level. https://ideas.repec.org/p/diw/diwwpp/dp884.html

34.

Khong

T. D.

Loch

Young

M. D.

(2020). Perceptions and responses to rising salinity intrusion in the mekong river delta: What drives a long-term community-based strategy? Science of The Total Environment, 711, Article 134759. https://doi.org/10.1016/j.scitotenv.2019.134759

35.

Kluza

Ziolo

Spoz

(2021). Innovation and environmental, social, and governance factors influencing sustainable business models: Meta-analysis. Journal of Cleaner Production, 303, Article 127015. https://doi.org/10.1016/j.jclepro.2021.127015

36.

Knox Lovell

C. A.

Pastor

J. T

. (1995). Units invariant and translation invariant dea models. Operations Research Letters, 18(3), 147–151. https://doi.org/10.1016/0167-6377(95)00044-5

37.

Kotabe

Zhao

(2002). A taxonomy of sourcing strategic types for mncs operating in China. Asia Pacific Journal of Management, 19(1), 11–27. https://doi.org/10.1023/A:1014835506271

38.

Kweh

Q. L.

W.-M.

Lin

Deng

Jr. (2022). Impact of research and development tax credits on the innovation and operational efficiencies of internet of things companies in taiwan. Annals of Operations Research, 315(2), 1217–1241. https://doi.org/10.1007/s10479-020-03880-6

39.

Lee

C. L.

Liang

(2024). The effect of carbon regulation initiatives on corporate ESG performance in real estate sector: International evidence. Journal of Cleaner Production, 453, Article 142188. https://doi.org/10.1016/j.jclepro.2024.142188

40.

Lee

Kwon

H.-B.

Pati

(2019). Exploring the relative impact of r&d and operational efficiency on performance: A sequential regression-neural network approach. Expert Systems With Applications, 137, 420–431. https://doi.org/10.1016/j.eswa.2019.07.026

41.

Lee

Z.-Y.

Pai

C.-C.

(2011). Operation analysis and performance assessment for tft-lcd manufacturers using improved dea. Expert Systems With Applications, 38(4), 4014–4024. https://doi.org/10.1016/j.eswa.2010.09.063

42.

W.-M.

Kweh

Q. L.

Ting

I. W. K.

Ren

(2023). How does stakeholder engagement through environmental, social, and governance affect eco-efficiency and profitability efficiency? Zooming into apple inc.’S counterparts. Business Strategy and the Environment, 32(1), 587–601. https://doi.org/10.1002/bse.3162

43.

See

K. F.

M.-M.

Zhao

(2021). Research efficiency analysis of China’s university faculty members: A modified meta-frontier DEA approach. Socio-Economic Planning Sciences, 76, Article 100944. https://doi.org/10.1016/j.seps.2020.100944

44.

Miralles-Quirós

M. M.

Miralles-Quirós

J. L.

Redondo Hernández

(2019). Esg performance and shareholder value creation in the banking industry: International differences. Sustainability, 11(5), 1404. https://www-mdpi-com-s.web.bisu.edu.cn/2071-1050/11/5/1404

45.

Nekhili

Boukadhaba

Nagati

Chtioui

(2021). Esg performance and market value: The moderating role of employee board representation. The International Journal of Human Resource Management, 32(14), 3061–3087. https://doi.org/10.1080/09585192.2019.1629989

46.

Nirino

Santoro

Miglietta

Quaglia

(2021). Corporate controversies and company’s financial performance: Exploring the moderating role of esg practices. Technological Forecasting and Social Change, 162, Article 120341. https://doi.org/10.1016/j.techfore.2020.120341

47.

O’Donnell

C. J.

Rao

D. S. P.

Battese

G. E.

(2008). Metafrontier frameworks for the study of firm-level efficiencies and technology ratios. Empirical Economics, 34(2), 231–255. https://doi.org/10.1007/s00181-007-0119-4

48.

Pakravan-Charvadeh

M. R.

Mohammadi-Nasrabadi

Gholamrezai

Vatanparast

Flora

Nabavi-Pelesaraei

(2021). The short-term effects of covid-19 outbreak on dietary diversity and food security status of iranian households (a case study in tehran province). Journal of Cleaner Production, 281, Article 124537. https://doi.org/10.1016/j.jclepro.2020.124537

49.

Pandit

Wasley

C. E.

Zach

(2011). The effect of research and development (R&D) inputs and outputs on the relation between the uncertainty of future operating performance and R&D expenditures. Journal of Accounting, Auditing & Finance, 26(1), 121–144. https://doi.org/10.1177/0148558x11400583

50.

Sharma

Thomas

(2008). Inter-country R&D efficiency analysis: An application of data envelopment analysis. Scientometrics, 76(3), 483–501. https://doi.org/10.1007/s11192-007-1896-4

51.

Sherman

H. D.

Ladino

(1995). Managing bank productivity using data envelopment analysis (dea). Interfaces, 25(2), 60–73. https://doi.org/10.1287/inte.25.2.60

52.

