Operational Competitiveness and the Relationship between Corporate Environmental and Financial Performance

Abstract

With increasing pressures on big businesses to expand performance objectives beyond financial metrics and to include social and environmental objectives, business organizations experience rising tension in balancing these various objectives. Oftentimes, subjective narratives can weigh in on the relative importance of competing objectives. This subjectivity is a contributing factor to findings of inconsistent and mixed results for the financial impact of an organization’s environmental performance in the prior literature. Our research effort seeks to provide a positivist perspective on the relationship between environmental performance and financial performance of companies. Also, given the importance of efficient operations for corporate success, we examine the influence of operational productivity on the environmental and financial performance relationship. Using a global dataset compiled from reputable sources, including 1738 unique firms spanning between the years 2011 and 2020, we find statistically significant results that indicate that lower carbon emissions are associated with higher profitability when a firm has competitively high operational productivity. Companies with operational productivity that is competitively low do not perform well financially when carbon emissions are low. Thus, our study fills a research gap in this domain by relying exclusively on a broad set of purely objective data and illuminating the importance of operational efficiency on the relationship between the environmental performance and financial performance of firms.

1. Introduction

Companies and national governments are speaking to and taking action on environmental and social initiatives to explicitly address organizational performance in these areas. For example, a notorious corporate polluter, Exxon, has pledged that its oil and gas production operations would reduce the intensity of emissions by 15% to 20% by 2025 (Matthews 2020). As a way to hear from investors and also receive feedback on its efforts with respect to carbon emissions, Unilever has committed to poll its shareholders every three years via a vote on its efforts to mitigate its impacts on the natural environment (Chaudhuri 2020).

At a macro-economic level, Germany and other European governments envision legislation to require companies to screen and police suppliers to assure their supply chains are not violating environmental and social rights standards (Sorge and Boston 2021). Firms affected by the proposed supply chain law in Germany argue that they will be at a competitive disadvantage due to the costs of being held accountable for supply chain actions beyond their immediate operations (Jennen 2021). On the other hand, Unilever CEO Alan Jope observed recently that, oftentimes, companies experience short-term expenses from efforts to improve the sustainability of their sourcing but that, in the long run, these investment expenses enable long-term savings. For example, from 2008 to 2020, Unilever saved USD 1.5 billion via sourcing sustainably (Brown 2021). Of course, organizational communications might not provide an objective view or clear accounting of the economic costs and benefits of meeting organizational goals in the areas of environmental and social performance.

Corporate social responsibility (CSR) and its main tenants of environmental, social, and governance (ESG) factors are areas of business research that have blossomed in academic and practitioner interest over the past 50 years. Beyond the fundamental imperative of corporations to generate capital returns for shareholders, CSR puts the onus on corporations to carry the mantle of being mindful of their performance in a multitude of ESG measures.

Milton Friedman’s Capitalism and Freedom (Friedman 1962) articulates how CSR can be subversive to free markets by distracting from “the one and only social responsibility of business—to use its resources and engage in activities designed to increase its profits so long as it stays within the rules of the game, which is to say, engages in open and free competition without deception or fraud”. Put another way, by splintering the purpose of corporations to perform well regarding the triple bottom line of people, planet, and profits, the driving purpose of corporations to provide economic returns for shareholders is diluted.

Furthermore, Orlitzky (2015) argues that the arc of CSR research has evolved from a positivist perspective to ideological, collectivist, and political perspectives as a way to influence corporate performance by adopting stakeholder interests rather than shareholder interests. For example, Orlitzky (2015) notes that attributes provided in KLD Research and Analytics, Inc.’s widely used KLD data frame express specific political views on corporate characteristics, while other KLD qualities are “merely good management practices”.

Thus, we seek to use audited, verified, well-defined, and quantitative data on two of the three dimensions of the triple bottom line—planet and profit—to provide a positivist perspective on firm performance. Given regulatory reporting requirements for corporations with regard to pollutants, we specifically examine the relationship between greenhouse gas (GHG) emissions of organizations and a key metric for corporate economic performance, return on sales (ROS), via publicly disclosed secondary data. Also, due to the fundamental relationship of productive output of both GHG and ROS, our study examines the operational competitiveness of firms via the effect of operational productivity (OP) on the relationship between GHG emissions and ROS.

Since this study measures the impact of the absolute levels of GHGs released during plant level operations (a direct indicator of environmental performance of the company), the interplay between emissions and financial returns will be conveniently decipherable by all managerial levels of the organization and will inform efforts from the production line to senior-level decision makers. GHG is a standardized metric that is widely used to measure the environmental impact of organizations, especially by environmental policy-making institutions like the United Nations Framework Convention on Climate Change (UNFCCC). Being a standard metric, it allows consistent measurement, which allows comparison across different organizations. Greenhouse gases are emitted from various types of organizational activity like manufacturing and transportation, making it a comprehensive indicator of overall impact of an organization on the environment. Furthermore, measuring the moderating role of OP involves the influence of the competitiveness of operations practices on the degree to which an organization’s financial performance is tied to its environmental performance. The inclusion of OP in the analysis provides deeper insights into how an organization’s operational competitiveness can work in concert to enhance or detract from the relationship between GHG and ROS.

Our research approach begins with designing a base model for the independent variable (GHG emission level) and the dependent variable (ROA). This basic model is extended by introducing moderating effects of OP on the relationship between GHG and ROS. To test our model, GHG emission data for the period 2011 to 2020 published by the Carbon Disclosure Project (CDP) are collected for this study. We include control variables of firm size, financial leverage, R&D intensity, SG&A intensity, and industry competitiveness. All the financial data required to calculate the variables used in this study are obtained from Compustat North America and Compustat Global, resulting in a global dataset.

