This is the accessible text file for GAO report number GAO-07-479 
entitled 'Clean water: Further Implementation and Better Cost Data 
Needed to Determine Impact of EPA's Storm Water Program on Communities' 
which was released on May 31, 2007. 

This text file was formatted by the U.S. Government Accountability 
Office (GAO) to be accessible to users with visual impairments, as part 
of a longer term project to improve GAO products' accessibility. Every 
attempt has been made to maintain the structural and data integrity of 
the original printed product. Accessibility features, such as text 
descriptions of tables, consecutively numbered footnotes placed at the 
end of the file, and the text of agency comment letters, are provided 
but may not exactly duplicate the presentation or format of the printed 
version. The portable document format (PDF) file is an exact electronic 
replica of the printed version. We welcome your feedback. Please E-mail 
your comments regarding the contents or accessibility features of this 
document to Webmaster@gao.gov. 

This is a work of the U.S. government and is not subject to copyright 
protection in the United States. It may be reproduced and distributed 
in its entirety without further permission from GAO. Because this work 
may contain copyrighted images or other material, permission from the 
copyright holder may be necessary if you wish to reproduce this 
material separately. 

Report to Congressional Requesters: 

United States Government Accountability Office: 

GAO: 

May 2007: 

Clean Water: 

Further Implementation and Better Cost Data Needed to Determine Impact 
of EPA's Storm Water Program on Communities: 

GAO-07-479: 

GAO Highlights: 

Highlights of GAO-07-479, a report to congressional requesters 

Why GAO Did This Study: 

Urban storm water runoff is a major contributor to the nation’s 
degraded waters. Under the Clean Water Act, the Environmental 
Protection Agency (EPA) established a program requiring communities to 
obtain permits and implement activities to control storm water 
pollution. EPA’s Phase I regulations (1990) applied to communities with 
populations of 100,000 or more, and its Phase II regulations (1999) 
covered smaller urban communities. Communities must report progress in 
meeting permit requirements. Some have raised concerns that storm water 
requirements impose an undue burden. To evaluate storm water program 
costs, EPA developed estimates for both phases. 

GAO was asked to (1) determine the progress in implementing the storm 
water program, (2) evaluate the extent to which the program burdens 
communities, (3) examine the accuracy of EPA’s cost estimates, and (4) 
examine the data available for assessing program burden. GAO collected 
data for all states and a sample of 130 communities, among other steps. 

What GAO Found: 

Storm water program implementation has been slow for both Phase I and 
II communities. The federal deadlines for permit applications were 
years ago—14 years for Phase I and 4 years for Phase II—but almost 11 
percent of all communities were not yet permitted as of fall 2006. In 
addition, litigation, among other reasons, delayed the issuance of some 
permits for years after the application deadlines. As a result, almost 
all Phase II and some Phase I communities are still in the early stages 
of program implementation. 

It is too early to determine the storm water program’s overall burden, 
but several factors influence the extent to which the program burdens a 
community. In particular, burden varies depending on whether 
communities (1) can use the flexibility built into EPA’s regulations to 
implement less expensive measures, or (2) are able to benefit from 
prior storm water management experience. Some communities may face a 
greater burden because of more stringent requirements set by EPA or the 
states, additional efforts required to address litigation over water 
quality, or because of barriers to obtaining funding for storm water 
activities. Storm water program burdens could increase in the future 
because, among other reasons, EPA or the states may reissue permits 
with more stringent requirements. 

Without an estimate of actual storm water program costs—or burden—GAO 
could not determine the accuracy of EPA’s cost estimates. However, GAO 
did identify methodological concerns that raise questions about the 
usefulness of these estimates for measuring the burden communities 
face. That is, the Phase I analysis was not designed to estimate 
national program costs, the Phase II analysis was based on survey data 
of questionable validity and reliability, and neither analysis excluded 
costs for activities that communities were implementing before the 
program. 

Any assessment of program burden will be hampered because EPA is not 
collecting complete and consistent data on communities’ activities and 
their costs. For example, only Phase I communities are required to 
include data on program costs and these data are often limited. Also, 
communities’ inconsistent reporting of activities makes it difficult to 
evaluate program implementation nationwide. Consequently, EPA will find 
it challenging to meet its goal to examine Phase II implementation 
starting in 2012. 

Figure: Sidewalk Plaque Discouraging Storm Drain Pollution: 

[See PDF for Image] 

Source: 

[End of figure] 

What GAO Recommends: 

GAO recommends that EPA issue guidance and consider regulatory changes 
so that communities report consistently on their efforts. EPA stated it 
has already taken some action, but agreed to take additional steps to 
collect better cost data. 

[Hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-07-479]. 

To view the full product, including the scope and methodology, click on 
the link above. For more information, contact John B. Stephenson at 
(202) 512-3841 or stephensonj@gao.gov. 

[End of section] 

Contents: 

Letter: 

Results in Brief: 

Background: 

Storm Water Permit Issuance Has Been Slow and Many Communities Have Not 
Fully Implemented Activities: 

Several Factors Influence the Extent to Which Storm Water Program 
Implementation Is or Could Be a Burden: 

Methodological Concerns Raise Questions about the Usefulness of EPA's 
Cost Estimates: 

Assessment of Program Burden Is Hampered by Limited and Inconsistent 
Data: 

Conclusions: 

Recommendation for Executive Action: 

Agency Comments and Our Evaluation: 

Appendix I: Briefing Provided to Congressional Requesters: 

Appendix II: Objectives, Scope, and Methodology: 

Appendix III: Detailed Analysis of Storm Water Program Implementation 
Data: 

Appendix IV: Comments from EPA: 

Appendix V: Key Contributors to This Report: 

Related GAO Products: 

Tables: 

Table 1: Examples of Storm Water BMPs: 

Table 2: EPA's Estimated Ranges of Phase I Per Capita Costs: 

Table 3: Phase I Storm Water Program Implementation Data, by State as 
of Fall 2006: 

Table 4: Phase II Storm Water Program Implementation Data, by State as 
of Fall 2006: 

Figures: 

Figure 1: The Storm Water Permitting Process: 

Figure 2: Issuance of Larger MS4s' Permits Has Taken up to 14 Years, 
and Some Are Still Not Issued: 

Figure 3: Many Phase II Permits Were Delayed or Are Not Yet Issued: 

Abbreviations: 

BMP: best management practice: 
CWSRF: Clean Water State Revolving Fund: 
EPA: Environmental Protection Agency: 
MS4: municipal separate storm sewer system: 
NAFSMA: National Association of Flood and Stormwater Management 
Agencies: 
NPDES: National Pollutant Discharge Elimination System: 
TMDL: Total Maximum Daily Load: 

United States Government Accountability Office: 
Washington, DC 20548: 

May 31, 2007: 

The Honorable James M. Inhofe: 
Ranking Member: 
Committee on Environment and Public Works: 
United States Senate: 

The Honorable Mike Crapo: 
United States Senate: 

Storm water from urban areas is a major contributor to the degradation 
of the nation's lakes, rivers, and streams. It runs off paved or other 
impervious areas into surface waters and may contain dangerous 
chemicals, harmful bacteria, debris, and other pollutants that can pose 
serious public health risks. As urban and suburban areas have expanded, 
so too has the amount of impervious surface. As a result, storm water 
runoff has increased in volume and velocity, which can alter the 
natural flow of water, harm aquatic ecosystems, and cause considerable 
property damage. 

