Impact of ISO Certifications on Corporate Financial Performance: Evidence from Istanbul Stock Exchange-Listed Manufacturing Companies

Abstract

The literature has reached a consensus that ISO standardization enhances the Environmental, Social, and Governance (ESG) performance of companies, which in turn has a positive effect on corporate financial performance (CFP). There is less understanding in terms of the effect of different certifications and underlying mechanisms between the effect of the ISO certification on the CFP. The purpose of this paper is to investigate the impact of different ISO certifications on the CFP of Turkish companies listed on the Istanbul Stock Exchange (BIST). Based on audited financial statements of a population of 148 manufacturing companies listed during 2010–2022 and using the generalized method of moments (GMM) technique, this study shows that the number of ISO certifications has a positive impact on return on asset (ROA) and Tobin’s Q, however, no direct effect on operational efficient and R&D intensity. While there is no effect of the occupational health and safety management systems certification on ROA and Tobin’s Q, the analysis brought forward that ROA seems to be positively affected by the standards referring to environmental, energy, quality, and information security management systems certification while Tobin’s Q is positively affected by the last two certifications.

1. Introduction

Today, sustainability has gained great importance with its economic, social, and environmental dimensions. Governments, local administrations, the private sector, NGOs, and academic institutions have a responsibility to contribute to sustainability. The private sector, especially the manufacturing industry, has a huge impact on the environment and society during sourcing, production processes, logistics, consumption, and activities at the end-of-life cycle. The industry should dampen the negative impacts of production with environmental, social, and governance (ESG) efforts. ISO standards improve the performance of firms by following managerial guidelines and certifications can provide quality, environmental, social, and governance benefits.

Standards can help reduce costs by promoting operational optimization in various areas. For example, quality certification significantly improves asset productivity by 26% [1]. Moreover, the industry can be held accountable through standards. Compliance with these standards can lead to increased legitimacy and reputation, which can enhance corporate financial performance (CFP). ISO certification can improve customer trust and sales and promote potential market opportunities by increasing international recognition [2]. Quality management systems facilitate the management of environmental processes, interactions with stakeholders, activities within the supply chain, and sustainable supply chain management (SSCM). Moreover, significant supply chain impacts are observed with the implementation of quality and environmental management standards than by either standard alone [3]. The implementation of ISO 50001 [4] has enabled enterprises to effectively achieve their energy conservation objectives [5]. Similarly, benefits multiply when environmental and energy management standards are implemented together since there are compatibilities between both standards [5].

Occupational health and safety issues have gained importance for multiple stakeholders such as shareholders due to costs associated with injury and illness cases, employees, and society in general. The implementation of a safety management standard can result in significant improvements in workplace safety, and signal superior average safety performance to buyers and suppliers [6]. Further evidence documents significant increases in abnormal performance on safety, sales growth, labor productivity, and profitability [7]. The information security standard (ISO 27001) [8] is associated with improvements in profitability, labor productivity, and sales performance [9]. ISO 27001 and financial performance exhibit a positive association. Furthermore, the beneficial influence of ISO 27001 on financial performance gradually grows with time [10].

In summary, ISO Certifications are believed to have lots of advantages such as cost reduction and better compliance with environmental and societal regulations due to operational improvements, increased sales, and better access to finance due to an improved reputation. Like any organizational transformation, to acquire the certification companies must bind limited resources such as capital, time, and human resources. Thus, it is imperative for companies to be confident that there will be a payoff of the certification before finalizing their investment decision. Therefore, this study focuses on the following research questions:

• What is the effect of ISO certification on manufacturing companies’ CFP?
• Which underlying mechanism explains the ISO certification-CFP relationship at best?
• What are the impacts of different ISO certifications on manufacturing companies’ CFP?

The literature advocates that ISO standards can enhance ESG performance, which in turn improves corporate financial performance with plausible explanations such as improved reputation and stakeholder engagement, improved internal processes, and resource management. Thus, companies are expected to benefit from ISO certification; however, to be accredited they must allocate scarce resources. Firms with available funds can pursue costly environmental or social activities [11]. More profitable firms are more likely to adopt international standards for environmental management (EMS) than less successful ones. The bi-directional relationship between ISO certification and CFP performance might cause selection bias, and the documented positive relationships are attributed to selection effects rather than treatment effects [12,13]. This inherent endogeneity in the ISO certification–CFP performance relationship should be considered in explanatory models.

To understand the effect of ISO certification on CFP a population of 148 manufacturing companies listed every year in BIST during a time span of 2010–2022 is employed. This study aims to conduct an analysis for the composite measure of the number of ISO certifications and specifically for certifications of ISO 9001 Quality management systems [14], ISO 14001 Environmental management systems [15], ISO 45001/OHSAS 18001 Occupational health and safety management systems [16], ISO 50001 Energy management [4], and ISO 26000 Social responsibility [17] and ISO 27001 [8] Information security management systems. The estimation procedure includes diversifying dependent variables such as R&D intensity, operational efficiency, accounting, and market-based financial performance measures for testing alternative theories explaining the underlying mechanism of the ISO certification and CFP relationship and applying the General Method of Moments (GMM) model to address the endogeneity issue caused by reverse causality between CFP and the ISO certification.

Since ISO Certification and CFP are not strictly exogenous but dependent on their own past observations, the econometric model should accommodate these dynamics. Moreover, there might be heteroskedasticity, autocorrelation, and cross-sectional dependency across companies. To evaluate the performance based on the performance of the previous period, GMM allows the dependent variables with a one-year lag to appear as explanatory variables in the model, while generating standard errors resilient to heteroskedasticity and autocorrelation. Since ISO certifications are updated every three years at most, the ISO certification with a one-year lag, two-year lag, and three-lag are incorporated as instrumental variables into the analysis.

While there is no evidence for improved R&D intensity and operational efficiency, ISO certification is associated with higher ROA and Tobin’s Q. While there is no effect of ISO 45001/OHSAS 18001 Occupational health and safety management systems certification on ROA and Tobin’s Q, ROA is positively affected by ISO 9001 Quality management systems, ISO 14001 Environmental management systems, ISO 5001 Energy management systems, and ISO 27001 Information security management systems certification and Tobin’s Q is positively affected by ISO 9001 and ISO 27001.

This study contributes to the existing literature by providing evidence on the impact of firm certification—namely quality, environmental, health and safety, and governance management systems—on corporate performance, in terms of the firm’s R&D intensity, operational efficiency, and accounting-based and market-based financial performance in the under-explored context of Türkiye with an emerging economy. The motivation of the study is to reveal the mechanisms underlying the ISO certification–CFP relationship, provide empirical evidence showcasing the potential ISO certification as a solution support tool, and encourage manufacturing companies to adopt ISO certifications and improve their ESG practice, which is an important step towards UN Sustainable Development Goals 8 and 12.

This paper is organized as follows. Section 2 presents the literature review on ISO certifications and CFP. Section 3 provides the estimation framework by describing the data set, variables, and models. Section 4 lays out the empirical application, presents the results, and discusses the theoretical and policy implications. Section 5 acknowledges limitations and concludes with future research opportunities.

