Equal Employment Opportunity in Supply Chains

Abstract

An equal employment opportunity (EEO) policy is an essential practice of workplace diversity, equity, and inclusion. Using a novel measure of workplace EEO based on the textual analysis of online job postings, we show a lead–lag pattern of EEO practice between principal customers and their dependent suppliers. The leadlag effect is more pronounced when principal customers enjoy greater bargaining power, consistent with the view that dependent suppliers adjust EEO policies to cater to customer preferences. We use the 2017 Diversity Campaign of the “Big Three” institutional investors as our main identification strategy to support a causal interpretation. We demonstrate that the diffusion of EEO policies along supply chains improves the likelihood of retaining female employees, implying the presence of social benefits in addition to economic advantages.

Keywords

Supply chain management equal employment opportunity (EEO)

1. Introduction

The attention on diversity, equity, and inclusion (DEI) has reached an unprecedented level recently, particularly in the wake of significant events such as the #MeToo and Black Lives Matter protests, which have sparked widespread social equality movements globally.¹ This rapidly growing attention is unsurprising given that DEI serves as a crucial practice in fostering an inclusive community where all individuals are esteemed and treated with respect. Moreover, DEI acts as a vital catalyst for corporate creativity and productivity, thereby generating substantial value for shareholders. From a standalone company’s perspective, researchers have shown that DEI is crucial to improve decision-making process (e.g., Kim and Starks, 2016; Bernile et al., 2018; Delis et al., 2017), engage and retain talents with diverse skills (e.g., Morgan and Várdy, 2009; Dezsö and Ross, 2012; Matsa and Miller, 2011), increase customer satisfaction and loyalty (e.g., Cen et al., 2022; Khan and Kalra, 2022), and reduce litigation risk related to legal compliance (e.g., Cumming et al., 2015; Adhikari et al., 2019; Wowak et al., 2021).

Equal employment opportunity (EEO), considered an essential component of DEI, prevents employers from “discriminating against a job applicant or an employee because of the person’s race, color, religion, sex (including pregnancy and related conditions, gender identity, and sexual orientation), national origin, age (40 or older), disability, or genetic information.”² Since recruitment is the starting point of employment, hiring with a sound EEO policy is the key to opening the “front door” to DEI. Specifically, an effective equal opportunity policy in recruitment removes systematic biases against minority groups and allows applicants from diverse backgrounds to gain job opportunities with their unique qualities. The respect for diversity and equity, as a natural outcome of practicing EEO in recruitment, nurtures a strong sense of belonging among employees (existing and new) and creates inclusive workplaces.

We argue in this paper that the importance of EEO extends beyond its impact on standalone firms. Over the past few decades, companies have realized that managing environmental, social, and governance (ESG) risks is crucial to maintaining sustainable and efficient supply chain relationships. While much emphasis has been placed on environmental and governance issues, the “social” aspect of ESG, particularly EEO-related labor issues, has received less attention in supply chain management.³ Recently, there has been a substantial shift in this trend. For example, Unilever set a goal to encourage its suppliers to participate in a fairer, more inclusive, and more equitable manner of conducting business.⁴ In 2014, Unilever introduced its Responsible Partner Policy, which requires its suppliers to enforce workplace equality. The mandatory requirements include “…no discrimination in employment (including hiring), any aspects of training, personal and professional development, compensation, advancement, discipline, employment termination or retirement.” In addition, Unilever mandates its suppliers to provide training for all workers, particularly managers, to prevent discrimination.⁵ In another example, PepsiCo leveraged the company’s scale and invited its suppliers to join PepsiCo’s Racial Equality Journey to create shared value, “a key pillar of its pep+ (PepsiCo Positive) transformation.”⁶

Despite strong corporate EEO initiatives in supply chain management, researchers have not fully understood whether and how such initiatives generate a systematic diffusion of EEO practices along the supply chain; if this diffusion does exist, what are the economic channels and outcomes of the EEO diffusion along the supply chain? This study fills these gaps in the literature. Taking advantage of a novel EEO measure based on textual analysis of EEO statements in online job postings, we study how firms interact with EEO practices in the supply chain setting and investigate the impact of EEO interaction on supply chain relationship improvement. We also explore whether supply chains that adopt EEO, that is, suppliers catering to their customers’ EEO preferences, can achieve both positive economic outcomes and social good.

While EEO is protected by laws and regulations in the United States,⁷ these laws and regulations do not require companies to include the “equal opportunity” label in their job postings. However, anecdotal evidence suggests that passionate EEO statements in job postings encourage potential employees, particularly those from minority groups, to apply.⁸ Cen et al. (2022) identify a significant cross-sectional difference in the EEO statements of US firms in their online job postings. They argue that, since EEO statements are public commitments that are verifiable ex-post, the sincerity and enthusiasm in EEO statements can accurately signal the level of corporate EEO preferences and practices. Moreover, while most existing EEO measures in the literature capture workplace equality in boardrooms or C-suits, the EEO measure in Cen et al. (2022) is based on EEO statements in job postings and, therefore, captures corporate workplace equality practices relevant to rank-and-file employees.⁹ We combine this measure with the customer–supplier relationship data from the Compustat Customer Segment File and empirically examine the diffusion of EEO policies along supply chains.

Our baseline results show that an increase in principal customers’ EEO scores is correlated with a future improvement in the EEO scores of their suppliers. We conduct two placebo tests to ensure that this pattern is specific to firm-level supply-chain relationships and not driven by industry-level economic links. First, we construct a placebo sample by keeping the actual customer firm and replacing the actual supplier with placebo suppliers that produce similar products. Second, we conduct a simulation exercise using 1,000 benchmark samples in which placebo suppliers are randomly chosen from the industry that the actual suppliers belong to. Both tests confirm that the empirical pattern in our baseline tests is not driven by industry-level links or spurious correlations driven by economic factors unrelated to supply-chain relationships.

Our main conjecture for the baseline results above is that suppliers adjust their EEO practices to cater to their customer preferences and thereby maintain and strengthen the relationships with principal customers. A potential endogeneity concern in our interpretation is that customer and supplier firms may adjust their EEO policies together as a result of common economic factors unobserved by researchers (i.e., the “omitted variable” issue). To alleviate this endogeneity concern, we take advantage of the 2017 Diversity Campaign of the “Big Three” institutional investors as an exogenous shock to the EEO policies of customer firms.¹⁰ We verify that customer firms with higher Big Three institutional ownership but fewer female directors (the “affected” firms) were disciplined by the Big Three institutions, following which their EEO scores improved more significantly than peer firms. Our results show that the EEO scores of affected customers’ suppliers experienced a more substantial increase than those of other suppliers, even after controlling for suppliers’ exposure to this campaign. Our results from this identification strategy are consistent with the causal claim that customers’ workplace equality practices affect the workplace equality of their suppliers.

We argue that suppliers have a strong economic incentive to adjust their workplace EEO policies and cater to the preferences of their principal customers because the purchases of principal customers account for a large proportion of their total sales. Therefore, whether principal customers can effectively play this implicit disciplinary role on EEO policies along the supply chain depends on their relative bargaining power with dependent suppliers. One of the most important determinants of customer bargaining power in bilateral trade relationships is the credibility of the threat to switch, that is, how costly it would be for a customer to switch to other “substitute” suppliers (Crook and Combs, 2007; Banerjee et al., 2008). We proxy customers’ bargaining power with two measures related to customer switching costs: product market competitiveness and product similarity in the supplier industry. Consistent with our conjecture, we find a much stronger lead–lag pattern in workplace EEO policies along the supply chain when suppliers face more intensive market competition or have products more similar to their industry peers, that is, when customers face a lower switching cost and, therefore, enjoy greater bargaining power.

Finally, we examine the real economic and social outcomes when EEO policies diffuse along supply chains. Specifically, we investigate the impact of EEO diffusion along supply chains on the retention of female employees. We measure female (male) employee retention by the average employment duration of female (male) employees at the end of the year. We show that, even after controlling for its own EEO score, a supplier is more likely to perform better in female employee retention if its customers have higher EEO scores. In comparison, suppliers’ own EEO scores and EEO diffusion along the supply chain do not have significant effects on male employee retention, which is consistent with anecdotal evidence that the effects of workplace equality practices are more pronounced in retaining female workers.¹¹

The rest of this paper is organized as follows. In Section 2, we discuss our contributions to various strands of related literature. Section 3 describes the data and variable construction. Section 4 presents the baseline results on the diffusion of EEO practices along supply chains, placebo tests to address alternative interpretations, and the identification strategy to address endogeneity issues. Section 5 presents our analysis of moderating mechanisms. Section 6 examines the real economic outcomes of EEO diffusion along the supply chain. Section 7 concludes our study.

