OFFICE OF INSPECTOR GENERAI
^   Evaluation Report
IV
             EPA Needs to Improve Tracking of
             National Petroleum Refinery
             Compliance Program Progress and
             Impacts
             Report No. 2004-P-00021
             June 22, 2004
                           1 *';»- .'i'." *" •' •''" '• %*"

-------
Report Contributors:
Kim Bryant
Katie Butler
Andrew death
Jeff Harris
Jeff Hart
Abbreviations

CO         Carbon Monoxide
EPA        Environmental Protection Agency
FCCU       Fluidized Catalytic Cracking Unit
GAO        General Accounting Office
H2S         Hydrogen Sulfide
ICIS        Integrated Compliance Information System
LDAR      Leak Detection and Repair
NESHAP    National Emissions Standards for Hazardous Air Pollutants
NSPS       New Source Performance Standards
NSR        New Source Review
NOX        Nitrogen Oxide
OECA      Office of Enforcement and Compliance Assurance
OIG        Office of Inspector General
OMB       Office of Management and Budget
PM         Paniculate Matter
PSD        Prevention of Significant Deterioration
PART       Program Assessment Rating Tool
SO2         Sulfur Dioxide
VOCs       Volatile Organic Compounds
Cover photo:  Image of petroleum refinery provided by EPA staff.

-------
                     UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                                  WASHINGTON, D.C. 20460
                                                                THE INSPECTOR GENERAL


                                    June 22, 2004
MEMORANDUM
SUBJECT:   EPA Needs to Improve Tracking of National Petroleum Refinery
             Compliance Program Progress and Impacts
             Report No. 2004-P-00021

TO:         Thomas V. Skinner
             Acting Assistant Administrator
             Office of Enforcement and Compliance Assurance
Attached is our final evaluation report regarding the Environmental Protection Agency's (EPA's)
national refinery compliance program. This report contains findings that describe problems and
lessons learned from the national petroleum refinery compliance program and corrective actions
the Office of Inspector General (OIG) recommends. This report represents our opinion, and
findings in this report do not necessarily represent the final EPA position. EPA managers will
make final determinations on matters in the report in accordance with established procedures.

Action Required

As the Action Official, EPA Manual 2750 requires you to provide this Office with a written
response within 90 days of the final report date.  The response should address all
recommendations.  For the corrective actions planned but not completed by the response date,
please describe ongoing actions and provide a timetable for completion.  If you disagree with a
recommendation, please provide alternative actions for addressing the findings reported. Our
team would like to work with your staff in developing the corrective action plan.  Please ask your
staff to contact Jeff Hart, Assignment Manager, at (303) 312-6169 for arrangements.

We appreciate the efforts of EPA officials in working with us to develop this report. If you or
your staff have any questions regarding this report, please contact me at (202) 566-0847 or Kwai
Chan, Assistant Inspector General for Program Evaluation, at (202) 566-0827.

                                                 ~
                                                           ?}
                                             Nikki L. Tinsley c

-------

-------
                     Executive Summary
             The Environmental Protection Agency's (EPA's) Office of Enforcement and
             Compliance Assurance (OECA) selected the petroleum refinery industry as a
             national enforcement priority in 1996 because refineries had the highest
             inspection-to-enforcement ratio of the 29 industry sectors ranked by EPA.

             The 145 operating petroleum refineries in the United States span 9 of EPA's 10
             regions and 33 States. Petroleum refineries account for significant releases of
             pollution into the environment. In 2001, refineries released over 35,000 tons of
             toxic air pollutants, with 75 percent released to the air, 24 percent to the water,
             and 1 percent to the land. These pollutants seriously impact human health and the
             environment, and include pollutants known or suspected to cause cancer or other
             serious human health effects.

Results in Brief

             EPA and the U.S. Department of Justice have developed and implemented an
             integrated refinery compliance strategy that addresses the most important
             noncompliance problems. EPA's national refinery compliance program began in
             1996, and over the last 8 years EPA implemented a succession of tools and
             strategies as its refinery program evolved and as EPA identified specific
             compliance problems.  EPA's integrated strategy includes compliance assistance,
             inspections, enforcement, and compliance incentives. As of March 2004, the
             program resulted in refineries agreeing to invest more than $1.9 billion in
             pollution control technologies, pay civil penalties of $36.8 million, and  implement
             supplemental environmental projects valued at approximately $25 million.
             Further, EPA projects the national refinery compliance program will result in
             annual reductions of approximately 44,000 tons of nitrogen oxide, 95,000 tons of
             sulfur dioxide, and significant amounts of other pollutants.

             However, OECA's performance measurement and reporting approach for the
             national petroleum refinery  program has not provided useful and reliable
           . information necessary to effectively implement, manage., evaluate, and
             continuously improve program results. OECA has not established and
             communicated clear goals, systematically monitored refinery program progress,
             reported actual outcomes, or tracked progress toward achievement of consent
             decree goals.  In addition, during consent decree implementation, EPA delays may
             have delayed emissions reductions and compromised compliance. OECA must
             resolve planning issues and  delays, and begin to measure outcomes, to ensure
             timely emissions reductions and to optimally protect human health and the
             environment, especially for  people living in the vicinity of refineries.

-------
             EPA learned several important lessons that it should apply throughout its refinery
             program and consider for other enforcement and compliance assurance programs.
             EPA effectively demonstrated some of these lessons learned in the refinery
             program, such as focusing on specific enforcement concerns, becoming
             knowledgeable about the industry, and encouraging EPA regional and
             headquarters staff to effectively work together. Other lessons learned that EPA
             needs to improve upon include the need to clearly communicate roles and
             responsibilities, meaningfully engage stakeholders throughout the process, and
             diligently oversee consent decree compliance.

Recommendations

             We made various recommendations to OECA related to the development of clear
             overall refinery program goals.  We also made recommendations to OECA to
             improve refinery consent decree implementation and tracking, and to ensure better
             measurement and reporting of refinery program outcomes.

Agency Comments and OIG Evaluation

             In its April 2,2004, comments on the draft report, OECA stated that the report
             will help EPA as it continues to implement the refinery program and other new
             programs or initiatives. OECA also stated that the report had several significant
             shortcomings. OECA agreed with 10 recommendations, disagreed with 5, and
             partially agreed with 3. We made changes to the report as we determined
             appropriate.  We include a summary of EPA's chapter-specific comments and our
             evaluation of those comments at the end of each chapter. We also provide as
             Appendix G the Agency's memorandum summarizing its overall comments,
             including its comments on the recommendations. Appendix H contains  our
             evaluation of those comments. OECA also provided us with detailed comments
             as an attachment to its summary memorandum. We have posted this attachment
             and  our evaluation of OECA's comments on our web site at
             http://www.epa.gov/oig/publications.htm.

-------
                    Table  of Contents
Executive Summary	    i
Preface	    1
 Chapters
     1     EPA's National Refinery Compliance Program
           Evolved to Become Fully Integrated	   3
     2     EPA Has Not Yet Demonstrated Environmental and
           Human Health Impacts of the Refinery Program	   13
     3     Refinery Program Lessons Learned	   29
 Appendices
     A    Details on Scope and Methodology  	   35
     B    Key Information on U.S. Petroleum Refineries 	   41
     C    Petroleum Refining Process Flow Chart 	   45
     D    Refinery Releases and Effects	   47
     E    National Refinery Program Time Line	-	   49
     F    Consent Decree Process Flow Chart	   51
     G    EPA Comments on the Official Draft Report 	   53
     H    OIG Evaluation of EPA Comments	   65
     I     Distribution	   69

-------

-------
                                  Preface
Purpose of Evaluation
             Enforcement and compliance assistance practices have seen considerable
             innovation in recent years. The Environmental Protection Agency's (EPA's)
             Office of Enforcement and Compliance Assurance (OECA) adopted a problem-
             based approach to addressing environmental problems.  This strategic approach
             includes: (1) giving up-front consideration to which combination of tools to use  '
             (e.g., enforcement, compliance assistance) when addressing particular
             environmental problems; and (2) encouraging up-front development of measures
             to assess progress and outcomes. According to an internal OECA evaluation,
             prior to using its problem-based approach, OECA implemented integrated
             strategies on an ad-hoc basis with limited measurable results.

             The Office of Inspector General (OIG) has planned a series of environmental
             enforcement evaluations with the overall objective of answering the following
             question: What impact have the enforcement and compliance assurance actions,
             activities, and policies of EPA and its partners had on the regulated community's
             compliance with environmental requirements and on protecting human health and
             the environment? At OECA's request, we agreed to first pilot-test our overall
             evaluation approach on a single priority area, hi consultation with OECA, we
             selected the petroleum refinery priority area for this pilot evaluation and
             developed objectives that would demonstrate the feasibility of our overall
             evaluation approach as well as provide meaningful insight on EPA's refinery
             program. Our specific objectives were to answer the following questions:

             1 .  What is the nature and extent of the regulated petroleum refinery universe?

             2.  To what extent have EPA (OECA, other program offices, and EPA regions),
                the U.S. Department of Justice, and EPA's partners (States and Tribes)
                developed  an integrated strategy (that considers compliance assistance,
                compliance incentives, inspections, and enforcement actions) to address
                priority noncompliance problems at petroleum refineries?

             3.  Does the performance measurement and reporting approach for petroleum
                refineries provide the information necessary to effectively implement, manage,
                evaluate, and improve OECA's petroleum refinery program?

             4.  Did EPA effectively implement and manage the petroleum refinery program?

             5.  What lessons can be learned from OECA's petroleum refinery program?
                                         1

-------
             We address questions 1 and 2 in Chapter 1, questions 3 and 4 in Chapter 2, and
             question 5 in Chapter 3.
Scope and Methodology
             We conducted our evaluation of EP A's refinery program between June 2003 and
             March 2004.  We performed our evaluation in accordance with Government
             Auditing Standards issued by the Comptroller General of the United States.

             Our evaluation focused on OECA's enforcement and compliance assurance
             activities at petroleum refineries from fiscal years 1996 through 2004. To answer
             our five objectives, we interviewed staff and collected and analyzed data from
             OECA, EPA's National Enforcement Investigations Center, EPA regions, States,
             industry, environmental groups, and the U.S.  Department of Justice.

             See Appendix A for a detailed description of our scope and methodology,
             including details on prior reviews.

-------
                             Chapter 1

      EPA's  National Refinery Compliance Program

             Evolved to Become Fully integrated	_

            EPA and the U.S. Department of Justice have developed and implemented an
            integrated refinery compliance strategy that addresses the most important
            noncompliance problems. Petroleum refineries account for significant releases of
            pollution into the environment.  EPA's national refinery compliance program
            began in 1996, and over the last 8 years EPA has implemented a succession of
            tools and strategies as its refinery program evolved and as EPA identified specific
            compliance problems. EPA's integrated strategy includes compliance assistance,
            inspections, enforcement, and compliance incentives. As of March 2004, the
            program has resulted in refineries agreeing to invest more than $1.9 billion in
            pollution control technologies, pay civil penalties of $36.8 million, and implement
            supplemental environmental projects valued at approximately $25 million.
            Further, EPA projects the national refinery compliance program will result in
            annual reductions of approximately 44,000 tons of nitrogen oxide (NOJ, 95,000
            tons of sulfur dioxide (SO2), and significant amounts of other pollutants.
Background
            Petroleum represents the single largest source of energy for the United States.
            Petroleum refineries span 9 of EPA's 10 regions and 33 States (there are no
            refineries on tribal lands).  See Figure 1.1.

                Figure 1.1: Location and Number of Refineries by State and EPA Region
                           \ piT X
                                             EPA REGION
                                                        f»-
                                                        l«f
                                    if       - - t- >..vr     j~s
-------
             Fifty-seven corporations oversaw 145 petroleum refineries operating in those
             33 States as of January 1,2003. These refineries processed a total of 16,5 million
             barrels of crude oil each calendar day.  See Appendix B for a complete listing of
             U.S. petroleum refinery companies and key information on each.

             Petroleum refineries account for significant releases of pollution into the
             environment.  The Clean Air Act, Clean Water Act, and Resource Conservation
             and Recovery Act regulate the majority of releases at petroleum refineries. In
             2001, refineries released over 35,000 tons of toxic air pollutants according to
             EPA's most current Sector Facility Indexing Project data, with 75 percent released
             to the air, 24 percent to the water, and 1 percent to the land.  In 1999, according to
             the most current data from EPA's AirData system, refineries released
             approximately 243,000 tons of NOX, 396,000 tons of SO2, and 412,000 tons of
             other common air pollutants (also known as "criteria air pollutants").

             Petroleum refining is the physical, thermal, and chemical separation of crude oil
             into its major components, which are further processed into a variety of finished
             petroleum products. Appendix C illustrates the complexity of the petroleum
             refining process, potential releases and release points, and the major applicable
             environmental regulations. For example, according to one refinery expert, an
             average petroleum refinery processing 120,000 barrels of crude daily can have
             100,000 connection points. Each connection point is a possible leaking emissions
             source.  The fluidized catalytic cracking units (FCCUs) and heaters and boilers
             represent the largest refinery emitters.  FCCUs use heat, pressure, and a catalyst to
             break larger hydrocarbon molecules into smaller ones.  FCCUs can emit several
             thousand tons of NOX and S02 per year.

             Petroleum refinery emissions seriously impact human health and the environment.
             In 2000, OECA reported that 45 percent of all refineries at that time were within
             3 miles of population centers containing 25,000 or more people, and 26 percent
             were within 3 miles of population centers containing 50,000 or more people.
             Appendix D provides a summary of the human health and environmental effects
             of the following common air pollutants released at refineries: volatile organic
             compounds (VOCs); SO2; NOX; particulate matter (PM); carbon monoxide (CO);
             hydrogen sulfide  (H2S); and toxic air pollutants. Toxic air pollutants include
             pollutants known or suspected to cause cancer or other serious human health
             effects.

Refinery Program Uses an Integrated Strategy

             OEC A's current refinery program incorporates various tools and strategies as part
             of an integrated effort to address the most important compliance problems. Using
             research, investigations, and national EPA experts, EPA identified four priority
             areas that addressed the most important noncompliance areas in the industry.

-------
EPA also used compliance assistance and incentives to educate the industry and
address noncompliance.

Identifying Priority Areas

During the initial stages of the program, OECA and regional officials used
multiple sources of information to identify refinery priority areas. These sources
included inspections, formal EPA information requests to refineries, and industry
trade journals. OECA and regional officials shared with each other the results of
these initial research efforts and began to focus (or target) investigations on the
noncompliance areas indicated by their research. EPA's national experts
continued gaining experience regarding compliance issues within the refinery
industry, and helped select the four Clean Air Act priority areas that became the
refinery program's focal point State, refinery industry, and environmental
interest groups generally agreed that these four priority areas represented the most
important noncompliance problems at petroleum refineries:
 2.  Flaring / New Source Performance Standards (NSPS)1

 SI  S-ftakQeteeHdrfaftdft^airfLDAi1?}          '     /""""     _V ""  ;  -  0\

 4,  Benzene Waste National Emissions Standards for Hazardous Air Pollutants (NESHAP)
Figure 1.2 briefly describes the primary problems, related regulations, and
solutions EPA and the refineries agreed to implement for each priority area.  The
agreement between EPA and refiners resulted in consent decrees that primarily
focused on these four priority areas.
  The refinery program addressed two different flaring programs: the elimination of acid gas
flaring; and the reduction of hydrocarbon flaring. The NSPS priority focused on ensuring that
companies complied with NSPS at their recovery plants, flares, and fuel gas combustion devices.

-------
           Figure 1.2: Problems, Related Regulations, and Solutions for Priority Areas
                                           He3m^adqfohly_^
:: r^uiwd by "the Clcaii:Air|Actj Also; EPA identified fttoblems with NSPS cconpliaiioc at iiUfiir f«coy«jy
;; jpii..^ ai^ fiirf gi» cdiii^
 ;Solntkm: Consent decrees require fkci)toto;:ide^(y loot causwof each flfliing incident; unplsmentj
 plans to ad^ws root oauSc» such as iiHtallingjisiw.equipnienJ, revising operating prooe*tote».:or:];•!;•!:••;
• • providing *"Tffifog; :(
 LD&R

 Prabkin: EPA found But fugitive emissions, or
 refine^ leaks, were two to ten times higher
 than reported by the facility  Ifunconeckd,
 &c^Ui« continijed to release fogidve emmtons
 of VOCfc and other hazardous a^poEutants
 The cumulative effect of fugitive emissions
 posed significant health and environmental
 Regulation: CAA Section 112  40CFRPart
 61. Subpart I. Pott €3, Subparto H and CC; and
 Part 60 Sntparts VV. OOO, and QQQ,

 Solution: Consent 
-------
             Compliance Assistance and Incentives Provided in Various Forms

             OECA conducted various compliance assistance and incentive activities
             including:

             •   Making presentations at industry       to^tkmtf tempter A&sktane*
                conferences describing              Acfivi«e^fert>te(«ftech«feal6ssiWtt:Bib*t
                noncompliance issues.              totfyt "1* fi$gulst£d eorhrfHjnity understand
                                                 a«d comply witfrertyirorvrnBntal regulations or
                _   ,  .      ,  ,.    .  .          Help other camFdJsnc&asalstapte^rovid^ri
             •   Developing and disseminating       aid ta m^i^ad cdramunfcy.
                Enforcement Alert newsletters on
                   ,   _,   »     ~      .  .          Definition of Compliance Incentive
                each of the four refinery priority     _ „,    .       .T  .. , ,
                                   ' r    '      Poliaes and programs that eliminate, reduce,
                areas-                            or waive penalties under certain conditions.

             •   Developing two compliance assistance guidance documents; one to help
                increase understanding of the Refining Maximum Achievable Control
                Technology Standard, and another to improve understanding of regulations
                covering benzene waste and transfer operations.

             •   Developing the slotted guidepole initiative that provided a compliance
                incentive for refinery companies to install  controls within specific time frames
                to address emissions from petroleum storage tanks. If companies complied
                with their schedules for installing acceptable controls, EPA would eliminate
                penalties.

Four Basic Phases Noted in National Refinery Compliance Program

             EPA's national refinery compliance program began in 1996 and evolved as EPA
             learned more about the noncompliance issues and applied a variety of tools and
             strategies to address those issues. While OECA did not specifically outline a
             phased program or develop a comprehensive master plan, in reviewing the
             strategy, it appeared that  four phases emerged. We used these phases to facilitate
             our description of the refinery program.  In many instances, activities overlapped
             as one phase continued while a new phase began. Appendix E provides a detailed
             time line of the refinery program's evolution.

             Phase t; Refineries Become a Natfonaf Enforcement Priority

             EPA had long conducted inspections and taken enforcement actions in the refinery
             industry. OECA began to focus significant attention on refinery compliance
             concerns in 1996 when refineries became an enforcement priority due to the
             industry's high rate of noncompliance and pollutant releases.  OECA reported in
             1996 that the refining sector had the highest inspection-to-enforcement ratio of the
             29 industry sectors ranked by EPA. In 1996, when compared to 496 other sectors,
             OECA ranked refineries number one for releases of VOCs, number two for SO2,
             number three for NO2 (a particular type of NOX), number four for PM10

-------
(a particular type of PM), and number five for CO releases. In its 1996/1997
Memorandum of Agreement guidance, OECA ranked petroleum refineries
number one for noncompliance and identified petroleum refining as one of three
national enforcement priorities. The Memorandum of Agreement guidance
generally sets forth the Agency's enforcement and compliance assurance priorities
and activities for a 2-year period.  Designation as an OECA enforcement priority
meant that an industry received special emphasis. OECA maintained refineries.as
a national priority and began working with regional officials to explore ways to
address compliance issues.
Although EPA identified refineries as a national priority, OECA officials said they
did not see significant improvement in the regional approach to addressing
compliance issues during 1996 and 1997. According to a senior OECA official,
after 2 years of typical regional inspections of the refinery industry, the
inspections did not identify significant national problems. However, OECA
obtained anecdotal information from OECA's National Enforcement
Investigations Center and regional investigations regarding problems at refineries.
OECA determined that it needed a more comprehensive approach to help assess
the extent of compliance problems identified in a few regions. OECA began
coordinating with the National Enforcement Investigations Center and regional
staff who were developing expertise regarding compliance issues at refineries.
The anecdotal information caused OECA to shift its focus from routine Clean Air
Act inspections that broadly assessed refinery compliance using checklists, to
more targeted,  resource-intensive investigations. These investigations focused on
assessing emissions released from certain industrial processes as opposed to
compliance with specific statutory requirements, and identified problems not
found during typical inspections. OECA officials stated that this phase of the
refinery program was extremely successful at developing new targeting and
investigative tools.

Phase lit: Program Shifts to Pursue Global Settlements

In 2000, OECA shifted the refinery  program focus to pursue voluntary global
settlements with refinery companies that resulted in consent decrees.2  OECA and
regional officials coordinated with the U.S. Department of Justice who took the
lead on all of the global settlement negotiations. OECA referred to these
settlements as "global" because they applied to all facilities owned by one
company. OECA's approach presented corporate officials with the option of
avoiding possible investigation and litigation. OECA's strategy included
2 OECA found success with global settlements in other industries. However, OECA officials
stated that their prior experience related to a much narrower set of issues and on a much smaller
scale than what they achieved under the national refinery compliance program.

                             8

-------
             coordinating with interested States and local authorities. States that signed
             consent decrees had their own legal claims against the settling refineries and
             received a share of the penalties paid by companies.  Along with EPA and the
             U.S. Department of Justice, 18 States or local authorities have signed consent
             decrees as of May 21,2004. Beginning in 2002, OECA began to shift its
             emphasis from pursuing new negotiations to concluding on-going investigations
             and negotiations, since over 80 percent of the domestic refining capacity universe
             had entered into global consent decrees, was in negotiations with EPA, and/or was
             under active investigation. As part of the shift in emphasis, OECA planned for
             regions and States to assume larger roles with new investigations and
             negotiations.

             Phase IV: Consent Decree Implementation Initiated

             Implementation of refinery consent decrees began anew and additional phase.
             The first two consent decrees entered the implementation phase in early 2001.
             The consent decrees span 8 to 10 years and  require coordination and
             communication among OECA, EPA regions, States, and industry. While the
             signing of a consent decree ends the settlement process for that company, it begins
             a new process of oversight and interaction by and between the parties. As of
             2004, OECA continues to conduct negotiations, assist regions in assuming a
             larger role with the refinery program, and work with refiners to implement
             consent decrees.
Settlements Projected to Result in Significant Emissions Reductions

             By March 2004, OECA had entered into 11 global settlements or consent decrees
             covering 42 of the 145 refineries. As shown in Table 1:1 below, the settlements
             covered 39 percent of total U.S. petroleum refining capacity.3
               Table 1.1 represents the settlements on the date of their entry by the Court. Since that date,
             companies may have sold some of their refineries and, while those refineries remain subject to the
             global settlements, they may be owned by separate refining entities. Consequently, this table may
             not easily reconcile with the information in Appendix B.

-------
              Table 1.1: Refinery Companies in Global Settlements
\->y' Company
Koch Petro)eum;Group •;;•::::*¥ - •^•^
•B&Bcplbrafibfta^^
Mbtiva; Enterprises LLC/Equilon
iEnfefpifa^WiBtF^i^Reflnlng'^p;
Marathon AshlanciPetrdl^um: LUG
Navajo Refining Company and s '&J;
Mbritaha:;Refinihg -Compa'ny^*^'
Conoco, jTicJf;;?;;; Ji^|l|-|^^||^Sfj
;L^:;o|||ii;;.| mi^m&iim
:&i8foriitefci&:;fiKz;?:iW£i*f^;\
,:li» ri O ' inCi. fwftflOX ):':::::':' ::.':':;/:•:;.'•;> -••:• -.•:••• •:;:•' ]'.o ';':::.;'.
CO3Sw3r'w9.9'0- Poifrt- OS' . C0^'>'^"'' '^'' '^'"'' ^:
^^^e^^^^gfjSjiil
Date Company
Signed
Settlement wltti
oecA
December 2000
January 2001
March 2001
May 2001
December 2001
December 2001
March 2003
October 2003
October 2003
October 2003
October 2003
Total Refineries under Decrees (Listed Above)
Total Refineries Not under Decrees
TOTALS
Number
of
Refineries
2
8
9
7
2
4
1
5
1
1
2
•fillMlM
:l^Mi:i
mm$$
Production
Capacity
(barrete par
cal.
-------
              The global settlements also relieved OECA from having to conduct resource-
              intensive investigations at each refinery a company owned.  According to OECA,
              a refinery-by-refinery, issue-by-issue approach, in which EPA conducted an
              individual inspection or investigation at each and every refinery followed by
              information requests, notices of violation, negotiations and/or litigation, could
              take many years and require resources beyond EPA's means.

              The refinery consent decrees require each company to take various actions over
              the next several years. These actions include implementing air pollution controls
              as  well as developing policies and procedures that go beyond compliance with
              existing regulations. As shown in Table 1.2, both the LDAR and benzene priority
              areas require companies to incorporate "enhanced" practices beyond regulatory
              requirements.  In addition, OECA and the companies agreed to test and use
              innovative technologies.

                            Table 1.2: Examples of Consent Decree Requirements
               NSR/PSD
               Fiaring/NSPS
               LDAR
               Benzene
                                            Requirement* In Consent Decrees
Install controls and Continuous Emissions Monitoring Systems to reduce
and measure SO2, NOx, PM, and CO from FCCUs, heaters, and boilers.
Implement program to investigate the cause of flaring incidents.
Conduct root cause analyses and take corrective actions.
Requires NSPS compliance at refinery's flares, sulfur recovery plants, and
fuel gas combustion devices.
Install new sulfur recovery units and tail gas control devices to ensure
compliance with emissions standards and good air pollution control
practices.
Implement enhanced monitoring program.                 ,
Train all refinery LOAR personnel annually.
Train all other refinery operations and maintenance personnel on aspects
of LDAR relevant to the employee's duties.
Monitor valves more frequently than regulations require.
Train all employees who draw benzene waste samples.
Establish standard operating procedures for all control equipment used to
comply with the benzene waste MESHAP.
Train all equipment operators oh procedures.
Audit all laboratories that perform analyses of benzene waste NESHAP
samples.
Conclusion
              EPA and its partners developed and implemented an integrated strategy that
              included compliance assistance, compliance incentives, inspections, and
              enforcement actions. EPA staff implemented a succession of tools and strategies
              as the program evolved. OECA's integrated strategy addresses the most important
              noncompliance problems at petroleum refineries. A significant phase and turning
                                            11

-------
             point in the refinery program's evolution came with the negotiation of voluntary
             global settlements. These settlements resulted in consent decrees between OECA
             and refinery companies owning multiple facilities.

Agency Comments and OIG Evaluation

             OECA stated that the background discussion in this chapter was vague and, at
             times, did not demonstrate an understanding of the refinery program. OECA
             stated that the background information implied that they directed the refinery
             program at all pollutants from refineries. In our opinion, the background
             information in this chapter provides a broad overview of the refinery industry,
             including some of the major processes, releases, and health and environmental
             impacts caused by refining. It describes the refinery industry as a whole and puts
             in perspective why EPA decided to monitor and address noncompliance within
             this industry. This chapter also describes how the refinery program focuses on
             four priority areas under the Clean Air Act and how the refinery program could
             reduce specific  pollutants  through companies' effectively implementing consent
             decrees.
                                         12

-------
                             Chapter 2
   EPA Has Not Yet Demonstrated Environmental and
     Human Health impacts of the Refinery Program
            OECA's performance measurement and reporting approach for the national
            petroleum refinery program has not provided useful and reliable information
            necessary to effectively implement, manage, evaluate, and continuously improve
            program implementation and results.  OECA has not established and
            communicated clear goals, systematically monitored refinery program progress,
            reported actual outcomes, or tracked progress toward achievement of consent
            decree goals, hi addition, during consent decree implementation, EPA delays may
            have delayed emissions reductions and compromised compliance. OECA must
            resolve planning issues and delays, and begin to measure outcomes to ensure
            timely emissions reductions and to optimally protect human health and the
            environment, especially for people living in the vicinity of refineries.

Program Lacked Clear Goals, Performance Measurement, and
Outcome Reporting

            OECA has not established clear program goals, performance measures, or a
            reporting system to track progress. Clear goals, performance measures, and a
            reporting system are essential to ensure that program managers have a systematic
            approach for gathering performance information for effective decision making.
            OECA has not clearly and precisely defined official program goals,  used
            performance measures to manage the refinery program, or reported actual
            environmental outcomes.

            Clear Goals, Performance Measures, and Reporting Essential

            Clear program goals, performance measures, and a reporting system to track
            progress are all essential to effectively implement, manage, evaluate, and improve
            any program. Government programs need a systematic approach for gathering
            performance data and reporting progress toward goals to ensure that program
            managers and Congress have performance information for decision making.

            The Office of Management and Budget's (OMB's) fiscal 2004 budget proposal for
            EPA stated that one of EPA's top two challenges included, "Tracking and
            demonstrating programs' effectiveness in achieving public health and ecosystem
            protection goals." To help improve program and funding decisions, OMB
            evaluated 11 EPA programs, including the civil enforcement program (which
            includes the refinery program), using OMB's Program Assessment Rating Tool
            (PART). OMB developed the PART to provide a consistent approach to

                                      13

-------
evaluating Federal programs during budget formulation. OMB found that the
absence of outcome-based performance data and, in some cases, any data,
hindered the Agency in evaluating the impacts of its programs on lie environment
and public health.  OMB recommended that EPA establish performance measures
focused on outcomes and efficiencies.

EPA agreed on the need to focus on program results and to use such data in
decision-making, but EPA appealed OMB's PART evaluation results. The
Deputy EPA Administrator said in a November 6, 2002, memo that, "One of my
goals as we implement the President's Management Agenda is to ... continuously
improve the results orientation of our programs and inform our decision-making
with high-quality information on program performance." However, the Deputy
Administrator also concluded that the PART instrument and process needed
dramatic overhaul to have any real value.5

EPA faces a challenge in the next few years to improve the linkage between its
program results and budget resources, which includes developing program
measures.  We found OMB's findings and observations concerning EPA's civil
enforcement program generally consistent with what we specifically found
concerning the refinery program.

In fiscal 2004, OEC A began moving toward a more performance-based approach
to program management as described in its December 18, 2002,
Recommendations for Improving OECA Planning, Priority Setting, and
Performance Measurement. The Assistant Administrator for OECA decided that
OECA should develop a performance-based approach for national priorities such
as the refinery program.  OECA's Recommendations document stated that
strategies "... should always include... a goal or set of goals and performance
measures that allow progress to be assessed...," and "these elements should be in
place before the implementation period  ... begins."

Program Goals Not Clearly Defined

OECA has not clearly and precisely defined official program goals, or ensured
mat everyone working on the refinery program in EPA headquarters and regional
offices and State offices have the same understanding of the goals.  Different
OECA officials referred to different goals and measures for the refinery program
at different times.  Table 2.1 summarizes various goals we identified for OECA's
refinery program.
5 We did not assess the PART tool or OMB's application of the tool.

                            14

-------
                 Table 2.1:  Various Refinery Program Goals
^\- " vVf Goal •"
• Develop innovative approaches to achieve
increased compliance within the industry
sector priorities
• Reduce emissions from refineries
• Bring the refineries into long-term compliance
(with the issues investigated)
• Ensure more consistent interpretations and
enforcement of regulations
• Obtain significantly reduced emissions/
discharges/releases by 20 percent
• Obtain substantially improved compliance
rates by 50 percent . •
• Significantly improve the industry compliance
rate and reduce the total emissions that
resulted from noncompliance
• Address 1 00 percent of refineries either
through consent decrees or other enforcement
actions
Soprce ' - \ ,
Fiscal Year 1996/1997 OECA
Memorandum of Agreement guidance
(National Priority Sector)
Fiscal Year 1998 OECA Accomplishments
Report (Refinery Strategy)
OECA National Sector Strategies
document - not dated (National Sector
Strategy)
Fiscal Year 2001 OECA Accomplishments
Report (Refinery Strategy)
2003 and 2004 interviews with senior
OECA managers
OECA managers and staff did not agree on program goals. Throughout our
evaluation, they cited multiple goals listed in Table 2.1, and did not formally
assess their progress toward meeting any of these goals.  For example:

•  OECA officials said Memorandum of Agreement documentation served as
   updates to the strategy. However, during our field work, OECA officials said
   that either the Memorandum of Agreement documents contained inaccurate
   information or that they were not familiar with that information.

•  While some managers stated that the 50-percent and 20-percent goals were the
   official program goals, other OECA managers disagreed.

•  While some managers stated that EPA had met the 50-percent and 20-percent
   goals, officials could not provide specific data to support their conclusion,

Two OECA senior executives told us that establishing and tracking an overall
program goal for the refinery program, or any other program, would require
additional resources. Both said that it was more important to use their limited
resources to take additional enforcement actions (e.g., seek additional consent
decrees) than it was to measure progress toward program goals. We believe that
without clear and well-defined program goals, it was difficult for OECA to
identify appropriate performance measures and track progress.
                             15

-------
Performance Measures Not Used to Manage Program

Although OECA developed some performance measures, it has not used them to
manage the refinery program. For example, while one early OECA strategy
document included nine goal areas and more than 50 separate output or activity
measures, OECA did not track or communicate progress toward these goals.
Performance information should feed back into the management of the refinery
program, but OECA could not take advantage of such feedback because OECA
lacked a formal system for capturing this information. OECA agreed and intends
to use appropriate performance measures and outcomes as part of the fiscal 2005
priority planning process.

In lieu of performance measures to manage the program, OECA officials held
national meetings and scheduled monthly conference calls with regional staff to
review program progress and discuss regional commitments to support the
strategy.  In addition, OECA and regional staff formed issue-specific work groups
to develop investigative tools, discuss issues, and devise solutions. Senior OECA
officials stated that they also tracked program-related activities through regular
telephone and electronic mail communications and "status of activities" charts.
However, an OECA official acknowledged that OECA had not tracked or
disseminated data for most of the activity or output measures OECA established.

Reporting Systems Not Focused on Environmental Outcomes

Refinery program reporting focused on projected rather than actual environmental
outcomes. OECA reported program results in two ways:

•  Press Releases. OECA used press releases to communicate the signing of
   consent decrees to the public. Press releases reported the projected emissions
   reductions at full implementation of consent decrees (consent decrees lasted
   8 to 10 years) and the dollars companies agreed to pay in penalties as a result
   of consent decrees.  OECA management did not plan to issue press releases or
   other reports to the public detailing the actual measured outcomes of consent
   decree implementation because OECA management did not believe the press
   would be interested.

•  Reporting to Congress. EPA used the Integrated Compliance Information
   System (ICIS) to report EPA enforcement program results to Congress under
   the Government Performance and Results Act. For refinery consent decrees,
   OECA input data into ICIS representing (1) the projected annual emissions
   reductions that would be realized once implementation was complete, (2) the
   dollar amount of penalties generated, and (3) the dollar value of required
   supplemental environmental projects. According to OECA, ICIS reporting
   was not designed to capture, and did not capture, information about
   environmental outcomes from the consent decrees, such as demonstrated
   environmental and human health benefits.
                            16

-------
             EPA used three systems to collect infonnation on consent decree implementation:
                                                             V
             •  Company data collection through consent decree reports.
             •  Monthly conference calls between EPA managers and staff working on
                consent decree implementation (described above).
             •  Contractor-developed consent decree tracking system (described in detail later
                in this chapter).

             However, EPA did not use these systems to demonstrate progress toward meeting
             consent decree goals. Consent decrees required companies to provide quarterly
             reports that included actual emissions data related to NSR and PSD issues - the
             most significant sources of refinery emissions reductions in consent decrees.
             OECA used the information to set some emissions limits that consent decrees did
             not specify. However, OECA did not use this information to monitor, verify, or
             report progress toward achieving consent decree goals.

OECA Has Not Tracked  Progress Toward Consent Decree Goals

             OECA should closely and regularly track actual emissions reductions and progress
             toward consent decree goals for three reasons: (1) refineries have a history of
             noncompliance (described in Chapter 1); (2) refineries emit toxic chemicals that
             affect human health and the environment (see Appendix D); and (3) some
             information used to develop consent decree emissions limits were based on
             estimates - facility estimates, pollution control equipment estimates, or both.

