The Real Effects of Conflict Minerals Disclosures

Bok Baik; Omri Even-Tov; Russell Han; David Park

doi:10.2139/ssrn.3908233

Baik, Bok;Even-Tov, Omri;Han, Russell;Park, David

2021-08-19 00:00:00

We examine the real effects of conflict minerals disclosures—mandatory social-responsibility disclosures intended to ameliorate conflicts fueled by mineral resources in the Democratic Republic of Congo (DRC). Section 1502 of the Dodd-Frank Act requires issuers to provide specialized disclosures on conflict minerals—tantalum, tin, tungsten, and gold (3TG)—that indicate whether 3TG are necessary to the functionality or production of their product(s) and whether they originate from mines in the DRC or a neighboring country (i.e., covered countries). We utilize hand-collected data from the conflict minerals reports and find that companies dissociate themselves from sourcing conflict minerals from the covered countries. We also find that the market reacts positively to responsible sourcing. The results imply that companies’ assessments of reputational costs due to greater public attention can lead to responsible sourcing. Finally, we find that after the conflict minerals disclosure rule, conflicts in covered countries’ mining regions decrease relative to those in non-covered countries. Overall, our results suggest that promoting supply chain transparency of conflict minerals alleviates the region’s resource curse. Keywords: Real effects, Dodd-Frank Act, Conflict minerals disclosures, Corporate social responsibility, Responsible sourcing, ESG Acknowledgements: We thank Christian Leuz, David Aboody, Brett Trueman, Woon-Oh Jung, Youngjo Lee, Michael Donohoe, Jong-Hag Choi, Woo-Jong Lee, Iny Hwang, Yong-Gyu Lee, Sunhwa Choi, Pureum Kim, George SheChih Chiu (discussant), Taeho Ko, Chris Bayer, Annie Signorelli (Apple), Sunwoo Hee- Yeon, Young Yoon, Jinsung Hwang, Kimberlyn George, Simmi Mookerjee (discussant), Sunyoung Park, Sun-Moon Jung, Sang-Mok Lee, Stacey Choy, Changri Cui, Jongha Kim, and seminar participants at Seoul National University, the Korean Accounting Association Conference 2021, the Korean Accounting Information Association Conference 2021, the University of Tennessee, and the 2021 American Accounting Association Annual Meeting. We acknowledge Alex Gunwoo Kim, Chanmu Park, Jaeho Yoo, and Sunho Yoo for their valuable research assistance. Bok Baik acknowledges financial support from the Institute of Management Research Seoul National University. Corresponding author. Electronic copy available at: https://ssrn.com/abstract=3908233 1. Introduction The extraction and distribution of Africa’s natural resources have been a long-standing source of conflict (U.N. Security Council, 2010) that bears economic, geopolitical, and humanitarian implications. The region’s possession of vast and valuable mineral reserves is often set against a landscape of heightened political volatility and corruption, weaker institutional structures and accountability, and poor social welfare metrics. Scholars have coined a term for this paradox: the “resource curse” (Auty, 1993; Sachs and Warner, 1995; Auty, 2001; Gylfason, 2001). The curse has been particularly pronounced in the Democratic Republic of Congo (DRC), where over 6 million inhabitants have died as a result of clashes over the country’s abundant natural resources (Business Social Responsibility, 2010). Concerned that Western corporations’ purchases of conflict minerals (i.e., tantalum, tin, tungsten, and gold, also referred to as 3TG) were exacerbating the crisis, in 2010, Congress enacted Section 1502 as part of the Dodd–Frank Wall Street Reform and Consumer Protection Act of 2010 (hereafter referred to as the “Dodd-Frank Act” or “the act”). This provision required Securities and Exchange Commission (SEC)-registered firms to file by June 2014 conflict minerals disclosures (CMDs) that indicate whether they source conflict minerals from the DRC or any of its nine neighboring countries, collectively known as the “covered countries.” The Enough Project, a nongovernmental organization (NGO) that played a leading role in advocating for Section 1502’s passage, claims that 3TG minerals are the most lucrative source of revenue to armed groups in Central Africa; in 2008 alone, armed groups earned an estimated $185 million from conflict minerals (Bafilemba et al., 2014). Section 1502 represents a marked shift from the SEC’s See https://www.bsr.org/reports/BSR_Conflict_Minerals_and_the_DRC.pdf Covered countries include the DRC and nine other “adjoining countries.” The term “adjoining countries” is defined in the Dodd-Frank Act as countries that share an internationally recognized border with the DRC. When the SEC issued the conflict minerals disclosure rule, these countries were Angola, Burundi, Central African Republic, the Republic of the Congo, Rwanda, South Sudan, Tanzania, Uganda, and Zambia. Electronic copy available at: https://ssrn.com/abstract=3908233 traditional role of investor protection to one of stakeholder protection (Lynn, 2011). The legislation broadened securities regulation to encompass greater accountability for SEC-registered firms in their sourcing practices. Specifically, by promoting supply chain transparency, the conflict minerals disclosures were intended to compel greater firm accountability and to curb the role of the conflict minerals trade in financing the region’s armed groups. Despite the importance of understanding the SEC’s regulation, limited empirical evidence exists. To fill the void in the literature, we investigate whether the conflict minerals disclosure mandate serves as an effective mechanism to achieve an international humanitarian goal. It is an open empirical question whether Section 1502’s conflict minerals disclosure requirement fulfills its goal of promoting sourcing transparency and corporate social responsibility, let alone whether it reduces conflict in covered countries. There are several reasons why the disclosure requirement may not be effective or that its direct effects may be unknown. First, companies have struggled to acquire and provide detailed information regarding the sources of their 3TG minerals, especially if they operate multi-layered supply chains (Kim and Davis, 2016). Second, even if companies discover that their 3TG minerals have been coming from covered countries, they must take numerous and often complex and costly steps to address the issue. Lastly, the SEC mandate does not impose a direct penalty for using conflict minerals. Thus, companies may continue to source conflict minerals from the covered countries without legal repercussions, particularly given that the mandate does not constitute an embargo on the trade of conflict minerals from all covered countries. Hence, it is uncertain whether we will find evidence of a decrease in conflicts given the number of reasonable arguments for why the CMD rule may not achieve its desired aim. To the extent that managers with limited information sets are able to collect, process, and See Appendix IV for OECD Due Diligence guidelines. However, issuers that make a false or misleading conflict minerals disclosure are liable to any person who purchased or sold a security at a price which was affected by the conflict minerals disclosure. Electronic copy available at: https://ssrn.com/abstract=3908233 disclose relevant information about their sourced minerals, it is probable that reputational concerns would spur changes in decision-making and/or behaviors that comport with the Congress’s overarching goal of conflict reduction (Hombach and Sellhorn, 2019). Specifically, reputational costs such as stakeholders’ activism (e.g., customers, investors, and employees) may serve to incentivize companies to comply with the conflict minerals disclosure rule and source responsibly. However, the pressure to conform to prevailing social norms can come into conflict with the actions that would be most operationally efficient. To this end, if the costs of altering suppliers outweigh the benefits of conferred reputation as a responsible sourcing firm, firms may continue to use conflict minerals that originate from the covered countries. Our study examines whether companies take action to source conflict-free minerals and, if so, whether those actions mitigate conflicts in the DRC and neighboring countries. To establish whether companies are able to provide detailed information about their sourced minerals and to assess decision-making regarding conflict-free minerals sourcing, we hand-collect data on the number of conflict-free smelters and refiners disclosed by all firms from the specialized disclosures (Form SDs) between 2014 and 2018. We supplement the data with a dataset purchased from Developmental International, which collects data from firms’ specialized disclosures between 2014 and 2016, respectively. Per Section 1502, all publicly traded firms are required to file specialized disclosures. If the firms believe that their products may contain minerals that are not DRC conflict-free, they must also file conflict minerals reports that specify the due diligence parameters of their responsible sourcing of those minerals. A prerequisite to our study is that companies are able to obtain and disclose information Companies face reputational costs from a perception of funding armed groups and thus conflicts in the covered countries. For example, student-led campus initiatives have encouraged universities (e.g., the University of California, Berkeley) to publicly support the conflict-free movement by not purchasing cell phones, laptops, and other devices that are known to finance war in the Democratic Republic of Congo and neighboring countries. Conflict-free minerals are defined as minerals that do not directly or indirectly benefit armed groups in the covered countries. Electronic copy available at: https://ssrn.com/abstract=3908233 about their supply chains as to whether they source conflict minerals. Our hand-collected data confirms that assumption. Specifically, our descriptive findings are consistent with Roychowdhury et al. (2019), which note that managers may learn new information only after a regulation forces them to collect it. Similarly, our descriptive evidence shows a gradual increase in the percentage of audited smelters, which is consistent with the extant literature that finds that compliance with financial reporting standards leads to improved corporate decisions by requiring managers to acquire information related to control systems (e.g., Cheng et al. 2018; Khan et al. 2019; Kang et al. 2021). Further, we provide evidence that companies take action due to reputational concerns. Specifically, our multivariate analysis using data from conflict minerals reports reveals that a shift toward responsible sourcing is more likely when a firm’s specialized disclosure garners greater public attention. We also find positive market reactions when firms disclose either a greater percentage of conflict-free smelters/refiners or a policy outlining dissociation from conflict minerals, which suggests that a commitment to responsible sourcing leads to greater shareholder wealth. These results indicate that reputational concerns may indeed compel corporations to dissociate from directly or indirectly contributing to the conflict and its financing of armed groups. Next, we focus on whether the mandatory disclosures have in fact ameliorated conflict in covered countries. Specifically, we investigate whether companies’ shift toward responsible sourcing results in human rights advancements and quality of life improvements for a neglected stakeholder group—civilians adversely affected by corporations’ externalities. We leverage data from the Armed Conflict Location & Event Database (ACLED) to obtain dates, locations, and types of conflict events from 2010 to 2019. Following Berman et al. (2017), we exploit a georeferenced panel dataset that divides the entire continent of Africa into about 10,000 For example, in 2017, Apple directed its suppliers to remove from its supply chain 10 smelters and refiners not willing to participate in, or complete, a Third Party Audit within given timelines. Apple requires its suppliers to ensure that conflict minerals smelters and refiners in its supply chain participate in Third Party Audit programs. Moreover, we note that 380 firms have leveraged the tools of the Responsible Minerals Initiative to eradicate smelters and refiners who are not in accordance with the responsible minerals assurance process. Electronic copy available at: https://ssrn.com/abstract=3908233 subnational units (i.e., square cells that are 55 × 55 kilometers in area at the equator) and compare changes in conflict incidence around mining areas of the covered countries relative to the contemporaneous changes in conflict incidence in the mining regions of the non-covered countries. Based on our analysis, we discover that Section 1502 achieves its intended humanitarian goal. Specifically, using a difference-in-differences research design, we find that relative to non- covered countries’ mining regions, conflicts in covered-countries’ mining regions decrease by roughly 15% after the conflict minerals disclosure mandate. To confirm that Section 1502 is effective against various types of conflicts, we stratify our sample into violent and non-violent conflicts. We document that both types decrease at a similar level. To alleviate the concern that the conflict minerals disclosure rule motivates the migration of mining areas’ armed groups to other regions, we examine whether conflict incidence spills over into non-mining areas. We find evidence that conflicts in the mining regions did not decrease at the expense of exacerbating violence elsewhere. Overall, we provide compelling evidence that Section 1502’s disclosure requirement motivates companies to dissociate themselves from conflict minerals sourcing in covered countries, and that this, in turn, results in conflict mitigation in the mining regions of covered countries. The effects of enhanced supply chain transparency appear to reverberate through transnational supply chains, thereby improving the humanitarian situation in the mining regions of covered countries. We examine parallel trends by using pre-treatment time period indicator variables and show that the conflict minerals disclosure mandate’s effect on the number of conflicts is statistically insignificant in the years prior to implementation but becomes negative and statistically significant The use of ACLED is in line with a large body of economics and political science literature that examines conflicts in Africa at a PRIO-GRID-level (see Besley and Reynal-Querol, 2014; Berman and Coutteneir, 2015; Berman et al., 2017; Harari and La Ferrara, 2018). ACLED is directed by Dr. Clionadh Raleigh (see Raleigh et al., 2010), and its credibility is widely recognized. Electronic copy available at: https://ssrn.com/abstract=3908233 after the conflict minerals disclosure mandate’s implementation. Our results are robust to a variety of sensitivity checks, ensuring the validity of the baseline results. Also, placebo tests further substantiate our findings. We are aware of at least two prior studies on Section 1502’s impact on the DRC’s conflicts. Parker and Vadheim (2017) and Stoop, Verpoorten, and Windt (2018) examine the legislation’s effects by using the post-period years of 2011–2012 and 2013–2015, respectively, and find that the regulation exacerbated the country’s conflicts. Importantly, their studies focus on the period following the act’s announcement (2010) but before 2014, which is the first year when firms were obligated to disclose whether their products contain 3TG that originate from the covered countries. Our study differs from the aforementioned studies in several ways. First, by using 2015 through 2019 as our post-period years, we extend the literature by focusing on the effective date of disclosure (i.e., after firms are required to document due diligence regarding conflict mineral sourcing). Second, the georeferenced panel data allows us to compare conflict incidents in the mining areas of the covered countries with those in mining areas of the non-covered countries. Put differently, our identification strategy specifies a more refined control group compared to that of Parker and Vadheim (2017), a study that compares the changes in the incidences of conflicts in the mining regions of DRC to the contemporaneous change in incidences of conflicts other non-mining regions of the DRC (e.g., capital area). We contribute to the extant literature in several ways. First, our study answers Christensen et al. (2021)’s call for more research on the consequences of Environmental, Social and Kim and Davis (2016) document that in 2014, only about 1% of corporations asserted that their products were “conflict-free beyond a reasonable doubt,” whereas 19% declared that they had no reason to believe their products contained DRC conflict minerals. The remaining 80% of corporations stated that they were unable to determine their raw materials’ countries of origin. Further, according to a 2015 Amnesty International dossier, 79% of firm reports did not meet the conflict minerals disclosure mandate’s minimum requirements in 2014, the first year of implementation. This may not be surprising, given that the SEC has estimated the cost of compliance at $3 billion to $4 billion, and the rule’s critics have deemed it as high as $16 billion during the first year of implementation alone (SEC 2012). This line of evidence suggests that the transition to responsible sourcing may be either slow to materialize or elusive. Electronic copy available at: https://ssrn.com/abstract=3908233 Governance (ESG) reporting (Christensen et al., 2017; Ioannou and Serafeim, 2017; Birkey et al., 2018; Grewal et al., 2019; Rauter, 2020; Krueger, 2021). Whereas most studies focus on traditional capital-market outcomes, we show that conflict minerals disclosures affect stakeholders beyond the scope of traditional capital-market participants. More broadly, we extend the literature on disclosures’ real effects (e.g., Leuz and Verrecchia 2000; Bens and Monahan, 2008; Zhang, 2009; Amir et al., 2010; Dyreng et al., 2016; Kanodia and Sapra, 2016; Roychowdhury et al., 2019) by identifying how companies make decisions in response to information revealed in non-financial disclosures that may influence non-financial stakeholders (e.g., consumers, suppliers, and communities that are adversely affected by corporations’ upstream operations). Second, we also add to the large body of economics and political science literature that examines the link between conflicts and natural resources (e.g., Berman et al., 2017) and that proposes solutions to human rights abuses in cross-border supply chains (e.g., Heffernan, 2016). We show that conflict minerals disclosures help mitigate externalities of firms’ upstream operations affecting a neglected stakeholder group, civilians of covered countries’ mining areas, and in so doing, highlight the real effects of mandating social responsibility disclosures. We also offer the channel of reputational costs to show that targeted transparency regulations compel greater corporate accountability and more responsible sourcing behaviors that align with society’s expectations. Third, comprehending the real effects of mandatory ESG reporting is increasingly important given the contemporary policy climate’s focus on social responsibility. For example, The terms “sustainability,” “environmental, social, and governance (ESG),” “non-financial,” and “corporate social responsibility” (CSR) reporting are used interchangeably to describe reports that focus on environmental, social, or corporate governance issues to varying degrees. One of the first widely publicized cases of human rights abuse in transnational supply chains occurred in the 1990s when the apparel and footwear sector was closely linked to sweatshops. Reports over substandard labor and human rights conditions related to well-known athletic apparel manufacturers were publicized by the media (e.g., CBS Evening News, 1996), prompting President Clinton to initiate the Apparel Industry Partnership in 1997, which introduced a workplace code of conduct and monitoring principles. Electronic copy available at: https://ssrn.com/abstract=3908233 U.S. federal agencies have been directed by President Biden to promote “social welfare, racial justice, environmental stewardship, human dignity, equity, and the interests of future generations.” Gary Gensler, SEC Chairman, has emphasized the agency’s focus on environmental, social, and governance (ESG) matters, and the Sustainability Accounting Standards Board (SASB) has issued standards to aid stakeholders’ application of ESG information in their decision-making processes. Moreover, the U.S. government is also assessing the efficacy of Section 1502. Specifically, the Government Accountability Office (GAO) is searching for performance indicators to assess progress toward the conflict minerals disclosures’ overarching goal of addressing armed groups’ exploitation of conflicts minerals. In addition, the European Union (EU) has also mandated ESG disclosures, including targeted Conflict Minerals Regulation that commenced in January 2021. Although our relatively narrow focus on Section 1502 of the Dodd-Frank Act could limit our findings’ generalizability, our study nonetheless provides direct evidence on the real effects of the conflict minerals disclosure mandate—a valuable insight given the aforementioned agenda and relevant initiatives. Our findings should be of interest to policy makers who suspect whether mandatory disclosure regimes can be an effective regulatory instrument to achieve social goals based on ex ante policy characteristics. Our study proceeds as follows. In section 2, we discuss the institutional background. In section 3, we present the literature review and hypotheses development. Section 4 contains our empirical results. We summarize and conclude our study in section 5. 2. Institutional Background 2.1 Section 1502 of the Dodd-Frank Act In response to the 2007–2009 financial crisis, Congress enacted the Dodd-Frank Act in See https://www.whitehouse.gov/briefing-room/presidential-actions/2021/01/20/modernizing-regulatory-review/ See https://www.sec.gov/files/esg-risk-alert.pdf See https://www.sasb.org/about/sasb-and-other-esg-frameworks/ See https://www.gao.gov/products/gao-20-595 See https://ec.europa.eu/trade/policy/in-focus/conflict-minerals-regulation/regulation-explained/ Electronic copy available at: https://ssrn.com/abstract=3908233 July 2010. Touted as the most comprehensive financial reform legislation since the 1930s, the Dodd-Frank Act aims to “promote the financial stability of the U.S. by improving accountability and transparency in the financial system, to end ‘too big to fail,’ to protect the American taxpayer by ending bailouts, to protect consumers from abusive financial services practices, and for other purposes.” In particular, Sections 1502, 1503, and 1504 relate to furthering social goals unrelated to the SEC’s traditional mission of investor protection and mitigating information friction. Section 1502 addresses the role of conflict minerals in the DRC’s human rights crisis. The conflict minerals disclosure rule requires firms to design, in good faith, a reasonable inquiry to determine whether their 3TG minerals are sourced from covered countries. To ensure compliance, the conflict minerals disclosure rule specifies that firms must file specialized disclosures for the first time by June 2014, and annually thereafter by May 31 of each year. The SEC also stipulates that firms that conduct reasonable country-of-origin investigation and thereby discover that their products may contain minerals that are not DRC conflict-free must also exhibit a conflict minerals report. The conflict minerals disclosure rule expands the corporate supply chain information environment to encourage more socially responsible sourcing. To fulfill this goal, it leverages public pressure and reputational concerns to incentivize corporate accountability. Put differently, as compared with an embargo that would prevent the covered countries’ export of natural resources entirely and indiscriminately, Section 1502 exploits supply chain transparency as a mechanism to Section 1503 requires corporations to include information regarding mine-safety performance in their financial reports. Section 1504 requires firms to disclose information on payments to foreign governments for resource- extraction activities. It is worth noting that conflict minerals are defined in the U.S. legislation as tantalum, tin, tungsten, and gold, regardless of their extraction location. For example, even if a company’s tin is extracted from Canada, because it is considered a conflict mineral, any company that uses it must disclose whether it is “DRC conflict-free” to meet the conflict minerals disclosure rule guidelines. For an example of a conflict minerals report, see https://www.apple.com/supplier-responsibility/pdf/Apple- Conflict-Minerals-Report.pdf Electronic copy available at: https://ssrn.com/abstract=3908233 discourage the sourcing of conflict minerals. Weil et al. (2013) refer to such rules as “targeted transparency” regulation, which compels firms to disclose standardized information to reduce specific risks, ameliorate negative externalities, minimize the given social costs associated with a product, and/or improve provision of public goods and services. Hence, the conflict minerals disclosure rule reflects Congress’s choice to use a securities law disclosure requirement to achieve a specific humanitarian objective. 2.2 The Resource Curse in the Democratic Republic of Congo The resource curse refers to the paradox that countries with the largest endowment of natural resources often suffer from greater corruption, grievances, economic distress, and/or conflicts (Auty, 1993; Sachs and Warner, 1995; Auty, 2001; Gylfason, 2001). The oil-rich Nigeria and the mineral-abundant Republic of Sierra Leone are both cases in point. Particularly, abundance of natural resources is known to attract armed groups while possibly allowing corrupt and/or incompetent governments to prolong their reigns through the extraction and sale of natural resources. With vast deposits of copper, diamonds, cobalt, petroleum, gold, silver, zinc, coltan, etc. DRC is noted for suffering gravely from the resource curse (U.N. Security Council, 2010). Specifically, a study by the United Nations Environment Programme in 2011 estimates that the DRC has an untapped deposit of minerals that are worth 24 trillion US dollars (United Nations Environment Programme, 2011). It is well-known that the DRC’s paramilitary groups exploit the trade of natural resources to finance their operations, and that those revenues fuel further violence (SEC, 2012; Prendergast and Bafilemba, 2018). The Enough Project, a non-government organization (NGO) that played a A New York Times article states that “Representative Jim McDermott, a Democrat from Seattle, had tried for years to regulate conflict minerals. Mr. McDermott, who served as a Foreign Service medical officer in central Africa during the 1980s, [toured] a hospital that was treating people wounded in the continuing civil war. After visiting with a group of rape victims, he said he was shocked by the human rights abuses. Mr. McDermott traced much of the suffering to armed soldiers who sold tantalum and other minerals to finance their war.” (New York Times, 2011) See: https://dealbook.nytimes.com/2011/07/13/unearthing-exotic-provisions-buried-in-dodd-frank/ Electronic copy available at: https://ssrn.com/abstract=3908233 leading role in advocating for Section 1502’s passage, claims that 3TG minerals are the most lucrative source of revenue for armed groups in Central Africa. The armed groups commit ongoing human rights abuses (e.g., the use of child labor has been reported in the DRC’s eastern regions). The abundance of natural resources in the DRC is the root cause of war, extreme violence and abject poverty. Collier et al. (2009) posit that armed groups assess operational costs to determine the financial feasibility of waging violence in a particular geographical location. Thus, per the Enough Project’s supposition, it is reasonable to expect that a decrease in demand for conflict minerals from covered countries may dampen the region’s armed groups’ revenues, thereby compromising the viability of their continuing domain over those areas. 