Our research effort culminating in this paper is the result of a project we began in 2021 and a working paper presented in 2022 (Amarasuriya and Burke 2022); as such, there is some content, especially in the introduction and the literature review sections, in this paper which is repeated from the previous working paper.

In the next section, we review the extant literature on the key elements of our analysis. In Section 3, there are details of data collection and measuring variables and an explanation of the methodology employed to collect and analyze the data. Section 4 details the analysis of results. Robustness checks are documented in Section 5. In the final section of our study, we discuss the findings, conclusion, and related areas for future work.

2. Literature Review

The academic literature to-date finds that the relationship between the level of emissions reported by an organization and its financial performance is ambiguous. A lack of consistent results about the relationship between environmental and economic output resulting from operational efforts makes it difficult for organizations to gauge the feasibility of engaging in sustainable processes. In part, the literature to-date suffers from the fact that the measurement of the environmental performance of a company and its impact on the organization’s finances can be shrouded by subjective political ideologies (Orlitzky 2015). Given the ambiguous and inconsistent results from the past literature pertaining to the impact of a firm’s environmental performance on their financial performance, there is a gap in the literature regarding the examination of the nature of the relationship existing between an organization’s environmental performance and financial performance with purely objective data. Hence, we address this gap by providing a positivist perspective on firm performance using audited and verified quantitative data. In addition, given the importance of the efficient management of operations within a company, we also investigate the influence of operational productivity on the environmental and financial performance relationship.

Therefore, a data-driven empirical examination of the relationship between direct GHG emissions and ROA can inform responses to managerial questions with respect to sustainability, such as “Is it financially worthwhile for our company to invest in controlling GHG emissions?”.

2.1. Environmental and Financial Performance

Klassen and Vachon (2003) study how supply chain collaboration and evaluation can help organizations improve their environmental management. Using data gathered from a sample of Canadian plants, they find that collaborations initiated by both organization and customer have a significant impact on the form and level of investments in environmental technologies used in regulating environmental impact across different segments of the supply chain. They find only limited evidence to support the argument that evaluation activities influence investment in environmental management. Similar to our study, Klassen and Vachon (2003) consider organizational environmental performance as the independent variable in their study, but they use evaluative activities and legal compliance to conceptualize the construct by leveraging a survey-based research design. Our study also includes companies from more diverse geographies than Klassen and Vachon (2003).

Hart and Ahuja (1996) study the impact of a reduction of selected chemical pollutants on the financial performance of companies. They collect data from the Toxic Release Inventory program and use the percentage change in the emissions efficiency index (the ratio of reported emissions in pounds to the revenue of companies) as the independent variable. They measure financial performance using ROA, and include control variables such as R&D intensity, leverage, and advertising intensity. The results of the study suggest that the positive impact of efforts invested in pollution control on a firm’s financial performance starts to deteriorate with time after the initiation period. Similar to the study conducted by Hart and Ahuja (1996), we investigate the environmental, operational, and financial performance of organizations. But, different to our study, Hart and Ahuja (1996) operationalize the environmental performance as a reduction in emissions, and operational and financial performance are both dependent variables. Additionally, in our study, we examine the moderating role of competitively derived operational productivity.

Chen and Delmas (2011) conduct a study on measuring corporate social performance of companies, which incorporates measuring corporate environmental performance. They state that there is no universally agreed upon scale or prioritization of environmental performance indicators existing for different stakeholders in the context of varying situations due to the diversity and dynamism of stakeholder attributes such as stakeholder preferences and perceptions. Accordingly, it is noted that many studies utilize directly observable indicators like energy consumption, greenhouse gas emissions, and pollutant emissions for the measurement of a firm’s environmental performance. Using data envelopment analysis (DEA), Chen and Delmas (2011) leverage KLD measures to establish environmental performance scores. However, GHG emissions are not a dimension included in the scores. Similar to the DEA approach for gauging environmental performance, our study utilizes a stochastic frontier approach (SFA) to measure operational productivity. Another difference in our study is the reliance on Scope 1 emissions as the gauge for the environmental performance of firms.

Other recent research on operations and supply chain management contains studies of disclosures related to low carbon emissions and policy design and implementation (Song et al. 2023; Park and Ravenel 2013). Many emission control studies are conducted at the firm level, sometimes in the context of a buyer–supplier relationship, and some conduct a dyadic level analysis (e.g., Song et al. 2023). Our study contributes to the literature by isolating the direct emissions of a firm and the respective financial performance in light of the firm’s competitive productivity.

Awaysheh et al. (2020) compare best-in-class firms with worst-in-class firms to examine the CSR–financial performance link, considering industry differences, time trends, and clustering and the potential existence of a nonlinear relationship between engagement in CSR activities and the financial performance of firms. Using KLD data for seven dimensions, Awaysheh et al. (2020) show that best-in-class performers of CSR activities have high operational and financial performance levels, and question whether “doing good” (engaging in CSR) leads to “doing well” (firm financial performance), or is it the other way round, since high profits give managers more liberty to invest in CSR activities without shirking their responsibility towards the shareholders. Furthermore, they show that the top 10% of firms in an industry, in terms of CSR performance, have higher operating performance levels (measured via operating income before depreciation divided by total assets) and higher relative market valuations (measured via Tobin’s Q). The results suggest that more profitable firms can invest more in CSR without compromising the interests of the shareholders. Ultimately, the results from Awaysheh et al. (2020) indicate that top CSR performers are not “doing good” just because they are “doing well”.