             OECA officials do not plan to regularly verify or monitor actual refinery
             emissions. A national OECA refinery expert said OECA would use emissions
             data to assess progress at 4 years into consent decrees and at the conclusion of the
             decrees.  However, two monitoring events over 8 years will not provide OECA
             with information about company or overall refinery program progress toward
             predicted emissions reductions, and would limit OECA's ability to modify and
             improve existing consent decrees or ongoing negotiations.

             Through consent decrees, EPA sets compliance schedules and emissions reduction
             goals for companies.  OECA officials said company self-certification processes
             assured them that companies remained on their compliance schedules. Consent
             decrees require companies to demonstrate that they  reach emissions reductions
             goals by the fourth year of consent decree implementation. However, during our
             evaluation, EPA could not demonstrate progress toward meeting emissions
             reductions goals in any priority area, as shown in Table 2.2, because EPA did not
             analyze information about company progress toward emissions reductions.
             Table 2.2 shows compliance status as "on a compliance schedule per consent
             decree requirements" and emissions status as "currently unknown" because
             OECA lacked data that demonstrated compliance or emissions status at the time
             of our evaluation.
                                         17

-------
                      Table 2.2: Current Status of Compliance and
                       Emissions Reductions in the Priority Areas
... Prtwjfty
Area
NSR/PSO
Flaring/
NSPS
LDAR
Benzene
" Current ';-
Compliance
Sla&is
On a
compliance
schedule per
consent
decree
requirements
Reaching
compliance
On a
compliance
schedule per
consent
decree
requirements
, Progress Toward
Emission?
; Reductions
Currently unknown
Currently unknown;
although National
technical lead for
flaring documented
a decline in flaring
events from 1 998
through early 2004.
Currently unknown
Potential
Measurement
Data/Sources
Monitoring data
reported to
OECA on a
quarterly basis;
4- and 8-year
assessments
Flaring Root
Cause Analyses
State
Inspections,
company
monitoring data,
AP-42" and other
estimated data
Issues Associated vrittt Measures
OECA is not assessing emissions
reductions in this area on an ongoing
basis. OECA could use actual
monitoring data to establish limits and
analyze catalyst tests, but companies
need not demonstrate emissions
reductions until the fourth year of a
consent decree.
Companies do not measure the
constituents of flare emissions.
Companies provide OECA root cause
analysis reports for each flaring
incident that OECA can then use to
measure the number of flaring
incidents. From the reports, OECA can
estimate the pollution resulting from the
flaring event.
Many States were not advised to track
the area and others were not tracking
due to implementation delays. OECA
must rely on company-developed
sampling and monitoring plans, and
4-year company-supplied audits.
             OECA indicated that OECA and States share responsibility for ensuring consent
             decree implementation and verifying refinery emissions. However, an OECA
             executive acknowledged that headquarters had not provided guidance to States or
             EPA regions indicating that inspections should take the four priority
             noncompliance areas into account. EPA and State staff did not generally include
             the priority noncompliance areas in typical Clean Air Act inspections.  These
             factors left little assurance that EPA or States would verify self-reported company
             progress under consent decrees.

Late and Absent EPA Responses to Consent Decree Documents
May Have Delayed Program Outcomes

             Refinery program outcomes will depend on successful consent decree
             implementation. Late and absent EPA responses to company consent decree
             documents may  have delayed company implementation of projects designed to
             reduce emissions and compromised compliance. EPA delays developed as soon
              AP-42 estimates are emission factors used to estimate emissions from a source when more
             reliable emissions data are not available. The emission factors used to develop AP-42 estimates
             are known to have limited accuracy.

                                        18

-------
as the implementation phase began, and persisted because OECA did not
effectively plan how it would manage and monitor consent decree
implementation. OECA took steps that reduced but did not eliminate the delays,
and did not take steps to address the delays in a timely fashion.

Program Outcomes Depend on Implementation

OECA needs to accurately track hundreds of consent decree milestones to ensure
that companies comply with consent decree requirements and that expected
emissions reductions actually occur. Appendix F provides a detailed flow chart of
the consent decree process. As depicted in the flow chart, after a company signs a
consent decree and the court enters the decree, the implementation phase of the
consent decree begins.  After any necessary EPA review and approval, the
company is expected to take the required actions that should result in reduced
emissions and improved environmental conditions.

OECA's plan for monitoring consent decree implementation called for EPA's
written responses to the hundreds of refinery  documents and reports required by
the consent decrees. Table 2.3 shows the types of company reports required by
consent decrees and identifies those reports requiring EPA responses.

               Table 2.3: Examples of Consent Decree-required
               Company Reports and Required EPA Responses
'fttaMrAtwhf
NSR/PSD

Flaring/ NSPS
Flaring/ NS PS
and any other
violations
Benzene

LDAR/
Benzene
Alt
- "- x - -"
Tttwi 
-------
designed to reduce emissions and compromised company compliance with
consent decrees. Figure 2.1 shows the timeliness of responses EPA issued and
reports provided by companies (these data were only available for July and
October 2003). Figure 2.2 shows the number of required EPA responses and the
number of responses EPA issued for July 2003, October 2003, and January 2004.
  Figure 2.1: Timeliness of EPA Responses and Company Reports
   Mandated by Consent Decrees for July 2003 and October 2003
                                  I Responses Issued
                                         ___Late
                              Q Average Days
                              13 Company Days late
                  July
            October
   Figure 2.2: Number of EPA Responses Issued Compared with
    Number of EPA Responses Not Issued per Consent Decree
    Requirements in July 2003, October 2003, and January 2004
                                         sponses Not Issued
                                 ^ Total Respor
              July
October
January
                             20

-------
We reviewed information from nine consent decrees well into implementation to
determine the extent and potential impacts of EPA response delays. We reviewed
information as of July 2003, October 2003, and January 2004. On average, the
consent decrees in our analysis required companies to submit a total of 2,191
reports (or 243 reports per company) requiring 421 EPA responses (or
approximately 47 responses per company) over the 8-year terms of their consent
decrees.

On average, late EPA responses delayed consent decree implementation by nearly
8 months between July 2003 and January 2004. In some cases, companies
proceeded with consent decree actions while awaiting EPA approval. In other
cases, companies awaited EPA responses before taking actions. OECA
management said they kept abreast of implementation activities through regular
conference calls with consent decree implementers.

Figure 2.3 demonstrates that most late and absent EPA responses related to
flaring/NSPS incidents (60 percent) and benzene handling or LDAR requirements
(29 percent).
      Figure 2.3: Required EPA Responses by Priority Area


                        Other NSR and PSD
           Benzene and     4%      7%
              LDAR
               29%
                                  Flaring and NSPS
                                        60%
To date, consent decree implementation is not complete and we did not discover
any instances where response delays affected emissions reductions during the
course of our evaluation.  However, regional staff, State staff, and refinery
company staff told us that EPA response delays might affect some company
actions under consent decrees.  For example:

      Acid Gas Flaring: In one consent decree, as of January 2004, EPA
      responded to 22 of 23 acid gas flaring reports an average of 273 days
      (9 months) late. OECA staff acknowledged that problems could arise if
      EPA did not respond to a company's acid gas flaring root cause analysis
      report. If a company did not correctly identify a root cause and EPA did

                            21

-------
      not respond in a timely fashion, companies might take inadequate
      corrective actions, and unnecessary emissions may have resulted.

•     NOX Control: One EPA consent decree press release stated that the decree
      would reduce the company's NOX and SO2 emissions by 50,000 tons each
      year. As of January 2004, EPA issued its response to 5 of 13 NOX control
      plans. For the three responses EPA issued that listed a company report
      date, EPA waited approximately 174 days (about 6 months) between
      receiving the company report and issuing a response. As of January 14,
      2004 (the date we last received the data), EPA had not approved the
      company's eight other plans. As a result, EPA had waited at least 478
      days (approximately 16 months) between receiving the company reports
      and responding to them. These delays could have resulted in continued
      emissions of tons of NOX to the surrounding community.

Late and absent EPA responses may have compromised company compliance
with consent decrees. EPA company leads said that, in most cases, companies
continued along their implementation schedules even if OECA did not issue
required responses to reports. In other cases, companies may have awaited EPA
responses before taking consent decree actions. Delays in EPA responses caused
one company to express concerns about remaining in compliance.  The company
questioned its compliance status in a July 31,2003, quarterly report to EPA
stating: "It is critical that [the company] receive responses from the EPA ... to
ensure compliance [with] ... the consent decree ... [The company] is still
awaiting EPA's comments."

Tracking Problems Developed and Persisted

Delays developed as soon as the implementation phase began and companies
began submitting reports to EPA.  EPA responded late to refinery documents and
reports 95  percent of the time as of July 2003, and 98 percent as of October 2003.
Because OECA eliminated tracking for due dates between October 2003 and
January 2004, we could not determine timeliness for company reports or EPA
responses after October 2003.

OECA quickly  found that it did not have the resources to efficiently address all of
the company reports arriving at EPA headquarters, particularly since several
consent decrees had similar implementation schedules, hi the face of the growing
document delays, OECA introduced an implementation plan and hired a
contractor to help with implementation tracking in January 2002. The
implementation plan outlined responsibilities for EPA headquarters, regional
staff, national technical leads, company leads, and external contractors, as shown
in Table 2.4.
                            22

-------
   Table 2,4: Key Roles and Responsibilities for Consent Decree Implementation
^;" Aster -
EPA Company
lead (1 IMd per
consent decree)
EPA National
Expert
(4 National
Experts)
EPA Regional
Staff (number of
staff varies by
region)
Contractor
OECA
ROl8/RB$poft5lbiilti*s
Each consent decree has a lead staff member who ensures timely review
of deliverables and implementation of all consent decree requirements.
Each priority area covered by consent decree has a national expert (also
known as national technical lead) who ensures national consistency across
all refinery consent decrees. EPA national experts also provide support to
regional staff on specific priority areas.
Regional staff responsible for consent decree implementation review all
company submissions directly involving their priority area and/or refinery,
recommend any necessary action, and coordinate with affected States as
appropriate.
Tracking contractor tracks consent decree deliverables and EPA's
responses to deliverables. A separate analysis contractor provides support
in evaluating control technology performance and setting appropriate
emissions limits under NSR/PSD regulations.
OECA's Air Enforcement Division makes determinations, issues approvals,
transmits comments, establishes emissions limits, and assesses stipulated
penalties based on consultations with appropriate company leads, national
experts, and regional staff.
The contractor developed an archive, a list of required deliverables, and a
document tracking system in close coordination with the EPA company leads by
late 2002. Meanwhile, the delays had grown as national technical leads attempted
to approve each individual EPA response, and OECA reported that its success
with obtaining settlements in the refinery sector had created a significant resource
drain. Company leads in EPA regions developed and maintained their own,
personal tracking systems and did not use the contractor's system to track
implementation.

OECA Took Steps to Address Problems But Did Not Eliminate Them

During our evaluation, OECA recognized that the document tracking delays
caused implementation problems and took steps that reduced but did not eliminate
the delays. OECA and regional EPA staff cited the requirement that four national
technical leads review each response and the Air Enforcement Division Director's
required approval as the reasons for delays.

OECA attempted to address the delays by training implementers, reallocating
staff, and making changes to consent decree design and implementation
requirements. For example, in a more recent consent decree, EPA required just
5 company reports and no EPA responses, compared with an average of 243
company reports and 47 EPA responses for the 9 consent decrees in our analysis.
                            23

-------
             OECA managers told us they disagreed with the contractor's interpretation of
             which items in consent decrees required an EPA response; OECA managers
             determined that the contractor-developed system included hundreds of items that
             did not require tracking.  Between July and October 2003, OECA eliminated
             requirements for 149 EPA responses, and eliminated tracking of the timeliness of
             both company reports and EPA responses from the tracking system. On the
             January 7,2004, company lead conference call, OECA management and company
             leads agreed to use the contractor's system as the principal tool for managing and
             tracking consent decree implementation. Because OECA eliminated tracking of
             timeliness, we could not determine the impact the changes had on EPA response
             time. Also, OECA could not use the tracking system to determine the timeliness
             of company reports or EPA responses.

             OECA did not accurately assess the resources required for consent decree
             implementation. EPA planning guidance stresses assessing the skills and number
             of personnel needed to implement programs.  More accurate planning that
             accounted for the specific monitoring and management requirements of each
             signed consent decree could have enabled OECA to avoid the document tracking
             delays, delays in EPA responses, and confusion about compliance.  OECA
             management and national technical leads familiar with the four priority areas
             when they negotiated consent decrees should have more accurately assessed the
             resources required to monitor and implement the steps outlined in consent
             decrees. Identifying resource needs during the planning phase would have
             highlighted resource limitations and EPA could have corrected problems before
             they caused delays in implementing the program.
Conclusion
             OECA did not clearly and precisely define official program goals and measures,
             or ensure the goals were consistently shared and clearly understood. Because
             refineries emit toxic chemicals, they should be closely and regularly monitored for
             compliance.  Although OECA officials used informal methods to track program
             progress, OECA would benefit from using more formal mechanisms to track and
             measure progress toward consent decree and overall refinery program goals.
             Implementation problems developed because OECA officials did not establish an
             accurate, detailed resource plan. Serious delays developed and persisted because
             OECA did not provide sufficient guidance to the contractor charged with
             developing the implementation tracking  system, and OECA did not reallocate its
             own resources to provide for implementation tracking. The ultimate success of
             the refinery program depends on effective management of consent decree
             implementation.
                                         24

-------
Recommendations
             To correct issues related to goals, performance measures, and reporting, we
             recommend that the Acting Assistant Administrator for Enforcement and
             Compliance Assurance:

             2-1.   Develop clear overall refinery program goals that allow for future
                   assessment or measurement and include timetables for accomplishment.

             2-2.   Instruct OECA refinery program managers to develop clear goals
                   specifically for the refinery program's implementation phase.

             2-3.   Ensure that all goals and performance measures are understood by
                   everyone involved in the national petroleum refinery program, including
                   all EPA and State staff involved in some portion of consent decree
                   implementation, and hold staff accountable for progress in performance
                   agreements.

             2-4.   Instruct OECA refinery program managers to develop reliable
                   performance measures to assess their progress toward meeting national
                   program goals. Specifically, managers should fully implement OECA's
                   performance-based approach to program management as described in its
                   December 18, 2002, Recommendations for Improving OECA Planning,
                   Priority Setting and Performance Measurement, which specifies
                   development of plans and reliable performance measures, to the remaining
                   phases of the petroleum refinery program.

             2-5.   Instruct OECA refinery program managers to gather,  analyze, and report
                   relevant program data to support overall OECA organizational decision
                   making and daily program decision making.

             To ensure Hie accurate measurement and reporting of refinery program outcomes,
             we recommend that the Acting Assistant Administrator for Enforcement and
             Compliance Assurance:

             2-6.   Instruct OECA managers to verify emissions reductions predicted in
                   consent decrees on a quarterly basis.  Verification might include
                   establishing a detailed monitoring system, which could contribute to
                   refinery program performance measurement

             To improve refinery consent decree implementation, we recommend that the
             Acting Assistant Administrator for Enforcement and Compliance Assurance:

             2-7.   Instruct its consent decree tracking contractor to resume tracking both
                   company due dates for reports and EPA response due dates so that OECA
                   and outside parties can easily track company and EPA responsiveness.

                                         25

-------
             2-8.    Revise and circulate a comprehensive tracking plan and system that
                    outlines specific roles and responsibilities for OECA staff, EPA regions,
                    State and local air pollution control agencies, and companies.

             2-9.    Provide additional training at the regional level, and hold regional experts
                    accountable for reviewing and responding to company reports. Require
                    national technical leads to spot-check responses from regional experts to
                    ensure national consistency.

             2-10.   Develop a formal feedback system to ensure that OECA's workforce and
                    managers have a common understanding of implementation
                    responsibilities, a common perspective on the status of implementation,
                    and the ability to expeditiously address implementation issues.

             To avoid similar issues in other enforcement and compliance initiatives, we
             recommend that the Acting Assistant Administrator for Enforcement and
             Compliance Assurance:

             2-11.   Ensure frequent and open communication between partners (States,
                    regions) and headquarters about responsibilities for executing portions of
                    strategies to quickly eliminate misconceptions or confusion.

             2-12.   As discussed with OECA managers, include consent decree
                    implementation in OECA priorities and strategic plans, allocating staff and
                    resources to implementation until OECA completely implements all
                    consent decrees.

             2-13.   Develop a plan for allocating negotiation and implementation resources.
                    Use resource planning in new initiatives to determine the predicted
                    workload associated with the initiative; allocate training, education, and
                    development resources; and provide for office-wide revaluation of the
                    resource plan.

Agency Comments and OIG Evaluation

             OECA agreed with 9 of our 13  recommendations above, and disagreed with 4.

             OECA disagreed with recommendation 2-6 because OECA officials believed mat
             quarterly monitoring was too frequent and did not take resource  limitations into
             account. OECA stated that these resources are better utilized if devoted to
             addressing compliance issues in other industry sectors, and that verifying
             emissions reductions twice over the life of a consent decree (once in the fourth
             year, and once at the conclusion) would provide EPA with sufficient  assurance
             that consent decrees led to environmental results. Because we do not believe that
             two monitoring events over the life of a consent decree provide sufficient
             assurance of results, we continue to recommend that OECA track available data
                                         26


-------
on its singular outcome measure for the program - change in tons of air emissions
- on a quarterly basis.

OECA disagreed with recommendation 2-7 because OECA does not see the
necessity of further revising the tracking system at this time. OECA said it has
made substantial progress in reducing the delays, indicating that the current
approach is having the desired result. We believe that the recommendation to
track company report due dates and EPA response due dates in addition to
company report receipt dates and EPA response issuance dates is crucial to
maintaining accountability for both companies and EPA in consent decree
implementation. To  avoid similar delays in the future, OECA should track
companies' and EPA's timeliness and identify and address problems as they arise.

OECA disagreed with recommendation 2-8 because OECA does not see the
necessity of creating a new tracking plan and system when EPA already has a
comprehensive consent decree tracking protocol implemented through its
contractor.  Despite OECA's current tracking plan and system, we found that
some EPA and State  staff were still unclear about their roles and responsibilities.
We recommend that  OECA revise as appropriate and circulate its existing
tracking plan and system that outline roles and responsibilities to all appropriate
stakeholders.

OECA disagreed with recommendation 2-10 because, in light of the small staffing
level and their overlapping responsibilities, they did not believe a formal feedback
system was necessary.  Should the number of staff substantially increase in the
future, they said they would consider a feedback system at that time.  Despite
OECA's monthly conference calls, our evaluation demonstrated management's
lack of awareness of some issues we raised in our evaluation.  We recommend
that OECA develop a formal feedback system so mat all staff and managers
working on the program can rely on a common system for making suggestions and
raising issues separate from their day-to-day interactions.

Our official draft report contained discussion about using a logic model to tie
program activities to outcomes, and a recommendation to use the logic model
developed in the course of the evaluation for current and future program
development. OECA disagreed with the draft report recommendation and
specifically stated that several of the short-term, intermediate,  and long-term
outcome measures in our logic model were not appropriate benchmarks for
judging the effectiveness of a compliance and enforcement program.  We chose to
delete this recommendation because a logic model is only one of several possible
means that OECA may employ to achieve the ends we advocate in
recommendations 2-1, 2-2 and 2-3 - that is, the agreement on and communication
of program goals.
                            27

-------
28

-------
                               Chapter 3
              Refinery Program  Lessons teamed
            EPA's refinery program received mixed reviews from States, industry, and
            environmental groups, ranging from positive to very negative.  For example, two
            companies with consent decrees viewed their relationship with OECA as more
            collaborative, but a major industry professional association believed EPA's
            refinery program severely damaged the Agency's relationship with the industry.
            Most of the stakeholders are taking a wait-and-see approach, believing the
            program's success depends on how well EPA maintains a level playing field
            within industry, and how well companies implement consent decrees.
            Representatives from environmental groups added that, in addition to effective
            implementation of consent decrees, they wanted EPA to make program results
            available to the public, particularly to those communities directly impacted by
            refineries.

            EPA learned several important lessons that it should apply throughout its refinery
            program as well as consider for other enforcement and compliance assurance
            programs. Overall, EPA has effectively used some tools for the refinery program,
            but needs to make improvements in other areas.  For example, EPA effectively
            focused on specific enforcement concerns and became knowledgeable about the
            industry, but EPA needs to clearly communicate roles and responsibilities. In
            addition, EPA needs to meaningfully engage stakeholders throughout the process,
            and diligently oversee consent decree compliance.

Refinery Program Offered Lessons Learned

            OECA, regional, State, and industry officials described lessons they learned from
            the refinery program that OECA could apply to other industries.
              Lesson
              Learned
Become Knowledgeable about the Industry and Its
Technical Processes
             Each industry has its own characteristics that impact the effectiveness of
             enforcement and compliance assurance programs.  OECA learned an important
             lesson through obtaining an understanding of petroleum refinery processes.
             According to one former senior OECA official, the industry did not generally
             view EPA as credible or very knowledgeable, and  believed that EPA staff could
             be easily misled or overwhelmed by technical details.

             A significant tactic of the refinery program involved pulling together EPA staff
             with knowledge about the refinery industry.  These staff attended training on the
             refinery process, reviewed trade journals, and met  with industry officials to learn

                                        29

-------
about refineries. In addition, OECA identified four national technical leads for
the priority areas and ensured national consistency in investigations and
negotiations. EPA regional and OECA staff emphasized that technical expertise
made them credible to the industry and helped immensely during negotiations.
Representatives from companies that signed consent decrees with OECA also said
EPA's industry knowledge increased the program's effectiveness.  OECA officials
stated that the marshaling of in-house, cross-regional expertise to investigate
refineries was one of the more remarkable elements of the refinery program.
 Lesson
 Learned
Bvitd th& Program wiifi Regional aod Headquarters Staff
Working Together
The refinery program demonstrated that a single program with experts from
headquarters and regional offices working together operated more effectively than
having each region and headquarters working alone. EPA regional officials
believed the refinery program's strategy to pull together staff with expertise to
lead the program regardless of geographic location served as a model for other
enforcement programs. One regional EPA official believed the refinery program's
approach prevented the 10 regions from conducting their own programs or having
OECA dictate the entire program. While OECA coordinated the effort, officials
said regional as well as OECA staff drove the program and contributed to its
success in identifying and addressing priority areas.
 Lesson
 Learned
Representatives from one company that signed a consent decree stated that EPA
obtained success with settlements because the refinery program had a senior EPA
executive who championed the program. Industry representatives said that having
a senior OECA executive who had specific knowledge about the issues, had
decision-making authority, talked with them about the program, and even
participated in negotiations, made a positive impact in how they reacted.
However, the senior OECA official who had championed the program from its
inception left EPA in February 2002. Some industry and EPA staff believed that
the departure of this former OECA executive slowed the program's progress in
obtaining additional settlements. One OECA official believed the slow progress
related more to difficult and contentious negotiations rather than the departure of
the prior senior executive. OECA replaced the prior official with another
executive, but limited resources and other priorities prevented the new executive
from taking a similar role with the refinery program. OECA officials stated that
resources often limited the extent to which a senior executive could become
actively involved in the day-to-day activities of a single national enforcement
program.
                            30

-------
 Lesson
 Learned
As described in Chapters 1 and 2, OECA engaged various stakeholders
throughout the program, including the U.S. Department of Justice and the States.
Stakeholders' were involved in settlement negotiations and negotiating final
consent decrees. However, some State officials said they wanted more
participation in the process. For example, officials from one State believed they
needed to aggressively try to participate in the negotiations or discussions would
take place without them.  Officials from another State complained about the lack
of adequate time to comment on proposed negotiation decisions. The officials
said lack of adequate time gave the appearance that their reviews were irrelevant
and would have no impact on the final EPA decision. Other State and local
officials commented that OECA did not:

•   Keep the State informed on the status of negotiations.
•   Keep the State involved during the negotiations or decision making process.
•   Adequately describe the benefits of participation to local officials.
•   Adequately describe the increased workload necessary to successfully
    implement and monitor consent decrees.

On the other hand, one OECA official said that many States did not want to
participate in the refinery program despite OECA's efforts to include them in
negotiations.

One group of stakeholders - the companies that signed consent decrees with
OECA - formed a "consenters" group to facilitate a relationship with OECA and
minimize the uncertainty around consent decree implementation. The group also
formed to facilitate compliance, learn from each other (when appropriate and
within anti-trust limitations), and reduce learning curves in implementing consent
decree provisions. OECA officials and a representative from a company in the
consenters' group said the group provided an effective means to communicate and
discuss issues among similarly situated companies.
 Lesson
 Learned
Clear communication is a critical component of any program. OECA's
Framework for a Problem-Based Approach to Integrated Strategies, dated
November 2002, emphasized the importance of communication and clearly
defining stakeholders' roles and responsibilities. While OECA effectively
communicated its four priority areas to the industry, the public, and other
stakeholders through various means (e.g., Enforcement Alert newsletters), OECA
did not effectively communicate other aspects of its program (particularly
environmental outcomes) as discussed in Chapter 2. OECA's Framework
document encourages the development of a communications plan for internal and

                             31

-------
             external stakeholders to ensure that OECA keeps stakeholders informed of the
             program's progress as well as roles and responsibilities.
              Lesson
              Learned
         tfre &x(fte$vttantt£$ti)ktt$b Incentives far Industry to
Participate                       /,     \
             OECA and regional officials believed their focus on emissions reductions rather
             than individual facility violations led to the refinery program's success. EPA's
             refinery program strategy included identifying and focusing on four priority areas
             under the Clean Air Act that represented the most significant compliance
             problems within the industry. OECA directed its pursuit of global settlements at
             obtaining industry's agreement on implementing controls and practices that are
             expected to achieve significant emissions reductions in these four priority areas.
             As part of its push to achieve significant emissions reductions, EPA encouraged
             and, in some instances, required refineries to develop and install new emission
             control technologies.  OECA officials stated that the refinery program pushed the
             use of the most advanced emissions control technologies available.

             OECA negotiated with refinery companies and offered an incentive in the form of
             relief from past liabilities in order to persuade the industry to sign consent
             decrees.  OECA worked with the industry's desire to obtain a level of certainty
             regarding regulatory risks with EPA's desire to significantly reduce emissions.
             The industry saw many complex regulations on the horizon and viewed
             participating in consent decrees as "good business" to limit liability and obtain a
             level of certainty regarding regulatory risks. As a result, EPA obtained consent
             decrees that included implementing controls expected to reduce emissions as well
             as requirements to go beyond compliance with regulations.
              Lesson
              Learned
Diligently Oversee Compliance With Negotiated Settlements and
Consent Decrees and Take Action When Provisions Are Not Met
             As discussed in Chapters 1 and 2, the implementation of refinery consent decrees
             represents one of the essential pieces of the refinery program. Without effective
             implementation, the refinery program may not result in anticipated emissions
             reductions and increased industry compliance. OECA officials expressed concern
             over the significant amount of resources required by headquarters, regional, and
             State offices to implement consent decrees while at the same time continuing
             other enforcement-related activities in other industries. While it will require
             careful priority-setting and resource allocation, OECA, regions, and States should
             ensure implementation of consent decree provisions and take appropriate
             enforcement action when necessary.
Conclusions
             OECA's refinery program resulted in several important lessons learned.  OECA
             and regional officials spent considerable time at the start of the program learning
                                          32

-------
             about the industry and developing expertise in the four areas on which the strategy
             focused.  Officials continued to gain industry knowledge throughout the refinery
             program. As an on-going program, however, the refinery program could continue
             to benefit from evaluating where it might apply some of the lessons learned. For
             example, OECA should improve communication with stakeholders, particularly
             given the feedback States and regional staff provided on the refinery program. In
             addition, OECA should designate a senior EPA executive to champion the
             consent decree implementation phase.

Recommendations

             We recommend that the Acting Assistant Administrator for Enforcement and
             Compliance Assurance:

             3-1.  Disseminate the lessons learned from the refinery program to EPA staff to
                  benefit other compliance efforts,  obtain additional feedback from
                  stakeholders - including  States, industry, and environmental groups - on
                  other lessons learned, and update relevant OECA guidance documents such
                  as OECA's Framework for a Problem-Based Approach to Integrated
                  Strategies or other appropriate documents for on-going and future industry-
                  specific enforcement programs.

             3-2.  Designate a senior OECA executive to assume the role of champion for the
                  refinery program to ensure (a) that refiners enter into consent decrees or face
                  appropriate alternative enforcement actions, and (b) that consent decrees are
                  effectively implemented.

             3-3.  Consider, on a case-by-case basis, designating a senior OECA executive to
                  assume the role of champion for each of the other enforcement priority
                  areas.  EPA and industry  officials should recognize the champion as
                  knowledgeable and as having the authority to make decisions related to the
                  priority area,

             3-4.  Develop a communications plan for refinery consent decree implementation.
                  The plan should clearly describe the roles and responsibilities of all
                  stakeholders, including refinery priority area experts and regional and State
                  officials.

Agency Comments and OIG Evaluation

             OECA  agreed in part with the first three recommendations in this chapter and
             fully agreed with the last recommendation.

             OECA  concurred in part with recommendation 3-1.  OECA disagreed that it
             needed  to revise fas Framework for a Problem-Based Approach to Integrated
             Strategies to reflect lessons learned from the refinery program. We view the

                                        33

-------
lessons learned from the refinery program as significant and believe OECA should
reflect them in a written document (either combined with lessons learned from
other initiatives or by itself) and distributed to stakeholders. We revised the
recommendation so that OECA can determine the appropriate document - either
the Framework document or another guidance document - to capture the lessons
learned and communicate them to stakeholders.

OECA also concurred in part with recommendation 3-2.  OECA agreed that a
senior enforcement official needed to manage national enforcement priorities and
OECA designated a senior official responsible for managing the refinery program.
OECA disagreed, however, that a senior OECA official needed to ensure lhat
EPA settle or litigate against all refiners in the industry. OECA officials stated
that after the refinery program reaches its goal of 50 percent increased compliance
and 20 percent decreased emissions, the program would return to the "core"
enforcement program, with primary enforcement and compliance assurance
program responsibility devolving to States and EPA regional offices. We
continue to recommend that while the refinery program remains as a national
priority, a senior OECA official maintain responsibility and accountability for the
program.

For recommendation 3-3, OECA stated that it would make the determination of
whether any particular priority area requires an OECA-designated "champion" on
a case-by-case basis as appropriate in light of all relevant facts and circumstances.
We agree, and revised this recommendation accordingly.
                             34

-------
                                                                       Appendix A

               Details on Scope and Methodology

We conducted our evaluation of EPA's national petroleum refinery program between June 2003
and March 2004.  We performed our evaluation in accordance with Government Auditing
Standards issued by the Comptroller General of the United States.

We began our evaluation of the petroleum refinery program at the request of OECA senior
managers and after we conducted extensive preliminary research on EPA's enforcement and
compliance assurance program. We began our general preliminary research on EPA's
enforcement and compliance assurance program in February 2002 to obtain information on
EPA's traditional and alternative approaches to enforcement and compliance assistance. Our
preliminary research included interviewing officials in EPA's OECA, Office of Research and
Development, Office of Environmental Information, and Office of the Chief Financial Officer.
We interviewed these officials to obtain information on EPA activities related to enforcement
and compliance assistance strategies.  We reviewed a variety of documentation describing
integrated strategies, compliance assistance grants, Memorandum of Agreement priorities,
Government Performance and Results Act reporting, and enforcement and compliance assistance
databases (Enforcement and Compliance History Online, Sector Facility Indexing Project, and
Integrated Data for Enforcement Analysis System).

We also interviewed officials at non-EPA organizations.  We met with officials at the General
Accounting Office (GAO), the National Academy of Public Administration, and the
Environmental  Council of the States to obtain information on prior reviews they performed on
enforcement and compliance assistance programs, and to obtain their perspectives on significant
concerns and issues related to enforcement and compliance assistance. In addition, we reviewed
reports they issued on enforcement and compliance  assistance issues.

We used the results of our initial preliminary research to develop objectives for four separate
evaluations of the enforcement and compliance assurance program. These four evaluations
became part of OIG's March 2003 Multi-Year Plan Fiscal 2003 - 2005.

In December 2002, we had held an entrance conference with OECA to begin one of the four
evaluations in the Multi-Year Plan. We designed the first evaluation to describe the regulated
universe, the compliance status of OECA's priorities, and the enforcement and compliance
assurance strategies EPA and its partners applied to each of the priorities. However,  OECA
senior managers said they found the scope of the evaluation too broad and thought it would not
provide meaningful results. After discussions with OECA senior management, we agreed to first
conduct a pilot  evaluation of one of the priorities. The pilot evaluation would review the
strategies and measures for one priority area and test our approach to answering the objectives in
all four evaluations.

From February  through May 2003, we performed additional research on OECA's priorities to
determine which priority to include in the pilot evaluation and to develop the specific evaluation
objectives. We met with OECA staff to determine the status of each priority and to obtain a brief

                                         35

-------
background on the issues prompting OECA to identify them as priorities. As a result of
additional research and analysis, we selected the petroleum refinery program as our pilot issue.
In addition to its status as an OECA national enforcement priority, OECA also had listed the
petroleum refinery as a separate priority industry sector for several years.

Nature and Extent of Petroleum Refinery Industry and Strategies Used

To understand the nature and extent of the petroleum refinery universe and what strategies EPA
and its partners developed to address compliance at refineries, we interviewed and collected
documents from OECA staff, EPA's National Enforcement Investigations Center, regions,
States, industry., environmental groups, and the U.S. Department of Justice.

We interviewed OECA, National Enforcement Investigations Center, and regional staff to
determine the compliance problems in the industry, the specific strategies and approaches used to
address the problems, and the industry's current compliance status. We interviewed officials in
EPA regions 5,6, 8, and 9 and State officials in Louisiana, Oklahoma, New Mexico, Minnesota,
Delaware, Texas, Arkansas, and Colorado. We also interviewed officials in the Northwest Air
Pollution Authority. We asked the regions, States, and local air authority to provide their
perspectives on the effectiveness of EPA's strategy to address refinery  compliance problems,
their level of participation in the strategy, and suggestions for improving the refinery strategy and
other enforcement and compliance strategies. We based our selection of regions on whether they
had a significant number of refineries within Iheir region.  We based our selection of States on
whether the State participated in a consent decree as part of EPA's strategy, the number of
refineries in the State, and recommendations from EPA officials and outside organizations.

We met with two petroleum refinery industry groups - the American Petroleum Institute and the
National Petrochemical  and Refiners Association - and asked for their perspectives on the
effectiveness of EPA's refinery strategy and their members' view of the strategy. In addition, we
spoke with representatives from two companies and a representative from the consenters group7
to obtain reasons why refineries did or did not enter into settlement agreements with EPA. We
chose not to interview officials of refineries that had not entered into agreements with EPA
because we did not want to interfere with then- ongoing negotiations with OECA.

We reviewed documentation on the petroleum refinery industry to better understand the industry,
its compliance issues, and impact on the environment. These documents included the following:

•  EPA Office of Compliance Sector Notebook Proj ect - Profile of the Petroleum Refining
   Industry, September 1995
•  Sector Facility Indexing Project
•  Scientific literature on the petroleum refining industry
             7 The consenters group consists of representatives from petroleum refinery companies that already
             signed consent decrees with EPA.

                                          36

-------
We also reviewed and analyzed documentation on EPA's refinery strategy, justification for its
approach and priority, and compliance assistance efforts. These documents included:

•  Memorandum of Agreement Guidance for fiscals 1996 through 2004
•  Enforcement Alert Newsletters
•  Internal documents describing the refinery strategy, goals, and performance measures

We attended the Refining Process Services Inc. 's Basics of Petroleum. Refining for
Non-Technical Personnel training to help us better understand the petroleum refining process.
We also toured three refineries - in Denver, Colorado; Whiting, Indiana; and Texas City, Texas -
to observe refineries with various capacities and emissions issues. We also met with officials at
each of these refineries to obtain feedback on how they viewed EPA's refinery strategy and the
consent decree process..

Goals, Performance Measures, and Reporting Approach for Petroleum Refineries

To evaluate the performance measurement and reporting approach for petroleum refineries and
whether the approach allowed EPA to effectively implement, manage, and improve its strategies
for refineries, we interviewed a variety of individuals and analyzed supporting documentation.
These individuals helped plan, implement, track, and/or manage EPA's strategies for the
petroleum refinery industry. These individuals included staff in OECA's Office of Regulatory
Enforcement, Office of Compliance, and OECA refinery issue experts in Headquarters, Region 5,
and the National Enforcement Investigations Center. We interviewed these individuals to obtain
information on OECA's  process to identify priorities, establish goals and performance measures,
and report accomplishments. We also interviewed them to determine whether they tracked their
progress in achieving the refinery goals and how they determined baselines for pollutant
emissions and reductions. In addition, we interviewed the prior Director of the Office of
Regulatory Enforcement to determine the refinery program's initial goals and objectives: We
also interviewed outside EPA experts to discuss environmental performance measurement.

