*&j\     Office of Inspector General
          Report of Review
     Special Review of EPA's Information
               Systems Program
                   Volume I

              E1SKG3-15-0098-4400038

                  March 24, 1994
                                      at toast 50% recycled fib*r

-------
Inspector General Division
  Conducting the Audit:
Region Covered:
Program Offices Involved:
Technical Assistance Division
Washington, D.C.
Agencywide
Agencywide

-------
     \
     8 UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                      WASHINGTON. D.C. 20460
                          MM? 24 (991             THE INSPECTOR GENERAL
MEMORANDUM
SUBJECT:  Special Review of EPA's Information Systems Program
          Special Report No. E1SKG3-15-0098-4400038
TO:       Carol M. Browifer
          Administrato
     Attached is our final report entitled "Special Review of
EPA's Information Systems Program."  The purpose of this review
was to respond to a September 27, 1993, request by the Chairman,
Subcommittee on Superfund, Recycling, and Solid Waste Management
of the Committee on Environment and Public Works, U.S. Senate.
In his request letter, the Subcommittee Chairman asked that we
perform a comprehensive management review of EPA's information
systems program, with as much specific attention on the Superfund
area as possible.

     Because of the high-level focus of the request and the time
constraints involved, we approached this review very differently
from past special reviews.  The most important difference was the
participation of staff members from EPA program offices jointly
with Office of Inspector General (OIG) employees in a team effort
to identify the root causes of EPA's information resources
management (IRM) problems and to develop cost-effective
solutions.

     The single most important theme we identified was that
historically, EPA has not treated information as a valuable,
strategic resource.  The review found that EPA's IRM problems are
the result of systemic deficiencies in the following areas: IRM
management and organizational structure; resource planning and
performance measurement; information system development; and data
management.  This report makes recommendations to address these
concerns .

     This special review report describes problems that the joint
review team identified and corrective actions the OIG recommends.
The report represents the opinion of the OIG.  Final
determinations on matters in this report will be made by EPA
managers in accordance with established EPA audit resolution
procedures.  Accordingly, the recommendations in this report do
not necessarily represent the final EPA position.

-------
     This report is addressed to you because of the Agencywide
impact of the IRM program and the need for you to be involved in
seeing to it that necessary corrective actions are properly
implemented.  We would recommend that you designate the Assistant
Administrator for Administration and Resources Management as the
action official to oversee the corrective actions being taken to
resolve the issues presented in this report.  He is required to
provide us, within 90 days, a written response describing the
actions the Agency has taken as a result of our recommendations.
If these actions will not be complete by that time, we ask that
the actions that are ongoing be described and a timetable
provided for their completion.  We have no objections to the
release of this report to the public.

     Should your staff have any questions or need additional
information regarding this report, please have them contact
Kenneth A. Konz, Assistant Inspector General for Audit, on
(202) 260-1106.
                                                Martin

Attachment
John C. Marl

-------
  	Special Review of EPA's Information Systems Program

                         EXECUTIVE SUMMARY


  PURPOSE. OBJECTIVES AMD METHODOLOGY

  This comprehensive management review of the  Environmental
  Protection Agency's (EPA)  information resources  management
  (IRM) program was undertaken in response to  a request from
  the Subcommittee  on Superfund, Recycling, and Solid Waste
  Management of the Committee on Environment and Public Works,
  U.S. Senate.  This request expressed concern about the "...
  apparent lack of  credible basic management information about
  where Agency  funds,  including Superfund money, are being
  spent and how,  and the questions raised about what the
  resulting accomplishments may be . . ."

  Because of the  strong base of prior audit work and studies in
  this area, one  objective of the review was to examine,
  validate, and synthesize issues raised in relevant prior
  Office of Inspector General (OIG)  and General Accounting
  Office  (GAO)  audit reports and Agency studies.   The second
  objective was to  identify any additional mission-related IRM
  problems, their root causes, and their solutions from an
  Agencywide perspective,  with specific attention  on the
  Superfund program.

  This management-oriented review focused on "big  picture,"
  whole program issues.   Working cooperatively,  a  joint
  OIG/Agency review team1  relied heavily on interviews and
  focus groups  to most fully leverage both review  team and
  Agency resources  within the allotted time.   The  groups
  responded to  previously identified problem statements from
  their perspectives,  and then discussed the IRM problems they
  encounter directly in their jobs,  as well as the problems'
  root causes and solutions.   These discussions brought forth a
  mixture of facts  and opinions.  The review team  also used
  recently issued audit reports and current audit  work to
  identify or confirm IRM problems and root causes.   Finally,
  the review team followed up on a number of significant
  recommendations made in OIG audit reports issued since fiscal
  1990 to determine the status of EPA's corrective actions.
  1 The review team consisted of representatives from the Office of Inspector General, the Office of Information Resources Management
(OIRM), and the Office of Solid Waste and Emergency Response (OSWER), which is responsible for the Superfund program.
                              Report No.  E1SKG3-15-0098-4400038

-------
	Special Review of EPA's Information Syatens Program

The review embodied three central assumptions: that the team
would obtain the best root cause statements and solutions
from front-line program office and  IRM representatives in the
Regions; that balance would be obtained by interviewing
members of the Headquarters IRM community; and that the joint
review team approach would help mitigate  the review's time
constraints.

BACKGROUND

A sound IRM infrastructure and efficient, responsive
information systems are critical to EPA's ability to provide
objective, reliable, and understandable information to help
achieve its environmental mission.  For instance, EPA's
information systems provide data:

  •  on management and oversight of the Superfund program;
  •  on nationwide air pollution and water quality;
  •  to track and report on Agency  financial management; and
  •  on the effects of pesticides and toxic substances on
     humans and the environment.

The IRM program at EPA has both centralized and decentralized
components.  The central components in the Office of
Administration and Resources Management (OARM) and the Office
of Policy, Planning and Evaluation  (OPPE)  provide certain
Agencywide IRM services, as well as IRM policy, planning, and
oversight.  The decentralized components  (i.e.)" the program
and Regional offices) implement and operate systems to meet
their needs.  While the responsibility to lead the use of
information and information technology to best meet the
Agency mission may be centralized,  most resources to
implement and manage information systems  are decentralized.

The IRM program has recently faced  significant challenges,
and is working to overcome them.  IRM planning and security
have been declared to the President as material weaknesses
for the Agency.  Corrective actions are underway in both
areas, are progressing well, and present  the promise of
fundamental improvements in the Agency's  management of
information.

Nevertheless, significant problems  remain.  A number of
recent reports and studies present  recommendations for these
problems, although the recommendations tend to focus on
specific systems and selected components  of the IRM program.
The intent of this review, however, was to discover root


                              it

                           Report No.  E1SKG3-15-0098-4400038

-------
	Special Review of EPA'3 Information Systems Program

causes of the most fundamental, cross-cutting IRM problems,
and to craft program-wide solutions affecting the entire
Agency.

RESULTS IN BRIEF

In recent years, EPA has been criticized for information
systems problems, many of which the Agency is in the process
of addressing.  These problems may be categorized as:

  •  concerns about the quality, integrity, and completeness
     of data, i.e., a lack of credible basic.management
     information about where and how Agency funds, including
     Superfund money, are being spent, and the
     accomplishments resulting from those expenditures;
  •  difficulties,in addressing cross-media pollution
     problems;
  •  significant cost overruns and delays in developing and
     implementing information systems;
  •  development of duplicate systems; and
  •  exposure of EPA's financial payment systems to
     unnecessary risks.

Our review has concluded that these problems are the result
of systemic deficiencies in the following areas: IRM
management and organizational structure; resource planning
and performance measurement; information system development;
and data management,  in our view, the Subcommittee's concern
stated in the request letter to the Inspector General cannot
be attributed to any one of these systemic deficiencies
alone, but results from all of them in combination.

The single most important theme identified by our review was
that the culture of information management at EPA does not
treat information as a valuable, strategic resource.  Well-
managed corporate assets are rigorously inventoried, tracked,
and maintained.  More importantly, they are used to leverage
a corporation's productivity.  But at EPA, information -
resources are rarely seen as being of strategic importance in
accomplishing the Agency's environmental mission.  The
transforming power of information is not being harnessed, and
the opportunity to use information to leverage the Agency's
other resources is being missed.

In response to previously identified IRM problems, EPA has
initiated a number of actions.  For example, the Office of
Information Resources Management  (OIRM) has drafted a revised


                             iii

                           Report No.  E1SKG3-15-0098-4400038

-------
  	Special Review of EPA's  Information Systems  Program

  charter which would establish the  "Executive Steering
  Committee for IRM" as  the decision-making vehicle for  top
  management  in IRM planning and oversight; the first meeting
  of  the committee is planned for  the second  quarter of  fiscal
  1994.   A new IRM Planning Group  has developed a proposed IRM
  strategic planning process; a high level portion of an IRM
  Strategic Plan2  is expected to be completed by  October 1994.
  Further, OARM plans to pilot test  a working capital fund in
  fiscal 1995,  which is  intended to  provide stable, .raulti-year
  funding for selected Agency activities, including major IRM
  initiatives;  implementation is scheduled for fiscal 1996.  In
  addition, the Agency has begun a number of  initiatives to
  help its information systems better assist  the environmental
  risk assessment process and to promote data sharing.
  Finally, EPA is converting hundreds of contractor positions
  to  Federal  employee positions, and many of  these are.
  tentatively planned to be automated data processing (ADP)-
  related.  This will help reduce  the Agency's vulnerability to
  the loss of IRM expertise.

  Although in many respects we believe the Agency is proceeding
  in  the right direction,  we still have concerns about EPA's
  ability and commitment to address  its long-standing IRM
  problems.

  PRINCIPAL ISSUES

  IRM Management and Organizational  Structure

  Virtually all of the interview and focus group participants
  believed that EPA's upper management does not understand how
  critical IRM is to environmental protection.   Because  of
  this,  information is not treated as a strategic resource, IRM
  is  not treated as a core function,  and information is  not
  viewed as a valuable tool to empower the public.  One  reason
  is  that top management is not perceived to  be accountable for
  IRM,  but instead is focused primarily on programmatic
  accomplishments.  Another factor is that EPA's IRM Steering
  Committee has not fulfilled an important role as a decision-
  2 We define the term, IRM Strategic Plan, as a detailed document which includes an overall IRM vision for the Agency, links
information technology to anticipated-program and mission needs-and reflects-budget constraints. In addition, the Plan should provide
detailed descriptions and costs of all major IRM initiatives—for example, major system development and modernization activities and
significant ADP acquisitions and contracts—in order to form the basis for budget requests. Hie Plan should be a 5-year plan, updated
annually, and includes functional requirements, current architecture, target architecture, and annual tactical plans with estimated budgets and
major milestone dates to achieve the target architecture. As such, this definition includes two documents the Agency separately calls its
IRM Strategic Plan and its Multi-year Implementation Plan.


                                   iv


                                Report  No.   E1SKG3-15-0098-4400038

-------
	Special Review of  EPA'a  Information Systems Program

making body, due to  its infrequent meetings and lower level
participation.

Lines of authority and communication for IRM matters are
confused at EPA.  Formal lines  of  authority are fragmented
between four offices under two  Assistant Administrators.
Interview and focus  group participants  consistently expressed
concerns about an inadequate communication infrastructure.
Of particular concern was the perception that the Office of
Information Resources Management and the National Data
Processing Division  (NDPD) make important unilateral
decisions without consulting their "customers," such as the
service cutbacks announced by NDPD for  fiscal 1993 and 1994
which some focus group participants said were developed
without any input from those Agency components using these
services.

The organizational placement, numbers,  and apparent
qualifications of the EPA employees with significant IRM
responsibilities are inconsistent  among Headquarters program
offices.  For example, two Assistant Administrators had no
in-house data processing staff, while one Assistant
Administrator had 45.  And these employees occupied 18
different job series positions, with over 42 percent in non-
computer specialist  positions.  This problem,  combined with
EPA's dependence on  support  contractors,  has left the Agency
vulnerable, to a loss of expertise.

Resource Pla.r>ninq and Performance  Measurement

Interview and focus  group participants  consistently cited the
lack of an Agencywide IRM Strategic Plan as a major
impediment to the success of IRM at EPA.   To maximize the
effectiveness of such a plan, however,  EPA as a whole needs
to have an Agency Business Plan3 which aligns  resources,
skills, and strategies  (including  information technology)
with EPA's mission.  Such a  Business Plan is now in
development, and is  expected to be completed in the Spring of
1994.  Concurrent with this,  a  draft portion of EPA's IRM
Strategic Plan--a high level IRM vision statement--is being
developed, with issuance expected  around the same time.

Another aspect of this problem  is  that  the IRM planning that
does exist is not integrated with  EPA's budgeting process.
3An Agency Business Plan is considered the same as an Agency Strategic Plan.


                               V
                           Report  No.   E1SKG3-15-0098-4400038

-------
_ Special Review of EPA7 a Information Sy at <•""." Program

This contributes to funding shortages for systems development
projects and to incremental, independent systems design and
development .

Unstable funding from one year to the next also was viewed as
contributing to systems development problems, due the lack of
a mechanism to ensure long-term financial support.  The
current 1-year budget cycle was viewed as the major reason
for this, along with the fact that system development
projects are not line budget items.  Longer-term budgeting
for these projects would establish more accountability for
managers in the development, implementation, and maintenance
of major information systems, and force managers to look at
the entire life cycle of a system.  For example, different
decisions may have been made during the development of EPA' s
Integrated Financial Management System  (IFMS) if management
had initially been aware of the total IFMS life cycle costs.
The initial estimate for IFMS was almost $8 million, but this
was only for the design, development, and implementation of
the system.  Our current audit has estimated that the costs
of the entire life cycle of IFMS over a 12 -year period, which
include maintenance and operations costs not included in the
original estimate, are actually over $200 million.

IRM planning is greatly complicated by EPA' s unbalanced
emphasis on short-term vs. long-term results.  While all
organizations have emergencies which must be addressed, it
was the perception of interview and focus group participants
that this occurs excessively at EPA.  Continually changing
Congressional information needs, new and changing
environmental regulations, and poor communications with
Congressional staff were viewed as keeping EPA in a generally
reactive mode .