In 1987, Congress amended the Clean Water Act, directing that the 
Administrator of the Environmental Protection Agency (EPA) establish a 
program to regulate storm water pollution. EPA developed the Storm 
Water Program as part of the National Pollutant Discharge Elimination 
System (NPDES) Program by issuing regulations in two phases: Phase I in 
1990 and Phase II in 1999. These regulations required communities with 
separate sewer systems for storm water to obtain permits and implement 
activities to control storm water runoff.[Footnote 1] The Phase I 
regulations generally applied to communities with populations of 
100,000 or more and identified 220 communities potentially subject to 
the program. The Phase II regulations generally applied to smaller 
communities in urban areas and identified 5,036 communities that could 
be required to obtain permit coverage.[Footnote 2] Communities had to 
apply for permits by specified dates early in the 1990s for Phase I and 
in 2003 for Phase II. In most cases, the states, authorized by EPA, 
issue 5-year permits that must then be renewed. The permits require 
communities to implement activities to control storm water runoff. 
These activities may include (1) educating the public about the impacts 
of runoff in order to discourage such practices as dumping used 
automobile oil into a storm drain; (2) street-sweeping; (3) structural 
practices, such as building retention ponds to slow or prevent the 
release of polluted storm water; or (4) disconnecting illicit plumbing 
connections to the storm sewer system. Communities are required to 
report periodically on their storm water activities. In some cases, 
communities may be able to obtain federal or state funds, such as Clean 
Water State Revolving Fund loans--low-interest loans that can be used 
for a variety of water quality projects--to help implement their storm 
water management programs. 

In 1990 and 1999, respectively, EPA analyzed the costs of implementing 
Phases I and II of the storm water program. EPA's Phase I analysis 
modeled program costs for a limited number of large and medium cities. 
It did not provide a national estimate of program costs. EPA's Phase II 
analysis estimated national program implementation costs based on 
survey data received from 56 communities. In its Phase II regulations, 
EPA set a goal of beginning to evaluate implementation of Phase II of 
the program in 2012. 

The Unfunded Mandates Reform Act of 1995 requires agencies to determine 
the cost of a significant regulatory action on state and local 
governments. However, the act does not require agencies to look at the 
cumulative costs of multiple regulations. Many communities responsible 
for implementing storm water program requirements are also responsible 
for implementing other regulations relating to water quality, such as 
drinking water regulations or specific pollution treatment programs. 
Consequently, states, local governments, water associations, and 
industry and academic experts have raised concerns that implementing 
EPA's storm water program imposes a burden on communities. For purposes 
of this report, we use burden to mean additional costs for implementing 
storm water control measures, increased administrative activities, 
reduced budget flexibility because of the need to divert resources from 
other governmental activities, actions related to litigation, and the 
influences of other regulatory programs on how storm water runoff is 
managed. 

In this context, you asked us to (1) identify the progress made in 
implementing the storm water program, (2) determine the extent to which 
the storm water program burdens communities, (3) evaluate the accuracy 
of EPA's cost estimates, and (4) examine the data available for future 
assessment of program burdens. On December 18, 2006, we briefed your 
staff on the results of our review; this report provides more 
information on highlights presented at this briefing. A copy of our 
December briefing is included as appendix I. 

To identify progress made in program implementation, as well as to 
examine the extent of program burden, we interviewed officials 
responsible for implementing the storm water program, including EPA, 
state, and local officials. For all 50 states, we obtained and analyzed 
storm water program data, such as the number of communities required to 
obtain permits, and the number that had obtained permits as of fall 
2006.[Footnote 3] For each state's data, we assessed the reliability of 
the data by (1) examining how the data were collected, processed, and 
maintained; (2) reviewing the data we received for discrepancies, such 
as duplicate entries; and (3) conducting detailed follow-up in cases 
where we found inconsistencies. Through these efforts, we determined 
that the data we collected were sufficiently reliable for the purposes 
of our review. In addition, we reviewed applicable statutes, 
regulations, guidance, and studies, and obtained the views of state and 
local government associations and academic and industry experts on 
program implementation. Furthermore, we analyzed EPA's cost estimates 
of the Phase I and II regulations and adjusted all dollars to 2006 
dollars. Finally, to obtain information on future data availability 
issues, we reviewed storm water program documents from a sample of 130 
communities to obtain information on their activities. Most of these 
communities (about 95 percent) were selected randomly, with the 
remainder chosen through a judgmental selection of communities that we 
believed had well-established storm water programs and good cost 
information based on interviews we conducted and studies we reviewed. 
Appendix II provides a more detailed description of our scope and 
methodology. We conducted our work between January 2006 and April 2007 
in accordance with generally accepted government auditing standards. 

Results in Brief: 

Implementation of the storm water program has been slow for both Phase 
I and II communities. While the Phase I and II federal application 
deadlines were years ago--14 years ago for Phase I and 4 years ago for 
Phase II--11 Phase I and 809 Phase II communities were not yet 
permitted as of fall 2006 (almost 11 percent of all communities). In 
addition, many permits were issued years after the application 
deadlines, in part because of litigation challenging the process for 
issuing these permits or the conditions that communities were to meet. 
As a result, almost all permitted Phase II and some permitted Phase I 
communities are still in the early stages of program implementation. 
That is, they are gathering information on the types of storm water 
pollution they face and determining which activities they plan to 
implement to address this pollution. Furthermore, some permitted 
communities may not be complying with their permit requirements. For 
example, recent audits sponsored by EPA Region 9 and California 
revealed that some communities were not adequately controlling storm 
water runoff at municipally owned and operated facilities. In one 
instance, a community's state-issued permit required it to reduce the 
discharge of pollutants into the storm sewer system to the maximum 
extent practicable. However, auditors found problems with compliance, 
such as street-sweeping debris that was dumped about 10 feet from a 
storm drain inlet at the community's municipal maintenance facility. 