Sun

C. C.

(2014). Assessing the relative efficiency and productivity growth of the taiwan led industry: Dea and malmquist indices application. Mathematical Problems in Engineering, 2014(1), Article 816801. https://doi.org/10.1155/2014/816801

53.

Tang

(2022). The effect of esg performance on corporate innovation in china: The mediating role of financial constraints and agency cost. Sustainability, 14(7), 3769. https://www-mdpi-com-s.web.bisu.edu.cn/2071-1050/14/7/3769

54.

Vantage Market Research. (2022, January 21). Optoelectronic components market: Global industry assessment & forecast. Retrieved February 27, 2023, from https://www.vantagemarketresearch.com/industry-report/optoelectronic-components-market-1154

55.

Velte

(2017). Does esg performance have an impact on financial performance? Evidence from germany. Journal of Global Responsibility, 8(2), 169–178. https://doi.org/10.1108/JGR-11-2016-0029

56.

Vladislav

Aleksandr

Lubov

Nataliya

Darya

Michael

(2016July). Product innovation efficiency of russian electronic industry: Dea approach and cluster analysis [Conference session]. 2016 2nd International Conference on Social Science and Higher Education. Sanya, China.

57.

Wang

C.-N.

H.-X. T.

Luo

S.-H.

Yang

Y.-S.

(2018). Using malmquist index to select the strategic partners in vertical integration: The case of taiwan led supply chain. Applied Economics Letters, 25(19), 1336–1341. https://doi.org/10.1080/13504851.2017.1420875

58.

Wang

Zhang

Lei

Qiu

(2021). Evaluation on the efficiency of led energy enterprises in china by employing the dea model. Mathematics, 9(19), 2356. https://www-mdpi-com-s.web.bisu.edu.cn/2227-7390/9/19/2356

59.

Wang

Zhao

Zhou

(2013). Energy efficiency and production technology heterogeneity in china: A meta-frontier dea approach. Economic Modelling, 35, 283–289. https://doi.org/10.1016/j.econmod.2013.07.017

60.

Xiao

Q.-W.

Tian

Ren

F.-R.

(2022). Efficiency assessment of electricity generation in China using meta-frontier data envelopment analysis: Cross-regional comparison based on different electricity generation energy sources. Energy Strategy Reviews, 39, Article 100767. https://doi.org/10.1016/j.esr.2021.100767

61.

Xie

Nozawa

Yagi

Fujii

Managi

(2019). Do environmental, social, and governance activities improve corporate financial performance? Business Strategy and the Environment, 28(2), 286–300. https://doi.org/10.1002/bse.2224

62.

Zhou

K. Z.

(2019). Deviant versus aspirational risk taking: The effects of performance feedback on bribery expenditure and r&d intensity. Academy of Management Journal, 62(4), 1226–1251. https://doi.org/10.5465/amj.2016.0749

63.

Yang

Notteboom

(2022). Chinese investment in overseas container terminals: The role of investor attributes in achieving a higher port competitiveness. Transport Policy, 118, 112–122. https://doi.org/10.1016/j.tranpol.2022.01.024

64.

Yang

Wei

Huang

Chen

(2018). Expected efficiency based on directional distance function in data envelopment analysis. Computers & Industrial Engineering, 125, 33–45. https://doi.org/10.1016/j.cie.2018.08.010

65.

M.-M.

Chen

L.-H.

(2020). A meta-frontier network data envelopment analysis approach for the measurement of technological bias with network production structure. Annals of Operations Research, 287(1), 495–514. https://doi.org/10.1007/s10479-019-03436-3

66.

Zhang

Zhou

Choi

(2013). Energy efficiency, co2 emission performance and technology gaps in fossil fuel electricity generation in korea: A meta-frontier non-radial directional distance functionanalysis. Energy Policy, 56, 653–662. https://doi.org/10.1016/j.enpol.2013.01.033

67.

Zhao

Guo

Yuan

Zhou

Kang

(2018). Esg and corporate financial performance: Empirical evidence from China’s listed power generation companies. Sustainability, 10(8), 2607. https://www-mdpi-com-s.web.bisu.edu.cn/2071-1050/10/8/2607

68.

Zheng

Lee

C. L.

(2025). A meta-analysis of ESG factors in the real estate investment trusts sector: Exploring their impacts on REITs performance. Engineering, Construction and Architectural Management. Advance online publication. https://doi.org/10.1108/ECAM-05-2024-0652

69.

Zhou

Liu

Luo

(2022). Sustainable development, esg performance and company market value: Mediating effect of financial performance. Business Strategy and the Environment, 31(7), 3371–3387. https://doi.org/10.1002/bse.3089

70.

Zumente

Bistrova

(2021). Esg importance for long-term shareholder value creation: Literature vs. Practice. Journal of Open Innovation: Technology, Market, and Complexity, 7(2), 127. https://www-mdpi-com-s.web.bisu.edu.cn/2199-8531/7/2/127