# g% \
%w:s
'% ,,01^°
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF INSPECTOR GENERAL
EPA Needs to Demonstrate
Whether It Has Achieved the
Goals It Set Under the National
Petroleum Refinery Initiative
Report No. 14-P-0184
April 15, 2014
Scan this mobile
code to learn more
about the EPA OIG.
-------
This is one of the U.S. Environmental Protection Agency Office of Inspector General's
products associated with climate change. For details on our other reports on climate
change, go to http://www.epa.gov/oig/climatechange.
Report Contributors: Kathlene Butler Julie Hamann
Allison Dutton Fredrick Light
Dan Engelberg Danielle Tesch
Abbreviations
ACS
Annual Commitment System
CAA
Clean Air Act
CEMS
Continuous Emissions Monitoring System
DOJ
U.S. Department of Justice
EPA
U.S. Environmental Protection Agency
FY
Fiscal year
GPRA
Government Performance and Results Act
NEI
National Enforcement Initiative
NEIC
National Enforcement Investigation Center
NOx
Nitrogen oxides
NPRI
National Petroleum Refinery Initiative
OECA
Office of Enforcement and Compliance Assurance
OIG
Office of Inspector General
so2
Sulfur dioxide
Cover photo: A petroleum refinery has multiple connection points that represent a
potential source of emissions leaking into the environment. (EPA photo)
Hotline
To report fraud, waste or abuse, contact us
through one of the following methods:
email:
OIG Hotline@.epa.qov
phone:
1-888-546-8740
fax:
1-202-566-2599
online:
http://www.epa.aov/oia/hotline.htm
write:
EPA Inspector General Hotline
1200 Pennsylvania Avenue, NW
Mailcode 2431T
Washington, DC 20460
Suggestions for Audits or Evaluations
To make suggestions for audits or evaluations,
contact us through one of the following methods:
email: OIG WEBCOMMENTS@epa.gov
phone: 1-202-566-2391
fax: 1-202-566-2599
online: http://www.epa.aov/oia/contact.html#Full Info
write: EPA Inspector General
1200 Pennsylvania Avenue, NW
Mailcode 241OT
Washington, DC 20460
-------
£0 STAf. 14-P-0184
s *. U.S. Environmental Protection Agency April 15,2014
£ "*
<
• UiOi d IVII Ul 111ICIIIdl r I vlCvll
¦¦ \ Office of Inspector General
At a Glance
Why We Did This Review
The objective of this evaluation
was to determine to what
extent the U.S. Environmental
Protection Agency's (EPA's)
enforcement actions led to
sustained compliance under
the National Petroleum
Refinery Initiative (NPRI, or
the initiative). The EPA
selected the petroleum refinery
sector as one of its National
Enforcement Initiatives (NEIs)
in 1996. The EPA intended that
its NPRI strategy's
companywide consent decrees,
or legally binding agreements,
would lead to improved
compliance and reduced
harmful air pollutants or
emissions as companies
changed environmental
management practices and
reduced their emissions. The
EPA officially concluded the
NPRI in 2007, when 80 percent
of the refining facilities were
under a consent decree.
This report addresses the
following EPA theme:
• Addressing climate change
and improving air quality.
For further information,
contact our public affairs office
at (202) 566-2391.
The full report is at:
www.epa.aov/oia/reports/2014/
20140415-14-P-0184.pdf
EPA Needs to Demonstrate Whether It Has Achieved
the Goals It Set Under the National Petroleum
Refinery Initiative
What We Found
By determining the
outcomes of the NPRI,
the EPA can strengthen
the likelihood of
success for future
initiatives and sustain
the desired benefits.
Under the NPRI, the EPA planned to increase
compliance and reduce emissions within the
petroleum refinery industry. However, the EPA
did not determine whether the NPRI achieved the
compliance goal it set. In 2006, the EPA assessed
whether companies under consent decree were
making progress toward the established emission-
reduction goal it set. However, since that time, the EPA has not analyzed the
available facility data to determine whether the initiative achieved the established
emissions-reduction goal. Work on the NPRI has declined since 2007, as the
EPA has reduced resources dedicated to the initiative. The EPA did not place the
same attention on monitoring initiative outcomes as it did on negotiating consent
decrees.
The EPA has replicated this enforcement model in other NEI sectors, such as the
stormwater initiative. The EPA needs to know whether this enforcement
approach produced the intended outcomes. By making this determination, the
EPA can strengthen the likelihood of success for future initiatives, and achieve
and sustain the desired reductions in risk to human health and the environment.
Recommendations and Planned Corrective Actions
We recommend that the EPA's Assistant Administrator for Enforcement and
Compliance Assurance develop and implement a plan to assess whether the
NPRI led to sustained improvement in compliance and sustained reductions in
pollution among refineries. We also recommend that the EPA report the results of
its efforts to the public.
The EPA agreed with our recommendations. The EPA responded that the
agency planned to post company-reported emission data to the public website as
consent decrees are completed. The EPA agreed to guide future NEIs to include
periodic evaluation. We agreed with the EPA's proposed corrective actions.
Three recommendations are resolved with corrective actions underway and one
recommendation is closed with corrective actions completed.
Noteworthy Achievements
The NPRI achieved broad industry coverage by addressing compliance problems
on a companywide basis as opposed to a facility-by-facility approach. The
companywide strategy used a proactive approach to solving compliance
problems by focusing on technology-based solutions to prevent noncompliance.
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
THE INSPECTOR GENERAL
April 15, 2014
MEMORANDUM
SUBJECT: EPA Needs to Demonstrate Whether It Has Achieved the Goals It Set Under the National
Petroleum Refinery Initiative
Report No. 14-P-0184
This is our report on the subject evaluation conducted by the Office of Inspector General (OIG) of the
U.S. Environmental Protection Agency (EPA). This report contains findings that describe the problems
the OIG has identified and corrective actions the OIG recommends. This report represents the opinion of
the OIG and does not necessarily represent the final EPA position. Final determinations on matters in
this report will be made by EPA managers in accordance with established audit resolution procedures.
Action Required
All recommendations are agreed to and resolved. Therefore, no final response to this report is needed.
If you wish to provide a final response to this report, it should be provided as an Adobe PDF file that
complies with the accessibility requirements of Section 508 of the Rehabilitation Act of 1973, as
amended. The final response should not contain data that you do not want to be released to the public;
if your response contains such data, you should identify the data for redaction or removal along with
corresponding justification. We will post this report to our website at http://www.epa.gov/oig.
If you or your staff have any questions regarding this report, please contact the Assistant Inspector
General for Program Evaluation, Carolyn Copper, at (202) 566-0829 or copper.carolyn@epa.gov;
or the Director for Water Evaluations, Dan Engelberg, at (202) 566-0830 or engelberg.dan@epa.gov.
FROM: Arthur A. Elkins Jr.
TO:
Cynthia Giles, Assistant Administrator
Office of Enforcement and Compliance Assurance
-------
EPA Needs to Demonstrate Whether It Has Achieved
the Goals It Set Under the National Petroleum Refinery Initiative
14-P-0184
Table of Contents
Chapters
1 Introduction 1
Purpose 1
Background 1
Noteworthy Achievements 5
Scope and Methodology 7
2 EPA Needs to Determine Whether the NPRI
Led to Sustained Compliance and Pollution Reduction 9
EPA Has Not Measured Progress Toward Achieving
the NPRI Compliance Goal, But Has Conducted
Periodic Follow-Up Inspections 9
EPA Made a One-Time Assessment of Facilities Emissions
Data, But Has Not Assessed Progress Toward Meeting
the NPRI Emission Goal Since 2006 10
EPA Has Not Established a Clear Process for Verifying
Compliance for Consent Decree Termination 11
EPA Reduced Its Resource Commitment to the NPRI,
Which Limited Its Monitoring Ability 12
Conclusion 13
Recommendations 14
Agency Comments and OIG Evaluation 14
Status of Recommendations and Potential Monetary Benefits 16
Appendices
A Timeline for the NPRI 17
B Agency Comments on Draft Report 18
C OIG Response to the Agency's Revised Corrective Action Plan 38
D Distribution 41
-------
Chapter 1
Introduction
Purpose
The objective of this evaluation was to determine the extent
to which the U.S. Environmental Protection Agency's
(EPA's) enforcement actions under the National Petroleum
Refinery Initiative (NPRI, or the initiative) led to sustained
compliance.
Background
Sustained compliance with environmental laws protects human health and the
environment. The principal goal of the EPA's enforcement activities is to bring
about a level of compliance that achieves the human health and environmental
benefits expected from environmental laws. The EPA's 2013 budget directed
about one-tenth of its annual budget to promoting compliance with environmental
laws ($830 million in fiscal year 2013).
The EPA strives to achieve sustained compliance in its programs by focusing on
three types of enforcement activities:
1. Compliance assistance—Activities designed to assist the regulated
community with understanding and complying with regulations.
2. Compliance monitoring—Activities designed to assess the compliance
status of the regulated community through inspections and other activities.
3. Enforcement actions—Legal actions taken by the EPA designed to bring
polluters back into compliance with the law.
The EPA believes these three components, together with enhanced coordination
of EPA and state actions, will lead to sustained improvement in compliance with
federal environmental laws.
EPA Focuses National Resources on Important Compliance and
Pollution Problems
The EPA operates a national enforcement program consisting of two major
components: its "core" program and its National Enforcement Initiatives (NEIs).1
The EPA's core enforcement program forms the bulk of the enforcement activity
1 Prior to February 2010, the EPA referred to national enforcement initiatives as "National Enforcement Priorities."
This is one of
the EPA OIG's
products
associated with
climate change.
14-P-0184
1
-------
for the Office of Enforcement and Compliance Assurance (OECA) and the EPA's
10 regions. Core enforcement responsibilities include the EPA's enforcement of
major environmental statutes like the Clean Air Act (CAA), Clean Water Act,
Resource Conservation and Recovery Act, and other environmental statutes.
Because the EPA's enforcement responsibilities cover millions of regulated sites
and many statutes, the OECA and the agency's regions use the NEIs to focus on
the most serious air, water, waste and chemical hazards. Through the NEIs, the
OECA directs resources to nationally significant enforcement issues, like
stormwater runoff, mineral processing, air toxics and petroleum refineries. The
OECA forms national teams to direct work and monitor progress toward
achieving the goals in individual industry sectors selected as initiatives. The EPA
regions also expend about one-third of their enforcement resources to implement
and monitor the progress of the initiatives (as evidenced by the full-time
equivalent positions assigned to these areas). Once the goals for individual sectors
are achieved, the EPA transitions the work back into the core enforcement
program.
Government Programs Should Be Monitored and Evaluated
The Government Performance and Results Act (GPRA) Modernization Act of
2010 requires performance assessments of government programs so Congress can
review agency performance. To guide agencies in assessing their performance,
GPRA requires agencies to set goals, measure performance against those goals,
and report publicly on their progress toward achieving those goals.
The OECA includes compliance monitoring as one of its essential tools for
achieving improved compliance. In addition, the EPA's congressional budget
justification for fiscal year (FY) 2013 stated that providing information about the
NEIs to the public is a key challenge for the agency. In the budget justification,
the EPA wrote that the agency needs to provide meaningful information to the
public about the progress being made under the NEIs to address priority
environmental risks and compliance problems.
In 2007, the OECA developed a strategic approach2 for achieving improved
compliance and better protection of human health and the environment. The guide
reflects GPRA requirements by instructing that initiatives include measures and
evaluation strategies to determine how well the measures and strategies help to
achieve the EPA's goals. The OECA uses this approach when developing
strategic plans for the individual national enforcement initiatives.
2 EPA 305-R-07-001, Guide for Addressing Environmental Problems: Using an Integrated Strategic Approach,
March 2007. This guide builds on a previously developed framework included as the guide's attachment A,
"Framework for a Problem-Based Approach to Integrated Strategies."
14-P-0184
2
-------
Refineries Present Widespread Compliance Challenges and
Emit Harmful Pollutants
Petroleum refineries account for significant releases of pollutants into the
environment during the complicated industrial process that refines crude oil into
petroleum products. Refineries emit 75 percent of their pollutants into the air.
These pollutants contribute to smog, acid rain, climate change, and
bioaccumulation in mammals and fish eaten by humans. These pollutants also
contribute to cardiovascular and respiratory disease, and cancer in humans.