EPA does not have adequate mechanisms in place to measure the
performance of the IRM program.  For example, EPA has not yet
built into its information systems sufficient measures of
environmental successes, accomplishments, and economic
benefits, which Congress and the public want.  Further, EPA
cannot totally account for its IRM expenditures on a system-
by- system basis; preliminary information developed in a
separate ongoing OIG audit shows that for 10 major
information systems, EPA had not reported correctly, by
system, fiscal 1993 costs to the Office of Management and
Budget totaling almost $29 million.  Nor can EPA measure the
success or failure of some critical information system
activities, such as software maintenance.
                           Report Ho.  E1SKG3-15-0098-4400038

-------
  	Special  Review of EPA's Information Systems Program

  Information Systems  Development

  Carefully defining the needs of all users and customers of an
  information system is  critical to its success.  But in our
  interviews and  focus groups,  managers and staff frequently
  stated that EPA's  systems developers* do not always work
  effectively with users and customers  (i.e., EPA program
  officials, Congress, and the States) to define needs and
  design good systems.

  Systems developers were specifically criticized for not
  really understanding how EPA's programs work or what program
  officials' needs were; some focus group members attributed
  this to EPA's heavy  use of contractors to develop systems.
  EPA's national  systems were also criticized for: meeting
  Headquarters needs but not Regional needs; having low levels
  of utility; not being  "friendly" to Regional users; and being
  overly complex.  A manager gave one example, saying that they
  could not do something as simple as go into IFMS and find out
  what their available travel balance was; the data was
  unavailable and unaccessible in a user-friendly fashion, so
  they had to ask an intermediary staff person to generate a
  report to find  out the answer.

  Inadequate consideration of customers' needs also applies to
  Congress and the States.   Focus group members believed that
  not working with Congress to anticipate their questions led
  to situations in which Congress asked good, straightforward
  questions about an environmental program's accomplishments,
  yet EPA's systems  could not answer them.  Likewise, not
  involving the States in the systems design process creates
  problems when the  States must then input data to,  or use data
  from, EPA's systems.

  On the other hand, even when needs are fully defined during
  systems design,  new  questions or needs can arise years later
  which the systems  were not designed to address.  This was
  perceived to lead  to unfair criticism.

  The vast majority  of Headquarters program and Regional
  offices do not  use a standard approach to manage software
  development and maintenance.   The primary reason appears to
  be the use of contractors for the majority of EPA's systems
  4 In this report, we use the term "system developers" to include contractor and Federal personnel from OIRM, program offices, and the
Regions who perform system development activities.


                                vii


                              Report No.  E1SKG3-15-0098-4400038

-------
  _ Special Review of EPA's  Information Systems Program

  development  projects; these contractors  are allowed to use
  their own  methodologies, which vary widely and do not always
  meet Federal requirements.  Further,  some of the steps in
  this standard approach are perceived  to  add time to the
  development  of a system, so some organizations tend to
  perform  them cursorily or not at all .  Participants in one
  focus group  summed it up by stating that "management often
  will not let us do it right, but there is always time to do
  it again."  Part of this process involves assessing whether
  older systems are obsolete, something EPA does not
  consistently do.

  Data  sharing is crucial to the Administrator' s four top
  priorities,  which cut across all environmental media.  Yet
  the link between data sharing and these  priorities has not
  been  widely recognized.

  EPA does not have the "structure"  (i.e.,  an information/data
  architecture,  data standards, and a Data Administration
  function)  to really share data Agencywide.   Interview and
  focus group participants attributed this to:  insufficient
  resources;  an EPA culture which does not value
  information/data as a strategic resource; EPA programs and
  States which want flexibility in defining some types of data;
  EPA' s media -based organizational structure;  program offices'
  reluctance to relinquish the control of  resources to another
  office;  and technical risks.

  Data  quality problems exist in many EPA  information systems,
  such  as  Superfund's main system5,  for which the OIG has
  reported significant inaccuracies for a  number of years.
  Another  example was provided by focus group participants, who
  said  that EPA's geographic information systems (GIS) showed
  certain  facilities in the middle of the  Atlantic  Ocean and at
  the North Pole.  These inaccuracies adversely affect both
  "business"  and environmental decision-making,  and are caused
  by such  things as frequently changing definitions of data,
  lack  of  "ownership" of the data by those who have to input
  it, lack of  a "data audit process" in the program offices,
  and EPA's inability to force the States  to  provide quality
  data .
  3 The Comprehensive Environmental Response, Compensation, and Liability Information System (CERCLJS) is Superfund's official
source of planning and accomplishment data and serves as the primary basis for strategic decision-making and site-by-site tracking of
cleanup activities.


                               viii


                              Report No.  E1SKG3-15-0098-4400038

-------
          Special Review of EPA's Information Systems Program
The OIG is making a number of recommendations affecting root
causes which cut across most, if not all, of the major IRM
problem categories.  By implementing certain recommendations,
the OIG believes the Agency can start to solve more than one
major problem.  More specifically, these recommendations
should go a long way in improving EPA's basic management
information about where Agency funds, including Superfund
money, are being spent and how, and what the resulting
accomplishments are.  The OIG has not prioritized these
recommendations, believing all to be so significant that they
must be addressed as soon -as possible.

A culture change is dramatically needed at EPA to view
information as a strategic resource and IRM as a core
function central to accomplishing the Agency's mission.  A
cornerstone of this change would be establishing a Chief
Information Officer  (CIO) with authority and responsibility
for the entire IRM program.  Other recommendations include
strengthening: the Senior IRM Official  (SIRMO) positions in
Headquarters and the Regions; the IRM organizations in the
program and Regional offices; and the Executive Steering
Committee for IRM.

The OIG is making recommendations which will substantially
improve IRM planning.  The IRM planning process should be
closely linked with the development of EPA's Business Plan,
and be integrated with the Agency's budget process.  Further,
the performance of information systems needs to be measured,
to quantify their support of the Agency's mission and goals.

EPA's management framework for building information systems
and increasing user involvement during the development
process needs to be strengthened.  Training mechanisms should
be established to close the expertise gap between those who
are knowledgeable about EPA programs and those who are
knowledgeable about IRM.

EPA needs an Agencywide information/data architecture and
data standards.  To help in this, a centralized Data
Administration function should be established with
appropriate authorities and responsibilities.  Finally, the
public should be given greater access to view EPA's non-
sensitive data.
                              IX

                           Report No.  E1SKG3-15-0098-4400038

-------
Special Review of EPA'a Information Systems Program
   (This page  intentionally left blank)
                 Report No.  E1SKG3-15-0098-4400038

-------
	Special Review of EPA's Information systems program

                      TABLE OF CONTENTS
                           VOLUME I

                                                         Page

EXECUTIVE SUMMARY	   i

CHAPTER

  1  INTRODUCTION	*	   1

     PURPOSE AND OBJECTIVES	   1

     SCOPE, METHODOLOGY, AND ASSUMPTIONS	   1

     BACKGROUND	   2

       Structure Of  IRM At EPA	   3

       Previously Identified Problems And
       Recommendations	   4

       Status Of Implementing Prior Recommendations	   5

  2  INFORMATION RESOURCES MANAGEMENT PROBLEMS AND
     UNDERLYING CAUSES	   6

     IRM MANAGEMENT  AND ORGANIZATIONAL STRUCTURE	   8

       Upper Management Does Not Adequately Understand
       And Participate  In Information Resources
       Management	   9

       Confusion Exists About Lines Of Authority And
       Communication Regarding IRM Issues	  13

       IRM Organizational Placement, Qualifications, And
       Responsibilities Are Inconsistent Throughout
       The Agency	  15

       IRM Policies  Are Not Effectively Implemented	  16

     RESOURCE PLANNING  AND PERFORMANCE MEASUREMENT	  17

       EPA Needs An  IRM Strategic Plan Driven By An
       Agencywide Business Plan	  17
                           Report NO.   E1SKG3-15-0098-4400038

-------
	Special Review of EPA's Information Systems program

                                                       page

     The Agency Has Not Integrated The Planning
     and Budgeting Processes	  18

     Too Much Emphasis Exists On Short-Term Results....  20

     EPA Has No Performance Measurement Criteria For
     The IRM Program	  22

   INFORMATION SYSTEMS DEVELOPMENT	  24

     Information Systems Are Designed Without Up-Front
     Agreement On Necessary Components	  24

     EPA Does Not Follow A Standard Approach To
     Managing The System Development Life Cycle
     Of Information Systems	  28

   DATA MANAGEMENT	  30

     The Agency Has Neither A Standard, Defined
     Information Architecture Nor A Centralized
     Data Administration Function To Facilitate
     Data Sharing	  31

     EPA Needs To Improve Its Quality Assurance
     Program	  36

 3  RECOMMENDATIONS	  40

     IRM Management and Organizational Structure	  41

     Resource Planning and Performance Measurement	  43

     Information Systems Development	  45

     Data Management	  46
                         Report NO.  E1SKG3-15-0098-4400038

-------
          Special Review of EPA's Information Systems Program
APPENDIXES

     APPENDIX  I:
     APPENDIX II:
     APPENDIX III;


     APPENDIX IV:

     APPENDIX V:

     APPENDIX VI:


     APPENDIX VII;
                      TABLE OF CONTENTS
                          VOLUME II
                                                         Page
               SENATE SUBCOMMITTEE ON SUPERFUND,
               RECYCLING, AND SOLID WASTE MANAGEMENT
               REQUEST TO INSPECTOR GENERAL
               FOR A SPECIAL IRM REVIEW	 1-1

               SENATE SUBCOMMITTEE ON SUPERFUND,
               RECYCLING, AND SOLID WASTE MANAGEMENT
               REQUEST TO EPA ADMINISTRATOR
               FOR A SPECIAL IRM REVIEW	II-l

               JOINT AGENCY AND OFFICE OF
               INSPECTOR GENERAL TEAM MEMBERS	III-l

               FOCUS GROUP SUMMARIES	 IV-1

               INTERVIEW SUMMARIES	  V-l

               AGENCY PARTICIPANTS IN FOCUS
               GROUPS AND INTERVIEWS	 VI-1

               DIG REPORTS, GAO REPORTS, EPA
               MANAGEMENT REPORTS, AND
               CONGRESSIONAL TESTIMONY.	VII-1
     APPENDIX VIII: APPROACH AND METHODOLOGY	VIII-1

     APPENDIX IX:   IMPLEMENTATION STATUS OF PRIOR
                    HIGH-PRIORITY RECOMMENDATIONS	 IX-1
                    CHART:  EPA OFFICES WITH MAJOR
                    IRM RESPONSIBILITIES	  X-l
APPENDIX X:


APPENDIX XI:


APPENDIX XII:


APPENDIX XIII: REPORT DISTRIBUTION	XIII-1
                    OIRM  ISSUE PAPER:  STRENGTHENING
                    IRM AT  EPA	 XI-1

                    GLOSSARY  OF ACRONYMS AND
                    ABBREVIATIONS	XII-1
                           Report No.   E1SKG3-15-0098-4400038

-------
Special Review of EPA's Information Systems Program
   (This  page  intentionally  left blank)
                 Report NO.  E1SKG3-15-0098-4400038

-------
	Special Review of EPA*a Information Systems Program

                           CHAPTER 1

                          INTRODUCTION

PURPOSE AND OBJECTIVES

This  comprehensive management review of EPA's IRM program was
undertaken in response to a request from the Subcommittee on
Superfund,  Recycling, and Solid Waste Management of the
Senate Committee on Environment and Public Works  (see
Appendix I)6.  The Subcommittee's  request expressed concern
'about the "... apparent lack of credible basic management
information about where Agency funds, including Superfund
money, are being spent and how, and the questions raised
about what the resulting accomplishments may be ..."  In
addition, the Subcommittee sent a letter to the EPA
Administrator (see Appendix II), requesting "... full
cooperation and participation from fiscal and information
system as well as program officials ..." during the course
of the review.

Because of the strong base of prior audit work and studies in
this  area,  one objective of the review was to examine,
validate, and synthesize issues raised in prior 016 and 6AO
audit reports and Agency studies that relate to the
Subcommittee's interest.  The second objective was to
identify any additional mission-related IRM problems, their
root  causes,  and their solutions from an Agencywide
perspective,  with specific attention on the Superfund
program.

SCOPE, METHODOLOGY. AND ASSUMPTIONS

The scope of the review was extremely comprehensive.  It was
a management-oriented review that looked at "big picture,"
whole program issues, rather than specific details or
systems.  Because of the nature of the Subcommittee's request
and the time constraints for the project, the methodology and
approach reflected a special review, rather than a formal
audit, of EPA's IRM program.  This review relied heavily upon
group interviews and focus groups to most fully leverage both
review team and Agency resources within the allotted time.
'All appendices are included in Volume n of the report, which is available upon request.
                            Report No.  E1SKG3-15-0098-4400038

-------
	Special Review of EPA'a  Information  Systems  Program

The methodology, in brief, began with  creation of  a  joint
OIG/Agency review team, consisting of  representatives  from
the OIG, OIRM, and the Office of Solid Waste and Emergency
Response  (OSWER), which is responsible for the Superfund
program (see Appendix III).  Working cooperatively,  the team
developed  an agreed-upon project plan, and then extracted
relevant problem statements from audits issued from  fiscal
1990  through fiscal 1993.  The team used these problem
statements to begin discussions with interviewees  and  focus
groups.  The groups responded to the problem statements from
their perspectives, and then discussed the IRM problems they
encounter  directly in their jobs,  as well as the problems'
root  causes and  solutions.  These  discussions  brought  forth  a
mixture of facts and opinions; it  should be noted  that OIRM
officials  may not necessarily agree with all the expressed
opinions.  After concluding and summarizing the interviews
and focus  groups (see Appendices IV, V, and VI), the review
team  followed up where needed, and synthesized the results.
The review team  also used recently issued audit reports,
studies, and other documents  (see  Appendix VII), and current
audit work to identify or confirm  IRM  problems  and root
causes.  Finally, the review team  followed up  on a number  of
significant recommendations made in audit reports  issued
since fiscal 1990 to determine the status of EPA's corrective
actions.   Complete details of the  review team's methodology
are provided in  Appendix VIII.