2. Literature Review on ISO Certifications and CFP

2.1. Scientometric Analysis

To capture as many relevant studies as possible, a structured keyword search on the ISI Web of Science (WoS) database is conducted. The ISI Web of Science (WoS) database has comprehensive coverage of high-impact journals published in Elsevier, IEEE, Springer, and Taylor & Francis, etc. The following keyword string (ISO certification) and (operational performance) or (financial performance) or (economic performance) is searched in the title, abstract, and keywords. The initial search query (1 March 2024) returned 569 articles in WoS. After filtering for peer-reviewed impact factor publications such as articles, proceeding papers, review articles, and book chapters and English-language articles, 551 remained for evaluation. The search is further refined by including articles strictly from Business, Business Finance, Economics, Engineering Industrial, Engineering Manufacturing, Environmental Studies, Management, and Operations Research Management Science fields and 334 articles remained for evaluation. Based on the screening of the title and the abstracts, the studies can be grouped under (i) Literature Reviews on ISO certification, (ii) determinants of ISO certification and critical success factors for adaptation, (iii) The environmental effects of ISO certification, and (iv) The influence of ISO certification on operational, financial, or economic performance. For the research purposes of the study, the focus is on the last group and 178 publications are shortlisted for full paper reading. Moreover, to capture the relevant studies written about the Turkish Context DergiPark Database (https://dergipark.org.tr/en/ accessed on 1 March 2024 ), the leading database for Turkish Academic Studies is searched. The keyword string (ISO certification) and (operational performance) or (financial performance) or (economic performance) is searched in the title, abstract, and keywords. The initial search query returned 21 articles. Since DergiPark Data does not offer filtering for academic fields, a title and abstract screening has been conducted and 10 publications are shortlisted for full paper reading. Figure 1 depicts the flow diagram for the identification of studies.

The shortlisted papers have been categorized according to certification. The international literature has dedicated most attention to ISO 9001 Certification, namely 82 studies, followed by ISO 14000 Certification with 60 studies. The third prominent research stream includes 17 studies, where multiple certifications are considered. The number of studies concerned only with ISO 50001, ISO 45001/OHSAS 18001, ISO 26000, and ISO 27001 are 6, 4, 2, and 2 respectively. Additional exclusion restrictions are imposed. In total, 11 literature review articles, 25 articles that conduct qualitative research such as case studies or where statistical inference is not possible due to small sample sizes, and 62 articles where data is gathered through surveys or data is cross-sectional have been excluded from further consideration. The remaining studies utilizing panel data have been screened for their methodology. A total of 19 papers that conduct event study methodology have been eliminated since the unit of analysis is certification announcement as opposed to firm-year observations. Furthermore, 21 studies, which rely on data sets where certified and non-certified companies are statistically matched with methods such as propensity score matching and coarsened exact matching and conduct Difference-in-Difference model are excluded. In

Table 1. Exclusion criteria for literature review.

Data Base: ISI Web of Science (WoS)	Reports Assessed for Eligibility	Reports Excluded for					Studies Included in Review
		Literature Review	Qualitative Studies & Small Sample Size	Survey & Cross Sectional Data	Event Study	Statistical Matching & Difference-In Difference Model
ISO 9001	82	4	9	41	6	11	11
ISO 14000	60	3	10	14	9	6	18
ISO 50001	6	0	3	1	1	0	1
ISO 45001/OHSAS 18001	4	0	1	0	2	1	0
ISO 26000	2	0	0	0	0	0	2
ISO 27001	2	0	0	0	1	1	0
Multiple Certification	17	4	2	6	0	2	3
Total	173	11	25	62	19	21	35
Data Base: DergiPark	Reports Assessed for Eligibility	Reports Excluded for					Studies Included in Review
		Literature Review	Qualitative Studies & Small Sample Size	Survey & Cross Sectional Data	Event Study	Statistical Matching & Difference-In Difference Model
ISO 9001	4		2	1			1
ISO 14000	2	1		1
ISO 50001	0
ISO 45001/OHSAS 18001	0
ISO 26000	1						1
ISO 27001	1			1
Multiple Certification	2			2
Total	10	1	2	5			2

, the number of eliminated studies are grouped according to certification type and imposed exclusion criteria. In total, 35 papers remain for full paper consideration.

2.2. Literature Review on ISO Certifications and CFP

Looking into the research aim three clusters emerge. The first cluster analyses the effect of ISO certification on performance. The second cluster aims to analyze the effect of adoption timing on performance. The third cluster aims to showcase the effect of ISO certification on performance by comparison before and after certification. The research on hand aligns with the first cluster. Since the research focus is to reveal the effect of the ISO certification on CFP, studies utilizing financial performance as the dependent variable and ISO certification as the independent variable are included in the final sample.

Different operationalizations of independent, dependent, and control variables, applied methodologies, and main findings are outlined. While the most common measure for ISO certification is a binary variable, ISO 9001 [19,20,21,22], ISO 14001 [6,23,24,25,26], ISO 27000 [27], and ISO 9001 and ISO 14001 [28], different approaches, such as composite measure regarding ISO 14001 [29], ISO 140001 certification and the duration [30], ISO 9001 integration strategy with not certified, certified for less than three years, certified for more than three years as possible strategies [31], integration strategy for multiple certifications with ISO 9001 and ISO 14001 [32], and ISO 9001, ISO 14001 and ISO 45001/OHSAS 18001 [33], are also found in the literature.

The operationalization of dependent variables can be grouped under operational performance measures, accounting-based, and market-based financial performance measures. Pure operational performance measures such as tons of CO₂ emitted by the company per unit of output [34], growth changes in production volume [2], and productivity change [20] can be used to test RBV and dynamic capabilities theories. Accounting-based financial measures such as cost of goods sold/sales revenue, earnings before interest, taxes, depreciation, and amortization (EBITDA) margin, return on asset (ROA), return on equity (ROE), and return on sales (ROS) capture both the effect certification on operational performance and financial outcomes because of the gain in legitimacy and reputation among stakeholders. Although both are concerned with the influence of certification on the stakeholders, signaling theory can be tested with accounting-based financial measures since the focus is more on the customers’ response to certification [19,21,22,23,24,25,33,35,36]. Legitimacy theory incorporates the investors’ response as well; thus, market-based financial performance measures such as Tobin’s Q and Marris ratio are appropriate [27,28,29,37,38].

The most common control variables are company-specific factors such as size (e.g., total assets, number of employees), age, leverage, research & development expenditure, total sales, ownership, board size, and board independence. Aligned with contingency theory industry, regional development, and country controls are added. While the effect of industry can be operationalized as dummy variables, a more nuanced approach to measure market-specific factors can be industry size and industry certification [2,21,25].

According to their research methodology, the papers can be grouped as comparison of means between certified and non-certified companies, regression analysis, panel data analysis, and generalized method of moments. For comparison of means between certified and non-certified companies, the Mann–Whitney non-parametric test is applied. Ref. [19] report significant performance differences between ISO 9001-certified companies and non-certified companies. Ref. [22] present similar results, whereas the results are more profound for manufacturing companies as opposed to service companies. Ref. [32] report significant performance differences between ISO 9001 and ISO 14001 certified companies and non-certified companies. Although with a cause–effect relationship limitation, this research stream supports that ISO-certified companies perform on average better compared to non-certified companies.

Utilizing the analysis of variance (ANOVA) technique renders mixed results, no significant effect [31], positive significant effect [23], and negative significant effect [39]. Studies conducting multiple regression techniques document no significant effect [37], positive significant effect [27], and negative significant effect [29]. Although the cause–effect relationship is statistically supported, the analysis is time invariant.

Panel studies can accommodate multi-dimensional data, cross-sectional data such as industry, and time series data such as financial performance involving measurements over time. However, there is no consensus on model selection. Random-effects panel data analysis renders positive effect [21,24] and no significant effect [36]. Fixed-effects panel data analysis documents positive effect [7,20,40] and a U-shaped relationship: costly in the short term and beneficial in the long term [30]. This result indicates that the time lag between certification and returns should be considered. The GMM technique attempts to capture the dynamic aspects and correct for the inherent endogeneity in the ISO certification and CFP relationship by incorporating the lag of the dependent variable into the analysis as a regressor [41]. The results are mixed: no significant effect [28], positive significant [2,33,34], and negative significant effect [38]. In Appendix A, the studies included in the literature review are summarized in terms of their data period, sample size, country, operationalization of dependent, independent, and control variables, applied methodology, and main findings.