A common operational measure of environmental performance of organizations is gauged by the level of toxic emissions of organizations, such as GHG, released by companies into the environment (Jacobs 2014; Hart and Ahuja 1996). Jacobs (2014) explores whether there is a change over time in the direction or magnitude of the effect of emission reductions by firms on their financial performance and market reactions to GHG emission reductions versus market reactions to non-GHG emission reductions. An event study methodology is employed to examine the effect of voluntary emission reduction announcements on stock market reaction. Jacobs (2014) finds that the market reaction to voluntary emission reductions decreases significantly over time and that, for GHG emission reductions, the market reaction is more positive than for non-GHG emission reductions. Jacobs (2014) considers the effect of voluntary announcements on GHG emission reduction by companies on their financial performance, which is operationalized via ROA. Our study focuses on how financial and environmental performance of publicly traded companies relate, especially with respect to operational dimensions within an organization’s control.

Researchers in the area of operational sustainability theorize via the resource-based view (RBV) that resources employed by a firm to control GHGs can constitute VRIN (valuable, rare, inimitable, non-substitutable) characteristics and lead to a competitive advantage for the firm (Jacobs 2014). Emission reductions in firms can result from labor intensive efforts such as total quality management practices like continuous improvement projects (Hart and Ahuja 1996). These practices can be embedded in organizational routines and depend on the knowledge of employees. Thus, resources used to control GHG emissions have aspects of inimitability. Russo and Fouts (1997) find that technologies used by a company to reduce GHG emissions will not be easily available to its competitors because they require specialized skills to operate. Hence, the resources associated with controlling GHG emissions can yield a competitive advantage to a company and generate returns in the form of increased profitability.

The transportation of raw materials to the manufacturing plant, conversion of raw materials to finished products, storage of raw materials, and work-in-progress and finished goods at the plant are some of the business processes that result in releasing GHGs to the atmosphere. These processes happen at the expense of organizational assets or resources like human capital, property, plant, and equipment. These productive processes are leveraged to create a value to the customer that will generate financial returns for the organization in terms of sales and profits.

Organizational resources involved in business processes that emit GHGs will be used to generate financial returns for the company, where they will also result in costs for the company in terms of energy inefficiencies. Thus, it is reasonable to model the relationship between GHG emissions and the financial returns of a company. Additionally, several variables like firm size, investment in R&D, advertising intensity, and financial leverage will have to be controlled to blunt the effect of differences in firm resources, commitments, and strategies (Hart and Ahuja 1996). Relationally, we expect that firms that generate higher GHG emissions (lower environmental performance) will be unable to reap beneficial financial performance. Therefore, we hypothesize the following:

H1. 

Low environmental performance will have a negative impact on firm financial performance.

2.2. Role of Operational Productivity

Operational productivity measured as the ratio of an output like sales of the firm to input resources like inventory, employees, property, plant, and equipment is found to elicit a condition for the relationship between the corporate social performance and the financial performance of companies. Productive companies have the necessary resources and technological base needed to enhance employee performance and leverage management practices that are needed to improve business performance (McKone et al. 2001). Jacobs et al. (2016) find that only the companies that excel in both OP and corporate social performance (the adoption of management practices to minimize the negative impact that firm operations have on society) can drive better financial performance than companies excelling in only one of them.

The reduction of the negative impact of operations on the environment via the efficient usage of natural resources and materials in manufacturing is studied by Rothenberg et al. (2001). More particularly, elements of lean production that encompass maintenance of minimal levels of inventory or buffer minimization, management practices, and policies that enable human resource management are examined in the previous literature for their role on improving the operational efficiency and environmental performance of the firm (Macduffie 1995; Rothenberg et al. 2001). Buffer minimization, which entails producing the maximum output using the minimum level of inventories, enables minimum end-of-process rework areas and work-in-process compared with plants without inventory efficiency. Employee efficiency, which is another component of operational productivity that we consider in our study, includes worker commitment, motivation, and skills, which are crucial for the operational success of the firm (Rothenberg et al. 2001).

These types of material efficiency and labor efficiency practices are deemed a part of pollution prevention technologies and management. In contrast to pollution prevention technologies, pollution control technologies improve the environmental performance of plants by minimizing the toxic emissions via treatment or disposal of harmful by-products or pollutants at the end of manufacturing processes (Rothenberg et al. 2001).

It has been found that, through material usage reduction combined with increased efficiency of plant equipment, manufacturing plants are able to reduce the emission of toxic volatile organic compounds (Aragón-Correa 1998). Firms with high environmental performance are found to combine pollution prevention technologies with pollution control technologies rather than depending solely on pollution prevention mechanisms. Evidence for this can be found in observing a 50% decrease in volatile organic compound emissions by introducing material usage reduction, where another 15% reduction in emissions is achieved by increasing the efficiency of the equipment used by an automobile assembly plant (Rothenberg et al. 2001).

According to the resource-based view of the firm, organizations gain competitive advantage by exploiting their internal strengths, which possess the VRIN characteristics of being valuable, rare, not imitable, and non-substitutable (Barney 1991). Firms with high productivity that generate more outputs from a unit of input compared with a firm with less productivity will likely possess VRIN resources, such as skillful employees who are capable of implementing activities related to environmental management and have support from organizational management in training and educating the workforce on green orientation of organizations, and organizational knowledge and technology (Paiva et al. 2008) needed to leverage GHG emission reduction practices effectively and efficiently.