3. Literature Review and Hypotheses Development A large body of literature investigates whether mandatory financial disclosures affect corporate decision-making (Leuz and Verrechia 2000; Healy and Palepu, 2001; Beyer et al., 2010; Leuz and Wysocki 2016; Roychowdhury et al., 2019; Healy and Serafeim, 2020). For instance, Bens and Monahan (2008) confirm that after FIN 46 required the consolidation of variable interest entities in financial statements, corporations reduced investments in those entities. Zhang (2009) finds that after SFAS 133 mandated that derivatives be recognized as either assets or liabilities at fair value, corporations decreased their speculative use of derivative instruments. Dambra et al. (2021) show that GASB 68’s implementation, which altered disclosure requirements for government pension obligations, led counties to reduce public welfare expenditure, employment, and salary expenses. Collectively, these studies document the real impact of mandatory financial As an additional example, non-state armed groups often use sexual violence to drive a local population away from an artisanal mining area. Artisanal and small-scale mining occurs in approximately 80 countries worldwide; it is particularly widespread in developing countries in Africa, Asia, Central and South America, and Oceania (World Bank). The World Bank estimates that about 100 million people (workers and their families) depend on artisanal mining, as compared with about 7 million people engaged in industrial mining. Artisanal and small-scale production accounts for 20% of global gold supply. Policymakers argue that the gradual eradication of artisanal mining may lead to the reduction of violence in the DRC (U.S. Secretary of State and USAID, 2011). Electronic copy available at: https://ssrn.com/abstract=3908233 disclosures. We review the literature on mandatory ESG reporting and their economic consequences. There is evidence that ESG-based disclosure regulations have real effects. For example, Christensen et al. (2017) show that when mine owners are obligated to disclose safety performance information in their 10-K filings, per Dodd-Frank Act Section 1503, they are incentivized to increase investments in that domain. Likewise, Ioannou and Serafeim (2017) document that the introduction of ESG disclosure requirements improves disclosure quantity and quality, as well as a firm’s value, thereby enhancing ESG development. In addition, Rauter (2020) provides evidence that detailed disclosures about extraction payments to foreign host governments, per Dodd-Frank Act Section 1504, reduce illicit payment practices. Mandatory ESG disclosure requirements can decrease negative externalities or foster positive ones. In this study, we contend that the conflict minerals disclosure requirement induces reputational costs as the channel through which the disclosure requirement nudges managers to reduce negative externalities. The reputational cost hypothesis (Benabou and Tirole, 2006) suggests that because the effect of opprobrium associated with the disclosure of irresponsibly sourced conflict minerals could result in fewer consumer sales (Wang et al., 2018), higher financing costs (Cao et al., 2015; Hartzmark and Sussman, 2019), increased employee wages (Novak and Bilinski, 2018), or greater political costs (Rauter, 2020), managers are incentivized to shift toward responsible sourcing to avoid such outcomes. We expect that the process of preparing CMDs (e.g., ascertaining the percentage of conflict-free smelters/refiners) will make firms more aware of what goes into their products and An alternative channel is managerial learning. Shroff 2017 suggests that managers acquire new information because of disclosure requirements. Steinmeier and Stich (2019) posit that requiring corporations to produce sustainability reports can enhance the efficiency of managers’ sustainability-related decisions. Along the same lines, Roychowdhury et al. (2019) note that managers may only learn new information if a regulation forces them to collect it. We cannot directly observe the priors of the managers nor discern whether they had material information about potential conflict inputs in their supply chain prior to the CMD rule. Thus, the proposed “learning channel” remains a conjecture that is not tested in the paper. Electronic copy available at: https://ssrn.com/abstract=3908233 escalate stakeholder pressure accordingly, thus prompting companies to dissociate from risky conflict minerals. We observe that various stakeholders are keenly aware of conflict minerals disclosures. For example, preparing conflict minerals disclosures increases the upstream visibility of a company’s supply chain, allowing managers to increase profits and sales. (Swift et al., 2019). Likewise, creditors give more trade credit to companies with more supply chain visibility (Ng et al., 2020). In addition, consumers, humanitarian groups, investors, and the public all use conflict minerals disclosures to evaluate corporations’ commitment to social responsibility. This practice is reflected by the comment letters submitted to Acting Chairman Michael S. Piwowar’s Reconsideration of Conflict Minerals Rule Implementation (2017). After carefully reviewing all 323 comments, we find that 270 supported the legislation’s implementation. Finally, the market incorporates the information in conflict minerals disclosures. Share prices drop for companies that source minerals from the DRC and adjoining countries, violate human rights, or provide ambiguous disclosures. Conversely, share prices appear to rise for firms that mitigate risks of sourcing conflict minerals (Elayan et al., 2021). We posit that SEC-issuers shift towards responsible sourcing as new information about the origins of their products’ inputs is disclosed by complying with Section 1502. To this end, Amnesty International (2015) notes that before the passage of Section 1502, most corporations did little to ensure they were sourcing 3TG responsibly. It stands to reason that in general, corporate managers and other stakeholders were unaware of the origins of the companies’ raw inputs prior to the passage of Section 1502 of the Dodd-Frank Act. To understand the process of compliance with Section 1502 and how different stakeholders can learn about companies’ sourcing behavior, we outline the due diligence requirements imposed over SEC-issuers. To begin, if 3TG are necessary to the functionality or production of the issuer’s product, the issuer must conduct a “reasonable country of origin inquiry” into the source of the designated minerals. After conducting Electronic copy available at: https://ssrn.com/abstract=3908233 the inquiry, the issuer reports whether the designated minerals used in its product originated from the covered countries in a specialized disclosure form (i.e., Form SD) by the end of May in the following year. If a company describes any of its products as “DRC conflict free” in its CMD, it must obtain an independent private sector audit on its due diligence; and if a company has products that are not described as “DRC conflict free” in its CMD, then it should disclose the smelters or refiners used to produce the conflict minerals and the efforts made to determine the mine or location of origin. Given the due diligence process, it stands to reason that stakeholders learn about the origins of companies’ rawest inputs from complying with Section 1502. To the extent that reputational costs affect corporate responsiveness, we anticipate subsequent dissociation from risky smelters and refiners. That is, we expect that as stakeholders learn about risky smelters and refiners within companies’ supply chains, reputational concerns motivate managers to strive for responsible sourcing. Hence, our first hypothesis focuses on dissociation (stated in the affirmative): H1: In response to the conflict minerals disclosures requirement, firms will dissociate from risky smelters and refiners that deal in conflict minerals that originate from the covered countries. To the extent that the conflict minerals disclosure regime successfully dampens the demand for conflict minerals that originate from the covered countries (i.e., the value of the spoils We have also considered reasonable arguments as to why the conflict minerals disclosure rule, despite its intentions, may not achieve its desired aim. First, rather than imposing direct sanctions on specific countries, the U.S. Congress leveraged Section 1502 to encourage corporations to dissociate from conflict minerals. Hence, considering that the regulation’s beneficiaries are civilians in the covered countries who may lack financial claims over the firm, it is possible that the real effects may not materialize due to the lack of substantive incentive to drive corporate behavioral sourcing changes. Secondly, the average firm may be unable to accurately conclude whether its products are “DRC conflict-free,” since it can be challenging for a corporation that operates multi-layered supply chains to track the origins of its rawest inputs (Kim and Davis, 2016; Islam and Van Staden, 2018). This difficulty can, for instance, encourage firms to elect to state that they were “unable to determine with certainty” whether the minerals do not originate from the covered countries without changing their resourcing behavior. If the costs of altering suppliers outweigh the benefits of attaining positive reputational standing, firms may continue to use conflict minerals that originate from the covered countries. Electronic copy available at: https://ssrn.com/abstract=3908233 that go to the armed groups from controlling the artisanal mining regions), incentives for rebel groups to commit acts of violence will drop as well. Given that the U.S. is the world’s single largest economy, accounting for more than 20 percent of global output (www.ustr.gov), we predict that a widespread change in resourcing behavior by firms can have real effects. Thus, responsible sourcing by corporations in the U.S. can decrease rebel groups’ incentives to resort to violence as the value of the contestable prize drops. Altogether, we expect managers to alter their sourcing behavior to confer with society’s expectations (i.e., reputation channel). This can, in turn, reduce the demand for conflict minerals that originate from the covered countries, dampening the incentives for armed groups to resort to violence to control the mining areas. We state our second hypothesis in the affirmative: H2: In response to the conflict minerals disclosure requirement, conflict incidence in the mining regions of the covered countries is alleviated. 4. Responsible Sourcing and its Impact on Conflicts in Covered Countries To test for the two hypotheses, we employ two datasets. In the first sub-section, exploiting hand-collected firm-level data, we examine the real effects of Dodd-Frank Act Section 1502 and observe whether firms shift to responsible sourcing. In the second sub-section, we use geopolitical data and investigate whether the conflict minerals disclosure requirement alleviates the resource curse in the covered countries. 4.1. Responsible Sourcing 4.1.1 Data and Descriptive Statistics It is worth noting that firms can shift from using conflict minerals that originate from the covered countries to using recycled/scrapped 3TG. Electronic copy available at: https://ssrn.com/abstract=3908233 To determine how public attention influences responsible sourcing behavior, we begin by leveraging hand-collected data from the specialized disclosures and the exhibited conflict minerals reports. As mentioned in Section 2.2, all firms are required to file a specialized disclosure. If a firm uses conflict minerals, they must also exhibit a conflict minerals report that includes (i) a description of the products manufactured or contracted to be manufactured that are not ‘‘DRC conflict-free,’’ (ii) the facilities (e.g., smelters and refiners) used to process the conflict minerals, (iii) the country of origin of the conflict minerals, and (iv) the efforts made to identify the mine or location of origin. We utilize hand-collected data from items in the conflict minerals reports for the period of 2014-2018. Specifically, we gather information on the total number of smelters and the number of verified, conflict-free smelters/refiners for all firms for which the conflict minerals disclosure rule applies for the sample period. Further, we leverage data from Developmental International, which is a not-for-profit organization that collected qualitative data from conflict minerals disclosures between 2014 and 2016. For instance, when firm i indicates that it has “implemented risk management plans, monitored and tracked performance of risk mitigations, and suspended or discontinued engagement with a supplier after failed attempts at risk mitigation or corrective action,” Developmental International codes that firm i has “dissociation policies in place” related to conflict minerals. We begin with all specialized disclosures between 2014 and 2018. This sample consists of 4,068 firm-year observations for 1,033 unique firms. We then match this sample with price information from the CRSP database and financial data from the Compustat annual file. After requiring non-missing information, our sample is limited to 1,956 firm-year observations for 776 unique firms. See Appendix III for examples of a specialized disclosure and conflict minerals report. Electronic copy available at: https://ssrn.com/abstract=3908233 To execute our empirical analysis, we use two measures to proxy for Responsible Sourcing. The first measure, % of Conflict-Free Smelters/Refiners, is the ratio of the number of conflict-free smelters/refiners over the number of total smelters/refiners. The second measure, Dissociation, is an indicator variable equal to one if there is evidence that a firm suspends its purchases from risky smelters and refiners, and equal to zero, otherwise. To examine market reactions to the specialized disclosures, we calculate the cumulative market-adjusted return for each firm in our sample in the 5 days surrounding the annual specialized disclosure (CAR[-2, +2]). Given that the specialized disclosure is filed at the end of May, separately from other mandatory filings (e.g., annual reports), there are very few concurrent firm-specific news events. However, to mitigate the effect of observations that might be contaminated with other news, CAR[-2, +2] is winsorized at the 1% level. We use the total number of a firm’s specialized disclosure downloaded by non- robots from EDGAR to measure Public Attention—our proxy for reputational risk. Panel A of Table 1 reports the descriptive statistics for the primary variables of our constructed sample. The average % of Conflict-Free Smelters/Refiners is 71.6% in our sample. Dissociation statements can be found in 52.2% of the disclosures in our sample. The average firm’s market-adjusted abnormal returns, CAR[-2, +2] is 0.3%. On average, a specialized disclosure is downloaded 85 times per year. The average firm has $9.293 billion in total assets, $11.157 billion in market value, and $6.730 billion in sales. Panel B of Table 1 provides descriptive evidence of changes in Responsible Sourcing by year. The average % of Conflict-Free Smelters/Refiners monotonically increases from 44.6% in The data with the % of Conflict-Free Smelters/Refiners measure was hand-collected from the conflict mineral reports by the authors and is available between 2014 and 2018. The data with the Dissociation measure was hand-collected and reviewed by Developmental International (DI), a not-for-profit organization that specializes in areas where law, business, and development intersect, and is available only between 2014 and 2016. See: https://www.developmentinternational.org. Results remain qualitatively unchanged when we truncate or do not winsorize the sample. Electronic copy available at: https://ssrn.com/abstract=3908233 2014 to 81.9% in 2018 and the cumulative level of firms that adopt dissociation policies (i.e., Dissociation) increases from 45% to 64.5% between 2014 and 2016. These trends demonstrate