The review embodied several assumptions.  One  assumption of
the methodology  was that the team  would obtain  the best root
cause statements and draft solutions from front-line program
office  and IRM representatives in  the  Regions.  The
methodology also assumed that balance  would be  obtained by
interviewing members of the Headquarters IRM community.  A
final assumption was that the joint review team approach
would help mitigate the review's time  constraints.

Due to  its limited scope,  this review  does not  represent an
audit in accordance with the Government Auditing Standards
(1988 revision)   issued by the Comptroller General  of the
United  States.   Our review was performed in accordance with
the procedures set forth in OIG Manual Chapter  150,  "Reports
of Review."

BACKGROUND

A sound IRM infrastructure and efficient, responsive
information systems are' critical to EPA's ability  to provide


                              Z~

                           Report  No.  E1SKG3-15-0098-4400038

-------
	Special Review of EPA's  Information  Systems  Program

objective, reliable, and understandable  information  to help
achieve its environmental mission.  EPA  is  an  information-
intensive organization, and its  information resources  are
critical to the  success of all program activities.   EPA has
over 500 information systems  as  well  as  computer models which
support its mission.   For instance, major information  systems
provide data:

  •  on management  and oversight of the  Super fund program;
  •  on nationwide  air pollution and  water  quality;
  •  to track and report on Agency financial management;
  •  for computerized  models  designed to evaluate acid rain
     and emissions; and
  •  on the effects of pesticides  and toxic substances on
     humans and  the environment.

The Federal government has been  making major investments in
IRM activities over the last  decade.   Annual expenditures now
exceed $20 billion  government-wide.   EPA also  is making
significant investments, totaling  an  estimated $265  million
in fiscal 1993.  This  figure  is  about five  times what  the
Agency spent ten years ago for IRM resources,  although EPA's
IRM budget has suffered cutbacks for  several years.

     Structure Of IRM  At EPA

The IRM program  at  EPA has both  centralized and decentralized
components.  The central components in OARM and OPPE provide
certain Agencywide  information management and  information
technology services, as well  as  IRM policy,  planning,  and
oversight.  The  decentralized components (i.e., the  program
and Regional offices)  implement  and operate systems  to meet
their needs.  While the responsibility to lead the use of
information and  information technology to best meet  the
Agency mission may  be  centralized, most  resources to
implement and manage information systems are decentralized,
residing in the  program offices  and Regions.

EPA's recent budgets have greatly  constrained  the centralized
components of EPA's IRM program..  This has  increased
pressures on the decentralized components to cover any
shortfalls in centralized IRM programs and  services.   Yet
these decentralized components are simultaneously receiving
their own budget cutbacks, essentially yielding a two-fold
cutback in IRM technology and services.  This  is especially
noticeable in the contracting area, because EPA's IRM  program
                               3

                            Report No.   E1SKG3-15-0098-4400038

-------
_ Special  Review of EPA's  Information  Systems  Program

relies  heavily  on  contractors  for  essential  technical
expertise.

Certain pressures  on the IRM program over the  past  few years
have  increased  the trend towards decentralization.   For
example, prior  problems with administration  of Agencywide  IRM
contracts  have  led program offices to proceed  rapidly  to
establish  their own contracts  for  essential  IRM services.  As
another example, because of advances in local  area- network
 (LAN) technology,  Headquarters and Regional  off ices now
operate and maintain their own basic desktop computing
platforms.  Technology is promoting decentralization of IRM,
and is  changing the role of the centrally-managed computing
resources .

The IRM program has recently faced significant challenges,
and is  working  to  overcome them.   IRM planning and  security
have  been  declared to the President as material weaknesses
for the Agency.  Corrective actions are underway in both
areas,  are progressing well, and present the promise of
fundamental improvements in the Agency's management of
information.

Nevertheless, significant problems remain.   For example, the
Agency  often manages information and information systems
reactively.  Congress asks questions of EPA  that EPA' s
systems cannot  easily answer.  In  part, this stems  from IRM
considerations  not  being addressed early enough during
development of  laws and regulations.  Some reactivity  derives
from  strict Congressional mandates, which are  then  followed
by broad questions.  For example,  the Clean  Water Act's best
practicable technology approach did not lead EPA to create
information systems whose primary  purpose was  to track
nationwide improvement in ambient  water quality.  It led EPA
to create information systems  to track permitting and
compliance activities.  Legislation that guides EPA towards
strategic use of information,  such as Title  III of  the
Superfund Amendments and Reauthorization Act of 1986,  is
rare.  This issue,   and many others, are discussed in this
report .

     Previousl  Identified Problems A*>d      mTnndai
Appendix VII provides a comprehensive list of reports,
studies, correspondence, and Congressional testimony from
fiscal 1990 through fiscal 1993 that are relevant to the
scope of the review.  These establish a solid foundation for


                              4

                           Report No.  E1SKG3-15-0098-4400038

-------
	Special Review of  EPA's  Information Systems  Program

the review and identify numerous problems  and
recommendations.

The review team categorized  the previously identified
problems into five major areas of  concern:

  •  organizational accountability;
  •  planning and resources;
  •  data integration and  data management;
  •  quality assurance and protection  of data; and
  •  communications throughout the IRM community.

The prior reports and studies present  recommendations  for all
of these areas, although the recommendations  tend to focus on
specific systems and selected components of the Agency IRM
program.  The review team  identified the most important of
these recommendations, and followed up to  see how fully the
Agency had implemented them.

     Status Of Implementing  Prior  Recommendations

Appendix IX presents details of the Agency's  efforts to
implement these high-priority prior recommendations.   In
general, the Agency shows  good progress in  implementing
certain recommendations  (e.g., those dealing  with the
Resource Access Control Facility  [RACF] and mainframe  access
security).  However, in some instances, resource constraints
appeared to be slowing progress somewhat  (e.g., in
establishing centralized data administration).  In addition,
some recommendations were  very long-term by nature, and are
not yet expected to be completed.   The recommendations for
improving aspects of IRM planning  are  a good  example of this.

All the specific recommendations were  designed to correct
particular previously identified problems.  The results show
that progress is occurring on numerous individual fronts.
The intent of this review, however, was to  discover root
causes of the most fundamental, cross-cutting IRM problems,
and to craft program-wide  solutions affecting the entire
Agency.  The remainder of  the report looks  at these broader
issues.
                              5

                           Report No.  E1SKG3-15-0098-4400038

-------
	Special Review of EPA'a Information Systems  Program

                          CHAPTER 2

          INFORMATION RESOURCES MRNAgEMPMT PROBLEMS
                    AND UNDERLYING CAUSES

The purpose of this chapter is not to simply list  IRM
problems, because they have already been well identified  in
various audit reports, studies, and testimonies.   Problems
are briefly discussed here only to provide a context for  the
analysis of root causes.  This, in turn, will serve to
document the need for the solutions and recommendations
presented in the next chapter.

Since fiscal 1990, 75 DIG and GAO reports, internal EPA and
contractor studies, and Congressional correspondence and
testimonies have criticized EPA for information systems
management problems, most of which the Agency is in the
process of addressing.  These problems may be categorized as
follows:  (1) concerns about the quality, integrity, and
completeness of data  (i.e., a lack of credible basic
management information about where and how Agency  funds,
including Superfund money, are being spent, and the
accomplishments resulting from those expenditures);
(2) difficulties in addressing cross-media pollution
problems; (3) significant cost overruns and delays in
developing and implementing information systems;  (4)
development of duplicate systems; and (5) exposure of  the
Agency's financial payment systems to unnecessary  access
risks.

For example, data quality and integrity problems have  been
particularly noticeable in EPA's Comprehensive Environmental
Response, Compensation and Liability Information System
(CERCLIS).  CERCLIS is the official source of planning and
accomplishment data and serves as the primary basis for
strategic decision-making and site-by-site tracking of
cleanup activities for the multi-billion dollar Superfund
program.  Yet,  in spite of its importance, the DIG has
repeatedly found major problems with CERCLIS, raising  serious
questions about the credibility of its financial and
accomplishments data.  For example:
                              6-

                           Report No.  E1SKG3-15-0098-4400038

-------
  	Special Review of EPA's Information Systems Program

    •   The last  four Superfund annual reports to Congress7
        have shown substantial error rates  on many crucial
        reported  program accomplishments, such as removal
        actions.

    •   A March 1990 OIG report8 concluded that  inadequate
        controls  existed over system documentation, software
        changes,  and testing.  The  report pointed out that the
        lack of these program quality assurance procedures led
        to unreliable reports which,  in turn,  caused users to
        abandon use of CERCLIS.

    •   A March 1992 OIG followup review9 of the 1990  report
        identified significant financial data errors,  as well as
        missing outlay data and audit trail  information  in
        CERCLIS.

  On March 29, 1993, GAO officials summarized EPA's Agencywide
  IRM  problems in a testimony before the Subcommittee on
  Legislation and National Security and the Subcommittee on
  Environment, Energy,  and Natural Resources,  Committee on
  Government Operations,  House of  Representatives.  One
  official stated that

        EPA is an agency with hundreds of information
        systems that are mostly separate and distinct, with •
        their own structures and purposes.   This plethora
        of systems impairs EPA's ability to  easily share
        mutually  beneficial information across program
        boundaries,  fosters data duplication,  and precludes
        more comprehensive,  cross-media assessments of
        environmental risk and solutions.  EPA's managers
        and analysts find many of the agency's automated
        systems too difficult to use or ill-designed to
        measure and assess environmental results, because
        few were  designed for this  purpose.   Instead,  the
  7 "Review of the Fiscal Year 1988 Superfund Report to Congress" (OIG Report No. E1SFF9-11-0015-0100227, issued March 28,
1990); "Review of the Fiscal Year 1989 Superfund Report, to Congress" (OIG Report No. E1SFFO-11-0018-1100026, issued October 18,
1990); 'Annual Superfund Report to the Congress for Fiscal 1990" (OIG Report No. PlSFFQ-11-0032-1100385, issued September 16,
1991); "Annual Superfund Report to the Congress for Fiscal 1991" (OIG Report No. PlSFFl-11-0026-2100660, issued September 30,
1992).


  ' "Report on CERCLIS Reporting" (OIG Report No. E1SFF9-15-0023-010087, issued March 12, 1990).
  9 "C
   "Special Review On Follow-up Of CERCLIS Reporting And Post-Implementation" (OIG Report No. E1SFG1-15-5001-2400027,
issued March 27, 1992).


                                    7~


                                Report No.  E1SKG3-15-0098-4400038

-------
	Special  Review of  EPA's  Information Systems Program

      information systems with  the  bulk  of  the agency's
      data  contain  activity-related data--chiefly
      designed to record,  count,  track,  and report  on
      such  items  as the number  of permits issued, levels
      of pollutants discovered, or  types of enforcement
      actions  taken.  Additionally,  data quality  and
      integrity remain a  chief  concern because of
      inattention to strong quality assurance  and data
      administration practices.   These problems,  I  might -
      add,  are not  necessarily  unique to EPA,  but are
      common across many  federal  agencies.

Our  current review has concluded that the  problems set forth
above are  the result of  systemic deficiencies in the
following  four general areas  (which combined  and expanded  on
the  five previously identified major areas of concern
identified on page 8) :   (1)  IRM  management and organizational
structure;  (2) resource  planning and performance measurement;
 (3)  information  system development; and (4) data management.
These underlying causes  are discussed in more detail in the
following  sections.  The causes  are not addressed  in any
priority order.  Instead, these  deficiencies  are all
considered to be significant barriers to an efficient and
effective  IRM program and should be addressed concurrently.
The  OIG's  recommendations for  correcting these deficiencies
are  set forth in Chapter 3 of  this  report.

IRM  MANAGEMENT AND ORGANIZATIONAL  STRUCTURE

EPA  has taken some steps  recently  to address  this  area of
concern.   For example,  the Assistant Administrator for
Administration and Resources Management met with the other
Assistant Administrators  in January 1994 to discuss  their
leadership role  in IRM and to  solicit their ideas.   Further,
OIRM has drafted a revised IRM Steering Committee  charter
which would establish the "Executive Steering Committee for
IRM"  as the decision-making vehicle for top management in  IRM
planning and  oversight.   The committee  has already met twice
since  the beginning  of the year.  Additionally,  the  Agency is
planning to take action  to address  some other organizational
and  communications issues as a result of OIRM's  "Model IRM
Study" to be  completed in May  1994.  Further,  the  Agency is
working toward converting hundreds  of contractor positions to
Federal employee positions, and  many of these are  tentatively
planned to be  ADP-related positions.  This will  contribute to
reducing the Agency's vulnerability to  the loss  of  IRM
expertise.


                               8

                           Report No.   E1SKG3-15-0098-4400038

-------
  	Special  Review of EPA's Information Systems Program

  We agree with these  actions and believe that they should go a
  long way in solving  the Agency's IRM problems and in
  achieving. Agencywide IRM goals.  Although we believe the
  Agency is generally  proceeding in the right direction, we
  still have concerns  about EPA's long-term ability and
  commitment to address these issues,  which will take years of
  continuing effort  to solve.  Our specific concerns are
  discussed below.


       Participate In  Information Resources Management

  During our interviews and focus group meetings, we repeatedly
  heard, from virtually all levels of Agency management and
  staff, that upper  management10 does not understand how
  critical IRM is to environmental protection.  This was a
  strong belief of the Headquarters focus group participants,
  who felt that senior management understands neither the
  linkage between information systems and mission
  accomplishment, nor  the role of data in framing options in
  the decision-making  process.  While this may not be the case
  with every top manager and executive at EPA, we nevertheless
  believe it is a very pervasive problem.

  Because of this, IRM is not treated as a core function,
  information/data is  not treated as a strategic resource,  and
  information/data is  not viewed as a valuable tool to empower
  the public  (i.e.,  environmental groups, labor, industry,
  etc.) to deal with environmental problems beyond what
  government can do.   Other Federal agencies--such as the
  Federal Aviation Administration, Social Security
  Administration, Internal Revenue Service, Patent and
  Trademark Office,  National Weather Service,  and Securities
  and Exchange Commission--have recognized the strategic value
  of information and treat IRM as a core business line.
  However, until EPA treats information as a key element of its
  business, it is losing the opportunity to use information to
  leverage its other resources.