We reviewed documentation on-the  refinery strategy goals, performance measures, and
accomplishment reports. We evaluated how EPA established baselines for its measures and how
it measured emissions reductions. • Our documentation review included the following:

•  Case Conclusion Data Sheet - Training Booklet, November 2000
•  OECA's Framework for a Problem-Based Approach to Integrated Strategies
•   11 Global Consent Decrees
*  Headquarters and Region 6 Consent Decree Implementation Tracking Databases
•  OECA's Accomplishment Reports for fiscals 1996 - 2002

In our interviews with OECA and regional officials, we obtained information on whether they
agreed on the refinery strategy goals and objectives and whether OECA met those goals. In
addition, we obtained information from States and local authority officials on other measures that
OECA could or should use to determine compliance and achievement of refinery strategy goals.
                                         37

-------
To determine how well OECA managed its consent decree implementation, we analyzed
OECA's consent decree tracking information. We retrieved nine company spreadsheets from
EPA's contractor on July 31,2003, October 16, 2003, and January 14, 2004.  The company
spreadsheets depicted the due dates and deliverables outlined in the consent decrees. From the
spreadsheets, we calculated the number of responses EPA had not issued and,the timeliness of
both company reports and EPA responses. By comparing data from July, October, and January,
we determined that company spreadsheet contents changed between the reporting dates;
therefore, we assessed the nature of changes that occurred, including comparing the number of
company reports and EPA responses required, the types of reports and EPA responses required,
and the fields tracked by the spreadsheets (for example, EPA eliminated using tables that tracked
the timeliness of company reports and EPA responses between October 2003 and January 2004).
                                                                        V
To obtain and document lessons learned that EPA could apply in other industries or that OECA
could make to improve the petroleum refining program, we interviewed staff at OECA, EPA's
National Enforcement Investigations Center, regions, States, industry, environmental groups, and
the U.S. Department of Justice.

Prior Audit and Evaluation Work

In our research, we found no previous audit or evaluation reports evaluating EPA's petroleum
refinery compliance program.  However, we identified EPA OIG and GAO reports listed below
with findings on performance measurement, monitoring, and tracking:

       EPA OIG: Compliance with Enforcement Instruments, Audit Report No.
       2001-P-00006, March 29,2001.

      We found that OECA's performance measures were not sufficient to determine the
      program's actual accomplishments. Consequently, we determined Congress had less
      useful performance data upon which to base its decision making. We also found that
       EPA regions did not always adequately monitor compliance with enforcement
      instruments (e.g., consent decrees) nor did they always consider further enforcement
       actions. We attributed ineffective monitoring primarily to the lack of: (1) guidance
       detailing how or when to  monitor enforcement instruments, and (2) emphasis OECA
       placed on monitoring.  Consequently,  OECA risked continued violations that would
       contribute to human and environmental health impacts, thus decreasing EPA's deterrence
       effect. In response, OECA concurred  that it and the regions can and should improve
      tracking and enforcing compliance with requirements in enforcement instruments.  At
      that time, we concluded that OECA had begun to take the steps necessary for OIG to
       close out the report.

       GAO: Environmental Protection: Wider Use of Advanced Technologies Can Improve
      Emissions Monitoring, Report No. GAO-01-313, June 22,2001.

       GAO found considerable variation in the compliance monitoring performed by stationary
       air pollution sources.  As  a result, regulators and regulated entities lacked certainty about
       whether air pollution sources maintained continuous compliance with the Clean Air Act.
                                         38

-------
GAO recommended that EPA encourage wider use of advanced air monitoring
technologies.  GAO also cited equipment manufacturers and regulators as stating thai,
without regulatory requirements, manufacturers had little incentive to bring new
monitoring technologies to market.  OECA responded that it had coordinated efforts to
explore ways to increase the application of advanced monitoring technologies with EPA's
Office of Air and Radiation and industry while minimizing the perception that
cooperation with OECA would lead to greater enforcement actions. OECA guaranteed
that they would not punish the industry for their willingness to advance the science of
monitoring except in the most egregious cases.

GAO: Air Pollution: EPA Should Improve Oversight of Emissions Reporting by Large
Facilities, Report No. GAO-01-46, April 6,2001.

GAO reported that EPA performed limited oversight of States' efforts to verify large
facilities' emissions reports.  GAO recommended that EPA improve its oversight of
States' review of emissions reports by evaluating the adequacy of these reviews and, if
necessary, strengthening them. In response, EPA stated that it offered tools and
encouragement for States to improve facility emissions estimates, but EPA had no basis
to impose a requirement on State or local permitting authorities to quality assure annual
emissions data.
                                   39

-------
40

-------
                                                                             Appendix B


        Key Information on U.S.  Petroleum Refineries
                                                                                         /
The refinery corporations are listed in order of refining capacity as of January 1,2003. The  .
corporations highlighted have signed consent decrees as of March 2004 that apply to at least
some of their refineries. The consent decrees covered each refinery the company owned and
operated on the date of consent decree entry by the court.  Since that date, companies may have
sold, bought, or closed one or more refineries, however, the consent decree follows a refinery
regardless of these actions. As a result, the "number of refineries under consent decrees" may
not align with January 1, 2003 data for "number of refineries" owned by a company as depicted
in the table.
 Exxon Mobil Corp
 jjfftd&frsi;^
 Valero Energy Corp
 Sunoco Inc.
 PDV America Inc.
 (includes Citgo Refining)
 Tesoro Petro Corp4
                      lisifiiK®
                      :•:•-:;•:•:•:•:•::•:•:•:y.:•:;; >-K:-;:-:-:-::"::-:':•::•:••::• :•:
                      .!!?TOffi^
                      :S:ip::;Ti-::;::Ma^
                                       :::|:i:;:^:;:^i;:^:^0:^^:^iii:?


                                       |l^nirj^|
6
12
                                         tmmm

6
1,808,000
1,247,362
                           730,000
                           698,300
 562,500
10.79
7.44
              4.36
              4.17
3.36
               Conoco purchased 12 of its 16 refineries after it entered into a consent decree with EPA.


               Two Chevron Texaco refineries not covered by the consent decree are asphalt refineries rather than
              petroleum refineries.


               Two Royal Dutch Shell GP refineries not covered by the consent decree are operated by Shell Chemical, •
              separate business unit from its fuels refineries.


               Tesoro purchased two refineries from BP, both of which are covered by BP's consent decree.


                                           '41

-------
                Corporation
                                 No. of
                               Refineries
No. of fceflnertes
 under Consent
    Decrees
  Capacity  .
 (barrels per
Calendar day)
 Percent
 of Total
Capacity
 Blackstone Group LP
                                                                  416,500
                                      2.49
 The Williams Co. (Premcor)5
                                                                  377,928
                                      2.26
ijffhiiyw^
 Lyondell Petrochemical Co.
                                                                  270.200
                                      1.61
 Chalmette Refining LLC
                                                                  182,500
                                      1.09
 TotalFinaElfSA
                                                                  175,068
                                      1.04
 Crown Central Petro Corp
                                                                  155,000
                                      0.92
 Orion Refining Corp"
                                                                  155,000
                                     0.92
 Sinclair Oil Corp
                                                                  152,195
                                      0.91
 Frontier Oil Corp
                                                                  149,000
                                     0.89

 Murphy Oil CorpT
                                                                  128,000
                                      0.76
 Farmland Industries Inc.'
                                                                  112,000
                                     0.67
                                                            •••-:-:'-  ---"-----;
 Giant Industries Inc.
                                                                   96,200
                                      0.57
 Calumet Lubricants Co. LP
                                                                   67,520
                                      0.40
                         i^ftjpC^panp
                                                              lltfSill
 United Refining
                                                                   65,000
                                      0.39
 Petro Star Inc.
                                                                   62,550
                                      0.37
 Alon USA Energy Inc
                                                                   58,500
                                      0.35
 Gary Williams Co.
                                                                   52,500
                                      0.31
 Paramount Acquisition Corp
                                                                   50,000
                                      0.30
 Placid Refining Co.
                                                                   48,500
                                      0.29
 Time Oil Co.
                                                                   44,350
                                      0.26
 Hunt Consolidated Inc.
                                                                   33,500
                                      0.20
 Transworid Oil USA Inc.
                                                                   29,400
                                      0.18
                                                                                                    non-
  EPA took enforcement action against Premcor under the refinery program and entered into a limited, n<
global settlement.  EPA reached similar limited, non-global settlements with Murphy, Farmland, Frontier,
Pennzoil, Crown, and NCRA (owned by Cenex Harvest States COOP).


  A consent decree with the State of Louisiana (patterned after EPA's settlements) covers Orion's refinery.


  EPA took an enforcement action against Murphy Oil for operations at one of its refineries that included
elements similar to the global refinery consent decrees.


  Farmland recently resolved its liability with EPA through a consent decree patterned after the global
refinery consent decrees.
                                                   42

-------
* ?
, , Corporation
Apex Oil Co. Inc.
Kern Oil & Refining Co.
San Joaquin Refining Co. Inc.
Flying J Inc.
Countrymark COOP Inc.
Southland Oil Corp
Silver Eagle Refining
Wyoming Refining Co.
- Age Refining & Marketing
American Refining Group Inc.
Greka Energy
World Oil Co.
Cross Oil & Refining Co. Inc.
Somerset Refinery Inc.
Young Refining Corp
Foreland Refining Corp
Oil Holding Inc.
Dow Chemical USA
litres : - - " ,-
No. of
Refineries
1
1
1
1
1
2
2
1
1
1
1
1
1
1
1
1
1
1
145
No. of Refineries
«utd«r Consent
Decrees ••


















42 :;
Capacity
- (barrels per
calendar day)
26,000
24.700
24,300
24,000
23,000
16,800
14,000
12.500
10,200
10,000
9,500
8,500
6,800
5,500
5,400
5,000
2,800
880
KjtSI,yf1t
Percent
of Total
Capacity
0.16
0.15
0.15
0.14
0.14
0.10
0.08
0.07
0.06
0.06
0.06
0.05
0.04
0.03
0.03
0.03
0.02
0.01
: too
43

-------
44

-------
                                                                               Appendix C
                  Petroleum Refining Process Flow Chart

      This flow chart illustrates the petroleum refinery process, potential releases, potential release
      points, and the major applicable environmental regulations.
MX
        . »ft ftwid aapoUedst**. dssctwlsdi*. Wtiahfjma awn *mnd (835,1*)!
        tnw«(4m^«eM<
-------
46

-------
                                                                     Appendix D
                  Refinery Releases and Effects
This table provides a summary of the human health and environmental effects of the following
common air pollutants released at refineries: volatile organic compounds (VOCs); sulfur dioxide
(SO2); nitrogen oxide (NOJ; paniculate matter (PM); carbon monoxide (CO); hydrogen sulfide
(H2S); and toxic air pollutants.  Toxic air pollutants include pollutants known or suspected to
cause cancer or other serious human health effects and include refinery releases such as benzene
and toluene. EPA does not consider S02; NOX, PM, CO, and H2S toxic air pollutants, but EPA
lists pollutants such as benzene as both VOCs and toxic air pollutants.
" Release* "
in
fi
1
3
X
111
1
1
Ul
Reacts with other chemicals to create PM that
can cause respiratory illness, aggravation of
heart conditions and asthma, permanent lung
damage, and premature death.
Aggravates respiratory conditions.
Reacts with other chemicals leading to
ground-level ozone and smog, which can
trigger respiratory problems.
Can cause health problems such as cancer.
Can cause reproductive, neurological,
developmental, respiratory, immune system,
and other health problems.
Reacts with common organic chemicals
forming toxins that may cause bio-mutations.
Affects cardiovascular system and can cause
problems within the central nervous system.
Causes visible impairments that may migrate
to sensitive areas such as National Parks.
Contributes to formation of acid rain, which
damages crops, trees, and buildings; and
increases acidity in soils, lakes, and streams.
Contributes to the formation of ground-level
ozone, which harms vegetation.
Contributes to global warming, which leads to
rising sea levels and other adverse changes
to plant and animal habitat.
Causes environmental hazards, including
concentration of toxic chemicals
(e.g., mercury) up the food chain.
Settles on ground and water, acidifying
streams and lakes, damaging forests and
farm crops, and depleting soil nutrients.
VOCs


/
/



/

/



S02
/






/
/



/
m*
/

/


/

/
/

/


m
/





/
/




/
CO







/

/



H,S

/




/

/




Toxic Air
Pollutants



/
/






/

                                       47

-------
48

-------
                                                                                     Appendix E
               National Refinery Program Time Line
        Date
1995
  ll
1997
1999
June
        September
Late 1997
        Late 1997
       iNbverriDar?
January
        February
        October
        November
                                                  Event
Petroleum refining identified as national enforcement priority in OECA's fiscal
1996/1997 Memorandum of Agreement Guidance. OECA required each Region to
develop compliance/enforcement strategies for the refinery enforcement priority.
               OECA issued the Profile of the Petroleum Refining Industry sector notebook.
OECA forms the national refinery workgroup with an initial conference structured to
discuss refinery issues and regional initiation of their own enforcement actions.  The
workgroup, comprised of OECA and regional staff, focused on investigations/
enforcement at petroleum refineries.
                OECA officials share environmental and enforcement concerns with the refinery
                industry, State/local officials, community groups, and environmental advocates. Based
                on publicly available information, OECA became aware that refinery capacity had
                increased substantially but relatively few refineries had applied for NSR permits.
                                  jfijei!»tirig:';p£tt^-|^
                       :The'i:>roup  e
                       ^XX.X"jiL^.::'i-.^:':':':..-;'-:r-or-'-'>'-civ':^i'-».:''_ »ii>::''»«»**V*l :/^\'ii»*i'iWii li^'^ilW^i^Wrl'^^XrXii^'^tiikrt •ft»Xii"»eAiHf':iri\»Ai«t'lrf'i *i 44Ai'4';*iii^'#il •":•:•.":•.• ''.•:•'•'•:
                                         suraglng and coordinating focused
                                         iritfalel&i:^
                       enforcement:efforts on friefsriority areas
                       !:|ega$jujj|sfc^


OECA issued an Enforcement Alert on complying with NSR and PSD.
                National meetings of EPA workgroup and senior enforcement managers to review
                investigation progress.
               OECA issued an Enforcement Alert on LDAR (fugitive emissions).
                EPA officials presented preliminary investigation results to the National Petrochemical
                and Refiners Association's annual environmental meeting.
                                                49

-------
  ,
2000
    ''
2001
2003
January 18
        March 21
        May 11
        Mid-2001
        December 20
March 11
        October 1
        October 16
                >.:> ;:::::•;:/; ;i^;i^^.V%.fcif :-f:;3,i::rts:^^
                Nattoriahmeefchg otsehiot^^

                        ;^uiniiary;Seerrienfe:ofyioiafip^

                        ;OEGAdscided;tQpurstfeglotet


                                       ;;ife^i^fne^;A
                                       rwo*:wlHi IMb^a^ i^
                                       S;"Agr^OTer^1n;|nnc^
                                       • hefdiwilH':
BP settlement filed with the court.
                Motiva, Equilon, and Shell Deer Park settlements filed with the court (25 percent of
                industry now under consent decrees; additional 25 percent in similar global settlement
                negotiations),
                Marathon Ashland Petroleum settlement filed with the court.
                Consenters Committee formed by refinery companies that had entered into global
                consent decrees with EPA.
                Conoco, Navajo Refining, and Montana Refining settlements each filed with the court
                (30 percent of industry now under consent decrees; additional 30 percent in similar
                global settlement negotiations).
                        ; Murphy;6i! Sett^
                        
-------
                                                             Appendix F
            Consent Decree Process Flow Chart
    Proposed
    Consent
   Decree (CD)
Company Report Re view and
Circulation Process f Implementation
Tracking System")
             Phase Begins Hers
developed by EPA DIG 02/2004
                                   51

-------
52

-------
                                                                     Appendix G

         EPA Comments on the Official Draft Report

                                            Apr. 2,2004
MEMORANDUM

SUBJECT:   The Office of Enforcement and Compliance Assurance's Agency Response to the
            Draft "Evaluation Report: Opportunities Exist to Improve and Replicate EPA's
            National Petroleum Refinery Compliance Program," dated March 5,2004

FROM:      Phyllis Harris Isl
            Acting Assistant Administrator
            Office of Enforcement and Compliance Assurance

TO:         Jeffrey K.  Harris
            Director
            Program Evaluation, Cross-Media Issues
            Office of Inspector General

Introduction

      Today, on behalf  of the Office of Enforcement and Compliance Assurance's (OECA),
and as the designated "Action Official," I am forwarding to you our consolidated "Agency
Response" (Response) regarding the Office of Inspector General's (OIG) draft "Evaluation
Report: Opportunities Exist to Improve and Replicate EPA's National Petroleum Refinery
Compliance Program," dated March 5,2004 (Evaluation Report). OECA has actively solicited
comments from Regions  5, 6,  8, and 9. Accordingly, the attached Response represents the
consolidated comments of OECA and those Regions that have provided comments.

      In accordance with the instructions provided in your March 5,2004 memorandum, the
Response addresses the factual accuracy of the draft Evaluation Report. Consistent with those
instructions, the Response also specifically indicates whether OECA concurs with each of the
recommendations proposed by the OIG.  Further, to the extent that action has already been
initiated or planned to address issues identified in the draft Evaluation Report, the Response
specifically identifies those actions that have been initiated or planned. Finally, your March 5,
2004 memorandum expressly  states that the "final report will include an assessment of [the]
comments" made in the Response. Consequently, I am specifically requesting that this
memorandum and the attached Response be attached to, and be made a part of, the final version
of the draft Evaluation Report.
                                        53

-------
       It is OECA's view that the final report should remain confidential in its entirety. The
draft Evaluation Report and Response are a road map of OECA's internal deliberations regarding
how, when, and under what circumstances it will deploy its array of enforcement tools to secure
compliance. Because targeting, settlement, and litigation under the Refinery Petroleum Initiative
(Initiative) remains extremely active and planning is currently underway for further
implementation during the fiscal year 2005 through 2007 time frame, it is our view that
disclosure of such deliberative product is not appropriate at this time. Moreover, from a practical
perspective, the publication of this report will make it far more difficult for EPA to reach
agreement with other refiners. At a minimum, the publication of this draft Evaluation Report
will offer potential settlers and their counsel arguments that they can advance in negotiations and,
more significantly, litigation that the conduct of the Initiative is inconsistent with the comments
made in the draft Evaluation Report and therefore unfair.

       The Evaluation Report expresses the views of the OIG only.  As summarized in this
memorandum and noted in greater detail in the attached Agency response, OIG's Report does not
represent the views of OECA regarding the Initiative.

General Comments

       It is apparent from the face of the draft Evaluation Report that it is the product of a
considerable amount of work and effort. Moreover, as indicated in the attached Response, there
are observations, recommendations and lessons learned made in the draft Evaluation Report that
will be helpful to EPA as it moves forward in the Initiative and as it develops and implements
new initiatives.

       However, as outlined below in this memorandum,  and with more specificity in the
attached Response, it is OECA's view that the draft Evaluation Report has several significant
shortcomings.  We highlight those shortcomings below:

       The draft Evaluation Report contains errors, omissions and misstatements To assist
you in the preparation of the final report we have highlighted those errors, omissions and mis-
statements in the attached Response by reference to the page, chapter and sentence in which they
appear. In addition, we have provided you with the specific comments that explain the bases for
our view that the identified sections of the draft Evaluation Report should be corrected.

       The draft Evaluation Report does not place the petroleum refining priority in its
proper historical context  The draft Evaluation Report does not reflect an understanding of the
many challenges that OECA identified and overcame as it established and sustained  priority
attention on mis industrial sector and its multi-media problems.  Rather, the draft Evaluation
Report gives the impression that sector-based and multi-media analysis and targeting,
environmental baseline and results measurement, and national planning and accountability
processes were fully mature in the mid-1990's when the first work on the Initiative commenced.
To the contrary, many of these features were in their early or conceptual  stages of development
when the petroleum refining sector first appeared on OECA's radar screen. The OECA
reorganization in 1994 marked a nearly complete overhaul of EPA's compliance and
enforcement business model, and to not put this report in that context is a big shortcoming.
                                         54

-------
       The draft Evaluation Report is unbalanced and does little to highlight the
Initiative's successes. The Petroleum Refinery Initiative is one of the most successful
enforcement initiatives ever undertaken by EPA. Since approximately January 2000, the date
lhat EPA began to formally engage petroleum refining companies in global settlement
discussions regarding their Clean Air Actnoncompliance, EPA has obtained settlements with 11
petroleum refiners representing almost 40% of the nation's domestic refining capacity and
covering 42 separate refineries for each of the major four substantive areas related to Clean Air
Act compliance.  The settlements contain substantial "beyond compliance" requirements, and
taken together, represent a breadth and depth of coverage not previously realized in the
enforcement program.

       The draft Evaluation Report does not demonstrate an appreciation for the
complexity of die issues EPA has successfully addressed under the Initiative and its
unprecedented scope. Prior to development of the Initiative, EPA had largely approached
enforcement on a facility-by-facility, issue-by-issue basis. The Initiative represents a radical
departure from this practice.  There are few industries as complex as the petroleum refining
industry and there are few regulatory programs as complex as the Clean Air Act.
Notwithstanding this complexity, under the Initiative, EPA successfully embraced the global
consent decree as a mechanism to secure permanent, consistent compliance with the Clean Air
Act on a company-wide basis.  Yet, the draft Evaluation Report only mentions this in passing.
Moreover, the substantive discussion does not appear to recognize or fully appreciate this
complexity as a factor in the development of OECA's consent decree implementation strategy.

       The draft Evaluation Report fails to account for the evolution of the Initiative. For
example, the OIG in the draft Evaluation Report appears to fault OECA for having an "absence
of strategic direction" for the petroleum refining priority. The OIG asserts that this is so because
some strategy documents do not have dates or a signature. We do not agree that from these facts
the OIG can conclude that OECA management did not have an idea of what it wanted to
accomplish strategically in this sector. As identified with further specificity in our comments, the
Initiative strategy did evolve over time as EPA learned more about the sector based on its
experience in the field. EPA learned, for example, that noncompliance in this sector was much
more significant and widespread than our original analysis suggested. The documents that OIG
should look to determine whether OECA had a strategic direction for this priority include not
only the petroleum refining sector strategy, but also the MOA guidance documents, the
individual regional MOAs which were agreed upon at the highest management levels in OECA
and the regions, and the MOA updates. Similarly, EPA learned, and continues to learn, from its
experience as the consent decree implementation phase of the Initiative continues to evolve.
Many of those lessons learned have already been incorporated into the consent decree
implementation process and  have resulted in a significant improvement in responding to consent
decree deliverables.      •             *

       The OIG fails to recognize the severe resource constraints under which the Initiative
operates, and the innovative approach EPA employed to overcome these resource
constraints. First, the OIG in the draft Evaluation Report criticizes the level of resources
committed to the Initiative. However, the OIG fails to consider the total level of resources
available to OECA's air enforcement program overall (covering the Refinery Initiative as well as

                                          55

-------
all other air enforcement activities).  Because this is nowhere taken into account, the critique of
the resource levels supporting the Initiative is wholly without context. Had the OIG included this
in the assessment, the leveraging benefits of the "global consent decree" approach would have
become apparent.

       Second, as a corollary proposition, and perhaps more importantly, the draft Evaluation
Report nowhere acknowledges or considers the level of resources that would have been required
to secure company-wide compliance and "beyond compliance" commitments had the innovative
"global" settlement approach not been used. A refinery-by-refinery, issue-by-issue approach in
which and individual inspection(s) or investigations) is conducted at each and every refinery
followed by information requests), not'ce(s) of violation, negotiations and/or litigation would
take many years from the inception of the investigation to a final resolution. The resources that
me Agency would have needed to expend under those circumstances to obtain company-wide
compliance under the New Source Review, new source performance standards, benzene waste,
and leak detection and repair programs would have been beyond our means. By way of
illustration, during the fiscal year 2001 and 2002 tuneframe, EPA conducted inspections or
investigations at 7 facilities for compliance with new source performance standards and leak
detection and repair program requirements. On just these two areas, EPA inspectors spent, on
average, approximately  366 hours (with a high of 702) at each facility.  Had EPA addressed the
refinery non-compliance issues in such a piecemeal fashion, using the high end of this range
(which is conservative since those investigations covered a narrower set of compliance issues),
we estimate that EPA would have expended nearly 15 full-time equivalents or 29,400 hours to
inspect each of the 42 refineries now covered by global consent decrees. This level of resources
is unavailable to OECA to devote solely to compliance in a single industry, without
compromising efforts elsewhere.

       Additionally, under the refinery-by-refinery, issue-by-issue approach, the scope of
injunctive relief at each separate facility would typically be limited to correcting those violations
identified during the inspection or investigation.  By leveraging activities at fewer faculties to
support a company-wide settlement on a broad range of emissions issues using the global
approach, EPA extended its reach and effectiveness far more efficiently than it otherwise would
have.  Rather than recognize OECA's approach in the Initiative as a creative solution in the face
of limited resources that was developed and spearheaded by those charged with accomplishing
results, the draft Evaluation Report overlooks mis almost completely.
Conclusion

       In an era of shrinking resources, it is OECA's view that the OIG unfairly criticizes a
creative and innovative approach to extending the reach of OECA's air enforcement program's
efforts to address the difficult compliance challenges presented by the petroleum refining
industry. The draft Evaluation Report lacks balance, overemphasizing shortcomings and failing
to provide a robust discussion of accomplishments.  Moreover, the OIG vastly underestimates the
complexity of refinery operations and of the controls necessary to reduce emissions, as well as of
the resource imph'cations of the evaluation's various recommendations. It also consistently
understates (or simply fails to give credit for) the "beyond compliance" aspects of the consent

                                          56

-------
decrees and of the benefits of a "global" approach to settlement.  Finally, the draft Evaluation
Report does not display a real understanding of the enforcement program as a whole, nor of how
the Initiative fits within that larger context. As a result, and without any basis for comparison,
the draft Evaluation Report draws unsupported and unsupportable conclusions regarding the
Initiative, the accomplishments achieved to date under the Initiative, and how those
accomplishments compare to other enforcement efforts.

Attachments
                                          57

-------
                       Responses to Proposed Recommendations

Chapter 3

Recommendations 3-1, "Develop clear overall refinery program goals that allow for future
assessment or measurement and include timetables for accomplishment"; 3-2, "Instruct OECA
refinery program managers to develop clear goals specifically for the refinery program's
implementation phase "; and 3-3, "Ensure that all goals and performance measures are
understood and shared by everyone involved in the national petroleum refinery program,
including all EPA and State staff involved in some portion of consent decree implementation."

 [This comment now refers to Recommendations 2-1, 2-2, and 2-3.]

       Concur. As discussed in detail in the preceding comments, OECA believes that the
refinery program's goals have been clearly articulated since the national strategy was initiated in
1998, and that as the program evolved they were further reflected in MOAs, etc. in the following
years. As a general matter., OECA agrees with these recommendations and will continue to
develop and articulate appropriate goals and performance measures.

Following the identification of refineries as an enforcement priority for FY96/97, OECA soon
recognized the need.for a comprehensive national strategy. It then developed a flexible,
integrated strategy (including sub-strategies) to address issues of widespread compliance and
enforcement concern at petroleum refineries.  The resulting 1998 strategy was developed in close
consultation and coordination with the Regions, the Office of Compliance and the Office of
Regulatory Enforcement's media-specific enforcement divisions. It has  remained largely
unchanged since then, with a focus on targeted investigations of "marquee" issues at petroleum
refineries and the goal of 50% improved compliance and 20% reduced emissions. The national
petroleum strategy and its implementation were regularly discussed at the staff level and
periodically reviewed by senior management in meetings, during conference calls and through
1he MOA process. Periodic progress updates have also been and will continue to be circulated to
OECA management and the regions, but the extent to which specific individuals clearly
understand the national strategy, including its sub-strategies, goals and objectives, may depend
on the level of their direct involvement in these processes and communications.

Recommendation 3-4, "Instruct OECA refinery program managers to use existing EPA, OECA,
and outside guidance to develop reliable performance measures to assess their progress toward
meeting national program goals. Specifically, managers should fully implement OECA's
performance-based approach to program management as described in its December 18, 2002,
Recommendations for Improving OECA Planning, Priority Setting and Performance
Measurement, which specifies development of plans and reliable performance measures, to the
remaining phases of the petroleum refinery program."

 [This comment now refers to Recommendation 2-4.]
                                          58

-------
       Concur. OECA has already begun to implement this recommendation (planned for prior
to the Evaluation), as priority planning process consistent with existing OECA guidance for FY
2005 has already been initiated.

Recommendation 3-5, "Validate and build upon the refinery program logic model we developed
during the evaluation, and consider developing similar program logic models for other OECA
programs to develop a clear consensus on program goals and how a program is intended to
work."

 [This recommendation was eliminated from the final draft]

       Non-concur. OECA does not agree that several of the short-term, intermediate and long-
term outcome measures in mis logic model are appropriate benchmarks for judging the
effectiveness of a compliance and enforcement program.  These goals - such as "increased
flexibility for refineries to expand or upgrade operations" - are not realistic or likely to be
obtained in an adversarial enforcement context. The absence of recommendations for how
OECA would benchmark the "before" conditions and measure changes overtime for these
ultimate outcomes means that there is no basis for OECA to determine whether these measures
are feasible. However, OECA does agree that it should use appropriate performance measures
and outcomes to measure performance under the Initiative, and intends to do so as part of the FY
2005 priority planning process.

Recommendation 3-6, "Instruct OECA managers to verify emissions reductions predicted in
consent decrees on a quarterly basis. Verification might include establishing a detailed
monitoring system, which could contribute to refinery program performance measurement."

 [This comment now refers to Recommendation 2-6.]

       Non-concur. As noted in the detailed comments on this issue, OECA does not believe
that this is an appropriate or effective use of resources. Furthermore, the recommendation
fundamentally misconceives the timing of reductions under the decrees, which does not happen
immediately upon lodging or entry of the decree (as is apparently assumed), with regular
reductions on a steady quarterly basis.  In part because these facilities are operating under court
order, and are required to submit reports and certify regarding their compliance with consent
decree requirements (punishable by contempt and/or criminal sanction), there are sufficient
indicia of reliability such that quarterly oversight of emissions reductions is riot necessary.
Significantly, the recommendation does not take into account the resource implications of this
level of monitoring - both with respect to those available for the Initiative (failing to recognize
that this work would need to be performed and/or reviewed by the same group of national experts
responsible for all other aspects of the Initiative), as well as those available to the air enforcement
program and OECA as a whole. Even if OECA agreed that this level of monitoring is
appropriate, it is not clear how this would be accomplished within the current resource levels and
in light of other priority activities. On balance, these resources are better utilized if devoted to
addressing compliance issues in other industry sectors. Notwithstanding the foregoing, OECA
agrees that it is important to track  emissions reductions under the consent decrees, as appropriate
given the consent decree milestone dates.

                                          59

-------
Recommendation 3-7, "Instruct OECA refinery program managers to gather, analyze, and
report relevant program data to support overall OECA organizational decision making, and
daily program decision making."

 [This comment now refers to Recommendation  2-5.]

       Concur.  As with Recommendations 3-1  through 3-4, OECA agrees with the principle
embodied in this recommendation, and will take steps to implement appropriate data gathering
and analysis to support program decisionmaking. However, in light of activities identified in
response to Recommendations 3-1 through 3-4, this recommendation appears redundant and
unnecessary.

Recommendation 4-1, "Instruct its consent decree tracking contractor to resume tracking both
company due dates for reports and EPA response due dates so that OECA and outside parties
can easily track company and EPA responsiveness."

 [This comment now refers to Recommendation  2-7.]

       Non-concur. As discussed in the detailed comments above, OECA does not agree that it
is necessary to further revise the tracking system at this time; this recommendation has been
overtaken by events. During the time that OIG was conducting its investigation, OECA itself
identified some deficiencies with its tracking system, and appropriate revisions were made (note
that due dates for reports and EPA responses continue to be tracked under each consent decree's
Master Inventory).  The critical issue is not simply tracking, but responding to those reports
requiring an EPA response. Changes have already been made to address this.  For example, in
January 2004, Matrix and Region 6 tracking systems were compared and verified for accuracy
and usefulness, and a single system was selected for implementation nationally, and access
provided to all parties responsible for consent decree implementation - including  companies. In
addition, changes to requirements for company  submittals have been made to subsequent
consent decrees to better manage the process. Furthermore, substantial progress has been made
to reduce the backlog, indicating that the current approach is having the  desired result.

Chapter 4

Recommendation 4-2, "Create a comprehensive tracking plan and system that outlines specific
responsibilities for OECA staff, EPA Regions, State and local air pollution control agencies, and
companies."

 [This comment now refers to Recommendation  2-8.]

       Non-concur. OECA does not agree that  this recommendation is necessary. Sixteen  states
are parties to global refinery consent decrees and currently receive copies of all consent decree
submissions that relate to each refinery within their states. OECA staff, EPA Regions and
state/local authorities who are parties to the consent decrees are and continue to be reflected in
the consent decree implementation plan. Specific tracking tools (e.g., Master Inventories and
Activity Lists) are circulated on a monthly basis to all necessary participants. As  discussed

                                         60

-------
above, EPA has a comprehensive consent decree tracking protocol that is being implemented
through our contractor.  Subject to claims of privilege and confidentiality, OECA does not object
to any interested non-party, including other states and local authorities, requesting tracking
information from Matrix at its own expense.

Recommendation 4-3, "Provide additional training at the regional level, and empower regional
experts to review and respond to company reports. Allow national technical leads to spot-check
responses from regional experts to ensure national consistency."

 (This comment now refers to Recommendation 2-9.]

       Concur.  OECA agrees with this recommendation, and has provided (and will continue to
provide) appropriate training as needed.

Recommendation 4-4, "Develop a formal feedback system to ensure that OECA's workforce and
managers have a common understanding of implementation responsibilities,  a common
perspective on the status of implementation, and the ability to expeditiously address
implementation issues."

 [This comment now refers to Recommendation 2-10.]

       Non-concur.  As explained in the detailed comments, in light of the small staffing level
and their overlapping responsibilities, a formal feedback system is not necessary. Should the
number of staff substantially increase in the future, a feedback system may be appropriate at that
time.

Recommendation 4-5, "Ensure frequent and open communication between partners (States,
regions) and headquarters about responsibilities for executing portions of strategies so that
misconceptions or confusion can quickly be eliminated. "

 [This comment now refers to Recommendation 2-11.]

       Concur.  OECA will continue to communicate with Initiative partners.

Recommendation 4-6, "As discussed with OECA managers, include consent decree
implementation in OECA priorities and strategic plans, allocating staff and resources to
implementation until OECA completely implements all consent decrees."

 [This comment now refers to Recommendation 2-12.]

       Concur.  OECA agrees with the principles underlying these recommendations, and efforts
have already begun for FY05 implementation on these matters.  OECA will continue to allocate
adequate resources to the Initiative whether it is identified as a national priority or part of the
core program.
                                          61

-------
Recommendation 4-7, "Develop a plan for allocating negotiation and implementation resources.
Use resource planning in new initiatives to determine the predicted workload associated with the
initiative; allocate training, education, and development resources; and provide for office-wide
revaluation of the resource plan. "

 [This comment now refers to Recommendation 2-13.]

       Concur. This recommendation has been overtaken by events (priority planning for
FY05), and implementation and other resources will be allocated in concert with other OECA
priorities and core programs requirements. Office-wide (and Region-wide) reevaluations are
considered as part of regular planning processes.

Chapter 5

Recommendation 5-1, "Disseminate the lessons learned from the refinery program to OECA  >
staff to benefit other compliance efforts, obtain additional feedback from stakeholders '-
including States, industry, and environmental groups - on other lessons learned, and update
OECA's  "Frameworkfor a Problem-Based Approach to Integrated Strategies "for on-going and
future industry-specific enforcement programs."