that corporations are able to obtain and disclose information about the conflict minerals in their supply chains as they meet the due diligence requirements of conflict minerals disclosures, which is a prerequisite in our study. Moreover, corporations’ gradual dissociation from smelters and refiners that source minerals from the covered countries provides preliminary evidence that supports our first hypothesis. (TABLE 1 ABOUT HERE) 4.1.2 Empirical Analysis of Responsible Sourcing To investigate Section 1502’s impact on firm’s sourcing behavior (a shift towards more responsible sourcing), we explore whether companies are more likely to take actions when public attention is more pronounced. We begin by estimating the following regression : 𝑏𝑖𝑙𝑅𝑒𝑝𝑜𝑛𝑠𝑠 𝑒 𝑖𝑛𝑆𝑜𝑔𝑐𝑢𝑟 = 𝛽 𝑖𝑐𝑃𝑢𝑏𝑙 𝑒𝑛𝑖𝑜𝑡𝑡𝐴𝑡𝑛 + 𝛽 𝑠𝑜𝑛𝑟𝑜𝑙𝐶𝑡 + 𝜂 + 𝜂 + 𝜀 (1) , , , Hombach and Sellhorn (2018) define reputational cost as all reductions in firm value resulting from a firm’s exposure to public scrutiny. We use Public Attention, the total number of firm i’s specialized disclosures downloaded by non-robots from EDGAR during the previous calendar year, as a proxy for reputational risk. We expect that companies will be subject to greater reputational costs with higher levels of Public Attention, following Dambra et al. (2021). As for control variables, we follow Kim and Davis (2016) and construct several firm characteristic We run an ordinary least squares regression when the 𝑜𝑛𝑅𝑒𝑠𝑝𝑠𝑏𝑖𝑒𝑙 𝑆𝑜𝑢𝑐𝑖𝑛𝑟𝑔 variable is % of Conflict-Free Smelters/Refiners and a probit regression when the 𝑜𝑅𝑒𝑠𝑝𝑛𝑏𝑠𝑖𝑙𝑒 𝑆𝑜𝑢𝑐𝑖𝑛𝑟𝑔 variable is Dissociation. The EDGAR Server Log data, which contain information on internet search traffic for EDGAR filings through SEC.gov, are produced by the SEC’s Division of Economic and Risk Analysis. For more information, visit https://www.sec.gov/dera/data/edgar-log-file-data-set.html. In an untabulated analysis, we find that there is a positive and statistically significant correlation between reputation risk measured by RepRisk and our Public Attention measure. RepRisk developed a proprietary algorithm to dynamically capture and quantify a company’s exposure to ESG news and stakeholder criticism to create the reputation risk metric (RepRisk, 2017). Electronic copy available at: https://ssrn.com/abstract=3908233 variables that could potentially confound the effects of public attention on 𝑜𝑏𝑠𝑝𝑛𝑠𝑖𝑒𝑅𝑒𝑙 𝑖𝑛𝑆𝑜𝑔𝑐𝑢𝑟 , such as profitability (e.g., ROA), revenues (e.g., Ln Sales), and total assets (e.g., Ln Total Assets). We account for growth by including the book-to-market ratio (BTM). We also control for cash (Ln Cash) and free cash flow (Free Cash Flow) to account for firms’ positions to alter their sourcing behavior. Also, we control for a firm’s capital adequacy using debt-to-assets (Leverage). All st th variables are defined in Appendix I. All continuous variables are winsorized at the 1 and 99 percentiles to mitigate the impact of outliers. In all regression specifications where we test the real effects of Section 1502 related to firm sourcing, we include Fama-French 12 industry fixed effects to account for time-invariant unobserved heterogeneity at the industry level and year fixed effects to account for annual shocks. Columns (1) and (2) of Table 2 present the results. The coefficient of interest, Public Attention, indicates that companies are more likely to responsibly source, as proxied by % of Conflict-Free Smelters/Refiners and Dissociation by 1.2% and 4.5%, respectively, when there is greater public attention related to their conflict minerals reports. These results suggest that companies’ shift toward responsible sourcing in year t is stronger for firms that receive greater public attention in the previous year, consistent with the reputational cost hypothesis, which postulates that managers integrate social expectations into their decisions to responsibly source. Overall, our evidence suggests the efficacy of the conflict minerals disclosure rule as a transparency measure to nudge firms towards responsible sourcing. (TABLE 2 ABOUT HERE) To better understand the motivation for corporate dissociation from risky smelters and refiners, we examine whether the disclosure of firms’ responsible sourcing results in positive or We multiply public attention by 100 to express the rate per 100 downloads/year. Thus, the economic interpretation is as follows: one hundred downloads of a firm’s specialized disclosure from EDGAR generates a 4.5% increase in the likelihood of Dissociation and 1.2% increase in % of Conflict-Free Smelters/Refiners. Electronic copy available at: https://ssrn.com/abstract=3908233 negative market reactions. On the one hand, responsible sourcing could be translated into value because of operating efficiencies (Swift et al. 2019), stronger brand and customer loyalty (Wang et al. 2018), lower financing costs (Cao et al. 2015), and employee engagement (Novak and Bilinski, 2018). All of these would result in a positive market reaction. On the other hand, we note that a firm’s responsible sourcing efforts could be performed at the expense of shareholders (Friedman, 1970). This could lead to a rise in firm costs, which would be disadvantageous in a competitive market and spur negative market reactions (Krueger, 2015). Moreover, Griffin et al. (2014) find negative market reactions to voluntary conflict minerals disclosures presumably due to firm costs. Hence, we study the market reactions to specialized disclosures that evidence a commitment to responsible sourcing. Table 3 reports the results for our univariate analysis. Panel A of Table 3 reports the differences in the mean values of selected variables between firms that do and do not indicate that they suspend their purchases from risky smelters and refiners. The results show that firms that dissociate (Dissociation=1) from risky smelters and refiners experience larger cumulative abnormal returns, provide significantly more conflict-free statements, are of larger size, and have lower book-to-market ratios than firms that do not dissociate (Dissociation=0). To mitigate concerns associated with the underlying differences between the two subsamples, we use the entropy-balancing approach. Following Hainmueller and Xu (2013), we match the mean, variance, and skewness (the first three moments) of the control group (Dissociation=0) to those of the treatment group (Dissociation=1). Our entropy-balanced results are presented in Panel B. These results show that there are no longer significant differences in the mean values of selected variables between firms with and without dissociation policies. (TABLE 3 ABOUT HERE) Next, we examine market reactions using a multivariate analysis by estimating the following regressions: Electronic copy available at: https://ssrn.com/abstract=3908233 CAR [-2, +2] = β + β Responsible Sourcing i,t 0 1 i,t + β Size + β BTM + β Leverage + ε (2) 2 i,t 3 i,t 4 i,t i,t. CAR[-2, +2] is calculated as the cumulative market-adjusted return for firm i in the 5 days surrounding the filing date t of the annual specialized disclosure. Following prior literature (Dhaliwal and Reynolds, 1994; Warfield et al., 1995), we control for the natural logarithm of market capitalization (Size), the book-to-market ratio (BTM), and the leverage ratio (Leverage) of firm i. Table 4 reports the regression results. Columns (1) and (2) show that the coefficient on % of Conflict-Free Smelters/Refiners is positive and statistically significant with and without including control variables. Columns (3) and (4) show that the coefficient on Dissociation is positive and statistically significant with and without including control variables. Specifically, we find that an increase in one standard deviation of % of Conflict-Free Smelters/Refiners (24.2%) leads to an increase of about 0.6% in market value. Moreover, we observe an increase of about 0.3% in market value for firms that disclose dissociation policies over the five-day window. Given that the average firm’s market value is $11.157 billion, an increase of 0.1% equates to $11.157 million. Overall, our results indicate that the market reacts positively to a responsible sourcing commitment, which suggests that the market views such a commitment as value-enhancing. (TABLE 4 ABOUT HERE) 4.2 Conflicts in Covered Countries 4.2.1 Data and Descriptive Statistics To analyze how the conflict minerals disclosure rule impacts conflict incidence in covered countries’ mining regions, we assemble a PRIO-GRID/year-level panel data set that includes 53 African countries represented by roughly 10,000 cells. Specifically, drawing upon Berman et al.’s The standardized regression coefficient estimate is generally calculated as follows: 𝐵 × [ , ] For % of Conflict-Free Smelters/Refiners, the calculation is 0.007 × = 0.059 Electronic copy available at: https://ssrn.com/abstract=3908233 (2017) design, we create a geo-referenced panel data set to divide the entire continent of Africa into 10,335 subnational units (i.e., square cells that are 55 × 55 kilometers in area at the equator) between 2010 and 2019. This data set covers virtually all of Africa’s terrestrial regions, with a resolution of 0.5 x 0.5 longitude and latitude across 10 years. We use mine-level data from Raw Material Data. The database provides the latitude and longitude of mining regions, which enables us to match these data to the geo-referenced panel data set. Thus, the geo-referenced panel data set also allows us to identify cells that are adjacent to a mining cell (i.e., Mining Region =1) (see Berman et al., 2017; Harari and La Ferrara, 2018). Next, following the extant economics and political science literature that examines conflicts in Africa (e.g., Berman et al., 2017; Harari and La Ferrara, 2018), we obtain data from the ACLED on the dates, locations, and types of conflict events between 2010 and 2019. Then, in accordance with prior research, we merge the ACLED data with the geo-referenced panel dataset upon collapsing conflict information at the cell/year-level. This methodology allows us to test changes in occurrences of conflicts at the cell/year-level. We remove the cells that correspond to the four nations in which no conflict events were reported during our sample period. Finally, we obtain information about country-level economic activities from the website theglobaleconomy.com, which contains over 300 indicators gathered from multiple official sources, such as the World Bank and the World Economic Forum. Specifically, we collect the following control variables: (i) GDP, (ii) GDP-per-capita, (iii) labor force, (iv) labor participation According to ACLED, the “process of ACLED coding assures that it is accurate, comprehensive, transparent, and regularly updated. Data are posted as they are complete, although there are ongoing checks to ensure the thoroughness of previously collected events. ACLED data are coded by a range of experienced researchers who collect information primarily from secondary source information and apply the guidelines outlined in the codebook to extract information from news reports. ACLED data are collected each week after individual researchers have scrutinized the information from reports; they are then aggregated and revised by the first coding reviewer, investigated and cross-checked by the second reviewer and then event notes and details are inspected by the third and final reviewer. The process is designed to assure: (1) Validity through intra- and inter-coder checks; (2) Accuracy to correct mistakes in coding; and (3) Relevance by determining whether each compiled event constitutes an act of political violence or protest.” These four countries include: Sao Tome and Principe, Rep. of Mauritius, Union of Comoros, and the Rep. of Cape Verde. See https://www.theglobaleconomy.com/download-data.php. Electronic copy available at: https://ssrn.com/abstract=3908233 37 rate, (v) male unemployment rate, and (vi) capital investment. To mitigate the effect of outliers, we winsorize all continuous variables at the 1% and 99% levels. We also exclude observations with missing control variables, which leads to the omission of four more nations. In our final step, we restrict our sample to include only mining regions, which results in a final total of 18,662 cell/year observations that includes 45 African countries. Table 5 summarizes the aforementioned process. (TABLE 5 ABOUT HERE) Panel A of Table 6 reports descriptive statistics for the primary variables of our constructed sample of mining regions between 2010 and 2019. The average raw number of Conflicts in our sample is 0.92. The average raw number of Violent Conflicts is 0.44 and the average raw number of Non-Violent Conflicts is 0.31. In our sample, the median GDP and GDP-per-capita are $29.31 billion and $1,743.65, respectively. The median country has 10.94-million Labor Force, $8.27- billion Capital Investment, Labor Participation Rate of 62.73%, and a Male Unemployment Rate of 8.11%. Panel B of Table 6 presents pairwise Pearson and Spearman correlations between Ln Conflicts and these country characteristics. Ln Conflicts is negatively correlated with GDP-per- capita and positively correlated with GDP, Labor Force, and Capital Investment. (TABLE 6 ABOUT HERE) 4.2.2 Empirical Analysis of Impact in Covered Countries Our objective is to determine whether Section 1502 of the Dodd-Frank Act achieved its goal of conflict reduction in the Democratic Republic of Congo and adjoining countries. We The extant literature documents various factors aside from natural resources that contribute to the prevalence of conflicts in Africa (e.g., Ross, 2004; Autesserre 2012; Geenen 2012; Seay 2012; Radley and Vogel 2015; Vogel and Raeymaekers 2016). These elements include: (i) weak and poorly functioning political institutions, (ii) ethnic fragmentation and polarization, and (iii) endemic poverty. We attempt to control for these constructs by using cell fixed effects. Electronic copy available at: https://ssrn.com/abstract=3908233 compare the changes in conflicts around the implementation of the conflict minerals disclosure regime for the “treated” mining regions of covered countries to those of the “control” mining regions of non-covered countries. If the conflict minerals disclosure rule reduces conflicts, we should observe a greater reduction in covered countries than in non-covered countries. We test our hypotheses using our final sample of mining regions in Africa between 2010 and 2019, which allows us to focus on conflicts that are specific to mining regions. To compare the changes in conflict incidence in the mining regions of covered countries to contemporaneous changes in the mining regions of non-covered countries after the conflict minerals disclosure regime becomes effective, we estimate the following regression: 𝐿𝑛 𝑠𝑙𝐶𝑖𝑜𝑐𝑛𝑡𝑓 = 𝛽 𝑃𝑜𝑠𝑡 𝐶𝑀𝐷 × 𝐶𝑜𝑣𝑒𝑑𝑒𝑟 𝐶𝑜𝑢𝑛𝑟𝑦𝑡 , , + 𝛽 𝑠𝑜𝐶𝑙𝑜𝑛𝑟𝑡 + 𝜂 + 𝜂 + 𝜀 , (3) , , where Ln Conflicts is the natural logarithm