Because many communities are still in the early stages of implementing 
the storm water program, it is too early to determine the overall 
program burden. However, several factors influence the extent to which 
storm water program implementation is currently a burden for 
communities or could become a burden in the future. For example: 

* Considerable flexibility is built into EPA's storm water regulations, 
which allow communities to choose the activities and levels of effort 
most appropriate to manage their storm water runoff. Therefore, one 
factor that may reduce the burden communities currently face is the 
extent to which they take advantage of this flexibility and choose less 
expensive measures for implementing their permits. For example, 
communities may choose to meet the program requirement to educate the 
public about the impacts of storm water runoff by having staff put 
together a brochure rather than by hiring a public relations firm to 
develop an elaborate media campaign. Other factors that may reduce 
program burdens for some communities include whether they already had 
storm water management activities in place, such as street-sweeping, or 
whether they are able to obtain federal or state funds to help 
implement permit requirements. 

* In contrast, other factors, such as whether permitting authorities 
include more stringent or specific conditions in storm water permits, 
may increase the burden communities currently face. For instance, one 
community reported that meeting a state requirement to increase the 
frequency with which it cleaned elements of its storm sewer system 
would increase West Nile Virus treatment costs, and draw staff away 
from other priorities, such as maintenance of its sanitary sewer 
system. Also, some communities face barriers to funding their 
activities. For example, communities in one state with hundreds of 
permittees have not been able to obtain Clean Water State Revolving 
Fund loans to help implement storm water activities because of a 
misperception regarding the eligibility of these activities for 
funding. 

* Furthermore, several factors may cause program burdens to increase in 
the future. For example, although EPA's program regulations are 
flexible, the agency instructs communities to expand or alter their 
storm water management activities, as needed, over successive permit 
terms to improve water quality. Some Phase I communities that have been 
implementing the storm water program for a longer period of time have 
already been reissued permits with more stringent or specific 
conditions. As EPA and state permitting authorities reissue permits for 
other communities, they may include additional requirements to help 
meet water quality standards--which could increase program burdens. 
Finally, although some communities have obtained federal funds to help 
manage storm water, continued reductions in the amount of federal funds 
potentially available for storm water projects--such as the nearly 20 
percent reduction in federal loan funds between 2004 and 2005--could 
cause communities to carry a greater share of program costs. 

Because we could not independently develop an estimate of actual storm 
water program costs, we could not conclusively determine whether EPA's 
1990 Phase I and 1999 Phase II analyses over-or underestimated these 
costs. However, we identified a number of methodological concerns that 
raise questions about the usefulness of EPA's estimates as measures of 
the burden communities face from implementing the program. For example, 
EPA's Phase I analysis, which estimated program costs for a small set 
of hypothetical cities under various scenarios, was not designed to 
estimate actual program costs for any specific community or provide a 
national estimate of program costs. Additionally, the Phase II cost 
analysis was largely based on data from only about 3.5 percent of the 
1,600 communities surveyed to identify the types of storm water 
activities they were conducting and the costs of these activities. In 
fact, only 56 officials returned information on activity costs. 
Moreover, many of the key survey questions were extremely complicated 
and subject to multiple interpretations, making it unlikely that 
communities could have responded with accurate information. Because of 
the small sample of data and concerns over their reliability, we do not 
believe that the Phase II survey data provide a valid and reliable 
estimate of program costs nationwide. Furthermore, EPA's Phase I and II 
analyses did not exclude costs of storm water activities that 
communities may have been conducting before the program, which could 
cause its analyses to overestimate incremental program costs. 

Any assessment of program burden will be hampered by limited and 
inconsistent data. EPA is not collecting complete and consistent data 
on communities' activities and their costs. While both Phase I and 
Phase II communities must submit reports on their storm water 
activities, only Phase I communities are required to include any 
information on storm water activity costs. Furthermore, the data in 
communities' reports are often limited. For example, one Phase II 
community's annual report was a one-page letter which stated that the 
community's estimated implementation status was "20-percent," but 
provided few additional details. Finally, because a number of factors 
influence the costs of implementing each of the many activities that 
may be part of a community's storm water management program, 
inconsistencies in reporting among different communities hamper a 
national evaluation of these costs. Consequently, EPA will find it 
difficult to assess implementation of either phase of the program, 
particularly to meet its goal to examine Phase II implementation 
starting in 2012. 

So that EPA can evaluate the implementation of the storm water program 
nationwide, we are recommending that the Administrator, EPA, issue 
program guidance and consider regulatory changes to ensure that 
communities provide consistent data on the scope, costs, and results of 
their efforts. 

In responding to a draft copy of this report, EPA recognized the 
importance of being able to assess the performance of the storm water 
program. On the basis of our recommendation, EPA said that it would 
investigate ways to gather better cost information through communities' 
annual reports. See appendix IV for EPA's written comments. EPA also 
offered technical comments, which we have incorporated as appropriate. 

Background: 

Pollutants and sediment carried by storm water, as well as the volume 
and temperature of runoff, can alter aquatic habitats and make it hard 
for fish and other organisms to survive. Some pollutants can also make 
fish and shellfish unsafe to eat. Reducing storm water pollution may 
help increase the number, size, and quality of fish and other 
organisms; which could provide benefits to those who value these 
resources for consumption, as well as for commercial, recreational, 
subsistence, and aesthetic purposes. Moreover, polluted storm water 
runoff can negatively impact those who use fresh and salt water areas 
for swimming and boating. Swimmers in water with high levels of 
bacteria have a greater risk of contracting gastrointestinal or 
respiratory illnesses. Reducing storm water pollution could also lead 
to fewer beach closings, and enhanced enjoyment of fresh and salt water 
areas. Finally, reducing the quantity and improving the quality of 
storm water runoff could help avoid costs to: 

* treat illnesses caused by contact with polluted runoff, 

* stabilize stream banks to limit erosion and prevent property damage, 

* repair water and sediment damage caused by flood events, 

* dredge waterways to maintain navigation channels, and: 

* treat or obtain alternate sources of drinking water. 

EPA's Office of Wastewater Management, within the Office of Water, 
leads and manages water quality improvement efforts under the NPDES 
program in partnership with EPA regional offices, states, and tribes. 
The NPDES program, created in 1972 under the Clean Water Act, 
authorized the Administrator to issue permits, according to conditions 
prescribed in regulation, for the discharge of pollutants from point 
sources.[Footnote 4] These point sources included factories or 
wastewater treatment plants that contributed pollutants directly into a 
body of water from a pipe or other conveyance. Neither the NPDES 
legislation nor the rules promulgated in 1973 specifically addressed 
storm water discharges. Under the 1987 amendments to the Clean Water 
Act,[Footnote 5] Congress required EPA to regulate storm water runoff 
that reaches municipal separate storm sewer systems (MS4s) as a point 
source of pollution.[Footnote 6] 