When the EPA was developing the initiative in 1996, the Toxics Release
Inventory showed refineries released over 66 million pounds of toxic pollutants
known or suspected to cause cancer or other serious human health effects. In the
EPA's most recent Toxics Release Inventory report from 2011 (released in
January 2013), refineries reported releasing over 56 million pounds of toxic
pollutants. Approximately 58 percent of these releases were to air through fugitive
air emissions or through stack air emissions. On average, the population within
3 miles of active petroleum refineries is approximately 40 percent minority, which
is one of several considerations when determining environmental justice concerns.
Between 1994 and 1995, the EPA conducted nationwide inspections of
109 petroleum refineries. The inspection results identified widespread CAA
compliance challenges, with violations in 70 percent of refineries. Through
inspections and additional research, the EPA identified four major areas where
refineries did not comply with the law:
1. Emissions from major refining units that were incorrectly permitted as
"minor" sources, were unpermitted, or did not have Best Available
Control Technology installed (New Source Review/Prevention of
Significant Deterioration requirements for fluidized catalytic cracking
units, heaters and boilers).
2. Fugitive emissions associated with leaks from refinery equipment,
e.g., valves, pumps and connectors (New Source Performance Standards
for leak detection and repair).
3. Uncontrolled and unreported benzene waste (National Emission Standards
for Hazardous Air Pollutants).
4. Use of flaring for routine purposes instead of on an emergency basis, and
indications that some emissions during emergency events were in excess
of applicable limits (New Source Performance Standards Subparts A & J).
14-P-0184
3
-------
NPRI Was Established to Address Industrywide Compliance
Problems and to Reduce Pollution
The EPA developed plans and devoted resources to improve compliance across
the refining sector in the four major areas identified by the agency. To achieve a
sustainable change in compliance within the petroleum refinery industry, the EPA
developed a strategy that would address the industry as a whole. In 1996, the EPA
selected the petroleum refinery sector as one of the agency's first national
enforcement initiatives (at the time called National Enforcement Priorities).
The EPA's overall goal for the NPRI was to bring U.S. refineries into long-term
sustained compliance with the CAA. The EPA believed the actions required by
consent decrees would lead to improved compliance and reduced emissions.
Because a single company may operate many individual refineries, the EPA
negotiated with companies instead of facilities under the NPRI. This enabled the
agency to cover all of a company's facilities in one negotiation. The negotiations
resulted in legally enforceable consent decrees, which established companywide
and facility-specific compliance expectations and reporting requirements. For
example, the companies agreed to change environmental management practices,
reduce their emissions, and provide the EPA with regular certified progress
reports for all company facilities. The certified progress reports provided the EPA
with information about refinery performance, emissions and progress toward
completing consent decree requirements. The consent decrees also included
requirements that a company's refineries must demonstrate they have paid all
stipulated penalties and achieved compliance with the established emissions limits
for 12 consecutive months. Once companies complete all consent decree
requirements, they may request termination of the decree by the federal court.
The EPA's I'etroleum Refining Performance-Based Strategy identified four goals
and measures for the NPRI (see table 1). In 2007, after the EPA achieved its goal
to cover 80 percent of the industry capacity with consent decrees (see Goal 1
listed in table 1), the agency returned enforcement of the petroleum refinery
sector back to the EPA's core enforcement program. Appendix A shows a
timeline of the petroleum refinery sector's inclusion as an initiative.
14-P-0184
4
-------
Table 1: Goals, baseline data and milestones for the NPRI
Goal No.
Goal text
Baseline data
Goal
1
80% of the domestic refining
capacity addressed through
settlement, by filed civil action
against a refinery, or referred to
the U.S. Department of Justice
for filing.
Total domestic capacity of
the refinery sector (2004):
142 refineries produced
17.4 billion barrels per day.
13.9 billion barrels per
day.
2
50% improvement in compliance
over the 1995 baseline.
Baseline noncompliance
rate (FYs 1994-1995): of
156 facilities inspected,
109, or 70%, of the facilities
were not in compliance.
55 of the 109 facilities
not in compliance
brought back into
compliance.
3
20% reduction in emissions of
S02 and NOx from the 1995
baseline
Emissions data (FY 1995):
NOx (380,641 short
tons/year); S02 (648,155
short tons/year).
205,759 short tons of
NOxand S02
combined per year.
4
100% of consent decree
deliverables to the EPA requiring
a response, with 75% responded
to by the agency within 90 days.
Receipt date of all
deliverables requiring a
response.
(Same as goal text).
Source: OIG summary of the OECA's Petroleum Refining Performance-Based Strategy.
NOx: Nitrogen oxides
SO2: Sulfur dioxide
The NPRI helped the EPA coordinate enforcement resources nationwide. The
EPA marshaled resources from the OECA, the EPA regions and the National
Enforcement Investigation Center (NEIC)3 to conduct research and to negotiate
with companies. The EPA established expectations for regional participation,
increased resource availability for the NPRI, and promoted the NPRI's
cooperative approach throughout the petroleum refinery sector. This approach
enabled the EPA to use a regional expert where the person would be the most
useful. For example, Region 5 air enforcement personnel traveled to Region 6 to
aid in negotiations.
Noteworthy Achievements
The NPRI achieved broad industry coverage by addressing compliance problems
on a companywide basis, as opposed to the traditional facility-by-facility
approach. By the end of January 2011, the EPA established consent decrees
covering 28 refining companies (105 refineries) that accounted for 90 percent of
3 The NEIC is the environmental forensic center for the EPA's enforcement programs. The NEIC is accredited for
field sampling, field measurements and monitoring, and laboratory measurement. The NEIC lias a unique role in
conducting complex criminal and civil enforcement investigations, and applied research and development to support
science for enforcement. The demands of specific enviromnental enforcement cases require nonstandard
methodologies. In response to these needs, the NEIC conducts and develops new methodologies and innovative
investigative strategies.
14-P-0184
5
-------
the national industrial capacity. Currently, only 30 operating refineries are not
parties in an NPRI consent decree.4
As an enforcement tool, the NPRI consent decrees furthered the objectives of the
CAA by requiring the refinery industry to develop pollutant controls more
stringent than the existing CAA requirements. In Chapter 2, we discuss the
importance of monitoring and evaluating the outcomes of the consent decrees
after they have been signed. By committing to install the technology controls
required by the NPRI consent decrees, participating companies made ambitious
commitments to reduce annual refinery emissions. For example, the EPA
estimated that companies under consent decree would cumulatively reduce
pollution by 93,000 tons of NOx and 256,000 tons of S02 annually.5 If achieved,
this would represent a 24 percent reduction in NOx and a 39 percent reduction in
S02 emissions annually from levels in FY 1995. The estimated annual reduction
in NOx is equivalent to the amount emitted by approximately 89 million cars
driving about 13,500 miles per year on average.6
The NPRI consent decrees employ a proactive approach to solving compliance
problems by focusing on technology-based solutions to prevent noncompliance,
reduce facility reporting burden, and enhance the EPA's ability to manage
compliance over this complex industry. The EPA's enforcement efforts under this
NEI drove improvements to existing control technologies and developments of
new technologies in the petroleum refineries sector. For example:
• Consent decrees require companies to install continuous emissions
monitoring systems (CEMS) on major emission sources. Once facilities
install CEMS, they can monitor actual data internally and report the data
to the EPA. The EPA can use the CEMS data to establish equipment-
specific and facility-specific emission limits. Facilities can also use CEMS
data to ensure they are in compliance with the terms of the consent decree.
• The more recent consent decrees also require fence-line monitoring.
These monitors provide near real-time reports of emission data on public
websites, thereby potentially enhancing transparency. The agency further
believes providing actual emission data to communities living close to
refining facilities will serve as a deterrent to serious noncompliance.
• Since 2010, consent decrees have required companies to install low-leak
valves when replacing existing valves at refineries. These valves virtually
4 The EPA posts the consent decrees on its website at http://www2.epa.gov/enforcement/petroleum-refinerv-
national-case-results.
5 In the United States, the terms ton and short ton (used in the EPA's goals for the NPRI) are interchangeable and are
equivalent to 2,000 pounds.
6 This calculation uses the EPA's 1999 Tier 2 tailpipe emissions standard of 0.07 grams of NOx pollutants emitted
per car, per mile driven.
14-P-0184
6
-------
eliminate pollutant leaks and reduce the need for the EPA to re-inspect the
facility for compliance with leak-detection protocols.
These technology-based compliance and monitoring tools enable facilities to
provide easily accessible emissions data, and increase the human health protection
for communities close to refining facilities.
Scope and Methodology
We conducted this evaluation in accordance with generally accepted government
auditing standards. Those standards require that we plan and perform our work to
obtain sufficient, appropriate evidence to provide a reasonable basis for our
findings and conclusions based on our audit objectives. We believe that the
evidence obtained provides a reasonable basis for our findings and conclusions
based on our evaluation objectives. We conducted our evaluation from
January 2012 through September 2013 (this assignment was suspended between
March 2012 and July 2012 to address other OIG priorities).
We selected the NPRI because it was an early national enforcement initiative, and
the EPA had moved the refinery program back to the core enforcement program.
As such, evaluating this NPRI would allow us to evaluate results from one of the
initiatives. OECA staff and managers agreed that evaluating the NPRI offered an
opportunity to evaluate the outcomes of an initiative.
To answer our objective, we reviewed documents, surveyed the EPA regions,
conducted interviews, and analyzed evidence of compliance from the EPA's
enforcement data and results. Specifically, we reviewed background documents
related to the EPA's development and management of the NPRI. We requested
information about refinery compliance with consent decrees from the EPA's nine
regions where refineries are located (Regions 2 through 10). We interviewed EPA
enforcement officials in headquarters and eight of the nine EPA regions with
refineries.7 We also interviewed experts at the NEIC who conduct many of the
inspections at refinery facilities. We did not speak with state officials for this
review because states generally did not participate in consent decree follow-up
inspections.
Prior Audit Coverage
EPA OIG Report No. 08-P-0278, EPA Has Initiated Strategic Planning for
Priority Enforcement Areas, but Key Elements Still Needed,
September 25, 2008: The OIG reviewed the EPA's strategies for the air toxics,
combined sewer overflow and mineral processing industry initiatives within the
NEI. The OIG reported that the OECA had instituted a strategic planning process
for these industry initiatives. Each strategy contained an overall goal, a problem
7 We did not interview enforcement officials in Region 7 because the consent decrees in this region were not signed
before 2007. Region 7 consent decrees were signed in 2009 and 2012.
14-P-0184
7
-------
statement, a description of the status of the priority area, and anticipated
environmental benefits. The EPA also described the facilities the agency would
address, the tools the EPA would use, and the OECA headquarters and regional
responsibilities. However, each of the strategic plans lacked key elements to
monitor progress and accomplishments. In some cases, the plans lacked detailed
exit plans (i.e., plans to move the industry initiatives back to the core enforcement
program). The OIG recommended that the OECA develop a policy requiring
strategy documents for industry initiatives and have these documents include a
full range of performance measures, exit plans and information about the role of
states. The OECA reported that it addressed this recommendation by developing a
template for performance-based strategies for future enforcement initiatives. The
OECA certified that it completed the other recommendation in December 2009.
EPA OIG Report No. 2004-P-00021, EPA Needs to Improve Tracking of
National Petroleum Refinery Compliance Program Progress and Impacts,
June 22, 2004: The OIG reported that the OECA's performance measurement
and reporting approach for the NPRI did not provide useful and reliable
information. The EPA needed this information to effectively implement, manage,
evaluate and continuously improve program results. In addition, the OECA had
not established and communicated clear goals, systematically monitored refinery
program progress, reported actual outcomes, or tracked progress toward achieving
consent decree goals. This report included 17 recommendations for improvement.
Among them, the OIG recommended that the OECA improve refinery consent
decree implementation and tracking, as well as ensure better measurement and
reporting of refinery program outcomes. On March 19, 2009, the OECA certified
that it had completed all corrective actions for the report. The OECA reported that
it fulfilled six recommendations through its 2004 National Program Managers
Guidance, which the OECA said included the goals developed as part of the
Petroleum Refining Performance Based Strategy (table 1).