These resources employed by efficient organizations will yield superior returns compared with less efficient organizations. Hence, if a firm with low OP engages in GHG emission reduction efforts, any financial return associated will be limited, since they do not utilize resources efficiently and effectively like firms with high OP do. Therefore, when firms are operationally productive, they leverage inputs, including those that aid in reducing GHG emission levels. This efficiency capability helps the firm to derive returns from engaging in GHG emission reduction practices (Jacobs et al. 2016). Based on this, we can expect firms with low OP to be less effective than high OP firms in terms of financial performance. Furthermore, high OP companies can gain financially while emitting lower levels of GHGs. Therefore, we hypothesize the following at the firm level:

H2. 

Higher operational competitiveness will be associated with higher firm financial performance.

H3. 

Higher operational competitiveness moderates the relationship between environmental performance and financial performance.

3. Data and Variables

The data to form the measures for this study are collected from two main sources. CDP (formally known as the Carbon Disclosure Project) is used to collect data on the independent variable (GHG Scope 1 emissions). All other financial data, including the measures used to calculate the dependent variable (ROS), control variables, and operational productivity (OP), are collected from Compustat North America and Compustat Global. The resulting sample of our study is 9414 firm year observations, with 1738 unique firms spanning between the years 2011 and 2020. Our sample includes multiple industries and multiple countries. A breakdown of our sample is found in

Table 1. Breakdown of samples by industry sector and region.

By 2-Digit NAICS Codes
		Firms	Firm-Yr
11	Agriculture, Forestry, Fishing, and Hunting	6	17
21	Mining, Quarrying, and Oil and Gas Extraction	107	520
22	Utilities	81	470
23	Construction	64	400
31	Manufacturing	159	863
32	Manufacturing	318	1859
33	Manufacturing	478	2570
42	Wholesale Trade	51	218
44	Retail Trade	61	318
45	Retail Trade	42	212
48	Transportation and Warehousing	75	379
49	Transportation and Warehousing	10	69
51	Information	97	513
52	Finance and Insurance	20	135
53	Real Estate and Rental and Leasing	12	46
54	Professional, Scientific, and Technical Services	73	398
56	Admin and Support and Waste Mgmt and Remediation Services	21	122
62	Health Care and Social Assistance	15	62
72	Accommodation and Food Services	33	174
99	Other	15	69
	Total	1738	9414
By Region
North America		443	2566
Europe		674	3816
Asia		427	2212
Other		194	820
Total		1738	9414

. For brevity, we do not show each individual country but rather the region in which the country is in (i.e., North America, Europe, Asia, and Other).

3.1. Dependent Variable

Following the prior literature, we capture financial performance (FP) as the return on sales. The return on sales is measured as the ratio between net income and total revenue.

3.2. Independent Variable

GHG emissions are categorized into three categories. Scope 1 refers to emissions from direct sources, e.g., the operations process. Scope 2 refers to emissions from indirect sources, e.g., using energy from the grid. Scope 3 refers to all other emissions outside of a firm’s boundary, e.g., business travel. In this study, we choose Scope 1 emissions as our primary independent variable because we posit that they are more aligned with OP than the other two measures. To allow for comparability between firms, we use Scope 1 emissions intensity, i.e., the ratio of Scope 1 emissions divided by total assets. Since GHG emissions are heterogeneous among different industries,1 we capture environmental performance (EP) as a normalized value based on the two-digit NAICS industry that a firm is in. It is important to note that this operationalization results in large values of EP representing worse environmental performance and vice versa.

3.3. Moderating Variable

Following prior studies (Li et al. 2021; Mishra et al. 2020; Yiu et al. 2020), we employ a stochastic frontier approach (SFA) to calculate a firm’s productive inefficiency. Operational productivity, OP, is then measured by the inverse of the firm’s productive inefficiency. We use operating income, capital expenditure, number of employees, and inventory for the SFA estimation, as shown below.

$$\ln \left(O{I}_{ijt}\right)={\beta }_o+{\beta }_1\ln \left(CAPE{X}_{ijt}\right)+{\beta }_2\ln \left(EM{P}_{ijt}\right)+{\beta }_3\ln \left(IN{V}_{ijt}\right)+V_{ijt}-U_{ijt}$$

(1)

$$O{E}_{ijt}=\exp \left(-U_{ijt}\right),$$

(2)

where $O{I}_{ijt}$ represents firm $i$’s operating income in a two-digit North American Industry Classification Standard (NAICS) industry $j$ in year $t$; $CapE{x}_{ijt}$ represents firm capital expenditure; $Em{p}_{ijt}$ captures the number of employees; $In{v}_{ijt}$ denotes a firm’s inventory; $U_{ijt}$ is specified as a nonnegative random variable assumed to be $iid~N(0,{\sigma }_{U_j}^2)$ and associated with technical inefficiency (Battese and Coelli 1995; Greene 2005); $V_{ijt}$ is an error term assumed to be $iid~N(0,{\sigma }_{V_j}^2)$ and distributed independently of $U_{ijt}$; $O{E}_{ijt}$ is bound between (0 and 1), with 1 representing $OE$ at the production frontier, with other values sitting below the frontier.

One drawback of utilizing efficiency estimates is that the calculations are usually completed with a relative comparable group (e.g., industry). To account for this, we calculate operational efficiency measures by excluding two-digit NAICS with less than five observations in any given year (Li et al. 2021; Yiu et al. 2020).

3.4. Control Variables

To mitigate influences of potentially confounding aspects of firms in this study, control variables used in our regressions are (1) firm size; (2) financial leverage; (3) R&D intensity; (4) SG&A intensity; and (5) industry competitiveness. We also include industry and year fixed effects.