       Concur in part, non-concur in part. OECA does not agree that revisions to the recently-
issued Framework for Problem-Based Approach to Integrated Strategies (November 2002)
("Framework") are needed to reflect lessons learned from the Initiative.  Rather, the lessons
learned from EPA's Refinery Initiative have informed and continue to inform the Agency's
evolving problem-based approach to solving environmental compliance problems. For example,
OECA is currently engaged with the Regions in developing performance-based strategies for
each of the national priorities selected for FY05-07. As part of that effort, EPA is reviewing and
refining, where appropriate, the goals and the strategies for the refinery initiative. In developing
these performance-based strategies, OECA and the Regions will be guided by the recently-issued
guidance, Template for Developing a Performance-Based Strategy for National Compliance and
Enforcement Priorities (Final Draft February 18, 2004) as well as the Framework. As EPA gains
more experience in the development and implementation of such strategies, we will refine
guidance on the use of such strategies where needed.

Recommendation 5-2, "Designate a senior OECA executive to assume the role of champion for
the refinery program to ensure (a) that all refiners enter into consent decrees or face appropriate
alternative enforcement actions, and (b) consent decrees are effectively implemented."

       Concur in part, non-concur in part. As noted above in the comments on the "Conclusion"
section, OECA does not agree that there is no "champion" for the Initiative. However, OECA
agrees with the need for national enforcement priorities to be managed by a senior enforcement
official (e.g., Division Directors or their Associates), working on a team with other senior
managers from EPA headquarters, regions and DOJ. Since 2002, the senior enforcement official
responsible for managing the refinery initiative has and continues to be the Associate Director of
ORE's Air Enforcement Division, who is now serving as the Acting Director of the Air
Enforcement Division. OECA agrees that the Air Enforcement Division Director is responsible

                                          62

-------
                                                                                                      I
for ensuring that (a) refineries enter into consent decrees or face appropriate enforcement action,
and (b) consent decrees are effectively implemented.  OECA does not agree that it is necessary
for EPA to settle with or litigate against all refiners in the industry under the Initiative. The goal
of the Initiative is and has been to increase compliance by 50% and decrease emissions from
refineries by 20%.  The "100%" goal suggested by OIG miscomprehends the purpose of a
"priority." Even after this is no longer a priority, further work in this area would be undertaken
through the "core" program guidance, including the potential for multi-regional priorities, as well
as the potential for State efforts.  OECA's work in an area is not just be driven by a coverage
number, but by whether there continues to be an appropriate federal role. OECA designates a set
of national priority  criteria (i.e., significant environmental benefit, pattern of noncompliance,
appropriate federal  role), and following the return of refineries to the "core" program certain
refineries may be better handled by States or as part of a multi-regional priority.  OlG's
suggestion of "all" refineries lacks the context of taking into account all of our regulatory
partners, and that certain types of facilities are best addressed at different levels.

Recommendation 5-3, "Consider designating a senior OECA executive to assume the role of
champion for each of the other enforcement priority areas. EPA and industry officials should
recognize the champion as knowledgeable and as having the authority to make decisions related
to the priority area."

       Concur in part, non-concur in part. The determination of whether any particular initiative
or priority area requires an OECA-designated "champion," and at what level, will be made on a
case-by-case basis as is appropriate in light of all relevant facts and circumstances.  For those
areas mat have been selected as national enforcement priorities for FY 2005, senior OECA and
regional management have been named as "champions for Ihe purpose of developing
performance-based strategies for each priority area. In addition, the OECA Planning Council,
which meets monthly, will regularly assess progress implementing, and results achieved through
the priority performance-based strategies. In order to ensure that adequate progress is being
made towards achieving priority goals the Planning Council will modify performance-based
strategies as needed, and make recommendations to the OECA Assistant Administrator for
flexibly deploying resources to address workforce gaps.

Recommendation 5-4, "Develop a communications plan for refinery consent decree
implementation. The plan should clearly describe the roles and responsibilities of all
stakeholders,  including refinery priority area experts and regional and State officials."

       Concur. OECA is already in the process of priority planning for FY 2005 (begun prior to
Ihe Evaluation), which will result in a revised performance-based strategy for the refinery sector.
The performance-based strategy for FY 2005 will outline the path forward in (a) completing the
refinery sector as a national priority and (b) ensuring that refineries governed by federal consent
decrees comply with the terms and conditions of their consent decrees.
                                          63

-------
64

-------
                                                                     Appendix H

               O/G Evaluation of EPA Comments

On April 2,2004, OECA provided us with a memorandum summarizing its overall comments,
including its comments on the recommendations.  We included the full text of OECA's summary
memorandum as Appendix G. In this Appendix, we highlight and evaluate specific comments
from OECA's summary.memorandum. We organized this Appendix along the same lines OECA
used in its summary memorandum. OECA also provided us with detailed comments as an
attachment to its memorandum. We have posted this attachment and our evaluation of OECA's
comments on our web site at http://www.epa.gov/oig/publications.htm.  We modified the text as
we determined appropriate based on OECA's detailed comments.

OECA stated that the report will help to EPA as it continues to implement the refinery program
and other initiatives. OECA also stated that the report had several significant shortcomings.
OECA stated that the report unfairly criticized a creative and innovative approach to address the
difficult compliance challenges presented by the petroleum refining industry.

OECA agreed with 10 of the 18 recommendations in our official draft report, partially agreed
with 3, and disagreed with 5. We included a summary of OECA's chapter-specific comments
and our evaluation of those comments at the end of each chapter.

We made various changes to the report as we determined appropriate based on OECA's
comments. We also eliminated some unnecessary detail, arid combined the information
previously in draft report Chapters 1 and 2 into a single  chapter (now Chapter 1), and Chapters 3
and 4 into a single chapter (now Chapter 2) to clarify our message. We renamed draft report
Chapter 5 as Chapter 3. In addition, we eliminated a recommendation concerning the use of a
logic model because the logic model is only one of several possible means that OECA may
employ to achieve the ends we advocate.

Report Confidentiality

OECA requested that the final report remain confidential in its entirety because the report
findings may adversely impact current and future negotiations with the refinery industry.  We
asked OECA to identify specific enforcement sensitive portions of the report, or portions where
the release would damage negotiations; OECA did not do so.  We believe the report provides an
accurate evaluation of the refinery program at a point in time and makes recommendations that
can improve program implementation and results.

Errors, Omissions, and Mis statements

OECA stated that the draft report contained errors, omissions, and misstatements. Where the
Agency clearly identified specific errors, omissions, or misstatements, and where we either
already had specific evidence supporting the Agency's suggested changes or where the Agency
                                       65

-------
provided specific evidence supporting different facts as part of its written comments, we made
appropriate changes to the report

Historical Context

OECA stated that the draft report did not place the petroleum refining priority in its proper
historical context, did not reflect an understanding of the many challenges that OECA overcame,
and gave the impression that all necessary management systems had fully matured when the
program began.  OECA also stated that we should mention the 1994 OECA reorganization,
which marked a nearly complete overhaul of EPA's compliance and enforcement business
model.

We believe the report places the petroleum refining priority in its proper historical context given
our evaluation scope and objectives.  We understand the challenges that OECA overcame and
believe the report reflects that fact. For example, in Chapter 1 of the report, we describe how
EPA shifted from routine Clean Air Act inspections to more targeted, resource-intensive
investigations that focused on carefully assessing emissions released as a result of refinery
processes.  We also believe mat OECA could have done a better job planning and implementing
the program whether or not its various management processes had fully matured in 1996 when
the refinery program began. Further, we did not state in the report that any of these processes had
fully matured, although we believe they should have matured sooner than they did.

Balance

OECA stated the draft report was unbalanced and did little to highlight the program's successes.
OECA stated that the settlements with refiners contain substantial "beyond compliance"
requirements and, taken together, represent a breadth and depth of coverage not previously
realized in the enforcement program.

We believe the report is well balanced and adequately highlights the program's success. For
example, we recognize in Chapter 1 that EPA obtained settlements with 11 petroleum refiners
representing 39 percent of the nation's domestic refining capacity and covering 42 separate
refineries and the settlements address each of the four priority areas under the refinery program.
Chapter 1 also recognizes that the settlements contain "beyond compliance" requirements and
describes OECA's compliance assistance and incentives developed as part of the refinery
program. In Chapter 3, we describe the lessons learned from the refinery program, such as
focusing on specific enforcement concerns, pulling together EPA staff with knowledge about the
industry, using in-house experts, focusing on the end result, and encouraging and requiring the
development of new emissions control technologies.

Complexity of the Issues

OECA stated that the draft report did not demonstrate an appreciation for the complexity of the
issues EPA successfully addressed and the refinery program's unprecedented  scope.
                                         66

-------
We believe the report demonstrates a keen appreciation for the complexity of the issues EPA
addressed. Without becoming too technical, the report provides sufficient and succinct
background on the complexity of the industry and the compliance issues addressed under the
refinery program and in the global consent decrees.  For example, Chapter 1 references Appendix
C that illustrates the complexity of the petroleum refining process, explains why OECA used
EPA national experts in investigating compliance, and summarizes Ihe four priority areas
addressed under the refinery program and the consent decree requirements for each priority area.
Readers should not interpret our succinct descriptions of the four priority areas to mean that we
did not appreciate or understand their complexities.  In addition, in Chapter 3, we describe how
refinery program staff applied technical expertise to gain knowledge of the industry and
compliance issues, and  to obtain credibility with the industry on its technical aspects.  We cannot
comment on the "unprecedented" scope of the refinery program because we did not compare it to
all other enforcement programs conducted by EPA.

Program Evolution

OECA stated that the draft report did not account for the program's evolution.  Our official draft
report referred to an "absence of strategic direction" for the refinery program and OECA stated
that, "We do not agree that from these facts that OIG can conclude that OECA management did
not have an idea of what it wanted to accomplish strategically in this sector." OECA stated that
the strategy evolved over time as EPA learned more about the sector based on its experience in
the field.  As evidence of its strategic direction, OECA suggested that we look at Memorandum
of Agreement guidance documents, Ihe individual regional Memorandum of Agreements agreed
upon at the highest management levels in OECA and the regions, and Memorandum of
Agreement updates. Similarly, OECA stated that EPA learned, and continues to learn, from its
experience as the consent decree implementation phase continues to evolve. OECA stated that it
has already incorporated many lessons learned into the consent decree implementation process,
resulting in significant improvement in responding to consent decree deliverables.

We believe the report appropriately communicates that the strategy evolved over time as EPA
learned more about the  refinery sector. The report recognizes that OECA learned from
implementation experiences and took steps to address its challenges. Chapter 1 clearly describes
the evolution of the refinery program from inspections to investigations, through global
settlements and consent decree implementation.

Although we do not believe the report conveyed that OECA "did not have an idea of what it
wanted to accomplish strategically in this sector," we believe EPA could have done a better job
of communicating  and documenting its goals and strategy. We discuss these issues in greater
detail in Chapter 2.

To determine whether OECA had a strategic direction for mis priority, we looked at not only the
petroleum refining sector strategy documents, but also the Memorandum of Agreement guidance
documents and updates OECA referred to in its comments. As we describe in Chapter 2, some
OECA officials told us  the information in the Memorandum of Agreement documents was not
entirely accurate or they were not familiar with the information. Based upon meetings with
                                         67

-------
senior OECA officials, we concluded that not all managers and staff used or even considered the
Memorandum of Agreement documents for planning and managing the refinery program.

Resource Constraints

OECA stated that we did not recognize the severe resource constraints under which the refinery
program operates and the innovative approach EPA employed to overcome these constraints.
OECA stated that we did not consider the total level of resources available to OECA's air
enforcement program and, had we done so,  we would have realized that OECA leveraged
benefits through its "global consent decree" approach. OECA stated that by leveraging activities
at fewer facilities to support company-wide settlements on a broad range of emission issues, EPA
extended its reach and effectiveness far more efficiently than it otherwise would have. OECA
stated it applied a creative solution in the face of limited resources.

We fully recognize the resource constraints under which this and all EPA programs operate, and
we believe the report accurately describes the approach EPA employed to overcome resource
constraints. We did not evaluate the amount of resources EPA chose to devote to the refinery
program compared to the total amount of resources OECA made available to its overall air
enforcement program or to other OECA programs.  While we  originally planned to conduct
evaluations of OECA's entire suite of enforcement priorities, OECA persuaded us to first pilot
our approach in a single priority area. In consultation with OECA, we chose the refinery sector
for our pilot We also consulted numerous times  with senior OECA officials over several   .
months at the beginning of this evaluation in determining the evaluation's scope and objectives.
Throughout our extensive consultations, OECA staff never suggested that we include among our
objectives a comparison of resources devoted to various other enforcement programs such as
OECA suggested in its comments.
                                         68

-------
                                                                       Appendix I

                                Distribution
EPA Headquarters

      Administrator
      Acting Assistant Administrator, Office of Enforcement and Compliance Assurance
      Assistant Administrator for Environmental Information
      Director, Office of Regulatory Enforcement, Office of Enforcement and Compliance
         Assurance
      Director, Office of Compliance, Office of Enforcement and Compliance Assurance
      Acting Director, Air Enforcement Division, Office of Regulatory Enforcement, Office of
         Enforcement and Compliance Assurance
      Comptroller (2731 A) (2724A)
      Agency Followup Official (the CFO) (271OA)
      Agency Audit Follow-up Coordinator (2724A)
      Audit Follow-up Coordinator, Office of Enforcement and Compliance Assurance
      Associate Administrator for Congressional and Intergovernmental Relations (1301 A)
      Acting Associate Administrator for Public Affairs (1101 A)
      Director, Office of Regional Operations (1108 A)
      Inspector General (2410)
EPA Regions

      Regional Administrators
      Regional Audit Follow-up Coordinators
                                        69

-------

-------
             EPA's Detailed Comments on Official
 Draft Refinery Compliance Program Evaluation Report
         and Office of Inspector General Response
On April 2,2004, EPA's Office of Enforcement and Compliance Assurance (OECA) provided
our office - the Office of Inspector General (OIG) - with a memorandum summarizing its
overall comments, including its comments on the recommendations, to our official draft report
on EPA's National Petroleum Refinery Compliance Program, dated March 5, 2004. We include
the full text of OECA's summary memorandum as Appendix G in our final report, EPA Needs to
Improve Tracking of National Petroleum Refinery Compliance Program Progress and Impacts,
Report No. 2004-P-00021, June 22,2004. In Appendix H of that report, we provide our
summary response.

OECA also provided us  detailed comments as an attachment to its summary memorandum.  We
have included in this document OECA's detailed comments and our response to selected
comments. OECA's comments are in black while our responses are in blue. We only responded
to comments not specifically addressed in our evaluation of OECA's summary memorandum, or
where we determined we needed to provide more details regarding our position on a particular
issue. In this document, we do not include our response to OECA's comments on the report
recommendations; we include our evaluation of those comments at the end of each chapter of the
final report.

We made various changes to the official draft report as we determined appropriate based on  '
OECA's comments. We also eliminated some unnecessary detail, and combined the information
previously in draft report Chapters 1 and 2 into a single chapter (now Chapter 1), and Chapters 3
and 4 into a single chapter (now Chapter 2) to clarify our message.  Chapter 5 became Chapter 3.
As a result of reorganizing the report, the page identifiers used in OECA's April 2,2004,
comments to the draft report will not match the page numbers in the final report.

OECA Summary Memorandum
Comments and OIG Evaluation

OECA stated that the report will help EPA as it continues to implement the national refinery
program and other initiatives.  OECA also stated that the report had several significant
shortcomings. OECA stated that the report unfairly criticized a creative and innovative approach
to address the difficult compliance challenges presented by the petroleum refining industry.

Report Confidentiality

OECA requested that the final report remain confidential in its entirety because the report
findings may adversely impact current and future negotiations with the refinery industry. We
asked OECA to identify specific enforcement-sensitive portions of the report, or portions where
the release would damage negotiations; OECA did not do so. We believe the report provides an


                                 Page 1 of 68

-------
accurate evaluation of the national refinery program at a point in time and makes
recommendations that can improve program implementation and results.

Errors, Omissions, and Misstatements

OECA stated that the draft report contained errors, omissions, and misstatements. Where the
Agency clearly identified specific errors, omissions, or misstatements, and where we either
already had specific evidence supporting the Agency's suggested changes or where the Agency
provided specific evidence supporting different facts as part of its written comments, we made
appropriate changes to the report,

Historical Context

OECA stated the draft report did not place the petroleum refining priority in its proper historical
context, did not reflect an understanding of the many challenges that OECA overcame, and gave
the impression that all necessary management systems had fully matured when the program
began. OECA also stated that we should mention the 1994 OECA reorganization, which marked
a nearly complete overhaul of EPA's compliance and enforcement business model.

We believe the report places the petroleum refining priority in its proper historical context given
our evaluation scope and objectives.  We understand the challenges that OECA overcame and
believe the report reflects mat'fact. For example, in Chapter 1, we describe how EPA shifted
from routine Clean Air Act inspections to more targeted, resource-intensive investigations that
focused on carefully assessing emissions released.  We also believe that OECA could have done
a better job planning and implementing  the program, whether or not its various management
processes had fully matured in 1996  when the national refinery program began.  Further, we did
not state in the report that any of these management processes had fully matured, although we
believe they should have matured sooner than they did.

Balance

OECA stated the draft report was unbalanced and did little to highlight the program's successes.
OECA stated that the settlements with refiners contain substantial "beyond  compliance"
requirements and, taken together, represent a breadth and depth of coverage not previously
realized in the enforcement program.

We believe the report is well balanced and adequately highlights the program's success. For
example, we recognize in Chapter 1 mat EPA obtained settlements with 11  petroleum refiners
representing 39 percent of the nation's domestic refining capacity and covering 42 separate
refineries,  hi addition, these settlements address each of the four priority areas under the
national refinery program. Chapter 1 also recognizes that the settlements contain "beyond
compliance" requirements and describes OECA's compliance assistance and incentives
developed as part of the national refinery program,  hi Chapter 3, we describe the lessons learned
from the refinery program, such as focusing on specific enforcement concerns, pulling together
EPA staff with knowledge about the industry, using in-house experts, focusing on the end result,
and encouraging and requiring the development of new emissions control technologies.


                                     Page 2 of 68

-------
Complexity of the Issues

OECA stated that the draft report did not demonstrate an appreciation for the complexity of the
issues EPA successfully addressed and the national refinery program's unprecedented scope.

We believe llie report demonstrates a keen appreciation for the complexity of issues EPA
addressed. Without becoming too technical, the report provides sufficient and succinct
background on the complexity of the industry and the compliance issues addressed under the
refiners' program and in the global consent decrees. For example, Chapter I references
Appendix C that illustrates the complexity of the petroleum refining process, explains why
OECA used EPA national experts in investigating compliance, and summarizes the four priority
areas addressed under the refinery' program and the consent decree requirements for each priority
area. Readers should not interpret our succinct descriptions of the four priority areas to mean
that we did not appreciate or understand their complexities.  In addition, in Chapter 3, we
describe how refinery program staff applied technical expertise to gain knowledge of the
industry and compliance issues, and to obtain credibility with the industry on its technical
aspects.  We cannot comment on the "unprecedented" scope of the refinery program because we
did not compare it to all other enforcement programs conducted by EPA.

Program Evolution

OECA stated that the draft report did not account for the program's evolution.  Our official draft
report referred to an "absence of strategic direction" for the refinery program, and OECA stated
that, "We do not agree that from these facts  that OTG can conclude that OECA management did
not have an idea of what it wanted to accomplish strategically in this sector." OECA stated that
the strategy evolved over time as  EPA learned more about the sector based on its experience in
the field.  As evidence of its strategic direction, OECA suggested that we look at Memorandum
of Agreement guidance documents, the individual regional Memorandum of Agreements agreed
upon at the highest management levels in OECA and the regions, and Memorandum of
Agreement updates. Similarly, OECA stated that EPA learned, and continues to learn, from its
experience as the consent decree implementation phase continues to evolve. OECA stated that it
has already incorporated many lessons learned into the consent decree implementation process,
resulting in significant improvement in responding to consent decree deliverables.

Wre believe the report appropriately communicates that the strategy evolved over time as EPA
learned more about the refinery sector.  The report recognizes that OECA learned from
implementation experiences and took steps to address its challenges. Chapter 1 clearly describes
the evolution of the refiner}' program from inspections to investigations, through global
settlements and consent decree implementation.

Although we do not believe the report conveyed that OECA "did not have an idea of what it
wanted to accomplish strategically in this sector," we believe EPA could have done a better job
of communicating and documenting its goals and strategy. We discuss these issues in greater
detail in Chapter 2.
                                     Page 3 of 68

-------
To determine whether OECA had a strategic direction for this priority, we looked at not only the
petroleum refining sector strategy documents, but also the Memorandum of Agreement guidance
documents and updates OECA referred to in its comments. As we describe in Chapter 2, some
OECA officials told us the information in the Memorandum of Agreement documents was not
entirely accurate or they were not familiar with the information. Based upon meetings with
senior OECA officials, we concluded that not all managers and staff used or even considered the
Memorandum of Agreement documents for planning and managing the refinery program.

Resource Constraints

OECA stated that we did not recognize the severe resource constraints under which the refinery
program operates and the innovative approach EPA employed to overcome these constraints.
OECA staled that we did not consider the total level of resources available to OECA's air
enforcement program and, had we done so, we would have realized that OECA leveraged
benefits through its "global consent decree" approach. OECA stated that by leveraging activities
at fewer facilities to support company-wide settlements on a broad range of emission issues,
EPA extended its reach and effectiveness far more efficiently than it otherwise would have.
OECA stated it applied a creative solution in the face of limited resources.

We fully recognize the resource constraints under which this and all EPA programs operate, and
we believe the report accurately describes the approach EPA employed to overcome resource
constraints. We did not evaluate the amount of resources EPA chose to devote to the refinery
program compared to the total amount of resources OECA made available to its overall air
enforcement program or to other OECA programs. While we originally planned to conduct
evaluations of OECA's entire suite of enforcement priorities, OECA persuaded us to first pilot
our approach in a single priority area.  In consultation with OECA, we chose the refinery sector
for our pilot. We also consulted numerous times with senior OECA officials over several
months at the beginning of this evaluation in determining the evaluation's scope and objectives.
Throughout our extensive consultations. OECA staff never suggested that we include among our
objectives a comparison of resources devoted to various other enforcement programs, as OECA
suggested in its comments.
                                     Page 4 of 68

-------
OECA Detailed Comments and OIG Response

Executive Summary

Page i, Introduction:

       The statement in the first sentence of the Evaluation Report (Evaluation) that "the
petroleum refinery industry ... ranked number one for noncompliance among hundreds of
industries" is inaccurate.  The correct statement is that the refinery sector ranked first among
"the 29 industry sectors tracked by EPA in 1996 (i.e., inspect'on-to-enforcement ratio)."

OIG Response: We modified the text as we determined appropriate.

Page i, Question 1:

       The phrase "toxic air pollutants are known to cause cancer" is an overstatement and
inaccurate. This should be revised to correctly state: 'Toxic air pollutants include pollutants
that are known or suspected to cause cancer or other serious human health effects."

OIG Response: We modified the text as we determined appropriate.

Page i, Question 2:

       The statement, "OECA's early planning documents indicate an absence of strategic
direction [because] three of the four documents OECA provided describing its strategy were
undated, unsigned, and without any evidence they were formally approved" is, as noted below
(and in more detail in the specific responses on each chapter), simply incorrect. A fundamental
error of the Evaluation is its failure to understand the larger MO A/priority planning process used
by OECA and Regions (and States), and how the Initiative fit within that larger context.  In
addition, OIG has established no causal connection between an undated document and a "lack"
of strategic direction.  This sentence should therefore be deleted and replaced with the following:

       "OECA's first planning document lacked strategic direction for identified national
       priorities, including petroleum refineries (MOA Guidance for FY96/97). Its  second
       planning document'corrected this deficiency and called for the development of a national
       petroleum refinery strategy in close coordination and consultation with affected EPA
       Regions (see, e.g., MOA Guidance for FY98/99). The resulting Petroleum Refining
       Sector Strategy was approved by the Director of OECA's Office of Regulatory
       Enforcement, distributed to and discussed with senior OECA and Regional management
       at a Washington meeting to review each national strategy that had been developed."
       The fourth sentence, which states "Nonetheless, an integrated strategy emerged that
included compliance assistance, inspections, enforcement, and compliance incentives that
addressed the most important or priority noncompliance problems" inaccurately infers that the
development of an integrated strategy for the petroleum refining sector was accidental or


                                     Page 5 of 68

-------
unplanned for. The development of an integrated strategy was deliberate and planned. For
accuracy, this sentence should be deleted and replaced with the following:

       "This integrated strategy developed for the refinery sector included compliance
       assistance, inspections, enforcement, and compliance incentives that addressed the most
       important or priority noncompliance problems."

       The final sentence, "OECA considers the program highly successful because, as of
March 2004, OECA projected the program would result in annual reductions of about 44,000
tons of NOx, 95,000 tons of SCh, and significant amounts of other pollutants," is problematic
and/or erroneous for a number of reasons. First is its use of the past tense, which is a problem
repeated throughout the Report. This tends to suggest that the Initiative is completed (it is not),
and mat the listed emission reductions have already been achieved by facilities currently subject
to consent decrees (implementation is at an early stage and is ongoing). Additionally, it confuses
matters that have already been done with those that are still to be done under the Initiative,
thereby inaccurately conveying the status of the Initiative and limiting the usefulness of the
OIG's Evaluation. This sentence should be replaced with the following:

       "The National Petroleum Initiative is highly successful because, as of March 2004, and
       based on the companies' estimates of emissions reductions they will achieve under the
       terms of the settlements, OECA projects that the program will result in annual reductions
       of about 44,000 tons of NOx, 95,000 tons of SOa, and significant amounts of other
       pollutants."

       Additionally, OIG's summary lacks historical perspective. OECA recommends that the
following be added:

       "To date, the National Petroleum Refinery Initiative is one of the most successful
       enforcement initiatives undertaken by EPA Since approximately January 2000, the date
       that EPA began to formally engage petroleum refining companies regarding their Clean
       Air Act non-compliance, EPA has obtained settlements with 11 petroleum refiners
       representing almost 40% of the nation's  domestic refining capacity and covering 42
       separate refineries for each of the major  four substantive areas related to Clean Air Act
       compliance. As of this date, OECA reports that it is engaged in similar settlement
       discussions with refiners representing an additional 40% of the domestic refining
       industry."

OIG Response: We deleted reference to "an absence of strategic direction." We also deleted the
entire discussion on OECA's undated, unsigned, and not formally approved strategy documents.
We believe OECA's documentation problems demonstrate how OECA's performance
measurement and reporting approach for the national refinery program has not provided useful
and reliable information necessary to effectively implement, manage, evaluate, and continuously
improve program implementation and results. Specifically, OECA has not established clear
program goals, performance measures, or a reporting system to track progress. OECA's lack of
undated, unsigned, and not formally approved documents further support our conclusion that
OECA lacked reliable and useful information to manage the national refinery program. We
discussed with EPA officials at the exit conference how documentation that is not signed, dated,

                                      Page  6 of 68

-------
or approved by senior officials may not be effectively used as a management tool. We decided
Chapter 2 of the report adequately describes significant management problems and we do not
need to include the documentation problem as another example.

We cannot comment on the national refinery program as "one of the most successful
enforcement initiatives undertaken by EPA" because we did not compare the program to all
other enforcement programs conducted by EPA. Further, OECA did not provide evidence to
support the national refinery program as one of its most successful initiatives.

Where appropriate, we modified the past tense in the report and made it clear that the national
refinery compliance program operates as an on-going program. In addition, we modified other
text as we determined appropriate.

Pages i-ii, Question 3:

      The assertion that OECA's measurement and reporting approach for Ihe national program
"did not provide useful and reliable information to effectively implement, manage, evaluate, and
improve the program" is not borne out by the facts.  OECA tracked and reported what it believed
was necessary to implement the focused approach for the refinery program.  Specific targets
were set in the sector strategy and then negotiated with the Regions for their final MOA
commitments. The progress of investigations in the areas of focus (LDAR, PSD/NSR, benzene,
flaring and sulfur recovery units) was, in fact, tracked on a near-real time basis through monthly
conference calls and periodic reports.

      The first sentence should therefore be revised to read as follows:

      "OECA's performance measurement and reporting approach for the national petroleum
      refinery program provided useful and reliable information necessary to effectively
      implement, manage, evaluate,  and improve the program."

      In addition, the statements that "OECA did not clearly and precisely define official
program goals and measures" and that "performance measurement and reporting systems were
ineffective for monitoring or reporting refinery program performance" are likewise unsupported
by the facts. OIG has erred in its assessment by inaccurately or inadequately taking the MOA
process into account and has failed to comprehend a major portion of me enforcement and
compliance program as a result. Under Hie MOAs, the Regions were required to provide end-of
year reports, which were to detail their accomplishments. Where reports were insufficient from
a quality standpoint, and where it could be determined that Regions did not meet their negotiated
goals, those were addressed in the subsequent MOA cycle. This dynamic is overlooked in the
Evaluation. Additionally, the draft Evaluation appears to be overly focused on the specific
numbers (such as pollution reduction estimates) in the strategy, but apparently confuses this with
the  goal to improve compliance for the four major areas that the Evaluation recognizes were
developed by OECA working closely with the Regions (see page 11).

      Accordingly, to state in the response to Question 3 of the Executive Summary that
"OECA did not have a consensus on what the program goals were" inaccurately implies that
                                     Page 7 of 68

-------
there was not then and is not now a general agreement as to what these goals are and were, and
what problems were the most significant to address.

       The balance of this paragraph should be revised to read as follows:

       "OECA established a variety of output or activity-related measures, and
       established an outcome-related measure for projected tons of emissions reduced.
       Although performance was regularly tracked and reported, existing EPA
     •  performance measurement and reporting systems were not used for monitoring or
       reporting refinery program performance, hi fiscal 2004, OECA began moving
       toward a more performance-based approach for all its activities, including
       enforcement initiatives. We made various recommendations to OECA related to
       developing and communicating clear refinery program goals and timetables that
       allow for future assessment  or measurement and made various other
       recommendations to ensure  OECA refinery program managers develop clear
       goals and gather needed data."

OIG Response: We disagree thatwe did not base our conclusion regarding OECA's
measurement and reporting approach on facts.  As described in Chapter 2 of our report, OECA's
performance measurement and reporting approach for the national refinery program has not
provided useful and reliable information necessary to effectively implement, manage, evaluate,
and continuously improve program implementation and results.  Specifically, OECA has not
established and communicated clear goals, systematically monitored refinery program progress,
reported actual outcomes, or tracked progress toward achievement of consent decree goals.
Although OECA officials used informal methods to track program progress, OECA would
benefit from using more formal mechanisms to measure progress toward consent decree and
overall refiners' program goals.

We also disagree with OECA that we did not factually support our conclusion that OECA did
not clearly and precisely define official program goals and measures.  As described in Chapter 2
of our report, EPA did not generally agree on the national refinery program goals, and OECA
officials referred to different goals and measures for the refinery program at different times!

We modified the text as we determined appropriate.

Page ii, Question 4:

       The second sentence, which states that "[djuring implementation, tracking problems
developed and persisted," incorrectly conveys that the tracking of reports and submissions was
the objective. In fact, the objective was responding in a timely manner to those company
submittals that required an Agency response. The sentence  should be corrected to state as
follows: "During implementation, response delays developed and persisted."

       There is no evidence to support the statement in the third sentence that actions on the part
of refineries that were necessary to reduce emissions were in fact delayed by a late EPA
response. Therefore, for accuracy the sentence should be substituted with the following:
                                     Page 8 of 68

-------
"Although these delays caused some delays in beginning sampling or paying stipulated penalties,
they did not delay company actions necessary to reduce emissions."

       The fourth and fifth sentences, which together state that there was a lack of prior
planning and that OECA only recently took steps to address tracking and implementation, are
inaccurate. These sentences should be deleted and substituted with the following:

       "OECA developed, in close consultation and coordination with affected Regions,
       a comprehensive plan for consent decree implementation, provided training
       opportunities for regional personnel to implement their consent decree
       responsibilities under that plan, and identified consent decree implementation as a
       priority resource concern in subsequent MOA Guidance.  See FY2003 MOA
       Guidance Update, p. 5 (June 2002); FY 2004/2005 MOA Guidance (June 2003);
       and FY 2004 MOA Guidance Update, p.  16 (July 2004).  OECA has taken
       numerous steps to address and resolve issues associated .with tracking of
       implementation of consent decrees.  This effort has, inter alia, manifested itself in
       the dramatic increase in the number of EPA responses issued in response to
       company submittals since the beginning of 2003."

OIG Response: We modified the text as we determined appropriate.

Page ii, Question 5:

       The "lessons learned" and recommendations identified by OIG following question 5
should be modified to reflect the comments EPA provides by way of response in the remainder
of this document. In this respect, the Evaluation should affirmatively acknowledge that the
recitation of "lessons learned" by OIG were, in fact, previously identified by OECA independent
of the Evaluation (e.g., having a "champion" for work in priority areas, identifying enforcement
concerns within an industry, etc.).

OIG Response: We obtained information on the lessons learned described in Chapter 3 of the
report from EPA, State, and industry officials, not only OECA officials.  While some OECA
officials may be aware of the lessons learned, other EPA staff, outside organizations, industry,
and the public may not.  We believe lessons learned are useful if they are shared among the
many stakeholders and acted upon. We believe mat if OECA already has awareness of the
lessons learned described in the report, it should have a head start in addressing those areas that
need improvement.

We modified the text as we determined appropriate.

Preface

       The Preface - in particular the first paragraph - suggests that OECA relies solely on a
"one size fits all" approach in its strategic planning efforts. OECA does not agree. A cursory
review of OECA's MO As and priority planning materials demonstrates that OECA employs a
wide range of approaches to address the diverse compliance issues presented by the numerous
and varied environmental enforcement programs administered by OECA.

                                     Page 9 of 68

-------
       Significantly, the statement that OECA's prior uses of an integrated strategy to address
environmental problems have been "ad hoc" and with "limited measurable results" is taken
directly from OECA's Interim Draft of the Guide to Implementing Integrated Strategies
Framework (June 2003 at p.2). That this is drawn from OECA's own Evaluation of integrated
strategies is nowhere attributed or cited in the draft Evaluation.  We emphasize this point
because the above-referenced statements are taken out of context and used selectively. When
viewed in its proper context, OECA's own critique can be understood as part of a series of pilots
for developing problem-based integrated strategies during  FY03 and FY04.  After those pilots
are completed, OECA's Office of Compliance plans to re-evaluate and revise the Framework.  It
is unhelpful for O1G to now use this recently-developed and still evolving Framework to
evaluate the adequacy of and process for developing an integrated strategy that was begun well
prior to this time (the Initiative's roots trace from 1996); a more useful evaluation  would have
relied on observations and critiques that OECAhad not itself already identified.

OIG Response: We modified the text as we determined appropriate.

Page 1, Question 2:

       The parenthetical in this question states that "compliance assistance, compliance
incentives, inspections and enforcement actions" are necessarily included in an integrated
strategy.  However, this is not accurate. The point of a "problem solving" approach and the use
of an integrated strategy is not to use all of the tools, but to first define the environmental and
compliance problem, and then to select which tool or combination of tools is most appropriate to
address that problem.  As phrased, the question is  symptomatic of OIG's erroneous "one size fits
all" perception of OECA's enforcement process noted above.

OIG Response: We modified the text as we determined appropriate.

Chapter 1

       The discussion in the "Petroleum Refining Process, Products, and Releases" section is
both vague and, at times, miscomprehends the Initiative. These errors are detailed in the
following comments.

OIG Response: We disagree with OECA that the section is vague  and mis-comprehends the
national refinery compliance program. The section provides background information on
refineries, including their products and releases. It provides the context for understanding the
refining process and its related pollutants.

Page 4, 2"* Paragraph:

       The fourth sentence of OIG's summary of the refining sector cites refinery emissions of
412,000 tons of "common air pollutants." However, the purpose for citing this data is both
unclear and unexplained, and it has no apparent connection for purposes of evaluating the
Initiative. The Initiative was plainly directed to reducing nitrogen oxides (NOx), sulfur dioxide
(SO2), paniculate matter (principally from fluidized catalytic cracking units and heaters and
boilers), fugitive benzene, and fugitive volatile organic compounds. It is not now (nor ever was)

                                     Page 10 of 68

-------
intended to address the entire suite of pollutants (e.g., ethylene) or other "common air pollutants"
whose source may be petroleum refineries, as apparently suggested by OIG.