14-P-0184
8
-------
Chapter 2
EPA Needs to Determine Whether the NPRI Led to
Sustained Compliance and Pollution Reduction
Though evidence of improvements exists, the EPA has not demonstrated whether
the NPRI resulted in sustained compliance or continued emission reductions in
the petroleum-refining sector. Consent decrees represented the first step toward
improving compliance and reducing harmful emissions in the refinery sector, and
the EPA set compliance improvement and pollution reduction goals for the NPRI.
However, the EPA has not measured progress toward achieving the compliance
goal it set for the NPRI, and, since 2006, the EPA has not evaluated progress
toward achieving the emissions goal. When the EPA returned the NPRI to the
agency's core enforcement program in 2007, the EPA reduced the amount of
resources for the initiative, which hampered its ability to oversee consent decree
implementation. The EPA needs to determine whether the NPRI achieved its
desired goals of addressing challenging compliance and emissions problems.
By making this determination, the EPA can strengthen the likelihood of success
for future national enforcement initiatives, and achieve and sustain the desired
reductions in risk to human health and the environment.
EPA Has Not Measured Progress Toward Achieving the NPRI
Compliance Goal, But Has Conducted Periodic Follow-Up Inspections
The EPA began the NPRI in 1996 after identifying a sectorwide CAA
noncompliance rate of 70 percent in 1995. The EPA set a compliance goal for the
NPRI that companies subject to NPRI consent decrees would achieve a 50 percent
improvement in compliance over the 1995 baseline. The 1995 baseline was based
on inspections of 156 refineries from August 1994 through August 1995.
Therefore, to determine whether sector compliance improved from the baseline as
a result of the NPRI, the EPA would need to conduct a round of inspections
similar to those conducted in 1994 and 1995. However, the EPA has not deployed
the resources to develop and implement such a protocol.
Instead, the EPA conducted periodic consent decree follow-up inspections.
Although the follow-up inspections did not allow the agency to determine
whether the NPRI led to sustained compliance improvements industrywide, they
provided the EPA with important information about compliance with consent
decree requirements at particular facilities. However, due to differences in the
number of refineries in each region and in resource availability to conduct
inspections, the EPA did not conduct follow-up inspections at facilities
consistently. Because the EPA regions prioritized other activities over NPRI
monitoring and evaluation, more than half of the facilities under consent decree
had not received follow-up inspections. In fact, between 2003 and 2012, Region 6
and the OECA conducted consent decree follow-up inspections at only 19 of the
14-P-0184
9
-------
35 consent decree facilities in Region 6. This is notable because Region 6
contains more refineries—and more refineries under consent decree—than any
other region.
Moreover, some of the nation's largest refineries went without any consent decree
follow-up inspection by the EPA. Although other types of federal or state
inspections8 may have occurred, data provided by the EPA's regional
enforcement personnel show that neither the OECA nor Region 6 conducted
consent decree follow-up inspections at the following facilities:
• Marathon Garyville, Louisiana, Refinery (490,000 barrels per day).
Consent decree signed 2001.
• ExxonMobil Beaumont, Texas, Refinery (344,500 barrels per day).
Consent decree signed 2005.
EPA officials said that although Region 6 has many more refineries, the region is
still using its staff to negotiate consent decrees with new companies. Generally,
EPA officials said that they prioritized consent decree follow-up inspections
based on the length of time a refinery has been subject to a consent decree,
whether the refinery has received other types of inspections, the results of any
other inspections, and input from regions on where to inspect. They also
conducted some random follow-up inspections.
EPA Made a One-Time Assessment of Facility Emissions Data,
But Has Not Assessed Progress Toward Meeting the NPRI Emission
Goal Since 2006
When signing consent decrees, the EPA estimated the emission reductions that
would occur when companies completed all consent decree actions. On the EPA's
website, the EPA reported that companies that were parties to consent decrees
would cumulatively reduce pollution by 205,759 short tons of NOx and SO2
combined per year. So that the EPA could track emissions, the consent decrees
required facilities to submit certified progress reports that included emission data.
In 2011, the EPA published a progress report on the NPRI that included a
one-time analysis of the emissions data self-reported by companies under consent
decree as of 2006.9 However, since that time, the EPA has not monitored the
progress that companies have made toward achieving the emission goals set in the
agency's NPRI strategy.
8 For example, under individual state Compliance Monitoring Strategies, major sources like these facilities are
required to be inspected on a periodic basis (between 2 to 5 years); such inspections include compliance with
consent decree requirements as reflected in CAA Title V permits, among other applicable permits.
9 The OIG did not assess the accuracy of self-reports as part of this evaluation.
14-P-0184
10
-------
The analysis of the 2006 emissions data showed that overall emissions from these
companies declined by almost 50 percent.10 However, facility-specific analysis
showed mixed results. Data from some facilities demonstrated that some would
likely meet or exceed anticipated emission-reduction goals. Emissions reported by
other facilities had increased since signing the consent decrees.
The 2006 assessment also did not show the progress made by all facilities under
the consent decree at the time. Many additional facilities signed consent decrees
with the EPA after the 2006 analysis. OECA officials also said that the 2006
analysis likely did not include some of the estimated emission reductions due to
the lag between finalizing more recent decrees and the installation of controls that
would achieve the emission reductions. For example, to reduce flaring emissions,
the implementation schedule for the 2005 ConocoPhillips settlement required
major equipment to be installed periodically through 2012 at some facilities,
which is after the 2006 emissions analysis was completed. The EPA said this
meant projected emissions reductions for the ConocoPhillips refinery may not
occur until 2015.
Enforcement officials said that resource constraints prevented them from
repeating the effort to estimate progress toward achieving emissions reductions
after the 2006 analysis. For example, due to a lack of funding, OECA officials
indicated they could not continue tracking pollution reductions in a manner
similar to their 2006 analysis based on emissions data. The OECA officials said
they do not have the budget to support the $36,000 that it would cost to update the
analysis of 2006 emission data.
EPA Has Not Established a Clear Process for Verifying Compliance
for Consent Decree Termination
EPA has not established a clear process for verifying compliance in advance of
consent decree termination. The court makes the final determination about
consent decree termination, and consent decrees require companies to certify
completion of all consent decree requirements under penalty of perjury or
contempt. However, the EPA verifies whether the company has, in fact,
completed all requirements.
An OECA official was not confident that the regions could verify compliance
with all consent decree requirements. OECA officials in charge of the NPRI
believed the process was clear, while enforcement personnel in five of the seven
regions we spoke with said they were not sure how the process would occur, or
that they would like additional guidance from OECA on termination.
10 In response to a 2004 OIG report recommendation, the OECA hired a contractor to verify predicted emission
reductions using facility progress self-reports. The OECA presents the results of the contractor's analysis in EPA
Enforcement: National Petroleum Refinery Initiative, February 11, 2011, draft (subject to revision). As of the date of
this report, this analysis was available at http://www2.epa.gov/enforcement/results-petroleum-refmerv-national-
initiative.
14-P-0184
11
-------
EPA Reduced Its Resource Commitment to the NPRI, Which Limited
Its Monitoring Ability
The OECA and EPA regional enforcement officials identified reduced resources
as a cause for not monitoring emission and compliance goals for the refinery
initiative. The EPA reduced the resources devoted to the NPRI even though the
work required to ensure the initiative resulted in the intended outcomes continued
and expanded. When the EPA concluded the NPRI in 2007, no consent decrees
had been terminated. In 2007, the EPA's 20 companywide consent decrees
covered 92 refineries. Moreover, after the NPRI ended, the EPA continued to
negotiate and sign consent decrees with additional refineries, eventually reaching
32 decrees covering 116 facilities by the end of 2012 (see appendix A). Each of
the 32 consent decrees included hundreds of activities, reports and results to
monitor and evaluate, and most decrees were designed to last for 10 or more
years.
Despite the continued and increasing oversight requirement, the OECA concluded
the NPRI in 2007 and shifted resources to work on other priorities. Enforcement
officials in the OECA and the regions told us the EPA focused fewer resources
and attention on overseeing the NPRI than the agency did on achieving industry
coverage. The OECA officials said that spending additional time on this sector
would be less productive than devoting those resources to a new priority area.
Regional personnel said that adequately monitoring implementation required a
resource commitment at least as large as the resources used to negotiate the
consent decrees. However, the regions shifted resources to work on other
priorities. Region officials told us and OECA officials agreed that the absence of
an Annual Commitment System (ACS)11 goal for the regions is an impediment to
regions investing significant resources to oversee refinery compliance. The
absence of an ACS goal concerning NPRI follow-up means that regions are more
likely to focus on other work associated with ACS goals than on providing
oversight of refineries.
The low number of EPA staff with technical expertise in the refinery sector has
also hampered oversight. Regulating the petroleum refinery process is complex
and requires technical knowledge of not only the refinery process but also the
complicated regulatory requirements under the CAA. The OECA does not have a
succession plan in place for refinery staff (including inspectors) leaving the
agency. Headquarters and region officials associated with the NPRI were
concerned that the EPA might not attract and retain qualified experts for the
EPA's refinery program. In 2012, the OECA had one inspector responsible for
ensuring the refineries comply with the New Source Review/Prevention of
Significant Deterioration consent decree components. In addition, only a few
11 The ACS is a performance module located within the EPA's Budget Automation System used to track annual
headquarters and regional performance commitment information and results.
14-P-0184
12
-------
engineers are qualified to conduct leak detection and repair and benzene
inspections. Regions also had very limited resources. For example, Region 9 had
one staff person responsible for overseeing consent decree compliance for
15 refineries and negotiating with companies not yet under a consent decree.
Officials from the OECA indicated that they would like to evaluate the impact of
the NPRI on compliance and emissions. However, they told us they cannot do this
due to resource constraints. With the EPA's current budgetary constraints, the
agency must make difficult decisions regarding resource allocation to maximize
the environmental and human health results achieved from its work. However,
without devoting resources to determining the outcomes of its programs, the EPA
cannot demonstrate whether the resources it commits have, in fact, led to the
intended and predicted environmental and human health outcomes.
Conclusion
Since beginning the NPRI, the EPA has adopted a similar consent decree approach
in other industry sectors selected as initiatives. For example, the EPA based its
stormwater and mineral processing initiatives on establishing consent decrees with
companies. Consent decrees in these sectors also target major compliance issues
and establish improved environmental management practices. Establishing a
demonstrated performance-based system within the NPRI would therefore provide
a model for how to accomplish this in other sectors.
To ensure it is operating an effective, results-based program, the EPA should be
assessing the success and outcomes of the NPRI before the multi-year consent
decrees have come to a close. Doing so enables the EPA and the regulated
industry to demonstrate success where it has occurred, identify lessons learned,
modify the approach to improve success rates, and enhance future enforcement
initiatives as needed.
The EPA's progress in signing consent decrees with companies representing over
90 percent of the national refinery capacity is an important achievement. The
EPA's 2006 evaluation of company-reported emission reduction results assessed
whether many facilities were making improvements. To sustain environmental
results and achievements and provide lessons for future enforcement initiatives,
the EPA needs to provide resources to regularly monitor and evaluate the
outcomes of the consent decrees.
By not taking action to assess the outcomes of the NPRI, the EPA misses
opportunities to design improved strategies that can be applied in other contexts,
and the agency risks falling short in being able to demonstrate that the NPRI
achieved its intended goals.
14-P-0184
13
-------
Recommendations
We recommend that the Assistant Administrator for Enforcement and Compliance
Assurance:
1. Develop and implement a plan to determine whether consent decrees
signed as part of the NPRI are leading to promised improvements in
compliance and sustained reductions in pollution. The plan should
establish procedures for verifying company completion of consent decree
requirements for termination, including company assessments of facility
pollution reductions and comparisons with estimated reductions.
2. Incorporate requirements that ease the resource burden on the EPA to
monitor refinery progress (e.g., CEMS and fence-line monitoring) into
future consent decrees or amendments to existing consent decrees under
the NPRI.
3. Inform the public about the extent to which the NPRI resulted in sustained
compliance improvement at facilities and reductions in emissions agreed
to in consent decrees, as highlighted as a key challenge for OECA in the
EPA's congressional budget justification.