3.4.1. Firm Size

When a company implements green practices, there will be purchases of intermediate goods and services related to the provision of green products. Large firms can extract scale economies via quantity or bulk purchase savings on input costs. They can also spread the costs of using human resources over numerous units of output (McWilliams and Siegel 2001). Large firms will likely be highly conscious of their public image because of the large scale and scope of any negative ramification from bad publicity. Hence, large firms typically possess economies of scale and also scope that enhance or depress the relationship between green practices and financial performance. Thus, firm size is included as a control variable in our regressions. Firm size (LNEMP) is captured by the number of employees in a firm, and we take the natural logarithm of the number of employees to address normality issues.

3.4.2. Financial Leverage

When firms are highly leveraged, they may have less capacity to invest in activities required to implement global environmental standards, green practices, and the development of green initiatives. Therefore, we control for this potential effect by including financial leverage (FIN_LEV) as a control variable in our study. FIN_LEV is calculated by dividing the long-term debt by total assets

3.4.3. R&D Intensity

Research and development (R&D) investment can result in both process and product innovations, which are each valued by some consumers. Therefore, product differentiations achieved through R&D can positively affect the provision of green attributes (McWilliams and Siegel 2001). Furthermore, product differentiation through the use of green initiatives, such as recycled products or reduction of GHG emissions, may include investment in research and development (R&D). Additionally, investments in R&D can enable organizations to improve their greenhouse gas performance (Koh et al. 2017). Also, employees of firms with higher investments in R&D are found to be better at learning from customer feedback, thereby increasing sales of the organization (Guo et al. 2020).

Hence, R&D intensity (RD_INT) is considered to be a control measure for firm performance, since R&D seeks to produce more successful products. Thus, higher performing firms tend to spend more on R&D. We capture RD_INT as the ratio of R&D expenditures divided by total assets. Since many of our sampled firms do not disclose R&D expenses, we adopt the methodology recommended by Koh and Reeb (2015) and substitute the absent R&D intensity values with the industry average at the three-digit SIC code level.

3.4.4. SG&A Intensity

We measure SG&A intensity (SGA_INT) as the ratio of SG&A expenses to total assets. SG&A intensive firms may underutilize their assets to control operational costs and vice versa. Thus, underutilization can lead to inefficiencies that can negatively affect a firm’s financial performance.

3.4.5. Industry Competitiveness

Firm financial performance can also be impacted by industry competitiveness. Firms in more competitive environments will either need to invest more to differentiate their products and services from their competitors or invest in improving efficiencies to reduce costs. In either scenario, firm financial performance can be negatively affected. The opposite would hold true for firms in less competitive industries. We capture industry competitiveness (HHI) using the Herfindahl–Hirschman index, taking the sum share square total of two-digit NAICS codes. We include

Table A1. Summary of variable operationalization.

Variable	Variable Long Name	Source	Calculation
ROS	Return on Sales	Compustat	Net income divided by total revenue
EP	Environmental Performance	CDP	Scope 1 intensity emissions normalized by 2-digit NAICS codes
OP	Operational Productivity	Compustat	Frontier analysis of 2-digit NAICS codes using CapEx, employee, and inventory, with outcome as operating income
LNEMP	Firm Size	Compustat	Natural logarithm of number of employees
FIN_LEV	Financial Leverage	Compustat	Total long-term debt divided by total assets
RD_INT	R&D Intensity	Compustat	R&D expenditures divided by total assets
SGA_INT	SGA Intensity	Compustat	SG&A expense divided by total assets
HHI	Herfindahl–Hirschman Index	Compustat	Sum share square total of 2-digit NAICS codes

in the Appendix A, which summarizes the operationalization of our study’s variables.

Table 2. Table for correlation and descriptive statistics.

Correlation Table
		(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
(1)	ROS	1.000
(2)	EP	−0.090	1.000
(3)	OP	0.087	0.019	1.000
(4)	LNEMP	0.020	0.031	0.057	1.000
(5)	FIN_LEV	−0.099	0.084	−0.008	0.171	1.000
(6)	RD_INT	0.120	−0.210	−0.154	−0.030	−0.158	1.000
(7)	SGA_INT	0.072	−0.175	0.062	−0.017	−0.143	0.375	1.000
(8)	HHI	−0.038	0.228	0.023	0.138	0.034	−0.199	−0.054	1.000
	Univariate Statistics	ROS	EP	OP	LNEMP	FIN_LEV	RD_INT	SGA_INT	HHI
	MEAN	0.067	0.092	0.641	1.697	0.261	0.028	0.188	0.048
	SD	0.145	0.198	0.249	2.394	0.161	0.053	0.139	0.065
	MIN	−2.621	0.000	0.000	−8.470	0.000	0.000	0.000	0.013
	MAX	5.957	1.000	0.999	7.741	1.846	0.531	0.961	0.688

contains the descriptive statistics of our variables. These data show that there is a negative correlation between EP and ROS, while the correlation between OP and ROS is positive. Further, the correlation matrix for the control variables suggests no major concerns for multicollinearity. In relation to the descriptive statistics, we see that the mean of EP is 0.092. We recall that EP is normalized, and thus a mean value closer to 0 than to 1 suggests that, while some companies are likely achieving high levels of EP, the average across all sampled firms remains modest. Table 2 also shows that the mean of OP is 0.641. This implies that, on average, firms exhibit a relatively high level of operational productivity.