       The footnote reference further compounds this mistake by implying that the "refinery
program" was directed to reducing emissions of all pollutants identified in the AirData system
(as noted above, it was not), that all such pollutant reductions are measurable (e.g., fugitive
emissions from multiple flanges, valves, and pipes are, by definition, not directly measurable),
and that it would be meaningful to measure such pollutant reductions at this point in the
implementation of the consent decrees.  With respect to this last point, the statement in the
footnote that it is not possible to estimate the extent that the Initiative "has actually reduced"
emissions (due to a lack of more current data in the AirData system), likewise indicates a
misunderstanding of the Initiative. Because this is written in the past tense, it suggests that the
Initiative is largely completed and emissions benefits are already realized. However, during the
course of OIG's investigation we pointed out on multiple occasions that many critical parts of
consent decree implementation are at their earliest stages, and that many emissions benefits that
will accrue under the decrees in the future have yet to be realized. These statements should be
corrected to reflect the bom the  purpose of the Initiative and its ongoing status.

       Additionally, the statement in the second to last sentence of mis paragraph that "EPA
develops regulations and OECA ensures compliance with these regulations alongside other
Federal agencies, States, and local authorities," seems to suggest that OECA is not part of EPA.
Moreover, it does not at all acknowledge the role of the Regions, nor of the role of the
Department of Justice in litigation matters. It should be revised to state that "OAQPS develops
regulations, and OECA, together with the Regions and the Department of Justice, ensure
compliance."

OIG Response: We modified the text as we determined appropriate.

Page 4, 3rd Paragraph:

      ' The opening sentence of this paragraph states that because of "the complexity of the
petroleum refining process ... OECA used experts in investigating refinery compliance." This
sentence repeats some of the problems already noted. First, this section of the report (as well as
many other portions) is written in  the past tense, implying that OECA's investigatory efforts
regarding refinery compliance have ceased.  In fact, EPA's investigation of refinery compliance
is active and ongoing. Second,  the reference to OECA's use of "experts" implies mat outside
experts were hired to investigate refineries. In fact, EPA (OECA, NEIC  and Regions) is
fortunate to have a top-notch national team of refinery experts in its employ.  One of the
remarkable elements of this Initiative, given its breadth, is the marshaling of in-house, cross-
regional expertise to investigate refineries. The text should be clear that EPA has used and
continues to use considerable in-house expertise (and made consistent with Chapter 2's
discussion of EPA's use of in-house experts).  This Evaluation should  also identify how EPA
developed its in-house "experts" - by enforcing PSD/NSR in the Pulp  and Paper industry,
conducting numerous BWN and LDAR inspections, taking enforcement actions on NSPS
compliance at sulfur recovery plants and against flaring, all at a time when those activities were
not encouraged.  The draft Evaluation tends to understate (or undervalue) the depth of
knowledge and range of experience required to become expert in this field.

                                      Page 11 of 68

-------
       Note that these vague and overbroad characterizations also illustrate some of the internal
inconsistencies in the report.  While the discussion in this section implies that the Initiative was
directed at all pollutants from refineries, the discussion in Chapter 2 notes the focus on certain
priority areas for enforcement. Similarly, while Chapter 2 elsewhere notes in passing that work
to implement changes at refineries required under the decrees is at an early stage, the basic
misunderstanding that Initiative is largely completed and emissions reductions have already been
realized is repeated in later chapters. These will need to be made consistent.

       The second column (SO^ of the last row on "Environmental Effects" should be checked
- SO2, in the form of acid rain, "travels and settles on ground and water, causing acidic streams
and lakes," etc.

O1G Response: We modified the text as we determined appropriate.  We disagree that the draft
report Chapter 1 had vague and broad characterizations. The information presented in the
chapter provides background on the various types of pollutants refineries generally emit and did
not imply that the "refinery program was directed to reducing emissions of all pollutants." The
report also describes the specific pollutants the national refinery compliance program  addressed.

Chapter 2

Page 7, Ist Paragraph:

       The reference in the third sentence, concerning OECA's use of "undated" planning
documents as "indicat[ing] an absence of strategic direction," is misleading.  It should be revised
as follows, to more accurately reflect both the process and the outcome:

       "While OECA's decision to implement a succession of tools and strategies evolved and
       was shared wilh the Regions, some of the early planning documents memorializing the
       use of the tools and the implementation of those strategies did not bear the imprimatur of
       formal approval, there is no evidence suggesting otherwise. To the extent that OECA's
       early planning documents did not indicate a strategic direction, this deficiency was
       corrected with the development of the Petroleum Refining Sector Strategy in 1998,
       reflecting an integrated approach (e.g., compliance assistance, focused investigations,
       enforcement and (later) compliance incentives) to address the most important or priority
       noncompliance problems under the Clean Air Act."

In any event, it is largely irrelevant whether the documents cited in the draft Evaluation were
signed or dated because, as explicitly acknowledged by OIG, an integrated strategy was, in fact,
developed for the Initiative. In addition, the following should be added as a footnote to the third
sentence:

       "This strategy was reflected in four documents: Petroleum Refining Sector Strategy
       (undated but issued simultaneously with its identified component sub-strategies) and
       Petroleum Refining Sector Strategy Components (April 6,1998); revised metrics for
       Petroleum Refining Sector Strategy Components (undated but issued in mid-to-late 1998
       as part of the NPMS pilot); and Petroleum Refining Sector 2000/2001 Strategy: executive

                                      Page 12 of 68

-------
       summary, components and outputs/outcomes (undated but issued in mid-to-late 1999 as
       part of the FY2000/2001 MOA process). Although we have no reason to believe
       otherwise, OECA was unable to provide documentary evidence as to when these
       documents were approved."

These revisions more accurately reflects what actually occurred, and is more consistent with the
second and fourth sentences of this paragraph, as well as with the "Conclusion" section of
Chapter 2.

       The fifth sentence, "OECA considers the program highly successful because, as of March
2004, OECA projected the program would result in annual reductions of about 44,000 tons of
NOx, 95,000 tons of SO2, and significant amounts of other pollutants," is written in the past tense
and implies that the Initiative is completed and emissions reductions already achieved; as such it
does not accurately reflect the current status. It should be revised as follows:

       "The National Petroleum Initiative is highly successful because, as of March 2004, and
       based on the companies' estimates of emissions reductions they will achieve under the
       terms of the settlements, OECA projects the program will result in annual reductions of
       about 44,000 tons of NOx, 95,000 tons of SO2, and significant amounts of other
       pollutants."

For the same reason, footnote 3 should be revised to state that OECA expects (not "expected")
that projected refinery emissions reductions will increase as the percentage of refineries covered
by consent decrees increases (not "increase").

OIG Response: We modified the text as we determined appropriate. As stated previously, in the
final report we deleted  the entire discussion on OECA's undated, unsigned, and unapproved
strategy documents because it served as just one example of the problems we found with
OECA's measurement  and reporting approach. We discussed with OECA officials at the exit
conference the importance of having dated, signed, and formally approved documentation in
order to use these documents as management tools. Chapter 2 adequately discusses the
management and reporting problems without the additional example of inadequate
documentatioa

Page 7,  1st Paragraph under "Evolution of Refinery Program ":

       The misleading reference to the use of "undated" planning documents (noted above) is
repeated in the fourth sentence, and for accuracy and consistency should be revised in
accordance with the above comment.

       The sixth sentence, "In addition, as the program evolved and OECA and regional staff
developed and implemented a variety  of tools and strategies (that later included compliance
incentives), OECA did not update or modify its written strategy documents," should clarify the
fact that the written strategy documents were not updated or modified "after issuance of the
Petroleum Refining Sector 2000/2001 Strategy in late 1999."  It is apparent from this statement
that OIG has not accounted for the role of the MOA in OECA's priority planning.
                                     Page 13 of 68

-------
       The seventh sentence, which states "In retrospect, it was easy to see that an integrated
program evolved despite the lack of updated and modified planning documents," is incorrect and
should be deleted.  In early 2000, OECA was heavily engaged in implementing the integrated
strategy. It would be factually wrong to suggest that the 1998 or 2000 strategies were not
integrated.  Moreover, the development of an integrated strategy was not accidental, as inferred
by the last two sentences of this paragraph, but was reflected in the inclusion of the global
strategy component in OECA's post-2000 priority planning documents - see MOA Guidance for
FY2002/2003 (June 2001), FY2003 (June 2002), FY 2004/2005 (June 2003) and FY 2004 (July
2004).

       In addition, the following should be added to the end of this paragraph:

       "The National Petroleum Refinery Initiative is one of the most successful enforcement
       initiatives undertaken by EPA. Since approximately January 2000, the date that EPA
       began to formally engage petroleum refining companies regarding their Clean Air Act
       non-compliance, EPA has obtained settlements with 11 petroleum refiners representing
       almost 40% of the nation's domestic refining capacity and covering 42 separate refineries
       for each of the major four substantive areas related to Clean Air Act compliance. The
       settlements contain substantial "beyond compliance" requirements, and altogether
       represents a breadth and depth of coverage not previously realized in the enforcement
       program."

OIG Response: We disagree that we did not account for the role of the Memorandum of
Agreement  in OECA's priority planning. We reviewed the Memorandum of Agreement as well
as other strategy documents. We discussed with OECA officials the priority planning process
and the information in the Memorandum of Agreement documents.  While OECA has stated that
the Memorandum of Agreement was used to help plan activities, our evaluation found that
OECA officials did not use the Memorandum of Agreement documents to manage the refinery
program, were not familiar with the information in the documents, or believed the information
was not entirely accurate.

We disagree with OECA's suggested revised paragraph. As stated previously, we cannot
comment on whether the national refinery compliance program serves as "one of the most
successful enforcement initiatives undertaken by EPA," since we did not compare the program
to all other enforcement programs conducted by EPA.

We modified the text as we determined appropriate.

Pages 7-8, under "Evolution of Refinery Program ":

       The basic argument in this discussion seems to be that because many documents do not
have dates,  there is an absence of strategic direction. If this is intended to suggest that there was
no documented management approval of the overall strategy, the draft Evaluation is incorrect, as
those exist and were provided as part of the Evaluation; OECA did not, however, document each
iteration of the evolving strategy. If the discussion in this section of the report is intended to
infer that OECA did not have an idea of what was to be accomplished strategically, the fact of
the strategy itself is evidence of the strategic direction for the refinery program. It is important

                                     Page 14 of 68

-------
to understand (and apparently overlooked by the Evaluation) that the strategy evolved over time,
from OECA's initial hypothesis to what was learned with field experience, which then informed
refinements over time of what the Initiative could and should accomplish. There is little to no
recognition of or credit given in the Evaluation to OECA's flexibility in incorporating "lessons
learned" as the strategy evolved; to have done otherwise (i.e., to establish a rigid strategy and set
of goals at the outset, and then fail to adjust those based on field experience) would have been
grossly inappropriate and discourage innovation.  The MO As reflect that experience and
basically serve as updates to the strategy. MOAs were agreed upon at the highest levels in
OECA. The four documents cited in the Evaluation reflect various stages and show revisions
over time as a result of field experience. However, the strategy cannot be looked at separately
but must be looked at in conjunction with the MOAs and MOA updates.  While the Evaluation
recognizes the MOA process, it fails to make the connection between the MOA and sector
strategy processes.

O1G Response: As stated previously, we have deleted the discussion on undated documentation
from the final report. However, as discussed previously, our evaluation found that some senior
OECA officials lacked familiarity with the information in the Memorandum of Agreement
documents and others believed the documents contained inaccurate information. The various
OECA officials responsible for managing and implementing the refinery program did not clearly
connect the Memorandum of Agreement and sector strategy processes, in Chapter 2, we
mention the problems with the Memorandum of Agreement documentation as part of OECA not
effectively managing the refinery program.

Page 8, 1st Paragraph under "Phase I":

      The last sentence states that "[i]n 1996, when compared to 496 other industries,  OECA
ranked refineries number one for releases of VOCs,...." This should be revised to reflect that
OECA's actual analysis was not done on an industry-by-industry basis, as inferred by this
statement, but on a sector-by-sector (3-digit SIC code) basis. It is correct to state that the
refinery sector ranked first among the 29 industry sectors ranked by EPA inl996 (see comment
Page i, Introduction).

OIG Response: We modified the text as we determined appropriate.

Table 2.1  and accompanying text (pages 8-10):

      Table 2.1 and the accompanying text describe the Initiative as comprised of "four distinct
Phases."  This is incorrect and should be revised. These phases are not, and have never been,
distinct. Refineries continue to be a national priority and compliance assistance continues, while
EPA's enforcement staff negotiates global settlements, prepares matters for litigation, and
implements the entered consent decrees. What are identified as the second through fourth phases
are steps taken to implement the priority.  Furthermore, the refinery  work is still ongoing, with
continuing investigations, efforts to pursue global settlements or litigation with additional
companies (as appropriate), and consent decree implementation. As drafted, this section
suggests that earlier "phases" have been completed. That this work is still ongoing needs to be
acknowledged in the report. More remarkably, it is from the same narrow but expert pool of
                                    Page 15 of 68

-------
personnel that EPA draws upon to perform the work associated with each of these overlapping
phases.

       As previously described to OIG representatives during the Evaluation, it is more accurate
to describe the national strategy as being comprised of four overlapping phases: a focus on
compliance assistance; a focus on investigations; a focus on global settlements; and a focus on
results and implementation. The compliance assistance phase involves considerable work, and
its importance cannot be overstated. It ensured national consistency on new source performance
standards (NSPS) applicability issues, included the issuance of, e.g., Enforcement Alerts,
Federal Register notices and the Storage Tank Emission Reduction Partnership Program
(developed in cooperation with the American Petroleum Institute), and established the detailed
approaches and legal theories that were later pursued in the investigations.

       The investigative phase, or what OIG characterizes as Phase II, was ongoing
simultaneously with the compliance assistance phase. At that time, investigations by the
Regions, were producing enforcement results and identifying common refinery problems.  These
investigations were also creating, within the industry, an understanding that there was
willingness by EPA to pursue violations of the Clean Air Act. Consequently, this phase of the
Initiative began with "on the ground" inspectors sharing knowledge and identifying common
problems; their results and expertise were then brought to management attention. The draft
Evaluation misreads this evolution by mis characterizing the Initiative as having been created
solely "top down," whereas the Initiative was in fact also a "bottom up" effort.

OIG Response: We modified the text as we determined appropriate, including clarifying that
activities under one phase overlapped with activities in other phases.

Page 8, 2nd Paragraph under  "Phase I":

       It is unclear what is intended by the fourth sentence: "As one of OECA's national
priorities, OECA expected regional office support, including the development of strategies to
address enforcement priorities." For clarity, it should be revised to state the following: "As one
of OECA's national priorities, the Regions were expected to develop their own strategies and
approaches to address enforcement priorities."

       The fifth sentence, "[i]n late 1997, OECA established a refinery compliance and
enforcement workgroup comprised of OECA and regional staff, to serve as a focal point for
sharing experiences, tools, and concerns, and to help develop strategies," is inaccurate,
misplaced and (as revised below) should appear in either Phase II (Focus on Compliance
Assistance) or Phase III (Focused Investigations, between the second and third sentence of that
paragraph):

       "In early 1998, OECA established a refinery compliance and enforcement workgroup
       comprised of OECA and regional staff, to serve  as a focal point for sharing experiences,
       tools, and concerns, and to develop a national petroleum refining sector strategy."

OIG Response: We modified the text as we determined  appropriate.
                                     Page 16 of 68

-------
Page 9, last sentence of "Phase I":

      The statement that the "designation of refineries as a national priority did not result in
increased refinery compliance" reflects a basic misunderstanding of the enforcement and
compliance program, as well as of the priority setting process.  At the time that this sector was
designated as a priority (circa 1995), OECA did not believe that mere designation of refineries as
a national priority, in and of itself, would result in increased refinery compliance. The
complexity of the noncompliance problems and expense of the solutions required investment of
considerable time, energy and resources to define and understand them; it is therefore unrealistic
to expect a substantial increase in compliance rates by 1997, as suggested by this discussion.
This expectation is also at odds with one of the Evaluation's recommendations (and is one of
many internal inconsistencies in the draft Evaluation) that OECA to spend time to "become
knowledgeable about the industry and its technical processes" at the outset of an Initiative.  (See
page 41, Lesson Learned #2.)

      Moreover, the expectation is that initially, by designating refineries as a national priority
and placing closer scrutiny on compliance issues  in the sector, the number of facilities detected
to be in noncompliance would tend to increase, not decrease. What the Evaluation fails to
consider is that it takes time for these facilities to return to compliance (as discussed in more
detail elsewhere in these comments). Therefore the last sentence in the Phase I discussion is not
meaningful, and should be deleted.

OTG Response: We disagree with OEC A's statement that we misunderstand the enforcement and
compliance program and the priority planning process. We understand that designating the
refinery sector as a priority would not, in and of itself, result in increased refinery compliance.
We clarified the statement to belter explain why EPA moved from Phase 1 (designation as a
national priority) to Phase 2 (using investigations to identify compliance problems).

Page 9, 1st Paragraph under "Phase II":

      The first sentence, "[although EPA identified refineries as a national priority in the
Agency's 1996/1997 Memorandum of Agreement Guidance, OECA officials said they did not
see significant improvement during 1996 and 1997" poorly conveys OECA's experience. A
better and more accurate way to characterize the introductory sentence of this section is that
"OECA officials did not see significant improvement in the approach to addressing compliance
issues in 1996 and 1997."

      The second sentence states, "[d]ue to the lack of regional progress, OECA began in 1998
to take a more comprehensive approach." However, as drafted this sentence is inaccurate and
unfairly criticizes the Regions for a lack of progress, and infers that there were no gains during
this time. In fact, NSPS and new source review (NSR) compliance issues were already
proceeding in cases such as Clark Cater Premcor) Blue Island, Illinois; Clark, Hartford, Illinois;
Murphy Oil, Superior, Wisconsin; and Shell, and Wood River, Illinois, among others. It was in
part the experience and insights gained in these efforts that indicated the need for a more
comprehensive approach. The sentence should therefore be revised as follows: "Because of the
breadth of the refinery sector and the complexity  of the technical issues, OECA began in 1998 to
take a more comprehensive approach."

                                     Page 17 of 68

-------
OIG Response: We modified the text as we determined appropriate.

Page 9, 2nd Paragraph under "Phase II":

       The statement that "the focused investigations only laid the groundwork for the global
refinery settlements that would follow" in the last sentence of this paragraph erroneously
suggests that mere is a direct cause and effect relationship between the investigations and global
settlements. While there is certainly an associative relationship, as described below, it is not as
direct as suggested by this discussion. Moreover, the level of effort to implement (e.g.,
developing new targeting and investigative tools) and the success achieved in this on-going
phase is nowhere mentioned or acknowledged.  This is an example of how the draft Evaluation
consistently understates the magnitude and challenges of the Initiative and of its
accomplishments.

OIG Response: We disagree that the report understates the magnitude and challenges of the
refmety program and its accomplishments. We believe the report adequately describes the
refinery program's challenges and accomplishments, as well as areas needing improvement. We
modified the text as we determined appropriate.

Page 9, 1st Paragraph under "Phase III":

       The discussion in this paragraph, that EPA decided to pursue global settlements with
refinery companies at the same time that two refiners approached EPA to settle noncompliance
concerns, is not accurate. In early 2000, OECA began to pursue national investigations of
selected refining companies; it was not until after EPA was approached by some of the
companies that OECA decided to pursue a global settlement approach.  As noted in the Report,
OECA had found success with this type of approach in other industries, but it should be
recognized mat OECA's prior experience related to a much narrower set of issues and on a much
smaller scale than attempted under the Initiative.

       The second sentence incorrectly implies that it was coincidental that "two of the largest
refining companies approached EPA" at or about the time that EPA decided to pursue global
settlements. The two refiners approached EPA because they understood that EPA was about to
focus its investigations on certain, unnamed refining companies by issuing each of those
companies information requests  under Section 114 of the Clean Air Act. At least some of the
companies approached EPA in an attempt to stop the issuance of those requests.

OIG Response: We modified the text as we determined appropriate.
Page 9, Td Paragraph under "Phase III":

      The third sentence indicates that OECA took the lead on all targeted investigations;
however, this is not true.  OECA coordinated its national experts and supported these
investigations, but it was not until OECA initiated national investigations against approximately
six refiners in early-to-mid 2000 that OECA took the lead and, even then, only on these national
                                     Page 18 of 68

-------
investigations. Moreover, the Department of Justice was and continues to be the lead in all
negotiations.

       The fourth sentence, stating that the "strategy presented corporate officials with the
option of settling out of court," erroneously suggests a cause-and-effect relationship between the
initiation of an investigation and achieving a global settlement. For accuracy, this sentence
should be revised to state: "OECA's approach presented corporate officials with the option of
avoiding possible investigation and litigation at one (or more) refineries on one (or more) issues
by pro-actively addressing all issues of concern at all its refineries."

       The fifth sentence incorrectly states that in the absence of a settlement, OECA would
otherwise continue to investigate all issues at all refineries. This sentence should be revised as
follows: "OECA would continue conducting investigations and taking enforcement actions as
necessary."

       OIG defines a "global" settlement as one that "applie[s] to all facilities owned by one
company." This is incomplete and, therefore, inaccurate.  The definition of "global" settlements
should be revised to reflect that they cover each of the four substantive areas of Clean Air Act
regulatory compliance that implicate petroleum refinery operations, and cover each of the
facilities owned by the refiner.

OIG Response: We modified the text as we determined appropriate.

Page 10, carryover paragraph, under "Phase III":

       This paragraph notes that "[s]tates that signed the consent decrees received a portion of
the penalties paid by the company." However, this is incomplete and reflects a lack of
understanding about State-Federal authorities.  For clarity and accuracy, the next to last sentence
should be revised to reflect that States that received a share of penalties had their own legal
claims against the settling refineries.

OIG Response: We modified the text as we determined appropriate.

Page 10, 1st Paragraph under "III":

       The first sentence of this paragraph begins, "[i]n fiscal year 2003, OECA shifted its
emphasis from pursuing negotiations with additional refinery companies to conclude on-going
investigations ...." For accuracy, the first sentence should be modified as follows: "Beginning
in fiscal year 2002, OECA began to shift its emphasis from pursuing negotiations with additional
refining companies to conclude on-going investigations and negotiations because 80 percent of
the domestic refining capacity universe had entered into global consent decrees, was in
negotiation with EPA, and/or under active investigation."

OIG Response: We modified the text as we determined appropriate.

Page 10,1st Paragraph of "Phase IV":

                                     Page 19 of 68

-------
        The mischaracterization that the four phases are "distinct" (noted above) is also evident
 in this paragraph. The first sentence should be modified as follows: "The implementation of
 refinery consent decrees began a new and additional phase of the refinery program."

       The third sentence states that there is anew process of "oversight, technical assistance,
 and enforcement of provisions." This is not necessarily accurate or universally true.  For
 example, consent decree enforcement is only necessary where terms of a decree have not been
 complied with.  Moreover, the sentence as drafted by OIG fails to account for a major aspect of
 the post-settlement process, which is a process of regular interaction and dealings among the
 parties.  To correct this error, it should be revised to state: "While the signing of a consent decree
 ends the settlement process for that company, it begins a new process of oversight and
 interaction by and between the parties."

       The last sentence of this paragraph erroneously suggests that inspections will occur
 "when the consent decrees and the current refinery program ended [sic]." The OIG misperceives
 both the Initiative and the enforcement program generally. This sentence should therefore be
 deleted, and the next to last sentence should be revised to  read as follows:

       "The implementation process continues at least until OECA ensures that companies have
       effectively implemented all consent decrees; concurrently with this process, EPA regions
       and States continue to inspect, monitor and take enforcement actions at these refineries
       and other industrial facilities as resources and other priorities allow."

 OIG Response: We modified the text as we determined appropriate.

 Page 10, last Paragraph:

 The last sentence of this paragraph,  which carries over to Page 11 and states that OECA used
 "compliance incentives to obtain the industry's interest in negotiating consent decrees and
 achieving compliance," is not accurate. In fact, it is not clear what OIG considers to be a
 "compliance incentive" in this context. The common understanding of a "compliance incentive"
 is not something that spurs interest in a settlement, but something that promotes voluntary
 compliance by regulated entities without the need for an enforcement response (e.g.,  NOV,
•administrative or civil judicial complaint, etc.). It would appear that OIG in its draft Evaluation
 has confused a "compliance incentive" (see, e.g., the listing of compliance incentives at
 www.epa.gov/compUance/incenti ves/index.h trnD with the incentive of a covenant-not-to-sue
 obtained in settlement. This sentence should be revised accordingly.

 OIG Response: We modified the text as we determined appropriate.

 Page 11, List of non-compliance problems:

 The list of non-compliance problems identified on page 11 ("New Source Review
 (NSR)/Prevention of Significant Deterioration (PSD); Flaring Incidents in Violation of New
 Source  Performance Standards (NSPS); Leak Detection and Repair (LDAR); Benzene Waste

                                     Page 20 of 68

-------
National Emission Standards for Hazardous Air Pollutants (NESHAP)"), and the accompanying
text is too narrow and therefore tends to understate the magnitude of the compliance issues
associated with those areas.  This is a problem repeated throughout the Report.  Furthermore, the
Report's discussions of these issues fail to identify that a major portion of each of the consent
decrees addresses refinery non-compliance with the "new source performance standards for
sulfur recovery plants and fuel gas combustion devices."  m fact, with the exception of the
installation of controls to meet NSR requirements at fluidized catalytic cracking units and
heaters and boilers, compliance with these obligations represents the largest capital investment
towards compliance made by a company under the consent decree. The requirements for
company compliance in these areas are plainly set forth in each of the eleven refinery consent
decrees on a refinery-by-refinery basis, and should be included in this Evaluation.

       In addition, the Report states that "inspections" and "formal EPA information requests"
shaped the investigations during the initial stages of the program.  However, as drafted OIG's
characterization omits a crucial factor: the experience and expertise of those that used these tools
during the initial stages. Research informed our targeting decisions, particularly for NSR/PSD,
but it was the experience of EPA's in-house expertise that informed the selection of the priority
areas, or "marquee issues": Ken Garing for LDAR and benzene; Panic McCoy  for NSPS and
flaring; and Pat Foley and others on NSR/PSD.

OIG Response: We modified the text as we determined appropriate.

Page 12, Figure 2.1:

       "Flaring" Box:  This box should be re-labeled as "New Source Performance Standards."
The box should include bullets for compliance at flares, sulfur recovery plants,  and fuel gas
combustion devices. Also, the "Solution" bullet is far too narrow and understates the
compliance issues, which indicates a lack of appreciation for the innovative and extensive nature
of the work required under the decrees. While the flaring protocol requires a company to
identify the root cause  of a flaring incident and to "outline plans" to EPA that addresses mat root
cause, it also requires that the plans be fully implemented and that the conditions giving rise to
the flaring incident (i.e., the root cause) be  addressed either through capital investment
(installation of new equipment) or institutional controls (revise standard operating procedure,
training, etc.). This is an example of the sort of "beyond compliance" measures obtained under
the Initiative that are overlooked, or simply not understood, throughout the report.  In addition,
consent decrees require compliance at all times for all sulfur recovery units within the plant
New sulfur recovery units and tail gas control devices are installed to ensure compliance with
emission standard and good air pollution control practice obligations. As drafted, the "solution"
suggests that analysis and planning -  but no implementation or controls - are required for
flaring.

       "LDAR" Box:  Similarly, the LDAR box fails to clearly identify the program as
"enhanced" - i.e., bringing a company beyond compliance with the applicable regulations.

       "Benzene" Box: As above, the benzene box fails to clearly identify the program as
"enhanced" - i.e., bringing a company beyond compliance with the applicable regulations. In
                                     Page 21 of 68

-------
addition, the parenthetical in the "problem" bullet identifies only two types, out of myriad of
potential types of waste streams. That parenthetical should use the signal "e.g.," to denote that
me two identified wastestreams are only examples.

       "NSR/PSD" Box:  The text in this box appears to emphasize monitoring of unit
performance as the most significant gain achieved under the NSR/PSD elements of the consent
decrees. While the significance of real time monitoring cannot be overstated, far and away the
most significant aspect of the companies' agreements to comply with NSR/PSD is the suite of
controls to be installed and implemented to reduce sulfur dioxide, nitrogen oxide, and particulate
matter emissions. This is not clearly conveyed by the statement that "new emission controls" are
required under the decrees, and therefore this should be revised to emphasize the controls aspect
of the NSR/PSD  programs that the companies have committed to implement.

OIG Response: We modified the text as we determined appropriate.

Page 13, ^Paragraph:

       The first sentence, which recites that "some" OECA officials characterized the
compliance assistance efforts as effective in getting the refinery industry's attention, while
"other" OECA officials said they did not see significant improvement, should be revised to note
that both OECA and Regional officials did not see significant improvement as a result of
compliance assistant efforts with this sector. This is consistent with the experience of the earlier
failure of the Common Sense Initiative to encourage better levels of performance with this
industry, and tends to support the conclusion that it was the initiation of a concerted enforcement
effort mat got this industry's attention. (See comment, page 8 Table 2.1.) The contrast between
this and the majority of the industry's positive attitude and good faith negotiations exhibited
within the Initiative stands in stark contrast.

OIG Response: We modified the text as we determined appropriate.

Page 14, 1st Paragraph:

       The discussion in this paragraph, concerning a "release" provided in settlement, and
"certainty" provided for "future potential liability" obtained under the consent decrees
(mischaracterized as "compliance incentives," repeating an earlier error - see comment page 10,
last paragraph), represents a fundamental misunderstanding both of enforcement in general, as
well as of what motivates companies to negotiate with the United States to resolve potential
liabilities. In particular, this paragraph mischaracterizes the nature and scope of the "covenant
not-to-sue" (inaccurately called a "release" in the report) granted by the United States.
Specifically:

•      Companies typically negotiate with the United States to avoid a potential federal civil,
       judicial enforcement action(s). The quid pro quo for reaching agreement with the United
       States is not a "release" but rather a "covenant" by the United States not to pursue the
       company in a separate administrative or judicial proceeding for the activities giving rise
       to the consent decree - i.e., the original violations that pre-date the consent decree.


                                      Page 22 of 68

-------
•      While it is true that it is unlikely that the United States would pursue a company for
       activities addressed by a consent decree, the United States always expressly reserves its
       statutory right to do so. Consequently, the United States has never agreed in the refinery
       consent decrees to resolve-the "future potential liability" of a company provided that it
       "remained in compliance with [its] consent decree requirements" as stated in the Report.

•      Finally, it is important to emphasize that the breadth of the United States' covenant not-
       to-sue is directly related to the scope of the injunctive relief.  Thus, as here, where a
       company agrees to implement a comprehensive program of injunctive relief that brings it
       well beyond compliance with all aspects of the NSPS and benzene leak detection
       regulations, the United States will extend a covenant not-to-sue regarding all aspects of
       that company's pre-consent decree compliance with those  regulations.  To the contrary,
       however, under the consent decrees a company receives a  covenant not-to-sue for
       NSR/PSD and new source performance standards for only those emission units that are
       specifically addressed by the consent decree.

OIG Response: We disagree with OECA that we misunderstand enforcement in general and
what motivates companies to negotiate with the United States to resolve potential liabilities. We
appreciate OECA providing us with the legal definition and description of "covenant not-to-
sue," and have included it as a footnote in the report.  We believe  that our use of the term
"release" still accurately describes this  aspect of the national refinery program and is the same
term that OECA officials used in discussions with us and in OECA's 1998 documented draft
national strategy.

Page 14, Table 2.2:

       This table, listing the names companies and number of refineries under "global"
settlements, is not accurate. It does not reflect the purchase and sale of refineries that are
covered by global settlements to other refiners. For example, BP Exploration has sold three of
its refineries, and while those refineries remain subject to the global  settlements they  are owned
by separate refining entities.  Consequently, this table cannot be reconciled with the information
contained in Appendix B. This could be addressed by a new heading or footnote that indicates
that the information in Table 2.2 represents the settlements on the date of their entry by the
Court.

OIG Response; We modified the text as we determined appropriate.

Page 15, Table 2.3:

       This table, describing "examples of consent decree requirements," needs to be revised for
consistency, accuracy and completeness. For consistency with earlier comments, the entry on
"flaring" should be re-labeled as "New Source Performance Standards." For accuracy and
completeness, the box should note that  the program relates to NSPS compliance at a refinery's
flares, sulfur recovery plants, and fuel gas combustion devices.  In addition, the box should be
expanded to include the control at and emissions monitoring of sulfur recovery plants.
                                      Page 23 of 68

-------
       The description of "flaring" in this table - and, in fact, throughout the report - is under-
representative. As a threshold matter, what is covered by "flaring" is not at all described.  For
accuracy, the "Flaring" box should be relabeled "NSPS SRP/ Flaring." Further, the general
discussion of "flaring" throughout the draft Evaluation fails to recognize that there are two
different flaring programs: the elimination of acid gas flaring; and the reduction (quasi-
voluntary) of hydrocarbon flaring. In addition, the NSPS component of this effort that are
devoted to ensuring that companies are NSPS at their sulfur recovery plants, flares, and fuel gas
combustion devices. The success of this effort is nowhere mentioned.  Under the entered
consent decrees there are now 43 sulfur recovery plants subject to NSPS Subpart J. The 43
sulfur recovery plants contain a total of 99 Claus trains.  Seven new Claus trains have or will be
installed under the consent decrees and 22 tail gas control units have or will be installed under
the decrees.

O1G Response: We modified the text as we determined appropriate.

Page 15, 3rd Paragraph:

       The discussion in this paragraph concerning "the  roles, responsibilities, and processes for
implementation of consent decrees," displays a lack of understanding about the fundamentals of
the enforcement process. Contrary to the assertion made in the fourth sentence that "[a]fter a
company signs a consent decree, the implementation phase of the consent decree begins," the
implementation phase of a consent decree does not in fact begin at company signature. The
implementation phase begins either on the date of consent decree lodging or entry depending on
Ihe specific language of the consent decree.  Typically, the vast majority of consent decree
obligations run from the date of entry of the consent decree.  Similarly, the statement in the third
sentence that "unsuccessful negotiations result in enforcement actions" is simply wrong.  It is
more accurate to state that in the event that negotiations break down, EPA would resume its
investigation, marshal its evidence and take whatever enforcement followup is appropriate in
light of the relevant facts and circumstances.

OIG Response: We modified the text as we determined appropriate.

Page 16, 1st Paragraph (carryover from Page 15):

       The fifth sentence provides, "[e]ach consent decree describes whether a company report
or action requires EPA review and approval." This is not accurate. For clarity and to avoid
confusion, this sentence should be revised to read as follows: "Each consent decree describes
whether a company report or action requires a formal EPA approval."

OIG Response: We modified the text as we determined appropriate.

Page 16, 1stParagraph under "Conclusion":

       The conclusion that OIG attempts to draw here regarding the absence of updated
documents, mat "[although OECA staff did not update its written strategy documents as the
                                     Page 24 of 68

-------
program evolved, the refinery program resulted in an integrated strategy," is both unsupported
and unsupportable.  OIG misconceives some fundamental aspects of the overall OECA priority
planning process, and of how the Initiative fits into that process.  While OECA did not continue
to produce written strategy documents signed by upper management, it should be recognized that
these strategies were pulled into the OECA MO As, and at that point OECA used the MO As to
manage the program. The MOA commitments and measures are approved at the upper
management level.

OIG Response: As described earlier, we deleted the discussion on the undated, unsigned, and
unapproved strategy documents. We disagree with OECA that we misconceived some
fundamental aspects of the priority planning process and how the national refinery program fits
into that process. As described in earlier sections, we reviewed the Memorandum of Agreement
documents and discussed the information in the documents with OECA officials.  Despite
OECA's assertion that they used the Memorandum of Agreement documents to manage the
program, OECA officials told us that they were not familiar with the information in the
documents or that the information was not entirely accurate.

Chapter 3

       As a general matter, Chapter 3 continues to focus on the strategy document, and ignores
the fact that OECA used the MO As to manage the refinery program. This failure to comprehend
the larger process leads to several erroneous conclusions, and internal inconsistencies in the
Evaluation, noted more specifically below. (See, e.g., comments at page 17,1st Paragraph; Page
20,1st Paragraph under "Clear Program Goals"; Page 21, Table 3.1.)