2. Literature Review

We examine several important social and labor aspects of sustainable (or responsible) supply chain management. The vast operations management literature on sustainable supply chain management explores environmental dimensions.¹² For example, Bowen et al. (2001) study the role of “green supply” management. Jira and Toffel (2013) and Blanco (2021) show that carbon emissions along the supply chain generate significant impacts on the nature and content of climate change disclosures.¹³ Lee and Klassen (2008) identify the key factors that contribute to the development of environmental management capabilities for small- and medium-sized suppliers in supply chains. Despite the early focus on environmental issues along the supply chain, recent studies have shifted their attention to other dimensions of sustainability, such as social and labor issues, by considering a more comprehensive set of stakeholders (Sodhi, 2015). For example, Soundararajan and Brammer (2018) use a longitudinal multiple case study method to study supply chains of the knitwear garment industry in South India and find that social sustainability requirements affect suppliers’ fairness perceptions and reciprocity. Using multiple case studies, Kalkanci et al. (2019) discuss the practices of inclusive supply chain innovation and highlight the importance of collaboration with other stakeholders to build an inclusive supply chain. Erhun et al. (2021) examine how Coca-Cola’s restructuring of its mango supply chain to support Haitian farms resulted in socially responsible procurement that reduced customers’ costs and lead time. Huq et al. (2016) use cases of the clothing industry to explore social management capabilities and their effects on sustainable supply chain management. In their theoretical framework, Villena and Gioia (2018) incorporate “people” as one of the sustainability dimensions for managing sustainable supply networks. Unlike previous studies that mainly rely on experiments, surveys, or case studies, we explore workplace equality issues along the supply chain based on granular archival information in almost all online job posts issued by US firms. We also provide a comprehensive set of verification tests to show that our measure is able to capture the quality of corporate DEI accurately and that our approach is generalizable.

We also contribute to the literature on the economic dynamics of the diffusion (cascade) of sustainability along the supply chain. Previous studies have shown that productivity (Serpa and Krishnan, 2018), quality (Zhu et al., 2007; Corbett, 2006), innovation (Chu et al., 2019), operating risks (Hertzel et al., 2008; Osadchiy et al., 2021), and credit risk (Agca et al., 2022; Wu et al., 2021) propagate along the supply chain. Not surprisingly, sustainability is no exception: although customer and supplier firms are separate business entities, the customer firms are still held responsible when suppliers experience sustainability-related incidences (Hartmann and Moeller, 2014). Since most severe environmental and social risks are concentrated among the lower-tier suppliers in the supply chain (Villena and Gioia, 2018), customer firms have a strong economic incentive to mitigate sustainability-related risks along the supply chain.¹⁴

Recent studies have started to focus on how the structure of supply chains affects the cascade of sustainability practices. Villena and Gioia (2018) establish a theoretical framework for the optimal structure of a sustainable supply chain network in which customers can cascade sustainable practices via direct or indirect control. Using a game theoretic model, Huang et al. (2022) show that the optimal cascading practice is either direct control, delegation, or no effort under different economic conditions. Empirically, researchers find that indirect control is usually more efficient (Lee and Klassen, 2008), especially when the first-tier suppliers possess more information about the production process (Huang et al., 2022) and the institutional environment (Soundararajan and Brammer, 2018) of the lower-tier suppliers.¹⁵ Under the setting of indirect control, researchers dig into the economic mechanisms that generate and facilitate sustainability cascades along the supply chain. For example, Porteous et al. (2015) develop a model showing that both suppliers’ economic incentives (“carrots”) and customers’ threat of potential penalties (“sticks”) reduce suppliers’ environmental and social violations. Consistent with their views, we suggest that the bargaining power of principal customers, based on credible threats of switching to other suppliers, plays an important role in the cascade of workplace equality practices along the supply chain. More importantly, we address the endogeneity issues (i.e., unobserved economic factors may affect workplace equality practices of both customers and suppliers) in this cascade by taking advantage of an identification strategy based on the Big Three Diversity Campaign and show that the workplace equality practices of principal customers causally affect those of their dependent suppliers.

Last but not least, our study is related to the existing literature on the real economic outcomes of sustainable supply chain management. Previous studies show that sustainable supply chain management boosts sales and profitability. Using game theory models, Buell and Kalkanci (2021) and Kalkanci and Plambeck (2020) show that transparency regarding suppliers’ sustainability performance improves sales and profitability for both the customer and the supplier. Parmigiani et al. (2011) establish a connection between sustainable supply chain management and the gaining of competitive advantage. Other studies connect supply chain management with cost efficiency. Based on multiple case studies, Erhun et al. (2021) suggest that socially responsible procurement reduces customers’ costs and lead time. Based on both model and survey evidence, Porteous et al. (2015) show that sustainable supply chain management is positively associated with operating cost reduction of the customers. Bird et al. (2019) show that labor codes of conduct imposed by customers improve suppliers’ labor practices using social audit data. Consistent with these findings in the literature, we find that the diffusion of workplace equality practices generates real economic outcomes for suppliers. Specifically, while workplace equality practices diffuse along the supply chain from customers to suppliers, suppliers’ improvement in social sustainability leads to an increase in the likelihood of retaining female workers.

3. Data and Variable Construction

3.1. DEI Measure: EEO Score

We construct the EEO scores to measure the levels of corporate diversity, equity, and inclusion of US firms based on the big data of online job postings from Burning Glass Technologies (BGT) following Cen et al. (2022). BGT is an analytic software company that collects and manages the world’s largest online job vacancy database.¹⁶ BGT web crawls job postings from more than 50,000 career and company websites, covering the near-universe online job postings of US firms. The job posting data set consists of 54 million job postings for the period 2010 to 2020. The employers in the BGT data are matched to firms in the Compustat database by a name-matching algorithm. The matching results are manually checked to ensure their accuracy and reliability.

While BGT data are widely used in the labor literature (e.g., Hershbein and Kahn, 2018), one unique feature of the BGT data used in Cen et al. (2022) is that the data contain not only the standardized and structured data items developed by BGT, but also the original unstructured texts of the job postings. The original text of job postings enables us to extract information embedded in the equal employment opportunity (EEO) statements of job postings.

Companies include EEO statements in their job postings to publicize their commitments to workplace EEO. Firms that genuinely embrace workplace EEO tend to make enthusiastic and original EEO statements in their job postings. Forexample, Gap Inc., which leads in the list of top 100 most diverse and inclusive companies rated by Refinitiv, has the following EEO statement in its job postings:¹⁷

Gap Inc. is an equal-opportunity employer and is committed to providing a workplace free from harassment and discrimination. We are committed to recruiting, hiring, training, and promoting qualified people of all backgrounds, and make all employment decisions without regard to any protected status. We have received numerous awards for our long-held commitment to equality and will continue to foster a diverse and inclusive environment of belonging.

However, simply stating that “EEO is law” should not be taken as a credit for the positive signal of the firm’s EEO commitment.¹⁸ Following Cen et al. (2022), we use the Valence Aware Dictionary and Sentiment Reasoner (VADER) algorithm from the natural language processing sentiment analysis toolkit to distinguish the actual attitudes toward EEO underlying the EEO statements of firms.¹⁹ The VADER is a lexicon-based and language rule-adjusted sentiment analysis algorithm that assigns sentences with sentiment scores ranging from $- 1$ to 1. A sentiment score of zero indicates that the underlying attitudes of the sentence are neutral, whereas a more positive sentiment score indicates a more positive attitude. We tag each EEO statement from job postings with the sentiment score rated by VADER. More genuine and enthusiastic EEO statements receive higher sentiment scores. For example, the sincere EEO statement of Gap Inc. receives a high sentiment score of 0.919, whereas the average sentiment score of all EEO statements of US public firms is 0.472.

The firm-year level EEO score is the normalized rank of the average sentiment score of EEO statements of all the job postings of the firm in the particular year:

E E O S c o r e_{i, t} = {rank}_{t} (\frac{\sum_{j \in i, t} E E O {S e n t i m e n t S c o r e}_{j}}{{N u m b e r o f J o b P o s t i n g s}_{i, t}}),

where

{EEO Sentiment Score}_{j}

is the sentiment score of the EEO statement of the posting

j

and

{Number of Job Postings}_{i, t}

is the total number of job postings issued by the firm

i

in year

t

. The EEO score ranges from 0 to 1, with a sample average of 0.5 by construction. The firms with higher EEO scores are those with higher levels of workplace EEO.

We provide examples of EEO statements and their corresponding EEO sentiment score rated by the VADER algorithm in Appendix Table A2. Take the EEO statement of Shell plc as an example:

Building on our core values, we aspire to sustain a diverse and inclusive culture where everyone feels respected and valued, from our employees to our customers and partners. A diverse workplace and an inclusive work environment are vital to our success, leading to greater innovation and better energy solutions.

The EEO sentiment score of this warm and genuine statement is high, at 0.973. In contrast, the EEO statement of Tyson Foods is plain and perfunctory, the EEO sentiment score of which is 0:

Tyson is an Equal Opportunity/Affirmative Action Employer. All qualified applicants will be considered without regard to race, national origin, color, religion, age, genetics, sex, sexual orientation, gender identity, disability, or veteran status.

When a supplier firm has multiple principal customers, we compute the lagged customer EEO score of the supplier firm i ( ${Cus EEO Score}_{i, t - 1}$ ) as the sales-weighted average of lagged customer EEO scores.

{C u s E E O S c o r e}_{i, t - 1} = \frac{\sum_{c \in i, t} {E E O S c o r e}_{c, t - 1} \times {S a l e s}_{i, c, t}}{\sum_{c \in i, t} {S a l e s}_{i, c, t}},

where

{EEO Score}_{c, t - 1}

is the EEO score of the principal customer c in year

t - 1

{Sales}_{i, c, t}

is the sales of the supplier i to the principal customer c in the year t.