3.5. Methodology

We use the following regression models to test our hypotheses. For Hypotheses 1 and 2, we use:

$$F{P}_{i,t}={\beta }_{1}O{P}_{i,t}+{\beta }_{2}E{P}_{i,t}+{\lambda }_1{Z}_{i,t}+{\mu }_i+{\tau }_t+ϵ$$

(3)

where $F{P}_{i,t}$ represents firm financial performance for firm i in year t. $O{P}_{i,t}$ and $E{P}_{i,t}$ are the main variables of interest. $Z_{i,t}$ is the set of controls described in Section 3.4. ${\mu }_i$ and ${\tau }_t$ are firm and year fixed effects.

To test Hypothesis 3, we interact both OP and EP, resulting in the following regression equation:

$$F{P}_{i,t}={\beta }_{1}O{P}_{i,t}+{\beta }_{2}E{P}_{i,t}+{\beta }_{3}O{P}_{i,t}\ast E{P}_{i,t}+{\lambda }_1{Z}_{i,t}+{\mu }_i+{\tau }_t+ϵ$$

(4)

Variables in Equation (4) are the same as in Equation (3). The additional inclusion of $O{P}_{i,t}\ast E{P}_{i,t}$ captures the moderating effect.

We employ random effects estimation to Equations (3) and (4) to account for the panel structure of our data. While there are some advantages to employing a fixed effects (within transformed) estimation, we choose random effects because the impact of environmental performance and operational productivity on financial firms is across firms and not within a firm (Ketokivi et al. 2021). As such, all the results in our main analysis are estimated via a random effects estimation, with cluster robust standard errors at the firm level.

4. Analysis and Results

In

Table 3. Main results using random effects estimation.

	(1)	(2)	(3)
DV	ROS	ROS	ROS
EP (H1)		−0.0734 ***	−0.150 ***
		(0.000)	(0.000)
OP (H2)		0.101 ***	0.0895 ***
		(0.000)	(0.000)
EP x OP (H3)			0.120 **
			(0.005)
LNEMP	0.00178	0.00118	0.00125
	(0.217)	(0.408)	(0.378)
FIN_LEV	−0.216 ***	−0.202 ***	−0.201 ***
	(0.000)	(0.000)	(0.000)
RD_INT	0.229 *	0.288 **	0.282 **
	(0.018)	(0.002)	(0.002)
SGA_INT	−0.136 ^	−0.128	−0.127
	(0.085)	(0.120)	(0.120)
HHI	−0.0424	−0.0300	−0.0127
	(0.744)	(0.816)	(0.922)
CONSTANT	0.195 *	0.130	0.120
	(0.018)	(0.102)	(0.133)
OBSERVATIONS	9414	9414	9414

Note: All estimations include industry and year fixed effects. p-values in parentheses ^ p < 0.10, * p < 0.05, ** p < 0.01, *** p < 0.001.

, we present the empirical results of testing our hypotheses. Column (1) contains results for a model with only control variables. We find that FIN_LEV $(\beta =-0.216; p<0.001$) and RD_INT $(\beta =0.229; p=0.018$) are statistically significant in effecting FP. These results suggest that the more financially leveraged a firm is, the worse its financial performance. Further, firms that invest more in R&D intensity demonstrate better financial performance. Column (2) reports the results for Equation (3), which tests both Hypotheses 1 and 2. According to Hypothesis 1, EP will negatively impact FP. The coefficient of EP $(\beta =-0.0734; p<0.001$) is negative and significant, suggesting that the more a firm emits compared with its industry peers, the worse its financial performance. To test Hypothesis 2, we assess whether higher OP will increase FP. The coefficient of OP $(\beta =0.101; p<0.001$) in column (2) is positive and significant. This result provides empirical evidence that a firm’s operational productivity positively impacts its financial performance.

Column (3) contains the empirical results for estimation Equation (4) and pertains to Hypothesis 3. We recall that we are interested in examining whether a negative moderating effect is present for EP and OP. The coefficient of EP × OP $(\beta =0.120; p=0.005$) suggests that better operational productivity can offset the negative impacts of environmental performance.

For a graphical interpretation of the results, the effect of different levels of OP on the relationship between EP and FP is given in Figure 1. We plot the effect of EP on FP at three different levels of OP (1st, 50th, and 99th percentile) and observe some boundary conditions. First, we see that the slope of EP reduces as OP increases. This suggests that the negative impact of EP on FP reduces as firms increase their OP. Given that none of the confidence bands overlap each other, we can say that the effect of EP on FP is statistically different at these three levels. The figure also highlights some boundary conditions; specially, when OE is at the first percentile, EP needs to be above 0.3 for statistical significance between EP and FP. We also observe that, when OE is at the 50th percentile, high levels of EP result in a nonsignificant relationship between EP and FP.

5. Robustness of Results

We conduct a series of robustness checks to ensure that our results are robust to different specifications.

5.1. Endogeneity

Endogeneity exists when there is a correlation between the independent variables and the error term. We address this issue through two methods. First, we estimate a fixed effects (within transformation) estimation. The within transformation estimates the panel data by demeaning all variables. This purges all time invariant unobservable fixed effects, which should alleviate some endogeneity concerns. Second, we employ Arellano and Bond’s system GMM model (Arellano and Bond 1991; Arellano and Bover 1995). The literature has used this approach to also address the endogeneity concern (Modi and Cantor 2021; Sartal et al. 2020; Ullah et al. 2018). For a given estimate equation, the system GMM approach establishes an equation system by stacking the original equation with the first differenced equation and solving them simultaneously. Blundell and Bond (1998) show that when the dependent variable is strongly path-dependent, the system GMM model yields efficient and unbiased estimates. For brevity, we summarize the results of both estimation techniques in Appendix A,

Table A2. Endogeneity concerns.