4. Ensure that plans for future NEIs include an evaluation component that
demonstrates the extent to which the NEI strategy achieves the goal(s) for
the NEI identified by the EPA.
Agency Comments and OIG Evaluation
We received comments on the draft report from the Assistant Administrator for
Enforcement and Compliance Assurance (appendix B). The agency largely agreed
with our recommendations and, in many cases, said actions it has already
underway achieve the purpose of the recommendations. In its response, the
OECA proposed additional corrective actions and estimated completion dates.
After subsequent communication with the OECA, we believe the proposed
actions are responsive (appendix C); three recommendations are resolved with
corrective actions underway and one recommendation is considered complete.
In response to recommendation 1, the OECA will issue a memorandum to the
regions reminding them of the requirements for termination and to work with the
Office of Civil Enforcement for confirmation that all termination requirements
have been met. The OECA said it is working with refiners seeking termination to
develop graphs of actual annual emission reductions achieved under the consent
decree at the time of termination. Additionally, the agency said it has long
planned to pull together and disseminate consent decree-specific emission
reduction data once a consent decree is terminated, thereby supplementing the
2006 Progress Report (recommendation 3).
14-P-0184
14
-------
The agency agreed with recommendation 2, saying it has begun implementing
lessons learned from the NPRI to address the resource burden from consent
decree implementation and oversight, including the January 2013 guidance issued
to the EPA's enforcement personnel. In its response, the OECA mentioned its
"Next Generation Compliance" paradigm, which takes advantage of new
information and monitoring technologies to enable the EPA, states and tribes to
get better compliance results.12 OECA also said it has a workgroup currently
exploring ways to use advanced monitoring, e-reporting, public transparency,
third-party verification and other tools in enforcement settlements. The agency
completed corrective actions and the recommendation is closed.
Finally, the agency agreed that it is important to evaluate the success of its NEIs
periodically (recommendation 4). The agency said that the level of information
about sources, as well as meaningful information about pollution loading and the
effectiveness of preventative measures, will vary by NEI sector. In the subsequent
communication, the agency proposed minor revisions to the recommendation
language (appendix C), which we accepted and incorporated into the final report.
The agency's response also included some technical comments, which were
incorporated into the final report as appropriate.
12 For more information on Next Generation Compliance, see OECA's FY 2014 National Program Managers
Guidance, at http://www2.epa.gov/planandbudget/national-program-manager-guidances.
14-P-0184
15
-------
Status of Recommendations and
Potential Monetary Benefits
RECOMMENDATIONS
Rec. Page
No. No.
Subject
Status1
Action Official
POTENTIAL
MONETARY BENEFITS
(in $000s)
Planned
Completion Claimed Agreed-To
Date Amount Amount
14 Develop and implement a plan to determine
whether consent decrees signed as part of the
NPRI are leading to promised improvements in
compliance and sustained reductions in
pollution. The plan should establish procedures
for verifying company completion of consent
decree requirements for termination, including
company assessments of facility pollution
reductions and comparisons with estimated
reductions.
Assistant Administrator
for Enforcement and
Compliance Assurance
5/30/14
14 Incorporate requirements that ease the resource
burden on the EPA to monitor refinery progress
(e.g., CEMS and fence-line monitoring) into
future consent decrees or amendments to
existing consent decrees under the NPRI.
14 Inform the public about the extent to which the
NPRI resulted in sustained compliance
improvement at facilities and reductions in
emissions agreed to in consent decrees, as
highlighted as a key challenge for OECA in the
EPA's congressional budget justification.
14 Ensure that plans for future NEIs include
an evaluation component that
demonstrates the extent to which the NEI
strategy achieves the goal(s) for the NEI
identified by the EPA.
Assistant Administrator 11/19/13
for Enforcement and
Compliance Assurance
Assistant Administrator 5/30/14
for Enforcement and
Compliance Assurance
Assistant Administrator 4/30/14
for Enforcement and
Compliance Assurance
0 = Recommendation is open with agreed-to corrective actions pending.
C = Recommendation is closed with all agreed-to actions completed.
U = Recommendation is unresolved with resolution efforts in progress.
14-P-0184 16
-------
Appendix A
Timeline for the NPRI
EPA inspects 109 petroleum refineries to identify compliance trends.
1995
1996
EPA designates the petroleum refinery sector as an NEI.
1 consent decree signed:
Koch (3 refineries)
2000
2001
2002
2003
1 consent decree signed:
CITGO (5 refineries)
2004
2005
2006
2007
2008
2009
2010
2012
2013
and
Beyond
EPA continues to bring other companies under consent decree. EPA also continues
oversight work for monitoring progress towards consent decree requirements.
2 consent decrees signed:
Frontier (2 refineries)
Wyoming(l refinery)
2 consent decrees signed:
Sinclair (3 refineries)
Holly (1 refinery)
2011
2 consent decrees signed:
Hovensa (1 refinery)
Western (1 refinery)
2 consent decrees signed:
Shell (3 refineries)
MurphyOil (2 refineries)
3 consent decrees signed:
Total Petrochemicals (1 refinery)
Hunt Refining (3 refineries)
Valero/Premcor (3 refineries)
EPA returns the petroleum refinery
sector back to the agency's core
enforcement program.
5 consent decrees signed:
Giant(2 refineries)
Sunoco (4 refineries)
Valero (14 refineries)
ExxonMobil (7 refineries)
ConocoPhillips (11 refineries)
5 consent decrees signed:
Lion (1 refinery)
Ergon (2 refineries)
Chevron (5 refineries)
CHS Inc. (Cenex) (1 refinery)
Coastal Eagle Point (1 refinery)
5 consent decrees signed:
BP (8 refineries)
Conoco (4 refineries)
Navajo-Montana (3 refineries)
Motiva-Equilon-Deer Park (9 refineries)
Marathon Ashland Petroleum (7 refineries)
4 consent decrees signed:
BP Whiting (1 refinery)
Hess Corporation (1 refinery)
Coffeyville ResourcesRefining& Marketing(l refinery)
Marathon Petroleum &Catlettsburg Refining (6 refineries)
Source: OIG analysis and summary.
14-P-0184
17
-------
Appendix B
Agency Comments on Draft Report
November 19, 2013
MEMORANDUM
SUBJECT: Response to the Office of Inspector General Draft Report: "Determining the
Outcomes of the National Petroleum Refinery Initiative May Identify
Opportunities to Strengthen Future National Enforcement Initiatives," dated
September 23, 2013, Report No. OPE- 2012-3065
FROM: Cynthia Giles
Assistant Administrator
TO: Carolyn Copper
Assistant Inspector General
Office of Program Evaluation
Thank you for the opportunity to respond to the draft findings and recommendations presented in
the Office of Inspector General (OIG) Draft Report, "Determining the Outcomes of the National
Petroleum Refinery Initiative May Identify Opportunities to Strengthen Future National
Enforcement Initiatives" (Draft Report). The Office of Enforcement and Compliance Assurance
(OECA)'s response starts with a summary, a response to the recommendations, followed by
more detailed comments on the Draft Report. As you will see, we largely agree with the
recommendations, and in many cases we think that the actions already underway achieve the
purpose of the recommendations For those Draft Report recommendations with which OECA
agrees, we propose corrective actions and estimated completion dates (below and in attached
corrective action plan) or, as applicable, that OECA continue to undertake actions consistent
with the recommendation.
Summary
As a threshold matter, we agree with the Draft Report's discussion of the importance of assessing
the effectiveness of the Agency's National Enforcement Initiatives (NEIs). As discussed in more
detail below, we believe that the available data strongly supports that the National Petroleum
Refinery (NPRI) has not only met but exceeded our goals for this work in reducing pollution that
affects communities across the country. We are also mindful in our work to implement the
current NEIs that we are operating in a time of constrained (and declining) budgets, and as a
result we must seek ways to utilize the resources available to the enforcement program in the
most efficient and effective way to monitor progress by those under consent decrees, and to
design approaches that are cost effective. In this respect, while we do not have the funding and
14-P-0184
18
-------
resources available to us to be able to undertake the additional evaluation of refinery emission
reductions in the way suggested in the Draft Report, we believe that the approach we are using
will be just as effective and achievable at less cost.
Response to Recommendations
The Draft Report's recommendations are generally workable, and as described in more detail
below OECA is already undertaking actions that are consistent with several of the
recommendations.
Recommendation #1: "Develop and implement a plan to determine whether consent decrees
signed as part of the NPRI are leading to promised improvements in compliance and sustained
reductions in pollution. The plan should establish procedures for verifying company completion
of consent decree requirements for termination, including company assessments of facility
pollution reductions and comparisons with estimated reductions."
OECA Response: The NPRI consent decrees already contain legally-binding provisions for
termination, and these will be an effective vehicle for verifying that the requirements of the
consent decree are met before the CD is terminated. These provisions include a certification of
completion of CD requirements by a refiner seeking termination (under penalty of peijury or
other sanction by the court), and confirmation that all required emission controls have been
installed, payment of all civil and stipulated penalties by a refiner, completion of and compliance
with all required injunctive relief required (including any Supplemental Environmental Projects),
receipt of all final permits incorporating the emission limits and standards established under the
CD, and operation in compliance with the emission limits established in the CD for at least one
year preceding termination. Because we are monitoring compliance with the consent decrees,
and will also be verifying compliance in connection with termination of the consent decrees, a
separate plan to accomplish these objectives is not necessary.
We agree that effective monitoring and termination are important, and OECA has already begun
working with refiners preparing for termination in a manner similar to this recommendation.
EPA, in conjunction with the U.S. Department of Justice and (as applicable) state co-plaintiffs,
follows the prescribed requirements for consent decree termination when evaluating whether a
defendant has met its legal obligations under the consent decree, which includes installation of
all required controls, receipt of final permits, operation in compliance with the consent decree,
etc. (all as specified in the consent decree). As part of this process, EPA, DOJ and applicable
state co-plaintiffs, at the time of termination, will confirm through consent-decree required
certified completion reports and independent verification that all major obligations have been
satisfied.
While achieving the total emission reductions that were estimated at the time settlement is not a
legally-required component of termination, as discussed further in the detailed comments, the
data indicates that emission benefits of NPRI settlements will be greater than the initial
estimates. In addition, EPA is working with refiners seeking termination to develop graphs of
annual emission reductions actually achieved under the consent decree at the time of termination.
14-P-0184
19
-------
However, if all legally required conditions are satisfied, EPA and DOJ cannot legally condition
the United States' consent to termination on such emissions reductions comparisons or data.
Proposed Correction Action: Within 45 days, OECA will issue a memorandum to the regions to
remind them of the requirements for termination as provided in refinery consent decrees, and in
particular to work with OCE for confirmation that all necessary requirements for termination
have been met. In addition, OECA will continue to work with refiners at the time of termination
of their consent decree to develop graphs of annual emission reductions obtained under their
consent decree at the time of termination.
Recommendation #2: "Incorporate requirements that ease the resource burden on the EPA to
monitor refinery progress (e.g., CEMS and fence-line monitoring) into future consent decrees or
amendments to existing consent decrees under the NPRI."
OECA Response: OECA agrees with this recommendation and has already begun implementing
lessons learned from the NPRI to address the resource burden from consent decree
implementation and oversight. Lessons learned from oversight of both NPRI and non-NPRI
consent decrees have been documented and disseminated in guidance issued to EPA enforcement
personnel. See "Guidance on Streamlining Oversight in Civil Settlements" (OECA Jan. 10,
2013). at http://intranet.epa.gov/oeca/oce/io/documents/fmalstreamliningmemol 1013.pdf. In
addition, CEMS and self-reporting of compliance have always been routinely required for all
major emission sources covered in NPRI consent decrees. Further, several recent settlements
have included provisions for fence-line monitoring. Provisions such as fenceline monitoring do
not necessarily reduce the burden on EPA of monitoring compliance with consent decrees, but
they do help to inform the community and to create pressure for the facility to comply; this is an
important part of our "Next Gen" approaches to drive better performance. In addition, OECA
currently has a headquarters/regional workgroup exploring ways to employ advanced
monitoring, e-reporting, public transparency, third-party verification and other tools in
enforcement settlements. These tools could both enhance compliance with consent decrees and
streamline EPA oversight of consent decrees.