3.2. Supply Chain Data and Sample Construction

We obtain the supply chain relationship data from the Compustat Customer Segment database, covering the period 1976–2019. Securities and Exchange Commission Regulation S-K requires all US public firms to disclose their principal customers representing more than 10% of the firm’s total sales.²⁰

Since the suppliers only report the names of their principal customers, the unique firm identifiers (i.e., GVKEYs) for customer firms are missing from the Compustat Customer Segment database. Suppliers may report various versions of customer names, including abbreviations, misspellings, and subsidiary names. To circumvent the data challenge, we use the Compustat customer segment company names-GVKEY link from Wharton Research Data Service (provided by Cen et al. (2017)) to match the names of principal customers with the unique GVKEYs in the Compustat database.

To construct the sample for our empirical analysis, we merge the Compustat Customer Segment File with the Center for Research in Security Prices (CRSP)-Compustat Merged File to obtain financial variables for both supplier and customer firms. The supply chain relationship data are merged with BGT data for the workplace EEO scores of suppliers and customers. Since the BGT data set spans the period 2010–2020, the lagged $EEO Score$ of customers is available for the years 2011–2020. Finally, following previous literature (e.g., Chang et al., 2022; Jain and Wu, 2023; Haw et al., 2023), we exclude suppliers and customers in the financial industries (Standard Industrial Classification [SIC] codes 6000–6999) and regulated industries (SIC codes 4900–4999). We remove financial industries mainly because many control variables for financial firms and nonfinancial firms are not comparable.²¹ We remove regulated industries from our sample since they face a significantly larger exposure to government regulations than firms in other industries.²² Our selection criterion above yields a sample of 5,106 supplier-year observations for the period 2011 to 2019, which covers 1,262 unique suppliers. We report our sample selection procedures in Appendix Table A3.

3.3. Summary Statistics

We present the summary statistics for the $EEO Scores$ and financial variables for suppliers in our sample in Table 1. Panel A of Table 1 shows that dependent suppliers have an average EEO score of 0.494. The principal customers have a higher average EEO score than the dependent suppliers. The average $Cus EEO Score$ is 20.9% higher than that of dependent suppliers, which is consistent with the notion that principal customers, which are typically large and reputable firms, have adopted better workplace equal opportunity policies than suppliers (e.g., Dai et al., 2021; Schiller, 2018).

Table 1.
Summary statistics.

Panel A. Summary statistics for main variables

Variables Mean $S D$ P25 Median P75 $N$

${E E O S c o r e}_{t}$ 0.494 0.294 0.149 0.484 0.761 5106

${C u s E E O S c o r e}_{t - 1}$ 0.597 0.189 0.486 0.612 0.719 5106

${I O}_{t}$ 0.693 2.275 0.446 0.746 0.903 5106

${L o g A s s e t}_{t}$ 6.782 1.851 5.512 6.745 8.043 5106

${T o b i n Q}_{t}$ 2.212 1.607 1.210 1.694 2.552 5106

${L e v e r a g e}_{t}$ 0.523 0.278 0.318 0.505 0.679 5106

${R O A}_{t}$ −0.046 0.227 −0.078 0.025 0.071 5106

${S a l e s G r o w t h}_{t}$ 0.168 0.610 −0.034 0.061 0.189 5106

Panel B. Autocorrelation of customer EEO scores

Cus EEO Score _t-1 Cus EEO Score _t-2 Cus EEO Score _t-3

${C u s E E O S c o r e}_{t - 1}$ 1

${C u s E E O S c o r e}_{t - 2}$ 0.521 1

(0.000)

${C u s E E O S c o r e}_{t - 3}$ 0.317 0.524 1

(0.000) (0.000)

Panel C. Pairwise correlation of independent variables

Cus EEO Score _t-1 IO _t LogAsset _t TobinQ _t Leverage _t ROA _t SalesGrowth _t

${C u s E E O S c o r e}_{t - 1}$ 1

${I O}_{t}$ −0.009 1

(0.457)

${L o g A s s e t}_{t}$ −0.040 0.068 1

(0.001) (0.000)

${T o b i n Q}_{t}$ 0.064 −0.001 −0.149 1

(0.000) (0.943) (0.000)

${L e v e r a g e}_{t}$ −0.013 0.011 0.206 0.048 1

(0.291) (0.370) (0.000) (0.000)

${R O A}_{t}$ −0.020 0.030 0.432 −0.163 −0.174 1

(0.094) (0.017) (0.000) (0.000) (0.000)

${S a l e s G r o w t h}_{t}$ 0.013 −0.013 −0.077 0.210 −0.025 −0.049 1

(0.296) (0.302) (0.000) (0.000) (0.039) (0.000)

Panel A. Summary statistics for main variables
${E E O S c o r e}_{t}$	0.494	0.294	0.149	0.484	0.761	5106
${C u s E E O S c o r e}_{t - 1}$	0.597	0.189	0.486	0.612	0.719	5106
${I O}_{t}$	0.693	2.275	0.446	0.746	0.903	5106
${L o g A s s e t}_{t}$	6.782	1.851	5.512	6.745	8.043	5106
${T o b i n Q}_{t}$	2.212	1.607	1.210	1.694	2.552	5106
${L e v e r a g e}_{t}$	0.523	0.278	0.318	0.505	0.679	5106
${R O A}_{t}$	−0.046	0.227	−0.078	0.025	0.071	5106
${S a l e s G r o w t h}_{t}$	0.168	0.610	−0.034	0.061	0.189	5106
Panel B. Autocorrelation of customer EEO scores
	Cus EEO Score _t-1	Cus EEO Score _t-2	Cus EEO Score _t-3
${C u s E E O S c o r e}_{t - 1}$	1
${C u s E E O S c o r e}_{t - 2}$	0.521	1
	(0.000)
${C u s E E O S c o r e}_{t - 3}$	0.317	0.524	1
	(0.000)	(0.000)
Panel C. Pairwise correlation of independent variables
	Cus EEO Score _t-1	IO _t	LogAsset _t	TobinQ _t	Leverage _t	ROA _t	SalesGrowth _t
${C u s E E O S c o r e}_{t - 1}$	1
${I O}_{t}$	−0.009	1
	(0.457)
${L o g A s s e t}_{t}$	−0.040	0.068	1
	(0.001)	(0.000)
${T o b i n Q}_{t}$	0.064	−0.001	−0.149	1
	(0.000)	(0.943)	(0.000)
${L e v e r a g e}_{t}$	−0.013	0.011	0.206	0.048	1
	(0.291)	(0.370)	(0.000)	(0.000)
${R O A}_{t}$	−0.020	0.030	0.432	−0.163	−0.174	1
	(0.094)	(0.017)	(0.000)	(0.000)	(0.000)
${S a l e s G r o w t h}_{t}$	0.013	−0.013	−0.077	0.210	−0.025	−0.049	1
	(0.296)	(0.302)	(0.000)	(0.000)	(0.039)	(0.000)

Note. EEO = equal employment opportunity; $S D$ = standard deviation.

We report the autocorrelations of $Cus EEO Score$ in Panel B of Table 1. The one-lag autocorrelation is 0.521, suggesting that the EEO measure is persistent but still contains a considerable amount of time-varying components. The two-lag autocorrelation is 0.317, which is consistent with the anecdotal evidence suggesting that workplace equality practices in US firms have undergone significant changes over the past decade.

Panel C of Table 1 presents pairwise correlations among independent variables. The pairwise correlations are generally small, ranging from −0.174 to 0.432. The low correlation coefficients between independent variables indicate that multicollinearity is not likely a major concern in our study.

4. Empirical Results

4.1. Baseline Tests

Our baseline sample is organized at the supplier-year level. Specifically, we regress the contemporaneous EEO score of supplier firms ( ${EEO Score}_{i, t}$ ) on the lagged EEO score of their principal customer firms ( ${Cus EEO Score}_{i, t - 1}$ ) with several contemporaneous controls for supplier firm characteristics. Our test specification is outlined in equation (1):

\begin{aligned} {E E O S c o r e}_{i, t} & = α + β \times {C u s E E O S c o r e}_{i, t - 1} + {C o n t r o l s}_{i, t} \\ + κ_{i} + γ_{t} + ϵ_{i, t}, \end{aligned}

(1)

where i denotes supplier and t denotes year.

Cus EEO {Score}_{i, t - 1}

is the sales-weighted average of customer EEO scores for the supplier i in year

t - 1

{Controls}_{i, t}

represents the contemporaneous control variables of supplier firms. Following Dai et al. (2021), we control for the commonly employed financial variables of suppliers, which include the fraction of shares held by institutional investors (

IO

), the natural logarithm of total assets (

LogAsset

), the ratio of the market value of total assets to the book value of total assets (

Tobin Q

), total debt scaled by total assets (

Leverage

), net income scaled by total assets (

ROA

), and one-year growth rate of total sales (

SalesGrowth

). These independent variables allow us to control for known correlations between ESG-related (or Corporate Social Responsibility-related) variables and common firm characteristics, including institutional ownership (e.g., Buchanan et al., 2018; Chen et al., 2020), financial leverage (e.g., Mishra and Modi, 2013; Hong et al., 2012), profitability (e.g., Lins et al., 2017; Liang and Renneboog, 2017), and firm value (Tobin’s Q) (e.g., Gao and Zhang, 2015; Albuquerque et al., 2019).²³ In addition to these control variables, we also include supplier (

κ_{i}

) and year-fixed effects (

γ_{t}

) to control for time-invariant supplier characteristics and common shocks that affect the entire economy. Standard errors, reported in parentheses, are clustered at the supplier industry (SIC4)-year level.