       This chapter reflects a fundamental misunderstanding and lack of appreciation for the
scope, innovation and good management practices (e.g., application of quality management
principles, demonstrated flexibility in adjusting to new challenges and information, lessons
learned, etc.) employed in the Initiative.  OIG assumes that only senior management can make
appropriate, informed judgements that must be formally approved and documented, and misses
the fact that the expert, multi-Regional team charged with implementation was also empowered
to make adjustments to the strategy.

       The draft Evaluation does not recognize or properly appreciate that the Initiative has been
a laboratory and driver for OECA innovation and has helped to stimulate demand for more
advanced emission  control technologies at refineries. Examples of some of these include:

•      Among the first integrated strategies developed - and, more significantly, that it was
applied to an entire industry sector on a nationwide basis, and that it addressed multiple
compliance issues simultaneously;
•      The Initiative was a driver for implementing the National Performance Measures Study,
       resulting in improved, better defined initiative outputs and outcomes;
       One of the first CAP-type program with identified, pre-approved compliance
       technologies;
•      Represents one of the most significant uses of cross-Regional, in-house national experts
       (a model for future national workforce planning efforts);


                                    Page 25 of 68

-------
•     The Initiative's push for "beyond compliance" by settling companies has served as a
      driver for the development of newer and better pollution control technologies;
•     The use of an innovative information collection process and related reporting mechanism
      that has enabled near real-time tracking of performance (e.g., monthly conference calls
      and regular status updates with the "consenter's group");
•     The use of electronic reporting by affected companies; and
•     The first internet-based consent decree tracking tool ("Navajo pilot" by Region 6).

      While not an exhaustive list, these and other elements of the Initiative, many attempted
for the first time or for the first time on this scale, are either unrecognized in the draft
Evaluation, or are noted only in passing.  As a result, the draft is unbalanced in its assessment,
discounting or overlooking the Initiative's scope, accomplishments and innovative use of new
approaches.

OIG Response: We disagree that we  ignored the Memorandum of Agreement documents.  As
stated previously, we evaluated Memorandum of Agreement documents for fiscal s 1996 through
2004 and discussed the information with OECA officials. During our discussions, we learned
that OECA officials did not use the Memorandum of Agreement documents to manage the
refinery program. In addition, we asked OECA staff and management for any and all
documentation of strategy planning, goal-setting and measurement. OECA officials never
mentioned Memorandum of Agreement documents in those discussions, mentioned them briefly
in passing, or referred to them as "inaccurate".

We disagree  with OECA that  the chapter reflected a misunderstanding and lack of appreciation
for the scope, innovation, and good management practices used in the program. We also
disagree with OECA that we assumed only senior management could make appropriate,
informed judgments. We believe that because OECA led the refinery program, senior OECA
officials maintain responsibility and accountability for ensuring that decisions regarding the
strategy are documented and clearly  understood.

We disagree  that the report is  unbalanced and discounts the refiners' program's scope,
accomplishments, and innovative use of new approaches. We describe the strategy's
accomplishments in the report, including some that OECA officials included in the list above.
However, we also reported areas needing improvement that OECA should address to ensure
timely emissions reductions and to optimally protect human health and the environment.

Page 17, 1st Paragraph:

      In Chapter 2, the Evaluation recognizes that OECA worked with the Regions to develop
the Refinery  strategy and its goals. However, the very first paragraph of Chapter 3 states that
OECA did not "precisely define official program goals and measures, or ensure the goals were
clearly and consistently shared with everyone working on the refinery program...." These two
statements are inconsistent and cannot be reconciled. As discussed earlier, the strategy was
developed with the Regions and had  Regional "buy-in." Once the sector strategies were
integrated into the MOA process, the MOA was the document where revisions to the strategy
and goals were made.
                                     Page 26 of 68

-------
OIG Response: We disagree.  OECA did not clearly and precisely define official program goals,
or ensure everyone working on the national refinery program had the same understanding of the
goals. We believe that despite OECA's claim in its response that it used the Memorandum of
Agreement documents to manage the program, our evaluation disclosed the opposite and that
staff referred to different goals for the refinery program.

Pages 18-19, under "Performance-BasedProgram Management":

      The discussion in this section assumes that the Initiative was solely a "top down"
priority. This is not correct.  As noted earlier (see comment page 8, Table 2.1), the evolution of
the Initiative was both "top down" and "bottom up." OIG therefore fails in its Evaluation to
take into account the evolution of OECA's learning and understanding of refinery operations that
was developed during the early years of the Initiative, and the effect of liability exposure and
litigation risk (which directly bears on the identification of goals and objectives).  In this
discussion, OIG appears to believe that achievable outcomes at refineries were known in
advance, and that injunctive relief is simply a "cut and paste" from one uniquely designed and
operated refinery to the next (another example of a lack of appreciation for and understanding of
both the enforcement process and the complexity of refineries and their associated compliance
issues).

OIG Response: We disagree.  We included this section to give OECA credit for moving toward
developing a performance-based strategy in fiscal 2004 designed to address many of the issues
we raised. We believe that OECA can develop outcomes at the beginning of a program and
modify them as necessary. As described in the report, OECA needed to develop clear goals for
the national refinery program, related performance measures, and measurable outputs and
outcomes. OECA's comments to the official draft report contradict an internal OECA December
18,2002, report, Recommendations for Improving OECA Planning, Priority Setting, and
Performance Measurement. The report recommended that strategies should include a goal or set
of goals and performance measures that allow progress to be assessed and these elements should
be in place before the implementation period begins.

Page 19, 3rd Paragraph:

      As discussed above, the statement in the second sentence that "OECA did not establish
consensus on outcome-related goals necessary to  adequately assess the progress of the national
refinery program" is incorrect. The Petroleum Refining Sector Strategy established outcome-
related goals and output measures for assessing progress under each of our marquee issues.
Identified output measures were used by the Regions to establish their MOA commitments in
1998-2000; these measures were then used by OECA to track performance under each marquee
issue: NSR/PSD investigations (Petroleum Refining Sector Strategy Component 2, Output 1 and
2); NSPS/flaring investigations (Petroleum Refining Sector Strategy Component 3, Output 1 and
2); LDAR investigations (Petroleum Refining Sector Strategy Component 4, Output 1,2,3 and
4); and benzene investigations (Petroleum Refining Sector Strategy Component 5, Output 1).
Information on each investigation was regularly collected, routinely shared, and informally
reported to enable near real-time tracking of Regional activities in support of the strategy. This
information also informed the providing of feedback on the conduct of these investigations (e.g.,
                                     Page 27 of 68

-------
problems, solutions and lessons learned by and between the team and its designated national
expert).

       The strategy's overall goals of 20% reduced emissions and 50% improved compliance
were to be measured against a 1997 baseline, using a 2000 accomplishments period (AIRS and
SNC data). However, it was soon recognized that the effort needed to perform effective
investigations to support reaching these goals would necessarily extend beyond 2000 and that the
level of major noncompliance (e.g., the NSR/PSD violations that are only determinable through
extensive file reviews and investigations) found was not adequately captured by SNC data
(which is largely based on traditional inspection activity). Therefore, information collection and
reporting was .tailored to track: referrals; probable referrals (i.e., preliminary indications of major
noncompliance); active investigations (i.e., initiated but no preliminary indications determined)
and inconclusive investigations. (Note that the fact that an investigation is inconclusive does not
mean the facility is in compliance, only that major not noncompliance was not determined, hi
many instances, significant (but not major) noncompliance was found and addressed by the
Regions under "inconclusive" investigations. OECA used this information as the principal tool
for tracking Regional progress under their MOA commitments through 2001/2002 because year-
end reporting by the Regions failed to provide this information.)
Page 20,1stParagraph under "Clear Program Goals":

       As an initial matter, this section contemplates that OECA can micro-manage decision-
making by each Region and State, as exhibited by statements such as OECA's inability to.
"ensure that everyone working on the refinery program in EPA headquarters offices, EPA
regional offices, and State offices" was working toward the same goals. This assumption fails to
take into account uniquely local issues and perspectives mat almost certainly impact refinery
compliance, as well as the widely divergent views that the Regions and States have regarding the
best approach for addressing refinery compliance. The MOA and priority setting process
(consistently misunderstood  or overlooked by OIG in the report) is intended to help address
these considerations.

       This discussion and critique in this section (that OECA did not clearly and precisely
define official program goals, or ensure that everyone working on the refinery program had the
same understanding of the goals) misses the mark by failing to understand the role of the
MOA/priority planning process, and how that work is executed. Every individual need not have
an understanding of each goal, provided that they-understand goals that apply to their work.
EPA Headquarters and the Regions refer to the MOA language, as the implementation of the
strategy was accomplished primarily through the MOA process. Accordingly, this paragraph is
incorrect and should be revised. OECA established clearly defined program goals (20% reduced
emissions and 50% improved compliance) in 1998 and restated these goals  for FY2000/2001.
The strategy remained unchanged and subsequent MOA guidance referred to it.

       OIG suggests consideration of a logic model  (Appendix F) as a design for the refinery
program. In discussions with OIG staff, OECA's Office of Compliance staff did not agree that
several of the short-term, intermediate and long-term outcome measures are appropriate
benchmarks for judging the effectiveness of a compliance and enforcement program. Goals such
as increasing flexibility for refineries to expand or upgrade operations, enhancing the
                                     Page 28 of 68

-------
environmental ethic in companies, improving the relationship/trust with the regulated
community, while perhaps laudable in theory, are not realistically likely to be obtained in an
adversarial enforcement context. Curiously, recommendations for how OECA would benchmark
the "before" conditions and measure changes over time for these ultimate outcomes are not
addressed in OIG's draft report. Absent any guidance from OIG, the suggestion is not practical
and there is no basis for OECA to determine whether these measures are feasible.

OIG Response: We disagree with OECA that it established clearly defined goals. As described
in our report, different OECA officials referred to different goals and measures for the refinery
program. Some OECA officials were not aware of the information when we discussed the
Memorandum of Agreement documents with them. Specifically, two top OECA officials
responsible for managing the refinery program were not familiar with the "20 percent reduced
emissions and 50 percent improved compliance" goals. These senior executives described
entirely different goals for the refinery program.

We deleted the logic model from our report since it represented just one of several possible
means that OECA may employ to achieve the ends we advocate in the report's
recommendations - that is, the agreement on and communication of program goals.

Page 21,  Table 3.1:

       This table, listing different articulations of refinery priority goals, creates unnecessary
confusion. As arrayed in the table, the entries create an impression that OECA's objectives
routinely shifted over time. However, the goals as paraphrased in the table not put in context,
which obscures their meaning. The first row relates to the initial designation of refineries as a
priority in FY96 and FY97 {i.e., reflecting planning done in 1995), which was prior to the
development of the integrated strategy, notes that the effort at that time was focused on
developing compliance strategies for the sector. Based on the experience gained during this
time, the problems and compliance challenges were better understood than when the sector was
initially designated a priority,  and resulted in a more robust Refinery Strategy starting in FY98.
This included the development of goals and the use of an integrated strategy (noted in the second
and third rows of the table). The fourth row, taken from OECA's 2001 Accomplishments Report,
is simply irrelevant, and indicates that OIG has failed to grasp some very elemental aspects of
OECA's program. Accomplishments Reports are written for an external (general public)
audience, and are not  used for internal OECA planning, although the Accomplishment Report
expresses the same themes of reduced emissions and improved compliance. OIG's mixing of
these separate documents prepared for different purposes and  for different audiences indicates a
lack of understanding of OECA's program,
                                                                                    "\
       Missing from the table entirely are references to OECA's documents that were relied on
for planning, specifically MO A language from the 1999 MO A update, and all subsequent
versions of Hie MOA. This is necessary to properly understand the goal for each MOA cycle as
the petroleum refining priority progressed. OECA adapted the goals and strategies as more was
learned, including what was working and what was not. The table therefore inaccurately
conveys program goals by inappropriately mixing references to planning and non-planning
documents, and omitting other planning documents entirely.  The table should be deleted.
                                     Page 29 of 68

-------
OIG Response: We disagree that the table should be deleted.  The table provides a good
illustration of the different goals for the national refinery program.  As we reviewed strategy
documentation (including the Memorandum of Agreement documents) and discussed refinery
program goals  with OECA and regional staff, we found that the program had various goals over
the course of the last 8 years. The descriptions of the goals varied by those we interviewed as
well as the various documents we reviewed. The table portrays what we found during our
evaluation - confusion over the goals, various descriptions of the goals, and a change in the
goals as the program evolved. While we understand that a program's goals may change over
time, at any one point in time, everyone involved should have the same understanding of the
goals. We found that was not the case.  We believe the refinery program should have clearly
defined goals.

Pages 20-21, "ClearProgram Goals":

       The last sentence of the second paragraph, "[i]n addition, the document has milestones
for fiscal years 1998 and 1999 and OECA did not update it with current objectives, measures,
and time frames," is incorrect As discussed above, OECA updated its national strategy in
calendar year 1999 to clarify that the 50% improved compliance goal was for "targeted [a.k.a
marquee] issues."

       The third paragraph erroneously equates the 2004 logic model with the Petroleum
Refining Sector Strategy.  The strategy identified two long-term outcome measures: reduced
emissions and improved compliance.  Other identified, logical outcomes reflected in Appendix F
have not yet been identified in the strategy or MOA Guidance, and related metrics/measures
have yet to be identified by OECA. In other words, to the extent that the draft Evaluation is
intended to assist in this and future Initiatives, the draft Evaluation misses the mark.

       With respect to the fourth paragraph on Page 21 concerning the use of company baseline
emissions data, OECA does not question the wisdom of using company baseline information to
measure success and progress under the consent decrees, provided that the baseline emissions
data is determined using accurate and precise emission measurement methods and techniques.
EPA and the refining companies have not relied, and should not rely, on emissions inventories as
a basis for establishing the baseline by which to track consent decree performance.

OIG Response: As discussed earlier, we deleted the logic model.

Page 22, 1st Paragraph under "Existing Measurement and Reporting Systems Were Ineffective":

      The first sentence states that "EPA performance measurement and reporting systems
were ineffective for monitoring or reporting refinery program performance," and as a result
criticizes OECA as being unable to "assure anyone, including itself, that companies complied
with regulations or with specific consent decree requirements	"  This is an erroneous
conclusion, without an accurate premise. The draft Evaluation errs by not acknowledging that
existing EPA systems were not, in fact, used for monitoring activities under the Initiative.
Instead, OECA created tools specifically tailored for managing performance by the Regions and


                                     Page 30 of 68

-------
implementation of the Initiative on a near real-time basis. Rather than state that the existing
systems were "ineffective," it is more accurate to state that they were "not used."

       OECA also strongly disagrees with the statement in the second sentence that "EPA could
not assure anyone, including itself, that companies complied with regulations or with specific
consent decree requirements in three of the four priority areas." This sentence should be struck
from the final report because it is not only inaccurate, but misconceives the enforcement and
consent decree process, the obligations of refineries to comply with applicable regulatory
requirements, with or without a consent decree, and suffers from the same "past tense"
perspective noted earlier. In particular:

       •      The terms of the consent decrees do not require that a company comply with
              applicable regulations as a matter of consent decree enforceability. That
              companies are required to comply with applicable regulations is required by the
              regulations themselves, and refining companies obligation vis-a-vis applicable
              regulations is the same as any other entity, notwithstanding the fact that the
              refining company is the subject of a consent decree;

       •      Under the terms of the consent decrees refining companies are repeatedly
              certifying compliance with the consent decree obligations. This provides the
              requisite assurance, particularly in light of the serious consequences for providing
              false information (fraud), that obligations are being met;

       •      The use of the past tense implies that the obligations and actions required under
              the consent decrees are already in place, whereas they are in fact at an early stage;
              and

       •      Under the terms of the consent decrees, it is the company's obligation to notify
              EPA when it is unable to meet a consent decree schedule or requirement.
              Therefore, it is perfectly reasonable for EPA to assume that the refining
              companies are meeting their consent decree obligations.  In fact, there are several
              instances where a refiner failed to meet a consent decree obligation in which it
              notified EPA and the non-compliance was addressed.

       Nevertheless, OECA does review reports and is in regular communication with each
company regarding its compliance with its consent decree.

       The third sentence of this paragraph, stating that "EPA was also unable to verify
emissions reductions in any priority area," is likewise inaccurate and misconceives the nature,
timing and impact of obligations under the decrees. On a number of occasions during its
investigation, OECA advised OIG's representatives that benzene and LDAR emissions
reductions under the consent decrees are difficult to quantify (i.e., verification is by means other
man direct monitoring, as implied by the Report).  This sentence (like many others in the Report)
is also written in the past tense, inaccurately suggesting that controls still to be installed under
the long-term schedules  under the decrees have already been implemented. In particular:

       •      First, emissions reductions under the benzene and leak detection programs

                                      Page 31 of 68

-------
             address refinery-wide fugitive emissions from thousands of emission points
             throughout the refinery, making it difficult to directly measure emissions
             reductions. This is not a shortcoming of the Initiative or of the obligations under
             the consent decrees to better control these emissions, as implied by the Report,
             but inherent in the nature of refineries themselves.

       •      Second, to date, it is too early in the consent decree implementation process to
             begin to measure consent decree emission benefits from NSR/PSD controlled
             units because the vast majority of the dates by which a company is required to
             install and the required controls has yet to pass.  As those controls come on line,
             EPA will be assessing the performance of the controls and the company's
             emissions from those units.

       •      Finally, the nature of flares and flaring incidents make it difficult (if not
             impossible) to measure directly emissions from flares either qualitatively or
             quantitatively. Nonetheless, the consent decrees require defendants to report each
             flaring incident, identify the pollutants flared, and the amount of that pollutant
             that was released. This is another example of where the report fails to
             acknowledge an important "beyond compliance" requirement of the consent
             decrees, that are not otherwise required by the regulations.

OIG Response: We modified the text as we determined appropriate. In particular, we modified
the report to delete the discussion on the systems that OECA did not use, clarify the systems that
OECA did use, and identify the areas needing improvement in these systems.

We do not imply that consent decrees require compliance any more than the typical regulatory
structure requires compliance.  However, the refinery program evolved because EPA found that
companies knowingly did not comply with regulations. Therefore, we maintain that verifying
compliance with refinery consent decrees should remain one of EPA's top priorities. Given the
characteristics and history of the refining industry, OECA should provide the Agency1, Congress,
the public, and other refiners with a high degree  of assurance that companies are complying with
consent decree obligatioas. We believe that self-reporting in tins sector does not provide
sufficient assurance that facilities are in compliance.

Page 22, Tdparagraph, Table 3.2:

       There are several aspects of this table that need to be revised to ensure its accuracy.
Specifically,

       NSR/PSD Priority Area. "Current Compliance Status" column: The entry should be
revised to read: "In compliance and/or on a compliance schedule." In addition, given the scope,
complexity, and multiple timetables across all facilities and consent decrees, it  is not possible to
generalize regarding the compliance status of 11 consent decrees as it relates to 42 separate
refineries. The entry in this column should be modified to reflect the fact that dates for consent
decree control obligations in many instances have yet to pass.  The sentence should be modified
to state as follows: "Many consent decree obligations are outstanding, including a number that
are not yet required to be implemented under .the terms of the consent decrees."

                                     Page 32 of 68

-------
       NSR/PSD Priority Area. "Emissions Status" column: The reference that emissions status
of the subject units is "unknown" is not accurate.  To the extent that a unit has not had consent
decree controls installed, its emissions status is reflected by its reported baseline, quarterly
emissions reports, and reporting under Title V (where applicable).  To the extent that this column
is intended to reflect performance of the controls required by the consent decrees, as noted
above, in many instances the dates for installation of controls has yet to pass and therefore there
is no relevant data or information to report.  This column should accurately reflect the foregoing.

       NSR/PSD Priority Area. "Issues Associated with Measures" column: This comment does
not accurately reflect how compliance with NSR/PSD requirements is measured. As noted
above, certification regarding compliance with applicable consent decree requirements occurs on
a regular basis. Under the terms of the consent decrees, the refiners submit regular reports
regarding the status of compliance-related activities, including their compliance with consent
decree requirements related to installation of fluidized catalytic cracking unit controls and
controls for heaters and boilers. Beyond the regular reports, the consent decree also requires the
settling refineries to notify EPA when they anticipate they will be unable to meet a consent
decree requirement or if they have in fact failed to meet a consent decree requirement
Consequently, OECA is verifying compliance with consent decree requirements on almost a real
time basis.  It is not clear in the report (nor is it at all discussed) the basis for the conclusion that
overall consent decree compliance is  assessed only once every four years. This statement is
incorrect.  It should be noted that a common element of the heater and boiler compliance
program is to report interim compliance in the fourth year of that eight year program.

       Flaring Priority Area. "Current Compliance Status" column: As discussed earlier, this
issue should more properly be identified as "NSPS/Flaring" to capture the work required to
ensure new source performance standards compliance at sulfur recovery plants and fuel gas
combustion devices. This entry should be revised to read: "In compliance and/or on a
compliance schedule."

       Flaring Priority Area. "Emissions  Status" column: As noted in the comments on the first
paragraph of this page (above), the nature of flares and  flaring incidents make it difficult if not
impossible to directly measure emissions from flares either qualitatively or quantitatively.
Nonetheless, the  consent decree requires that a company report every flaring incident, identify
the pollutants flared, and the amount  of that pollutant that was released. The settling refiners
regularly report mis information to EPA.  OIG representatives were advised as part of the
investigation that OECA is currently  engaged in a project to assess flare performance time at
individual refineries, company-wide,  and  across the industry. It is also important to understand
the nature of the noncompliance concern with flaring (improper use on a "routine" basis, as
opposed an allowed "episodic" basis  only), and the beyond compliance nature of the actions
required under the decrees. Furthermore, this column also exhibits some of the internal
inconsistency noted earlier - it is not  possible to reconcile the "unknown" reference in this
column with the statement in the "Issues Associated with Measures" column that "[fjrom the
reports, OECA can estimate the pollution resulting from the flaring event." In addition, since
"5-year flaring histories" from each company are submitted under the decrees, this entry should
be revised accordingly.
                                     Page 33 of 68

-------
       LDAR and Benzene Priority Areas:  The statements under this heading tend to treat
States as a monolith, whereas individual States have shown varying degrees of interest in these
areas. To the extent that a State is a consent decree signatory it receives the same reports as
EPA, and it can assess performance of the consent decree requirements. Moreover, it should be
noted that many States may not have either a benzene or leak detection program; therefore, these
States cannot be used to assess refinery performance in these areas. Both of these important
considerations are overlooked in this discussion. It should be noted further that since entry of
the consent decrees, and notwithstanding EPA's active solicitation of the States to join EPA in
the consent decree process, OEC A has received few, if any, reports from the States inquiring
about refinery compliance in these areas. Finally, it should be noted that it is perfectly
appropriate for EPA to rely on company-developed sampling and monitoring plans to assess
compliance under the consent decrees. This is a time-tested and reliable approach, which forms
the basis for monitoring compliance and assessing performance under many environmental
programs, including the underlying benzene and leak detection programs.

       LDAR and Benzene Priority Areas "Current Compliance Status" column: Under the
terms of the consent decrees, these refineries are "in compliance and/or on a compliance
schedule." These entries should be revised accordingly.

       LDAR and Benzene Priority Areas "Emissions Status": As discussed above, emission
status is generally irrelevant. Accordingly, these entries should be revised to read: "N/A."

       LDAR and Benzene Priority Areas "Potential Measurement": This entry should be
revised to read "AP-42, inspections and company monitoring data."

OIG Response: We modified the text as we  determined appropriate.

We disagree with OECA that the report inaccurately references NSR/PSD emission status as
"unknown". Our interviews with a national EPA expert and senior OECA management
confirmed that some company baselines provided to EPA were inaccurate, some were based on
AP-42 estimates (emission factors used to estimate emissions that are known to have limited
accuracy), and some were considered accurate. We have modified the column heading to
"Progress Toward Emission Reductions," but maintain our conclusion that emissions status is
currently unknown.

We disagree with OECA about the accuracy of comments we have in the NSR/PSD "Issues
Associated with Measures" column.  Interviews with the national EPA expert in this area and
with senior OECA managers provided the information that OECA would verify compliance with
NSR/PSD requirements in the fourth year and at the conclusion of the consent decree; thus, once
every four years.  In interviews with the national EPA expert and senior OECA managers, we
requested information about monitoring plans for NSR/PSD emissions under consent decrees.
We were repeatedly told that "there was no  plan" to verify reductions on any schedule other Jhan
once every four years.

We disagree with OEC A that the LDAR and benzene priority areas discussed in the table "treats
States as a monolith." We intended the table to serve as a summary of current compliance
achievements in each of the priority areas. We chose not to discuss specific details but rather

                                     Page 34 of 68

-------
summarize general compliance information.  A senior OEC A manager told us that States that did
not actively participate in the consent decree process were not advised to focus on compliance in
these areas in state inspections.

We disagree with OECAthat the LDAR and benzene "Emissions Status" column is generally
irrelevant and should read "N/A."  OECA chose emissions from leaks and benzene emissions as
priority areas because they found that facilities emitted significantly higher amounts of
pollutants in these areas than reported by facilities.

Page 23,  "ICIS Data Did Not Accurately Report Results":

       The Report begins the discussion of ICIS data with a conclusion:

       "ICIS data did not serve as an accurate reporting mechanism for three reasons:
       (1) OECA included projected rather than actual emissions reductions, so ICIS
       data did not measure results; (2) OECA could not document that they input
       emissions reduction data timely, and (3)  OECA used an inaccurate (straight-line)
       method for estimating annual emissions reductions for multi-year consent
       decrees. Therefore, OECA may have taken credit in ICIS for refinery emissions
       reductions and other accomplishments that had not yet occurred or been verified."

However, this conclusion, and the  entire Section that precedes it, demonstrates a fundamental
misunderstanding of reporting under ICIS.

       OIG criticizes ICIS data for not listing actual emission reductions and for its use of a
straight-line projection of emission reductions over time, rather than some other (presumably
more accurate) method. This criticism misunderstands the purpose of ICIS.  This database is not
intended as the repository for emission inventory information from regulated sources.  That
function is met by EPA's NET, NTI and TRI databases. ICIS serves a wholly different function.
Due to ICIS's requirement to document enforcement information soon after the conclusion of
cases, knowledge of the actual emission reductions that will be achieved in the future, especially
for technology-forcing initiatives such as the Refinery Initiative, are necessarily prospective.
Furthermore, it is unclear what benefit would be gained by increasing the resolution of such a
prospective prediction of emission reductions by using anything more complex than a straight-
line model. Once emission control equipment is installed at a facility, the resulting reduction in
emissions will be reflected in NET, NTI and TRI,

       Additionally, the draft Evaluation criticizes OECA because it may have taken credit in
ICIS for emissions reductions that have yet to occur. As noted above, this is not the purpose of
the ICIS database. During the Evaluation OECA stated to OIG investigators that the estimates
were of what the annual emission reduction would be at full implementation of the consent
decrees. After full implementation of the consent decrees, such annual reductions would
continue for the life of the emission units from which these reductions are being made. OECA
focused its resources more on achieving results than on frequently measuring emission
reductions for purposes of entry into ICIS. It should be noted that for the foregoing reasons,
OECA's estimate of the annual emissions reductions grossly underestimates the overall
emissions reduction caused by the  consent decrees.

                                     Page 35 of 68

-------
       OECA acknowledges that certain consent decree data and information should have been
entered at or about the date of lodging of the consent decrees.  When OECA management
became aware that certain consent decree data had not been entered in a timely fashion it
immediately directed OECA and Regional staff to make the appropriate entries.  All consent
decrees are now entered in ICIS upon their lodging. While OECA agrees that it is appropriate to
note mat some entries were not made in a timely fashion, we recommend that the Evaluation
Report note also that OECA has taken steps to ensure that the consent decrees are entered into
ICIS in a timely fashion, as was the case with the five most recent consent decrees.

       The fourth paragraph of this section, stating that OECA "should have determined annual
projected reductions by claiming the reductions in the year OECA estimated the  reductions
would actually occur," is inaccurate and should be modified. OECA's estimates of emissions
reductions from the consent decrees reflects an estimate on a tons per year basis of the emissions
reductions that are realized once the consent decree is fully implemented. Given the breadth of
the consent decrees, the complexity of the programs covered by the consent decrees, and the
large number of emissions units at which the emissions are expected to occur, any other method
of estimating emissions reductions (e.g., annualized to reflect consent decree obligations in a
specific calendar year) would be exceedingly difficult and  require a large amount of OECA's
limited resources.  The suggestion for a more resource-intensive approach is unrealistic given the
available resources and add little to ensuring overall compliance objectives (both with respect to
the Initiative as well as the program overall). In any event, OIG's discussion misconceives the
purpose of ICIS, which is not to track emissions data on a per-y ear basis.

       The fourth sentence of the last paragraph of this section, which states that "the vast
majority of the reductions would occur in the latter years of the decrees," is not accurate. The
rule of thumb that OECA followed in negotiating the consent decrees is that two-thirds of the
emissions reductions from the installation of controls would be realized in the first four years of
the consent decrees.  The consent decree with BP is an example of this practice.

OIG Response: We modified the text as we determined appropriate.

We disagree that the ICIS section demonstrates a fundamental misunderstanding of reporting
under ICIS.  We modified the discussion of ICIS in the report to better clarify how OECA used
ICIS.  Specifically, we describe that, according to OECA, ICIS reporting was not designed to
capture, and  did not capture, information about environmental outcomes from the consent
decrees, such as demonstrated environmental and human health benefits.

We believe OECA's comment that we were inaccurate in our statement, "the vast majority of the
reductions would occur in the latter years of the decrees," contradicts other statements in
OECA's response in which they indicate it is too early for  OECA to demonstrate consent decree
emissions reductions. To use OECA's example, if BP realized two-thirds of the  emissions
reductions in the first four years, OECA could begin demonstrating initial emissions reductions
at BP facilities in January 2001.

Pages 23-24 under "SNC Rates Did Not Provide Useful Information":

       The draft Report discusses the use of SNC rates as  a measure of whether  the Initiative is

                                     Page 36 of 68

-------
productive, concluding that SNC rates did not indicate increased compliance in the refinery
sector. OIG's underlying assumption is that a successful initiative will lower SNC rates. As
noted elsewhere in OECA's comments, this is not a valid assumption. OIG's unfounded reliance
on SNC rates reveals another misunderstanding of the enforcement program; OIG's failure to
acknowledge why OECA does not rely on SNC rates to determine whether the Initiative is
successful likewise indicates OIG's lack of understanding.

       First, EPA chose the refining sector in part because of the very high relative SNC rates
(compared to other sectors). By focusing on this sector, EPA's efforts served to uncover
additional problems that had not been identified previously - thus raising SNC rates even higher.
This result indicates that the targeting and selection of this sector was a success (i.e., that the
decision to target refineries was valid and uncovered extensive noncompliance) and that
additional attention would reveal more compliance problems. OECA believes that, as a result of
the targeting work done prior to identifying a sector as a priority, any priority sector will see an
increase in SNC rates.

       In addition (and as noted elsewhere), while SNC rates might be expected to drop in the
long term, under the Clean Air Act, facilities operating under long-term consent decrees remain
in SNC (or "High Priority Violation") status until all conditions of the consent decrees are met,
all penalties paid, and any supplemental environmental projects are completed. Because of the
extensive nature of the mjunctive relief obtained by EPA in the consent decrees, which will take
many years to fully implement, a decrease in the SNC rate will not happen in the short term.
Because OIG misunderstands mis, the reliance on SNC rates as a measure of the Initiative's
success is misplaced and erroneous.

OIG Response: We disagree with OECA that we relied on Significant Non-Compliance rates and
that we had an underlying assumption that a successful refinery program will lower Significant
Non-Compliance rates.  Since OECA initially used Significant Non-Compliance rates to identify
refineries as a priority, we believed it was important to explain why Significant Non-Compliance
rates could not be used as a measure of increased compliance.  We recognize OECA does not
rely on Significant Non-Compliance rates to determine the refinery program's success.
Nevertheless, we deleted this section from the report to eliminate any confusion,

Page 24,  Table 3.3 and accompanying text:

       While OIG is correct that Clean Air Act SNC  rates as reported in EPA databases does not
yield useful information for consent decree tracking (and, as noted above,  it was not used for
such under the Initiative), the conclusion drawn by OIG in the draft Evaluation that compliance
•worsened under the Initiative because SNC rates increased between 1998 and 2003 is incorrect.
As noted earlier, the increased attention and focus on compliance issues would be  expected to
better identity noncompliance, which would lead to an expected initial increase in SNC rates.
These rates will go down over time, as the obligations under consent decrees are fulfilled: A
function of the SNC reporting criteria is that a facility remains listed until  a consent decree is
terminated. Accordingly, global refinery consent decrees affecting 42 refineries are likely to
remain identified as a SNC until well into the next decade, even where they are otherwise in
compliance with their obligations under the consent decrees. Put simply, SNC rates cannot (and
are not) used to  assess consent decree compliance; the draft Evaluation errs by doing so.

                                     Page 37 of 68

-------
       As discussed above, OEC A developed tools tailored to its investigative activities under
the Initiative by identifying and tracking "major" noncompliance with one of our four marquee
issues: failure to obtain a permit and install BACT controls (NSR/PSD); failure to include entire
units and hundreds of components in routine monitoring (LDAR); statistically significant
disparities between company reported and EPA determined leak rates (LDAR): identification of
refineries that erroneously thought they were exempt from control requirements or in compliance
with their selected compliance option (benzene). Other marquee issue violations may exist and
are Ihen pursued at a refinery; these actions are part of the Regions' core activities but are not
tracked under the Initiative.

OIG Response: As described previously, we deleted the section on Significant Non-Compliance
rates in the report We initially sought to demonstrate how we attempted to independently assess
the impact of the refinery program throughout the course of our evaluation. We looked to the
systems used by OECA to document results to Congress and prioritize work, including
Significant Non-Compliance rates.  However, we deleted the section on Significant Non-
Compliance rates from the report to eliminate any confusion on their use for determining
compliance in the refiners' industry.

Pages 24-25, Carryover Paragraph:

       The first sentence of this paragraph, that the "consent decree tracking system did not
provide accurate, reliable information about refinery company performance," should be deleted
because it is incorrect.  The paragraph should begin with the second sentence,  and the third
sentence should be modified to read:

       "In addition to the consent decree tracking system, EPA used standard consent decree
       monitoring techniques to assess companies' performance under the consent decrees,
       including review of a companies certification of compliance reports submitted to EPA
       and regular communications with the refining companies regarding its  performance."

       The last two sentences of this paragraph discussing consent decree tracking from 2001
through 2003 are in error and reflect a misunderstanding of what was required to implement
global refinery consent decrees. They should be deleted and the following substituted:

"EPA entered into four global refinery consent decrees in mid-2001. A large volume of
materials was submitted by the settling companies. By way of response by the Regions and
Headquarters, a plan for implementing consent decrees was developed, discussed and approved
in early 2002. A contractor was retained to assist in managing the flow of information and
aggressive actions were taken under that plan immediately thereafter.  Within  12 months of
entering into the first global settlement, OECA had a functioning system for monitoring and
managing consent decree implementation. OECA has provided  training for regional consent
decree implementers and has since conducted monthly conference calls to identify and promptly
resolve implementation problems with its contractor and identified company leads (i.e.,
individuals responsible for ensuring that a company complies with its consent decree and that
EPA takes timely action on all approval requests). Despite these and other efforts identified in
Chapter 4, EPA response delays persisted as more consent decrees were entered into and the
                                     Page 38 of  68

-------
volume of materials submitted increased significantly."

OIG Response: We disagree with OECA's suggested revisions.  We believe we have accurately
described the history of consent decree monitoring issues,

Page 25, "QuarterlyReports":

       The first sentence of this paragraph, that companies "routinely provide actual emissions
data related to NSR and PSD compliance," is misleading and/or inaccurate. This sentence should
be deleted.

       The third sentence, asserting that "OECA did not use available consent decree quarterly
reports to monitor, verify, or report consent decree progress," is also incorrect. Company leads
routinely monitor all quarterly reports and identify (if necessary) whether the company is
encountering compliance difficulties.  Since these reports are submitted under certifications with
attached criminal penalties (5-year felonies under 18 U.S.C. § 1001) and potential civil sanctions
(contempt of court), it is both appropriate and reasonable to generally accept the accuracy of the
companies' representations.  Thus, mis sentence should be revised to read: "OECA used
available quarterly reports to  monitor, verify and report consent decree progress."