Because OECA is already implementing actions consistent with this recommendation for NPRI
consent decrees and has issued guidance applicable to all consent decrees, no further corrective
action is needed.
Recommendation #3: "Inform the public about the extent to which the NPRI resulted in
sustained compliance improvement at facilities and reductions in emissions agreed to in consent
decrees, as highlighted as a key challenge for OECA in EPA's congressional budget
justification."
OECA Response: OECA agrees with this recommendation and, as noted in the response to
Recommendation #1, OECA has long planned to pull together and disseminate consent decree-
specific emission reduction data once all of an individual consent decree's requirements have
been completed. This would supplement the initial summary of emission reductions achieved by
settling refiners through 2005, and posted to EPA's website. See "EPA Enforcement: National
14-P-0184
20
-------
Petroleum Refinery Initiative" at http://www2.epa.gov/enforcement/results-petroleum-refinery-
national-initiative.
Proposed Corrective Action: As provided in the proposed corrective action for Recommendation
#1, OECA will continue to work with refiners at the time of termination of their consent decree
to develop graphs of annual emission reductions obtained under a consent decree at the time of
termination.
Recommendation #4: "Ensure that plans for future industry initiatives of the NEI [National
Enforcement Initiative] include a monitoring and evaluation component that demonstrates the
extent to which the NEI strategy leads to improvement in compliance and human-health risks
identified by the EPA."
OECA Response: OECA agrees that it is important to evaluate the success of our NEIs based on
the goals set for the NEI from the outset, and, as noted below, we are already making substantial
efforts in doing this. We agree that we need to both learn from our experiences in implementing
the NEIs, and be transparent about what we have achieved through the NEIs.
We are concerned, however, with this recommendation as written, and propose that it be
modified. First, it would be better to clarify the recommendation to focus on assuring that, for
future NEIs, OECA assesses whether our NEI effort has achieved the goal(s) stated in the NEI
strategy. We note that the level of information we have about the sources covered by NEIs
varies, so the extent to which we can set meaningful numeric goals will also vary by subject.
Meaningful information about pollution loading or the effect of preventive measures outlined in
an NEI will also vary. For clearly defined universes with reasonably good emissions
information, like the Refinery initiative, more specific information is available. Many other NEIs
will not have that level of detail, and obtaining it can be prohibitively expensive or even
impossible. Our experience with the current round of NEIs has helped us to understand the
challenges of measuring some of these outcomes, and this is a topic where we continue to learn.
Second, the means for collecting the information upon which to evaluate NEI success will not
always come through a government monitoring effort. For example, compliance data might be
available through self-reported information, as with NPDES discharge monitoring reports, or
may be obtained through consent decree-required reporting. (Note also that EPA must assure
that we do not run afoul of the Paperwork Reduction Act when collecting information on the
impact of our NEI efforts.) Inspections are not the only way to determine emissions or
compliance, nor are they likely to be the most cost effective in many instances; in a time of
increasingly limited inspection resources, we need to be sure that we use inspectors where they
can make the most difference.
Therefore, we suggest revising this recommendation to provide as follows: "Ensure that plans for
future National Enforcement Initiatives include an evaluation component that demonstrates the
extent to which the NEI strategy has achieved the goal(s) for the NEI identified by the EPA."
Proposed Corrective Action: By April 30, 2014, the OECA Office of Compliance will produce
guidance that requires the strategies for future National Enforcement Initiatives to include an
14-P-0184
21
-------
evaluation component for determining, where feasible, the extent to which the Initiative achieved
the goals established in the strategy.
Comments on the Report
1. Summary Page
The first paragraph of the Summary (under the "What We Found" heading) states as follows:
"The U.S. Environmental Protection Agency does not have evidence to show that the
agency's National Petroleum Refinery Initiative has resulted in emission reductions or
sustained compliance."
This is an incorrect statement. EPA does have information and evidence of emission reductions
resulting from each refinery's compliance with an NPRI consent decree, as detailed in periodic
(quarterly or semi-annually) consent decree implementation progress reports. These periodic
progress reports were made available as part of the OIG's evaluation.
As discussed in more detail below, it appears that the discussion in the Draft Report is referring
to a 2006 "Results of the Petroleum Refinery National Initiative" report (the "2006 progress
report") that was prepared by OCE. As part of that progress report, the emission reduction data
from individual progress reports were aggregated to produce year-by-year charts showing
reductions in emissions of sulfur dioxide (SO2) and nitrogen oxide (NOx) for the settlements that
had been reached at that time. Because this was an expensive and time-intensive effort that
required significant contractor resources (which we appreciate is acknowledged later in the Draft
Report), and because OCE's extramural funding levels have declined by nearly half since
FY2008, a comprehensive update to the 2006 progress report is not an affordable option.
Therefore, instead of diverting these declining resources away from oversight of existing consent
decrees and prosecution of new cases, OCE determined to aggregate this information and make it
public in conjunction with the completion (termination) of an individual refiner's consent decree
to show the full extent of reductions achieved under the consent decree. Thus, it is not correct to
state that EPA does not have this information under the NPRI consent decrees; instead, for cost
reasons, we have deferred pulling this information together in a manner similar to the 2006
progress report until the completion of a refiner's obligations under its consent decree.
In addition, data reported by the petroleum refining sector to EPA's National Emissions
Inventory (NEI)1 demonstrates a significant and steady decline in emissions in this sector since
1999, as shown in the following chart (in tons per year)2:
1 The NEI is a comprehensive and detailed estimate of air emissions of both criteria and hazardous air pollutants
from all air emissions sources. The NEI is prepared every three years by the Office of Air and Radiation, and the
data is made publicly available on EPA's website. See generally http://www.epa.gov/ttn/chief/eiinformation.html.
2 The first NPRI settlements were reached in 2000 and 2001, so as compared to 1999 NEI data, the emissions
reductions from these first settlements would begin to show up in the NEI data for 2002. The 1999, 2002 and 2005
NEI data for emissions of S02 and NOx shown in the chart is that reported to SIC code 2911 (petroleum refining);
2008 and 2011 NEI data for the same pollutants is that for the "industrial processes-petroleum refining" sector.
14-P-0184
22
-------
700,000
600,000
500,000
400,000
300,000
200,000
100,000
IS02
INOx
1999
2002
2005
2008
2011
Accordingly, the Agency has evidence indicating that sources in this sector that are required to
implement NPRI consent decrees have been steadily and dramatically reducing their emissions
over this period of time.3
OIG Response: We agree that the EPA collected evidence of emissions reductions in its 2006 effort.
However, the agency has not continued to measure emission reductions resulting from the NPRI. To
evaluate the outcome of this multi-decade project, the EPA needs to regularly evaluate the program in
the aggregate so that it can make modifications along the way to improve its environmental outcomes.
We also agree that the EPA collected evidence of compliance with the consent decrees. However, this
evidence has not demonstrated that facilities under consent decrees are, in fact, complying with the
elements for which the consent decrees were signed. We understand the EPA's reticence to diverting
resources away from other refinery work to monitor the success of this effort, but maintain that to
follow Office of Management and Budget guidance on operating government programs the EPA needs
to demonstrate that the resources put into this program have resulted in environmental
improvements.
The first, second and third paragraphs of the Summary Page go on to state:
3 In addition, one of the NPRI goals was to obtain consent decree commitments that were estimated to reduce NOx
and S02 emissions by 20% over pre-NPRI levels. NEI data indicates that emissions in this sector have declined by
approximately 75% since 1999, a far greater reduction outcome than OECA had set as a goal for the initiative. In
this respect, while enforcement is not claimed as the only reason for this decline in emissions, NEI data does
illustrate that emissions were in fact declining as expected as the NPRI consent decrees were implemented. It is also
worth noting that nearly the entire sector is covered by NPRI consent decrees - 80% by FY05 and currently over
90% - and no new federal regulations, such as new NSPS standards, were promulgated during this time period that
would account for this level of reductions.
14-P-0184
23
-------
"One of the goals of the NPRI was to increase compliance by 50 percent over its baseline
level. The EPA needs to determine whether the NPRI achieved the intended compliance
and emission outcomes. . . . Without analyzing the available data, however, the EPA
cannot determine whether the NPRI actually achieved established goals. . . . [T]o assess
the success of this strategy, the EPA must determine whether the strategy achieved
compliance, coverage, and emission-reduction goals, and reduced the environmental and
human-health risks associated with noncompliance."
The NPRI was concluded as an initiative and returned to the "core" program upon meeting each
of the four goals specified for completion as part of NPRI's overall strategy. The Draft Report
briefly summarizes two of these goals - the compliance improvement and emission reduction
goals - in Chapter 1 (on page 5), but the summary somewhat mis-states the goals. Specifically,
these two goals were set for the NPRI as follows:
• Compliance Improvement Goal: This goal was defined by the Strategy as a 50%
improvement in compliance over the 1995 baseline (which showed that 109 of the 156
facilities then in operation in the petroleum refining industry were in violation during the
period from August 1994 through August 1995, or a 70% noncompliance rate in the
industry). Thus, to improve compliance by 50%, the Strategy specified that 55 refineries
needed to be under consent decrees that would address past noncompliance by the end of
FY2005. Principally due to the successful achievement of settlements with 3 large
refiners in 2005 (ConocoPhillips, Valero and ExxonMobil, covering 32 refineries), a total
of 87 refineries - or 55.7% - were under consent decrees by the end of that fiscal year.
Thus, the compliance goal as specified by the Strategy was not only achieved, but was
exceeded.
OIG Response: We agree that these consent decrees constitute an important accomplishment.
However, as a matter of evaluating the outcomes of the NPRI as an environmental program, having a
facility under consent decree does not necessarily mean that the facility is in compliance. The agency's
interpretation is that a facility under consent decree is considered "in compliance." We do not see the
consent decree as the end of the line. When the EPA has returned to do on-site inspections, including
consent decree follow-up inspections, it has found issues with consent decree requirements or new
compliance issues at the facilities under consent decree.
• Emission Reduction Goal: This goal - a 20% reduction in NOx and SO2 combined - was
defined as securing commitments in enforceable consent decrees for new controls and
management practices that would be estimated to achieve a 20% reduction from 1995
emission levels, or just over 200,000 tons per year of SO2 and NOx. By the end of FY05,
settlements had secured commitments for an estimated reduction of 250,000 tons per
year, once all controls were implemented. Thus, the reduction goal as specified in the
Strategy was also not only achieved, but was exceeded by 50,000 tons per year.
Therefore, the evidence supports the view that the NPRI did successfully achieve the goals
established for this Strategy.
14-P-0184
24
-------
Based on the discussion elsewhere in the Draft Report, it appears that the Report is focusing on
whether the expected emission reductions that were estimated at the time of settlement will be
realized upon full implementation of the controls and requirements of each consent decree.4 As
noted below in the discussion of the "Noteworthy Achievements" of the NPRI (on Page 5 of the
Draft Report), the total emission reductions that are estimated at the time of settlement for all
NPRI consent decrees is 349,00 tons (93,000 tons of NOx and 256,000 of S02). NEI data for
this sector indicates that reductions in line with these initial estimates are in fact being achieved,
with a reduction of approximately 495,000 tons overall between 1999 and 2011 (more than
165,000 tons of NOx and 332,000 tons of SO2).
OIG Response: During the course of our evaluation, EPA staff informed us that National Emissions
Inventory data was not an accurate source for emissions data for the NPRI. These data are generally
based on estimates made by facilities using emission factors and do not account for differences in
production. As part of the consent decree requirements, facilities are required to regularly submit
emissions data to the EPA. We believe it makes sense to use the data submitted by refineries for the
purpose of calculating actual reductions.
EPA agrees that compliance with the terms of the NPRI consent decrees and the achievement of
these reductions are important objectives, and accordingly the successful implementation of and
compliance with consent decree-prescribed controls to reduce emissions is required prior to
termination of a refiner's consent decree. As noted above, providing a demonstration of the
emissions benefits of refiner's work to implement its obligations under its consent decree is one
of the planned components of consent decree termination.
2. Chapter One
Page 1: The discussion under "Background" is an unduly narrow characterization of the work
EPA does to increase compliance and reduce unlawful pollution. We suggest that this section be
revised to read as follows:
"Assuring compliance with environmental laws is an integral part of EPA's Strategic
Plan (available at http://www2.epa.gov/planandbudget) to protect human health and the
environment. The EPA's 2013 budget directs about one-tenth of its annual budget to
promoting compliance with environmental laws ($830 million in fiscal year 2013).