EPA met this requirement by establishing the NPDES Storm Water Program, 
which regulated storm water discharges from MS4s in two 
phases.[Footnote 7] In November 1990, EPA issued regulations for Phase 
I of the program, requiring that large and medium MS4s obtain permits 
for their storm water discharges. Large MS4s were those serving 
populations of 250,000 or more, while medium MS4s were those serving 
100,000 or more but less than 250,000. In a February 2000 report to 
Congress on the implementation of Phase I, EPA reported that storm 
water pollution from these systems was considered to be the greatest 
threat to water quality.[Footnote 8] EPA's Phase I regulations required 
that applications for storm water permits be submitted in two parts. 
Among other things, the first part was to describe the characteristics 
of local storm water pollution and identify existing controls. The 
second part was to include a proposed storm water management program, 
and was due in November 1992 and May 1993 for large and medium MS4s, 
respectively. In December 1999, EPA issued regulations for Phase II of 
the program, which required other systems--generally smaller MS4s in 
urban areas--to obtain permits for their storm water discharges. 
Applications for these smaller systems were more streamlined and were 
generally due in March 2003. EPA has authorized most states to issue 
and enforce storm water permits.[Footnote 9] EPA and state permitting 
authorities are authorized to include additional MS4s in the storm 
water program, beyond those designated automatically by EPA 
regulations.[Footnote 10] Most Phase I MS4s have been issued individual 
permits, while most Phase II MS4s have received coverage under general 
permits that EPA and state permit authorities issued for entire groups, 
such as all Phase II MS4s in a state.[Footnote 11] Figure 1 shows the 
processes for issuing individual and general storm water 
permits.[Footnote 12] 

Figure 1: The Storm Water Permitting Process: 

[See PDF for image] 

Source: GAO analysis of EPA guidance. 

[End of figure] 

Under the Clean Water Act, permits issued to MS4s must require controls 
that reduce storm water pollution to the maximum extent practicable. 
EPA intentionally did not include a precise definition of what it means 
to reduce storm water pollution to the maximum extent practicable in 
order to provide EPA and state permitting authorities with flexibility 
in developing MS4s' permits. As a result, EPA's storm water regulations 
largely do not identify specific controls that MS4s must implement. 
Instead, the regulations direct MS4s to meet their permit requirements 
by implementing storm water management programs that may include best 
management practices (BMPs) in the following categories: public 
education, public involvement, illicit discharge detection and 
elimination, construction site runoff, post-construction runoff, and 
pollution prevention from commercial, industrial, and residential 
areas, as well as from municipal operations. Storm water BMPs include, 
among other things, prohibiting certain practices or establishing 
maintenance or other management procedures to reduce or prevent storm 
water runoff and/or pollutants from reaching receiving waters. Table 1 
shows examples of BMPs that MS4s may implement in each of these 
categories.[Footnote 13] 

Table 1: Examples of Storm Water BMPs: 

BMP category: Public education--implement a program to educate the 
public about the impacts of storm water pollution and how to mitigate 
them; 
Examples of BMPs: 
* Distribute inserts in utility bills to help educate the public about 
the storm water impacts of certain activities, such as not cleaning up 
pet waste; 
* Stencil storm drains to increase public awareness of where storm 
water goes to help prevent dumping of waste into storm drains. 

BMP category: Public involvement--implement a program to include the 
public in developing, implementing, and reviewing an MS4's storm water 
management program; 
Examples of BMPs: 
* Enlist community groups in helping to clean up local streams; 
* Use volunteers to stencil storm drains. 

BMP category: Illicit discharge detection and elimination--develop and 
implement a program to prevent or eliminate discharges from entering an 
MS4 that are not composed entirely of storm water[A]; 
Examples of BMPs: 
* Map storm sewer systems to identify the locations of outfalls; 
* Use dye or other types of testing to trace the source of illicit 
connections to the storm sewer system. 

BMP category: Construction site runoff--develop and implement a program 
to reduce pollutants in storm water runoff from construction 
activities; 
Examples of BMPs: 
* Train and certify construction site contractors; 
* Conduct inspections of construction sites to ensure BMPs are properly 
installed and maintained. 

BMP category: Post-construction runoff--develop, implement, and enforce 
a program to prevent or minimize storm water runoff impacts from new 
development and redevelopment projects; 
Examples of BMPs: 
* Review the design plans of new development sites to minimize 
additional storm water runoff; 
* Ensure that structural controls to capture and help treat runoff at 
new or redeveloped sites (e.g., swales, ponds, or wetlands) are 
maintained. 

BMP category: Pollution prevention/municipal good housekeeping-- 
develop and implement programs to reduce or prevent pollution from 
commercial, industrial, and residential areas, and municipal 
operations; 
Examples of BMPs: 
* Periodically clean catch basins or sweep streets and parking lots to 
help eliminate trash and other pollutants from entering the storm sewer 
system; 
* Establish municipal vehicle maintenance procedures to prevent surface 
water pollution. 

Source: GAO analysis of EPA regulations and guidance. 

[A] Illicit discharges can enter an MS4 through (1) piping mistakenly 
or deliberately connected to storm drains, or (2) infiltration from 
cracked sanitary systems, spills, or dumping. 

[End of table] 

As part of their storm water management programs, some MS4s may also 
have to implement certain activities as a result of other environmental 
requirements, such as those stemming from the Total Maximum Daily Load 
(TMDL) Program, EPA's Combined Sewer Overflow Control Policy, and the 
Endangered Species Act. Specifically: 

* The TMDL Program is a Clean Water Act program requiring states to (1) 
set the maximum amount of a pollutant that a waterbody can receive 
while still meeting water quality standards, and (2) allocate specific 
amounts of the pollutant to individual entities that contribute it to 
the waterbody.[Footnote 14] As a result of a TMDL for a pollutant 
contained in storm water runoff, an MS4 may have to implement certain 
activities, such as treating storm water runoff to ensure that the 
amount of pollution it discharges does not exceed the TMDL. 

* EPA's 1994 Combined Sewer Overflow Control Policy provides guidance 
to permitting authorities and municipalities on how to meet the Clean 
Water Act's pollution control goals through NPDES permits issued to 
combined sewer systems. The policy requires communities with combined 
sewer systems to take action to address overflows caused when the 
amount of storm water entering pipes already carrying sewage exceeds 
the capacity of the system. Actions that communities take to adhere to 
this policy, such as separating their sanitary and storm sewer systems, 
can affect their storm water programs by increasing the size of the MS4 
and requiring controls for storm water runoff previously discharged 
through the combined sewer system. 

* The Endangered Species Act of 1973 prohibits, among other things, 
actions that harm endangered species within the United States.[Footnote 
15] Actions that harm endangered species might include discharging 
pollution that kills these species or destroys their habitats. As a 
result, an MS4 may have to take steps through its storm water program 
to protect endangered species, such as improving the quality of storm 
water discharges to certain streams to prevent damage to these species' 
breeding, feeding, or sheltering areas. 

