Note: EPA no longer updates this information, but it may be useful as a reference or resource.
Dialogue to Explore the Use of EPA Data in Financial and Investment Analysis
New York, NY
June 19, 2008
The Environmental Protection Agency (EPA) and the New York Society of Security Analysts (NYSSA) co-hosted a dialogue on the use of EPA data and its utility to investment analysis and decision making. This dialogue, held June 19, 2008, was a continuation of previous discussions and stemmed from EPA’s desire to bring the use of its data to the mainstream financial analyst.
EPA wishes to better understand how its data and related systems can provide additional value to users in the financial services industry. The emphasis throughout the dialogue was on sharing information and perspectives by those who produce, disseminate, process, analyze, and use EPA’s environmental management and performance data.
Meeting participants included current EPA staff members (including headquarters and regional offices), previous EPA staff members now in the financial community, financial community members, former financial community members, insurance community members, mutual fund representatives, research vendors, investment consultants, pension fund representatives, academics, consultants to EPA, journalists, and representatives of regulated entities.
Participants were encouraged to seek out their counterparts from industries with which they are not familiar and begin to dialogue with each other. The meeting also offered an opportunity for EPA to learn about the questions financial community clients are asking so as to better understand the needs of the financial community.
Questions used to guide the discussion included the following:
- Where and when are financial markets considering environmental factors?
- Will the use of these factors increase or decrease?
- Do we know in which areas their use will increase?
- Are there specific tools the financial community needs to make decisions?
- How can EPA support financial sector decision making with existing expertise and data and what new data and expertise can it develop?
EPA representatives provided an overview of the Agency’s major data bases and systems. Important points included the following:
- Much of EPA’s data collection activity is mandated by statute, e.g., the Clean Air Act, Clean Water Act, and Superfund (Comprehensive Environmental Response, Compensation, and Liability Act), and is implemented at the regional and state levels. Accordingly, the data collected by EPA may reside in a number of different places and exist in different forms and formats. These distributed data often are not shared with EPA headquarters. Further, the program offices “own” the data, and control which other EPA and non-EPA entities have access to particular data sets.
- EPA data comes from multiple sources, including permit applications, inventory submissions, accident reporting, and monitoring, with up to 95 percent of data being self-reported or from delegated programs. Much of the Agency’s data collection is at arms length.
- Currently there are 2.4 million facilities under EPA regulation (air, water, etc.) that must report data. Related data comes from state-regulated facilities (often reported by the states), EPA regional offices and labs, continuous monitoring stations, and reports of incidents (which may be provided by the general public).
- Challenges are raised by differences in terminology and whether data reflect forward looking (e.g., permitting process), concurrent (e.g., Clean Air markets), or historical (e.g., TRI, Superfund, brownfields, remediation) situations. Most EPA data are historical or lagging data.
- Information gaps, i.e., data that are not collected by EPA, include use of underground storage tanks (which likely number between one and four million), above ground storage tanks (estimated at one million or more), as well as spills (the U.S. Coast Guard maintains records).
- EPA is taking steps to develop more centralized and standardized access to its data bases. One major initiative is the Facility Registry System (FRS), which brings place-based information into one database, with 79 different environmental classifications and 19 broader program classifications. Currently, data are collected from 40 states into this data base. FRS data are used by TRI Explorer, ECHO, EnviroFacts, and RSEI, among other data bases of potential interest to the financial community. FRS is EPA’s largest database, with 2.4 million records. FRS functions much like a telephone book, in that it contains only limited summary information, though it also allows a focus on specific locations such as regulated outfalls and stacks. It contains company names (legal and alternative), location, mailing address, point of contact, industrial codes (SIC/NAICS), and Dun & Bradstreet numbers (not always populated), as well as a mapping function.
- Currently, parent company information is not routinely collected; tax ID numbers are never collected, and Dun & Bradstreet numbers are rarely collected.
- Terms used by EPA may be different than those used in the financial community (e.g., open/closed, active/inactive, permanently closed/closed). Once data are in some regulatory systems, these data are never taken out, so users should be careful about inferring, for example, facility operating status from these data.
The perspective of financial community users of EPA and other environmental data was provided through panel discussions, and extensive audience interaction. Major issues raised are presented below, organized by session.
The User Perspective
The next session explored how environmental data support business decision making in an investment/underwriting context, and discussed the utility of EPA data in formulating these decisions.
A representative of a major provider of environmental insurance described the information needs of his firm in underwriting of new policies and responding to claims. Underwriting work occurs very rapidly, with a typical required turnaround of about one day. Claims result from existing policies, and are addressed using a set of established frameworks to evaluate key engineering and regulatory requirements, among other factors, to determine the company’s responsibility to pay.