EPA employs a broad range of tools and approaches to assure and increase regulatory
compliance. EPA brings civil or criminal enforcement actions to address violations or
require regulated entities to clean up pollution. The Agency monitors compliance to
assure regulated entities obey applicable laws and regulations. It empowers communities
and the public through transparency and accountability by disclosing compliance and
performance information on regulated entities and government. It also empowers
communities by providing analytical tools that better display critical information, and
provide users with enhanced analytical and targeting capabilities. EPA helps builds our
4 In this respect, it should be understood that these reductions are estimates only, and not enforceable requirements
for a specific tons-per-year reduction.
14-P-0184
25
-------
state and tribal partners' compliance and enforcement capacity and oversees their
performance. EPA also helps businesses, federal facilities, local governments and tribes
meet environmental regulatory requirements through compliance assistance.
Robust enforcement is critically important for addressing violations and promoting
deterrence. However, enforcement alone will not be enough to achieve compliance
results that protect public health and the environment and ensure that businesses that
comply with the law are not disadvantaged relative to companies that do not play by the
rules. Therefore, EPA also is investing in a new paradigm called "Next Generation
Compliance" to improve compliance and reduce pollution. Next Generation Compliance
takes advantage of new information and monitoring technologies to enable EPA, states,
and tribes to get better compliance results and tackle today's compliance challenges. It
includes:
• Designing more effective regulations and permits that are easier to implement and
produce higher compliance and improved environmental outcomes.
• Using advanced emissions/pollutant detection technology so regulated entities,
government, and the public can more easily see and respond to pollutant
discharges, environmental conditions, and noncompliance.
• Requiring electronic reporting by regulated entities to generate more accurate,
complete and timely information on pollution sources, pollution, and compliance.
• Making the information we have today more accessible, and making new
information obtained from advanced emissions monitoring and electronic
reporting publicly available to expand transparency and improve the performance
of government and regulated entities.
• Developing and using innovative enforcement approaches to achieve more
widespread compliance and to help to increase the effectiveness of our
compliance work, such as making greater use of targeted deterrence approaches,
and self- and third-party certification tools.5"
Page 3: The third paragraph under the heading "Refineries Present Widespread Compliance
Challenges and Emit Harmful Pollutants" somewhat incorrectly summarizes the four "marquee
issue" areas of focus under the NPRI. The Draft Report states:
"Through inspections and additional research, the EPA identified four major areas where
refineries did not comply with the law:
1. Emissions from new or updated major refining units without a permit.
2. Fugitive emissions from refinery leaks.
3. Uncontrolled and unreported benzene waste.
5 See "OECA's FY2014 National Program Manager (NPM) Guidance" (June 23, 2013), at 10-11, available at
http://www2.epa.gov/sites/production/files/documents/fV14oecanpmguidance.pdf.
14-P-0184
26
-------
4. Use of flaring for routine purposes instead of for only venting dangerous gases."
Both for clarity and accuracy, this should be revised to more accurately identify the marquee
issue areas. Specifically:
"Through inspections and research, EPA identified the four most significant sources of
noncompliant emissions from refineries:
1. New Source Review/Prevention of Significant Deterioration (NSR/PSD)
requirements for FCCUs, heaters and boilers.
2. New Source Performance Standards (NSPS) Leak Detection and Repair
requirements Leaks from refinery equipment, such as valves, pumps, and
connectors.
3. National Emission Standards for Hazardous Air Pollutants requirements for
management of benzene wastes at refineries.
4. Uncontrolled and routine flaring (NSPS Subparts A&J)."
This revision, especially with respect to #1 and #4, is recommended because the largest sources
of refinery emissions covered by NPRI consent decrees are not always "unpermitted" - instead,
most were likely improperly permitted (e.g., incorrectly permitted as "minor" sources for
NSR/PSD purposes, did not have controls that represented required Best Available Control
Technology, etc.), in addition to being unpermitted. With respect to flaring, the issue addressed
by the NPRI is not whether flares should be used "only [for] venting dangerous gases" - by
definition and design, flares are only used for this purpose. Instead, the NPRI focused on
whether flares were being used routinely instead of on an emergency basis, and when used
whether flares were emitting at levels in excess of applicable limits.
Page 5: The discussion under the heading "Noteworthy Achievements" should be revised for
technical accuracy. Specifically, the second paragraph, third sentence states:
"For example, companies cumulatively pledged to reduce pollution by 93,000 tons of
NOx and 256,000 tons of SO2 annually" (emphasis added).
As noted in the discussion of the "Summary Page" above, the NPRI consent decrees specify a
variety of enforceable control equipment, new (lower) emission levels, and improved
management practices covering the four marquee issues; however, the consent decrees do not
specify certain levels of emission reductions as enforceable requirements.6 As discussed above,
the cumulative emission reductions cited are estimates (not enforceable "pledges") of the
expected emissions benefits from these controls.
Page 6: The listed bullet points could benefit from some technical corrections, as follows:
First bullet: This bullet point states that "many" companies have agreed to install low-
leaking valves as part of their enhanced LDAR program. In fact, this is a relatively
recent development, first agreed to as part of the settlement with Murphy Oil (2010), and
included in refinery settlements reached since that time. This technology did not
6 There is a limited exception for a specified reduction in NOx emissions, in tons per year, from refinery heaters and
boilers. See, e.g., Chevron (2003), 133; Citgo (2004) 154; Sinclair (2008), 142; Murphy (2010), 131.
14-P-0184
27
-------
previously exist, but is an example of the "technology-forcing" impact of the NPRI, in
which EPA's enforcement efforts have effectively driven improvements to existing
control technologies and the development of new technologies in this industrial sector.
(Other examples under the NPRI include the development of more effective pollutant-
reducing catalyst additives for use in FCCUs and, more recently, the use of cutting-edge
infrared imaging technologies for flare "efficiency" controls, as in the Marathon and BP
Whiting settlements in 2012 and the Shell Deer Park settlement in 2013.)
Second bullet: This bullet point states that "most" consent decrees require installation of
continuous emission monitors (CEMS). In fact, all consent decrees require installation of
CEMS on major emission sources (e.g., FCCUs, large heaters, etc.).
Third bullet: This bullet point states that "some" consent decrees include requirements
for fenceline monitoring of emissions. Solely for purposes of clarification, this is a recent
development (similar to the discussion of low-leaking valves, above) that was first agreed
to in the Murphy Oil settlement in 2010, and included in several other settlements agreed
to since that time.
With respect to each of these technical corrections to the points on Page 6, it is important to
underscore that OECA agrees with Draft Report's conclusion that these kinds of provisions
increase transparency (in the case of fencline monitoring) and enhance EPA's ability to manage
compliance in this complex industry by enabling EPA, facilities, and the public to more readily
access emission data as well as provide protections for those living near these facilities.
3. Chapter Two
Pages 9-10: We do not think the evidence supports the statements on Page 9 that "EPA has not
demonstrated whether the NPRI resulted in sustained compliance or emission reductions in the
petroleum-refining sector," or that "since 2006, the EPA has not evaluated progress toward
achieving these outcomes through the consent decrees," or similar statements elsewhere on pages
9 and 10. Even though OECA does not have the funds to do a formal update of the 2006
progress report, we do have the means to measure success by aggregating data demonstrating
emission reductions obtained as the result of the completion of consent decree commitments:
These can be handled on a refiner-specific basis as part of the consent decree-termination
process. To date, one settlement (the 2003 Coastal Eagle Point settlement) has been terminated
in its entirety,7 and others are nearing termination in the next several years. OECA has
determined that the use of its (declining) resources is better put toward ensuring that prior to
termination all consent decree requirements are completed and the expected emission reduction
benefits are realized, rather than putting those resources toward a general updating of the 2006
7 For business reasons, the Coastal refinery was shut down and dismantled in 2010. Accordingly, there are no
further emissions from refinery operations at this facility.
14-P-0184
28
-------
progress report.8 Finally, and perhaps most importantly, EPA's NEI data demonstrates the
significant decrease in emissions in this sector during the term of the NPRI, and the record of
enforcement of NPRI consent decree requirements likewise demonstrates that OECA's oversight
has been successfully achieving compliance and obtaining expected emission benefits.
Throughout the report, including on Pages 9-10, the content and tone suggest that relying on self-
reported data is a somehow questionable practice. Our experience is that self-reporting is a very
important and powerful tool in the compliance arsenal. Increased use of well structured self-
reporting strategies will be an important part of our strategy to increase compliance while we are
reducing the size of our staff. It is not the case that inspections are the only, or even sometimes
the best, way to accurately gauge compliance or the success of pollution reduction efforts, and to
the extent that the Draft Report suggests otherwise we strongly recommend you reconsider.
Technological innovation is making new approaches to compliance and pollution verification
possible, and we hope that the OIG will be updating its reviews and recommendations, just as we
are updating our practices, to reflect the power of these new approaches. It is important to
recognize and seize opportunities presented by new technologies at any time, but especially so in
a time of declining resources, and not adhere to prior practices when new approaches promise
better and more efficient results. One example of the use of self-reporting in the context of this
initiative is demonstrated by the September 2013 enforcement action involving the Total (2007)
NPRI consent decree. In that action, the $8.75 million penalty obtained for consent decree
noncompliance, plus additional injunctive relief, was based primarily on EPA's review of
required company reporting. See "Texas Refinery Will Pay $8.75 Million for Failing to Comply
with Enforcement Settlement to Resolve Air Violations," at
http://vosemite.epa.gov/opa/admpress.nsf/d0cf6618525a9efb85257359003fb69d/cQ7d422e06400
be685257bec005fal 81! opendocument (press release);
http://www2.epa.gov/sites/production/files/2013-10/documents/lstamendtotal-cd.pdf (consent
decree amendment requiring corrective action to address noncompliance); and
http://www2.epa.gov/sites/production/files/2013-10/documents/total-stipulationandorder.pdf
(order requiring payment of penalties).
OIG Response: We did not evaluate the veracity of self-reported data and have made that clear in the
final report. However, we found instances where the EPA inspectors found violations during follow-up
inspections at refineries. We concluded that although self-reports are an essential part of compliance
and enforcement, continuing, periodic EPA inspections play an important role as well.
The Draft Report also ignores the variety of inspections that EPA does, as well as important
work done by the states. Within this sector, EPA has conducted some type of consent decree
compliance inspections at refineries every year nationwide, in addition to inspections at
refineries not yet under consent decree. Second, and as illustrated in more detail below, EPA
relies on and works in conjunction with its state partners (many of whom are consent decree co-
8 One of the four interlocking consent decrees covering the Shell family of refineries (Motiva. Shell Deer Park and
Equilon) covers only the requirements applicable to refinery heaters and boilers. While several of the refineries
subject to this consent decree have been sold to other refiners. Shell has completed its obligations under the Heaters
and Boilers consent decree at the refineries it still owns, which has been terminated as it applies to Shell. Consistent
with OECA's plans for termination, when available EPA will post graphs showing emissions reductions obtained
under this consent decree at http://www2.epa.gov/enforcement/motiva-enterprises-llc-refinerv-settlement.
14-P-0184
29
-------
plaintiffs) for inspections and oversight at refineries. The Agency relies on a number of tools
and partners to confirm compliance with applicable requirements at refineries; an on-site
inspection is just one of these tools, but is not the sole method of reliably ensuring compliance or
obtaining compliance information from a regulated facility. In addition, EPA is building on such
self-reporting as part of its "Next Generation Compliance" goal to achieve greater compliance
and reduce pollution using advances in monitoring and information technologies, e-reporting and
other measures. See, e.g., "Enforcement Goals: Next Generation Compliance" at
http://www2.epa.gov/enforcement/enforcement-goals; see also "Next Generation Compliance"
at http://www.epa.gov/compliance/data/svstems/icis/vmeeting/vmeeting6a-panel.pdf. In addition,
the suggestion that EPA should deploy its inspection resources "to determine whether
compliance changed under consent decrees" (regardless of risk) would have the unfortunate
effect of diverting inspection resources away from higher-risk and higher-priority facilities.