The baseline regression is reported in column (1) of Table 2. The coefficient of ${Cus EEO Score}_{i, t - 1}$ in column (1) (0.048) suggests that a one standard deviation increase in the customer EEO score is correlated with an upward adjustment of 0.009 in the supplier EEO score in the following year, which translates into a 1.8% increase from the mean EEO score of suppliers. Among other control variables, the supplier’s firm size and institutional ownership of suppliers are positively and significantly correlated with supplier EEO scores. This is intuitive as larger companies with higher levels of institutional ownership are likely to pay more attention to workplace equality when facing greater media and investor scrutiny(e.g., Dyck et al., 2019; Chen et al., 2020).²⁴

Table 2.

Diffusion of workplace EEO practice along supply chains.

	(1)	(2)
Variables	${E E O S c o r e}_{t}$	${E E O S c o r e}_{t}$
${C u s E E O S c o r e}_{t - 1}$	0.048^**	0.068^**
	(0.023)	(0.030)
${E S G S c o r e}_{t - 1}$		0.005
		(0.018)
${I O}_{t}$	0.063^**	0.062
	(0.030)	(0.039)
${L o g A s s e t}_{t}$	0.041^***	0.012
	(0.013)	(0.016)
${T o b i n Q}_{t}$	−0.002	−0.004
	(0.005)	(0.006)
${L e v e r a g e}_{t}$	−0.035	−0.044
	(0.029)	(0.037)
${R O A}_{t}$	−0.003	−0.026
	(0.031)	(0.044)
${S a l e s G r o w t h}_{t}$	0.000	0.007
	(0.008)	(0.013)
Supplier firm FE	Yes	Yes
Year FE	Yes	Yes
Observations	4,820	3,164
R-squared	0.544	0.516
Within R-squared	0.009	0.005

Note. EEO = equal employment opportunity; FE = fixed effect. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% level, respectively.

Next, we carry out a series of robustness checks of our baseline results reported in column (1). First, we repeat our baseline test with an additional control variable, the (lagged) KLD-MSCI ESG measure of suppliers, into our test specification. We include this measure to verify whether our baseline effect is specific to the diffusion of corporate workplace practice along the supply chain. One possibility that we want to rule out is that the EEO practice is highly correlated with other ESG policies, and, therefore, the baseline result captures the correlation of aggregated ESG policies between customers and suppliers. We report this test in column (2) of Table 2. Our results in column (2) suggest that (1) suppliers’ EEO scores are positively correlated with their aggregated ESG scores, although this correlation is not statistically significant, and (2) including the aggregated ESG measure of suppliers does not weaken our baseline results.

In the second set of tests reported in Online Appendix Table OA1, we show that our baseline results are robust to the inclusion of different combinations of control variables. In the third set of tests reported in Online Appendix Table OA2, we replace the year fixed effects with industry-year fixed effects to control for time-varying industry-specific shocks. The main patterns of our baseline test are unaffected.

It is important to point out that the lagged structure of our baseline specification does not guarantee a causal interpretation. We face three layers of endogeneity concerns. First, we need to investigate whether the EEO diffusion along the supply chain is relationship-specific; for example, the correlation above may be driven by vertical economic links at the industry level. Second, it is possible that the correlation of the EEO scores between supply chain partners is driven by a selection effect; for example, principal customers systematically select suppliers with good EEO practices at the relationship establishment. Third, the correlation of the EEO scores between supply chain partners may be driven by common (and time-varying) economic factors that researchers do not observe (i.e., the “omitted variable” concern). We will address these three endogeneity concerns in the following subsections.

4.2. Placebo Tests

We address the first endogeneity concern by conducting the following placebo tests. We first construct a placebo sample as a benchmark for interpreting our findings. In the placebo sample, we keep the actual customer firm in each supply chain relationship and replace the actual supplier with placebo suppliers. For each supplier, we identify the corresponding placebo suppliers based on product similarity. We obtain product similarity score following Hoberg and Phillips (2016).²⁵ Using the product description in 10-K filings, Hoberg and Phillips (2016) conduct pair-wise textual similarity analysis and calculate pair-wise product similarity scores for US public firms, which measures how similar the products of the firm are to products of every other US public firm. They define the two firms to be product-market competitors if the pair-wise product similarity score between the two firms exceeds the minimum threshold.²⁶ A higher product similarity score between firm i and firm j indicates that firm i and firm j produce more similar products and, henceforth, have a higher propensity to serve the same customer firm. Similar to the formation of our baseline sample, when a placebo supplier has multiple principal customers, we compute the sales-weighted average of customer EEO score using the sales of the corresponding actual supplier.

Table 3 presents the results of our placebo tests. The test specification is exactly the same as that in column (1) of Table 2 where actual suppliers are replaced by placebo suppliers. In columns (1)–(3), each actual supplier is matched with the one, two, or three placebo suppliers that have the highest product similarity to the actual supplier but with no disclosed supply chain relationship with the actual customers.²⁷ From these placebo test results, we can see that the customer EEO score is not related to its placebo suppliers, suggesting the diffusion of EEO practice along supply chains is driven by firm-specific relationships rather than industry-level economic links.

Table 3.
Placebo tests: Diffusion of workforce EEO practice along supply chains.

(1) (2) (3)

Variables ${E E O S c o r e}_{t}$ ${E E O S c o r e}_{t}$ ${E E O S c o r e}_{t}$

${C u s E E O S c o r e}_{t - 1}$ 0.018 0.007 0.014

(0.025) (0.021) (0.021)

${I O}_{t}$ 0.002^* 0.003^* 0.003^*

(0.001) (0.001) (0.001)

${L o g A s s e t}_{t}$ 0.037^* 0.042^* 0.046^*

(0.013) (0.010) (0.009)

${T o b i n Q}_{t}$ −0.007* −0.002 −0.002

(0.004) (0.003) (0.003)

${L e v e r a g e}_{t}$ 0.011 0.019 −0.008

(0.030) (0.024) (0.021)

${R O A}_{t}$ 0.006 0.003 −0.026

(0.037) (0.031) (0.026)

${S a l e s G r o w t h}_{t}$ −0.012* −0.010^** −0.003

(0.007) (0.005) (0.006)

Supplier firm FE Yes Yes Yes

Year FE Yes Yes Yes

Observations 4,124 6,167 7,514

R-squared 0.571 0.554 0.546

Within R-squared 0.008 0.006 0.007

	(1)	(2)	(3)
${C u s E E O S c o r e}_{t - 1}$	0.018	0.007	0.014
	(0.025)	(0.021)	(0.021)
${I O}_{t}$	0.002^***	0.003^***	0.003^***
	(0.001)	(0.001)	(0.001)
${L o g A s s e t}_{t}$	0.037^***	0.042^***	0.046^***
	(0.013)	(0.010)	(0.009)
${T o b i n Q}_{t}$	−0.007*	−0.002	−0.002
	(0.004)	(0.003)	(0.003)
${L e v e r a g e}_{t}$	0.011	0.019	−0.008
	(0.030)	(0.024)	(0.021)
${R O A}_{t}$	0.006	0.003	−0.026
	(0.037)	(0.031)	(0.026)
${S a l e s G r o w t h}_{t}$	−0.012*	−0.010^**	−0.003
	(0.007)	(0.005)	(0.006)
Supplier firm FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Observations	4,124	6,167	7,514
R-squared	0.571	0.554	0.546
Within R-squared	0.008	0.006	0.007

Note. EEO = equal employment opportunity; FE = fixed effects. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% level, respectively.

A potential concern with the above placebo tests is that the results may be specific to the matching technique based on product similarity that we use to identify placebo suppliers. To address this concern, we further carry out a simulation test using 1,000 benchmark samples as follows: First, for each actual customer, we randomly select a simulated supplier from the same SIC2 industry that the actual supplier belongs to. We only require that the simulated supplier has no direct economic link with the actual customer. Second, we form a benchmark sample of simulated relationships using the actual customers and their simulated suppliers. When a simulated supplier has multiple principal customers, we aggregate the sample to the supplier-year level by computing the sales-weighted average of customer EEO scores. Third, we repeat this process 1,000 times to form 1,000 simulated benchmark samples. Finally, we reestimate our baseline specification for all 1,000 simulated benchmark samples and generate 1,000 coefficient estimates with corresponding $t$ -statistics.

The main advantage of this simulation exercise is that we do not rely on any specific choices of models or parameters. Furthermore, given the large number of simulations, it is unlikely that the coefficient in the actual relationship sample can “beat” most coefficients based on simulated benchmark samples by luck. Essentially, this technique allows us to see where the coefficient ( $t$ -statistics) of ${Cus EEO Score}_{i, t - 1}$ in our baseline test would be while it is placed into the distribution of 1,000 coefficients ( $t$ -statistics) based on simulated benchmark samples. The vertical dashed line in panel A of Figure 1 represents the coefficient estimate of 0.048 in our baseline regression. Panel A of Figure 1 shows that only 0.1% coefficients of ${Cus EEO Score}_{i, t - 1}$ in our 1,000 simulations are able to beat the coefficient in the actual supply chain relationship sample. Panel B of Figure 1 provides the distribution of $t$ -statistics corresponding to these 1,000 coefficients. The distribution shows that only 1.2% of these 1,000 coefficients from simulations are positive and statistically significant at the 5% level (i.e., $t$ -statistics are $\geq$ 1.96). The vertical dashed line in panel B of Figure 1 represents the $t$ -statistics for the actual coefficient (2.09) in our baseline regression reported in column (1) of Table 2. Only 1% of $t$ -statistics from the 1,000 simulations are >2.09. These results, once again, confirm that (1) the diffusion of EEO practice along supply chains is not due to industry-level links or any mechanical correlations driven by other economic factors and (2) the pattern in our placebo tests is not specific to certain matching techniques for identifying placebo suppliers.