2.3. Theoretical Background and Hypothesis Development

While institutional theory proposes that organizations adopt ISO standards to meet external pressures from stakeholders such as customers, regulators, and investors [42], signaling theory suggests that certification with a management standard may reduce information asymmetries between producers and consumers, especially about underlying supplier attributes, thereby generating a competitive advantage for certified companies [43]. Based on the organizational legitimacy perspective, certification will enhance stakeholder approval and support and improve access to funding [44]. The resource-based view (RBV) suggests that the adoption of ISO Certification can improve a company’s resource base by enhancing its focus on internal resources to identify assets, capabilities, and competencies that have the most potential to provide significant competitive advantage [40,45,46]. Dynamic capabilities theory advises that the adoption of ISO certifications can enhance a company’s dynamic capabilities, such as innovation and adaptability, which can lead to improved CFP over time [47].

H1: 

According to signaling theory, ISO certification leads to improved accounting-based financial performance.

H2: 

Based on the organizational legitimacy perspective, ISO certification leads to improved market-based financial performance.

H3: 

According to the resource-based view, ISO certification leads to improved operational performance.

H4: 

According to dynamic capabilities theory, ISO certification leads to improved R&D performance.

Although different underlying mechanisms all advocate that ISO certification affects CFP positively, the slack resources theory supports the recursive relationship. Due to the cost associated with certification and the uncertainty of their returns, it is expected that companies with slack resources are more likely to invest in certification [48,49].

H5: 

According to slack resources theory, companies with slack resources are more likely to adopt ISO certification.

Companies with certain characteristics such as size, ownership structure, and age are more likely to acquire ISO certification and become more profitable [50]. Two interrelated problems have to be addressed: a selection problem generated by the relationship between the unobserved company characteristics and ISO certification; and a simultaneity problem generated by the relationship between ISO Certification and CFP.

H6: 

Companies with certain characteristics such as size and age are more likely to acquire ISO certification.

Furthermore, the contingency theory recommends that the relationship between ISO certification and CFP depends on the organizational and regulatory context, such as the culture of the company, industry, level of economic development of the country, market environment, and the intensity of competition. CFP can improve in some contexts but may not have a significant impact in others [51].

H7: 

According to the contingency theory, subsectors influence the ISO certification–CFP relationship.

ISO 9001, ISO 14001, and ISO 27001 certifications are mostly associated with positive returns, whereas multiple certifications have more intricate dynamics. For example, while simultaneous integration strategy has a positive impact, sequential integration strategy negatively affects firm performance [32].

H8: 

Different ISO certifications have unique impacts on CFP.

Figure 2 depicts the theoretical framework for ISO certifications and CFP.

3. Estimation Framework

3.1. Data Set

In total, 211 manufacturing companies were listed on the BIST as of 2022. However, due the reasons such as mergers & acquisitions or bankruptcy, not all companies are active every year between 2010 and 2022. To ensure that the dataset is balanced, 65 production companies, that is, 30.8% of the dataset, were excluded from the dataset. Companies with data from 2010 to 2022 were examined every year, and a panel data set was created by compiling the information of 146 companies identified for 13 years. Common method bias is a serious and problematic issue that can jeopardize the validity of research results. It creates a correlation between variables and can lead to an incorrect estimation of the predicted relationship. To avoid common method bias, it would be an important solution to create dependent and independent variables from information from as many different sources as possible [52]. For this reason, various sources were used. Whether the companies have ISO certifications for the years 2010–2022 has been compiled through the Public Disclosure Platform—Kamuyu Aydınlatma Platformu (KAP)—by scanning publicly disclosed reports such as sustainability reports, corporate responsibility reports, annual reports, and corporate governance compliance reports on the official websites. Since some companies have multiple plants/sites certified, similar to previous research [53,54,55] the first adoption is considered. Company-specific data such as total assets, total long-term debt, number of employees, total equity, R&D expenses, Cost of Goods sold, net income, and total sales have been downloaded from Wharton Research Data Services (WRDS) COMPUSTAT. In addition, the two-digit Standard Industry Classification (SIC) codes, which indicate the core activity of the companies, were also downloaded from WRDS COMPUSTAT. Companies are grouped according to the sectors defined by KAP with the help of SIC codes. Annual market value is obtained from the Central Securities Depository & Trade Repository of Türkiye Data Services.

Table 2. Sectoral breakdown of the sample.

Sectors	SIC Codes	Supply Chain Position			n = 146
		OEM	Tier 1	Tier 2	Total	%
Food and Kindred Products	20, 21	21	1	4	26	17.81%
Textile Mill Products, Apparel and other Finished Products, Leather and Leather Products	22, 23, 31	9	4	8	21	14.38%
Lumber and Wood Products, Furniture and Fixtures	24, 25	3	0	0	3	2.05%
Paper Allied Products, Printing, Publishing, and Allied Industries	26, 27	4	4	3	11	7.53%
Chemicals and Allied Products, Petroleum Refining and Related Industries, Rubber and Miscellaneous Plastic Products	28, 29, 30	14	2	8	24	16.44%
Stone, Slay, Glass and Concrete Products	32	4	4	5	13	8.90%
Primary Metal Industries	33	0	3	7	10	6.85%
Fabricated Metal Products, Industrial and Commercial Machinery, Electronic and other Electrical Equipment, Transportation Equipment	34, 35, 36, 37, 38	17	12	6	35	23.97%
Miscellaneous Manufacturing Industries	39	2	1	0	3	2.05%
Total		74	31	41	146	100%

shows the distribution of companies by sector. In addition, the supply chain tier of the companies, namely business-to-consumer or business-to-business trading status, was compiled through KAP and by scanning the annual reports on official websites. While 74 companies are original equipment manufacturers (OEM), 31 companies and 41 companies are first-tier and second-tier suppliers, respectively.

3.2. Variables

The independent variables ISO Certification are binary variables and indicate whether the company has certification in the year examined.

$$\mathrm{I}\mathrm{S}\mathrm{O} 9001=\left\{\begin{array}{l}1 if ISO 9001 certified\\ 0 otherwise\end{array}\right.$$

(1)

$$\mathrm{I}\mathrm{S}\mathrm{O} 14001=\left\{\begin{array}{l}1 if ISO 14001 certified\\ 0 otherwise\end{array}\right.$$

(2)

$$\mathrm{I}\mathrm{S}\mathrm{O} 45001=\left\{\begin{array}{c}1 if ISO 45001/OHSAS 18001 certified\\ 0 otherwise\end{array}\right.$$

(3)

$$\mathrm{I}\mathrm{S}\mathrm{O} 50001=\left\{\begin{array}{c}1 if ISO 50001 certified\\ 0 otherwise\end{array}\right.$$

(4)

$$\mathrm{I}\mathrm{S}\mathrm{O} 27001=\left\{\begin{array}{c}1 if ISO 27001 certified\\ 0 otherwise\end{array}\right.$$

(5)

$$Number of ISO certificates=\mathrm{I}\mathrm{S}\mathrm{O} 9001+\mathrm{I}\mathrm{S}\mathrm{O} 14001+\mathrm{I}\mathrm{S}\mathrm{O} 45001+\mathrm{I}\mathrm{S}\mathrm{O} 50001+\mathrm{I}\mathrm{S}\mathrm{O} 27001$$

(6)

As literature has focused its attention on ISO 9001, in Turkish practice the most common certification is ISO 9001, followed by ISO 14001 and ISO 45001/OHSAS 18001. All certifications show a positive trend. Figure 3 depicts the certification dynamics throughout the period 2010–2022. During the observed period, five companies have never been certified. Most of the companies renewed their certification once the certificate expired. Only four companies did not renew their certification. In total, 21 companies followed single certification strategy throughout the observation period: 20 of them are holding ISO 9001, and only one company ISO 14001. A total of 66 companies acquired single certification at first and gradually switched to multiple certification strategy. In total, 52 companies followed the multiple certification strategy, and 30 of them acquired all certifications at the end of 2022.