       The use and utility of quarterly reports appears misunderstood, as reflected by the
simplistic conclusion stated in the last sentence of this paragraph.

       In addition, the criticism of OECA for failing to plan to regularly verify or monitor actual
refinery emissions reflects a misunderstanding of the intent of the Initiative. EPA intended that
consent decree requirements would become an integral part of the compliance requirements for
the affected facilities, with most requirements actually being included in state-issued permits.
As such, it would be the permitting authority's responsibility to verify and monitor emissions.
OECA believes that it would  have been inappropriate to assume this co-regulator's
responsibility.

OIG Response: We disagree that our criticism of OECA not planning to regularly verify or
monitor actual emissions reflects a misunderstanding.  We assessed whether OECA used
emissions data required by consent decrees to track emissions reductions on an on-going basis -
that is, through development and use of an emissions database or other system that could log
emissions data as it was generated by, or received from, facilities.  We found that, even though
an EPA national  refinery expert and an OECA official indicated that quarterly reports include
emissions statements, EPA did not use quarterly reports to monitor, verify, and report consent
decree progress.

We believe that as long as the refinery program remains a national program, OECA maintains
responsibility for tracking emissions reductions. To date, OECA has only projected emissions
reductions and, until OECA can verify actual emissions reductions, we do not believe EPA can
determine whether the refinery program has achieved reduced emissions.

"InformalMethods"Page 25, 1stParagraph:


                                      Page 39 of 68

-------
       The second sentence of the first paragraph, stating that "OECA did not use [activity or
output measures] for monitoring and managing the program" may accurately reflect the
understanding of an OECA official, but as discussed in considerable detail above, it is incorrect

       The discussion in the third paragraph concerning the lack of a formal "feedback system"
for "capturing" performance information under decrees to "feed back into the consent decree
negotiation process," reflects a lack of basic familiarity with some important aspects of the
Initiative.  Since (as noted several times above) the same national experts that ensure national
consistency in consent decree implementation are also the lead EPA consent decree negotiators,
there is no need for a separate formal feedback loop. Therefore it is inaccurate to conclude that
performance information did not "feed back into the consent decree negotiation" process.  A
formal feedback system may have been appropriate if the Initiative had relied on separate groups
of people for these functions, however. In this case, there is no reason to do so.

OIG Response: We modified the text as we determined appropriate.

Page 26, 1st Paragraph:

       For the reasons discussed above (see, e.g., comments for Page 22,2nd paragraph), the first
sentence's assertion mat "OECA has not verified emissions reductions from consent decree
implementation" is inaccurate and should be struck or modified to reflect the fact that it is not
yet possible to verify emissions  reductions from consent decree implementation. The second
sentence's suggestion to verify consent decree compliance should be modified to reflect that
OECA intends to develop  a system to capture emissions reductions achieved through
implementation of the consent decrees. The third sentence (noting the importance of verifying
mat actual emissions match projected estimates) implies that companies may have overstated
their anticipated emission  reductions; this is not true. In light of the severe consequences for
making false or misleading statements to the public, including the investment community under
SEC requirements, their estimates are probably low and their actual emission reductions are
likely to be somewhat higher than projected.

OIG Response: We disagree.  Whether OECA finds estimates too high or too low is immaterial
to the statement; rather the statement demonstrates that, for a number of reasons, OECA should
not consider or report estimates  as "pounds of pollutants reduced" by the program.  Rather, only
actual, verified emissions data compared against actual, verified baselines will provide accurate
information about pollutants reduced under consent decrees.

Page 26, T* and 3rd Paragraphs:

       The discussion in these paragraphs that OECA did not have accurate baseline emissions
data and therefore relied on data of .questionable reliability does not accurately characterize the
OECA manager's statements regarding a company's baseline information.  What the manager
stated was that EPA  cannot rely on emissions inventory data provided by a company to a State
for the purpose of assessing a fee to determine that company's baseline emissions to decide what
control measures a company might implement under a consent decree. The OECA manager
stated that this was because emissions inventory information is not as accurate as desired as it is
frequently based on, inter alia, emissions factors (which do not directly measure emissions from

                                     Page 40 of 68

-------
process units) and infrequent stack tests. For that reason, the OECA manager stated that OECA
demanded from a company, and received at the earliest stages of negotiations, the company's
best baseline information.  That OECA manager further stated that OECA reviewed the baseline
information to determine what methods of measurement were used as a predicate for the
baseline. Where OECA determined that the baseline information was reliable, it was used as a
basis for negotiations. In this respect, the Evaluation Report mistakenly characterizes data from
"stack tests" as not "actual monitoring data," This is incorrect as stack tests reflect "actual
monitoring data"; estimates, however, are admittedly less reliable for making informed decisions
based on individual unit emissions. Where OECA determined that the baseline information was
not reliable, OECA demanded that a more representative baseline be developed. In several
instances, EPA even required the development of better baseline information as  a term and
condition of the consent decrees, where the structure of the consent decree so allowed. These
paragraphs should be revised accordingly.

       As noted above, the verification of actual emissions reductions is the responsibility of
state and local permitting officials, not OECA.  The report quotes a 2001 General Accounting
Office (GAO) report mat found fault with EPA's oversight of state and local agencies in mis
area However, the GAO report only provides impetus for enhanced oversight of state and local
programs; nowhere does GAO suggest mat EPA's enforcement and compliance staff supplant
these state and local efforts.

OIG Response: We disagree with OECA that our discussion in the report regarding baseline
emissions data does not accurately characterize the OECA manager's statements regarding a
company's baseline information. We did not derive our statement on baseline emissions data
solely from conversations with one individual.  Other OECA staff and managers also discussed
the inaccuracies of baseline data provided by companies.

We also disagree that State and local permitting officials maintain sole responsibility for
verifying actual emissions reductions. EPA identified reduced emissions as a primary goal of
the refinery program and, as long as the refinery program remains a national EPA priority, then
OECA should retain accountability for the program achieving projected  reduced emissions.

Page 26, 4th Paragraph:

       The first sentence's assertion that "OECA also used manufacturers' estimates to predict
emissions reductions" is incorrect - manufacturers' estimates were not used to predict emission
or establish emission limits. Rather, they were used by OECA "to inform its decision-making
and negotiating positions." The second sentence should be modified to read "Preliminarily,
during consent decree implementation, OECA found that some new technologies performed well
and delivered the predicted reductions, while others did not."  This discussion does not
acknowledge that OECA has yet to reach any definitive decisions on the relative merits of the
control technologies identified in the consent decrees.

OIG Response: We modified the text as we determined appropriate.

Page 26, 5th Paragraph:


                                    Page 41 of 68

-------
       While the statement in the first sentence that "States were responsible for monitoring
refinery emissions using existing inspection programs to look for facility violations of consent
decree provisions" may accurately reflect what an individual OECA official said, it is incorrect.
OECA, as well as those states that have joined in these consent decrees, have a shared
responsibility for ensuring consent decree implementation. As to this paragraph's conclusion, it
is important to note that MOA Guidance has repeatedly emphasized the importance of the states
to the refinery initiative and identified expectations of the Regions in dealing with their states.
See e.g., MOA Guidance for FY 2002/2003, FY2003 and FY2004.

OIG Response: We modified the text as we determined appropriate.

Page 27, "Conclusion":

       Since this section repeats the same or similar statements commented on earlier and for
the same reasons above-identified, this section should be rewritten entirely.

       The suggestion that EPA should more regularly and more closely monitor consent
decrees for compliance is misplaced.  As discussed elsewhere in the comments to the draft
Evaluation, this suggestion neither accounts for the resources available to OECA as a whole nor
shows an appreciation for the large amount of work accomplished by the small group
responsible for work under each of the overlapping phases of the Initiative. It also fails to
appreciate the innovative "certification" approach used under the consent decrees as an
assurance of certainty and accuracy of the actions taken by refineries operating under decrees in
order to maximize available resources. Because the Evaluation does not take these factors into
consideration, the suggestion would functionally require OECA to divert resources from
addressing significant environmental issues in other industrial sectors.  The draft Evaluation fails
to recognize this reality.

       The suggestion in the third paragraph of this section displays a basic misunderstanding of
OECA's role and function. The report criticizes OECA for not establishing a process for
tracking trends in human health and environmental outputs and outcomes. This is not OECA's
function, and this achievement was never intended to be a goal of the Refinery Initiative.
Enforcement initiatives derive their justification from a fundamental tenet that compliance with
environmental laws and regulations will protect human health and the environment.  The
responsibility for establishing rules that are sufficiently protective, and for tracking trends and
environmental outcomes  and outputs, falls upon the states and the EPA program offices, such as
the Office of Air and Radiation.  OECA never intended to usurp that role.

OIG Response: As discussed with senior OECA officials, we understand that OECA is
experiencing resource strain and has multiple responsibilities. However, the objective for this
evaluation (as recommended by senior OECA management) was to evaluate the impact of the
petroleum refiner)' program alone, and offer suggestions for improving the program. Resource
strain does not diminish the importance of verifying the actual impacts (human health and
environmental improvements) achieved by the program:

We disagree that our report displays a basic misunderstanding of OECA's role and function.
Our report does not suggest usurping State or other EPA program responsibilities. However.

                                     Page 42 of 68

-------
since we believe the overall goal for all EPA programs is protection of human health and the
environment, all program activities, outputs, and outcomes should align with that goal and
eventually demonstrate success at protecting human health and the environment.

Page 28, "Recommendations":

       3-1 (Develop clear overall refinery program goals that allow for future assessment or
measurement and include timetables for accomplishment): 3-2 (Instruct OECA refinery program
managers to develop clear goals specifically for the refinery program's implementation phase):
and 3-3 (Ensure that all goals and performance measures are understood and shared bv everyone
involved in the national petroleum refinery program, including all EPA and State staff involved
in some portion of consent decree implementation). Concur. As discussed in detail in the
preceding comments, OECA believes that the refinery program's goals have been clearly
articulated since the national strategy was initiated in 1998, and that as the program evolved they
were further reflected in MOAs, etc. in the following years. As a general matter, OECA agrees
with these recommendations and will continue to develop and articulate appropriate goals and
performance measures.

       Following the identification of refineries as an enforcement priority for FY96/97, OECA
soon recognized the need for a comprehensive national strategy. It then developed a flexible,
integrated strategy (including sub-strategies) to address issues of widespread compliance and
enforcement concern at petroleum refineries. The resulting 1998  strategy was developed in
close consultation and coordination with the Regions, the Office of Compliance and the Office
of Regulatory Enforcement's media-specific enforcement divisions. It has remained largely
unchanged since then, with a focus on targeted investigations of "marquee" issues at petroleum
refineries and the goal of 50% improved compliance and 20% reduced emissions. The national
petroleum strategy and its implementation were regularly discussed at the staff level and
periodically reviewed by senior management in meetings, during conference calls and through
the MOA process. Periodic progress updates have also been and will continue to be circulated to
OECA management and the regions, but the extent to which specific individuals clearly
understand the national strategy, including its sub-strategies, goals and objectives, may depend
on the level of their direct involvement in these processes and communications.

       3-4 (Instruct OECA refinery program managers to use existing EPA. OECA. and outside
guidance to develop reliable performance measures to assess their progress toward meeting
national program goals.  Specifically, managers should fully implement OECA's performance-
based approach to program management as described in its December 18.2002.
Recommendations for Improving OECA Planning.  Priority Setting and Performance
Measurement, which specifies development of plans and reliable performance measures, to the
remaining phases of the petroleum refinery program): Concur.  OECA has already begun to
implement this recommendation (planned for prior to the Evaluation), as priority planning
process consistent with existing OECA guidance for FY 2005 has already been initiated.

       3-5 (Validate and build upon the refinery program logic model we developed during the
evaluation, and consider developing similar program logic models for other OECA programs to
develop a clear consensus on program goals and how a program is intended to work.): Non-
concur. OECA does not agree that several of the short-term, intermediate and long-term

                                     Page 43 of 68

-------
outcome measures in this logic model are appropriate benchmarks for judging the effectiveness
of a compliance and enforcement program.  These goals - such as "increased flexibility for
refineries to expand or upgrade operations" - are not realistic or likely to be obtained in an
adversarial enforcement context.  The absence of recommendations for how OECA would
benchmark the "before" conditions and measure changes over time for these ultimate outcomes
means that there is no basis for OECA to determine whether these measures are feasible.
However, OECA does agree that it should use appropriate performance measures and outcomes
to measure performance under the Initiative, and intends to do so as part of the FY 2005 priority
planning process.

       3-6 (Instruct OECA managers to verify emissions reductions predicted in consent decrees
on a quarterly basis. Verification might include establishing a detailed monitoring system, which
could contribute to refinery program performance measurement.'): Non-concur. As noted in the
detailed comments on this issue, OECA does not believe that this is an appropriate or effective
use of resources. Furthermore, the recommendation fundamentally misconceives the timing of
reductions under the decrees, which does not happen immediately upon lodging or entry of the
decree (as is apparently assumed), with regular reductions on a steady quarterly basis,  hi part
because these facilities are operating under court order, and are required to submit reports and
certify regarding their compliance with consent decree requirements (punishable by contempt
and/or criminal sanction), mere are sufficient indicia of reliability such that quarterly oversight
of emissions reductions is not necessary. Significantly, the recommendation does not take into
account the resource implications of this level of monitoring - both with respect to those
available for the Initiative (failing to recognize mat this work would need to be performed and/or
reviewed by the same group of national  experts responsible for all other aspects of the Initiative),
as well as those available to the air enforcement program and OECA as a whole. Even if OECA
agreed that this level of monitoring is appropriate, it is not clear how this would be accomplished
within the current resource levels and in light of other priority activities. On balance, these
resources are better utilized if devoted to addressing compliance issues in other industry sectors.
Notwithstanding the foregoing, OECA agrees that it is important to track emissions reductions
under the consent decrees, as appropriate given the consent decree milestone dates.

       3-7 (Instruct OECA refineryprogram managers to gatfaerr analyze,, and report relevant
program data to support overall QECA organizational decision making, and daily program
decision making'): Concur. As with Recommendations 3-1  through 3-4, OECA agrees with the
principle embodied in this recommendation, and will take steps to implement appropriate data •
gathering and analysis to support program decisionmaking.  However, in light of activities
identified in response to Recommendations 3-1 through  3-4, this recommendation appears
redundant and unnecessary.

Chapter 4

Page 29,1st Paragraph:

       For accuracy, the first sentence's statement, that "tracking problems developed  and
persisted,", should be modified as follows: "During early refinery consent decree
implementation, tracking problems developed and were  addressed."  By early-to-mid 2002
OECA had a fully functioning tracking system.

                                     Page 44 of 68  .

-------
       The second sentence states: "Tracking problems occurred because OECA did not
effectively plan how it would manage and monitor consent decree implementation, and because
OECA did not address implementation problems in a timely fashion." This is incorrect; more
importantly, it misconceives the importance of reporting requiring a response under the consent
decrees, and should be corrected. The issue is not "tracking" it is "responding." Company leads,
national experts and Regional personnel always knew what was submitted and what required an
EPA response. The plans for consent decree implementation identified roles, responsibilities
and expectations for action by the Regions; the process of monthly conference calls ensured that
all implementation issues were identified promptly and resolved (if possible). Although
identified as a national priority in MOA Guidance for FY2003 (June 2002) and FY2004 (July
2003), the unfortunate reality is that most Regions failed to make adequate resources available
for consent decree implementation. From the outset, OECA has aggressively attacked this
problem.

       A critical omission throughout the draft report is a general lack of appreciation for the
innovative and highly successful leveraging employed by OECA that allowed compliance issues
to be addressed company-wide, vastly extending the effectiveness of the limited resources
available for the effort. Therefore, in order to provide the context that is critically necessary to
fully understand and appreciate both the efforts and accomplishments of the Initiative, the
following should be added to the beginning of the first paragraph:

       "Negotiation and resolution of refiners' potential liability under the Clean Air Act
       proved to be an effective strategy for addressing refinery compliance issues on a
       company-wide basis, and without expending the considerable resources that
       would have been necessary to undertake and complete an exhaustive investigation
       of each individual refinery, and all associated follow up actions. EPA achieved
       considerable success early in the global settlement phase of the strategy. By May
       2001, EPA had negotiated and lodged consent decrees with four separate refiners.
       As discussed in Chapter 2, these consent decrees had a sweeping scope covering
       Clean  Air Act compliance in the four priority areas at 26 individual refineries.
       EPA did not anticipate its success. Consequently, EPA did not move as quickly
       as it should have to develop and implement a strategy to ensure oversight of the
       refiners of implementation of their consent decrees. Nonetheless, an
       implementation strategy did develop."

       OIG's statement in the last sentence, that "OECA must resolve remaining problems to
ensure timely  emissions reductions," overlooks actions already taken, and therefore should be
deleted and replaced with the following:

       "OECA has taken numerous steps to address and resolve issues associated with
       responding to consent decree deliverables. This effort has, inter alia, manifested itself in
       the significant increase in the number of EPA responses issued in response to company
       submittals since the beginning of 2003 fiscal year.  At bottom, the implementation of
       refinery global settlements requires considerable resources. Because of competing
       priorities, EPA has not been able to devote the resources it needs to ensure that timely
       responses are made to companies' submittals."

                                     Page 45 of 68

-------
OIG Response: We disagree with OECA that it had a fully functioning tracking system by early-
to-rnid 2002. January 2004 was three years into implementation of the first consent decrees.
This was nearly half-way through implementation for companies that signed consent decrees in
2000 and 2001, though it may have been early in implementation for other consent decrees.  For
this reason, we believe it may mislead readers to use the term "early refinery consent decree
implementation". As of January 2004, OECA had still not issued 108 responses required by
consent decrees, or 28 percent of required responses.  Responses were missing for all companies
except one and covered all priority areas. OECA worked with the contractor to redefine the
tracking and response problems, but we did not find that OECA "addressed" the tracking
problem in an efficient, clearly explained, and written fashion. Rather, OECA worked with the
contractor on a case-by-case basis (as indicated in e-mails) to redefine which items required an
EPA response. In the report, we acknowledge OECA's efforts to alleviate both the document
backlog (response problem) and the tracking system deficiencies.  OECA managers and an
employee of the contractor told us that the contractor-developed tracking system was not
operational until late 2002.  Further, the same managers and contractor staff told us that OECA
officials disagreed with how the contractor developed the system, and that implementers did not
directly use the system for mis reason, thus the system was not "fully functional."

We disagree that we incorrectly describe the tracking problems. We found implementation
problems with both OECA's responses and with tracking.  Two OECA managers detailed
problems related to development and use of the contractor tracking product. One OECA
manager told us that company leads, national experts, and Regional personnel always knew what
was submitted and what required an EPA response. However, problems related to development
and use of the contractor tracking product still existed, and OECA should apply lessons learned
from the system's development to modifying the product.

We disagree with OECA's suggested revised paragraph regarding the actions OECA has taken to
address and resolve issues associated with responding to consent decree deliverables. We
believe our report describes the most significant steps OECA took to resolve the problems and
accurately describes the difficulty in determining the degree of improvement.

Page 29, 2nd Paragraph:

       For accuracy, the third sentence, erroneously asserting that "[djuring the first 3 to 4 years
of consent decree implementation, companies installed emissions monitoring equipment," etc.,
should be modified to reflect the fact that refinery implementation has been ongoing for a period
of "2 to 3 years," and not "3 to 4 years" as currently drafted.  The Koch settlement was the first
settlement that was completed and it was not entered by the Court until April 1,2001.  In
addition, it is the emission monitors, and not the controls, that are to be tested and calibrated
during this time.  As noted above, comparatively, the bulk of compliance dates for installing
controls required by the consent decrees have not been reached  as of the date of this report.  On
the other hand, the consent decrees do require the companies to install emissions monitoring
devices at the earliest stages of consent decree implementation.
                                     Page 46 of 68

-------
OIG Response: We disagree. The third sentence refers generally to the activities that consent
decrees required of companies in the first three to four years of implementation, not to the
current state of implementation for any one consent decree.

Page 29, 3rd Paragraph:

       With respect to OECA's plan for monitoring consent decree implementation (the topic of
this paragraph), some context is important to understand the approach taken. The model for
company submittals and EPA responses followed in the refinery consent decrees is identical to
the model followed in crafting consent decrees generally. That is, where an issue cannot finally
be resolved through negotiations either because an event has yet to occur (e.g., a company-
required audit) or because the level of emissions reduction to be achieved can not be agreed
upon (e.g., emissions levels from a particular emissions unit given a particular innovative control
technology), a protocol is established and set forth in the consent decree by which the company
must prepare and submit the results addressing that particular issue. EPA always endeavors to
resolve the most contentious issues as  part of the consent decree, while universally  requiring
consent decree reporting of the most significant issues. Significantly, what distinguishes the
global refinery settlements from the typical consent decree are their breadth and complexity.
There are few industries as complex as the petroleum refining industry and there are few statutes
as complex as the Clean Air Act.  The global consent decrees address both on a company-wide
basis. Yet the discussion of the consent decrees in the Evaluation only notes this in passing, but
the substantive discussion does not appear to recognize or fully appreciate this complexity as a
factor in the development of OECA's consent decree implementation strategy.

       Although the focus of the discussion is understandably on NSR/PSD controls, each of the
other marquee issues was similarly engaged in this process. The fourth sentence should
therefore be revised to read: "Through this process, companies in consent decrees and EPA
national technical leads and other EPA staff collaborated to ensure that companies took actions
that would lead to the emissions reductions and improved environmental performance projected
by consent decrees."

OIG Response: We modified the text as we determined appropriate.

Page 30, Table 4.1:

       As currently drafted, Table 4.1's summation of consent decree-required reports
oversimplifies the nature and extent of the reporting. Therefore, the heading of Table 4.1 should
be modified to reflect that the table is for illustrative purposes only and does not provide either
an exhaustive or specific list of the reports that are required by the consent decrees.

       Root cause reports are not for "illegal" acid gas flaring incidents but for "all" major acid
gas flaring incidents. Flaring is an accepted and appropriate activity to avoid catastrophic
events; whether OECA considers them legal  or not is highly dependent on the circumstances and
a legal theory for requiring actions upstream of certain "affected facilities," as defined under the
new source performance standards.  OIG's report appears to assume that all flaring is illegal; this
is inaccurate and reflects OIG's misunderstanding of the "flaring" rules.


                                      Page 47 of 68

-------
       For accuracy, the table should be revised to reflect that EPA does not receive reports on
"installing" control equipment but on "the intended design of'such control equipment.

       The only report on abatement procedures concerns benzene EOL sampling; the table
should be revised to delete reference to LDAR and Flaring.

OIG Response: We modified the text as we determined appropriate.

Page 30, TdParagraph under "Tracking Problems":

       The draft Evaluation does not accurately reflect that backlogs did not exist until
companies began submitting reports that required an EPA response. Although only a handful of
such reports were submitted in 2001, the volume of other submissions was substantial and
growing. A major driver for the consent decree implementation plan was to manage this
process, paper and decision-making. It is incorrect to suggest that response backlogs developed
in early-to-mid 2001; backlogs only started to grow in early 2002.

       The statement in the fourth sentence, that there were 26 full-time equivalents devoted to
refinery implementation tasks, is not accurate, and the basis for this estimate is neither explained
nor understood.  While mere may have been 26 people nationwide that provided input into the
refinery consent decree implementation process from time-to-time, only the two staff members
at EPA headquarters and the three national technical leads (who were not one of the two staff
members) and perhaps an additional one or two individuals could be considered to be devoted
exclusively to the Refinery Initiative work. Moreover, it is important to emphasize that these 5
people spent a considerable amount of time devoted to other tasks, including negotiating consent
decrees and developing cases. As measured against the volume of work, the accomplishments
achieved (the  global agreements) by this small group is laudatory; this, however, is not
recognized in the draft report. In fact, as drafted, the report tends to suggest that the Refinery
Initiative was the only air enforcement work undertaken during this time. The report thus fails to
recognize or appreciate the Initiative work was - and continues to be - undertaken along with
major air enforcement efforts in other areas.

       Additionally, contrary to the Report's erroneous statement otherwise, the national
technical leads did not "officially approve" EPA responses to company submittals. Under the
language of the consent decrees only the Air Enforcement Division Director, or his delegate,
could "officially approve" EPA responses. The national technical leads' primary function was to
ensure consistency amongst and between the regions in responding to substantive technical
issues.

       A significant oversight in the draft Evaluation is the absence of any mention of the
importance of national consistency  for the implementation of these uniquely large, complex, and
comprehensive consent decrees. Companies took a leap of faith to join with EPA in these global
settlements and are expending nearly $2 billion in capital and will be spending several hundred
million dollars of year in operating  and other expenses. The success of the program demanded
that our decisions be consistent and apply uniformly  to all refineries owned by all the consenting
companies.  The OIG Evaluation seems to place a premium on tracking and related activities,
and devalues (or ignores) the critical importance of national consistency for both OECA and the

                                     Page 48 of  68

-------
settling companies.

OIG Response: We modified the text as we determined appropriate.

We disagree that the 26 full-time equivalents (FTEs) is not accurate. We derived the FIB
information directly from an OECA-developed document entitled Petroleum Refinery Initiative
FTEs, dated January 22, 2004, which we asked OECA to compile. The document shows 16.65
FTEs for 1998, 22.2 for 1999, 25.35 for 2000, 25.25 for 2001, 26.85 for 2002, and 26.15 for
2003. This provides an average of 24 FTEs over the period 1998 through 2003.

The report acknowledges refinery program accomplishments, but we believe that the program's
accomplishments will only be fully realized when the emissions reductions and other associated
environmental benefits (outcomes) can be demonstrated at the facility or community levels.

We believe it is generally expected that a "global" program will be implemented consistently.
We do not feel it necessary to emphasize this point  OECA correctly asserts that we placed a
premium on tracking  and monitoring the implementation and results of consent decrees. As
previously stated, while reaching global settlements has been an accomplishment, EPA needs to
assure Congress and the public mat these settlements achieve emissions reductions and other
environmental benefits predicted in consent decrees.

Page 31, 1st Paragraph:

       The report does not adequately address the cause of the backlog, nor of the steps already
taken to address it.  The backlog of responses to  company suhmittals was a function of three
structural issues:  First, the consent decrees required that company reports be delivered to the
region in which the particular refinery was located; Second, the consent decree implementation
plan put the onus on the regions to develop responses to company submittals in consultation with
the national technical leads, and; Third, the consent decrees required that the Air Enforcement
Division's Division Director approve each of EPA's responses, including those prepared in the
regional offices.  The effect of these three factors was to create a bottleneck that OECA
acknowledges delayed responses to company submittals. However, the Evaluation fails to
acknowledge the steps OECA has taken, and continues to take, to address these issues,
specifically:

•      First, in parallel with retaining a contractor to track consent decree deliverables, OECA
       collected and delivered the company submittals that had been transmitted to the various
       regional offices and headquarters; verified that all such deliverables were accounted for;
       and forwarded those deliverables to OECA's contractor to catalogue and enter into a
       consent decree tracking  system;

•      Second, OECA conducted negotiations with each of the refiners to modify the refiner's
       consent decree reporting obligations to ensure that documents were in both electronic and
       hard copy formats and delivered to the regional and EPA headquarters offices as well as
       to EPA's contractor;

•      Third, OECA  working with the contractor developed an electronic early notice system to

                                   . Page 49 of 68

-------
      notify those regions and national technical leads of a company subnrittal that required its
      review;

•     Fourth, OECA retained a contractor to provide technical support to the national leads in
      evaluating and preparing responses to technically complex reports;

•     Fifth, OECA organized and conducted national meetings to train regional staff, address
      overarching technical issues, and to encourage, organize and coordinate responses by and
      amongst the regional offices; and

•     Sixth, OECA actively coordinated with the Regions to ensure there were adequate staff
      resources in place to respond to company reports.

      Taken together, all of these steps resulted in significant improvement in EPA's response
time and a marked improvement in the number of reports issued by EPA.  Nonetheless, the task
of responding to company submittals is an important issue that requires constant attention. This
paragraph should be revised to read as follows:

      "Also in January 2002 and as identified in the implementation plan for consent
      decrees, OECA hired a contractor to serve as EPA's consent decree archive and to
      develop electronic databases for managing the process, its paper and decision-
      making. OECA provided the contractor with the consent decrees and all
      documents received to date. It then tasked the-contractor with developing a
      document archive and tracking system that cataloged all required company
      reports, all received company reports and all required EPA responses. The
      contractor developed an archive, a list of required deliverables and a document
      tracking system by late 2002, in close coordination with the several company
      leads. Some company leads in EPA regions had developed and continued to use
      their own, personal tracking to satisfy their responsibilities; others used the
      contractor-developed tracking system."

      In addition, to further improve response time, EPA continues to take additional steps to
further improve the process, including: Eliminating, for future consent decrees, the requirement
that the Air Enforcement Division Director's approval be obtained for EPA responses that
address leak detection, benzene, and flaring related issues; evaluating the consent decree
reporting obligations to eliminate unnecessary reports and approvals; better defining the universe
of submittals that require review by the national experts; and working with the Regions and the
States to identify greater efficiencies in the implementation phase.  The report should be revised
to reflect these measures.

O1G Response: We modified the text as we determined appropriate. While we acknowledge
some of the steps EPA has taken to address the backlog, we also identify additional steps needed
to correct tracking delays.

Page 31, footnote 11:

      The footnote is vague as drafted, and is therefore potentially misleading. Specifically, it

                                      Page 50 of 68

-------
is not clear from the footnote the bases for the conclusion that one consent decree required 585
reports to be submitted. It is also not clear from the footnote whether each (or how many) of the
reports that are included within the 585 are ihose requiring an EPA response.

OIG Response: We modified the text as we determined appropriate.

Figure 4.2 Page 32:

       The text of tfie report indicates that 149 items were erroneously identified as requiring a
response in July 2003. These erroneous reports should be removed from the totals: 272
"responses missing." This is probably still an overstatement. Had OIG afforded OECA
additional time to prepare this Response, we would be able to provide additional feedback.

OIG Response: We disagree.  We analyzed tracking information available to OECA managers
and consent decree implementers at three points in time.  Because the contractor provided
information about the 149 items in July, we believe it is important to characterize the backlog at
that time.

Page 32-33, Carryover Paragraph:
                                                                i
       This discussion (and the Evaluation in general) does not acknowledge that OECA
managers recognized the issue of potential for a backlog as early as April 2002, when it
specifically brought on board a manager to oversee refinery-related matters. However, because
of resource constraints and sheer volume of deliverables, it was not until September 2002 that
EPA's contractor was in a position to meaningfully track consent decree deliverables, and it was
shortly after that time that EPA redoubled its efforts to aggressively address the backlog.  By
failing to acknowledge this, the report implies that no action was taken to attempt to address the
backlog until much later. In fact, the data represented in Figures 4.1 and 4.2 both show a
significant improvement by EPA in addressing the backlog of deliverables.

       In addition, the fourth sentence is in error to the extent that it suggests that it was not
until July 2003 that it was "agreed that a backlog existed." The fact of a backlog was recognized
and steps were taken in early 2002 to address it by making the Regions responsible for drafting
initial responses and by providing training to the Regions that spring to ensure that they could
fulfill their responsibilities.

OIG Response: We disagree.  The report acknowledges thai OECA managers recognized the
potential for a backlog and took steps to reduce but not eliminate the backlog.

Page 33:

       The draft Report does not reflect the multiple, iterative aspects of our on-going efforts to
improve response and tracking of company reports requiring an EPA response, resulting in a
mischaracterization of those efforts.

       Based upon their knowledge of the specific terms and requirements of their assigned
consent decrees, company leads had managed consent decree implementation by using

                                      Page 51 of 68

-------
contractor-supplied Master Inventories that reflect multiple data elements, including when a
company report is due, when it was submitted, whether an EPA response is required, when an
EPA response is due (under self-imposed deadlines) and when EPA issued its response. As
earlier described to OIG, EPA created a new database to better track reports requiring an EPA
response in Hie second half of 2003.

       An initial Activity List was  created by EPA's contractor in July by sorting the Master
Inventories by whether an EPA response is required and whether a response had been issued.
This initial list was thereafter refined because some action items did not, in fact, require an EPA
response and because others required only one EPA response to multiple company submissions
(e.g., although a flaring incident report may result in multiple follow up and completion reports,
only one EPA response is required for an incident).  Refined/revised Activity Lists were vetted
with the company leads on January  7 and February 4, 2004. The format, accuracy and
completeness of these lists were confirmed, and it was then agreed that Activity Lists would be
used as the principle tool for tracking and managing items requiring an EPA response.

       EPA's contractor issues both a comprehensive Master Inventory and a more limited
Activity List to each company lead  every month:

•      Master Inventories enable company leads to determine whether a company report was
       timely submitted and an EPA response was timely issued.

•      Activity Lists enable company leads to quickly identify all reports requiring an EPA
       response by issue (e.g.,  NSR/PSD and Flaring), when they were received and whether
       "new" items have been  added.

These reports enable company leads to ensure compliance (e.g., timely company submissions)
and to engage in triage and focus on those action items most in need of a prompt response
because of their potential importance to future company actions or their being exceedingly
"late."

       EPA has recognized and continues to recognize that it should respond to all company
submissions in a timely and appropriate manner.  EPA has aggressively addressed its consent
decree implementation responsibilities. Despite these efforts, however, backlogs continued to
increase through 2002 (e.g., an analysis of current databases indicates that EPA had responded to
only 103 of 237 (43%) company submissions requiring a response as of December 31,2002).
EPA has and is focused on working off the backlog and anticipates reducing its response time
through continuing, concerted action.  Indeed, an  analysis of current databases indicates that
EPA had responded to 261 of 354 (74%) submissions requiring a response as of December 31,
2003.  Although improvements have been made, additional improvement is needed.

OIG Response: We modified the text as we determined appropriate.

Page 34, 1st paragraph:

       The first sentence's assertion that "[l]ate and absent EPA responses delayed company
implementation of projects designed to reduce emissions" is factually unsupported. OECA is

                                     Page 52 of 68

-------
unaware of any instance in which a delay in a required response delayed emissions reductions.

       The fourth sentence of this paragraph, asserting that "[a]bsent EPA responses also
compromised company compliance with consent decrees," likewise needs to be corrected. An
"absent EPA response" cannot compromise a defendant's compliance with a consent decree in
that its obligation to comply is wholly independent of an EPA response - that is, compliance is
required regardless of that response.

OIG Response; We disagree that we did not factually support our statement on how late and
absent EPA responses delayed project implementation. We discussed this point with OECA
managers and regional staff. We believe all action items included in consent decrees are
intended to work toward emissions reductions and oilier environmental benefits. We determined
that, although we found no examples of response delays affecting emissions reductions, in some
cases, response delays caused companies to delay implementation of a consent decree action that
would subsequently lead to an emissions reduction. OECA should recognize that responding to
company reports in a timely fashion would help ensure timely emissions reductions and other
environmental benefits from consent decrees, and that the potential impact of delayed responses
could include delayed emissions reductions.

Page 34, 2nd Paragraph:

       The second sentence's discussion of late responses to reports on "flaring incidents (60
percent), benzene handling, or LDAR requirements (29 percent); a small proportion (7 percent)
related to NSR and PSD requirements" should be revised to recognize that these delays are in
components of the "beyond compliance" or "enhanced program" and to specify their generally
limited nature. They do not deal with regulatory compliance.  For example, with respect to
flaring (and noting the two separate flaring programs covered by the initiative), the
overwhelming majority of the flaring reports with delayed responses from EPA are hydrocarbon
flaring, which is a "beyond compliance" component (see comment, page 15 Table 2.3). The acid
gas flaring program, by comparison, has the "stick" of stipulated penalties and therefore the
response to those letters is much more critical.  EPA is effectively current on responding to the
(more critical) acid gas flaring reports.  In this respect, in terms of allocation of scarce
enforcement resources, the "Consenter" refiners are correcting the hydrocarbon flaring problems
without Agency input, Therefore, for accuracy the sentence should be revised as follows:

       "Figure 4.3 demonstrates that most late and absent EPA responses related to
       flaring incidents (beyond compliance - 60%), benzene quarterly sampling plans
       (beyond compliance -20+%), LDAR database specifications (beyond compliance
       - 5%); and a small proportion (7 percent) related to NSR and PSD requirements,  •
       and 4 percent related to non-priority areas."