Figure 1.

The distribution of 1,000 coefficient estimates and $t$ -statistics from simulation. (A) The distribution of 1,000 coefficient estimates from simulation. (B) The distribution of 1,000 $t$ -statistics from simulation.

4.3. Other Endogeneity Issues

The placebo test shows that our baseline result is not driven by industry-level economic links. However, this test does not address all endogeneity concerns, especially the selection issue and any unobserved common factors that affect the EEO practices of customer and supplier firms simultaneously.

We first run a selection test to investigate whether suppliers with better EEO practices are more likely to be selected by principal customers. The results of this test are reported in Online Appendix Table OA3. Specifically, for any recently added supplier firms in year t, we identify all comparable firms in the same SIC2 industry in year $t - 1$ that were not dependent suppliers of any firms and we test which firm characteristics would affect the choices of customer firms. Our results suggest that several firm characteristics, including firm size, institutional ownership, and leverage, affect a firm’s likelihood of being selected by principal customer firms. However, we find no evidence supporting the conjecture that, in our sample period, EEO practice is among the major criteria for supplier selection.²⁸

Next, we formally address the remaining endogeneity concerns by exploiting an identification strategy based on the Big Three Diversity Campaign in 2017. The “Big Three” refers to three major institutional investors, including BlackRock, Vanguard, and State Street. The Big Three collectively represent 75% of indexed funds. In nearly 90% of S&P 500 companies, either Blackrock, Vanguard, or State Street holds the largest stake.²⁹ Moreover, the Big Three have disproportionately high voting power as not all investors exercise their voting rights, making them “the most important players in corporate America.”³⁰ On International Women’s Day (March 8) of 2017, State Street launched its “Fearless Girls” campaign to encourage firms to add female directors. This initiative was subsequently echoed and supported by BlackRock in July and by Vanguard in August 2017. After launching this initiative, the Big Three began to vote against firms with lower gender diversity. The Big Three’s diversity campaign had a substantial impact on boards’ gender diversity (Gormley et al., 2023).

In this identification strategy, we define the “affected” (i.e., treated) firms as those whose percentage of outstanding shares held by the “Big Three” institutions is higher than the sample average but for which the number of female directors is lower than the sample average. The firms in the benchmark sample either have low institutional ownership by the Big Three (i.e., so that the “Big Three” institutions do not have sufficient stakes to discipline them) or a high level of female directors (i.e., so that the “Big Three” institutions have no reasons to discipline them). Based on this identification strategy, we estimate the following difference-in-differences regression:

\begin{aligned} {E E O S c o r e}_{i, t} & = α + β_{1} \times {C u s A f f e c t e d}_{i, t} \times {P o s t}_{t} \\ + β_{2} \times {C u s A f f e c t e d}_{i, t} \\ + {C o n t r o l s}_{i, t} + κ_{i} + γ_{t} + ϵ_{i, t}, \end{aligned}

(2)

where i denotes supplier and t denotes year.

{Cus Affected}_{i, t}

is the number of the principal customers of the supplier i that are affected by the Big Three’s gender diversity campaign.

{Post}_{t}

is a dummy variable that is equal to 1 for the years 2017 and onwards.³¹

{Controls}_{i, t}

includes all control variables in the baseline regression. In some tests, we also exclude the observations for which the suppliers are directly affected by the campaign themselves. Other test specifications are identical to those in column (1) of Table 2.

Results in Table 4 suggest that the EEO policies propagate along the supply chain after the exogenous increase in customers’ EEO levels. Suppliers with customers affected by the 2017 Big Three Diversity Campaign experienced a larger increase in EEO scores after 2017 relative to other supplier firms. The effect is sizable in economic magnitude: the additional increase in the EEO scores of suppliers with one more affected customer relative to their peers is 0.07, which represents 14.2% of the average EEO score in the full sample. The results are robust to excluding suppliers that are directly affected by the Big Three Diversity Campaign themselves (column (2) of Table 4). We also show in Online Appendix Table OA4 that our results are robust to including industry-year fixed effects. Overall, our results based on the 2017 Big Three Diversity Campaign are consistent with the causal claim that the workplace equality practices of customers affect subsequent workplace equality levels of their suppliers.

Table 4.

Exogenous shock to customers’ workforce equality: 2017 Big Three gender diversity campaign.

	(1)	(2)
	All suppliers	Unaffected suppliers
Variables	${E E O S c o r e}_{t}$	${E E O S c o r e}_{t}$
${C u s A f f e c t}_{t} \times {P o s t}_{t}$	0.070^**	0.085*
	(0.035)	(0.044)
${C u s A f f e c t}_{t}$	0.011	0.005
	(0.013)	(0.015)
${I O}_{t}$	0.061*	0.050
	(0.031)	(0.033)
${L o g A s s e t}_{t}$	0.048^***	0.052^***
	(0.014)	(0.018)
${T o b i n Q}_{t}$	0.001	−0.000
	(0.005)	(0.007)
${L e v e r a g e}_{t}$	−0.041	−0.063
	(0.035)	(0.041)
${R O A}_{t}$	0.003	−0.022
	(0.038)	(0.048)
${S a l e s G r o w t h}_{t}$	−0.004	−0.000
	(0.009)	(0.015)
Supplier firm FE	Yes	Yes
Year FE	Yes	Yes
Observations	4,280	3,074
R-squared	0.542	0.582
Within R-squared	0.011	0.012

Note. EEO = equal employment opportunity; FE = fixed effects. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% level, respectively.

5. Mechanism Analysis

Our main point is that suppliers, in order to maintain and strengthen relationships with their principal customers, adjust their workplace equal opportunity policies to cater to the preferences of their principal customers. Implicitly, principal customer firms play a disciplinary role while they pay attention to the labor-related risks along supply chains. The effectiveness of this implicit disciplinary role depends on the bargaining power of principal customers with their dependent suppliers.³²

We proxy customers’ bargaining power with their suppliers by customer switching cost, that is, how costly it is for a customer to replace its existing supplier with a “substitute” supplier. Following Hoberg and Phillips (2016), we construct two measures for the customer switching cost: the number of product-market competitors and product similarity in the supplier industry. The number of product market competitors is closely related to the market share of the firm: the more product competitors a firm has, the less likely the firm occupies a large market share. Intuitively, we argue that when a supplier firm belongs to an industry with a larger number of competitors or a higher level of product similarity, its customer firms face a lower cost to replace the supplier firm and, therefore, the supplier is more likely to adjust its workplace EEO policies to cater to the preferences or requirements of its customers.

Results of cross-sectional partitions based on the customer switching cost are reported in Table 5. In this test, a supplier belongs to the HighRival (LowRival) group when the average number of product rivals of the supplier ranks above (below) the median for all suppliers. Similarly, a supplier belongs to the HighProductSim (LowProductSim) group when the average product similarity of the supplier ranks above (below) the median for suppliers. We carry out the baseline test (i.e., column (1) of Table 2) separately for two subsamples and test the statistical difference between the main coefficients of interest (i.e., the coefficients of

{Cus EEO Score}_{i, t - 1}

) from these subsamples.

Table 5.
Diffusion of workplace EEO practice along supply chains: Cross-sectional partitions based on customer switching cost.

	(1)	(2)	(3)	(4)
	HighRival	LowRival	HighProductSim	LowProductSim
Variables	${E E O S c o r e}_{t}$	${E E O S c o r e}_{t}$	${E E O S c o r e}_{t}$	${E E O S c o r e}_{t}$
${C u s E E O S c o r e}_{t - 1}$	0.084^**	−0.038	0.083^**	−0.037
	(0.034)	(0.037)	(0.033)	(0.038)
${I O}_{t}$	0.024	0.088^**	0.018	0.120^***
	(0.044)	(0.042)	(0.043)	(0.042)
${L o g A s s e t}_{t}$	0.041^***	0.019	0.045^***	0.024
	(0.015)	(0.022)	(0.015)	(0.023)
${T o b i n Q}_{t}$	−0.005	−0.001	−0.007	0.004
	(0.006)	(0.011)	(0.006)	(0.010)
${L e v e r a g e}_{t}$	0.003	−0.131^**	−0.013	−0.087
	(0.033)	(0.057)	(0.034)	(0.054)
${R O A}_{t}$	−0.009	0.007	−0.026	−0.000
	(0.035)	(0.072)	(0.037)	(0.075)
${S a l e s G r o w t h}_{t}$	−0.006	0.046^***	−0.004	0.027*
	(0.009)	(0.017)	(0.009)	(0.015)
Difference	0.122^***		0.120^***
Supplier firm FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Observations	2,327	2,338	2,342	2,330
R-squared	0.560	0.554	0.566	0.564
Within R-squared	0.010	0.013	0.012	0.011

Note. EEO = equal employment opportunity; FE = fixed effects. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% level, respectively.