  This is just one manifestation of a "culture" prpblem at EPA,
  which has relegated  to secondary importance a number of
  crucial management activities because they are not perceived
  to be sufficiently related-to EPA's environmental mission.
  10 When asked to define 'upper management," focus group members decided it was Headquarters Office Directors and above, and
Regional Branch Chiefs and above.

                                 9

                              Report No.  E1SKG3-15-0098-4400038

-------
	Special Review of EPA's  Information  Systems  Program

As  the  Inspector General testified on March 29,  1993,  before
the Subcommittee on Legislation and National Security,  and
the Subcommittee on Environment, Energy and Natural
Resources, Committee on Government Operations, House of
Representatives,

      Just  as  contract management at SPA has been viewed
      as secondary to the Agency's  environmental  mission,
      information resources management has also been
      pushed into the background and not given  the
      attention it deserves.  These are both problems in
      the "culture" at EPA which must be changed.

EPA's culture lias limited top management's central direction
and control of IRM.  This is exemplified by things like an
ineffective IRM Steering Committee which was not mandated to
make substantive decisions, inadequate IRM policies and
procedures, and insufficient communications within EPA's IRM
community, all of which we discuss later in this chapter.
Overall, the  lack of central direction and control-of  IRM has
significantly contributed to EPA's many long-standing  IRM
problems.

Part of this  problem is that top management officials  are not
perceived  to  be accountable for IRM.  Instead, focus group
members stated that upper management is accountable for
things  other  than IRM, primarily programmatic
accomplishments,  and as a result pays insufficient attention
to  IRM.  One  SIRMO stated that IRM is buried with other
resources management issues, helping to make management less
directly accountable for it.  Even the SIRMOs  are affected,
because almost all of them have many other major duties which
seem to take  precedence over IRM.

Another factor pertains to EPA's IRM Steering  Committee.
Such a  committee is the accepted method for top  management to
provide leadership and direction to ensure effective and
efficient use of information resources.  Yet,  historically,
EPA  rarely held any IRM Steering.Committee meetings (an
average  of 1  to 2 meetings a year  were held between fiscal
1985  and 1992, and only l meeting  in fiscal 1993), and when
they  were held the Assistant Administrators or their Deputies
did not attend.  During these meetings, very few decisions
were made because the committee acted almost exclusively as
an information-exchange body:  As  a result, EPA's IRM
Steering Committee has not effectively fulfilled what  should
have been an  important role as a decision-making body.


                              10

                           Report  Ho.  E1SKG3-15-0098-4400038

-------
  	Special Review of EPA's  Information Systems Program

  We are convinced that a culture  change is needed at EPA to
  recognize  the crucial role that  IRM  and information/data play
  in accomplishing the Agency's mission.   But changing EPA's
  current  IRM culture will require long-term commitment from
  top management.   For example, in 1992  GAO conducted a
  study11 on changing organizational  cultures, with the
  assistance of five academics who had researched the subject
  and of officials of nine large private industry companies.
  The study  reported that

       The experts generally agreed  that a culture change
       is  a  long-term effort that  takes  at least 5 to 10
       years to complete.  Company officials believe that
       two key techniques are of prime importance to a
       successful  culture change:

             Top management must be totally committed to
             the change in both words and actions.

             Organizations must provide training that
             promotes and develops  skills  related to their
             desired values and beliefs.

  One critical demonstration of such a culture  change would be
  the appointment  of a CIO for the Agency.   On  May 6,  1993, in
  testimony  before the Legislation and National Security
  Subcommittee,  and the Environment, Energy,  and. Natural
  Resources  Subcommittee, Committee on Government Operations,
  House of Representatives, GAO officials stated that

       Department  of Environmental Protection12  clearly
       needs strong, competent leadership and direction to
       tackle its  information management  problems.   The
       appointment of a Chief Information Officer who is
       familiar with the uses of information technology in
       simplifying and streamlining organizational
       practices and who can devote full-time attention to
       these issues offers real advantages.
  11 'Techniques Companies Use to Perpetuate or Change Beliefs and Values' (GAO Report No. GAO/NSIAD-92-105, issued February
27, 1992).

  12 GAO uses this term in discussing possible EPA elevation to cabinet level.


                                 11


                              Report NO.  E1SKG3-15-0098-4400038

-------
  	Special Review of EPA'a Information Systems Program

  Further,  a September 1993 GAO report13 stated

        Currently,  the agency's designated official for IRM
        has  many other responsibilities ...  We have
        previously testified that information management
        should not be viewed as a subset of facilities
        management or administration; it needs to be
        recognized and dealt with at a strategic level.  We
        further testified that having a Chief Information
        Officer at EPA who reports to the Administrator and
        can  devote full-time attention to IRM issues,
        combined with the adoption of proven, disciplined
        practices for managing information resources, is a
        sound investment and can provide major benefits.

  The  Assistant Administrator for Administration and Resources
  Management does not agree with appointing a separate CIO.   In
  response  to the September 1993 GAO report, he stated that as
  the  Designated Senior Official for IRM (DSO)  and Assistant
  Administrator for Administration and Resources Management,  he
  reports directly to the Administrator.   In this capacity, and
  because of his unique position of managing the Agency's
  budget and finances as well as IRM,  he believes he can
  effectively influence IRM investments at all levels to meet
  Agency goals.

  However,  we agree with GAO that EPA needs a separate CIO with
  full-time IRM responsibilities, because running an effective
  IRM  program is substantially more than just having
  "influence" over IRM investments.  Program officials we met
  with throughout Headquarters and in the Regions cited the
  lack of knowledgeable,  experienced,  forceful leadership in
  the  IRM arena as a significant problem,  and strongly
  supported the creation of a separate CIO.  Considering the
  hundreds  of millions of dollars of annual IRM expenditures
  and  the extent of IRM deficiencies identified in all the
  reports,  studies,  and testimonies over many years, creation
  of a separate CIO who can devote his/her undivided attention
  to IRM is imperative.   We believe it is unrealistic to expect
  the  Assistant  Administrator for Administration and Resources
  Management,  who is burdened with major responsibilities in
  other areas (i.e.,  Chief Financial Officer (CFO) duties,
  human resources  .management.,  acquisition ..management, grants
  " "EPA's Plans to Improve Longstanding Information Resources Management Problems" (GAO Report No. GAO/AIMD-93-8, issued
September 16, 1993).
                                12


                              Report No.   E1SKG3-15-0098-4400038

-------
_ Special Review of EPA's Information Systems  Program

and debarment administration,  facilities management  and
services, and other Agency management and administration
functions) to devote  the attention necessary to solving one
of the Agency's most  significant and oldest problems.
Furthermore, even though the current Assistant Administrator
for Administration  and Resources Management has taken  an
active interest in  IRM,  solving the Agency's IRM problems  is
a long-term activity  and will undoubtedly transcend  this
administration and, therefore,  deserves continuous and long-
term top management attention and involvement.   Finally, the
House bill on EPA cabinet-level elevation,  H.R.  3425
sponsored by Congressman Conyers,  specifically calls for a
separate CIO..

     Confusion Exists About Lines Of Authority And
                    Regarding IRM Issues
Although offices  with IRM responsibilities  generally
cooperate with one another on primary IRM matters,  formal
lines of authority for IRM matters are fragmented between
four offices under two Assistant Administrators  (see Appendix
X) which are not  in the direct chain of command  with one
another.  For  example:

  •  OIRM reports directly to the Assistant Administrator for
     Administration and Resources Management  and has the
     primary functional responsibility for  IRM..policy
     development  and overall management of  the Agency's IRM
     program.   (Note:  Appendix XI contains  a  detailed
     description  of EPA's IRM program,  provided  by  OIRM
     officials.)

  •  The Office of Regulatory Management and  Evaluation
      (ORME) , which reports directly to the  Assistant
     Administrator for Policy,  Planning,  and  Evaluation,
     evaluates and reviews all Agency information collection
     requests  and activities,  evaluates Agency management and
     uses of data for decision-making,  and  is piloting
     increased use of Electronic Data Interchange (EDI) .

  •  NDPD,  which  reports to the Director, OARM-Research
     Triangle  Park,  NC,  is responsible within the authority
     of OIRM for:  (1)  planning,  oversight,  management,
     operation, and acquisition of all ADP  resources in the
     Agency; and  (2)  computing and telecommunications
     services .
                              13

                            Report No.   E1SKG3-15-0098-4400038

-------
  	Special Review of EPA's Information Systems  Program

     •  The Information Resources  Management Division  (IRMD)-
        Cincinnati reports to the  Director,  OARM-Cincinnati, OH,
        and is responsible for operating the Agency's disaster
        recovery facility for highly sensitive systems under the
        technical direction of NDPD.

  This fragmented organizational  structure has been previously
  criticized by program officials and in recent audit reports
  and studies.  For example, a .1992 DIG audit report14 stated
  that

        The current fragmented structure does not lend
        itself to a centralized consistent approach to
        accomplishing Agencywide IRM support.  This has
        resulted in duplication of efforts,  failure to
        address key IRM areas, and confusion from the
        program offices as to who  is providing IRM
        leadership and direction ...  In our view, a
        direct chain of command between OIRM and NDPD is
        needed in order to establish strong centralized IRM
        leadership .  .  .

  Additionally,  a 1991 General Services Administration
  report15 predicted that the organizational separation of  IRM
  activities in EPA could create  problems in the future.

  The lack of an adequate communication infrastructure was
  consistently identified during  our review as a"root cause for
  many of the IRM problems.  Agency employees do not always
  know in what directions OIRM and NDPD are headed, and  these
  two components do not always solicit the input of users to
  their decisions and directions.   The perception of some
  program officials is that NDPD  and OIRM make unilateral
  decisions without consulting their "customers" or assessing
  the impacts.   For example,

     •  Some focus group members stated that in fiscal 1993 NDPD
        made unilateral decisions  about its  budget cuts
        affecting service to program offices, which resulted in
        considerable criticism.  Headquarters OARM officials had
        to spend a great deal of time listening to the reactions
        of the program offices and,  after the fact, learning
  " "COMPUTER SYSTEMS INTEGRITY: EPA Must Fully Address Longstanding Information Resources Management Problems"
(DIG Report No. E1NMF1-15-0032-2100641, issued September 28, 1992).


  13 "Information Resources Procurement and Management Review, Environmental Protection Agency" (issued 1991).
                              Report No.   E1SKG3-15-0098-4400038

-------
    	Special Review of  EPA's Information Systems Program

     what their most  important  needs were.   As a result,  NDPD
     restructured its service cutbacks.   Nevertheless,  some
     focus group members stated that NDPD repeated this
     process for fiscal  1994, making additional budget cuts
     without consulting  customers again.

     Some focus group stated that the NDPD officials  making
     architectural decisions do not  understand how computers
     are used in the  Agency  or  by the States/local
     governments.

     Many Agency officials also complained that they  were not
     consulted about  the drastic cutbacks in IRM training
     which will adversely impact their ability to implement
     new technologies.
IRM functions are not  treated consistently throughout  the
Agency  (i.e., organizational  placement,  and qualifications
and responsibilities of  staff).   For example,  an April  1993
OIG survey found the following.

  •  The responsibility  for IRM  within major Headquarters
     offices belongs to  individual  sub-offices in  five  cases,
     a  SIRMO in one case,  a LAN  administrator  in one office,
     an information management office in one case,  and  a
     program operations  office in another.

  •  The number of in-house ADP  staff in Headquarters program
     offices16 varies  from no  staff  in two offices  to 45
     staff in another  office.  The  number of in-house ADP
     staff in the Regions  varies from no staff in  two regions
     to 50 staff in another region.

  •  ADP functions are normally  performed by "computer
     specialists"  (job series "334"),  which is accepted
     government-wide as  the job  series qualified for
     performing these  types of duties.  However, based  on
     their job series, the qualifications of EPA's in-house
     ADP staff vary greatly.   For example,  the staff in the
     Headquarters program  offices occupy 18 different job
     series positions  with over  42  percent in  non-computer
" "Program offices" do not include OARM or OPPE.
                              15
                            Report No.   E1SKG3-15-0098-4400038

-------
       	Special Review of EPA.'s Information Systems  Program

        specialist positions.  Regional ADP staff occupy  11
        different job series positions,  with over 37 percent in
        non-computer specialist positions.   While we realize
        that some of these positions provide other types  of IRM
        support,  our concern is that the job series positions
        are inconsistent throughout  the Agency and, in many
        cases,  are completely unrelated to IRM (e.g.,
        Environmental Protection  Specialists,  Hydrologists,
        Scientists, Contract Specialists,  Clerks,..and
        Secretaries).
                                  .17
  According to a recent GAO study

        Neither the senior-level  IRM executive nor the
        organization can be effective without adequate
        staff resources imbued with  disciplined approaches
        for both managing the current technology base and
        carrying out strategic planning for new technology.
        Providing these resources  is difficult for agencies
        in lean budget times.  But this is a good place to
        make an investment since  the effective use of
        information technology can leverage major benefits
        in operational efficiency  and service to the
        public.
  A September 1992 OIG report18  identified that OIRM:  (1) had
  directives  that were incomplete and outdated; (2) had not
  developed certain key IRM standards;  and (3)  had not always
  distinguished between formal  or mandatory policies,
  standards,  and procedures, and informal or optional guidance
  when  issuing IRM documents.   In response to that report, EPA
  officials inventoried all their IRM policies, standards,
  procedures,  and guidance, created  a prioritized listing of
  additional  policies needed, and formally issued a number of
  key IRM  documents as temporary directives through the
  Agency's directives system.   They  are currently developing a
  number of policies, including revised systems life cycle
  guidance and information security-related policies.
  17 "Information Management and Technology Issues' (GAO Report No. GAO/OCG-93-5TR, issued December 1992).


  " "COMPUTER SYSTEMS INTEGRITY: EPA Must Fully Address Longstanding Information Resources Management Problems"
(OIG Report No. E1NMF1-15-0032-2100641, issued September 28, 1992).