OIG Response: We disagree.  Consent decrees are legal agreements between companies and
EPA, and all actions agreed to in the decrees thus become "requirements" of the consent decree.
We believe that adding a discussion in this report section about elements in consent decrees that
are beyond compliance would be confusing and irrelevant.
Page 34, Figure 4.3:
                                     Page 53 of 68

-------
       This chart on "EPA responses by Priority Area" is lacking needed detail. In particular, it
should reflect the universe of reports considered in developing the percentages. In addition, the
figure should also separate benzene reports from LDAR reports, and acid gas and sour water
stripper gas flaring reports from hydrocarbon flaring reports.  This lack of specificity detracts
from the usefulness of the data represented.

OIG Response: We disagree that the chart on "EPA Responses by Priority Area" lacks needed
detail.  We intended this chart to demonstrate that the majority of required EPA responses
related to flaring and NSPS.  The level of specificity contained within the contractor tracking
reports we obtained did not allow for accurate parsing of types of flaring reports. In some cases,
the item delineated a particular type of flare, while in others the item just listed "flaring
incident".  We separated the responses out in this way so that each of the three national EPA
experts' areas was contained within one portion of die chart.

Page 35:

       Acid Gas Flaring - The discussion in this section attempts to draw a conclusion that a
late response is the cause of a facility's failure to correct the "root cause" of a flaring incident.
There is no demonstrated causal link. Under the terms of the consent decree, refiners are
required to identify the root cause of a flaring incident and implement corrective action to
address the root cause, independent of EPA's response.  In fact, the purpose of a flaring incident
report is for the facility to identify the results of the Evaluation, and the corrective action
undertaken or to be undertaken. While EPA should respond in a timely fashion to a flaring
incident report, and prior to September 2002 EPA had not acquitted itself well in that area, there
is no causal or associated link between a refiner failing to correctly identify and correct the root
cause and EPA's late response. OECA is not aware of any instances where mis-identification of
a cause for flaring led to subsequent flaring events. This paragraph should be revised
accordingly.

       NOx control - OIG suggests that, during the time that EPA's response was pending, the
refineries would have installed control equipment. The implication is that, had the EPA
response been timely, the public would not have been exposed to excess  emissions. This
suggestion is clearly erroneous. Almost without exception, no NOx controls  were slated to come
online during the time that EPA responses were pending, thus EPA delays in  responding cannot
be construed as somehow leading to delayed compliance.  This paragraph further fails to account
for the quality of the refiner's report or the reasonableness of the NOx plan proposed.  Many of
the proposed NOx control plans are being actively evaluated, and are the subject of ongoing
negotiations with the companies in an attempt to avoid dispute resolution and potential court
proceedings regarding the reasonableness  of the limits proposed. In EPA's view, many of the
plans provided by the refiners, and the limits proposed therein, were inadequate, unreasonable,
or both. The discussion in this paragraph tends to suggest that a timely approval of a report that
has not been determined to be adequate is preferable to  efforts to craft one that is adequate.

       Reporting Consent Decree Outcomes -  OIG suggests that delayed Case Conclusion Data
Sheets led to inaccurate characterization of yearly emission reductions associated with consent
decrees. However, as stated above, the tracking of such emission reductions  is not the function
                                     Page 54 of 68

-------
of OECA databases, so delays in entry had no effect on the appropriate characterization of
emission reductions from the initiative

OIG Response: We modified the text as we determined appropriate for the discussion on acid
gas flaring. We did not intend to indicate any relationship between facilities' not correcting root
causes and EPA's late responses, rather, this shows one of three "examples of potential impacts."
OECA staff told us that mis-identification and, moreover, delayed EPA responses could impact
flaring incidents in this way. We are not aware of any instances where ttiis has occurred, but it is
a potential effect of delayed EPA responses that OECA acknowledged, and an example of how
EPA response time could impact program  outcomes.

We believe that had EPA timely responded, the companies could have timely proceeded with
properly developed action plans and may have begun installing pollution control equipment.
More timely EPA review and approval would minimize the possibility and the amount of time
that refineries might expose the public to excess emissions. We recognize the importance of the
quality of refiners' reports. However, we believe that 478 days (approximately 16 months)
between receiving a company report and issuing a response pushes the limit of reasonableness.

Pages 35-36:

      The Report alleges that the inadequacy of the tracking process instigated by EPA led to
delays in company compliance with their consent decrees. However, while it would have been
problematic for OECA to immediately document company compliance with the consent decrees,
on substantive consent decree requirements no delays would have occurred.  The stakes were so
high for both the companies and EPA on these requirements that frequent communications and
meetings between the parties ensured that  significant compliance deadlines were met as
scheduled.

      The discussion on these pages also seems to assume that, under the decrees, it is EPA's
obligation to demonstrate a facility's state of compliance. However, this is incorrect.  As noted
above, the consent decree reporting requirements obligate a company to identify when it is in
non-compliance with a consent decree requirement. In light of the consequences for
noncompliance (and the incentive to identify non-compliance), absent such a notice it is both
reasonable and logical to assume that a company is, in fact, in compliance. EPA justifiably
relies on this assumption, because the refinery consent decrees, like all other consent decrees,
provide incentives to a defendant in the form of stipulated penalties to both comply with
significant consent decree milestones or requirements and report that it has done so or not (as the
case may be).

      It is important to note that in  many instances - with the exception of NSR/PSD related
requirements - it was the Region's (not OECA's) responsibility  to formulate and draft responses
and to track compliance.  The tracking contractor was retained to provide the Regions with a tool
for doing so. A substantial portion of the delay in failing to respond to a company submittal
occurred well prior to the delivery of a draft response to the national leads or headquarters

OIG Response: We disagree mat EPA can assume a refinery's compliance just because they are
under a consent decree. The refiner)' program began because: (1) EPA identified disparities

                                     Page 55 of 68

-------
between refinery operations as reported to EPA through the NSR/PSD permitting process and
actual refinery operations; and (2) EPA identified refineries as number one for noncompliance
among 29 industry sectors ranked by EPA in 1996. While we acknowledge that companies bear
responsibility for reporting violations under consent decrees, the industry's history of
non-disclosure and noncompliance warrant EPA oversight of company self-reporting and
additional scrutiny of refinery emissions.

Page 36,  "Better Planning" discussion:

       The Report suggests that more accurate planning would have enabled OECA to avoid the
document backlog.  However, the reason for the backlog was  not insufficient planning, but rather
insufficient staff. OECA deliberately expended its scarce resources in pursuing further consent
decrees from other refineries, rather than tracking existing agreements, thereby ensuring
additional emission reductions. This decision affected consent decree document processing, but
it had a minimal effect on the timing and success of the emission reductions required by the
consent decrees, because companies were aware of the severe legal consequences of failure to
comply with consent decree requirements.

       The first paragraph of this discussion states that monitoring the implementation of
consent decrees is a very common activity for OECA, and therefore OECA should have known
what would be involved and should have been able to plan accordingly. However, the scope and
complexity of refinery  consent decrees far exceeded that of any  other agreements OECA reached
in the past, so past agreements  would have provided no basis for accurate planning of resource
needs in the refinery initiative.

       An important contextual point to recognize in the discussion of bottlenecks,
notwithstanding OECA's experience with consent decrees, is  that the scope of the refinery
consent decrees is unprecedented (a factor not adequately taken into account by OIG).  In
addition, while OECA could have taken steps to prepare for consent decree implementation
sooner, at the time no one could have or should have anticipated the success  we achieved in the
period  of time in which it was achieved.  These should be recognized as factors (not excuses)
that should be acknowledged as part of the Evaluation (in this discussion as weir as elsewhere in
the report) for the Evaluation to be both fair and balanced.

       By late 2001 OECA recognized that additional resources were needed for consent decree
implementation, as reflected in the Attachment 1 to the consent decree implementation plan.
The national experts also recognized this need, as reflected in the plan's requirement that the
Regions (not the national experts) prepare the initial draft response. Moreover, this resource
need was repeatedly and continuously identified thereafter. See FY2003 MOA Guidance (June
2002) and FY2004 MOA Guidance (July 2003).  In other words, the fact that an increased
workload under the consent decrees was coming was recognized. Resource choices forced
decisions on the aspects of implementation that were the most critical - a factor not fully
accounted for in the draft Evaluation.  (Note also that there have been no requests from States
regarding consent decree tracking needs, etc., and prior to review of this report OECA is
unaware of any such request.)
                                     Page 56 of 68

-------
       It is important to recognize that the most substantial portion of the delay in responding to
company submittals occurred at the Regions well prior to the delivery of the Region-prepared
draft responses to either the national technical leads or EPA headquarters staff.  This is not to say
that there were no delays attributable to the national technical lead or EPA headquarters; there
was. OEC A recognizes that efficiencies need to be achieved all around., and we are continuously
striving to do so: As noted above, the early success of the global approach was  unexpected and
its scale was without precedent The failure of OIG to acknowledge this success, coupled with
the overly-critical view of the implementation difficulties that were directly attributable to this
success, is indicative of the lack of balance in OlG's Evaluation.

       The statement at the bottom of Page 36 and carrying over to Page 37 that OECA did not
begin to address delays until "late 2003" is inaccurate and should be corrected, as that effort had
been ongoing for more than one year. In fact, that conclusion conflicts with earlier statements in
the report, including when OECA began its efforts with its tracking contractor (see page 31). It
should be revised as follows: "Delays in responding to company reports became acute in mid-
2002 and persisted into 2004. OECA developed a plan for consent decree implementation in
January 2002 and began addressing anticipated delays in early 2002, but Regional resource
limitations precluded  its plan from being implemented fully, resulting in growing backlogs.
More recently, OECA is using contractors to assist in the work assigned to the Regions under
that plan and anticipates eliminating its backlog by mid-year."

OIG Response:  We modified the text as we determined appropriate.

We disagree that insufficient staff caused the backlog. OECA knew how many  staff it had
available for consent decree negotiations and consent decree implementation, and, as noted
above, OECA made active decisions to devote staff to negotiations at the expense of
implementation. Better planning for  staff utilization to successfully accomplish both  consent
decree negotiations and consent decree implementation could have alleviated the document
backlog and associated delays.

We disagree with OECA that past agreements provided no basis for accurate planning of
resource needs in the refinery initiative since the scope and complexity of refiners' consent
decrees exceeded that of any other agreements. In our opinion, the national and regional EPA
experts involved in negotiation had the necessary' subject area expertise to at least roughly
estimate the time required of EPA to  implement individual consent decree actions, and OECA
management could have developed time and resource requirements for each consent decree
using those estimates.

Page 37:

       Training for Implementers - Training by national leads and implementers occurred, well
prior to 2003 - specifically, in the spring of 2002. This should be corrected for  accuracy.

       Revisions to Tracking System -  This discussion is incomplete as it does not
acknowledge OECA managers' assessment of backlogs, which occurred on an ongoing basis,
and did not begin in 'late 2003" as indicated in the report, but much earlier.  It is also relevant
(but not recognized in the report) that OECA had been working with the contractor prior to the

                                      Page 57 of 68

-------
OIG Evaluation to better tailor the tracking system to reflect consent decree requirements that
required an EPA response.

       In fairness it should be noted that during late 2003, OECA had received preliminary
feedback from IG investigators regarding the tracking system, had acknowledged the difficulties,
and set out to correct perceived and actual deficiencies.  This work was not occurring in a
vacuum.

OIG Response: We disagree that the discussion on revisions related to the tracking system is
incomplete.  The report accurately summarizes the events related to the tracking system,
including OECA's steps to address continuing problems. We made other changes to the text as
we determined appropriate.

Page 38, 1st Paragraph:

       The conclusion of this paragraph, that "OECA eliminated tracking of the timeliness of
both company reports and EPA responses from the tracking system," is incorrect.  OECA's
contractor maintains three databases and generates three reports for each consent decree: (1) a
Deliverables Tracking Table that identifies when deliverables will be due (a planning document
that is no longer used); (2) a Master Inventory that includes all data fields, including when a
report is due from a defendant, when it is received and when an EPA response is due (the
principal tool for company leads ensuring that a company is in consent decree compliance); and
(3) an Activity List of open action items that informs all team members of what is  open and
needs attention, including the highlighting of all "new" action items (i.e., those received,
archived and inventoried since the last Activity List). The Report should be revised accordingly.

OIG Response: We disagree.  The master inventory lists that OECA and the contractor provided
to us listed company report submittal dates and EPA response dates. The lists did not include
company report due dates or EPA response due dates.

Conclusion and Recommendations

       4-1 (Instruct its consent decree tracking contractor to resume tracking both company due
dates for reports and EPA response due dates  so that OECA and outside parties can easily track
company and EPA responsiveness):  Non-concur. As discussed in the detailed comments above,
OECA does not agree that it is necessary to further revise the tracking system at this time; this
recommendation has been overtaken by events.  During the time that OIG was conducting its
investigation, OECA itself identified some deficiencies with its tracking system, and appropriate
revisions were made (note that due dates for reports and EPA responses continue to be tracked
under each consent decree's Master Inventory). The critical issue is not simply tracking, but
responding to those reports requiring an EPA response.  Changes have already been made to
address this.  For example, in January 2004, Matrix and Region 6 tracking systems were
compared and verified for accuracy and usefulness, and a single system was selected for
implementation nationally, and access provided to all parties responsible for consent decree
implementation - including companies. In addition, changes to requirements for company
submittals have been made to subsequent consent decrees to better manage the process.
                                     Page 58 of 68

-------
Furthermore, substantial progress has been made to reduce the backlog, indicating that the
current approach is having the desired result.

       4-2 (Create a comprehensive tracking plan and system that outlines specific
responsibilities for OECA staff. EPA Regions. State and local air pollution control agencies, and
companies):  Non-concur. OECA does not agree that this recommendation is necessary. Sixteen
states are parties to global refinery consent decrees and currently receive copies of all consent
decree submissions that relate to each refinery within their states. OECA staff, EPA Regions and
state/local authorities who are parties to the consent decrees are and continue to be reflected in
the consent decree implementation plan.  Specific tracking tools (e.g., Master Inventories and
Activity Lists) are circulated on a monthly basis to all necessary participants. As discussed
above, EPA has a comprehensive consent decree tracking protocol that is being implemented
through our contractor.  Subject to claims of privilege and confidentiality, OECA does not object
to any interested non-party, including other states and local authorities, requesting tracking
information from Matrix at its own expense.

       4-3 (Provide additional training at the regional level, and empower regional experts to
review and respond to company reports. Allow national technical leads to spot-check responses
from regional experts to ensure national consistency): Concur. OECA agrees with this
recommendation, and has provided (and will continue to provide) appropriate training as needed.

       4-4 (Developa formal feedbacksystem to ensure that  QECA's workforce andmanaggrs
have a common understanding of implementation responsibilities, a common perspective on the
status of implementation, and the ability to expeditiously address implementation issues):  Non-
concur. As explained in the detailed comments, in light of the small staffing level and their
overlapping responsibilities, a formal feedback system is not necessary.  Should the number of
staff substantially increase in the future, a feedback system may be appropriate at that time.
       4-5 (Ensure frequent and open communication between partners (States, regions) and
headquarters about responsibilities for executing portions of strategies so that misconceptions or
confusion can quickly be eliminated): Concur.  OECA will continue to communicate with
Initiative partners.

       4-6 (As discussed with OECA managers, include consent decree implementation in
OECA priorities and strategic plans, allocating staff and resources to implementation until
OECA completely implements all consent decrees.):  Concur. OECA agrees with the principles
underlying these recommendations, and efforts have already begun for FY05 implementation on
these matters. OECA will continue to allocate adequate resources to the Initiative whether it is
identified as a national priority or part of the core program.

       4-7 (Develop a plan for allocating negotiation and implementation resources. Use
resource planning in new initiatives to determine the predicted workload associated with the
initiative; allocate training, education, and development resources: and provide for office-wide
reevaluation of the resource plan): Concur.  This recommendation has been overtaken by events
(priority planning for FY05), and implementation and other resources will be allocated in
concert with other OECA priorities and core programs requirements. Office-wide (and Region-

                                     Page 59 of 68

-------
wide) reevaluations are considered as part of regular planning processes.

Chapter 5

Page 41, 1st Paragraph:

       The Evaluation states that the refinery initiative received "mixed reviews" from the
various stakeholders ranging from "positive to very negative." While we understand the value of
soliciting such views from stakeholders and would encourage that OIG solicit views in Ihe
future, we question the necessity of highlighting the views of one "major industry professional
association," which has been openly hostile to the Refinery Initiative since its inception.  It is
important, to ensure balance, to identify that as the position of the industry group. It should also
be considered that an industry group's adverse comment could be legitimately considered an
indicator of the success of the Initiative.

OIG Response: We do not believe we need to modify the paragraph. We obtained views from
various stakeholders and accurately state that their comments ranged from very positive to very
negative without disclosing .from which specific stakeholder those comments came.

Page 41,1st Lesson Learned (Identify Enforcement Concerns within an Industry):

       A major lesson learned that is not otherwise captured is the importance of developing
issue-specific strategies and investigative techniques. Identifying an issue as major started this
process, it did not end it. Accordingly, the second sentence should be revised to  add new
language at the end:  "Focusing on specific enforcement concerns allowed OECA to direct its
limited resources to address an industry's most significant compliance problems  through the
development and use of new tools and cost-effective investigative methods.

OIG Response: We have combined this lesson learned with another lesson. We modified the
text as we determined appropriate.

Page 42, 3rd Lesson Learned (Build the Program as a Whole with Regional and Headquarters
Staff):

       This section wholly discounts the role of the Department of Justice in these cases (a
persistent shortcoming of the Evaluation as a whole). The pace of negotiations is frequently
influenced by Justice Department attorneys who themselves have competing demands for their
time.  Ignoring the role of the lead negotiator in the draft Evaluation is a significant omission
(and an additional indication of a basic lack of understanding of the enforcement process).

OIG Response:  We did not discount the role of the U.S. Department of Justice (DOT) in the
refinery cases. We disagree that we lack understanding of the enforcement process because we
did not include DOJ in this lesson learned. This lesson learned highlights the unique working
relationship between EPA regional and headquarters staff on the national refinery program.
Stakeholders did not describe DOJ as part of mat unique relationship.  Further, when we
obtained information from DOJ staff on their roles and responsibilities regarding the national
refinery program, they stated that they had a similar role to that in prior enforcement cases. In

                                     Page 60 of 68

-------
 addition, they stated that their coordination and collaboration with EPA was typical of other
 similar initiatives.

 Page 42, 4th Lesson Learned (Designation of a "Champion "):

       The discussion of this "lesson" fails to recognize that there is a senior manager
 "champion" for the Initiative (the Air Enforcement Division Director), and fails to appreciate
 what goes into the choice of which particular manager is appropriate in a given set of
 circumstances.  Note also that the end of the  second sentence should be revised to read as
 follows: "Industry representatives said that having a senior OECA executive who had specific
 knowledge about the issues, had decision-making authority, talked with them about the program,
 and even participated in negotiations, made a positive impact in how they reacted."

       This recommendation also fails to understand both the history of the Initiative and who
 served as the "champion" during a given period of time, and why.  Beginning in January 2000,
 the Director of ORE oversaw the enforcement-related elements of the Initiative.  The Director of
 ORE assumed these responsibilities because  the Director of the Air Enforcement Division was
 fully engaged in the ongoing Utility Initiative. Since February 2002, the enforcement phase of
 the Initiative has been championed by the Acting Associate Air Division Director, now the
 Acting Director of the Air Enforcement Division.  At all critical times there has been an
 appropriately designated "champion," taking all relevant factors and circumstances into account

 OIG Response: We disagree that we did not appreciate what goes into the choice of a
 "champion",  This lesson describes the views of EPA and industry stakeholders on how an
 effective "champion" can benefit a priority such as the national refinery program. We believe an
 effective "champion" represents more than a title assigned to a senior manager, but rather an
 individual who keeps the priority moving toward its goal.

 Page 44, 7th Lesson Learned (Focus on End Result):

       As written, the importance of certainty is identified only with respect to regulatory
 compliance/risks. From a company perspective, this certainty is much more for other purposes,
 primarily capital planning.  Accordingly, it is suggested that the final sentence  be revised as
 follows: "The industry  saw many complex regulations on the horizon and viewed participating
 hi consent decrees as 'good business' to provide certainty and to inform their capital planning
processes."

 OIG Response: We have combined this lesson learned with another lesson. We modified the
 text as we determined appropriate.

 Page 44, 8th Lesson Learned (Diligent Oversight of Consent Decrees):

       The statement that oversight of the consent decrees is the "last piece" of the refinery
 program is incorrect  The unstated assumption is that all refineries will eventually voluntarily
 sign up to install specific control devices, etc., after a couple of rounds of discussion, and that me
 process is largely a "cookie cutter" operation. The settlement process is far from a "cookie


                                      Page 61 of 68

-------
cutter" operation.  The last piece of the refinery program is the inevitable litigation with the
outlying companies that refuse to settle.

       A principal reason why the global settlement process has been so successful is that it
recognizes that reduced emissions and improved practices are of benefit to the company as they
demonstrate their commitment to the surrounding community. Thus, the second sentence should
be revised as follows: "Without effective implementation, anticipated emission reductions and
increased industry compliance may not be realized."

       The third sentence's statement that "OECA, regions, and States should ensure
implementation of consent decree provisions and, if not, take appropriate enforcement action"
misses the mark on what consent decree implementation, as under the global refinery consent
decrees, requires.  It is not checking the checker but to take positive action for implementing the
decrees (i.e., issue approvals and otherwise take action). Thus, "to oversee the consent decree
implementation" should be replaced with "to implement the consent decrees."

OIG Response: We modified the text as we determined appropriate.

Page 45, Conclusions:

       The tone and tenor of the conclusion is symptomatic of OIG's failure to grasp what was
accomplished (and what is still being accomplished) by the Initiative, and particularly the
unbalanced view exhibited throughout the report. For example, the first and third sentences tend
to suggest (and, more importantly, do not recognize) the dynamic nature of the Initiative, nor do
they properly credit OECA for having already incorporated into the Initiative the lessons learned
in it is early stages.  Similarly, the second sentence gives passing reference to the need for an
effective enforcement program "despite limited resources." However, Ihis cannot be squared
with the unrealistic resource implications of many of OIG's recommendations (particularly in
the absence of any consideration of OECA's overall resources program-wide).  The fourth
sentence repeats a mistake - that there is no senior OECA "champion" for the Initiative - that
was pointed out to OIG investigators on numerous prior occasions.  The last two sentences also
repeat a misperception that is both noted several times in the detailed comments above as well as
pointed out to OIG during the investigation, that the Initiative is  still ongoing, that
implementation under the decrees is at an early stage, and that much of the new control
equipment (such as FCCUs) which will result in significant emission reductions is not yet
required to be installed. What OIG fails to appreciate is that the changes to refinery operations
and controls under the decrees require significant capital expenditures, and will be taking place
over a period of years under long-term implementation schedules (i.e.,  emissions reductions are
still to be realized in future years). Because OIG's characterization is written in the past tense, it
infers that the Initiative is completed, controls installed and emission reductions already realized.

OIG Response:  We disagree that we did not grasp what the refinery program accomplished and
that the report is unbalanced. We believe the report adequately describes the refinery program's
accomplishments as well as areas for improvement. We modified tiie conclusion to
acknowledge that OECA incorporated some of the lessons learned.  However, OECA has not
effectively  incorporated other lessons learned into its refinery' program, such as effective
                                      Page 62 of 68

-------
communication among stakeholders. We kept our recommendation that OECA develop a
communications plan for the consent decree implementation phase of the program.

We believe OECA's designation of the Air Enforcement Division Director as the "champion"
for the refinery program represents an important first step in ensuring the success of the refinery
program. However, the lesson learned goes beyond just assigning a label to a senior official.
The lesson includes that the senior official ensures the achievement of program goals and the
communication of results to all stakeholders.

Recommendations

      5-1 (Disseminate the lessons learned from fee refinery program to OECA staff to
benefit other compliance efforts, obtain additional feedback from stakeholders — including
States, industry, and environmental groups - on other lessons learned, and update OECA's
Framework for a Problem-Based Approach to Integrated Strategies for on-going and future
industry-specific enforcement programsV  Concur in part, non-concur in part. OECA does not
agree that revisions to the recently-issued Framework for Problem-Based Approach to
Integrated Strategies (November 2002) ("Framework") are needed to reflect lessons learned
from the Initiative.  Rather, the lessons learned from EPA's Refinery Initiative have informed
and continue to inform the Agency's evolving problem-based approach to solving environmental
compliance problems.  For example, OECA is currently engaged with the Regions in developing
performance-based strategies for each of the national priorities selected for FY05-07. As part of
that effort, EPA is reviewing and refining, where appropriate, the goals and the strategies for the
refinery initiative. In developing these performance-based strategies, OECA and the Regions
will be guided by the recently-issued guidance, Template for Developing a Performance-Based
Strategy for National Compliance and Enforcement Priorities (Final Draft February 18, 2004) as
well as the Framework. As EPA gains more experience in the development and implementation
of such strategies, we will refine guidance on the use of such strategies where needed.

5-2 (Designate a senior OECA executive to assume the role of champion for the refinery
program to ensure (a) that all refiners enter into consent decrees or face appropriate alternative
enforcement actions, and (bVconsent decrees are effectively implemented):  Concur in part, non-
concur in part. As noted above in the comments on the "Conclusion" section, OECA does not
agree that there is no "champion" for the Initiative. However, OECA agrees with the need for
national enforcement priorities to be managed by a senior enforcement official (e.g., Division
Directors or their Associates), working on a team with other senior managers from EPA
headquarters, regions and DOJ.  Since 2002, Ihe senior enforcement official responsible for
managing the refinery initiative has and continues to be the Associate Director of ORE'S Air
Enforcement Division, who is now serving as the Acting Director of the Air Enforcement
Division. OECA agrees that the Air Enforcement Division Director is responsible for ensuring
that (a) refineries enter into consent decrees or face appropriate enforcement action, and (b)
consent  decrees are effectively implemented. OECA does not agree that it is necessary for EPA
to settle  with or litigate against all refiners in the industry under the Initiative. The goal of the
Initiative is  and has been to increase compliance by 50% and decrease emissions from refineries
by 20%.  The "100%" goal suggested by OIG miscomprehends the purpose of a "priority." Even
after this is no longer a priority, further work in this area would be undertaken through the
"core" program guidance, including the potential for multi-regional priorities, as well as the
                                     Page 63 of 68

-------
potential for State efforts.  OECA's work in an area is not just be driven by a coverage number,
but by whether there continues to be an appropriate federal role.  OECA designates a set of
national priority criteria (i.e., significant environmental benefit, pattern of noncompliance,
appropriate federal role), and following the return of refineries to the "core" program certain
refineries may be better handled by States or as part of a multi-regional priority.  OIG's
suggestion of "all" refineries lacks the context of taking into account all of our regulatory
partners, and that certain types of facilities are best addressed at different levels.
       5-3 (Consider designating a senior OECA executive to assume the role of champion for
each of the other enforcement priority areas. EPA and industry officials should recognize the
champion as knowledgeable and as having the authority to make decisions related to the priority
area): Concur in part, non-concur in part.  The determination of whether any particular initiative
or priority area requires an OECA-designated "champion," and at what level, will be made on a
case-by-case basis as is appropriate in light of all relevant facts and circumstances. For those
areas that have been selected as national enforcement priorities for FY 2005, senior OECA and
regional management have been named as "champions" for the purpose of developing
performance-based strategies for each priority area OECA anticipates that EPA senior
enforcement management will also name a lead manager to be responsible for the
implementation of each performance-based strategy.

       5-4 (Develop a communications plan for refinery consent decree implementation.  The
plan should clearly describe the roles and responsibilities of all stakeholders, including refinery
priority area experts and regional and State officials): Concur. OECA is already in the process
of priority planning for FY 2005 (begun prior to the Evaluation), which will result in a revised
performance-based strategy for the refinery sector. The performance-based strategy for FY 2005
will outline the path forward in (a) completing the refinery sector as a national priority and (b)
ensuring mat refineries governed by federal consent decrees comply with the terms and
conditions of their consent decrees.

Appendix A

Page 50, 1st Paragraph:

       The statement at the end of this paragraph is incorrect. EPA did not eliminate tracking
the timeliness of company reports and EPA responses between October 2003 and January 2004.
They remain on both the Deliverables Tracking Chart (a forward-looking planning document)
and the Master Inventory (a comprehensive tracking tool).  They were only deleted from the
Activity Lists, an extract from the Master Inventory that identifies outstanding action items.

OIG Response: We disagree.  EPA eliminated using tables that tracked the timeliness of
company reports and EPA responses between October 2003 and January 2004.
                                     Page 64 of 68

-------
Appendix B

Pages 53-54:

       The information relayed in the table must be footnoted or explained to better reflect that
each of the 11 consent decrees covered each of the refiner's petroleum refineries owned and
operated by that refiner on the date of consent decree entry. Thus, EPA recommends the Table
be modified to add footnotes or endnotes, or corrected, as follows:

       ConocoPhillips - The table should reflect that 12 of its 16 refineries were purchased after
       Conoco resolved its liability to the United States under its global settlement.

       ExxonMobil - This should not be highlighted as there is no consent decree at this time.

       ChevronTexaco - The table should reflect that the two refineries not covered by its
       global settlement with the United States are asphalt, not petroleum, refineries.

       Roval Dutch Shell GP — The table should reflect that the two refineries not covered by a
       global settlement are operated by Shell Chemical, a separate business unit from its fuels
       refineries.

       Citgo - This is omitted from the table, but should be included.

       Tesoro - The table should reflect that Tesoro purchased two refineries from BP, both of
       which are covered by BP's global settlement with the United States.

       El Paso - The table should be corrected to reflect that El Paso (which should be identified
       as Coastal  Eagle Point Refining) has only one refinery, not two.

       Orion - The table should reflect that this refinery is covered by a global settlement
       (patterned  after the United States' settlements) with the State of Louisiana.

       Murphy. Farmland. Premcor. Pennzoil. Crown. Frontier, and NCRA - (1) The table
       should reflect that each of these refiners were the subject of individual enforcement
       action under the Initiative and entered into limited, non-global settlements. (2) The table
       should reflect that the United States pursued an enforcement action against Murphy to
       judgment for operations at one of its refineries and that the relief obtained there includes
       elements patterned after Ihe global settlements. (3) The table should reflect that
       Farmland,  a company in bankruptcy, has recently resolved its liability with the United
       States through a consent decree that is patterned after the global settlements.

OIG Response: We modified the text as we determined appropriate.

Appendix C

       No comment
                                     Page 65 of 68

-------
Appendix D

Pages 57-58:

       For the April, 1998 entry, it is not clear what is meant by the sentence "Implementation
phase of the national refineiy program began." Additionally, the table is difficult to reconcile.
with Table 2.1 "Major Phases of Refinery Program." The table puts undue emphasis on the
dates that die consent decrees are lodged. Few obligations arise on the date the decree is lodged,
as the terms of the settlement have no legal force unless and until the decree is entered by the
court. Therefore, at a minimum the Table should reflect the date of both lodging and entry.

For clarity and accuracy:

       Early 1998 should indicate that monthly calls and annual meetings occurred from 1998
       "through" 2000.
•      February 1999 should refer to national meetings "to review investigation progress."
       February 2000 should identify that meeting was to "review progress" and to discuss
       national investigations.
•      Early 2000 should indicate that corporate officials given the option "to resolve all issues
       of widespread compliance/enforcement concern to EPA." It is not accurate to suggest
       that they were then under threat of enforcement.
•      December 20. 2001 should parenthetically indicate, as indicated in the March 21,2001
       entry that "(30 percent of industry then under consent decree; additional 30 percent in
       similar global settlement negotiations)."
•      October 16.2003 should parenthetically indicate, as indicated in the March 21,2001
       entry that "(40 percent of industry then under consent decree; additional 40 percent in
       similar global settlement negotiations)."

OIG Response:  We modified the text as we determined appropriate.

Appendix E
                                      *.
Page 59:

       The "Consent Decree Process Flow Chart" should reflect that it is the date of entry, not
lodging, that a consent decree becomes a live legal document for purposes of the Court. It is
also the date from which the vast majority of the refiner's consent decree obligations flow.

For accuracy and to conform the chart to reality, the following changes should be made:

*      Negotiation should be replaced with "Global Consent Decree Opportunity"; based on a
       willingness to pursue or reject that opportunity, the chart branches.
•      Lawsuit/Other Enforcement Action should be replaced with "Continue/Initiate
       Investigations) and Take Enforcement Action"; most global consent decrees were not
       based on mature investigations. If the opportunity had been rejected, we would have
       been required to complete the investigations and, only if noncompliance then found,
       would we be able to file a lawsuit or take other enforcement action.

                                     Page 66 of 68

-------
•      Company Report to EPA on Action should delete "on Action;" most reports are after the
       fact; some are to approve future action.
•      No EPA Response Required leads directly to Environmental Result and does not result in
       Company Action (except in the unusual case).
•      Environmental Result has no required return loop.  However, if one is to be reflected, it
       should not be to "Negotiation" but to "Planned Consent Decree Action" (e.g., approval of
       catalyst leads to optimization study that results in proposed catalyst addition rates for the
       demonstration period that (when approved by EPA) leads to..., etc.).

OIG Response: We modified the text as  we determined appropriate.

Appendix F

Page 61:

"National Refinery Program Logic Model - Outputs and Later Activities"

       "$ of Penalties" is referenced under the heading of "Measures" but it is unclear
       what it is intended to measure. If it is intended as a qualitative assessment of the
       success of the initiative, then we would recommend that amount and scope of
       injunctive relief and the amount of supplemental environmental projects also be
       included under the "Measures" heading.

"National Refinery Program Logic Model - "Short Term Outcomes and Later Activities,"
"Intermediate Outcomes," and "Long-term Outcomes"

       We recommend that the three boxes be combined to create two boxes -
       "Immediate Outcomes" and "Long-term Outcomes." Many of the outcomes
       identified as "short-term" could also be considered "intermediate" and "long-
       term" outcomes (e.g., effective monitoring, improved relations). We also
       recommend and mat the "logic" model expressly state that outcomes identified as
       immediate are listed because they are first realized immediately after entry of the
       consent decrees, although they are very likely to translate to long-term outcomes
       as well.

       The "measures" associated with the "short-term," "intermediate," and 'long-
       term" outcomes are not well-defined, well-considered, well-understood or
       measurable.  For example, "accuracy of [emissions] inventories" is not, and can
       not be, a measure of success of consent decree programs.  That the States  do a
       poor job of ensuring accurate emissions inventories or that the Clean Air Act does
       not require universally verifiable methods of measuring emissions from each and
       every emission unit are matters well beyond the scope of the Initiative and the
       consent decrees. Likewise, the "measures" of improved "enforcement
       credibility," "improvement of relations," and "enhanced ethic" are not susceptible
       to measurement.
                                     Page 67 of 68

-------
       For accuracy and completeness and as described above, the following Measures should
be revised as indicated:  '

•      Outputs and Later Activities - add "# of referrals and probably referrals"
•      Outputs and Later Activities - revise to read "% of Refining Capacity under Global
       Consent Decrees; in Global Consent Decree Negotiations; under Non-Global Consent
       Decrees; in Litigation; and Under Investigation."
•      Short Term Outcomes and Later Activities - revise Extent of Monitoring & Accuracy of
       Inventories to read "# of CEMs installed and stack tests required under consent decrees
       and %of Refining Capacity and Revised NOx and SO2 Baselines."
•      Improvement of Relations should be replaced with "# of state parties to consent decrees
       and # of states participating in joint marquee issue investigations."

Intermediate Outcomes - include the following measures:

       "# of WGS, SCRs, SNCRs and ULNBs installed under consent decrees" and "# of
       FCCUs utilizing catalyst additive control technologies."
•      "# of refineries implementing enhanced LDAR program"; "# of refineries implementing
       enhanced benzene program"; and "# or refineries implementing enhanced NSPS/flaring
       program (e.g., root cause failure analyses)."
•      "# of Consenters Group meetings to which EPA is invited" and "# of consent decree
       amendments approved by EPA."

Long Term Outcomes - Include the following compliance measure:

•      "# of refineries found to be in maj or violation of LDAR, benzene, NSPS/flaring and/or
       NSR/PSD requirements" and "# of refineries under global consent decrees thereafter
       determined to be in major violation of LDAR, benzene, NSPS/flaring and/or NSR/PSD
       requirements."

OIG Response: We have deleted the logic model appendix from the report.
                                    Page 68 of 68

-------