Results in Table 5 are consistent with our conjecture. Specifically, we have a much stronger lead–lag pattern in workplace EEO policies between customers and suppliers in the $HighRival$ ( $HighProductSim$ ) group, relative to that of the $LowRival$ ( $LowProductSim$ ) group. The coefficients of ${Cus EEO Score}_{i, t - 1}$ in the $HighRival$ and $HighProductSim$ groups are positive and statistically significant at the 5% level, while the coefficients of the $LowRival$ and $LowProductSim$ groups are statistically not different from zero. The statistical differences between coefficients of ${Cus EEO Score}_{i, t - 1}$ in both partitions are statistically significant at the 1% level. Overall, our results support the conjecture that when customers are able to exercise stronger bargaining power, their suppliers are more likely to adjust their EEO policies to cater to the preferences of the customers.

6. Real Economic Outcomes: Retention of Female Employees

In previous sections, we have shown the EEO diffusion along the supply chain and established the causality that suppliers adopt EEO policies to cater to their principal customers’ preferences. In this section, we focus on the real economic outcomes of EEO diffusion along the supply chain. Specifically, we examine whether the diffusion of workplace EEO practices from customers to suppliers affects how suppliers retain female workers. This test is motivated by anecdotal evidence that the effects of workplace equality practices are more pronounced in retaining female employees.³³

In this test, we retrieve the career histories of male and female workers from the Lightcast US Profile database. In this database, we are able to see the duration (i.e., starting and ending year months) of each individual job for both male and female workers. Our main conjecture is that, after controlling for the contemporaneous EEO level of supplier firms themselves, the lagged EEO level of customers would still affect how long female workers would be retained in the workforce of suppliers. We test this conjecture with the following model:

\begin{aligned} {L o g F e m a l e R T}_{i, t} & = α + β_{1} \times {E E O S c o r e}_{i, t} \\ + β_{2} \times {C u s E E O S c o r e}_{i, t - 1} \\ + {C o n t r o l s}_{i, t} + κ_{i} + γ_{t} + ϵ_{i, t}, \end{aligned}

(3)

where

{L o g F e m a l e R T}_{i, t}

is the natural logarithm of the average career duration of female employees in the supplier i in year t.

\begin{aligned} {L o g F e m a l e R T}_{i, t} = L o g (\frac{\sum_{p \in n} {R e t e n t i o n}_{p, i, t}}{{N u m b e r o f F e m a l e E m p l o y e e}_{i, t}}), \end{aligned}

(4)

where i denotes the supplier firm, p denotes the female employee, and t denotes year.

{Retention}_{p, i, t}

is the number of years between the year t and first year that this employee p works in supplier i. We use

{LogMaleRT}_{i, t}

as the dependent variable in placebo tests.

{LogMaleRT}_{i, t}

is similarly constructed to measure the average career duration of male employees in the supplier firm i in year t.

To account for the impact of suppliers’ own EEO practices on the likelihood of retaining female employees, we include the contemporaneous EEO score of the supplier as a control variable. We also control for the natural logarithm of the total number of employees of the supplier ( ${LogEmployee}_{i, t}$ ) in some test specifications. Other control variables are the same as those in Table 2. Standard errors, reported in parentheses, are clustered at the supplier industry (SIC4)-year level.

Table 6 presents the results for the effect of workplace EEO diffusion on the career duration of female and male employees in supplier firms. The coefficient of ${EEO Score}_{t}$ is positive and statistically significant at the 5% level in column (1), indicating that firms with higher EEO scores are likely to retain female employees for a longer duration. A one standard deviation increase in EEO score is associated with a 1.3% increase in the average years of female employee retention. In comparison, the coefficient of ${EEO Score}_{t}$ is insignificant in column (2), suggesting that workplace equality practices do not have significant effects on retaining male workers.

Table 6.

Customer EEO diffusion and supplier employee retention.

	(1)	(2)
Variables	${L o g F e m a l e R T}_{t}$	${L o g M a l e R T}_{t}$
${E E O S c o r e}_{t}$	0.043^**	0.015
	(0.019)	(0.018)
${C u s E E O S c o r e}_{t - 1}$	0.049*	0.040
	(0.030)	(0.026)
${I O}_{t}$	0.079^**	0.075^**
	(0.032)	(0.031)
${L o g A s s e t}_{t}$	−0.087^***	−0.058^***
	(0.016)	(0.014)
${T o b i n Q}_{t}$	−0.017^***	−0.004
	(0.007)	(0.006)
${L e v e r a g e}_{t}$	0.025	−0.034
	(0.040)	(0.033)
${R O A}_{t}$	−0.017	−0.047
	(0.043)	(0.039)
${S a l e s G r o w t h}_{t}$	−0.018	−0.022^**
	(0.012)	(0.011)
Supplier firm FE	Yes	Yes
Year FE	Yes	Yes
Observations	4,023	4,131
R-squared	0.877	0.891
Within R-squared	0.024	0.017

Note. EEO = equal employment opportunity; FE = fixed effects. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% level, respectively.

Our main interest lies in the coefficients of $Cus EEO {Score}_{t - 1}$ , which measure the effect of lagged customer EEO score on suppliers’ female worker retention (after controlling for the EEO practices of suppliers themselves). The result in column (1) shows that a one standard deviation increase in customer EEO score is associated with a 1% $(0.049 \times 0.189)$ increase in the average number of years of female employee retention of the supplier. Similar to the results in column (1), customers’ lagged EEO scores do not have a significant effect on male employee retention of the suppliers.

As shown in Online Appendix Table OA5, our results are robust to including supplier industry-year fixed effects. Our results are also robust to different model specifications.³⁴ One caveat here is that we do not intend to claim causality in our tests of real economic outcomes in Section 6. Instead, we suggest that our results are consistent with the notion that the diffusion of EEO policies along the supply chain may generate real economic outcomes, particularly for the dependent suppliers.

7. Conclusion

EEO policies play a crucial role in fostering workplace DEI. Firms are increasingly recognizing the importance of integrating EEO initiatives into their business models to build a sustainable competitive advantage in the marketplace. Despite the growing recognition of EEO’s importance, its influence on supply chains remains largely unexplored as an empirical question. We use a unique dataset that combines job posting textual analyses and supply chain information, to illuminate how EEO practices propagate throughout supply chain relationships.

Our research reveals a compelling pattern of EEO diffusion (cascade) in the supply chain. That is, the EEO policies of downstream companies affect the EEO practices of their upstream suppliers. This effect is amplified when the bargaining power of downstream companies is greater. Furthermore, we find that the diffusion of EEO policies along the supply chain generates important economic outcomes. Specifically, the diffusion of good EEO policies from customers to suppliers helps suppliers retain female employees. In conclusion, our study demonstrates that EEO diffusion along the supply chain is an important economic dynamic to ensure “doing well by doing good” in sustainable supply chain management.

While the significance of EEO transcends regional or industry boundaries, our findings regarding the EEO diffusion along supply chains may have important policy implications. That said, an increase in a company’s EEO score could have ripple effects across the entire supply chain network. Therefore, forward-thinking companies that prioritize sustainable development can enhance overall social welfare and supply chain performance by promoting EEO policies. Policymakers should also be aware that EEO practices can interact with and spill over into supply chains and generate greater positive social returns. For companies that have not fully implemented EEO policies, our results show compelling evidence that adopting such policies is not only socially responsible but also supported by supply chain economic incentives, warranting strategic considerations.

Supplemental Material

sj-pdf-1-pao-10.1177_10591478241238968 - Supplemental material for Equal Employment Opportunity in Supply Chains

Supplemental material, sj-pdf-1-pao-10.1177_10591478241238968 for Equal Employment Opportunity in Supply Chains by Ling Cen, Yanru Han and Jing Wu in Production and Operations Management

Footnotes

Appendix

Table A1.

Variable definition.

Variable	Definition
Dependent variables
${E E O S c o r e}_{t}$	EEO score of supplier s in year t.
	EEO score is constructed using EEO statements from online job postings.
${L o g F e m a l e R e t e n t i o n}_{t}$	The natural logarithm of the average number of years of female employee retention of the supplier in year t
${L o g M a l e R e t e n t i o n}_{t}$	The natural logarithm of the average number of years of male employee retention of the supplier in year t
Independent variables
$I O$	Institutional ownership as measured by the percentage of shares outstanding held by 13F institutional investors.
${L n A s s e t}_{t}$	The natural logarithm of total assets of the firm.

Variable	Definition
${T o b i n Q}_{t}$	Market value of equity divided by book value of equity.
${L e v e r a g e}_{t}$	Total liability divided by total assets.
$R O A$	Returns on assets.
${S a l e s G r o w t h}_{t}$	The growth rate of sales.
${C u s A f f e c t}_{t}$	The number of the principal customers that are affected by the Big Three’s gender diversity campaign.The firm is affected by the campaign if its Big Three ownership is higher than the industry-year medianBig Three ownership at the end of 2016 and its board female ratio is lower than the industry-year median board female ratio in year t.
$L o g E m p l o y e e$	The natural logarithm of the total number of employees of the firm in the year.
${N s u i t}_{t}$	The number of employment discrimination lawsuits against the firm in the year.
$D u m m y ({N s u i t}_{t})$	A dummy variable that is equal to 1 if the number of employment discrimination lawsuits against the firm in the year is positive.
$L o g (1 + {N s u i t}_{t})$	The natural logarithm of 1 plus the number of employment discrimination lawsuits against the firm in the year.
Moderating variables
$H i g h R i v a l$	A dummy variable that is equal to 1 if the average number of product rivals of the supplier within its relationship duration ranks above https://hobergphillips.tuck.dartmouth.edu/industryclass.htm following Hoberg and Phillips (2016).
$L o w R i v a l$	A dummy variable that is equal to 1 if the average number of product rivals of the supplier within its relationship duration ranks below the median among all suppliers.
$H i g h P r o d u c t S i m$	A dummy variable that is equal to 1 if the average product similarity of the supplier within its relationship duration ranks above the median among all suppliers. The product similarity score is retrieved from https://hobergphillips.tuck.dartmouth.edu/industryclass.htm following Hoberg and Phillips (2016).
$L o w P r o d u c t S i m$	A dummy variable that is equal to 1 if the average product similarity of the supplier within its relationship duration ranks below the median among all suppliers. The product similarity score is retrieved from https://hobergphillips.tuck.dartmouth.edu/industryclass.htm following Hoberg and Phillips (2016).