of (1 + the number of conflicts) in a cell k during year t. Post CMD is an indicator variable equal to one for years after 2014, and equal to zero, otherwise. Covered Country is an indicator variable equal to one if country c is designated as a covered country under Section 1502, and equal to zero, otherwise. Our main coefficient of interest is the interaction Post CMD × Covered Country, which estimates the change in the level of conflict incidence in mining regions of covered countries relative to the contemporaneous changes in the level of conflict incidence in mining regions of non-covered countries. We also include cell fixed effects (i.e., 𝜂 ) to account for time-invariant unobserved heterogeneity at the cell level. To this end, our identification strategy exploits within-cell variations in conflicts while controlling for While the specialized disclosure first became public at the end of May 2014, we set our post indicator variable to begin in 2015. This is because, consistent with prior literature on conflict minerals disclosures, we find that firms need time to acquire supply chain information. Further, a 2019 report by the U.S. Government Accountability Office reports that the percentage of companies that could determine the origins of their inputs was 30% in 2014 and about 50% in 2015 (GAO-19-607). Berman et al. (2017) note that “cell fixed effects are included with the purpose of controlling for time-invariant co- determinants of violence and mining at the local level – such as weak state capacity and property rights enforcement in remote places or latent political instability (e.g., ethnic cleavages).” Electronic copy available at: https://ssrn.com/abstract=3908233 40 time-invariant unobservables at the cell-level. We also include year fixed effects (i.e., 𝜂 ). Furthermore, we control for various theoretical determinants of conflicts, such as national wealth and male unemployment rate. For instance, we follow prior literature and control for GDP-per- capita at the country level (Besley and Reynal-Querol, 2014). Moreover, we control for social influences on crime, such as Labor Force and Labor Participation Rate. In addition, we control for Capital Investment, since both public and private investments can be used to offset social issues related to unemployment. Lastly, we cluster standard errors by country, given that our treatment effect is at the country level (see Bertrand et al., 2004). Appendix I presents the detailed variable definitions. ACLED identifies conflicts as either violent or non-violent. Using our sample of mining regions, we examine whether the rule had an effect on both types of conflicts after the conflict minerals disclosure rule’s implementation. To evaluate whether Section 1502 fulfilled its broader goal to promote peace, we examine the association between different types of conflicts and the conflict minerals disclosure rule by estimating the following regression: 𝐿𝑛 (𝑁𝑜𝑛 )𝑛𝑉𝑖𝑜𝑒𝑡𝑙 𝐶𝑜𝑛𝑠𝑓𝑙𝑖𝑐𝑡 = 𝛽 𝑠𝑡𝑃𝑜 𝐶 𝑀𝐷 × 𝑑𝑒𝐶𝑜𝑣𝑟𝑒 𝑡𝑟𝑦𝐶𝑜𝑢𝑛 , , (4) + 𝛽 𝑟𝑜𝑙𝑠𝐶𝑜𝑛𝑡 + 𝜂 + 𝜂 + 𝜀 , , Columns (1) and (2) of Table 7 present the results of estimating equation (3) before and after controlling for country characteristics and cell and year fixed effects, respectively. Over our time series, we observe a 17.7% (𝑒 − 1) increase in the number of conflicts, as evidenced by a positive and statistically significant coefficient on Post. Moving on to relative effects, we focus on the interaction term Post × Covered Country, which captures the conflict minerals disclosure regime’s incremental effect. Column (1) shows that the number of conflicts declined for covered countries relative to uncovered countries following the conflict minerals disclosure rule (coeff. = We cannot completely rule out the possibility that conflict reporting is biased toward certain types of countries, regions, or events, since some regions might have better media coverage. However, our empirical methodology makes it unlikely that this aspect affects our results, as structural differences in media coverage, or more generally, in the reporting of events, is captured by cell and country-year fixed effects. We also show that our results are obtained across events of different types. Electronic copy available at: https://ssrn.com/abstract=3908233 . -0.098, p-value < 0.01). This coefficient suggests a 9.3% ( 𝑒 − 1) decline in conflict incidence in the covered countries’ mining regions relative to those of the non-covered countries. We find a slightly stronger decline of 15.1% (𝑒 − 1) in Column (2) after including several control variables, as well as year and cell fixed effects. Adding the Post and Post × Covered Country coefficients in tandem demonstrates that although there is a significant increase in conflicts of mining regions in non-covered countries, they increase to a lesser degree in those of covered countries, suggesting that the conflict minerals disclosure rule helps to mitigate conflict escalation. Next, we examine whether the conflict minerals disclosure rule effectively contained both violent and non-violent conflicts. As reflected in Section 1502, the legislation was intended to promote the humanitarian goal and bring peace by alleviating the DRC’s conflict, which has been partially financed by the exploitation and trade of conflict minerals originating in the DRC. To test the rule’s impact regarding those broader aims, we rely on the ACLED classification of conflicts into violent and non-violent ones to examine whether both types of conflicts are mitigated. Then, in equation (4), we replace the outcome variable in equation (3) with the natural log of violent and non-violent conflicts, respectively. Columns (3) and (4) of Table 7 present the results of estimating equation (4) for violent conflicts, where Columns (5) and (6) present the results for non-violent conflicts. The coefficients of interest in Columns (4) and (6) suggest a 9.70% (𝑒 − 1) decline in the incidence of violent conflicts in the post-disclosure period and an 8.42% (𝑒 − 1) decline in the incidence of non- violent conflicts. Overall, our results show that the conflict minerals disclosure rule effectively mitigates both types of conflicts. (TABLE 7 ABOUT HERE) Violent and non-violent conflicts are defined in Appendix I. Examples of conflicts are presented in Appendix V. Electronic copy available at: https://ssrn.com/abstract=3908233 4.2.3 Spillover Test To alleviate the concern that the conflict minerals disclosure rule motivated armed groups to move away from mining areas into other regions, we perform an additional test to examine whether incidences of conflicts in the mining areas spilled over to other areas. We test for this by re-estimating equations (3) and (4) using the non-mining region sample. Our results, reported in Table 8, suggest that the enhanced supply chain transparency obtained by firms did not affect non- mining regions in covered countries relative to non-covered countries. Specifically, the coefficients on Post × Covered Country are statistically insignificant across total conflicts as well as both violent and n0on-violent conflicts. (TABLE 8 ABOUT HERE) 4.2.4 Parallel Trends and Robustness We plot the log-adjusted conflicts for each of the three groups in our analyses (all conflicts, violent conflicts, and non-violent conflicts) between the years 2010 and 2019 in Panels A through C of Figure 1. These plots serve two purposes. First, they provide preliminary visual evidence documenting the effect of the conflict minerals disclosure rule. Second, they support the crucial assumption underlying the validity of our DiD estimation: that conflicts in covered countries and those in non-covered countries would have followed parallel trends in the rule’s absence (Bertrand et al., 2004, Gow et al., 2016; Cuny, Even-Tov, and Watts, 2021). All three panels of Figure 1 show that both the treatment group of covered countries and the control group of non-covered countries followed parallel trends in the pre-conflict minerals disclosure period. These groups begin to diverge in 2015, when there is a decrease in the number of conflicts in covered countries and an increase in the number of conflicts in non-covered countries. There are two possible reasons for Consistent with standard difference-in-differences approaches (i.e., Angrist and Pischke [2009]), our identifying assumption is that in the absence of treatment, the difference in levels (and not the percentage changes) would remain the same across groups. Electronic copy available at: https://ssrn.com/abstract=3908233 the increase in covered countries in 2016: election-related political unrest (e.g., President Joseph Kabila’s refusal to step down at the official end of his mandate) and other instabilities related to ethnic conflicts and disputes with the central government. After 2016, the number of conflicts increases at a slower pace in covered countries (and is stagnant for violent conflicts) relative to non-covered countries. Overall, Figure 1 supports the parallel trends assumption and provides preliminary evidence that the conflict minerals disclosure requirement contributes to a reduction in conflicts of mining regions in covered countries relative to those of non-covered countries. Furthermore, this comparative decrease appears to be driven by the maintenance of conflict levels of mining regions in covered countries (except for violent conflicts in 2016), while the number of conflicts continues to increase in mining regions of non-covered countries. (FIGURE 1 ABOUT HERE) We employ three additional tests to support the parallel trends assumption. First, we follow Angrist and Pischke (2009) and Lechner (2011) and test the parallel trends assumption by using pre-treatment time period indicator variables. To do so, we replace the Post × Covered Country indicator in equations (3) and (4) with separate interactions for each of the years in our sample (except for 2014, which serves as the benchmark). The results of this analysis are reported in Table 9. In Figure 2, we graph the Year × Covered coefficient estimates and their corresponding 95% confidence intervals. In support of the parallel-trends assumption, the estimated treatment effects in the pre-treatment period are close to zero and statistically indistinguishable from the benchmark period. In the post-period, consistent with conflict minerals disclosure regime-dampening conflicts in covered countries, treated cells exhibit a gradually decreasing level of conflicts relative to control cells. The gradual decrease in the coefficient estimate is consistent with the notion that it See https://acleddata.com/2016/12/09/democratic-republic-of-congo-december-2016-update/ Electronic copy available at: https://ssrn.com/abstract=3908233 is unlikely for armed groups to abruptly halt their activities. (TABLE 9 ABOUT HERE) (FIGURE 2 ABOUT HERE) In our second analysis, we follow Serfling (2016) and add linear time trends specific to covered and non-covered countries, respectively, to equations (3) and (4). These group-specific time trends control for the possibility that the number of conflicts (also violent conflicts and non- violent conflicts) in covered countries and those in non-covered countries trend differently throughout the sample period due to coincidence with the adoption of conflict minerals disclosure regime. As shown in Table 10, our results are statistically and economically similar after controlling for these trends, which suggests that there is no pre-treatment trend. (TABLE 10 ABOUT HERE) Third, we follow prior studies and perform a placebo test (Amore and Bennedsen, 2013; Derrien and Kecskes, 2013; Gilje and Taillard, 2016; Almeida et al., 2017). Specifically, we re- estimate equations (3) and (4) but re-define our Post indicator to be equal to one for fiscal years after 2011, instead of 2014, while focusing on the sample period between 2010 and 2014 (i.e., pre- Section 1502). The lack of statistical significance across all columns reported in Table 11 confirms that the group was similar prior to 2014 and, thus, that the parallel trends assumption is justified. (TABLE 11 ABOUT HERE) 5. Conclusion The SEC passed Section 1502 of the Dodd-Frank Act, which promulgates the conflict minerals disclosure rule that requires firms to exercise due diligence with respect to the source and chain of custody of the materials that are known for fueling conflicts. Despite the importance of the conflict minerals disclosure rule and the debate about the consequences of the disclosure rule, the literature is largely silent on whether the conflict minerals disclosure mandate has effectively Electronic copy available at: https://ssrn.com/abstract=3908233 reduced human rights abuses. As such, we examine the following two questions: (1) Does enhanced supply chain transparency motivate corporations to dissociate from smelters or refiners that deal in conflict minerals irresponsibly sourced from the covered countries? (2) Do incidences of conflicts in the mining regions of the covered countries decrease after the conflict minerals disclosure requirement becomes effective? We find that firms gradually dissociate from risky smelters and refiners and shift toward responsible sourcing, presumably due to reputational concerns. We also find that the market reacts positively to responsible sourcing. Using PRIO-GRID/year panel data for the period of 2010-2019, we show the conflict decreasing effects of Section 1502 of the Dodd-Frank Act after 2015. Further, we document evidence that the conflicts did not spillover to other regions within the nation. Collectively, we provide evidence that enhanced transparency in the U.S., by promoting responsible sourcing behavior among firms, reduces violence in the DRC and its neighboring countries. Caveats are in order. Conflict minerals are not the only drivers of conflict for the nations that we study. Specifically, the extant literature proposes three other factors that might prolong conflicts in Africa: weak and poorly functioning political institutions, ethnic fragmentation and polarization, and endemic poverty (see Ross 2004). While we have closely followed prior literature (e.g., cell-fixed effects, controlling for GDP, male unemployment rate, etc.), it is difficult to accurately control for all of the determinants of civil conflict. Also, we do not proclaim to have documented a complete cessation to conflicts. While our empirical evidence suggests the conflict minerals disclosure regime advanced human rights and improved the quality of life in the covered countries, the United Nations continues to report serious violations of human rights as of 2019, and according to a 2019 report by the GAO, members of the Congolese national military and police Electronic copy available at: https://ssrn.com/abstract=3908233 continue to derive illegal revenues from smuggling and illicit taxation of minerals from eastern Congolese mines. Our empirical evidence on the real effects of conflict minerals disclosures provides a rationale for evidence-based policymaking (Leuz 2018). First, our findings suggest that transparency can nudge companies to change their sourcing and alleviate unintended consequences of irresponsible sourcing. The DRC supplies approximately 65% of the world’s cobalt, which is powering the transition to clean energy through batteries and other technological innovations. To this end, the addition of cobalt as a conflict mineral aligns with the spirit of the conflict minerals disclosure rule. Lastly, as the Government Accountability Office recommends, performance indicators for assessing progress toward the aims of the conflict minerals disclosure rule should be accounted for. The U.S. Government can leverage data on conflicts, non-governmental organizations’ efforts, and academic research to evaluate the impact of conflict minerals disclosures. See Deberdt, Raphael. 