Figure 4 depicts percentage of certified companies by industry and year for all certifications. The most popular certification in Food and Kindred Products industry is ISO 9001. In total, 92% of businesses are ISO 9001 certified by 2022. ISO 9001 is closely followed by ISO 14001, but as of 2019, despite the slow start, the proportion of companies with ISO 45001/OHSAS 18001 has surpassed that of ISO 14001 companies. In the Textile Mill Products, Apparel and other Finished Products, Leather and Leather Products industry, ISO 90001, ISO 140001, and ISO 50001 are the most common certifications. In the Lumber and Wood Products, Furniture and Fixtures industry, as of 2016, every company in the sector has an ISO 45001/OHSAS 18001 certificate; by 2020, all companies have an ISO 14001 certificate. The third widely used certification in the sector is ISO 9001. In Paper Allied Products, Printing, Publishing, and Allied Industries, ISO 9001 is the most widely used certification. However, by 2022, 82% of firms have ISO 9001 certification. ISO 9001 is closely followed by ISO 14001, but as of 2022, the ratio of companies with ISO 45001 has exceeded the ratio of those with ISO 14001. The most popular certifications in the Chemicals and Allied Products, Petroleum Refining and Related Industries, Rubber and Miscellaneous Plastic Products industry are ISO 14001 and ISO 9001. By 2022, 79% of the companies have ISO 9001 and ISO 140001, 67% have ISO 45001/OHSAS 18001, 54% have ISO 27001, and 42% have ISO 50001 certification. By 2022, 77% of companies in the Stone, Slay, Glass and Concrete Products industry are ISO 9001 and ISO 140001, 69% of the companies ISO 45001/OHSAS 18001 and ISO 50001, and 46% of the companies ISO 27001 certified. In the Primary Metal Industries, ISO 14001, ISO 45001/OHSAS 18001, and ISO 9001 certificates are the most widely used certifications. As of 2019, all the companies hold ISO 14001 and ISO 45001 certifications, with 90% of the companies having obtained ISO 9001 and ISO 50001 certifications by 2022. Moreover, one of the two businesses in the sector acquired ISO 27001 certification by 2022. The most used certifications for Fabricated Metal Products, Industrial and Commercial Machinery, Electronic and other Electrical Equipment, Transportation Equipment industry are ISO 14001 and ISO 9001, with ISO 45001/OHSAS 18001 gaining traction in 2022. With 60% of the companies holding ISO 27001 certification in 2022, this industry is the most widely ISO 27001 certification-used industry. Lastly, in the Miscellaneous Manufacturing Industries in 2019, two out of three companies in the industry had ISO 140001 and ISO 45001/OHSAS 18001 certifications, whereas in 2022, all companies have ISO 140001, ISO 45001/OHSAS 18001, and ISO 9001 certifications.

Since the aim of the study is to reveal the effect of the ISO certification on the CFP and to understand the underlying mechanisms, different dependent variables were employed. To test dynamic capabilities theory R&D intensity, resource-based view GOGS/Revenue, signaling theory accounting-based measure ROA, and organizational legitimacy perspective market-based measure Tobin’s Q were determined as the dependent variables. R&D intensity is calculated by dividing the company’s R&D expenses by its revenue. COGS/Revenue is calculated by dividing the company’s Cost of Goods Sold by its revenue. The smaller the value it takes, the better the operational efficiency. ROA is calculated by dividing the company’s net income by its total assets. Tobin’s Q is calculated as market value divided by total assets.

$$R\&D intensity={\displaystyle \frac{R\&D Expenses}{Reveue}}$$

(7)

$${\displaystyle \frac{COGS}{Revenue}}={\displaystyle \frac{Cost of Goods Sold}{ Revenue}}$$

(8)

$$ROA={\displaystyle \frac{Net income}{Total Assets}}$$

(9)

$$Tobin’s Q={\displaystyle \frac{Market value}{Total Assets}}$$

(10)

When choosing control variables, consideration should be given to underidentification resulting from omitted control variables and overidentification resulting from too many control variables. Control variables were selected as parsimoniously and comprehensively as possible and were collected in two main groups. Company-specific factors and market factors that may affect the decision to obtain ISO certifications and CFP are considered. Company size is a factor that can affect CFP, and larger companies are more likely to obtain ISO certifications because their stakeholders and regulatory bodies expect them to obtain certifications and they have the financial strength to invest in certifications. Company size is measured with total assets and number of employees. Company age is a factor that can affect ISO certification and CFP. Long-established companies might be more likely to obtain ISO certifications, since they have accumulated slack resources to invest in certification over the years and established a corporate culture which fosters the obtainment of certification. Company age is calculated as the difference between year examined and founding year of the company. Company age is included in the analysis using the natural logarithm, since age is skewed to the right. The leverage shows how much of a company’s assets are financed by debt. Leverage is calculated as the ratio of the company’s total debt to its total assets. By adding sub-sector controls to the analysis, it is aimed to prevent the comparison of companies in different sectors which might have different environmental regulations. The control variable supply chain position is intended to capture the different dynamics of business-to-consumer or business-to-business trading. In addition to company characteristics, industry is controlled by dummy variables.

The average ROA is 0.04684%, which indicates that financially good companies are not overrepresented in the sample. The companies in the final sample have on average 2463 employees. By 2022, the average age of the companies in the sample is 50.26 years. On average, 70.84% of the companies in a sub-industry are ISO 9001 certified, 53.74% of the companies in a sub-industry are ISO 14001 certified, and 42.31% of the companies in a sub-industry are ISO 45001/OHSAS 18001 certified.

Table 3. Descriptive statistics.

	Observations	Mean	Std. Dev.	Min	Max	Skewness	Kurtosis
Total Assets (Million TL)	1897	3079.635	11,608.39	1.134	174,893.6	0.0000	0.0000
Stakeholder Equity (Million TL)	1898	1344.398	5595.906	−390.7246	118,662.2	0.0000	0.0000
Revenue (Million TL)	1894	2837.542	14,836.77	0	481,764.7	0.0000	0.0000
Net Income (Million TL)	1894	256.8924	1478.861	−2422.925	41,260.57	0.0000	0.0000
Cost of Goods Sold (GOGS) (Million TL)	1872	2239.435	12,530.75	0	418,987.7	0.0000	0.0000
Marketvalue (Million TL)	1809	2799.828	11,507.43	2.368651	248,999.6	0.0000	0.0000
Number of Employees (Thousand)	1882	2.463083	30.86817	0.001	1330	0.0000	0.0000
Long Term Debt (Million TL)	1898	439.3414	1900.053	0	28,324.38	0.0000	0.0000
R&D Expenditure (Million TL)	1888	14.93951	81.91762	0	1532.88	0.0000	0.0000
Age by 2022	146	50.26	15.23	12	95	0.7945	0.2510
R&D Expenditure/revenue	1873	0.004794	0.018231	0	0.303121	0.0000	0.0000
GOGS/Revenue	1855	0.7782437	0.2473318	0	4.509441	0.0000	0.0000
ROA	1893	0.04684	0.120191	−2.05094	1.57487	0.0000	0.0000
ROE	1894	0.045196	1.4451	−48.9169	24.1917	0.0000	0.0000
Tobin’s Q	1782	1.293835	2.710669	0.0574678	61.20941	0.0000	0.0000
Leverage (Long Term Dept/Total Asset)	1897	0.098771	0.140826	0	3.103579	0.0000	0.0000
Age (ln age)	1898	3.731494	0.4536431	0	4.564348	0.0000	0.0000

presents descriptive statistics on the dataset.