Note. EEO = equal employment opportunity.

Table A2.

Examples of EEO statements.

EEO statement	Company	EEO score
Building on our core values, we aspire to sustain a diverse and inclusive culture where everyone feels respected and valued, from our employees to our customers and partners. A diverse workforce and an inclusive work environment are vital to our success, leading to greater innovation and better energy solutions.	Shell	0.9732
Unilever is an organization committed to diversity and inclusion to drive our business results and create a better future every day for our diverse employees, global consumers, partners, and communities. We believe a diverse workforce allows us to match our growth ambitions and drive inclusion across the business. All qualified applicants will receive consideration for employment without regard to race, color, religion, sex, sexual orientation, gender identity, age, national origin, or protected veteran status and will not be discriminated against on the basis of disability. For more information, please see Equal Employment Opportunity Posters	Unilever	0.943
NVIDIA is committed to fostering a diverse work environment and proud to be an equal-opportunity employer. As we highly value diversity in our current and future employees, we do not discriminate (including in our hiring and promotion practices) on the basis of race, religion, color, national origin, gender, gender expression, sexual orientation, age, marital status, veteran status, disability status or any other characteristic protected by law.	NVIDIA	0.9117
It is the policy of The PNC Financial Services Group, Inc. (“PNC”), in compliance with the law, to afford equal opportunity to all applicants and existing employees without regard to race, color, religion, national origin, gender, sexual orientation, gender identity/expression, age, ancestry, marital status, genetic information, family medical history, disability, protected veteran status, or any other basis that would be in violation of any applicable ordinance or law.	PNC	0.3612
HCSC is committed to maintaining a work environment that is free from any and all forms of unlawful discrimination and harassment.	HCSC	0.2263
Tyson is an Equal Opportunity/Affirmative Action Employer. All qualified applicants will be considered without regard to race, national origin, color, religion, age, genetics, sex, sexual orientation, gender identity, disability or veteran status.	Tyson Foods	0
Agilysys is an equal-opportunity employer. In compliance with Federal and State EEO laws, qualified applicants are considered for all positions without regard to race, color, gender, religion, national origin, ancestry, place of birth, age, marital status, sexual orientation, disability, or veteran status.	Agilysys Inc.	0
Five Star prohibits discrimination against any employee or applicant for employment with regard to race, color, religion, sex, sexual orientation, gender identity or expression, national origin, age, marital status, genetic information, disability, or veteran status.	Five Star	0

Note. EEO = equal employment opportunity; VADER = valence aware dictionary and sentiment reasoner. This table presents examples of EEO statements and their EEO sentiment scores from the VADER algorithm.

Table A3.

Sample selection.

Descriptions	Firm-year observations	Unique firms
Supplier firm-year observations in Compustat Segment Files from 2011 to 2019 with at least one nonmissing customer GVKEY.	11,468	2,381
Less supplier firm-year observations:
With missing CRSP-Compustat merged data	(2,226)	(399)
With missing control variables	(841)	(106)
Missing job postings	(2,418)	(394)
Less supplier firm-year observations in regulated industry (SIC: 4900–4999)	(169)	(48)
Less supplier firm-year observations in financial industry (SIC: 6000–6999)	(708)	(172)
Subtotal	5,106	1,262
For regressions with supplier firm fixed-effects:
Less singleton observations	(286)	(286)
Final sample	4,820	976

Note. CRSP= Center for Research in Security Prices; SIC = Standard Industrial Classification.

Declaration of Conflicting Interests

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

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Ling Cen acknowledges the funding support from GRF 14506522 offered by the Hong Kong Research Grants Council. Jing Wu acknowledges the funding support from GRF 14505320 offered by the Hong Kong Research Grants Council.

ORCID iD

Jing Wu

Supplemental Material

Supplemental material for this article is available online ().

Notes

How to cite this article

Cen L, Han Y and Wu J (2025) Equal Employment Opportunity in Supply Chains. Production and Operations Management 34(3): 524–540.

References

Adhikari

Agrawal

Malm

(2019) Do women managers keep firms out of trouble? Evidence from corporate litigation and policies. Journal of Accounting and Economics 67(1): 202–225.

Agca

Babich

Birge

, et al. (2022) Credit shock propagation along supply chains: Evidence from the CDS market. Management Science 68(9): 6506–6538.

Albuquerque

Koskinen

Zhang

(2019) Corporate social responsibility and firm risk: Theory and empirical evidence. Management Science 65(10): 4451–4469.

Banerjee

Dasgupta

Kim

(2008) Buyer–supplier relationships and the stakeholder theory of capital structure. Journal of Finance 63(5): 2507–2552.

Benerjee

Dasgupta

Shi

(2021) The dark side of stakeholder influence: The surprising effect of customer fraud on suppliers. Working Paper.

Bernile

Bhagwat

Yonker

(2018) Board diversity, firm risk, and corporate policies. Journal of Financial Economics 127(3): 588–612.

Bird

Jodi

Michael

(2019) Coupling labor codes of conduct and supplier labor practices: The role of internal structural conditions. Organization Science 30(4): 847–867.

Blanco

(2021) Supply chain carbon footprinting and climate change disclosures of global firms. Production and Operations Management 30(9): 3143–3160.

Blanco

(2022) A classification of carbon abatement opportunities of global firms. Manufacturing & Service Operations Management 24(5): 2648–2665.

10.

Bowen

Cousins

Lamming

, et al. (2001) The role of supply management capabilities in green supply. Production and Operations Management 10(2): 174–189.

11.

Buchanan

Cao

Chen

(2018) Corporate social responsibility, firm value, and influential institutional ownership. Journal of Corporate Finance 52: 73–95.

12.

Buell

Kalkanci

(2021) How transparency into internal and external responsibility initiatives influences consumer choice. Management Science 67(2): 932–950.

13.

Cen

Han

Liu

, et al. (2022) Rise of conscious consumers: Impact of corporate workplace equality on household spending. Working Paper.

14.

Cen

Maydew

Zhang

, et al. (2017) Customer–supplier relationships and corporate tax avoidance. Journal of Financial Economics 123(2): 377–394.

15.

Chang

Liu

Mashruwala

(2022) Customer bargaining power, strategic fit, and supplier performance. Production and Operations Management 31(4): 1492–1509.

16.

Chen

Dong

Lin

(2020) Institutional shareholders and corporate social responsibility. Journal of Financial Economics 135(2): 483–504.

17.

Chu

Tian

Wang

(2019) Corporate innovation along the supply chain. Management Science 65(6): 2445–2466.

18.

Corbett

(2006) Global diffusion of ISO 9000 certification through supply chains. Manufacturing & Service Operations Management 8(4): 330–350.

19.

Corbett

Klassen

(2006) Extending the horizons: Environmental excellence as key to improving operations. Manufacturing & Service Operations Management 8(1): 5–22.

20.

Crook

Combs

(2007) Sources and consequences of bargaining power in supply chains. Journal of Operations Management 25(2): 546–555.

21.

Cumming

Leung

Rui

(2015) Gender diversity and securities fraud. Academy of Management Journal 58(5): 1572–1593.

22.

Dai

Liang

(2021) Socially responsible corporate customers. Journal of Financial Economics 142(2): 598–626.

23.

Delis

Gaganis

Hasan

, et al. (2017) The effect of board directors from countries with different genetic diversity levels on corporate performance. Management Science 63(1): 231–249.

24.

Dezsö

Ross

(2012) Does female representation in top management improve firm performance? A panel data investigation. Strategic Management Journal 33(9): 1072–1089.

25.

Dyck

Lins

Roth

, et al. (2019) Do institutional investors drive corporate social responsibility? International evidence. Journal of Financial Economics 131(3): 693–714.

26.

Erhun

Kraft

Wijnsma

(2021) Sustainable triple-a supply chains. Production and Operations Management 30(3): 644–655.

27.

Gao

Zhang

(2015) Firms’ earnings smoothing, corporate social responsibility, and valuation. Journal of Corporate Finance 32: 108–127.

28.

Giannetti

Burkart

Ellingsen

(2011) What you sell is what you lend? Explaining trade credit contracts. Review of Financial Studies 24(4): 1261–1298.

29.

Gormley

Gupta

Matsa

, et al. (2023) The Big Three and board gender diversity: The effectiveness of shareholder voice. Journal of Financial Economics 149(2): 323–348.

30.

Hartmann

Moeller

(2014) Chain liability in multitier supply chains? Responsibility attributions for unsustainable supplier behavior. Journal of Operations Management 32(5): 281–294.

31.

Haw

Swink

Zhang

(2023) The role of direct equity ownership in supply chains. Journal of Operations Management 69(4): 586–615.