	(1)	(2)	(3)	(4)
DV	ROS	ROS	ROS	ROS
Model	FE	FE	AB	AB
EP (H1)	−0.0545 *	−0.115 *	−0.0558 ***	−0.162 ***
	(0.027)	(0.012)	(0.000)	(0.000)
OP (H2)	0.104 ***	0.0948 **	0.0784 ***	0.0644 ***
	(0.000)	(0.002)	(0.000)	(0.000)
EP x OP (H3)		0.0958 ^		0.159 ***
		(0.062)		(0.001)
Lag ROS			0.384 ***	0.382 ***
			(0.000)	(0.000)
LNEMP	−0.0153	−0.0155	0.00124	0.00132
	(0.186)	(0.181)	(0.264)	(0.233)
FIN_LEV	−0.284 ***	−0.284 ***	−0.0721 ***	−0.0732 ***
	(0.000)	(0.000)	(0.000)	(0.000)
RD_INT	−0.434 ^	−0.444 ^	0.209 ***	0.195 ***
	(0.085)	(0.074)	(0.000)	(0.000)
SGA_INT	−0.519	−0.524	−0.00340	−0.00157
	(0.248)	(0.245)	(0.816)	(0.913)
HHI	−0.00838	0.00685	−0.0951	−0.0774
	(0.950)	(0.960)	(0.524)	(0.606)
CONSTANT	0.223 *	0.229 *	0.112	0.102
	(0.028)	(0.027)	(0.138)	(0.185)
OBSERVATIONS	9414	9414	7249	7249

Note: All estimations in Table A2 and Table A3 include industry and year fixed effects. p-values in parentheses ^ p < 0.10, * p < 0.05, ** p < 0.01, *** p < 0.001.

. In sum, the results provide consistent support for Hypotheses 1, 2, and 3.

5.2. Alternative Measures

We further test the sensitivity of our results by different operationalizations of the dependent variable, independent variables, and control variables. Instead of using return on sales as a proxy for financial performance, we use return on assets. Following our theoretical arguments, we operationalize EP as the normalized value of Scope 1 emissions. For robustness, we operationalize EP as the normalized value of combined Scope 1 and Scope 2 emissions. To ensure that operational efficiency measured by the stochastic frontier analysis does not drive our findings, we alternate the operationalization of operational efficiency using the notion of fixed asset efficiency (Fu and Jacobs 2022). We measure a firm’s fixed asset efficiency by first taking the ratio of revenue over total property plant and equipment and normalizing it by industry (at the NAICS two-digit level). Lastly, academics have suggested that the inclusion of control variables can improve the statistical significance of the main independent variables. To account for this concern, we estimate our results without any control variables.

Table A3. Alternative measures.

	(1)	(2)	(3) 1	(4) 1	(5) 2	(6) 2	(7)	(8)
DV	ROA	ROA	ROS	ROS	ROS	ROS	ROS	ROS
Model	RE	RE	RE	RE	RE	RE	RE	RE
EP (H1)	−0.0172 **	−0.0389 **	−0.0612 ***	−0.144 ***	−0.0691 ***	−0.0837 ***	−0.0672 ***	−0.146 ***
	(0.009)	(0.007)	(0.000)	(0.001)	(0.000)	(0.000)	(0.000)	(0.000)
OP (H2)	0.0759 ***	0.0726 ***	0.103 ***	0.0876 ***	0.105 ***	0.0967 ***	0.105 ***	0.0923 ***
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
EPxOP (H3)		0.0340 *		0.131 **		0.121 *		0.123 **
		(0.041)		(0.005)		(0.046)		(0.006)
LNEMP	0.00118 *	0.00120 *	0.00129	0.00138	0.000881	0.000953
	(0.041)	(0.037)	(0.394)	(0.358)	(0.528)	(0.494)
FIN_LEV	−0.153 ***	−0.152 ***	−0.210 ***	−0.209 ***	−0.203 ***	−0.203 ***
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
RD_INT	0.00264	0.00185	0.273 **	0.264 **	0.137	0.142
	(0.950)	(0.964)	(0.007)	(0.008)	(0.162)	(0.147)
SGA_INT	−0.0452 *	−0.0451 *	−0.142	−0.140	−0.136 ^	−0.137 ^
	(0.018)	(0.017)	(0.134)	(0.133)	(0.081)	(0.079)
HHI	0.0232	0.0282	−0.0343	−0.0287	−0.0391	−0.0424
	(0.461)	(0.370)	(0.793)	(0.827)	(0.780)	(0.759)
CONS	0.0606 ^	0.0579 ^	0.129	0.126	0.165 ^	0.147	0.0194 ^	0.0272 *
	(0.072)	(0.082)	(0.115)	(0.123)	(0.078)	(0.128)	(0.057)	(0.013)
OBS	9414	9414	9334	9334	9206	9206	9414	9414

Note: all estimations in B.1 and B.2 include industry and year fixed effects. p-Values in parentheses ^ p < 0.10, * p < 0.05, ** p < 0.01, *** p < 0.001. ¹ Columns (3) and (4) are estimated using Scope 1 and 2 intensity emissions normalized at the 2-digit NAICS for EP. ² Columns (5) and (6) are estimated using fixed asset efficiency, the ratio of OI/PPE normalized at the 2-digit NAICS2, for OP.

contains results for the above procedures, which show consistent support for our main findings.

Table 4. Summary of robustness results.