2021. Baseline Study of Artisanal and Small-Scale Cobalt Mining in the Democratic Republic of Congo See https://www.gao.gov/products/gao-20-595 Electronic copy available at: https://ssrn.com/abstract=3908233 APPENDIX I Variable Definitions Variable Definition Responsible Sourcing Variables: % Conflict-Free Smelters/Refiners The ratio of the number of conflict-free smelters/refiners over the number of total smelters/refiners. Dissociation An indicator variable equal to one if there is evidence that a firm suspends its purchases from risky smelters and refiners, and equal to zero, otherwise. Public Attention The number of EDGAR downloads of specialized disclosures from the previous year. CAR The average firm’s market-adjusted abnormal returns cumulated over a 5-day window around the specialized disclosure. Size The natural logarithm of (Market Capitalization) BTM Book Value of Equity / Market Value of Equity Free Cash Flow (Net Cash Flow from Operating Activities – Dividends) / Total Assets ROA Earnings Before Income Tax, Depreciation, and Assets / Total Assets Leverage (Long-Term Debt + Current Liabilities) / Total Assets Ln Cash The natural logarithm of (1 + Cash) Ln Sales The natural logarithm of (Total Annual Sales) Ln Total Assets The natural logarithm of (Total Assets) Electronic copy available at: https://ssrn.com/abstract=3908233 Variable Definition Mining Region Variables Ln Conflicts The natural logarithm of (one plus the number of conflicts) in cell k of year t. This is our main dependent variable and reflects the conflict incidence aggregated at the cell/year- level. Ln Violent The natural logarithm of (one plus the number of violent conflicts) in cell k of year t. The ACLED defines Violent conflicts refer to battles (violent interactions between two organized armed groups), explosions (one-sided violent events in which the tactic creates asymmetry by disabling target response), violence against civilians (violent events where an organized armed group deliberately inflicts violence upon unarmed non-combatants), and riots (violent events where demonstrators or mobs engage in disruptive acts or disorganized acts of violence against property or people). Ln Non-Violent The natural logarithm of (one plus the number of non-violent conflicts) in cell k of year t. The ACLED defines non-violent conflicts as protests (a public demonstration against a political entity, government institution, policy or group in which the participants are not violent) and strategic developments (accounts for often non-violent activity by conflict and other agents within the context of the war/dispute. Recruitment, looting, and arrests are included). Post CMD A dummy variable that takes the value of one starting from the year 2015, zero otherwise (i.e., between 2015 and 2019 in our main specification). Mining Region A dummy variable that equals the value of one if at least a single mine was recorded in any of the cells that surround cell k, zero otherwise, following the specification of Berman et al. (2017). Covered Country Defined by Section 1502 of the Dodd-Frank Act as the DRC and nine other “adjoining countries:” Angola, Burundi, Central African Republic, the Republic of the Congo, Rwanda, South Sudan, Tanzania, Uganda, and Zambia. GDP The sum of gross value added by all resident producers in the economy plus any product taxes and minus any subsidies not included in the value of the products. Data are in current U.S. dollars. Electronic copy available at: https://ssrn.com/abstract=3908233 GDP-per-capita Gross domestic product divided by mid-year population. GDP is the sum of gross value added by all resident producers in the economy plus any product taxes and minus any subsidies not included in the value of the products. Data are in current U.S. dollars. Labor Force The population aged 15 and older that is able to work. Labor Participation Rate The proportion of the population age 15 and older that is economically active: all people who supply labor for the production of goods and services during a specified period. Male Unemployment Rate The share of the male labor force that is without work but available for and seeking employment. Capital Investment Outlays on additions to the economy’s fixed assets plus net changes in the level of inventories. Data are in current U.S. dollars. Electronic copy available at: https://ssrn.com/abstract=3908233 APPENDIX II Covered Countries and Non-Covered Countries of Africa Appendix II shows the Democratic Republic of Congo and its nine adjoining countries— collectively referred to as “covered countries” under Section 1502 of the Dodd-Frank Act—in dark grey and all non-covered countries in light grey. Electronic copy available at: https://ssrn.com/abstract=3908233 APPENDIX III Form Specialized Disclosure Source: https://www.sec.gov/Archives/edgar/data/0000320193/000119312517159397/d383904dsd.htm Electronic copy available at: https://ssrn.com/abstract=3908233 APPENDIX III - Continued Exhibit Conflict Minerals Report Source: https://www.sec.gov/Archives/edgar/data/320193/000119312518073716/d538673dex101.htm Electronic copy available at: https://ssrn.com/abstract=3908233 APPENDIX IV OECD Due Diligence Guidance for Responsible Mineral Supply Chains from Conflict-Affected and High-Risk Areas STEP 1: Establish strong company management systems A) Adopt, and clearly communicate to suppliers and the public, a company policy for the supply chain of minerals originating from conflict-affected and high-risk areas. This policy should incorporate the standards against which due diligence is to be conducted. B) Structure internal management to support supply chain due diligence. C) Establish a system of controls and transparency over the mineral supply chain. This includes a chain of custody or a traceability system or the identification of upstream actors in the supply chain. This may be implemented through participation in industry-driven programs. D) Strengthen company engagement with suppliers. A supply chain policy should be incorporated into contracts and/or agreements with suppliers. Where possible, assist suppliers in building capacities with a view to improving due diligence performance. E) Establish a company-level, or industry-wide, grievance mechanism as an early-warning risk-awareness system. STEP 2: Identify and assess risks in the supply chain A) Identify risks in the supply chain. B) Assess risks of adverse impacts in light of the standards of the supply chain. STEP 3: Design and implement a strategy to respond to identified risks A) Report findings of the supply chain risk assessment to the designated senior management of the company. B) Devise and adopt a risk management plan. Devise a strategy for risk management by either i) continuing trade throughout the course of measurable risk mitigation efforts; ii) temporarily suspending trade while pursuing ongoing measurable risk mitigation; or iii) disengaging with a supplier after failed attempts at mitigation or where a company deems risk mitigation not feasible or unacceptable. Consider the ability to influence, and where necessary take steps to build leverage, over suppliers who can most effectively prevent or mitigate the identified risk. If companies pursue risk mitigation efforts while continuing trade or temporarily suspending trade, they should consult with suppliers and affected stakeholders, including local and central government authorities, international or civil society organizations and affected third parties, where appropriate, and agree on the strategy for measurable risk mitigation in the risk management plan. Companies may draw on suggested measures and indicators to design conflict and high-risk sensitive strategies for mitigation in the risk management plan and measure progressive improvement. C) Implement the risk management plan, monitor and track performance of risk mitigation efforts and report back to designated senior management. This may be done in cooperation and/or consultation with local and central government authorities, upstream companies, international or civil society organizations and affected third-parties where the risk management plan is implemented and monitored in conflict-affected and high- risk areas. D) Undertake additional fact and risk assessments for risks requiring mitigation, or after a change of circumstances. STEP 4: Carry out independent third-party audit of smelter/refiner’s due diligence practices STEP 5: Report annually on supply chain due diligence Source: OECD iLibrary | OECD Due Diligence Guidance for Responsible Supply Chains of Minerals from Conflict-Affected and High-Risk Areas: Third Edition (oecd-ilibrary.org) Electronic copy available at: https://ssrn.com/abstract=3908233 APPENDIX V Examples of Conflicts  On 24 Oct 2018, ADF militiamen attacked Bakaiku neighborhood of Oicha rural commune. They killed three people and burned houses and vehicles before being repelled by the FARDC. The attackers also looted goats, drugs and other property.  On 12 May 2015, as many as 37 people were killed by armed men wielding axes and machetes in two villages, Sabu and Mapiki, in the Mbau area. Thousands of people have fled following the violence. The attack is attributed to the ADF.  On 11 January 2013, ADF-NALU attacked civilians in Kiravo, Mbau. They pillaged and raped 2 women.  On 15 January 2013, FDLR rebels carried out 28 rapes in Lubero throughout January.  On 03 May 2011, FDLR rebels and Mayi Mayi militia (DRC) members attack civilians around the Hauts Plateaux of Bijombo. Reports of looting and raping.  On 19 February 2011, FLDR rebels are witnessed committing 56 rapes in a town in Sud-Kivu in four days.  On 15 April 2013, M23 & FDLR rebels are carrying out abuses on civilians including looting, sexual harassment and violence.  On 15 and 16 July 2015, FRPI elements allegedly raped three women and two girls at Koni village near Aveba.  On 08 March 2015, more than 100 FRPI fighters armed with guns and knives, attacked the Lagabo IDP camp. They reportedly seriously injured 12 people with machetes and raped a woman who refused to help them carry looted goods for them. Soldiers later arrived but the men had already left the area.  On 01 September 2012, LRA raided Balifondo and Zobe Mbari villages of Bangassou in Mbomou on 1 September, abducting 55 people, mostly girls, including 41 in Balifondo. 52 girls were released a week later by the LRA. They were reportedly sexually abused while being detained.  On 13 March 2011, Eight civilians were killed and dozens abducted during an attack on a village by Lords Resistance Army rebels in the Central African Republic mining town of Nzako on Sunday. Between 30 and 50 people were abducted by the rebels, and dozens of properties were looted and/or burnt. 3 more people died while being detained by the group.  On 18 June 2017, 10 LRA fighters attacked a mining camp near Gangala. They looted gold, diamonds and money, and abducted 8 artisanal miners.  On 06 June 2017, LRA forces armed with automatic weapons raided 2 communities near Gangala, abducting 10 civilians. They then looted food, gold, and diamonds from the mine near Gangala. Source: ACLED (acleddata.com) Electronic copy available at: https://ssrn.com/abstract=3908233 REFERENCES Almeida, H., Cunha, I., Ferreira, M. A., & Restrepo, F. 2017. The real effects of credit ratings: The sovereign ceiling channel. The Journal of Finance, 72(1), pp.249-290. Amir, E., Guan, Y. and Oswald, D., 2010. The effect of pension accounting on corporate pension asset allocation. Review of Accounting Studies, 15(2), pp.345-366. Amnesty International, 2015. Amnesty International Report 2015/16. Amore, M. D., & Bennedsen, M., 2013. The value of local political connections in a low- corruption environment. Journal of Financial Economics, 110(2), pp.387-402. Angrist, J.D. and Pischke, J.S., 2009. Instrumental variables in action: sometimes you get what you need. Mostly harmless econometrics: an empiricist’s companion, pp.113-220. Autesserre, S., 2012. Dangerous tales: Dominant narratives on the Congo and their unintended consequences. African Affairs, 111(443), pp.202-222. Auty, R.M., 1993. Sustaining development in mineral economies: The resource curse thesis. London: Rout-ledge. Auty, R.M., 2001. The political economy of resource-driven growth. European economic review, 45(4-6), pp.839-846. Benabou, R. and Tirole, J. 2006. Incentives and Prosocial Behavior. American Economic Review 96 (5), pp.1652–1678. Bens, D.A. and Monahan, S.J., 2008. Altering investment decisions to manage financial reporting outcomes: Asset‐backed commercial paper conduits and FIN 46. Journal of Accounting Research, 46(5), pp.1017-1055. Berman, N. and Couttenier, M., 2015. External shocks, internal shots: the geography of civil conflicts. Review of Economics and Statistics, 97(4), pp.758-776. Berman, N., Couttenier, M., Rohner, D. and Thoenig, M., 2017. This mine is mine! How minerals fuel conflicts in Africa. American Economic Review, 107(6), pp.1564-1610. Bertrand, M., Duflo, E. and Mullainathan, S., 2004. How much should we trust differences-in- differences estimates?. The Quarterly journal of economics, 119(1), pp.249-275. Besley, T. and Reynal-Querol, M., 2014. The legacy of historical conflict: Evidence from Africa. American Political Science Review, pp.319-336. Beyer, A., Cohen, D.A., Lys, T.Z. and Walther, B.R., 2010. The financial reporting environment: Review of the recent literature. Journal of accounting and economics, 50(2-3), pp.296-343. Birkey, R.N., Guidry, R.P., Islam, M.A. and Patten, D.M., 2018. Mandated social disclosure: An analysis of the response to the California Transparency in Supply Chains Act of 2010. Journal of Business Ethics, 152(3), pp.827-841. Cao, Y., Myers, J.N., Myers, L.A. and Omer, T.C., 2015. Company reputation and the cost of equity capital. Review of Accounting Studies, 20(1), pp.42-81. Cheng, Q., Cho, Y.J. and Yang, H., 2018. Financial reporting changes and the internal information environment: Evidence from SFAS 142. Review of Accounting Studies, 23(1), pp.347-383. Christensen, H.B., Floyd, E., Liu, L.Y. and Maffett, M., 2017. The real effects of mandated Electronic copy available at: https://ssrn.com/abstract=3908233 information on social responsibility in financial reports: Evidence from mine-safety records. Journal of Accounting and Economics, 64(2-3), pp.284-304. Christensen, H.B., Hail, L. and Leuz, C., 2021. Mandatory CSR and sustainability reporting: economic analysis and literature review. Review of Accounting Studies, pp.1-73. Collier, P., Hoeffler, A. and Rohner, D., 2009. Beyond greed and grievance: feasibility and civil war. Oxford Economic papers, 61(1), pp.1-27. Cuny, C., Even‐Tov, O., & Watts, E. M., 2021. From implicit to explicit: The impact of disclosure requirements on hidden transaction costs. Journal of Accounting Research, 59(1), pp.215-242. Dambra, M., Even-Tov, O., & Naughton, J. The economic consequences of GASB financial statement disclosure. Working Paper. 