                                 If


                              Report No.   E1SKG3-15-0098-4400038

-------
_ Special Review of  EPA's Information Systems  Program

While EPA is making progress in developing and updating  IRM
policy and guidance documents,  IRM directives are  not  being
adequately enforced and Agency officials  are  not always  aware
of these directives.  Agency officials  attribute this
primarily to insufficient top management  involvement and
oversight over  IRM  and  to the lack of adequate training.  In
the decentralized environment within EPA,  the IRM  community
requires an effective policy structure, complete with
adequate enforcement measures,  to direct  Agency IRM efforts
in the pursuit  of Agency goals.
RESOURCE PLANNING AMP PERFORMANCE

In response  to prior OIG and GAO report  recommendations, OIRM
created an IRM Planning Group (IRMPG)  responsible  for
implementing a corrective action plan.   This group prepared a
proposed IRM strategic planning process  which was  discussed
with the top Agency managers at a meeting in January 1994.
The proposed planning process includes a structured
methodology  and  involves input from a  wide range of internal
and external groups.   A high level IRM strategic planning
document is  scheduled to be  completed  by October 1994 .  This
document will contain a high level mission statement, vision
statements,  operating principles,  and  key implementation
strategies.   The key implementation strategies will define
overall Agency steps to be taken to implement the  vision
statements,  with scope,  timing,  and cost implications
discussed.   The  vision section of the  document is  planned to
be substantially complete in March 1994.   In addition, OARM
plans to pilot test a working capital  fund in fiscal 1995 and
implement it beginning in the fiscal 1996 budget cycle.  The
fund is intended to provide  stable,  multi-year funding for
selected Agency  resources and initiatives,  including IRM
initiatives.  However,  our review identified a number of
serious concerns which need  to be addressed in this IRM
planning and budgeting process.

     EPA Needs An IRM Strategic Plan Driven Bv An  Aoencvwide
     Business Plan

The need -for a comprehensive,  integrated IRM strategic
planning process which is tied to the  mission needs of the
Agency has-been  a long- standing concern  at EPA.  Part of the
problem in the past has  been that EPA  has not analyzed its
"business" processes and developed an  Agencywide Business, or
Strategic, Plan.  Prior to completing  an IRM Strategic Plan,
EPA needs to develop an Agencywide Business Plan which aligns


                              17

                           Report No.  E1SKG3-15-0098-4400038

-------
	Special Review of EPA's Information Systems Program

resources,  skills,  and strategies with mission objectives.
Where  the  strategies include information technology, they
should be  reflected in the Business Plan.  Without such a
Business Plan,  the effectiveness and efficiency of
information technology cannot be maximized.  The subsequent,
more detailed IRM Strategic Plan should then be tied to and
support this Business Plan.

According  to a GAO government-wide study19,

     Successful modernization is based on a strategic
     analysis of what the agency needs to accomplish,
     where it is now,  and where it must be at future
     points in order to meet its goals.  While most
     agencies have mission statements that define
     general goals,  many agency leaders do not follow
     through on the next step:  analyzing current
     business processes to learn where they are breaking
     down  and how they should be restructured to achieve
     fundamental,  long-term improvements in business
     practices and service to the public.  Such analysis
     is exactly what many successful private sector
     organizations are doing .  .  .  When an agency does
     not first analyze its business processes and
     determine where improvements should be focused, it
     short-circuits technology's potential to
     dramatically increase productivity and reduce
     costs.

According  to our discussions with officials in OPPE and in
OIRM,  EPA plans to complete an Agency Business Plan by March
1994,  and  a draft of a portion (i.e.,  a high level IRM vision
statement  for the Agency)  of the IRM Strategic Plan at the
same time.   These officials also told us that they are
coordinating with each other to ensure that both plans are
linked.

     The Agrencv Has Not Integrated The Planning And Budgeting
     Processes

EPA has not  established an integrated strategic IRM planning
and budgeting process  which is  updated on an annual basis to
help acquire, manage,  and use its computer resources.  As a
result, the  Agency  and its components lack a sufficient
 "Information Management and Technology Issues" (GAO Report No. GAO/OCG-93-STR, issued December 1992).
                           Report No.   E1SKG3-15-0098-4400038

-------
  	Special Review of  EPA'a Information Systems Program

  mechanism to make information system funding decisions, thus
  subjecting system development projects to funding shortages
  and to  incremental, independent design and development.

  A 1992  OIG report20 criticized the Agency for not
  integrating these two processes.   In addition, a recent
  Office  of Management and Budget (OMB)  report21 criticized
  the Agency for not using strategic plans to drive the budget
  process.   According to that  report,

       Several [EPA] managers  thought that planned
       integration efforts of  financial and related
       systems .could not be  supported by current funding
       levels.  However, there appears to be no intention
       to request special funding for future efforts that
       may be under-supported.   This also relates to the
       findings in strategic planning.  An effective
       strategic plan with resources identified, should
       drive the budget process and result in the
       proactive efforts to  secure  necessary resources,
       where warranted.

  Agency  officials agree that  this  linkage of IRM planning and
  budgeting is an important  issue.   IRMPG is performing a study
  to determine the methodology for  integrating the IRM planning
  and budgeting processes.

  One of  the more significant  barriers to the successful
  development and implementation of major information systems
  is unstable funding from one year to the next, due to the
  lack of a mechanism to ensure long-term financial support.
  The major reason for this  is that the current 1-year budget
  cycle is not consistent with multi-year technology
  development activities.  For example,  the Integrated Contract
  Management System and IFMS are long-term system development
  efforts which are highly dependent upon multi-year funding.
  In addition, system development projects are not line budget
  items--funding is hidden in  program budgets.  When program
  budgets get cut, the system  development projects usually
  suffer.
  20 'Computer Systems Integrity: EPA Must Fully Address Longstanding Information Resources Management Problems" (OIG Report
No. E1NMFM 5-0032-2100641, issued September 28, 1992).

  21 'Report on the Financial Systems Planning Process: The Environmental Protection Agency' (issued September 20, 1993).


                                 19


                              Report No.   E1SKG3-15-0098-4400038

-------
	Special Review of EPA's Information Systems  Program

Another benefit of long-term budgeting and accounting  for
information systems would be to establish accountability for
managers in the development, implementation, and maintenance
of major information systems.  It would provide a valuable
tool for decision-making and force managers to look at the
entire life cycle  (i.e., initiation, design, development,
implementation, operations, maintenance, and retirement) of  a
system.  If top management was aware of the life cycle costs
when making decisions for initiating long-term system
development projects, different decisions might be made.  For
example, for IFMS, the initial estimate was almost $8
million, but this was only for the design, development, and
implementation of the system.  Our current IFMS audit has
estimated that the costs of the entire life cycle of IFMS
over a 12-year period, which include maintenance and
operations costs  (such as timesharing and telecommunication
costs) not included in the original estimate, are actually
over $200 million.
A more balanced emphasis between long-term and short-term
results is needed.  At present, the emergency of the moment
is frequently the driving factor in much of EPA's operations.
Participants in one of the focus groups, commenting on the
rapid coming and going of "major" initiatives, jokingly
referred to them as the "initiative du jour."  While all
organizations have crises that must be addressed, it is the
perception that this occurs excessively at EPA, and has
helped create an enormous short-term focus very detrimental
to long-range planning, especially for information systems.

Continually changing Congressional information needs, new and
changing environmental regulations affecting changes in
information systems, and poor communications with
Congressional staff generally keep EPA in a reactive mode.
For example, EPA reports to at least 110 Congressional
committees and subcommittees; this level of oversight is
second only to the Department of Defense.  In addition, EPA
is currently tracking 576 bills introduced in Congress that
have environmental aspects.   In both of these areas,
oversight and pending legislation, EPA is frequently asked
for quick turnaround information and immediate answers to
                              20

                           Report No.  E1SKG3-15-0098-4400038

-------
  	Special Review of EPA's  Information Systems Program

  complex problems.  According to  the  GAO study mentioned
  above22,

       Pressure for quick solutions  to complex problems
       works against strategic planning--the heart of
       effective IRM.  The political process pushes both
       the  Congress and executive  agencies to focus
       strongly on achieving near-term results at the
       expense of the long term.

  The frequent issuance of new regulations,  and regular
  revisions of existing ones, adversely impact .Agency
  information systems.  For example, the  Agency's Permit
  Compliance System  (PCS) is increasingly cumbersome to use
  because it requires detailed understanding of over 2,500
  parameters for issuing permits;  many of these parameters have
  been introduced due to revised regulations.   Another example
  is the Agency's Toxic Release Inventory System (TRIS),  which
  cost only $300,000 to develop, but now  costs $1.2 million
  each year for maintenance due to frequent  regulatory changes.

  EPA's IRM Strategic Plan will also be subject to changing
  programmatic needs.  For example,  EPA's information systems
  reflect its media-specific legislative  and regulatory
  framework.   Yet during the tenure  of the prior Administrator,
  EPA began to put increased emphasis  on  a cross-media approach
  to pollution problems; the new Administrator has continued
  this approach.   But according to a 1992 GAO report23,


       EPA  cannot readily bring together  and correlate
       data from its various programs--such  as air,  water,
       hazardous waste,  and pesticides--in order to assess
       environmental risks comprehensively or identify and
       target the most important enforcement priorities.

  This deficiency demonstrates the need to better integrate
  program offices'  data and systems.   Therefore,  the IRM
  Strategic Plan must be flexible  enough  to  deal with these
  changing  programmatic needs.  As. long-term needs change,  the
  changes need to be reflected in  the  Plan,  and annually
  0 "Information Management and Technology Issues' (GAO Report No. GAO/OCG-93-5TR, issued December 1992).

  " "EPA Needs a Better Strategy to Manage Its Cross-Media Information" (GAO Report No. GAO/IMTEC-92-14, issued April 2,
1992).

                                 21

                              Report No.  E1SKG3-15-0098-4400038

-------
   	Special Review  of  EPA'3 Information Systems  Program

   reviewing and updating the Plan would facilitate this
   process.

        EPA Has No Performance  Measu**«*"*»"<: Criteria For The IRM
        Program

   EPA does not have adequate mechanisms in place to measure the
   overall performance of the IRM program.  One critical  aspect
   of performance measurement related to information-systems
   pertains to the information/data they contain.  The Agency
   does not currently build  into  its information systems
   sufficient measures of environmental successes,
   accomplishments, and economic  benefits.  Yet this is exactly
   the kind of information that Congress and the public want.
   In 1993 GAO testified to  Congress24 that

        EPA lacks not only performance measures but also
        the information necessary to establish these
        measures and to assess  the effectiveness of its
        programs in improving or  protecting environmental
        quality.  Although environmental programs are meant
        to clean up or prevent  unacceptable levels of
        pollution,  EPA has not  had the information with
        which to judge the success of its programs.

   EPA's Strategic Targeted Activities for Results System
   (STARS)  is the Agency's official  accomplishments reporting
   system.   However,  this system  accumulates activity or  "bean
   counting" measures (e.g., compliance rates,  permits issued,
   inspections completed, penalties  collected,  etc.)  rather  than
   trends/outcomes (e.g., how much have we improved the quality
   of water  or reduced air pollution?).   An OPPE official told
   us that  STARS II is currently  being planned and will get
   closer to identifying outcome-related information.   The
   system will link resources to  accomplishments and will
   represent all program offices.   Information systems can place
   this  information/data at a user's fingertips.  However, this
   system is still  in the conceptual stage.

  Another aspect of  program performance involves accounting for
  expenditures and the accomplishments resulting from them.  In
  the first place,  EPA cannot  adequately account for IRM
  expenditures ,on a  system by  system .basis.  In our current
  24 "Creation Of A Department of Environmental Protection", GAO Testimony before the Subcommittee on Legislation and National
Security and the Subcommittee on  Environment, Energy, and Natural Resources, Committee on Government Operations, House of
Representative on May 6, 1993.


                                 22


                              Report No.  E1SKG3-15-0098-4400038

-------
_ Special  Review of EPA's Information Systems Program

audit of EPA's  software maintenance activities,  preliminary
information  is  showing that EPA program officials and
information  system managers do not consider,  by system,  the
life cycle costs of major information systems which support
the mission  of  the Agency.   When asked for cost information,
most system  managers were only able to provide information  on
contract dollars spent by their program office in support of
the system.   For 10 major information systems reviewed in
that audit,  preliminary information shows  that EPA did not
correctly report almost $29 million in fiscal 1993 system
costs to OMB.   Instead of reporting these  costs by individual
system, EPA  reported them in the aggregate.

Nor can EPA  measure the success or failure of some critical
information  system activities.   For example,  for the  10
systems mentioned  above,  which account for approximately $18
million of reported annual  EPA IRM maintenance expenditures,
five Assistant  Administrators reported that their offices do
not measure  the success or  failure of software maintenance.
The other four  Assistant Administrators reported that  they
have success factors for software maintenance.   However, our
preliminary  evaluation of the factors is that the measures  of
success are  not quantifiable.

The importance  of  performance measures is  gaining long-
deserved recognition.   For  example,  the "Government
Performance  and Results Act of 1993"  requires that over the
next few years  EPA set program goals,  measure program
performance  against those goals,  and report publicly on their
progress.  Also, the report of Vice President Gore's National
Performance  Review stated that

     We must hold  every organization and individual
     accountable for clearly understood, feasible
     outcomes.  Accountability for results will  replace
     "command and  control"  as the way we manage
     government .  .  .  Agencies will start  measuring and
     reporting  on  their past goals and performance as
     part of their 1996 budget requests.

Finally, the Administrator's March 10,  1993,  memorandum to
all employees stated that

     To address resource management problems  I am taking
     three courses of action:  (1)  construction of a
     rigorous system of accountability;  (2) establish-
     ment of clear standards of discipline with serious
                            Report No.   E1SKG3-15-0098-4400038

-------
   	Special Review of EPA's Information Systems  Program

        consequences for failure to meet the standards; and
        (3)  development of an overall management scheme
        based on clear goals, work plans and milestones for
        measuring success.

   Without performance measures, EPA does not have an adequate
   mechanism for providing credible basic management information
   about where Agency funds are being spent on system
   development activities and the resulting accomplishments.
   EPA has made some progress in recent years in the information
   systems development area.  For example, the Agency has
   implemented the Systems Development Center, which is a
   contractor-run operation using formal, disciplined
   development methods to improve EPA systems25.  As previously
   discussed,  in response to a recent audit report, EPA
   officials have improved many of their IRM policies,
   standards,  procedures,  and guidance, and are working on
   others,  including their systems life cycle policies.
   Although we believe the Agency is generally proceeding in  the
   right  direction,  we have identified a number of barriers that
   EPA must address when developing and implementing these
   policies.