Category	Alternative Specification	Support Hypo	Table
Endogeneity	Fixed Effects Regression	H1 H2 H3	Table A2
	Arellano and Bond System GMM	H1 H2 H3	Table A2
Alternative Measures	FP Measured as Return on Assets	H1 H2 H3	Table A3
	EP Measured as Normalized Scope 1 and 2 Intensity	H1 H2 H3	Table A3
	OP Measured as Fixed Asset Efficiency	H1 H2 H3	Table A3
	No Controls	H1 H2 H3	Table A3

summarizes the specifications and corresponding results of our robustness tests.

6. Discussion and Future Work

6.1. Discussion

We initiated our study with the motivation of managerial actions toward sustainability pressures from various stakeholders on the efforts of publicly traded companies to meet financial performance expectations of shareholders. Following Hypothesis 1, our findings indicate that companies can satisfy stakeholder concerns about environmental issues while also being financially successful for their shareholders. However, as indicated by the results supporting Hypothesis 3, the tension between negative environmental effects and business operations to achieve positive financial outcomes are governed by how well firms manage other operational aspects of production. Our findings supporting Hypothesis 2 denote that firms need to be operationally productive to achieve financial benefits in relation to lowering their direct GHG emissions. In particular, with respect to the moderating effect proposed according to Hypothesis 3, efforts to reduce GHG emissions will yield high financial returns for companies while in the presence of high operational productivity levels. Companies with low operational productivity tend to miss out on financial benefits associated with controlling GHG emissions. Thus, threats to the financial outcomes of firms by initiatives to improve environmental performance targets on firms via elective action, activist shareholders, or government policies can be mitigated in part when firms are operationally productive compared with industry peers.

Our data-driven empirical examination of the relationships between GHG emissions and financial performance in light of operational productivity provides a positivist approach for organizations interested in the feasibility of pursuing returns for shareholders and sustainable practices. From an emissions mitigation perspective, high GHG emission levels of companies are associated with poor financial performance. Therefore, environmental effects of productive processes should be considered by managers because it is a responsible attitude both environmentally and financially. A key take-away is that, when engaging in green practices, whether due to stakeholder pressures, policies of governments, or to acquire high financial returns, firms should focus on achieving competitively high operational productivity. Green-oriented initiatives to control GHG emissions are only truly sustainable if the firm remains a going concern, and the firm’s financial benefits associated with this friendlier environmental practice are moderated by the firm’s operational productivity. Therefore, operations managers and executives should be mindful of maintaining high OP levels when trying to achieve low GHG emissions and also perform well financially.

Since our study includes a direct measurement of environmental performance of firms in the form of absolute levels of GHGs emitted during Scope 1 firm operations, our results can be conveniently comprehended across all hierarchical levels of firms and can inform all employees about how the environmental performance of the company impacts the operational returns of the company. Operational managers can stress the point that increases in GHG emissions lower the profitability of the company. Strategic managers can interpret the results of our study and understand that, no matter the efforts invested in working towards improving the environmental performance of the company or the sustainability agenda followed by the company, in order to encourage returns from such investments in the form of improved financial performance, the company should have high operational productivity levels. Therefore, our study signals to strategy makers that inventory, labor, and fixed asset productivity levels of the company should be high if the company is seeking financial returns from their sustainability agenda. With the resource-based view grounded in the competitiveness of operational productivity, the firm will garner the optimal conditions for reduced GHG levels to provide financial benefits for the company.

Our analysis of the main effect between firms’ environmental performance and their profitability turns out to be positive and significant, since a firm with lower GHG emissions or higher environmental performance will have a high financial performance, suggesting that firms with greater environmental performance can attract skilled workers and improve their reputation as “green” organizations, resulting in a greater market presence for the firms (Jacobs et al. 2016). Firms with high environmental performance and operational productivity may also be able to attract greater customer attention via positive advertising. Hence, this suggests that the greater productivity of green entities can be translated to improved bottom lines that include people, planet, and profit (Jacobs et al. 2016). The fundamental contribution of our moderation results to the existing literature on the green performance of firms is the identification of the crucial role of operations in leveraging the green or environmental performance of the company for its financial benefit. Even if the main role of managers and executives is to improve the productivity of the company, many decision makers might not understand how OP facilitates performance gains for the firm via the environmental–financial performance link. Therefore, our findings illuminate the important role played by operations management in garnering a sustainable relationship between a firm’s environmental performance and its financial livelihood.

In summary, our research has several important findings. We highlight the essential role of operational productivity for the relational dynamics of the “profit” and “planet” aspects of the triple bottom line performance of the firm. We find that the win–win paradigm of strong financial performance and strong environmental performance happens when the firm is operationally efficient. In other words, firms that “do well and do good” are competitively positioned as highly productive organizations.

6.2. Future Work

Furthermore, future studies can extend our study to incorporate other aspects of ESG metrics by addressing social and governing aspects of the firm. As Rothenberg et al. (2001) indicate, input efficiency improvements include practices like buffer minimizations, management systems, and human resource management practices that are elements of lean management. In our study, we discuss how such practices result in different efficiencies like inventory and labor efficiencies that improve the firm’s overall OP score. Therefore, future researchers can design studies to evaluate the relationship between the people and the profit aspects of performance as moderated by OP. Further, our approach to firm environmental performance can also be examined along with financial, social, and governing performance measures to provide a comprehensive triple bottom line perspective of firm competitiveness in light of OP. Additionally, future work can involve different metrics to gauge environmental performance like KLD measures, which can be benchmarked for competitiveness (similar to how we calculate the OP score in our study).

Extending our approach beyond U.S. data to a global scope can incorporate various national or international regulatory aspects of firm performance. Similarly, due to data availability, we restrict our study to the direct greenhouse gas emissions of firms. Thus, other areas for future work include testing the generalizability of our results with sample data from various countries and with variables that reach beyond the firm.