2021. Derrien, F., & Kecskés, A. 2013. The real effects of financial shocks: Evidence from exogenous changes in analyst coverage. The Journal of Finance, 68(4), pp.1407-1440. Dhaliwal, D. S., & Reynolds, S. S. 1994. The effect of the default risk of debt on the earnings response coefficient. Accounting Review, pp.412-419. Dyreng, S.D., Hoopes, J.L. and Wilde, J.H., 2016. Public pressure and corporate tax behavior. Journal of Accounting Research, 54(1), pp.147-186. Elayan, F.A., Brown, K., Li, J. and Chen, Y., 2021. The market response to mandatory conflict mineral disclosures. Journal of Business Ethics, pp.1-30. ELM Consulting Group International LLC. 2011. Elm summary, analysis of OECD conflict minerals pilot study report for downstream companies. Available at: https://elmconsultinggroup.wordpress.com/. Friedman, M., 1970. The social responsibility of business is to increase its profits. New York Times Magazine 32 (13), pp.122–126. Geenen, S., 2012. A dangerous bet: The challenges of formalizing artisanal mining in the Democratic Republic of Congo. Resources Policy, 37(3), pp.322-330. Gilje, E. P., & Taillard, J. P. 2016. Do private firms invest differently than public firms? Taking cues from the natural gas industry. The Journal of Finance, 71(4), pp.1733-1778. Gow, I. D., Larcker, D. F., & Reiss, P. C. 2016. Causal inference in accounting research. Journal of Accounting Research, 54(2), pp.477-523. Grewal, J., Riedl, E.J. and Serafeim, G., 2019. Market reaction to mandatory nonfinancial disclosure. Management Science, 65(7), pp.3061-3084. Griffin, P.A., Lont, D.H. and Sun, Y., 2014. Supply chain sustainability: evidence on conflict minerals. Pacific Accounting Review. Gylfason, T., 2001. Natural resources, education, and economic development. European economic review, 45(4-6), pp.847-859. Hainmueller, J., & Xu, Y., 2013. Ebalance: A Stata package for entropy balancing. Journal of Statistical Software, 54(7). Harari, M. and Ferrara, E.L., 2018. Conflict, climate, and cells: a disaggregated analysis. Review of Economics and Statistics, 100(4), pp.594-608. Electronic copy available at: https://ssrn.com/abstract=3908233 Hartzmark, S. M. and Sussman, A. B., 2019. Do Investors Value Sustainability? A Natural Experiment Examining Ranking and Fund Flows. Journal of Finance 74 (6), pp.2789–2837. Healy, P.M. and Palepu, K.G., 2001. Information asymmetry, corporate disclosure, and the capital markets: A review of the empirical disclosure literature. Journal of accounting and economics, 31(1-3), pp.405-440. Healy, P.M. and Serafeim, G., 2020. Voluntary, self-regulatory, and mandatory disclosure of oil and gas company payments to foreign governments. Accounting Horizons, 34(1), pp.111-129. Heffernan, I., 2016. Peace diamonds: Combating civil war with a diamond certification scheme. Hombach, K. and Sellhorn, T., 2018. Firm value effects of targeted disclosure regulation: The role of reputational costs. Available at SSRN 3204505. Hombach, K. and Sellhorn, T., 2019. Shaping corporate actions through targeted transparency regulation: A framework and review of extant evidence. Schmalenbach Business Review, 71(2), pp.137-168. International Resource Panel, United Nations Environment Programme. Sustainable Consumption and Production Branch, 2011. Decoupling natural resource use and environmental impacts from economic growth. UNEP/Earthprint. Ioannou, I. and Serafeim, G., 2017. The consequences of mandatory corporate sustainability reporting. Harvard Business School research working paper, (11-100). Islam, M.A. and Van Staden, C.J., 2018. Social movement NGOs and the comprehensiveness of conflict mineral disclosures: evidence from global companies. Accounting, Organizations and Society, 65, pp.1-19. Kang, J., Loumioti. M. and Wittenberg-Moerman, R., 2021. The harmonization of lending standards within banks through mandated loan-level transparency. Journal of Accounting and Economics, forthcoming Kanodia, C. and Sapra, H., 2016. A real effects perspective to accounting measurement and disclosure: Implications and insights for future research. Journal of Accounting Research, 54(2), pp.623-676. Khan, U., Ryan, S.G. and Varma, A., 2019. Fair value versus amortized cost measurement and the timeliness of other-than-temporary impairments: Evidence from the insurance industry. The Accounting Review, 94(6), pp.285-307. Kim, Y.H. and Davis, G.F., 2016. Challenges for global supply chain sustainability: Evidence from conflict minerals reports. Academy of Management Journal, 59(6), pp.1896-1916. Krueger, P., 2015. Corporate goodness and shareholder wealth. Journal of financial economics, 115(2), pp.304-329. Krueger, P., Sautner, Z., Tang, D.Y. and Zhong, R., 2021. The effects of mandatory ESG disclosure around the world. Available at SSRN 3832745. Lechner, M., 2011. The estimation of causal effects by difference-in-difference methods. Foundations and Trends (R) in Econometrics, 4(3), pp.165-224. Leuz, C., 2018. Evidence-based policymaking: promise, challenges and opportunities for accounting and financial markets research. Accounting and Business Research, 48(5), Electronic copy available at: https://ssrn.com/abstract=3908233 pp.582-608. Leuz, C. and Verrechia, R., 2000. The economic consequences of increased disclosure. Journal of Accounting Research, 38, pp.91-123 Leuz, C. and Wysocki, P.D., 2016. The economics of disclosure and financial reporting regulation: Evidence and suggestions for future research. Journal of Accounting Research, 54(2), pp.525- Lynn, D.M., 2011. The Dodd-Frank Act's specialized corporate disclosure: Using the securities laws to address public policy issues. J. Bus. & Tech. L., 6, pp.327. Ng, J., Yeung, A.C. and Zhang, J.J., 2020. Supply chain visibility and trade credit: Evidence from supply chain-related conflict minerals disclosures. Novak, J. and Bilinski, P. 2018. Social Stigma and Executive Compensation. Journal of Banking and Finance 96, pp.169–184. Parker, D.P. and Vadheim, B., 2017. Resource cursed or policy cursed? US regulation of conflict minerals and violence in the Congo. Journal of the Association of Environmental and Resource Economists, 4(1), pp.1-49. Prendergast, J. and Bafilemba, F., 2018. Congo stories: Battling five centuries of exploitation and greed. Hachette UK. Radley, B. and Vogel, C., 2015. Fighting windmills in Eastern Congo? The ambiguous impact of the ‘conflict minerals’ movement. The Extractive industries and society, 2(3), pp.406-410. Rauter, T., 2020. The effect of mandatory extraction payment disclosures on corporate payment and investment policies abroad. Journal of Accounting Research, 58(5), pp.1075-1116. RepRisk, 2017. RepRisk: Datasets, methodology, and data elements for WRDS subscribers. Ross, M.L., 2004. What do we know about natural resources and civil war?. Journal of peace research, 41(3), pp.337-356. Roychowdhury, S., Shroff, N. and Verdi, R.S., 2019. The effects of financial reporting and disclosure on corporate investment: A review. Journal of Accounting and Economics, 68(2-3), pp.101246. Sachs, J.D. and Warner, A.M., 1995. Natural resource abundance and economic growth (No. w5398). National Bureau of Economic Research. Seay, L., 2012. What’s wrong with Dodd-Frank 1502? Conflict minerals, civilian livelihoods, and the unintended consequences of western advocacy. Center for Global Development Working Paper, (284). Securities and Exchange Commission (SEC), 2012. Final Rule: Conflict minerals (November). https://www.sec.gov/rules/final/2012/34-67716.pdf Serfling, M. 2016. Firing costs and capital structure decisions. The Journal of Finance, 71(5), pp.2239-2286. Shroff, N., 2017. Corporate investment and changes in GAAP. Review of Accounting Studies, 22(1), pp.1-63. Steinmeier, M. and Stich, M., 2019. Does sustainability assurance improve managerial investment Electronic copy available at: https://ssrn.com/abstract=3908233 decisions?. European Accounting Review, 28(1), pp.177-209. Stoop, N., Verpoorten, M. and Van der Windt, P., 2018. More legislation, more violence? The impact of Dodd-Frank in the DRC. PloS one, 13(8), p.e0201783. Swift, C., Guide Jr, V.D.R. and Muthulingam, S., 2019. Does supply chain visibility affect operating performance? Evidence from conflict minerals disclosures. Journal of Operations Management, 65(5), pp.406-429. Vogel, C. and Raeymaekers, T., 2016. Terr (it) or (ies) of peace? The Congolese mining frontier and the fight against “conflict minerals”. Antipode, 48(4), pp.1102-1121. Warfield, T. D., Wild, J. J., & Wild, K. L., 1995. Managerial ownership, accounting choices, and informativeness of earnings. Journal of Accounting and Economics, 20(1), pp.61-91. Weil, D., Fung, A., Graham, M. and Fagotto, E., 2006. The effectiveness of regulatory disclosure policies. Journal of Policy Analysis and Management: The Journal of the Association for Public Policy Analysis and Management, 25(1), pp.155-181. Zhang, H., 2009. Effect of derivative accounting rules on corporate risk-management behavior. Journal of accounting and economics, 47(3), pp.244-264. Electronic copy available at: https://ssrn.com/abstract=3908233 Figure 1 Time-Series Variation in Conflicts for Covered or Non-Covered Countries This figure presents plots the log-adjusted conflicts in covered countries (solid line, left y-axis) and non-covered countries (dashed-line, right y-axis) over our sample period. Panel A shows the variation in the natural log-adjusted number of all conflicts. Panel B shows the variation in natural log-adjusted violent conflicts. Panel C shows the variation in natural log adjusted non-violent conflicts. For each panel, years are displayed on the x-axis. The vertical dashed line partitions our sample by pre- and post-Conflict Minerals Disclosure (CMD) years (i.e., 2010 to 2014 are pre- CMD, and 2015 to 2019 are post-CMD). All variables are defined in Appendix I. Panel A. Ln Conflicts Electronic copy available at: https://ssrn.com/abstract=3908233 Panel B. Ln Violent Panel C. Ln Non-Violent Electronic copy available at: https://ssrn.com/abstract=3908233 Figure 2 Coefficients on Year Dummy x Covered Country of Table 9 This figure presents coefficient estimates and 95% confidence intervals for OLS regressions estimating the effects of the post-CMD decrease in conflicts for each year (See Table 9). Panel A plots the coefficients of column 1 of Table 9. Panel B plots the coefficients of column 2 of Table 9. Panel C plots the coefficients of column 3 of Table 9. All variables are defined in Appendix I. Panel A. Ln Conflicts Electronic copy available at: https://ssrn.com/abstract=3908233 Panel B. Ln Violent Panel C. Ln Non-Violent Electronic copy available at: https://ssrn.com/abstract=3908233 TABLE 1 Descriptive Statistics Panel A provides summary statistics for the variables used in our hand-collected sample. Panel B provides the distribution of the Responsible Sourcing measures by year. All variables are defined in Appendix I. All continuous variables are winsorized at the 1st and 99th percentiles. The sample period for our % of Conflict-Free Smelters/Refiners measure is between 2014 and 2018, while the sample period for the Dissociation measure is between 2014 and 2016 based on data from Developmental International. PANEL A – Descriptive Statistics Variable n Mean S.D. Q1 Median Q3 % of Conflict-Free 1,354 0.716 0.242 0.593 0.776 0.904 Smelters/Refiners Public Attention 1,248 84.546 109.424 34.000 53.000 88.000 Dissociation 1,956 0.522 0.500 0.000 1.000 1.000 CAR 1,956 0.003 0.029 -0.010 0.002 0.016 Size 1,956 7.404 2.139 6.045 7.464 8.828 BTM 1,956 0.622 0.267 0.425 0.595 0.788 Free Cash Flow 1,956 0.049 0.123 0.032 0.069 0.105 ROA 1,956 0.090 0.143 0.066 0.113 0.156 Leverage 1,956 0.432 0.213 0.279 0.423 0.551 Ln Cash 1,956 0.182 0.161 0.059 0.138 0.252 Ln Sales 1,956 7.094 2.021 5.770 7.300 8.430 Ln Total Assets 1,956 7.284 2.051 5.886 7.457 8.637 PANEL B – Responsible Sourcing by Year Variable n Mean S.D. n Mean S.D. % of Conflict-Free Responsible Sourcing Dissociation Smelters/Refiners 2014 147 0.446 0.265 669 0.450 0.498 2015 251 0.637 0.256 656 0.477 0.500 2016 323 0.726 0.215 631 0.645 0.479 2017 317 0.792 0.186 2018 316 0.819 0.172 Total 1,354 0.716 0.242 1,956 0.522 0.500 Electronic copy available at: https://ssrn.com/abstract=3908233 TABLE 2 Public Attention and Responsible Sourcing This table reports the regression results for the Responsible Sourcing variables on Public Attention. Column (1) reports the regression of % of Conflict-Free Smelters/Refiners, the ratio of the number of conflict-free smelters/refiners over the number of total smelters/refiners, on Public Attention with controls. Column (2) reports the probit regression of Dissociation, an indicator of evidence that a firm has already suspended or intends to suspend its purchases from problematic smelters and refiners, on Public Attention with controls. The variable of interest is Public Attention, the previous year’s number of EDGAR downloads of a firm’s specialized disclosure. All variables are defined in Appendix I. Standard errors are clustered at the industry level. All continuous variables are winsorized at the 1st and 99th percentiles. T-statistics are reported in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively. Dependent Variables % of Conflict-Free Smelters/Refiners Dissociation (1) (2) Independent Variables Public Attention 0.012** 0.045*** 2.35 3.34 Free Cash Flow 0.122 0.493 1.78 0.85 ROA -0.072 -0.498 (-0.75) (-1.15) Ln Cash -0.038 -0.343 (-0.30) (-0.97) Leverage 0.064** -0.082 2.67 (-0.52) BTM -0.01 -0.242*** (-0.43) (-2.78) Ln Sales -0.012 0.143 (-0.45) 1.26 Ln Total Assets -0.003 -0.062 -0.18 (-0.62) Industry Fixed Effects Yes Yes Year Fixed Effects Yes Yes 0.085 0.052 Pseudo/Adjusted R Observations 920 1,248 Electronic copy available at: https://ssrn.com/abstract=3908233 TABLE 3 Univariate Analysis of Conflict Minerals Disclosures by Dissociation