While EPA would agree that conducting routine compliance inspections at all covered refineries
on a regular basis would be likely to yield a statistically significant analysis and evaluation of
whether refiners' compliance changed under consent decrees, this is not feasible in light of
current and expected future resource constraints, and is also not the intent or purpose of
compliance inspections.
This section also states that "between 2003 and 2012, Region 6 and the OECA conducted
inspections at only 19 of the 35 consent decree facilities in Region 6." This tends to suggest that
each refinery under consent decree should have been inspected by EPA during this time.
However, this overlooks several important facts and methods used by EPA to assure compliance:
1. Throughout the 2003 to 2012 time period, negotiations covering all these refineries were
either underway or had recently been completed {i.e., the number of consent decree-
covered refineries is not static during this period). In 2003, for example, only 14
refineries were under an NPRI consent decree. In 2005, settlements were reached with
refiners covering an additional 19 refineries. Another 5 refineries were covered under
consent decrees reached between 2007 and 2012. On-site inspections by EPA were
conducted at more than half of these refineries during this time period, which is
consistent with a risk-based approach to inspections9 and the use of other tools for
determining consent decree compliance.
2. EPA is not the sole inspection or enforcement authority, as states also play a significant
role in addition to EPA. The Draft Report incorrectly cites Marathon's Garyville LA and
ExxonMobil's Beaumont TX refineries as two that had not been inspected. Under the
applicable state Compliance Monitoring Strategy (CMS), as major sources these facilities
are required to be inspected on a periodic basis (between 2 to 5 years); such inspections
include compliance with consent decree requirements as reflected in Title V permits,
among other applicable permits. Consistent with these CMS requirements, and as
9 Because the installation of major controls under NPRI consent decrees are typically large, multi-year construction
projects, a consent decree-compliance inspection conducted shortly after a settlement is reached, and before major
implementation work has begun, would not be a wise use of inspection resources, and progress at these stages can be
more efficiently tracked through review of required implementation reporting. Instead, and consistent with the risk-
based approach to the use of inspection resources, on-site compliance inspections are more commonly employed to
determine compliance with major consent decree milestones (which may be some years after settlement).
14-P-0184
30
-------
documented in EPA's "Enforcement & Compliance History Online" Database (ECHO),
Texas has conducted several hundred Partial Compliance Evaluation (PCEs),10 Title V
compliance reviews and on-site inspections at the Beaumont refinery, and EPA has
conducted an additional 20 PCEs during this time (see http://www.epa-echo.gov/cgi-
bin/get 1 cReport.cgi9tool=echo&IDNuinber=4824500018). Similarly, Louisiana has
conducted nearly 90 PCEs, Title V compliance reviews and on-site inspections at the
Garyville refinery, and EPA has conducted an additional 15 PCEs as well (see
http://www.epa-echo.gov/cgi-bin/getlcReport.cgi?tool=echo&IDNumber=2209500013).
OIG Response: After reviewing information for these two refineries we still maintain that our
statement is correct. These refineries have not received a compliance-based inspection from
the EPA. For example, according to the agency's analysis of 2006 emission data, the
Beaumont Exxon refinery had milestones for hard limits for S02 to be established toward the
end of 2005. Additionally, there was a milestone for "hardware limits effective" for NOx in
2009. We would expect a compliance-based inspection by the EPA to check these milestones.
3. The regular course of dealing between EPA and refiners under consent decrees also
results in actions to address noncompliance, without the need for formal inspections or
the use of other enforcement tools. For example, disclosures by ExxonMobil of
violations at its Beaumont refinery (and other refineries) resulted in the assessment of
stipulated penalties and commitments for additional injunctive relief in 2008 (see
http://www2.epa.gov/sites/production/files/documents/exxonmobil-stipulation-
agreel208.pdf and http://www2.epa.gov/sites/production/files/documents/3rdmod-
exxonmobill208-cd.pdf). In addition, refiners anticipating a compliance issue have been
encouraged by EPA work out a resolution with the Agency in advance, without the need
for more formal processes (such as a Notice of Violation or other enforcement followup);
the outcome of these prophylactic efforts to address the anticipated compliance problem
also typically results in consent decree modifications that yield additional emission
reductions (see, e.g., http://www2.epa.gov/enforcement/second-amendment-consent-
decree-sinclair-tulsa-refining-company-sinclair-wvoming).
4. As noted above in the example of the Total consent decree, required consent decree self-
reporting is also used to assure compliance.
In other words, the Draft Report tends to focus on on-site compliance inspections as the principal
or only means of reliably ensuring compliance; however, the DraftReport should recognize that
such inspections are not the sole method of compliance oversight at consent decree-covered
refineries, and that EPA uses a number of effective tools to perform this oversight function.
Pages 10-11: The Draft Report's discussion of EPA's 2006 progress report states:
"Company-specific analysis showed mixed results. Data from some companies
demonstrated the companies would likely meet or exceed anticipated emission-reduction
goals. Emission data from other companies had increased since signing the consent
decrees."
111 See "Inspections and Evaluations," at http://www.epa.gov/compliance/monitoring/inspections.
14-P-0184
31
-------
The basis for the draft report's statement above is not clear. As a threshold matter, the consent
decrees for which emissions reductions had been achieved at the time of the 2006 progress report
were still in the early stages of implementation (because of the extent and complexity of consent
decree requirements, most implementation schedules run for in excess of 10-12 years or more);
several covered in the 2006 progress report were in only the first or second year of
implementation {i.e., a significant amount of work still remained to be done in later years under
these consent decrees). Even at this early stage of implementation under the 7 settlements that
had been reached at that time, overall emissions from covered refineries were reduced by nearly
50% from pre-settlement levels.11 In two instances, as noted in the Draft Report, some emissions
had increased from pre-settlement levels (Coastal - FCCU NOx, and CHS - FCCU SO2), which
is of limited relevance given that the dates for installation of equipment and controls to lower
emissions from these units were still in the future at the time of the 2006 progress report. See
Coastal consent decree ]}15 (FCCU NOx controls required by 2008 or 2010, depending on
control technology selected); CHS consent decree ]}33 (FCCU SO2 controls required by the end
of 2007 or 2009, depending on the control technology selected). Nevertheless, overall emissions
from all refineries under settlement at that time were cumulatively lower than their pre-
settlement levels. This data does not seem to support a "mixed results" conclusion - it shows
immediate, if not yet complete, progress in achieving expected emission reductions.
The achievement of nearly half of the estimated emission reductions from these early settlements
by the initial stages of implementation also demonstrates the inaccuracy of the Draft Report's
statement at the top of Page 10 that "EPA's press releases may have overstated the successes of
the agency's NPRI strategy." To the contrary, it tends to suggest that EPA's estimates were
conservative and likely understated the emission benefits of these consent decrees. Although the
Draft Report's discussion of EPA's emission reduction estimates is focused on settlements
reached through 2005, data and information available since that time supports the conclusion that
these estimates are conservative and likely understate total reductions, rather than to overstate
them. As noted above, NEI data from the 1999 to 2011 time period shows a reduction in
emissions from this sector that is exceeds EPA's total emission reduction estimates for all NPRI
settlements, including those reached at the end of or after the 1999-2011 timeframe12 (that is,
once the emission reductions from these more recent settlements are accounted for, total
emission reductions in this sector will be even greater than EPA's initial estimates from the
settlements).
The following paragraph from the Draft Report also appears to acknowledge the likelihood that
estimates at the time of settlement would tend to understate actual reductions:
"However, the 2006 emissions analysis did not show the progress made by all facilities
under the consent decree at the time. . . . The OECA officials also said that the 2006
11 Emissions of S02, NOx, PM and CO were approximately 360,000 tons peryear in 1997; by 2005 these emissions
were approximately 190,000 tons peryear. See
http://www2.epa.gov/sites/production/files/documents/refinervinitiative-powerDoint021111.pdf (slide 28).
12 Settlements reached at the end of or after this period include Shell Chemical and Murphy (2010); Hovensa and
Western (2011); Coffevville (2012); and Big West and Countrvmark (2013).
14-P-0184
32
-------
analysis likely missed some emission reductions due to the lag between the finalization of
the decree and the installation of controls at the refineries. For example, the
implementation schedule for the 2005 ConocoPhillips settlement requires major
equipment to reduce flaring emissions to be installed periodically through 2012 at some
facilities."
Accordingly, we suggest that the Draft Report should be revised to reflect this discussion. In
addition, a minor revision to the second full paragraph on Page 11 is needed for accuracy. As
noted, the 2006 progress report did not include emissions reductions for every settlement that
had been reached through 2005, as the Citgo. ConocoPhillips. Sunoco. Valero and ExxonMobil
settlements had been reached late in the year. However, rather than "missing" emission
reductions from those settlements, those consent decrees were too early in their implementation
schedules to quantify or report emission reductions that had been achieved at that time (a
suggested revision would be that "the 2006 analysis did not include some emission reductions
due to the lag between the finalization of the most recent settlements at that time and the
installation of controls at the refineries covered by those newer decrees").
Page 11: The Draft Report includes the following statement in the second full paragraph:
"An agency official stressed that achieving emission reductions was more important than
achieving compliance with consent decree requirements. However, enforcement officials
said that due to resource constraints, they discontinued this effort after the 2006
contractor's analysis."
There are two concerns with the quoted passage.
• The first sentence is not an accurate statement of OECA's objective or intent with respect
to refinery settlements. While achieving consent decree-required emission reductions is
one of the objectives of the NPRI, the suggestion that this objective is "more important"
than consent decree compliance does not follow, as the two are necessarily linked -
compliance with emission reduction requirements in NPRI consent decrees is necessary
to achieve the bargained-for emission benefits. We suggest this internally inconsistent
statement is either a misunderstanding or an erroneous attribution.13
• The second sentence appears to suggest that "this effort" (to ensure the achievement of
emission reductions) was "discontinued" after 2006 is not correct. EPA has continued to
oversee and enforce existing consent decrees to ensure compliance with those
requirements,14 as well as to continue to bring other refiners that had not settled as of
13 In this respect, context is important. As noted earlier in the comments on Page 10, EPA has encouraged refiners
anticipating a compliance problem with their consent decree to discuss such issues with EPA in advance. These
joint efforts typically result in consent decree amendments addressing the prospective problem and obtaining
additional emission reductions. See, e.g., 4th Amendment to the Motiva "Heaters and Boilers" consent decree
(2005).
14 Oversight and enforcement of NPRI consent decrees includes the following modifications to consent decree
requirements that provide additional emission benefits, over and above those contained in the original settlement
agreement: BP (3 amendments plus one new consent decree for the Whiting, Indiana refinery); ConocoPhillips (4
14-P-0184
33
-------
2006 under NPRI consent decrees.15 What OECA decided not to continue after 2006 was
to regularly update the 2006 progress report; this is unrelated to the continuing efforts to
obtain additional emission reductions in this sector. The discontinued updating of the
2006 progress report due to budgetary considerations is unrelated to the issue of efforts to
oversee and enforce implementation of the NPRI consent decrees.
Page 11: The second full paragraph concludes with the following statement regarding the
decision to discontinue updating the 2006 progress report:
"[W]e believe it [discontinuing the updating effort] does not provide interim information
that would allow for mid-course corrections if facilities are not meeting requirements."
As noted above, the 2006 progress report and EPA's continuing efforts to ensure compliance
with NPRI consent decrees are unrelated. In fact, the evidence demonstrates the opposite - that
the Agency is taking action to make such "mid-course corrections" on a consent decree-specific
basis in appropriate circumstances for those refiners that are not meeting their enforceable
obligations, as noted above. In addition, and perhaps more importantly, aggregated annual data
on total emission reductions cannot be used to identify noncompliance with any specific consent
decree requirement; only consent decree oversight activities can confirm compliance (see
discussion above).
Page 11: The last paragraph in this section states as follows:
"By taking these steps [to update the 2006 progress report], the OECA could better
determine the success of the NPRI, and modify the NPRI by determining what factors
prevented facilities from realizing the projected emission reductions. Because the EPA
has not used these data to track whether emission reductions were occurring, it is unclear
whether the companies implementing consent decrees are making the progress
promised."