MS4s outline their storm water management programs in plans that they 
submit to their permitting authority. According to EPA regulations, 
MS4s have up to 5 years to fully implement their programs. In addition, 
MS4s may work with other entities including the state, neighboring 
MS4s, or community groups to implement their storm water management 
programs. These relationships may be established by MS4s applying for 
permit coverage together as co-permittees, or through agreements 
allowing individual permittees to share responsibility for storm water 
activities. For those MS4s that establish co-permittee relationships, 
one MS4 may assume the position of a lead permittee and implement 
activities, such as monitoring, public education, and reporting on 
behalf of the co-permittees. 

MS4s must submit annual reports on their storm water management efforts 
that include information on (1) the status of compliance with permit 
conditions, (2) proposed changes or revisions to the storm water 
management program, and (3) results of information collected and 
analyzed, including monitoring data, if any, during the reporting 
period.[Footnote 16] In addition, Phase I MS4s must submit information 
on annual program expenditures and a budget for the following year. 
Phase II MS4s also must submit (1) an assessment of the appropriateness 
of their BMPs and progress towards achieving identified measurable 
goals in each of the six BMP categories, (2) a summary of the storm 
water activities planned for the next reporting cycle, and (3) notice 
if the MS4 is relying on another entity to satisfy any of the permit 
obligations. 

In some cases, MS4s may obtain federal funding to help implement storm 
water management efforts. The primary source of federal funds for storm 
water activities is the Clean Water State Revolving Fund (CWSRF), which 
was created by the 1987 amendments to the Clean Water Act as a source 
of independent, permanent, low-cost financing for a wide range of 
efforts to protect or improve water quality.[Footnote 17] However, from 
July 1987 to June 2006, only $346 million, or less than 1 percent of 
all CWSRF funding for wastewater projects, was loaned for storm water 
projects.[Footnote 18] Other types of federal assistance that may be 
available to MS4s include grant and technical assistance programs. 
Finally, assistance provided through other federal agencies' programs 
may have a storm water component. For example, federal community 
development block grants may be provided for activities that include 
some aspect of storm water runoff management. 

Storm Water Permit Issuance Has Been Slow and Many Communities Have Not 
Fully Implemented Activities: 

Storm water program implementation has been slow for both Phase I and 
II MS4s. Although the Phase I and II federal application deadlines for 
storm water permits passed years ago--around 14 years ago for Phase I 
and 4 years ago for Phase II--nearly 11 percent of MS4s have not yet 
received permits, and thus are not required to implement storm water 
management activities. Moreover, for many MS4s, litigation over the 
process of issuing storm water permits or their conditions, among other 
reasons, delayed program implementation for years after the initial 
permit application deadlines. As a result, almost all Phase II and some 
Phase I MS4s are still in the early stages of implementing their first 
5-year permits. Furthermore, some permitted MS4s may not be complying 
with their permit requirements. 

A Large Number of Communities Are Still Not Permitted and Many Permits 
Were Issued Years after Application Deadlines: 

As of fall 2006, a substantial number of MS4s--11 Phase I and an 
estimated 809 Phase II MS4s (nearly 11 percent)--had not yet been 
issued permits. Furthermore, this problem is widespread. More than half 
of all permitting authorities nationwide still had not issued all of 
their permits. Some permitting authorities, including Texas, 
Washington, and EPA Region 6, had not issued any Phase II permits 
because they had not yet developed final permits,[Footnote 19] or (in 
the case of Region 6) the permit had not yet gone into effect.[Footnote 
20] For instance, while one Texas official indicated that most issues 
with the state's Phase II permit had been resolved as of October 2006, 
another official anticipated that the permit would not be final until 
mid-2007. Appendix III provides more detailed data on the status of 
storm water program implementation, by state. 

While EPA and state permitting authorities issued storm water permits 
to most MS4s by the fall of 2006, in many cases these permits were 
issued years after the initial application deadlines. For example, as 
shown in figure 2, although some Phase I MS4s' initial permits were 
issued by 1993, many other Phase I MS4s did not receive their initial 
permits until years later. Ultimately, almost 93 percent of permitted 
Phase I MS4s were not permitted until after 1993. 

Figure 2: Issuance of Larger MS4s' Permits Has Taken up to 14 Years, 
and Some Are Still Not Issued: 

[See PDF for image] 

Source: GAO analysis of EPA and state data. 

Note: These data are for lead Phase I MS4s as of fall 2006. Phase I 
permit applications were due by November 1992 for large MS4s, and May 
1993 for medium MS4s. The percentage for 1993 represents 21 MS4s that 
first received permit coverage prior to or in 1993. The other 
percentages represent MS4s that received permit coverage in each year 
as follows: 55 MS4s in 1994, 45 MS4s in 1995, 30 MS4s in 1996, 34 MS4s 
in 1997, 35 MS4s in 1998, 10 MS4s in 1999, 23 MS4s in 2000, 4 MS4s in 
2001, 3 MS4s in 2002, 8 MS4s in 2003, 22 MS4s in 2004, 1 MS4 in 2005, 
and 4 MS4s in 2006. 644 MS4s received coverage as Phase I co- 
permittees, 25 MS4s received waivers, and 11 MS4s had not received 
either a permit or a waiver. 

[End of figure] 

Furthermore, as figure 3 shows, some Phase II MS4s did not receive 
permit coverage until years after the 2003 permit application deadline, 
and a substantial number are still not permitted.[Footnote 21] 

Figure 3: Many Phase II Permits Were Delayed or Are Not Yet Issued: 

[See PDF for image] 

Source: GAO analysis of EPA and state data. 

Note: These data are for lead Phase II MS4s as of fall 2006. Phase II 
permit applications were due by March 2003. The percentage for 2003 
represents 2,283 MS4s that first received permit coverage prior to or 
in 2003. The other percentages represent MS4s that received permit 
coverage in each year as follows: 1,881 MS4s in 2004, 306 MS4s in 2005, 
and 153 MS4s in 2006. 437 MS4s received coverage as Phase II co- 
permittees, 758 MS4s received waivers, and 809 MS4s had not received 
either a permit or a waiver. 

[End of figure] 

These delays in issuing MS4s their Phase I and II permits were caused, 
in part, by legal challenges to both the process of issuing storm water 
permits and the conditions of these permits. Several EPA, state, and 
other storm water experts we spoke with said that legal challenges had 
complicated implementation, and many of these individuals noted that 
the legal challenges limited permitting authorities' ability to issue 
permits on time. One challenge in particular--Environmental Defense 
Center v. EPA, decided in 2003 by the U.S. Court of Appeals for the 9TH 
Circuit--significantly affected the Phase II permitting process for 
MS4s in some states.[Footnote 22] At issue was whether permit 
authorities had to submit MS4s' permit coverage documents for public 
review and comment. To streamline the permit application process, EPA's 
Phase II regulations authorized the use of general permits for MS4s. 
However, the regulations did not require the notices of intent that 
MS4s submitted for coverage under these permits to be subject to public 
notice and comment. The court held that EPA's failure to require the 
review of these documents violated the Clean Water Act. To implement 
the court's decision, EPA issued guidance in April 2004 that instructed 
permitting authorities to make MS4s' notices of intent available to the 
public, provide the public with an opportunity to request a hearing, 
and review the notices of intent to ensure their consistency with the 
permit. 