       Information Svp+-«»™g &re Designed Without Up-Front
       Agreement On Necessary Components

   A  critical  factor in the long-term success of an information
   system is the careful definition of all users' and customers'
   needs  early in its design.   This involves understanding who
   all  the  users and customers will be, and communicating
   effectively with them to identify what they will need from
   the  system.   These needs must then be addressed by the system
   builders.   Otherwise,  a number of adverse effects can occur,
   such as  the system not being used by many of its intended
   beneficiaries,  users who are responsible for entering data
   into the system not being motivated to ensure data quality,
   and  significant criticisms being leveled at the system by
   users  or customers who were not consulted during the design
   process.
  " The Agency has recognized the value of the Center and the process, and intends to continue them after expiration of the current
contracts.


                                 24


                              Report No.  E1SKG3-15-0098-4400038

-------
	Special  Review of EPA's Information Systems Program

In both our Headquarters and Regional focus groups  and
interviews, managers  and staff frequently stated that EPA's
systems developers do not always work effectively with users
and customers  to design useful systems.   These users  and
customers include  EPA program officials,  Congress,  and the
States.

     EPA Program Officials

Overall, systems developers were criticized for not really
understanding  how  EPA's programs work and what their  needs
are.  Some focus group members attributed this to EPA's  heavy
use of outside contractors for system development work.

Focus group members were also concerned that Agencywide  (or
"national") systems are usually designed  to meet only EPA
Headquarters needs, so they frequently do not meet  Regional
needs.  This is especially burdensome to  the Regions  when
they are required  to  devote the resources to input  much, or
all, of the data to these systems,  and yet cannot'effectively
use them to help manage their programs.   This impairs the
Regions' motivation to ensure the quality of the data in
these systems,  since  the Regions do not have a sense  of
"ownership" of the data.   Some Regional focus group members
added that these systems need to be designed to do  more  than
just generate  answers to questions for Headquarters
personnel; they also  need to be useful tools for Regional
managers and staff.

Other specific criticisms of many of the  national systems
also appear to be  due to inadequate consideration of  users'
and customers'  needs.  For example,  members of one  Regional
focus group discussed what they termed the national systems'
"low levels of utility and friendliness to Regional users."
A good example of  this was given by a SIRMO,  who said that
they could not do  something as simple as  go into IFMS and
find out what  their available travel balance was.   IFMS
contains lots  of data,  but it is not available and  accessible
in a user-friendly fashion.   So the typical EPA manager  must
ask an intermediary staff person to generate a report to find
the basic data needed.   Some of the reasons Agency  officials
gave for this  include the following.

  •  Better identification is needed of "who the system  must
     be friendly/useful for."
                              25

                            Report No.   E1SKG3-15-0098-4400038

-------
	Special Review of EPA.1 a Information Systems  Program

  •  The universe of potential system users has expanded  over
     time to include everyone with a personal computer on
     their desk.

  •  It costs more money to make systems really user-
     friendly.

  •  Program staff and systems developers do not work
     together on what is needed and what use is to be  made of
     the data input to the system.

  •  Headquarters, or Agencywide, needs are different  from
     individual Regional needs.

On the other hand, this same SIRMO cited TRIS as a success
story, in which the developers gave more consideration to
users' needs and user-friendliness.  Another example is the
development of the Public Water Supply System.

Some Regional focus group members also expressed concerns
about the complexity of the national systems, saying they
were overly complicated and too difficult for managers  and
staff to use.  They stated that the Regions frequently have
to develop their own systems because the national systems
cannot produce the reports the Regions need, changes to the
systems are difficult and take too long due to the process of
obtaining nationwide consensus, and, as previously mentioned,
the systems developers often have insufficient program
knowledge.

     Congress

Many focus group participants believed that EPA does not do a
good enough job of working with Congress to define their
needs and anticipate their questions.  Clearly, this
parallels the concerns expressed about understanding the
needs of the program offices.  Focus group members felt this
often led to situations in which Congress asked good,  •
straightforward questions about an environmental program's
progress or accomplishments, yet EPA's systems simply  could
not answer them.

     The States

This criticism is valid for the States, as well.  Focus group
members,  especially in the Regions, stated that the States
were too often not involved in the design of EPA's systems.


                              2-6

                           Report No.  E1SKG3-15-0098-4400038

-------
  	Special Review of EPA's Information Systems Program

  One SIRMO stated that they frequently hear from the States,
  local governments,  and even the public, that EPA data is not
  easy to access  arid that,  when accessed, does not serve their
  needs.  The  SIRMO attributed this to the fact that users and
  customers like  these groups were not involved in the design
  of the particular systems.  This can be especially
  troublesome  when the States are expected to input data to
  certain systems; as with the EPA Regions, when they feel a
  lack of ownership,  yet must devote resources to entering
  data, the motivation to ensure data quality is adversely
  affected.

  Overall, information needs analyses and cost/benefit studies
  are not being adequately conducted.  As a result, systems are
  being delivered to users that do not meet their needs, and no
  mechanisms are  in place to adequately identify costs and
  benefits for managers to make informed decisions about
  information  systems.  For example, a recent GAO report2* on
  the Office of Prevention,  Pesticides, and Toxic Substances
  (OPPTS) stated  that

       OPPTS'  officewide systems do not meet users' needs
       for screening large numbers of existing chemicals
       and setting review priorities.  Instead,  these
       systems largely serve either as an index to
       available  chemical studies or to track the progress
       of chemicals through OPPTS' review process.  These
       systems do not meet users' needs because OPPTS
       never  (1)  assessed and validated users'  needs for
       chemical data--such as the type of data needed or
       how quickly such data should be retrieved,  or
        (2) developed information systems to satisfy these
       data requirements.  Without such analysis of user
       needs,  OPPTS scientists and contractors must follow
       cumbersome and time-consuming manual steps when
       screening  large numbers of chemicals.

  It was clear in all our focus groups and discussions that
  developers of national systems need to better recognize the
  roles and needs of the Regions, Congress, and the States in
  implementing and maintaining information systems.
  * "EPA TOXIC SUBSTANCES PROGRAM: Long-standing Information Planning Problems Must Be Addressed" (GAO Report No.

GAO/AIMD-94-25, issued November 1993).


                                 27


                              Report No.  E1SKG3-15-0098-4400038

-------
	Special Review of EPA's  Information  Systems  Program

On the other hand, it was also clear that even when needs  are
fully defined during the systems design process,  completely
new questions or needs can be raised years later  which the
systems were legitimately not designed to address.  This,
too, has led to criticisms of EPA's systems, but  these
criticisms were generally viewed as unfair.  For  example,
members of one Regional focus group cited the  Resource
Conservation and Recovery Information System (RCRIS) as a
system on which the developers spent a great deal of money to
thoroughly identify all likely needs during its design.  They
believed it was an excellent example of a collaborative
effort between EPA, the States, and other users to define
requirements.  Nevertheless, with  the passage  of  time,  new
requirements arose which have been difficult to implement,
resulting in complaints that the system does not meet  users'
needs.
The Federal government and the private sector generally use
standard, internally consistent approaches to developing and
maintaining information systems.  A critical part of the
standard system development life cycle (SDLC) process is an
adequate requirements analysis, discussed in the preceding
section.  The SDLC process also addresses other key
components of systems development and maintenance, such as
risk analyses, system decision papers, test plans, user
manuals, and operations and maintenance manuals.  In fact, a
standard approach to the entire life of information systems
is so important that OMB Circular A-130 specifically requires
compliance with Federal SDLC standards, unless an agency can
show it is cost-beneficial to use an alternative approach.

However, a recent OIG survey conducted as part of our audit
of EPA's software maintenance practices showed that the vast
majority of Headquarters program and Regional offices do not
                              28

                           Report No.  E1SKG3-15-0098-4400038

-------
  	Special  Review of EPA's  Information Systems Program

  use  a standard approach to manage  software development and
  maintenance.  The  responses to the survey indicated:

    •   88 percent of Headquarters program offices and 60
        percent of Regional offices do not utilize a standard
        SDLC methodology27 when designing,  developing, and
        modifying information systems.

    •   77 percent of Headquarters program offices and 70
        percent of Regional offices do not utilize a mandatory
        SDLC approach28 in managing information systems.

  The  primary reason for this appears to  be that the majority
  of EPA's system development projects  are contracted out,  and
  contractors are allowed to use their  own system development
  methodologies which vary widely and do  not necessarily meet
  minimum Federal requirements.  In  addition,  contracting and
  project officers responsible for monitoring these contractors
  do not always have the knowledge and  ability to effectively
  determine whether  the contractors  are meeting minimum Federal
  standards.

  Another important  reason is that EPA's  IRM policy, standards,
  and  guidance documents relating to the  SDLC are not being •
  adequately enforced,  and Agency officials are not always
  aware of these directives.  As previously mentioned,  focus
  group and interview participants attributed this primarily to
  insufficient top management involvement and oversight over
  IRM  and to the lack of adequate training.

  Some of the steps  in the standard  SDLC  process are perceived
  to add time to the development of  a system.   Consequently,
  some organizations tend to perform them only cursorily,
  shortcut them, or  avoid them entirely.   This can lead to
  system problems which must then be corrected later in the
  process.  Participants in one focus group summed it up by
  stating that "management often will not let  us (i.e.,  give us
  the  time to) do it right,  but there is  always time to do it
  again."  This sometimes hasty approach  to systems development
  can  also lead to duplicate systems.   An example of this was
  given by a Headquarters focus group participant,  who said
  " Includes feasibility studies;cost/benefit analyses; requirements and conversion studies; decision papers; accepted programming
techniques; test plans; program, system, and acceptance testing; etc.

  a A management approach which continuously manages and evaluates an information system over the entire life cycle (i.e., design,
development, implementation, operations, maintenance, and extinction).


                                 29

                               Report No.  E1SKG3-15-0098-4400038

-------
  	Special Review of EPA*a Information  Systems Program

  that a Congressional mandate to report on contaminant
  sediments by a certain date led to the creation  of  an
  "emergency data base," despite the fact that a system was
  already in development which could have provided the data but
  would have taken longer than the mandated date to complete.

  Information systems need to be periodically evaluated as part
  of  the SDLC process to assess whether they are obsolete.  EPA
  does not,  however,  consistently do this.  For  example,  in our
  audit of software maintenance activities, we found  that 4 of
  the 10 systems we reviewed are over 7 years old,  which is one
  of  the Federal criteria29 for making a system a candidate
  for redesign.   Although EPA's IRM Policy Manual  requires
  program officials to periodically review all software
  resources to identify obsolete software and evaluate what to
  do  about it,  we found no evidence that EPA had assessed the
  obsolescence of these four systems.

  Obsolescence can make software maintenance more  difficult and
  costly.   For example, CERCLIS is almost seven  years old and
  is  maintained on an archaic data base management system,
  System 2000.   We conducted an informal survey  of five
  technical recruiting and consulting firms to determine the
  availability of programmers with System 2000 skills and
  industry demand for such skills.  None of the  firms has had  a
  request  for such skills in the last five years.   These firms
  characterized System 2000 as archaic and obsolete,  and said
  it  is difficult to find programmers with current System 2000
  experience.   As a result,  maintaining CERCLIS  is very costly.
  For example,  fiscal 1993 maintenance costs for CERCLIS were
  estimated at $3.9 million.

  DATA MANAGEMENT

  EPA has  begun a number of initiatives in recent  years to help
  its information systems better address issues  such  as
  environmental  monitoring and environmental risk  assessment.
  These initiatives include the creation of an Information
  Management/Data Administration (IM/DA)  group in  May 1992.
  The IM/DA group is  beginning to develop an IM/DA discipline
  and infrastructure  at EPA.   This will serve as the  backbone
  for coordination of integrated development and/or use of
  shared data  and, information systerns.-aligned.with EPA goals
  29 Federal Information Processing Standards Publication 106 defines programming code over seven years old as one of the characteristics
which makes systems candidates for redesign.


                                30


                              Report No.  E1SKG3-15-0098-4400038

-------
  	Special Review of EPA's Information Systems Program

  and strategies.  As such,  it should enable  integration of
  data and information  systems across environmental programs.

  Also, EPA has initiated projects to promote data sharing.
  For example,  the Agency developed and enhanced the Facility
  Index System (FINDS) , which provides a central index to the
  facilities tracked by various EPA databases--RCRIS,  CERCLIS
  PCS, THIS,  and others.   In 1990, EPA initiated a project
  known as Gateway/Envirofacts that is expected to apply GIS
  technology.30  The Gateway project was initiated primarily
  to research technical aspects of cross-media  data
  integration,  and is still a developmental system.   A new
  database,  Envirofacts,  was developed under  the Gateway
  project  and contains  excerpts of individual program systems
  databases--PCS, TRIS, CERCLIS,  and FINDS.   EPA plans to use
  GIS mapping capabilities with the Gateway/Envirofacts user
  interface to display  information geographically.   Another
  project,  the Integrated Data for Enforcement  Analysis .(IDEA),
  was initiated in 1990 to integrate data across programs to do
  analysis by industry, by corporation, and by  facility31.

  These efforts are producing valuable integration tools.
  However,  the Agency still needs to address  certain barriers
  to efficiently share  mutually beneficial information across
  program  boundaries, eliminate data duplication,  and allow for
  more comprehensive, cross-media assessments of environmental
  and health risks and  solutions.

       The Agency Has Neither A Standard, Defined Information

       Function To Facilitate Data Sharing

  Data sharing is crucial to the Administrator's four top
  priorities: ecosystem protection; pollution prevention;
  environmental justice/equity; and,  less directly,  partnership
  building.   These priorities cut across all  environmental
  media, but the link between data sharing and  these priorities
  has not  been widely recognized.

  To successfully share or integrate data on  a  large scale,  a
  standard,  defined "information architecture,"  along with
,  * A GB is a computer system that captures, stores, displays, analyzes, and models natural and artificial environments using data
referenced to locations on the earth's surface.

  31 Certafn aspects of IDEA (i.e., documentation and testing of the system software, and training users) have been reported to the
President as material weaknesses for the last two years. However, this should not detract from the intent of the system.