The major EPA sources used to support these activities are ECHO and EnviroFacts. ECHO is used by analysts to document a potential insured’s compliance history (e.g., numbers and types of violations, inspections). As a general matter, one or a small number of minor administrative violations are not of concern from an underwriting standpoint, but multiple and/or serious violations are. EnviroFacts is a gateway to seven underlying databases, which can be searched concurrently with one query, enabling significant time savings. Insurance analysts use the data retrieved to infer possible liabilities from both current and future operations.
Analysts also use EPA’s remedial action databases to assess other issues, such as possible changes in land use (in the context of cleanup standards) and successful application of emerging remediation technologies. To find data on the latter, insurance analysts explore the relevant links on EPA CLU-IN, ITRC, and other web sites maintained by EPA and other federal government entities.
“There are no substitutes for CERCLIS, TRI, BRS and other EPA datasets.”
A representative of an environmental information provider with an extensive financial services industry clientele indicated that he uses several EPA databases in his work, including the Toxics Release Inventory (TRI), Comprehensive Environmental Response, Compensation, Liability Information System (CERCLIS), Biennial Reporting System (BRS), Energy Star, and Performance Track. He reviews TRI data, which are normalized for comparison purposes by dividing total TRI releases by sales. He also identifies key chemicals and assesses the possibility of catastrophic events by reviewing EPA’s Risk Screening Environmental Indicators (RSEI) model. He suggested that EPA and other environmental data are often used qualitatively, as part of his firm’s development of its overall company ratings.
To put the perceived value of EPA data into context, he also stated that there are no substitutes for some of EPA’s major data sets, including CERCLIS, TRI, and BRS. There also are few analogues internationally, though Japan does maintain a database that has certain similarities to TRI.
An academic researcher noted several intrinsic limitations in using and interpreting EPA data. These include the regulatory cut-offs (measurement/reporting thresholds) that exist in many programs. This leads to biased samples, and in turn, a tendency to confound or confuse the effects of policy or other stimuli, and inaccurate metrics, among other undesirable effects.
Another representative of a data provider described how her company uses EPA and other environmental data as part of its work in preparing profiles about 21,000 companies according to about 12 types of risk, which include environmental, social, and governance (ESG) risk. Primary consumers of these profiles are socially responsible investing (SRI) firms. Her firm collects and tracks data on four major environmental endpoints for approximately 2,200 U.S. companies:
- Superfund site liabilities
- Compliance, and
- TRI emissions.
EPA data bases are the major source of the data used to evaluate these endpoints.
Major Challenges in Using EPA Data
- Data consistency, quality, and ease of access
- Limited or lack of availability of certain types of data
- Timing of data releases
- Data formats and documentation (e.g., Adobe Acrobat files are difficult to manipulate)
- Comparisons across geographies (e.g., U.S. vs. Canada, Japan, and/or Australia), and
- Matching of company identifiers across records and data bases.
She also gave an example of how important an ESG risk overlay can be to an investor, by describing how the risk continuum within the energy industry becomes reordered when supply concerns and possible limits on greenhouse gas emissions are considered. That is, otherwise low-risk firms can become substantially more risky if they are particularly vulnerable to either supply interruptions or future greenhouse gas emission restrictions.
She suggested that EPA make a wider array of data available to the public via download. Currently, her firm uses the Freedom of Information Act (FOIA) process to obtain certain types of data, a process that is expensive and time-consuming.
The meeting’s keynote speaker was a senior official within EPA’s Office of Environmental Information (OEI), who provided remarks during lunch. She identified several factors driving EPA’s efforts to improve access to its data and information. These include a general increase in interest in environmental information and the explosion of the Internet and computer-based tools that enable users to personalize and customize information. She stated that EPA is working to address these drivers, and looking for ways in which to actively anticipate the needs and preferences of the public and proactively develop and deliver environmental data to those who need it.
She also shared the general themes that EPA has heard during a series of listening sessions with groups that use EPA data. Users want raw data so that they can interpret it themselves. Data needs to be available more quickly and be searchable.
She encouraged participants to ask hard questions and think big. While solutions are coming in the future, they may take time to implement. She also encouraged participants to share their suggestions about changes that can be made in the meantime to improve access to and use of data.
Following lunch, a moderated discussion focusing on opportunities to improve the value of EPA Data to financial community users took place. The session moderator voiced the hope that the discussion would take a big picture view of the issues related to EPA data, and asked more specifically whether there is enough EPA data to cause people to use it to make decisions.
A panelist summarized the major threads of the discussion in the morning session. These included the value of and annoyance with using EPA data, which was felt to be directly related to what each company did with the data. Those who use data for specific firms or facilities experience in fewer problems with data aggregation. In contrast, those who look at environmental performance as it relates to overall management and performance of firms often believe that the data require excessive amounts of time to aggregate and process before they can be analyzed. Also, while the data are quite valuable, aggregation of data may result in bias as, for example, when firms drop out of an EPA monitoring program because of good performance; this good performance is then not reflected in EPA’s data. Perhaps most significantly, there are long-standing concerns that the typical use of environmental data is focused on the data that exist, rather than on those that are most important.