Permitting authorities' responses to this decision delayed permit 
issuance for some MS4s. In particular, permitting authorities for a few 
states decided to change their Phase II permitting processes and issue 
individual permits--rather than general permits. An EPA Region 10 
official expressed concern over the number of legal challenges that had 
been filed against permitting authorities' use of general permits to 
address local environmental issues, and cited it as a key reason for 
these decisions. For example, EPA Region 10 was about to issue a 
general permit for Phase II MS4s in Alaska and Idaho when the 2003 
decision was issued. The EPA Region 10 official reported that, as a 
result of the decision, the region withdrew its draft permit and 
decided to issue watershed-based individual permits to MS4s to address 
specific water quality concerns. As a result of the extensive planning, 
consultation, and negotiation involved in issuing these permits 
individually, as of fall 2006 Region 10 had not issued permits to any 
Phase II MS4s in Idaho, although it had issued such permits to Phase II 
MS4s in Alaska.[Footnote 23] Similarly, an Oregon official noted that 
the state initially planned on issuing a Phase II general permit. 
Because of concern over the impact of the 2003 decision, the state 
decided to withdraw its draft general permit and issue individual 
permits to Phase II MS4s. 

Several EPA and state officials also noted that legal challenges to the 
specific conditions of storm water permits delayed permit issuance. For 
instance, an EPA Region 10 official reported that the issuance of 
Washington's Phase II general permits was delayed because environmental 
groups filed legal challenges against general permits the state had 
previously drafted for construction sites and industrial facilities. 
The environmental groups claimed that these general permits did not 
adequately address impaired waters. The Region 10 official said that 
these challenges caused Washington to postpone issuance of general 
permits for Phase II MS4s until it was able to resolve the groups' 
concerns about how to deal with impaired waters in the context of a 
general permit. Conversely, a South Carolina official said that cities 
in that state challenged the state's Phase II general permit because 
they considered the permit requirements to be too strict. 

Other factors also contributed to delays in issuing storm water permits 
to MS4s. For example, officials for two permitting authorities we spoke 
with linked the lack of staff resources available for implementing the 
storm water program to competing priorities, such as implementing the 
TMDL Program and other Clean Water Act efforts. Finally, some MS4s may 
not yet have received permit coverage because of questions about 
whether they should be included in the program. For example, Missouri 
has a few MS4s where the portion of the population served by the MS4 
inside the urban area is actually less than the population threshold 
set by the federal regulations. However, these MS4s may also serve 
people outside the urban area, bringing their total population served 
above the federal threshold. Missouri officials said that they are 
still trying to determine whether the state should require these MS4s 
to obtain permit coverage. 

Many Communities Have Not Fully Developed Storm Water Programs: 

Because many MS4s are still in the early stages of implementing their 
storm water permit requirements, the results of their efforts cannot 
generally be assessed. Nearly all Phase II MS4s (4,589, or over 99 
percent of those permitted) and some Phase I MS4s (38, or almost 13 
percent of those permitted) received their initial permits in 2002 or 
after, and therefore are still in their first permit terms. According 
to EPA regulations, MS4s are typically not responsible for full program 
implementation until the end of their first 5-year permit term. 
Therefore, state officials and storm water experts reported that they 
expect MS4s to take some time to gather information and develop their 
programs. For example, Virginia officials said that Phase II MS4s in 
that state would not have their programs fully developed and 
implemented until the end of the first permit cycle in 2007. 
Furthermore, because MS4s are still developing their storm water 
management programs, some EPA enforcement officials and state officials 
reported that they could not yet assess the degree to which MS4s are in 
compliance with the Phase II regulations. For example, a Connecticut 
official noted that some Connecticut MS4s had only submitted monitoring 
results on the types of pollution they face and did not yet have 
complete annual reports. 

Some permitting authorities who have attempted to assess the status of 
their MS4s' programs found instances in which MS4s may not be complying 
with the conditions of their permits. State and other storm water 
experts we spoke with noted that there is a wide variation in the level 
of noncompliance among permittees--ranging from MS4s that are 
inadequately implementing storm water BMPs to MS4s that are potentially 
violating their permits. For example, auditors for EPA Region 9 and the 
state of California recently discovered, among other things, that some 
MS4s (1) had not developed storm water management plans, (2) were not 
properly performing an adequate number of inspections to enforce their 
storm water ordinances, and (3) were lax in implementing BMPs at 
publicly owned construction sites. They also found that some MS4s were 
not adequately controlling storm water runoff at municipally owned and 
operated facilities, such as maintenance yards. In one case, an MS4's 
state-issued permit required it to reduce the discharge of pollutants 
into the storm sewer system to the maximum extent practicable. However, 
auditors found that street-sweeping debris was dumped about 10 feet 
from a storm drain inlet at one of this MS4's facilities. 

Several Factors Influence the Extent to Which Storm Water Program 
Implementation Is or Could Be a Burden: 

Because many MS4s are still in their first permit cycle, it is too 
early to determine the overall program burden. However, several factors 
influence the extent to which implementing the storm water program 
burdens an MS4. Some of these factors may lessen the burden MS4s face 
from implementing the storm water program, while other factors may 
increase the extent to which the program is a burden on MS4s. In the 
future, several factors could increase the program's burden on MS4s. 

Regulatory Flexibility, Preexisting Activities, and Communities' 
Decisions Can Help Reduce Burden: 

According to EPA, the maximum extent practicable standard gives MS4s 
the flexibility to design their storm water programs using BMPs that 
require varying levels of effort and cost. MS4s may choose to implement 
lower cost, nonstructural BMPs over higher cost structural activities. 
In addition, MS4s can choose less costly routes to implement specific 
BMPs. For example, MS4s could be in compliance with permit requirements 
if they: 

* educate the public by having a few staff members put together a storm 
water brochure rather than hiring a public relations firm to develop an 
elaborate media campaign, or: 

* map their storm sewer systems by purchasing a map from a gas station 
and asking volunteers to place dots where the storm water outfalls are 
located rather than creating a sophisticated geographic information 
system map. 

MS4s can also count storm water management efforts they were 
undertaking before implementing the program towards meeting their 
permit requirements. In some cases we found the following: 

* States already had regulations to control runoff. Several state 
officials reported that they had state storm water regulations in place 
prior to the storm water program. Some of these state programs began as 
early as the 1970s. 