                                  31

                               Report No.  E1SKG3-15-0098-4400038

-------
  	Special Review of EPA's Information Systems  Program

  organization-wide data standards, are necessary.  An
  information architecture models the data and processes of  the
  organization,  and presents a stable basis for developing the
  processes  and information used by the organization to
  accomplish its mission.   The data architecture, a major
  component  of an information architecture, is a key, but
  expensive,  ingredient in determining the data and application
  software needed to manage the organization's programs.  Part
  of developing the data architecture involves establishing
  data  standards,  which means setting out official,
  organization-wide definitions of what terms mean and how they
  are to be  entered into information systems.  A final key
  factor is  having a strong,  centralized Data Administration
  function to oversee the development and implementation of  the
  data  architecture and data standards.

  However, EPA has not established such an Agencywide
  information  architecture,  sufficient data standards -,  or Data
  Administration function.   Although the formation of the IM/DA
  group is a step in the right direction,  this OIRM group has
  no representation from EPA program offices.  Furthermore,
  this  group has not been established as a permanent
  organization and does not have adequate funding, staffing
   (five staff members),  and authority needed to centralize
  control over the Agency's decentralized systems development
  activities.

  As a  result,  EPA systems  continue to be independently
  designed and managed and  lack both EPA and government-wide
  data  standards (e.g.,  standard definitions, common user
  interfaces,  core data sets,  and analytical tools).   A  recent
  GAO report32 stated

        EPA's  lack of sound  data management practices has
        resulted in difficulties in accessing and sharing
        data.   Supporting EPA's new strategies calling for
        cross-media assessments of environmental and health
        problems  requires data management practices that
        will  allow users to  easily access data from
        multiple  existing systems.  This will require the
        development of standard definitions, common user
        interfaces,  core data sets,  and analytical
        tools  .  .  .
  32 "EPA's Plans to Improve Longstanding Information Resources Management Problems" (GAO Report No. GAO/AIMD-93-8, issued
September 16, 1993).


                                32


                              Report No.   E1SKG3-15-0098-4400038

-------
	Special Review of EPA's Information Systems  Program

Despite the activities  we have outlined above,  focus group
and interview participants generally believed that  limited
national initiatives  exist to support data sharing  and data
management.  As a result,  some Regions are embarking on their
own integration initiatives.   For example,  members  of  one
focus group told us that two States  in their Region are
moving away from a single-medium approach to pollution
prevention, and one of  these States  is integrating  its
information regarding air,  water,  and permit activities for
its top 400 facilities.

Focus group and interview participants identified a number of
important reasons why data sharing has not yet  been pervasive
at EPA, and why these critical elements of successful  data
sharing are not present.

     Resource Constraints

Lack of resources has contributed heavily to EPA's
inattention to building a good information architecture with
shared/integrated data.   In fact,  participants  in focus
groups and interviews commented that budget items for
implementing data standards are the  first ones  cut  when IRM
cutbacks occur, severely harming data integration efforts.
One example given of  harmful  budget  cuts was the State/EPA
Data Management Program,  which some  officials called
excellent.  They said that State and EPA people had worked
together on small multi-media types  of projects, and the
participants were very  enthusiastic.   The project apparently
facilitated very good State-EPA communications.  However,  it
was suspended due to  budget cutbacks in fiscal  1993.

     Differing Views  of  Data

The other major cause,  discussed at  length earlier  in  this
chapter, is that the  Agency culture  does not value  data.  Its
strategic importance  to the Agency is minimized, so large-
scale efforts to maximize its usefulness have not been deemed
warranted.

Also, some Agency components  view the current management of
data as working against  data  standards and,  therefore,  data
sharing.  For example,  members of one focus group described
how some definitions  of  programmatic outputs continually
change.  OIG audits have seen this first-hand,  particularly
in the Superfund program,  where official definitions of some
programmatic accomplishments,  such as the completion of a


                              33

                            Report No.   E1SKG3-15-0098-4400038

-------
	Special Review of EPA's  Information  Systems Program

removal action  or a preliminary  assessment,  can change each
year.

Superfund  officials in one focus group  further stated that
the premise  that all Superfund sites have  common traits is a
false assumption.  Instead, they said that each site  has
unique characteristics that make nationwide  definitions,  or
standardization, difficult.

In addition,  some program officials view their data very
parochially.  Some are only interested  in  their own
information  needs, and so are reluctant to spend their
resources  on behalf of secondary users  or  customers.

Finally, participants in one focus group also  mentioned that
even the States want flexibility in how some terms are
defined.   This  is due, at least  in part, to  the fact  that the
States may already have their own definitions  of data and do
not want to  change them just to  satisfy EPA's  individual,
compartmentalized systems.  This creates inconsistencies  and
makes data aggregation very difficult.  The  large number of
users  (e.g.,  EPA officials, Congress, the  States, the public,
industry,  even  foreign governments, groups and individuals,
etc.) for  many  of EPA's systems  and/or data  greatly
complicates  EPA's ability to establish  and enforce data
standards.

     Media-based Approach

The media-based organizational structure of  the Agency does
not encourage information system designs which use data
integration  to  address risk-based environmental  problems.
Furthermore,  systems tend to lack common data  elements (e.g.,
facility identifier numbers)  which would facilitate use
across organizational lines.   Instead,  systems are developed
for .single media concerns.

Focus group participants mentioned a number  of causes for
this, such as the needs from media to media  are different,
and the separate media-based programs do not encourage cross-
media communications.  However,  the fundamental  cause is  that
these systems are based on Federal statutes  that are
themselves compartmentalized with distinct budget
authorizations,  allocations,  and oversight,  and so drive  the
Agency in that direction.
                              34

                           Report No.  E1SKG3-15-0098-4400038

-------
	Special  Review of EPA's Information Systems Program

On the other hand,  one group noted that while the statutes
are, indeed, partly to blame,  EPA also must share the blame.
EPA has not been very aggressive in pursuing data
integration, yet there is nothing in these statutes  which
prohibits this.

     Program Offices'  Desire For Contro?

EPA managers and staff told us that program offices  are
reluctant to relinquish resources and the control of those
resources to another office.  This creates risks  (i.e., the
potential loss of  those resources) .  Other risks  are also
involved, such as  the technical risks to a program office of
having its systems development efforts combined with those of
another office.

At the heart of  this is the issue of trust.   Participants in
one interview group told us that even when data integration
efforts do take  place,  such as FINDS,  they are inadequately
supported by the Agency,  and the program offices  become
reluctant to participate.

For example, the Office of Enforcement (OE)  is aware of the
Agency's data integration activities and met with IM/DA over
a year ago.  But because of concerns about the speed of the
data integration group's progress and about  the size and
capability of the  group,  OE has taken the initiative to begin
development of its own data dictionary and data standards.
Yet, some OE systems (e.g.,  IDEA and the  Enforcement Docket
System) are national systems that depend on information from
other Agency systems,  and would benefit greatly from an
Agencywide data  dictionary and common data standards.

Without an Agencywide information architecture and data
standards, the ability of Agency components  to access and
share information  needed to accomplish EPA's mission will
continue to be inhibited.   Further,  it is doubtful that
Agencywide data  sharing,  which will require  the enforcement
of common data standards throughout EPA,  will ever be
realized unless  the enforcement becomes centralized  under one
group.  A strong data administration office  is critical to
the success of the Agencywide  data integration goals.  The
director of this office (i.e.,  the Data Administrator) needs
to be a strong leader with knowledge,  experience,  and
authority, and should report,  preferably,  directly to the
Agency's CIO.
                              35

                            Report No.  E1SKG3-15-0098-4400038

-------
          Special Review of EPA's Information Systems Program
EPA is an information/data-intensive agency, and accurate,
complete, and timely data are critical.  Agency managers must
use an enormous amount of data for both "business" (i.e.,
financial, contracting, etc.) and scientific decision-making.
Other interested parties, such as Congress, other Federal
agencies, and the States, also make decisions based on EPA's
data.  Much of this data comes from EPA's information.
systems.  However, problems have been previously reported
with the quality of both the data in these information
systems, and the processes used to build them.

It should be noted that some of the participants in our
review felt that the data in EPA systems were sometimes
unfairly criticized.  They pointed out that EPA systems have
generally been designed for a specific purpose and usually do
not support secondary users  (e.g., Congress and the general
public) adequately.  When secondary users query EPA systems,
the desired information is very seldom available.  This
creates the perception that the data is bad, even when the
data may be fine for the primary users.

Regardless of whether the data truly is good or bad,  we
learned that EPA employees and Congress have,  over time,
stopped asking many important questions because EPA simply
does not have the data to answer them or cannot get the
needed information out of its systems.  As a result,  we were
told that making decisions based on limited amounts of
information has become a general course of business.   For
example, some focus group members stated that

  •  The lack of an interface between CERCLIS and IFMS has
     made it difficult to match cost data in IFMS with
     obligations and planning data in Superfund.  As a
     result, EPA may be missing opportunities to recover
     Federal money spent in the Superfund program.

  •  PCS provides information on wastewater permit limits
     versus discharges to assess compliance.  It does not
     bring in wetlands, population impacts, ecosystems, and
     the proximity of discharges to well supplies.
     Therefore,  permit and enforcement decisions may be based
     on limited available information.
                             36-

                           Report No.  E1SKG3-15-0098-4400038

-------
  	Special Review of EPA's Information Systems  Program

    •  The  Storage and Retrieval of Water Quality System
       cannot show how much contamination is coming  into a body
       of water.

  In addition to not having some information available,  the
  data accuracy of EPA's  current information systems has been
  criticized.  For example,  a recent GAO report" stated  that

       After more than 3  years of effort and systems
       investments costing more than $14 million, EPA
       cannot easily assemble accurate, reliable, and
       complete information on chemicals in the
       reregistration process.   This has compounded the
       already difficult  task facing the agency in meeting
       pesticide reregistration deadlines imposed by the
       Congress . . . EPA's information management
       problems are traceable to inadequate systems
       planning and poor  data management .  .  .  the
       consistency, accuracy,  and completeness of
       information within the new systems is also
       jeopardized because EPA has not implemented data
       management procedures to ensure data consistency
       and  quality.

  Another example was presented in a 1993 DIG report34, which
  questioned 30 percent of a sample of 650  Superfund
  accomplishments claimed in CERCLIS.   The  entire gamut  of
  cleanup achievements was involved,  in each of the four
  Regions audited.  The report attributed the problem to:
  (1) official definitions of accomplishments were not met;
  (2) source documents could not be located to support the
  claimed accomplishments;  (3)  the accomplishments were
  duplicatively claimed;  (4)  the accomplishments were claimed
  in the wrong fiscal year;  and (5)  there were data entry
  errors.

  Finally,  participants in our focus groups and interviews
  cited examples of data  quality problems in other systems,
  such as:
  » -PESTICIDES: Information Systems Improvements Essential For EPA's Reregistration Efforts" (GAO Report No. GAO/IMTEC-93-

5, issued November 23, 1992).

  « 'Consolidated Report Regarding Fiscal 1992 CERCLIS Data' (DIG Audit Report No. E1SFF3-11-0016-3100392, issued September

29, 1993).

                                 37


                              Report No.  E1SKG3-15-0098-4400038

-------
	Special Review of EPA's Information Systems  Program

  •  GIS showed the locations of certain facilities  in the
     middle of the Atlantic Ocean and at the North Pole.

  •  25 FINDS numbers were assigned to the same facility
     under the air program.

These officials attributed the data inaccuracies to  a  number
of causes.

  •  Data definitions cause inaccuracies because of  frequent
     changes and misinterpretation.

  •  The lack of consistent data dictionaries (e.g., multiple
     data dictionaries exist for common data elements  with
     different data definitions) contribute to inaccuracies.

  •  As discussed earlier, for some national systems little
     or no motivation exists to ensure data quality.   No
     clear ownership of data exists--data is input by  non-
     stakeholders (i.e., Regions and States).

  •  Data quality is less accurate for "environmental" data
     and is subject to many qualifiers (season,  age, sex,
     species, etc.)  which must be interpreted..  Environmental
     data is expensive to collect and limited in amount,
     geography, year, etc.

  •  Some focus group participants stated that~no "data audit
     process" exists in the EPA program offices to ensure
     data elements and definitions within individual systems
     are consistent and adhered to.

  •  The States sometimes input questionable data to the
     systems due to unclear input instructions and lack of
     resources.  This is complicated by EPA's inability to
     force the States to provide good quality data.  One
     example of the last point concerned the Aerometric
     Information Retrieval System (AIRS), in which we  were
     told that one major State refused to input zip  codes
     into the system, and EPA was powerless to force them to
     do so.

Agency officials involved in data integration projects
believe that the need for data integration will become more
apparent and the data quality of current and future  infor-
mation systems will improve dramatically as EPA increases
public access to more of its information resources.  Some


                              38

                           Report No.  E1SKG3-15-0098-4400038

-------
	Special Review of  EPA's  Information Systems Program

interview participants  believed EPA should facilitate this
access by embracing  the "information highway"  concept and
making available to  the public all information that  is of
interest to the public  and is  not  officially declared to be
sensitive.  These officials  believed the  Agency will eventu-
ally be forced in this  direction.   For example, the Vice-
President's National Performance Review strongly encourages
agencies to provide  information to the customer (i.e.,
general public) on what they have  paid for.   Additionally,
the proposed  House bill on EPA cabinet-level elevation  (H.R.
3425 sponsored by Congressman  Conyers)  specifically  calls for
enhancing and promoting public access to  EPA information.
                               39

                            Report No.  E1SKG3-15-0098-4400038

-------
	Special Review of EPA's Information Systems Program

                          CHAPTER 3

                       RECOMMENDATIONS

Our review identified problems in many areas of EPA's IRM
program.  These problems exist for many reasons, and involve
all facets of the Agency's operation.  Consequently,
recommendations have been developed which will affect all
aspects of EPA's IRM program, and which take the Agency's
streamlining initiatives into consideration.

Many of these recommendations address root causes which cut
across the problem categories in Chapter 2.  So by
implementing 'certain recommendations, the Agency can impact
more than one category.  However, the recommendations are
presented consistent with the categories' order of
presentation in Chapter 2.   The recommendations are not in
any priority order of importance, and many should be
implemented concurrently.