3.3. Model

The ISO certification and CFP relationship is dynamic in nature. The Generalized-Method-of-Moments (GMM) panel estimation dynamic relationships are characterized by the presence of a lagged dependent variable as regressor. Since ISO Certification and CFP are not strictly exogenous but dependent on their own past observations, the dependent variables with a one-year lag also appear in the model, so that it is possible to understand the performance based on the performance of the previous period. The certifications are renewed at most every three years, so the effect of ISO certification on CFP is analyzed for t − 1 and controlled for t − 2 and t − 3 [49]. The GMM model method allows controlling the unobserved firm-specific effects correlated with the regressors, while controlling heteroskedasticity and autocorrelation within firms. Moreover, no specific distribution is assumed for its estimation [41].

Model 1 is represented as.

$$\begin{gathered} {CFP}_{it}={\beta }_0{CFP}_{it-i}+{\beta }_1{number of certifications}_{it-1}+{\beta }_2{number of certifications}_{it-2}+ \\ {\beta }_3{number of ISO certifications}_{it-3}+{\beta }_4{size}_{it}+{\beta }_5{leverage}_{it}+{\beta }_6{age}_{it}+{\beta }_7{industry}_{it}+{\beta }_8{tier}_{it}+{\mu }_t+v_{it} \end{gathered}$$

(11)

$$v_{it}={\epsilon }_i+{\sigma }_{it}$$

(12)

where I = 1, …, n represents the companies, t = 1, …, T represents the time periods, CFP represents the dependents variables R&D intensity, COGS/Revenue, ROA, and Tobin’s Q, ${\beta }_0$, …, ${\beta }_k$ are the coefficients of the variables to be estimated, the error term is composed of a random element $v_{it}$, which can vary between companies and time periods, and the individual effect ${\mu }_t$ specific to each company and invariant over time.

Model 2 intends to measure the effect of different certifications on CFP.

$$\begin{gathered} {CFP}_{it}={\beta }_0{CFP}_{it-i}+{\beta }_1{ISO9001}_{it-1}+{\beta }_2{ISO9001}_{it-2}+{\beta }_3{ISO9001}_{it-3}+{\beta }_4{ISO14001}_{it-1}+{\beta }_5{ISO14001}_{it-2}+ \\ {\beta }_6{ISO14001}_{it-3}+{\beta }_7{ISO45001}_{it-1}+{\beta }_8{ISO45001}_{it-2}+{\beta }_9{ISO45001}_{it-3}+{\beta }_{10}{ISO50001}_{it-1}+ \\ {\beta }_{11}{ISO50001}_{it-2}+{\beta }_{12}{ISO50001}_{it-3}+{\beta }_{13}{ISO27001}_{it-1}+{\beta }_{14}{ISO27001}_{it-2}+{\beta }_{15}{ISO27001}_{it-3}+{\beta }_{16}{size}_{it}+ \\ {\beta }_{17}{leverage}_{it}+{\beta }_{18}{age}_{it}+{\beta }_{19}{industry}_{it}+{\beta }_{20}{tier}_{it}+{\mu }_t+v_{it} \end{gathered}$$

(13)

$$v_{it}={\epsilon }_i+{\sigma }_{it}$$

(14)

where I = 1, …, n represents the companies, t = 1, …, T represents the time periods, CFP represents the dependents variables R&D intensity, COGS/Revenue, ROA, and Tobin’s Q, ${\beta }_0$, …, ${\beta }_k$ are the coefficients of the variables to be estimated, the error term is composed of a random element $v_{it}$, which can vary between companies and time periods, and the individual effect ${\mu }_t$ specific to each company and invariant over time.

4. Results and Discussion

4.1. Impact of Number of ISO Certifications on CFP

The results of the two-step system-GMM estimators are summarized in

Table 4. Impact of number of ISO Certificates on CFP.

	Model 1.1 (R&D Intensity)				Model 1.2 (GOGS/Revenue)
	Coefficient	Robust Std. Err.	Z	p > |z|	Coefficient	Robust Std. Err.	z	p >|z|
dependent variable.lag	0.9164655	0.0913792	10.03	0.000	−1.013724	3.224766	−0.31	0.753
Number of certificates	−0.0000871	0.0001377	−0.63	0.527	−0.0712818	0.1621543	−0.44	0.660
Leverage	0.0004668	0.0006543	0.71	0.476	−0.3885818	1.118042	−0.35	0.728
Number of employees	6.27 × 10−6	0.0000247	0.25	0.799	−0.0090246	0.0228205	−0.40	0.693
Lnage	0.0000766	0.0002531	0.30	0.762	0.0361254	0.2793436	0.13	0.897
tier_1	−0.0000915	0.000503	−0.18	0.856	−0.3283477	0.6719115	−0.49	0.625
tier_2	−0.0001806	0.0004711	−0.38	0.701	−0.2583642	0.6512045	−0.40	0.692
_cons	0.000084	0.0010413	0.08	0.936	1.523519	2.694314	0.57	0.572
Number of Observations	1554				1556
Hansen J statistic	0.886, Chi-sq(1) p-val = 0.6421				0.793, Chi-sq(1) p-val = 0.6728
AR (1) p value	z = −1.09 Pr > z = 0.2752				z = −0.91 Pr > z = 0.3611
AR (2) p value	z = −1.43 Pr > z = 0.1516				z = −0.91 Pr > z = 0.3611
	Model 1.3 (ROA)				Model 1.4 (Tobins’ Q)
	Coefficient	Robust Std. Err.	z	p > |z|	Coefficient	Robust Std. Err.	z	p >|z|
dependent variable.lag	0.5520088	0.0817362	6.75	0.000	0.6076748	0.0922316	6.59	0.000
Number of certificates	0.0115833	0.0023409	4.95	0.000	0.1086216	0.0401953	2.70	0.007
Leverage	−0.1118339	0.0317571	−3.52	0.000	−1.316951	0.6421696	−2.05	0.040
Number of employees	−0.0012794	0.0003529	−3.63	0.000	−0.0426905	0.0096442	−4.43	0.000
Lnage	0.0144103	0.0065979	2.18	0.029	0.4155525	0.1869997	2.22	0.026
tier_1	0.0181414	0.0072843	2.49	0.013	0.2380513	0.1456362	1.63	0.102
tier_2	0.0074475	0.0065956	1.13	0.259	0.1885341	0.111777	1.69	0.092
_cons	−0.0411942	0.0292477	−1.41	0.159	−0.9932993	0.6480631	−1.53	0.125
Number of Observations	1578				1504
Hansen J statistic	4.681, Chi-sq(1) p-val = 0.0963				2.083, Chi-sq(1) p-val = 0.3530
AR (1) p value	z = −1.40 Pr > z = 0.1629				z = −2.29 Pr > z = 0.0221
AR (2) p value	z = 1.81 Pr > z = 0.0706				z = 1.77 Pr > z = 0.0766

industry effects are included in all estimations.