* MS4s were already implementing comprehensive storm water management 
programs. For example, officials in Austin, Texas, said that by 1991, 
virtually the entire city was required to have structural storm water 
controls, and that almost all of the controls that it now uses to meet 
the conditions of its storm water permit were in place prior to the 
permit. In addition, a local official in Florida said that some MS4s in 
his region were implementing storm water management programs long 
before they were covered under an NPDES permit. 

* MS4s conducted individual storm water management activities. For 
example, several officials reported that MS4s were conducting street- 
sweeping, reviewing construction site plans, or collecting household 
hazardous waste prior to the storm water program. 

Federal and state program funds and/or dedicated funding sources can 
also be used to reduce the burden of implementing the storm water 
program. As of June 30, 2006, 19 states had provided nearly $346 
million in low-interest loans from their CWSRF programs for storm water 
projects. Several state officials also reported that their states 
established programs to provide financial support to MS4s to implement 
storm water activities. For example, a New Jersey official said the 
state allotted $12 million in grants to help MS4s establish their 
programs, and a Rhode Island official noted that most MS4s in the state 
took advantage of a state grant to implement their storm water 
programs. In addition to obtaining federal or state funds, officials 
and other experts reported that MS4s can reduce program burdens by 
creating a dedicated funding source, such as a storm water utility--a 
fee-based funding mechanism. With a dedicated funding source, MS4s may 
reduce the annual budgetary competition with other local funding 
priorities. At the same time, fee-based utilities help to ensure equity 
by linking fees to the demands placed on the storm drain system caused 
by the runoff from a particular property. 

MS4s may also reduce the burden they face by sharing program 
responsibilities with the state, co-permittees, or other entities. For 
example, Maine officials said MS4s in that state shared costs by 
implementing a statewide mass media campaign on storm water management. 
The Phase II regulation encouraged MS4s to work cooperatively with 
other Phase II MS4s or with a Phase I MS4 to avoid duplicative efforts 
and take advantage of economies of scale. For example, a Maryland 
official said that some Phase I Maryland counties have taken on 
responsibility for helping nearby Phase II MS4s implement the program. 
In one instance, a county is implementing the storm water requirements 
for Phase II MS4s within the county in exchange for funding 
contributions from the MS4s. The county is also submitting annual 
reports for these MS4s. 

In addition, MS4s may plan public projects in ways that minimize 
additional expenditures and reduce program burden. One storm water 
expert we spoke with reported that MS4s can avoid large-scale 
structural control costs associated with new development sites by 
planning to capture and filter storm water runoff at its source through 
environmentally friendly techniques, such as conservation design. MS4s 
can also help reduce storm water costs by scheduling construction 
activity for a time of year when the erosion potential of a site is 
relatively low. Additionally, some cities have begun to take steps to 
manage storm water runoff before it gets into the storm sewer system-- 
thereby avoiding costly infrastructure investments. For instance, 
Portland, Oregon, recently implemented a program to redirect 
residential downspouts to decrease the amount of storm water that 
drains into city sewers. City officials noted that, to date, 44,000 
residential downspouts have been redirected, preventing 1 billion 
gallons, or approximately one-tenth of the city's storm water, from 
draining into sewers. By redirecting storm water onto lawns and 
gardens, the city was able to avoid spending billions of dollars on 
structural storm water systems and treatment plants. 

Finally, MS4s may not experience significant program burden if, as is 
often the case, developers and contractors have primary responsibility 
for implementing BMPs at new and redeveloped sites. According to some 
storm water experts we spoke with, developers and contractors--and not 
MS4s--are implementing the more expensive structural BMPs. Such 
controls may be needed to meet certain water quality standards. 
Moreover, experts reported that some MS4s are reducing burden by 
negotiating with homeowners' associations or other groups to maintain 
these BMPs. 

Additional State Permit Requirements and Litigation, Among Other 
Factors, May Increase Program Burden: 

Other factors may increase burdens. First, some states may require that 
MS4s implement more stringent or specific storm water permit 
requirements than envisioned by federal regulations to address local 
water quality concerns. Several state officials noted that such 
requirements can (1) widen the scope of the program, (2) establish 
additional program goals, (3) institute implementation deadlines, or 
(4) delineate minimum acceptable activity levels and measurable goals, 
which can increase the burden for MS4s. For example, Washington's storm 
water program includes all of the portions of MS4s--not just the 
portion in the census-defined urban areas as required by the federal 
regulations--as well as the growth areas of unincorporated counties. In 
addition, one of Washington's Phase II general permits includes a 
requirement that MS4s adopt an ordinance or other enforceable mechanism 
to regulate runoff from new and redeveloped sites that conforms to the 
state's standards within 2-1/2 years of receiving permit coverage. The 
permit also requires that MS4s institute a process to review 
development plans, inspect sites, and enforce ordinance compliance 
within the same time frame. Specific requirements, such as those to 
increase the frequency of activities, can burden MS4s. For instance, 
one Wisconsin MS4 reported that a state requirement to increase the 
frequency of cleaning catch basins (which are part of its storm sewer 
system) will impose a burden. Additional manpower is needed during 
warmer months because the city can only clean catch basins during 
portions of the year when the temperature is above freezing. Also, 
shifting these resources away from current sanitary sewer maintenance 
may expose the city to more sanitary sewer overflows than in the past. 
In addition, while previously an annual larvicide treatment effectively 
treated West Nile Virus, multiple applications will now be necessary 
because of the increased catch basin cleaning. 

Second, permitting authorities have made adjustments to their 
permitting procedures in response to legal challenges that have 
affected program requirements for some MS4s. For instance, an EPA 
Region 10 official said that the decision to issue watershed-based 
individual permits in lieu of a Phase II general permit in response to 
the 2003 court decision has resulted in more prescriptive permits than 
were required by the Phase II rule. Similarly, legal challenges to the 
conditions of permits have required permitting authorities to adjust 
these conditions. For example, a Minnesota official said that an 
environmental group challenged the state's initial Phase II general 
permit, arguing that the permit would not prevent water quality from 
being degraded in certain waters designated by the state as highly 
valued. As a result of this challenge, Minnesota revised its permit to 
include additional requirements for communities discharging to these 
waters, such as evaluating their BMPs to determine whether any 
adjustments are needed to prevent storm water pollution that could 
degrade water quality. Such provisions could require these communities 
to take actions that nondesignated communities do not have to take. 
Legal challenges have also resulted in consent decrees that require 
MS4s to increase efforts to control storm water runoff. One storm water 
expert we spoke with stated that some MS4s do not take the program 
seriously until they are forced to do so under a court order. 