Meeting participants offered several suggestions for how EPA could potentially make transformative improvements in its data design, compilation, and display to make these data indispensable. One was that EPA should define metrics that could identify environmental liabilities. From an investment stand point, investors are interested in companies that have good environmental policies and practices. Data on this would be useful. Another was to consider changing the format of EPA reports, especially those used for financial analysis. It is important that data can be readily pulled into spreadsheets or other commonly used tools.
Another panelist contended that fundamental financial analysis probably already takes environmental issues into account, but perhaps using different language than used within the ESG/SRI (environmental, social, and governance or socially responsible investing) communities. Sometimes the companies themselves are not producing or releasing environmental information in a useful way. He also emphasized the importance of understanding the relationship between environmental performance and financial performance of companies. He postulated that once financial institutions understand the relationship of financial performance to environmental performance to EPA data (a “chain of relationships”), EPA’s data will be more useful. He also stressed the importance of subgroups and stakeholder groups using a common language to describe the same ideas. Much analysis depends on the context (reflected in metadata), and should be based on understand resulting from looking at multiple data sets to develop a more nuanced context. This more integrative approach is particularly important because each EPA data set offers a very narrow view of company activity related to the environment.
Another issue that was raised during this discussion had to do with the adequacy of existing regulations regarding environmental disclosure and whether additional regulation might be warranted. Several participants suggested that current laws are not strongly enforced, and in particular, that the U.S. Securities and Exchange Commission could do more to both enforce disclosure requirements and collect and share data. One panelist said that it is important to get EPA data into SEC reports as the only widely available information addressing company-level environmental liabilities. He and other participants reiterated their desire for EPA to link company names with their parent companies.
A meeting participant asked those in attendance how many analysts inquire whether companies are meeting their environmental targets, and one individual answered affirmatively. Another added that there are sources for identifying targets, but not necessarily for identifying success in meeting them. A data provider indicated that his firm uses a 100-point scale to address these issues, while another said this information is really dependent on voluntary disclosure by companies.
Several participants said it was not EPA’s responsibility to provide analysis to the financial community if protecting human health and environment is EPA’s mission; however, EPA should help the financial community understand the context of its data in comparing behaviors reported in the data. A panelist added that as EPA changes its report formats and provides information analysts can use, companies will respond to the ratings the analysts produce.
The meeting ended with two senior EPA officials representing OEI and the Office of Policy, Economics, and Innovation. They expressed an interest in both continuing the dialog and defining some near-term actions that could be taken to improve the access to and utility of EPA data. They informed participants that EPA plans to release a draft document on the national dialogue on improving access to environmental information later this year. They also acknowledged that EPA needs to work with the financial community to leverage other people and organizations to support its broader environmental mission. He specifically noted that it may be possible to address the parent company identification issue in the TRI data base, and suggested that this might be an example of a concrete, productive step forward that could be accomplished relatively quickly.
They also pointed out several additional issues that may be worthy of further thought and discussion. These included how best to deal with large, important issues that are affected by EPA decisions, e.g., granting a waiver to a state program (or not) but are not routinely captured in EPA (or other) databases. Such decisions and their consequences may be far more important in assessing company/sector risks and opportunities than, for example, TRI data. How best to capture and report such information is, however, largely unexplored territory. They also acknowledged many of the previously offered suggestions regarding EPA’s data quality (e.g., obsolete fields in ECHO); specific concerns with TRI (data not aggregated or organized by firm); lack of comprehensive access tools (i.e., a system analogous to SEC’s Edgar Online); translation and terminology (consistency and clarity); and EPA’s role in self-disclosure by the regulated community.
Comments made by participants related to the situation concerning the use of environmental information internationally (e.g., China); how to access and interpret management quality information from EPA leadership and other programs; and linking data points to firms, but not (necessarily) organizing the data by firm.
The meeting ended with general agreement that the meeting addressed technical and normative issues, short and long term issues, points where innovation can occur, and points of interest where EPA and the financial community can help one another. The EPA and financial industry organizers concluded that it is now time to break down into smaller working groups and solicit volunteers willing to participate in further dialogues.
Financial Services Industry
Summary of Financial Services Industry Representatives
Approximate Number of Participants
|EPA/other regulatory agency||23|
Types of Environmental Information
The financial services industry participants indicated that the following EPA databases were most commonly used within their firms and sectors:
- Biennial Reporting System (BRS)
- CERCLIS (Superfund site information)
- Enforcement & Compliance History Online (ECHO)
- Energy Star
- EPA CLU-IN (hazardous waste cleanup information)
- Integrated Risk Information System (IRIS)
- Performance Track
- Risk Screening Environmental Indicators (RSEI)
- Toxics Release Inventory (TRI)