The first group of recommendations addresses weaknesses in
IRM management and necessary changes to the Agency's
organizational structure.  Recommended improvements focus on
establishing a separate CIO function, strengthening the SIRMO
position and IRM organizations in the program and Regional
offices, and strengthening the Executive Steering Committee
for IRM.

The second group of recommendations addresses weaknesses in
the way in which the Agency plans for and measures the uses
of information resources.  Recommendations address
development of an IRM Strategic Plan, linking the IRM
Strategic Plan with the Agency Business Plan, integration of
the planning and budget processes,  and the establishment of
performance measurement criteria for the IRM program.

The third group of recommendations addresses specific
concerns related to the development of information systems.
Key areas of recommendations include development of policies
and procedures to strengthen EPA's management framework for
developing information systems and for increasing user
involvement during development activities, and establishment
of training.mechanisms to.close.the gap in.expertise between
those who are knowledgeable about EPA programs and those who
are knowledgeable about IRM:
                              40

                           Report No.  E1SKG3-15-0098-4400038

-------
	Special  Review of EPA's Information Systems Program

The final group  of recommendations addresses the management
of information/data.   Recommended improvements focus on
developing an Agencywide information/data architecture,
establishing a centralized data administration function,  and
increasing public  access to EPA's non-sensitive
information/data.

The DIG recommends that the Administrator take the following
actions to correct the most significant problems in EPA's IRM
program, and to  facilitate the program's long-term success.

IRM Management and Organizational Structure

1.   Establish a separate senior level position,  the CIO,  at
     the Assistant Administrator level.   Consolidate all  IRM
     . functions under this position and do not encumber  it
     with any other responsibilities outside of IRM.  The
     CIO's responsibilities should include,  but not be
     limited to,  the following:  (a)  establishing what IRM
     capabilities  are necessary to support programs for
     accomplishing the mission of the Agency,  including IRM
     policies, architectures,  and standards;  (b)  coordinating
     with EPA's  CFO to help ensure that financial and other
     information systems used in preparing EPA's  financial
     statements  are effectively designed,  developed, and
     implemented;  (c)  developing policies,  standards, and
     mechanisms  to communicate information management goals,
     priorities,  and practices effectively throughout the
     Agency;  (d)  working to enhance top management's
     knowledge and understanding of IRM,  particularly that it
     is a core Agency function and that  data is a strategic
     EPA resource; and (e)  designing and implementing a
     process by  which the Agency initiates,  approves, and
     evaluates major automated information systems  at key
     milestones.

2.   Establish clear lines of authority under the CIO,
     especially  between OIRM and NDPD-

3    Approve and support the proposed new charter for the
     Executive Steering Committee for IRM and require the
     active participation of the permanent members  of the
     Committee.

4    Direct the  CIO to conduct a study of the SIRMO function
  '   to evaluate the success of its current operation in
     terms of the  requirements and responsibilities of  the


                              41

                            Report No.  E1SKG3-15-0098-4400038

-------
	Special Review of EPA's Information Systems Program

     function, and the differences in responsibilities and
     organizational placement between the various
     Headquarters program offices and Regional offices.  The
     study should articulate what is expected of and needed
     from the function by establishing a baseline of
     requirements, responsibilities, qualifications, and
     organizational placement of the function.  To make the
     SIRMO function truly effective, the study should then
     assess the time required to implement the functional
     responsibilities and recommend any additional staff
     required to support the function.

5.   Direct the CIO to develop,  in coordination with the
     Office of Human Resources Management and the Agency
     streamlining efforts,  a generic position description
     outlining the duties and responsibilities of program and
     Regional office SIRMOs based upon the study.  The
     position description should establish minimum
     qualifications for staffing this position.

6.   Direct the CIO to develop,  in coordination with the
     Office of Human Resources Management,  generic
     performance standards for program and Regional SIRMOs
     based upon the study.   These standards should reflect
     the responsibilities outlined in the position
     description.

7.   Direct the Assistant Administrators, Regional
     Administrators,  and equivalent office heads to ensure
     that the SIRMOs in their organizations meet and maintain
     the qualifications of the position as defined in the
     position description and are evaluated using the
     standards developed by the CIO.

8.   Direct the Assistant Administrators, Regional
     Administrators,  and equivalent office heads to establish
     an IRM advisory committee in each Headquarters program
     office and Region to:  (a)  provide a forum for sharing
     information about the following types of IRM issues--
     program,  State,  Regional,  and national initiatives;
     budget issues; information technology issues; training;
     planning; policy formulation,- etc.; and (b) support
     senior, management .participation. on..the Executive
     Steering Committee for IRM.  The CIO should develop a
     generic charter for these advisory committees,
     identifying purpose,  mission, membership (should include
     representation from each major division/office and


                             42

                           Report No.   E1SKG3-15-0098-4400038

-------
         	   ^	— J — —' i  '•  • <•• >»3J<» VMM*

     States), and  a  meeting schedule.   The SIRMOs would serve
     as the Chair  of these advisory committees.

9.   Direct the CIO  to:  (a)  provide and/or identify  training
     vehicles for  enhancing the IRM knowledge of  Agency
     employees, such as  the ADP exchange program  offered  by
     OARM-RTP;  (b) develop a comprehensive IRM component
     within the Agency's core training curriculum for  both
     managers and  staff;  (c)  encourage OARM's IRM personnel
     to take training in Agency programmatic  activities/  and
     (d) provide basic training in proper procedures for
     developing information systems.

10.  Direct the Assistant Administrators,  Regional
     Administrators,  and equivalent office heads  to  ensure
     that their employees are appropriately trained  in
     information management and information technology
     disciplines.

Resource Planning  and Performance  Measmr^ingnt,

11.  Ensure that general  IRM goals and objectives  in relation
     to the Agency mission are addressed in the Business
     Plan.

12.  Give the CIO  overall responsibility for  developing,
     maintaining,  and overseeing the implementation  of EPA's
     IRM Strategic Plan,  which should:  (a) be "based  on the
     Agency Business  Plan;  (b)  include implementation plans
     for all major IRM initiatives (e.g.,  major system
     development and  modernization activities and  significant
     ADP acquisitions and contracts) with  detailed
     descriptions, significant milestones  and target dates,
     life cycle costs, and explanations  of how  the
     initiatives support  the  Agency mission;  and  (c)  incor-
     porate component plans  from the program  and Regional
     offices, the  responsibility for which resides with the
     Assistant Administrators and  Regional Administrators.

13   Direct the CIO  to ensure that the IRM Strategic Plan:
     (a) continues .to support efforts  to integrate the
     budgetary and resource  allocation process; (b) has a
     planning horizon of  a minimum of  five years;  (c) is
     updated at least annually with associated  changes based
     on accomplishments,  budget authorizations  and
     allocations,  revisions  of life cycle  cost  estimates,  and
     actual expenditures;  (d)  is flexible  enough to deal with


                              43

                           Report  No.  E1SKG3-15-0098-4400038

-------
	Special Review of EPA's Information Systems Program

     changing program and technology needs;  (e) anticipates
     and projects future potential IRM needs  (data,
     information, and technology) and the future direction of
     technology, distinguishes these from known needs, and
     establishes an appropriate priority level;
     (f) accommodates OMB, GAO, Congressional, and internal
     budgetary reporting requirements; and  (g) takes into
     account legislative, regulatory, and guidance
     development and changes which affect IRM.

14.  Direct the CIO to include the "stakeholders"  (program,
     Regional, and State officials associated with both IRM
     and programmatic activities) in the information
     resources planning and implementation processes.

15.  Require that the Agency's IRM Strategic Plan be reviewed
     and approved by the Executive Steering Committee for IRM
     and the CIO.

16.  Direct the Assistant Administrators, Regional
     Administrators,  and equivalent office heads to reduce,
     where possible,  reliance on contractors in order to
     reduce EPA's vulnerability to loss of IRM expertise.

17.  Direct the CIO,  in coordination with the CFO,  to examine
     whether the Agency can fund IRM on a multi-year basis.
     The examination should evaluate funding of: (a)  IRM
     initiatives; (b)  implementation of Agency policies and
     standards; (c)  ADP technology; and (d)  systems
     development, operations and maintenance.  The Agency
     survey of resources expended for IRM activities should
     be modified as necessary to facilitate this examination.

18.  Direct the CFO to establish direct line items in the
     Agency budget process for IRM activities as a necessary
     component of meeting the mission of the Agency.   The
     line budget items should reflect the IRM budgets for
     both Headquarters and Regional components,  and should be
     tied to implementation plans for all significant IRM
     initiatives.

19.  Direct all Agency components to ensure that their:
     (a) major systems'  development,  operations, and
     maintenance costs are addressed explicitly in their
     individual organizations' IRM plans and; (b)  systems are
     built,  enhanced,  and operated only in accordance with
     those plans,-  and 
-------
	Special  Review of EPA's Information Systems Program

     in response to the Agency's survey of resources expended
     for IRM  (i.e.,  "Exhibit 43" data call)  is certified for
     accuracy.

20.  Establish  an  ongoing system to measure and report on
     specific IRM  program performance from policy,
     operational,  and economic standpoints.   This system
     should include performance measurement criteria for
     information systems and other IRM activities,  which
     will:  (a)  measure operational results against mission-
     related  goals for which the information systems were
     developed;  (b)  evaluate and compare actual life cycle
     costs to budgeted life cycle costs; and (c)  compare
     planned  benefits to actual benefits.   The initial
     measurement criteria and goals should be based on
     established plans and budgets.  Subsequent changes  in
     performance criteria based on implementation should be
     reflected  in  plan revisions.

Information Svpfc**"1" Development

21.  Direct the CIO to establish controls  to ensure that  EPA
     employees  and contractors follow established Federal
     SDLC standards throughout the life cycle of  information
     systems.

22.  Direct the CIO to continue efforts to develop  and revise
     clear IRM  policies,  procedures,  and standards.   In  those
     instances  in  which Agencywide policies  and standards
     cannot be  resolved within the context of the green
     border review process,  expeditiously  bring those
     policies and  standards before the Executive  Steering
     Committee  for IRM for resolution.

23.  Direct the CIO to ensure that the new IRM policies
     require  that  on all major system development/enhancement
     projects:  (a)  primary and secondary users be identified
     and involved  throughout the system life cycle  process,-
     (b) detailed  needs analyses be conducted and updated
     periodically,  and include the needs of  primary and
     secondary  users;  (c)  cost-benefit analyses be  conducted
     and updated periodically;  and (d)  system decision papers
     be prepared at significant decision points  in  the life
     cycle of the  system for approval by the Executive
     Steering Committee for IRM.
                              45

                            Report No.  E1SKG3-15-0098-4400038

-------
_ Special Review of EPA's Information Systems  Program

24.  Direct the Assistant Administrators, Regional
     Administrators, and equivalent office heads to  establish
     a system life for each major system.  These systems
     should be reviewed periodically prior to the end  of
     their identified life to:  (a) assess relationships among
     the systems;  (b) identify areas of excessive overlap;
     and (c) determine which systems need to be retired or
     enhanced .

D     anagoinort t
25.  Establish a formal Information/Data Administration
     Program under the CIO.  This program should be
     responsible for:  (a) developing, overseeing, and
     enforcing a centralized repository of mission-related
     standardized data elements for every EPA program in
     Headquarters and the Regions; (b) developing and
     implementing a standard, defined information/data
     architecture; and (c) working with other Agency
     components to implement data integration throughout the
     Agency to reduce unnecessary duplicate data in
     information systems developed for individual programs.

26.  Appoint a Data Administrator with the knowledge,
     experience, and authority to effectively manage the
     Information/Data Administration Program.  The Data
     Administrator should report directly to the CIO.

27.  Direct the Data Administrator to: (a)  develop and
     promulgate Agencywide standards and management controls
     for data resources,-   (b) develop, oversee, and enforce
     centralized coordination of mission-related data
     standardization efforts for every EPA program,-
     (c) develop and manage an Agencywide information
     architecture; and (d) coordinate and control
     information/data integration efforts throughout the
     Agency .

28.  Direct the Data Administrator to work with the National
     Institute of Standards and Technology,  other Federal
     agencies,  and non- Federal organizations, as appropriate,
     in the development of data standards.

29.  Direct the Data Administrator to create a central
     repository to manage and control logical data models,
     architectures,  and standard data element documentation,
                             46

                           Report No.  E1SKG3-15-0098-4400038

-------
	Special  Review of EPA's Information Systems Program

     and to make this information easily accessible within
     and outside the Agency.

30.  Ensure that the Data Administrator participates in the
     IRM strategic planning and budgeting process,  develops
     an Agencywide strategic information/data plan,  works  to
     see that  sufficient funds are allocated for
     information/data management activities through the
     budget process,  and participates in the information
     collection process.

31.  Require that  every proposed new information/data
     collection and modification activity be submitted by  the
     SIRMOs to the Data Administrator for comparison to
     existing  data models,  consistency with data standards,
     .assurance of  data quality plans,  and reasonableness of
     cost versus benefits and impacts analysis.

32.  Require that  all significant information system
     development activity be accompanied by the  development
     of data collection plans,  data quality assurance plans,
     data cost and impact analyses,  and survey designs.
     Ensure that these plans,  analyses,  and designs  are
     reviewed  and  approved by the Data Administrator prior to
     the initiation of development activity.

33.  Direct the CIO to work with Agency components to:
      (a) identify,  to the maximum extent practicable,  EPA
     information/data that is of interest to the public and
     is not officially declared to be sensitive;  (b)  provide
     improved  access for the public to view (but not add to,
     delete, or otherwise modify)  this information/data,- and
      (c) ensure that,  to the maximum extent possible,  Agency-
     developed systems provide easy access,  help screens,  on-
     line definitions,  and on-line documentation.

34.  Direct the CIO to assist in the establishment of
     electronic data interchange capability within EPA for
     collection, transmission,  and dissemination of  data.

35   Direct senior program managers to provide sufficient
     resources on  a continuing basis to:  (a)  successfully
     perform any proposed new data collection endeavors; and
      (b) accommodate adherence to EPA information/data
     management policies.
                              47

                            Report No.  E1SKG3-15-0098-4400038

-------
Special Review of EPA's Information Systems Program
   (This page intentionally left blank)
                   48



                Report No.  E1SKG3-15-0098-4400038

-------