. All models contain control variables firm size (employee), leverage, age (ln_age), industry, and supply chain position. Model 1.1, Model 1.2, Model 1.2, and Model 1.4 have R&D intensity, COGS/Revenue, ROA, and Tobin’s Q as dependent variables, respectively. First, the validity of the models is assessed, followed by the estimated coefficients. If the Hansen J statistic for overidentifying restrictions is non-significant, the validity of the instruments exogeneity assumption is accepted [56]. To examine the validity of a strong exogeneity assumption, estimates autocorrelation under the null hypothesis that the error terms of two different periods are uncorrelated. If AR (2) is non-significant at the 0.05 level, there is no evidence of serial first-order correlation and no evidence to invalidate instruments through autocorrelation. The AR (1) is significant by construction, so its significance is uninformative in the validity assessment of the GMM model. Thus, if AR (1) is non-significant, the null hypothesis of no first-order serial correlation in first differences AR (1) test should not be rejected, and if AR (2) is non-significant, the null hypothesis of no higher-order serial correlation in first differences AR (2) should be rejected [57].

The dynamic models for each CFP indicator R&D intensity, COGS/Revenue, ROA, and Tobin’s Q are valid according to the assumptions of the two-step system-GMM estimators. Finally, there are no major concerns regarding the instrument count, which is considerably smaller compared to the sample size in all cases.

In all model specifications except Model 1.2, the lag of the dependent variable is significant and positive, validating the endogeneity assumption and confirming that applying the GMM model is indeed appropriate. The number of ISO certificates does not affect the R&D intensity and COGS/Revenue, whereas the number of ISO certificates has a positive and significant effect on ROA and Tobin’s Q. In Models 1.1 and 1.2 the control variables are insignificant. In Models 1.3. and 1.4, leverage and number of employees have a negative and significant effect on CFP, whereas firm age has a positive significant effect.

The empirical literature primarily uses R&D intensity as a control variable. While [30] finds a negative association between R&D intensity and CFP, [38] documents no significant effect on environmental performance and a positive effect on CFP. Ref. [48] use R&D intensity as an explanatory variable and show that it has a favorable impact on ISO 9000 certification. Since the dependent variable in this study is R&D intensity, the insignificant results are not very comparable with the literature.

Using pure operational performance measures [20,34] and operational performance and accounting-based financial measures [19,22,35] the majority of studies find a positive relationship between certification and performance, whereas [25] documents no significant relationship. According to [25] the effect of certification on companies’ CFP is insignificant because their operating costs burden increases while their marketing efficiency and managerial efficiency improve.

Using accounting-based CFP measures, the literature confirms the positive effect of ISO 9001 Certification in Türkiye [21] and ISO 14001 in Malaysia and Columbia [23,24] except in Indonesia [36]. Another study conducted in Indonesia had similar negative results regarding market-based measures [37]. Thus, this contradictory observation can be related to country effects. The positive and significant result between ISO Certification and ROA in Model 1.3 is consistent with the literature. Since both increased sales and decreased expenses can have an impact on ROA, signaling theory is not the only explanation for the ROA results. Combining results from Model 1.2 and Model 1.3, the signaling theory dynamic is more plausible, even though the RBV dynamic is within probability.

The results with market-based measures do not converge. The findings range from negative significant effect [29,38], inconclusive results [28], no direct relationship but mediating effect between environmental management and environmental performance [37], and positive significant effect [7,27]. The positive and significant result between ISO Certification and Tobin’s Q in Model 1.4 indicates that ISO certification improves stakeholders’ support and funding.

On the one hand, the result of R&D intensity and COGS/Revenue do not support dynamic capabilities theory and resource-based view, respectively. On the other hand, results regarding ROA and Tobin’s Q support signaling theory and organizational legitimacy perspective. Establishing the positive and significant relationship between ISO certification and ROA and between ISO certification and Tobin’s Q, to understand the influence of different ISO certifications, further investigation is conducted on ROA and Tobin’s Q.

4.2. Impact of Different ISO Certifications on CFP

In order to test “H8: Different ISO certifications have unique impacts on CFP.” the binary variables (ISO 9001, ISO 14001, ISO 45001, ISO 50001, ISO 27001) are utilized as explanatory variables, and ROA and Tobin’s Q as dependent variables in

Table 5. Impact of different ISO certifications on ROA.

	Model 2.1				Model 2.2
	Coefficient	Robust Std. Err.	Z	p > |z|	Coefficient	Robust Std. Err.	Z	p > |z|
roa.lag	0.5607376	0.0809877	6.92	0.000	0.5614217	0.0815476	6.88	0.000
ISO 9001	0.0154015	0.0062461	2.47	0.014	0.0153144	0.006311	2.43	0.015
ISO 14001	0.0133234	0.0082635	1.61	0.107	0.0161586	0.0083233	1.94	0.052
ISO 45001	0.0076847	0.007421	1.04	0.300
ISO 50001	0.0100436	0.0070965	1.42	0.157	0.0118694	0.0067144	1.77	0.077
ISO 27001	0.0113846	0.0074897	1.52	0.129	−0.1099235	0.0313568	−3.51	0.000
Leverage	−0.111983	0.0308079	−3.63	0.000	−0.0012523	0.000356	−3.52	0.000
Number of employees	−0.0012635	0.0003541	−3.57	0.000	−0.0012523	0.0003561	−3.52	0.000
lnage	0.0144944	0.00645	2.25	0.025	0.0147369	0.006543	2.25	0.024
tier_1	0.0170392	0.0071405	2.39	0.017	0.0165838	0.0073803	2.25	0.025
tier_2	0.0062727	0.0068192	0.92	0.358	0.0067138	0.0068101	0.99	0.324
Constant	−0.0395708	0.0287254	−1.38	0.168	−0.0386828	0.0290922	−1.33	0.184
industry	industry effects are included in the estimations				industry effects are included in the estimations
Number of Observations	1437				1437
Hansen J statistic	9.796, Chi-sq(10) p-val = 0.4585				9.652, Chi-sq(8) p-val = 0.2903
AR (1) p value	z = −1.59 Pr > z = 0.1123				z = −1.65 Pr > z = 0.0983
AR (2) p value	z = 1.74 Pr > z = 0.0815				z = 1.74 Pr > z = 0.0824

and

Table 6. Impact of different ISO certifications on Tobin’s Q.

	Model 2.3				Model 2.4
	Coefficient	Robust Std. Err.	Z	p > |z|	Coefficient	Robust Std. Err.	z	p > |z|
TobinsQ.lag	0.5652461	0.0853422	6.62	0.000	0.536038	0.0888145	6.04	0.000
ISO 9001	0.1822074	0.0891764	2.04	0.041	0.1877296	0.0888575	2.11	0.035
ISO 14001	−0.0800243	0.1412061	−0.57	0.571	−0.1542967	0.144231	−1.07	0.285
ISO 45001	0.0856484	0.1369664	0.63	0.532
ISO 50001	−0.1126909	0.0987593	−1.14	0.254	−0.1234731	0.1009166	−1.22	0.221
ISO 27001	0.1606696	0.1117754	1.44	0.151	0.198241	0.1080534	1.83	0.067
Leverage	−1.222295	0.5838707	−2.09	0.036	−1.32432	0.5872056	−2.26	0.024
Number of employees	−0.0260502	0.0059737	−4.36	0.000	−0.0305983	0.0068743	−4.45	0.000
lnage	0.34378	0.1348504	2.55	0.011	0.6563635	0.2835178	2.32	0.021
tier_1	0.2223122	0.1370581	1.62	0.105	0.2295241	0.1379038	1.66	0.096
tier_2	0.2463476	0.1146058	2.15	0.032	0.282484	0.115056	2.46	0.014
Constant	−0.6307342	0.4730502	−1.33	0.182	−1.624504	0.942284	−1.72	0.085
Industry	industry effects are included in the estimations				industry effects are included in the estimations
Number of Observations	1378				1378
Hansen J statistic	14.958, Chi-sq(10) p-val = 0.1336				13.484, Chi-sq(8) p-val =0.0963
AR (1) p value	z = −1.06 Pr > z = 0.2898				z = −0.70 Pr > z = 0.4833
AR (2) p value	z = 1.79 Pr > z = 0.0733				z = 1.85 Pr > z = 0.0646

, respectively. Model 2.1 and Model 2.3 contain all control variables and the full set of ISO certifications. Model 2.2 and Model 2.4 contain all control variables and a set of ISO certifications excluding ISO 45001 Certification.