32.

Hershbein

Kahn

(2018) Do recessions accelerate routine-biased technological change? Evidence from vacancy postings. American Economic Review 108(7): 1737–72.

33.

Hertzel

Officer

, et al. (2008) Inter-firm linkages and the wealth effects of financial distress along the supply chain. Journal of Financial Economics 87(2): 374–387.

34.

Hoberg

Phillips

(2016) Text-based network industries and endogenous product differentiation. Journal of Political Economy 124(5): 1423–1465.

35.

Hong

Kubik

Scheinkman

(2012) Financial constraints on corporate goodness. Working Paper.

36.

Huang

Song

Swinney

(2022) Managing social responsibility in multitier supply chains. Manufacturing & Service Operations Management 24(6): 2843–2862.

37.

Hui

Klasa

Yeung

(2012) Corporate suppliers and customers and accounting conservatism. Journal of Accounting and Economics 53(1): 115–135.

38.

Huq

Chowdhury

Klassen

(2016) Social management capabilities of multinational buying firms and their emerging market suppliers: An exploratory study of the clothing industry. Journal of Operations Management 46: 19–37.

39.

Jain

(2023) Can global sourcing strategy predict stock returns? Manufacturing & Service Operations Management 25(4): 1357–1375.

40.

Jira

Toffel

(2013) Engaging supply chains in climate change. Manufacturing & Service Operations Management 15(4): 559–577.

41.

Kalkanci

Plambeck

(2020) Reveal the supplier list? A trade-off in capacity vs. responsibility. Manufacturing & Service Operations Management 22(6): 1251–1267.

42.

Kalkanci

Rahmani

Toktay

(2019) The role of inclusive innovation in promoting social sustainability. Production and Operations Management 28(12): 2960–2982.

43.

Khan

Kalra

(2022) It’s good to be different: How diversity impacts judgments of moral behavior. Journal of Consumer Research 49(2): 177–201.

44.

Kim

Starks

(2016) Gender diversity on corporate boards: Do women contribute unique skills? American Economic Review 106(5): 267–71.

45.

Lee

Klassen

(2008) Drivers and enablers that foster environmental management capabilities in small-and medium-sized suppliers in supply chains. Production and Operations Management 17(6): 573–586.

46.

Liang

Renneboog

(2017) Corporate donations and shareholder value. Oxford Review of Economic Policy 33(2): 278–316.

47.

Lins

Servaes

Tamayo

(2017) Social capital, trust, and firm performance: The value of corporate social responsibility during the financial crisis. Journal of Finance 72(4): 1785–1824.

48.

Matsa

Miller

(2011) Chipping away at the glass ceiling: Gender spillovers in corporate leadership. American Economic Review 101(3): 635–639.

49.

Mishra

Modi

(2013) Positive and negative corporate social responsibility, financial leverage, and idiosyncratic risk. Journal of Business Ethics 117: 431–448.

50.

Morgan

Várdy

(2009) Diversity in the workplace. American Economic Review 99(1): 472–485.

51.

Osadchiy

Schmidt

(2021) The bullwhip effect in supply networks. Management Science 67(10): 6153–6173.

52.

Parmigiani

Klassen

Russo

(2011) Efficiency meets accountability: Performance implications of supply chain configuration, control, and capabilities. Journal of Operations Management 29(3): 212–223.

53.

Porteous

Rammohan

Lee

(2015) Carrots or sticks? Improving social and environmental compliance at suppliers through incentives and penalties. Production and Operations Management 24(9): 1402–1413.

54.

Schiller

(2018) Global supply-chain networks and corporate social responsibility. Working Paper.

55.

Serpa

Krishnan

(2018) The impact of supply chains on firm-level productivity. Management Science 64(2): 511–532.

56.

Sodhi

(2015) Conceptualizing social responsibility in operations via stakeholder resource-based view. Production and Operations Management 24(9): 1375–1389.

57.

Soundararajan

Brammer

(2018) Developing country sub-supplier responses to social sustainability requirements of intermediaries: Exploring the influence of framing on fairness perceptions and reciprocity. Journal of Operations Management 58: 42–58.

58.

Veronica

(2019) The missing link? The strategic role of procurement in building sustainable supply networks. Production and Operations Management 28(5): 1149–1172.

59.

Villena

Gioia

(2018) On the riskiness of lower-tier suppliers: Managing sustainability in supply networks. Journal of Operations Management 64: 65–87.

60.

Wilhelm

Villena

(2021) Cascading sustainability in multi-tier supply chains: When do chinese suppliers adopt sustainable procurement? Production and Operations Management 30(11): 4198–4218.

61.

Wowak

Ball

Post

Ketchen

DJ Jr

(2021) The influence of female directors on product recall decisions. Manufacturing & Service Operations Management 23(4): 895–913.

62.

Zhang

Zhou

(2021) Credit rating prediction through supply chains: A machine learning approach. Production and Operations Management 31(4): 1613–1629.

63.

Zhu

Zhang

Tsung

(2007) Pushing quality improvement along supply chains. Management science 53(3): 421–436.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.18 MB

Panel A. Summary statistics for main variables
Variables	Mean	$S D$	P25	Median	P75	$N$
${E E O S c o r e}_{t}$	0.494	0.294	0.149	0.484	0.761	5106
${C u s E E O S c o r e}_{t - 1}$	0.597	0.189	0.486	0.612	0.719	5106
${I O}_{t}$	0.693	2.275	0.446	0.746	0.903	5106
${L o g A s s e t}_{t}$	6.782	1.851	5.512	6.745	8.043	5106
${T o b i n Q}_{t}$	2.212	1.607	1.210	1.694	2.552	5106
${L e v e r a g e}_{t}$	0.523	0.278	0.318	0.505	0.679	5106
${R O A}_{t}$	−0.046	0.227	−0.078	0.025	0.071	5106
${S a l e s G r o w t h}_{t}$	0.168	0.610	−0.034	0.061	0.189	5106
Panel B. Autocorrelation of customer EEO scores
	Cus EEO Score _t-1	Cus EEO Score _t-2	Cus EEO Score _t-3
${C u s E E O S c o r e}_{t - 1}$	1
${C u s E E O S c o r e}_{t - 2}$	0.521	1
	(0.000)
${C u s E E O S c o r e}_{t - 3}$	0.317	0.524	1
	(0.000)	(0.000)
Panel C. Pairwise correlation of independent variables
	Cus EEO Score _t-1	IO _t	LogAsset _t	TobinQ _t	Leverage _t	ROA _t	SalesGrowth _t
${C u s E E O S c o r e}_{t - 1}$	1
${I O}_{t}$	−0.009	1
	(0.457)
${L o g A s s e t}_{t}$	−0.040	0.068	1
	(0.001)	(0.000)
${T o b i n Q}_{t}$	0.064	−0.001	−0.149	1
	(0.000)	(0.943)	(0.000)
${L e v e r a g e}_{t}$	−0.013	0.011	0.206	0.048	1
	(0.291)	(0.370)	(0.000)	(0.000)
${R O A}_{t}$	−0.020	0.030	0.432	−0.163	−0.174	1
	(0.094)	(0.017)	(0.000)	(0.000)	(0.000)
${S a l e s G r o w t h}_{t}$	0.013	−0.013	−0.077	0.210	−0.025	−0.049	1
	(0.296)	(0.302)	(0.000)	(0.000)	(0.039)	(0.000)

	(1)	(2)	(3)
Variables	${E E O S c o r e}_{t}$	${E E O S c o r e}_{t}$	${E E O S c o r e}_{t}$
${C u s E E O S c o r e}_{t - 1}$	0.018	0.007	0.014
	(0.025)	(0.021)	(0.021)
${I O}_{t}$	0.002^***	0.003^***	0.003^***
	(0.001)	(0.001)	(0.001)
${L o g A s s e t}_{t}$	0.037^***	0.042^***	0.046^***
	(0.013)	(0.010)	(0.009)
${T o b i n Q}_{t}$	−0.007*	−0.002	−0.002
	(0.004)	(0.003)	(0.003)
${L e v e r a g e}_{t}$	0.011	0.019	−0.008
	(0.030)	(0.024)	(0.021)
${R O A}_{t}$	0.006	0.003	−0.026
	(0.037)	(0.031)	(0.026)
${S a l e s G r o w t h}_{t}$	−0.012*	−0.010^**	−0.003
	(0.007)	(0.005)	(0.006)
Supplier firm FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Observations	4,124	6,167	7,514
R-squared	0.571	0.554	0.546
Within R-squared	0.008	0.006	0.007

Equal Employment Opportunity in Supply Chains

Abstract

Keywords

1. Introduction

2. Literature Review

3. Data and Variable Construction

3.1. DEI Measure: EEO Score

3.2. Supply Chain Data and Sample Construction

3.3. Summary Statistics

4.1. Baseline Tests

Table 5. Diffusion of workplace EEO practice along supply chains: Cross-sectional partitions based on customer switching cost.

Supplemental Material

sj-pdf-1-pao-10.1177_10591478241238968 - Supplemental material for Equal Employment Opportunity in Supply Chains

Footnotes

Appendix

Declaration of Conflicting Interests

Funding

ORCID iD

Supplemental Material

Notes

How to cite this article

References

Supplementary Material

Table 5.
Diffusion of workplace EEO practice along supply chains: Cross-sectional partitions based on customer switching cost.