OECA fully agrees that the 2006 progress report was a valuable effort to show the overall
progress made up that time, which is why it was initially undertaken. If resources allowed,
updating the report would also be interesting, although, as noted above, we already know from
the NEI data for this sector that there have been significant reductions in emissions since 1999.
However, for all the reasons described above, we think there is a way to verify compliance and
to evaluate the overall effectiveness of the initiative in a much more cost effective manner
through the already required consent decree termination process.
Page 12: The Draft Report suggests that there is not a process for conducting terminations of the
consent decrees. In fact, the consent decrees each contain a full section specifying the legal
amendments); ExxonMobil (4 amendments); Flint Hills (Koch) (2 amendments); Marathon (2 amendments); Total
(1 amendment); Sinclair (4 amendments); and Sunoco (3 amendments).
15 New NPRI consent decrees agreed-to after 2006 include Total. Hunt. Valero/Premcor. Sinclair. Holly. Frontier.
Wyoming Refining. Shell Chemical. Murphy Oil. Hovensa. Western. Hess. Coffewille. Countrvmark. and Big
West.
14-P-0184
34
-------
requirements for termination. A separate process is not necessary. The termination process as
specified in NPRI consent decrees includes requirements for certifications of completion of all
consent decree requirements by refiners (under penalty of perjury or contempt), payment of
stipulated penalties, and a requirement for full compliance with the consent decree for the year
prior to termination, among other requirements. In addition, EPA - and, where applicable, the
state co-plaintiff(s) - independently verify a refiner's compliance with these provisions. It is not
clear whether the Draft Report is suggesting new or different requirements for termination, but of
course the provisions of the consent decree govern, and new provisions cannot now be added.
If what the OIG is saying is that some personnel were not familiar with the consent decree
requirements for termination, this may reflect that some regions have not yet had a consent
decree eligible for termination and so had not as yet refamiliarized themselves with those
requirements. OECA is working directly with those regions affected by terminations as they
come due. However, we agree that a reminder about consent decree termination provisions
would be useful and we are intending to issue a memorandum to the regions refreshing their
understanding of these provisions.
Page 13: The third full paragraph expresses concerns regarding EPA's ability to oversee NPRI
consent decrees because the Agency is currently understaffed with sufficient technical expertise
in this sector, concerns regarding the Agency's ability to attract and retain qualified experts in
this area, and concerns stemming from the very limited resources available in both Headquarters
and in Regions. OECA agrees that declining staffing levels and extramural resources for the
enforcement program is a serious constraint, both for enforcement in this sector as well as
overall. We are doing the best we can with what we have available to us for oversight and
enforcement of these consent decrees, and in response to declining resources we have shifted
strategies to reduce the resource burden and to focus on key enforcement and oversight needs to
assure compliance and reduction of emissions, which necessarily results in cutting back on work
that consumes extramural resources but is not essential to achieving compliance or emissions
reductions.16
Page 14: The first paragraph under "Conclusion" states as follows:
"The EPA should be assessing the success and outcomes of the NPRI. Doing so enables
the EPA and the regulated industry to demonstrate success where it has occurred, identify
lessons learned, modify this enforcement approach to improve success rates, and enhance
future enforcement initiatives as needed. Noncompliance and pollution at refineries
particularly affect communities located close to refineries. Noncompliance could lead to
emergencies and upsets at facilities, which could endanger refinery workers and
communities."
OECA entirely agrees but would add that we think we are doing this already. We agree that the
work already done in this sector is very important for protecting communities from harmful
16 Other efforts to reduce the NPRI resource burden include streamline (where possible) consent decree reporting
requirements, mandated "review and approval" requirements, etc. See "Guidance on Streamlining Oversight in
Civil Settlements" (OECA Jan. 10, 2013)
http ://intranet.epa. gov/oeca/oce/io/documents/finalstreamliningmemo 11013 .pdf.
14-P-0184
35
-------
pollution, and as you know we are continuing to focus on this sector for other emissions, as
evidenced by our recent cases addressing hazardous emissions from flaring. We have learned
many things from our work in this sector that we are applying elsewhere. For example, the
efficiencies realized from a "global" approach for addressing environmental compliance on a
company-wide basis (where possible), as opposed to employing a unit-by-unit/facility-by-facility
approach, has been extended into other efforts. On the other hand, the costly and time-
consuming "test and set" approach to establishing emission limits, which was employed
extensively in early NPRI consent decrees, is now generally avoided in favor of setting "hard" or
numeric limits in later consent decrees, and "test and set" is used only where absolutely
necessary.17 As noted above, other lessons learned from the NPRI have been incorporated into
guidance for all consent decrees.18
Finally, OECA agrees that upset events and excess emissions can have an adverse impact on
local communities - one of the reasons for the selection of the NPRI was the significant
environmental justice impact and proximity of refineries to overburdened communities.
However, for accuracy and completeness the Draft Report should recognize that "upsets" or
emissions from the malfunction of emission sources may not be violations of applicable EPA
regulations. See, e.g., 40 C.F.R. § 60.102a(h).
Contact Information
If you have any questions or concerns regarding this response, please contact the OECA Audit
Liaison, Gwendolyn Spriggs, at 202-564-2439.
Attachment
cc: Lawrence Starfield, OECA
Susan Shinkman, OECA/OCE
Pam Mazakas, OECA/OCE
John Fogarty, OECA/OCE
Phillip Brooks, OECA/OCE
Lauren Kabler, OECA/OCE
Gwendolyn Spriggs, OECA/OAP
17 Compare the 2001 BP consent decree at ** 14.A-14.F. pages 14-23 (extensive, multi-year provisions for
establishing FCCU limits) with the 2013 Big West consent decree at "12. page 14 and " 23. page 16 (immediately
effective limits with no lengthy process).
18 See "Guidance on Streamlining Oversight in Civil Settlements" (OECA Jan. 10, 2013),
http ://intranet.epa. gov/oeca/oce/io/documents/finalstreamliningmemo 11013 .pdf.
14-P-0184
36
-------
Corrective Action Plan
Determining the Outcomes of the National Petroleum Refinery Initiative May Identify
Opportunities to Strengthen Future National Enforcement Initiatives dated
Report No. OPE-2012-3065, September 23, 2013
Recommendation
Corrective Action
Target Date
1. Develop and implement a plan to
determine whether consent decrees
signed as part of the NPRI are leading
to promised improvements in
compliance and sustained reductions in
pollution. The plan should establish
procedures for verifying company
completion of consent decree
requirements for termination, including
company assessments of facility
pollution reductions and comparisons
with estimated reductions
OECA will issue a memorandum to
the regions to remind them of the
requirements for termination as
provided in refinery consent decrees,
and in particular to work with OCE
for confirmation that all necessary
requirements for termination have
been met. In addition, OECA will
continue to work with refiners at the
time of termination of their consent
decree to develop graphs of annual
emission reductions obtained under a
consent decree at the time of
termination.
45 days from the date
of a final report.
2. Incorporate requirements that ease
the resource burden on the EPA to
monitor refinery progress (e.g., CEMS
and fence- line monitoring) into future
consent decrees or amendments to
existing consent decrees under the
NPRI.
Consent decrees include requirements
and guidance issued in lanuary 2013
to enforcement personnel for
incorporating resource-burden
requirements into consent decrees.
OECA considers this
corrective action
complete.
3. Inform the public about the extent to
which the NPRI resulted in sustained
compliance improvement at facilities
and reductions in emissions agreed to
in consent decrees, as highlighted as a
key challenge for OECA in EPA's
congressional budget justification.
OECA will continue to work with
refiners at the time of termination of
their consent decree to develop graphs
of annual emission reductions
obtained under a consent decree at the
time of termination.
45 days from date of a
final report (included as
part of proposed
corrective action for
Recommendation #1).
4. Ensure that plans for future industry
initiatives of the NEI include a
monitoring and evaluation component
that demonstrates the extent to which
the NEI strategy leads to improvement
in compliance and human-health risks
identified by the EPA.*
OECA will produce guidance that
requires the strategies for future
National Enforcement Initiatives to
include an evaluation component for
determining, where feasible, the
extent to which the Initiative achieved
the goals established in the strategy.
April 30,2014
* Text of this recommendation subject to change pending approval by the OIG of EPA's
proposed alternative text.
14-P-0184
37
-------
Appendix C
OIG Response to the Agency's Revised
Corrective Action Plan
OIG assessment of
EPA response and
Completion
EPA response and
OIG recommendation
corrective action
date
corrective action
1. Develop and
"OECA will issue a
45 days from
We agree with the EPA's
implement a plan to
memorandum to the
the date of a
proposed actions. This
determine whether
regions to remind them of
final report.
recommendation is resolved.
consent decrees
the requirements for
signed as part of the
termination as provided in
NPRI are leading to
refinery consent decrees,
promised
and in particular to work
improvements in
with OCE for confirmation
compliance and
that all necessary
sustained reductions
requirements for
in pollution. The plan
termination have been
should establish
met. In addition, OECA will
procedures for
continue to work with
verifying company
refiners at the time of
completion of consent
termination of their
decree requirements
consent decree to develop
for termination,
graphs of annual emission
including company
reductions obtained under
assessments of facility
a consent decree at the
pollution reductions
time of termination."
and comparisons with
estimated reductions.
2. Incorporate
"Consent decrees include
OECA considers
After reviewing the January
requirements that
requirements and
this corrective
2013 guidance issued to
ease the resource
guidance issued in January
action
enforcement personnel, we
burden on the EPA to
2013 to enforcement
complete.
agree this corrective action is
monitor refinery
personnel for
complete. This
progress (e.g., CEMS
incorporating resource-
recommendation is resolved.
and fence-line
burden requirements into
monitoring) into
consent decrees."
future consent
decrees or
amendments to
existing consent
decrees under the
NPRI.
14-P-0184
38
-------
OIG assessment of
EPA response and
Completion
EPA response and
OIG recommendation
corrective action
date
corrective action
3. Inform the public
"As provided in the
45 days from
We agree with the EPA's
about the extent to
proposed corrective
date of a final
proposed actions. This
which the NPRI
action for
report
recommendation is resolved.
resulted in sustained
Recommendation #1,
(included as
compliance
OECA will continue to
part of
improvement at
work with refiners at the
proposed
facilities and
time of termination of
corrective
reductions in
their consent decree to
action for
emissions agreed to in
develop graphs of annual
Recommen-
consent decrees, as
emission reductions
dation #1).
highlighted as a key
obtained under a consent
challenge for OECA in
decree and OECA will post
the EPA's
this information on EPA's
congressional budget
public website."
justification.
NOTE: In a subsequent
communication, the EPA
clarified that it has long
planned to pull together
and post on the EPA's
website consent decree-
specific emission
reduction data once all of
an individual consent
decree's requirements
have been completed.
This would supplement
the initial summary of
emission reductions
achieved by settling
refiners through 2005,
and posted to the EPA's
website.
14-P-0184
39
-------
OIG assessment of
EPA response and
Completion
EPA response and
OIG recommendation
corrective action
date
corrective action
4. Ensure that plans for
"OECA will produce
April 30, 2014
We agree with the EPA's
future NEIs include an
guidance that requires the
proposed actions. This
evaluation component
strategies for future
recommendation is resolved.
that demonstrates the
National Enforcement
extent to which the
Initiatives to include an
NEI strategy achieves
evaluation component for
the goal(s) for the NEI
determining, where
identified by the EPA.
feasible, the extent to
which the Initiative
achieved the goals
established in the
strategy."
NOTE: In a subsequent
communication, the EPA
clarified its corrective
action by saying it will
produce guidance that
requires the strategies for
future NEIs to include an
evaluation component for
periodically determining,
where feasible, the extent
to which the initiative is
achieving the goals
established in the
strategy. When
appropriate and
practicable, the strategies
will employ outcome
measures to examine the
success of the initiative.
14-P-0184
40
-------
Appendix D
Distribution
Office of the Administrator
Assistant Administrator for Enforcement and Compliance Assurance
Agency Follow-Up Official (the CFO)
Agency Follow-Up Coordinator
General Counsel
Associate Administrator for Congressional and Intergovernmental Relations
Associate Administrator for External Affairs and Environmental Education
Principal Deputy Assistant Administrator for Enforcement and Compliance Assurance
Audit Follow-Up Coordinator, Office of Enforcement and Compliance Assurance
14-P-0184
41
------- |