First, the validity of the models is assessed, followed by the estimated coefficients. If the Hansen J statistic for overidentifying restrictions is non-significant, the validity of the instruments’ exogeneity assumption is accepted [56]. To test for autocorrelation aside from the fixed effects, the Arellano–Bond test is applied to the second-order correlation. If AR (2) is non-significant at 0.05 level, there is no evidence to invalidate instruments through autocorrelation, and no evidence of serial first-order correlation. The AR (1) is significant by construction, so its significance is uninformative in the validity assessment of the GMM model [57]. The dynamic models for both ROA and Tobin’s Q are valid according to the assumptions of the two-step system-GMM estimators. Finally, there are no major concerns regarding the instrument count, which is considerably smaller compared to the sample size in all cases.

In all model specifications lag of the dependent variable is significant and positive, validating the endogeneity assumption and confirming that applying the GMM model is indeed appropriate. The impact of ISO 9001 certification on ROA and Tobin’s Q is significant and positive in all model specifications. Although ISO 45001 is the 3rd most widespread standard, ISO 45001 has no significant effect on either ROA or Tobin’s Q. Thus, it is the first standard to eliminate from the model constellation.

In Models 2.1 and 2.3. the coefficient regarding ISO 14001 is insignificant. With a leaner constellation in Model 2.2, the positive and moderately significant coefficient (p > |z|= 0.052) indicates that ISO 14001 has positive influence on ROA, while the insignificant coefficient (p > |z| = 0.285) in Model 2.4 indicates that ISO14001 has no influence on Tobin’s Q. Thus, ISO 14001 certification has moderately positive and significant effect on ROA as seen in Model 2.2, whereas Tobin’s Q is not affected by ISO 14001 certification as seen in Model 2.4

The effect of ISO 50001 certification on ROA is insignificant in Model 2.1, however with a leaner constellation in Model 2.2 it becomes positive and moderately significant. In Model 2.3 and Model 2.4, the effect of ISO 50001 certification on Tobin’s Q is insignificant. The effect of ISO 27001 certification on ROA and Tobin’s Q is positive and moderately significant, with a leaner constellation in Model 2.1 it becomes significant at the 0.05 level.

The results of control variable leverage (ratio of long-term debt to total assets of the company) in all model specifications are consistent with finance literature. As expected, higher leverage indicates a negative relationship with ROA and Tobin’s Q. The significant and negative results for the number of employees in all specifications indicate that companies with a large workforce are less efficient than their counterparts. ROA and Tobin’s Q are affected by the supply chain position. ROA is significantly higher for tier 1 than OEM, whereas there is no significant difference between OEM and tier 2. Tobin’s Q is significantly higher for tier 2 than OEM, whereas there is no significant difference between OEM and tier 1.

While ISO 45001/OHSAS 18001 certification has no effect on ROA or Tobin’s Q, ISO 9001, ISO 14001, ISO 5001, and ISO 27001 certification have a positive impact on ROA, while ISO 9001 and ISO 27001 have a positive impact on Tobin’s Q. These results are in line with a positive significant effect of ISO 9001, and ISO 140001 when combined with ISO 9001, and no significant effect of ISO 45001/OHSAS 18001 on ROA documented by [11]. The positive and significant results regarding ROA in Model 2.1 and Model 2.2 support signaling theory. Companies with ISO 9001, ISO 14001, ISO 50001, and ISO 27001 generate more ROA compared to non-certified companies. Based on the organizational legitimacy perspective, certification enhances stakeholder approval and support and improves access to funding. The positive and significant results regarding Tobin’s Q in Model 2.3 and Model 2.4 results support the organizational legitimacy perspective and indicate that investors respond to ISO 9001 and ISO 27001 certification. While consumers respond to economic, environmental, and governance dimensions, investors value economic and governance dimensions.

4.3. Robustness

To evaluate whether the findings are robust, the amplitude of the coefficients and their significance should be checked with another model. Due to the inherent endogeneity of the ISO Certification–CFP relationship, the error term could potentially exhibit heteroskedasticity, autocorrelation, and cross-sectional dependency across companies. To address these issues, the Driscoll and Kraay estimator, which generates standard errors resilient to heteroskedasticity and autocorrelation while considering the presence of cross-sectional dependence [58,59], is applied. The estimations in Model 1.1 to Model 1.4 in Table 4, and Model 2.1 in Table 5 and 2.3 in

Table 7. Findings with fixed-effect Driscoll–Kraay regression.

	Model 1.1 (R&D Intensity)	Model 1.2 (GOGS/Revenue)	Model 1.3 (ROA)	Model 1.4 (Tobins’ Q)	Model 2.1 (ROA)	Model 2.3 (Tobins’ Q)
Number of certificates	−0.0000673	−0.3389804	0.0099449 ***	0.0634517
ISO 9001					0.0090738 *	0.226618 **
ISO 14001					0.0150345 ***	−0.1237565
ISO 45001					0.0048694	0.3384095
ISO 50001					0.0061371 *	−0.2619176
ISO 27001					0.0166118 ***	0.1724123
dependent variable.lag	0.8505042 ***	−6.408547	0.5695793 ***	0.5929838 ***	0.569488 ***	0.5865831 ***
Leverage	0.0001237	−1.703022	−0.105576 ***	−1.405957 ***	−0.1099927 ***	−1.352853 ***
Number of employees	0.0000388	−0.0434484	−0.001071	−0.0405156 ***	−0.0009245	−0.0339429 ***
lnage	−0.0000135	0.2097061	0.0143774 **	0.690438 *	0.0144684 **	0.6287225 *
tier_1	0.0003776	−1.491006	0.0119395 *	0.297591 **	0.0131486 **	0.2892051 **
tier_2	−0.0001857	−1.334499	0.0037172	0.2427555 **	0.0052754 **	0.2744535 ***
Constant	0.0004785	5.536562	−0.0348519	−1.846177	−0.032749	−1.70171
industry	industry effects are included in all estimations
Number of Observations	1559	1561	1583	1509	1583	1509

*** p < 0.01, ** p < 0.05, * p < 0.1.

are replicated with the inclusion of the lagged dependent variable. The GMM model and Driscoll and Kraay regression estimate the same effect for the specifications Model 1.1, Model 1.2, Model 1.3, and Model 2.1 whereas Model 1.4 and Model 2.3 are partially validated. The effect of ISO Certification on R&D performance and operational performance remains insignificant. The results indicate a robust positive relationship between ISO certification and ROA, while the results with Tobin’s Q are partially validated. The findings support that different ISO certifications have unique impacts on CFP, namely all ISO Certifications except ISO 45001/OHSAS 18001 have positive effects on ROA. The results for Tobin’s Q replicate the positive effect of ISO 9001 but not ISO 27001.