This table reports the means and differences-in-means of variables used in analysis, partitioned by whether there is evidence that a firm has already suspended or intends to suspend its purchases from problematic smelters using the dissociation dummy (Dissociation). Panel A (Panel B) provides our descriptive statistics across our unweighted sample variables (entropy-balanced weighted control variables) between firms classified as Dissociation = 1 and those classified as Dissociation = 0. All variables are defined in Appendix I. All continuous variables are winsorized at the 1st and 99th percentiles. T-statistics are reported in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively. PANEL A: Univariate Analysis Dissociation = 0 Dissociation = 1 Mean S.D. Mean S.D. Difference Variable (1) (2) (3) (4) (3) – (1) CAR 0.001 0.030 0.004 0.028 0.003** (2.33) Size 7.193 2.186 7.598 2.076 0.405*** (4.20) BTM 0.640 0.276 0.606 0.258 -0.033*** (-2.77) Leverage 0.432 0.218 0.432 0.207 0.000 (0.04) PANEL B: Univariate Analysis After Entropy Balancing Dissociation = 0 Dissociation = 1 Mean S.D. Mean S.D. Difference Variable (1) (2) (3) (4) (3) – (1) Size 7.598 2.257 7.598 2.076 0.000 (0.00) BTM 0.618 0.280 0.606 0.258 -0.012 (-1.01) Leverage 0.432 0.221 0.432 0.207 0.000 (0.00) Electronic copy available at: https://ssrn.com/abstract=3908233 TABLE 4 Market Reactions to Conflict Minerals Disclosures This table reports the regression results for the five-day market-adjusted cumulative abnormal return (CAR) on the Responsible Sourcing variables. Columns (1) and (3) report the regressions of CAR on the Responsible Sourcing variables without controls. Columns (2) and (4) report the regression results including the control variables. Columns (3) and (4) report the results for weighted ordinary least squares regressions where we both control for and entropy balance the mean, variance, and skewness between the two groups of all control variables. The sample period for our % of Conflict-Free Smelters/Refiners measure is between 2014 and 2018, while the sample period for the Dissociation measure is between 2014 and 2016 based on data from Developmental International. All variables are defined in Appendix I. Standard errors are clustered at the firm and year level. All continuous variables are winsorized at the 1st and 99th percentiles. T-statistics are reported in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively. Dependent variable = CAR (1) (2) (3) (4) % of Conflict-Free 0.007** 0.007** Smelters/Refiners (2.27) (2.15) Dissociation 0.003** 0.003** (2.37) (2.35) Size -0.001 -0.000 (-1.03) (-0.59) BTM -0.004 -0.001 (-1.14) (-0.34) Leverage 0.004 0.004 (0.94) (1.21) 0.003 0.003 0.002 0.002 Adjusted R Observations 1,340 1,340 1,956 1,956 Electronic copy available at: https://ssrn.com/abstract=3908233 TABLE 5 Sample Construction of Grid-Level Analysis This table provides details on the construction of our final sample of mining region cells. Number of Observations All 10,335 cells over a 10-year sample period between 2010 and 2019 103,350 Less: Cells that correspond to countries in which ACLED does not (110) report conflicts Cells with missing control variables (13,462) Restrict to mining regions only (71,116) Final sample 18,662 Electronic copy available at: https://ssrn.com/abstract=3908233 TABLE 6 Descriptive Statistics of Mining Regions Panel A provides descriptive statistics of mining regions for the years 2010–2019. Panel B provides the pairwise Pearson (Spearman) correlations among variables of mining regions in the upper (lower) triangular region. Correlations that are statistically significant at the 5% level of significance are reported in bold. All variables are defined in Appendix I. All continuous variables are winsorized at the 1st and 99th percentiles. PANEL A – Descriptive Statistics of Our Final Sample of Mining Regions Variable Mean S.D. Min Q1 Median Q3 Max Ln Conflicts 0.267 0.639 0 0 0 0 3.367 Ln Violent 0.181 0.476 0 0 0 0 3.045 Ln Non-Violent 0.131 0.405 0 0 0 0 2.303 GDP (billion USD) 107.503 138.139 0.85 13.450 29.31 161.210 568.5 GPD-per-capita (USD) 2949.835 2428.904 234.24 787.240 1743.85 5303.310 8279.6 Labor Force (mil. people) 12.929 11.397 0.32 5.610 10.94 20.360 58.4 Labor Participation Rate 63.081 12.649 41.15 53.050 62.73 74.680 86.65 Capital Investment (billion USD) 24.801 28.817 0 3.110 8.57 41.010 97.4 Male Unemployment Rate 10.280 7.449 0.39 4.760 8.11 14.670 26.39 PANEL B – Pairwise Correlations of Mining Regions Variable (1) (2) (3) (4) (5) (6) (7) (8) (9) (1) Ln Conflicts 0.91 0.86 0.08 -0.02 0.09 -0.02 0.06 0.00 (2) Ln Violent 0.88 0.61 0.08 -0.03 0.10 0.02 0.04 -0.01 (3) Ln Non-Violent 0.76 0.51 0.06 0.01 0.04 -0.06 0.05 0.02 (4) GDP 0.09 0.08 0.08 0.65 0.63 -0.47 0.92 0.63 (5) GDP-per-capita -0.01 -0.03 0.03 0.56 0.05 -0.43 0.65 0.84 (6) Labor Force 0.09 0.10 0.05 0.78 0.01 -0.09 0.54 0.04 (7) Labor Participation Rate 0.00 0.03 -0.05 -0.39 -0.47 -0.11 -0.55 -0.47 (8) Capital Investment 0.04 0.03 0.05 0.93 0.57 0.71 -0.49 0.56 (9) Male Unemployment Rate -0.05 -0.06 -0.01 0.43 0.81 -0.02 -0.59 0.45 Electronic copy available at: https://ssrn.com/abstract=3908233 TABLE 7 Conflicts in Mining Regions around Conflict Minerals Disclosures: DID Regression Analysis This table reports the regression results of the differential changes in the number of conflicts pre- and post-Conflict Minerals Disclosure based on whether the country is covered under Section 1502 of the Dodd-Frank Act. The sample period spans from January 1, 2010 to December 31, 2019. The dependent variable, Ln Conflicts (Ln Violent) {Ln Non-Violent}, is the natural log of (1 + the number of all (violent) {non-violent} conflicts) in cell k during year t. The variable of interest is Post × Covered Country. Post is an indicator variable equal to one for years after 2014. Covered Country is an indicator variable equal to one if the cell is within the Democratic Republic of Congo or an adjoining country. Columns (2), (4), and (6) include cell and year fixed effects. All variables are defined in Appendix I. Standard errors are clustered at the country level. All continuous variables are winsorized at the 1st and 99th percentiles. T-statistics are reported in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively. Ln Conflicts Ln Violent Ln Non-Violent Independent Variables (1) (2) (3) (4) (5) (6) Post 0.163*** 0.100*** 0.090*** (16.08) (13.14) (14.05) Covered Country -0.079*** -0.037*** -0.061*** (-4.63) (-2.93) (-5.59) Post × Covered Country -0.098*** -0.164*** -0.050*** -0.102*** -0.056*** -0.088*** (-3.93) (-5.59) (-2.70) (-3.70) (-3.56) (-4.52) GDP 0.000 0.000 0.000 (0.17) (0.12) (0.07) GDP-per-capita -0.000** -0.000 -0.000* (-2.19) (-1.07) (-2.03) Labor Force 0.009 0.016 0.000 (0.73) (1.55) (0.03) Labor Participation Rate -0.017*** -0.009* -0.010** (-2.96) (-1.91) (-2.67) Capital Investment 0.007** 0.003 0.005*** (2.09) (1.24) (3.07) Male Unemployment Rate -0.017* -0.006 -0.010 (-1.87) (-0.94) (-1.62) Electronic copy available at: https://ssrn.com/abstract=3908233 Cell Fixed Effects No Yes No Yes No Yes Year Fixed Effects No Yes No Yes No Yes Adjusted R 0.019 0.609 0.012 0.523 0.017 0.567 Observations 18,662 18,662 18,662 18,662 18,662 18,662 Electronic copy available at: https://ssrn.com/abstract=3908233 TABLE 8 Spillover Tests This table reports whether there are spillovers of different conflict types into non-mining regions of Africa around the conflict minerals disclosures. The sample period spans January 1, 2010 to December 31, 2019. The dependent variable, Ln Conflicts, is the natural log of (1 + the number of conflicts) in cell k during year t. The dependent variable, Ln Violent (Ln Non-Violent), is the natural log of (1 + the number of violent (non-violent) conflicts) in cell k during year t. The variable of interest is Post × Covered Country. Post is an indicator variable equal to one for years after 2014. Covered Country is an indicator variable equal to one if the cell is within the Democratic Republic of Congo or an adjoining country. All variables are defined in Appendix I. Standard errors are clustered at the country level. All continuous variables are winsorized at the 1st and 99th percentiles. T-statistics are reported in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively. Non-Mining Regions Ln Conflicts Ln Violent Ln Non-Violent Independent Variables (1) (2) (3) Post × Covered Country -0.043 -0.021 -0.016 (-1.34) (-0.82) (-0.76) GDP 0.001** 0.001*** 0.000 (2.25) (2.77) (1.34) GDP-per-capita 0.000 0.000 -0.000 (0.40) (0.93) (-1.25) Labor Force 0.015** 0.010 0.011*** (2.52) (1.60) (3.76) Labor Participation Rate -0.013 -0.010 -0.004 (-0.96) (-0.95) (-0.72) Capital Investment -0.003* -0.003* -0.000 (-1.68) (-1.99) (-0.69) Male Unemployment Rate 0.007 0.008 0.003 (0.52) (0.66) (0.61) Cell Fixed Effects Yes Yes Yes Year Fixed Effects Yes Yes Yes Adjusted R 0.680 0.633 0.622 Observations 71,116 71,116 71,116 Electronic copy available at: https://ssrn.com/abstract=3908233 TABLE 9 Time-Period Indicator Tests This table checks for the parallel trend assumption using time-period indicator variables. The sample period spans from January 1, 2010 to December 31, 2019. The dependent variable, Ln Conflicts, is the natural log of (1 + the number of conflicts) in cell k during year t. The dependent variable, Ln Violent (Ln Non-Violent), is the natural log of (1 + the number of violent (non-violent) conflicts) in cell k during year t. Covered Country is an indicator variable equal to one if the cell is within the Democratic Republic of Congo or an adjoining country. Year is a set of indicator variables for each year in our sample period. We estimate the model from Table 7 but replace the indicator variable Post × Covered Country with separate interactions for each of the years in our sample (except for 2014, which serves as the benchmark). All variables are defined in Appendix I. Standard errors are clustered at the country level. All continuous variables are winsorized at the 1st and 99th percentiles. T-statistics are reported in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively. Ln Conflicts Ln Violent Ln Non-Violent Independent Variables (1) (2) (3) Year 2010 × 0.094* 0.057 0.040 Covered Country (1.91) (1.34) (1.50) Year 2011 × 0.063 0.050 0.018 Covered Country (1.14) (1.04) (0.67) Year 2012 × -0.027 -0.013 -0.024 Covered Country (-0.66) (-0.41) (-1.33) Year 2013 × 0.010 0.026 -0.017 Covered Country (0.56) (1.30) (-1.27) Year 2014 × Dropped Dropped Dropped Covered Country Year 2015 × -0.100*** -0.062** -0.057** Covered Country (-2.96) (-2.47) (-2.52) Year 2016 × -0.059 -0.006 -0.072*** Covered Country (-1.28) (-0.16) (-2.77) Year 2017 × -0.099* -0.056 -0.068** Covered Country (-2.01) (-1.37) (-2.32) Year 2018 × -0.194*** -0.139** -0.092*** Covered Country (-3.13) (-2.59) (-3.00) Year 2019 × -0.374*** -0.218*** -0.202*** Covered Country (-4.58) (-3.19) (-2.94) Controls Yes Yes Yes Cell Fixed Effects Yes Yes Yes Year Fixed Effects Yes Yes Yes Adjusted R 0.611 0.524 0.568 Observations 18,662 18,662 18,662 Electronic copy available at: https://ssrn.com/abstract=3908233 TABLE 10 Group Time Trends This table reports the robustness checks of controlling for group-specific linear time trends. The sample period spans from January 1, 2010 to December 31, 2019. The dependent variable, Ln Conflicts, is the natural log of (1 + the number of conflicts) in cell k during year t. The dependent variable, Ln Violent (Ln Non-Violent), is the natural log of (1 + the number of violent (non-violent) conflicts) in cell k during year t. The variable of interest is Post × Covered Country. Post is an indicator variable equal to one for years after 2014. Covered Country is an indicator variable equal to one if the cell is within the Democratic Republic of Congo or and its adjoining countries, as defined by the Act. All variables are defined in Appendix I. Standard errors are clustered at the country level. All continuous variables are winsorized at the 1st and 99th percentiles. T-statistics are reported in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively. Ln Conflicts Ln Violent Ln Non-Violent Independent Variables (1) (2) (3) Post -0.011 0.009 -0.022 (-0.40) (0.56) (-1.01) Post × Covered Country -0.168*** -0.104*** -0.090*** (-5.45) (-3.64) (-4.62) Time Trends Yes Yes Yes Controls Yes Yes Yes Cell Fixed Effects Yes Yes Yes Year Fixed Effects No No No Adjusted R 0.606 0.522 0.564 Observations 18,662 18,662 18,662 Electronic copy available at: https://ssrn.com/abstract=3908233 TABLE 11 Placebo Tests This table reports our placebo test results. The sample period spans from January 1, 2010 to December 31, 2014. The dependent variable, Ln Conflicts, is the natural log of (1 + the number of conflicts) in cell k during year t. The dependent variable, Ln Violent (Ln Non-Violent), is the natural log of (1 + the number of violent (non-violent) conflicts) in cell k during year t. The variable of interest is Post × Covered Country. Covered Country is an indicator variable equal to one if the cell is within the Democratic Republic of Congo or an adjoining country. Post is an indicator variable equal to one starting after 2012 in columns (1), (2), and (3). All variables are defined in Appendix I. Standard errors are clustered at the country level. All continuous variables are winsorized at the 1st and 99th percentiles. T-statistics are reported in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively. Ln Conflicts Ln Violent Ln Non-Violent Independent Post = 2013-2014 Post = 2013-2014 Post = 2013-2014 Variables (1) (2) (3) Post × 0.048 0.052 0.012 Covered (1.07) (1.43) (0.47) Country Controls Yes Yes Yes Cell Fixed Effects Yes Yes Yes Year Fixed Effects Yes Yes Yes Adjusted R 0.604 0.515 0.571 Observations 7,672 7,672 7,672 Electronic copy available at: https://ssrn.com/abstract=3908233

http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png

ARN Conferences & Meetings SSRN

http://www.deepdyve.com/lp/ssrn/the-real-effects-of-conflict-minerals-disclosures-z3Z0i6UaDO

The Real Effects of Conflict Minerals Disclosures

Loading next page...

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

eISSN: ARCH-4059
DOI: 10.2139/ssrn.3908233
Publisher site: See Article on Publisher Site

Abstract

Journal

ARN Conferences & Meetings – SSRN

Published: Aug 19, 2021

Keywords: Real effects, Dodd-Frank Act, Conflict minerals disclosures, Corporate social responsibility, Responsible sourcing, ESG

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

The Real Effects of Conflict Minerals Disclosures

The Real Effects of Conflict Minerals Disclosures

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

The Real Effects of Conflict Minerals Disclosures

The Real Effects of Conflict Minerals Disclosures

References

Abstract

Journal

Recommended Articles

There are no references for this article.

Our policy towards the use of cookies