The Driscoll and Kraay estimator is used for robustness checks since it is capable of generating standard errors resilient to heteroskedasticity and autocorrelation while considering the presence of cross-sectional dependence. In this estimation, the bidirectional causality is controlled by incorporating past performance as an explanatory variable, however, the influence of past ISO certifications on performance cannot be incorporated into the analysis. In GMM, both ISO certification and financial performance dynamics can be accounted for by incorporating a one-year lag of dependent variable as an explanatory variable and ISO certification with a one-year lag, two-year lag, and three-lag as instrumental variables into the analysis. Moreover, GMM generates standard errors resilient to heteroskedasticity and autocorrelation. GMM is clearly better suited to conduct analysis on hand compared to the Driscoll and Kraay estimator. Since there is a need for robustness checks, the second-best alternative approach was to employ the Driscoll and Kraay estimator. Even though the ISO certification dynamics cannot be incorporated, using the Driscoll and Kraay estimator supported most of the results; only the Tobin’s Q results are weaker. This occurrence might be caused also by the sample since data availability for Tobin’s Q was limited compared to R&D intensity, GOGS/revenue, and ROA.

4.4. Theoretical and Policy Implications

The results for COGS/Revenue and R&D intensity do not support RBV and dynamic capabilities theory; however, these theories should not be rejected right away. It is expected that ISO certifications have a positive effect on operational performance, e.g., ISO9001 decreases costs associated with rework and scrap and OHSAS 18001/ISO 45001 decreases the occurrence and severity of occupational accidents, which have a direct effect on CFP in terms of paid leave and compensation for accidents resulting in loss of limb or fatality. One plausible explanation is that cost increases in material, labor, and overheads may mask the savings due to improvements in operational performance through certification. Another explanation is that the massive benefits of signaling outweigh the benefits of process innovation. A similar observation in the emerging economy context is that ISO 9000 certification increases the total amount of sales, but it decreases the productivity and profitability of certified service firms [60]. In particular, the signaling effect is more prominent in emerging economies. Where institutions are weak, the certification acts as a substitute for the regulatory role [61].

It is expected that ISO certifications have a positive effect on R&D. On the one hand, ISO certifications might facilitate a culture of conformance improvement rather than performance improvement, which might lead to a negative effect on product innovation performance [62]. On the other hand, R&D efforts are composed of new technologies R&D, new product/ new model R&D, and ongoing R&D. While new technologies R&D and ongoing R&D show increasing or decreasing trends, new product/ new model R&D are cyclic with varying cycle lengths, e.g., new model launches every 5–10 years for the automotive industry and seasonal for the textile industry. More nuanced results can be achieved by incorporating sector-specific R&D dynamics.

Accounting-based measure ROA can be affected by increases in sales and decreases in costs. Thus, the results of ROA cannot only be attributed to signaling theory. However, the results indicate that, although the RBV dynamic is within probability, the signaling theory dynamic is most likely. Since the coefficient in Table 4 Model 1.2 (GOGS/revenue) is insignificant, the cost reduction route is ruled out and it is inferred that ISO certifications improve revenue, which indicates that the underlying mechanism is signaling theory. The revenue of the company increases since ISO certifications improve companies’ reputations and consumers prefer certified companies as opposed to non-certified companies.

The results of Tobin’s Q are evidence for the organizational legitimacy perspective. The significant and positive coefficient in Table 4 Model 1.4 indicates that ISO certifications have a positive impact on market-based performance measure Tobin’s Q. From this observation we can infer that ISO certifications improve companies’ reputation, and investors favor the certified companies compared to non-certified companies. Certified companies have better chances to access financial resources, and the organizational legitimacy perspective is supported.

ROA is positively affected by ISO 9001, ISO 14001, ISO 5001, and ISO 27001 certification, whereas Tobin’s Q is positively affected only by ISO 9001 and ISO 27001. While there is no effect of ISO 45001/OHSAS 18001 certification on ROA and Tobin’s Q, occupational health and safety certification should be viewed as a moral commitment by the company to the United Nations Global Compact and as a recognition of such commitment should be required by the regulatory bodies of Capital Markets Board of Türkiye- Sermaye Piyasası Kurulu (SPK) and Istanbul Stock Exchange as a prerequisite for listing on BIST. The theoretical findings of the study are summarized in

Table 8. Summary of theoretical findings.

Hypotheses	Findings
H1: According to signaling theory, ISO certification leads to improved accounting-based financial performance.	Supported
H2: Based on the organizational legitimacy perspective, ISO certification leads to improved market-based financial performance.	partially supported
H3: According to resource-based view, ISO certification leads to improved operational performance.	not supported
H4: According to dynamic capabilities theory, ISO certification leads to improved R&D performance.	not supported
H5: According to slack resources theory, companies with slack resources are more likely to adopt ISO certification.	Supported
H6: Companies with certain characteristics such as size and age are more likely to acquire ISO certification.	Supported
H7: According to the contingency theory, subsectors influence the ISO certification-CFP relationship.	not supported
H8: Different ISO certifications have unique impacts on CFP	Supported

.

5. Conclusions

The purpose of this paper was to investigate the impact of different ISO certifications on corporate financial performance based on audited financial statements of 148 manufacturing companies listed in BIST during 2010–2022. The theoretical foundation indicates that ISO certification and CFP have a reciprocal relationship. Thus, the application of the GMM model was essential in resolving the endogeneity issue resulting from the reverse causality between the CFP and ISO certification. This study shows that ISO certifications have a positive effect on ROA and Tobin’s Q, with no significant effect on COGS/Revenue and R&D intensity. Moreover, this study decomposes the effect of different ISO certifications. The positive and significant effect of ISO 9001, ISO14001, ISO5001, and ISO 27001 on ROA support signaling theory and indicate that consumers respond to economic, environmental, and governance dimensions. The results regarding Tobin’s Q indicate that investors respond to ISO 9001 and ISO 27001 certifications and support the organizational legitimacy perspective. While consumers respond to economic, environmental, and governance dimensions, investors value economic and governance dimensions.

This study identifies ISO certification as an effective management tool to signal to stakeholders the company’s commitment to quality, environment, health and safety, and governance. However, the expected process improvement and product innovation could not be documented. ISO certifications might not be the panacea to ESG issues as one has hoped for. This could be attributed to companies giving too much attention to obtaining the ISO certification but thereafter failing to maintain the required standards. Thus, companies should internalize ISO standards as part of a long-term continuous improvement philosophy, instead of just acquiring an ISO certificate. Nevertheless, this study aims to encourage manufacturing companies to adopt ISO certifications and improve their ESG practice, which is an important step towards UN Sustainable Development Goals 8 and 12.

This study builds on the notion that ISO certifications enhance ESG, but this positive influence can be different in each sector and even in each country. Since, the results are for search goods, generalizations for experience goods, and credence goods should be interpreted cautiously. The findings are limited to the Turkish manufacturing industry and its surrounding institutional context and should be generalized to emerging countries’ contexts with this consideration. Inferences on countries with similar GDP per capita and legislations can be made. Other company characteristics (e.g., board diversity, ownership structure, percentage of exports) and specific economic conditions can influence the capacity of companies to adopt ISO standards. Theories were tested separately; methods such as structural equation modeling, which make parametric assumptions about linkages among variables, can be applied and theoretical linkages can be estimated. Therefore, more comprehensive studies are recommended to overcome the shortcomings of the current analysis.