Third, state and other storm water officials said that requirements 
incorporated into storm water permits as a result of other regulatory 
programs--such as the TMDL Program, Combined Sewer Overflow Policy, or 
Endangered Species Act--have increased program burdens. For example, a 
Washington official reported that TMDL requirements for one local 
drainage area include, among other things, monitoring for fecal 
coliform concentrations at certain places and times, and doing 
additional source tracing. Also, one expert noted that incorporating 
TMDL requirements into storm water permits can be costly if they lead 
to expensive controls such as piping storm water to a wastewater 
treatment plant. Officials in Portland, Oregon, said that their storm 
water costs are higher because activities the city is implementing as a 
result of the Combined Sewer Overflow Policy are being implemented 
citywide--even in areas where the storm and sanitary sewers are 
separate. Further, Austin, Texas, officials reported that the city is 
taking additional steps through its storm water activities, such as 
purchasing land, to address endangered species concerns. 

Fourth, resource limitations can make program implementation more 
burdensome. Specifically, federal funding for storm water activities 
has either decreased or been eliminated in recent years. For example, 
according to EPA, federal allocations to state CWSRF grants decreased 
by about $259 million (nearly 20 percent) between 2004 and 
2005.[Footnote 24] Also, EPA officials noted that the water quality 
cooperative agreements grants program--otherwise known as the 104(b)(3) 
grants program--is ending. The officials said that at the peak of the 
program, the agency was providing around $19 million in grants. Program 
funds were used, in part, to support innovative storm water projects. 
Additionally, officials for two permitting authorities said they lack 
sufficient staff to provide technical assistance to MS4s as they 
develop their storm water management programs. For instance, an Oregon 
official noted that the state did not have any full-time storm water 
staff until 4 years ago, which limited the state's ability to review 
MS4s' storm water management documents in any meaningful way. Storm 
water officials we spoke with also cited limited technical expertise 
and staff availability in Phase II MS4s as an implementation burden. 

Fifth, some MS4s may face barriers to obtaining additional resources, 
such as CWSRF loans, to help implement storm water activities. EPA 
storm water and CWSRF program officials stated that smaller MS4s may 
not be aware of the availability of CWSRF loans. The officials 
attributed this lack of awareness to a variety of factors, including 
the limited outreach on the eligibility of storm water projects for 
CWSRF loans that has been conducted to date, as well as the 
inexperience of local officials who are just beginning to develop storm 
water programs.[Footnote 25] In addition, confusion caused by unclear 
language in EPA guidance could affect some MS4s' ability to obtain 
CWSRF loans. For example, one EPA guidance document lists a number of 
storm water activities that may be eligible for CWSRF loans, such as 
rehabilitating a storm sewer system, constructing a wetland, or 
purchasing a street sweeper, as long as such efforts are undertaken by 
a publicly owned system. However, another section of the same guidance 
indicates that an MS4 may be eligible for CWSRF funding "so long as it 
is a publicly owned treatment works." While this guidance is consistent 
with federal requirements that entities receiving certain CWSRF loans 
be publicly owned, a Publicly Owned Treatment Works is a specific type 
of facility regulated under the Clean Water Act and is not part of an 
MS4. According to one state official, the state had not issued any 
loans for storm water activities under this category of CWSRF loans 
because the official believed that loans were only authorized for a 
publicly owned treatment work. Furthermore, many smaller Phase II MS4s 
may lack dedicated funding sources. However, these MS4s need to show 
that they have sufficient revenue streams to pay back a CWSRF loan in 
order to be eligible to receive such a loan. In some cases, the ability 
of these MS4s to establish utilities and other dedicated funding 
sources is limited by a lack of authority or local opposition. For 
example, according to a California official, a 67 percent vote of 
public approval is required in that state for certain new taxes and 
fees. 

Finally, MS4s' characteristics--environmental or geographic conditions 
and age of infrastructure--can increase the burden of implementing the 
storm water program. State and other storm water experts we spoke with 
noted that the quality of local receiving waters, whether degraded or 
highly valued, is an important factor in determining if an MS4 has to 
implement additional BMPs. Geographic conditions, such as the 
characteristics of local soils or topography, may affect which BMPs an 
MS4 selects. For example, if the composition of an MS4's soil does not 
allow storm water to infiltrate, the MS4 may need to construct a more 
costly BMP. In addition, older MS4s will have higher illicit discharge 
detection and elimination program costs than newer MS4s because older 
infrastructure can mean more incidences of illicit connections and more 
pipes needing repair. 

Storm Water Program May Impose Additional Burdens in the Future: 

Four factors could increase MS4s' storm water program burden in the 
future. First, burdens could increase as EPA and state permitting 
authorities reissue permits with more stringent or specific 
requirements than they have done to date. Future permits may be more 
rigorous because, as EPA described in publication of its regulations, 
compliance with the Clean Water Act's standard of preventing storm 
water pollution to the maximum extent practicable should be an 
iterative process in which MS4s refine their efforts over time as they 
consider current conditions and the effectiveness of their BMPs. For 
example, a California official said that with each generation of Phase 
I permits the state has included more specific requirements, such as 
(1) identifying the BMPs MS4s have to implement, (2) establishing which 
pollutants MS4s should focus on reducing, and (3) detailing the number 
of times MS4s must inspect a BMP. Similarly, EPA Region 5 officials 
noted that states in the region that issued their permits at the 
beginning of Phase II did not include specific conditions, while those 
states that issued their permits more recently included specific 
conditions, such as dates by which MS4s had to complete certain 
activities. The officials said that as permits expire and are reissued, 
they expect that more states will make their permit conditions and 
compliance dates more specific. EPA and state officials and other storm 
water experts also said that permits could become more burdensome as 
requirements get added from future TMDLs. While TMDLs have been 
required by the Clean Water Act since 1972, until recently many had not 
been developed. As a result, according to EPA, the agency is under 
court order or consent decrees in many states to ensure that TMDLs are 
established, either by the state or EPA. Several of the officials we 
spoke with noted that they expect additional storm water permit 
requirements as more TMDLs are developed. Future legal challenges to 
storm water permits could also increase permit requirements, as they 
have done in the past, according to EPA and state permitting officials. 

Second, more aggressive permit enforcement could cause some MS4s to 
increase their efforts to implement the program--potentially increasing 
the burdens that they face. With some exceptions, enforcement of MS4s' 
storm water permits has been limited to date. EPA Region 5 officials 
said that early on, the agency's role in implementing the storm water 
program was to make sure that state permitting authorities adopted the 
necessary rules, issued permits to MS4s, and kept the permits current. 
According to one storm water expert, the lack of permit enforcement 
has, in part, contributed to some MS4s' noncompliance with their permit 
conditions. However, officials with EPA's Office of Enforcement and 
Compliance Assurance (the Office), told us they expect to emphasize the 
enforcement of storm water permits in the future. For example, while 
officials stated that the Office had conducted only 30 Phase I MS4 
audits as of November 2006, one official said that EPA is taking steps 
to increase compliance monitoring of Phase I and II MS4s by, for 
example, modifying audit guidance and procedures to increase the 
efficiency of audits and conducting additional training for EPA 
regional staff. The official said that these steps are designed to help 
the agency increase the number of MS4 audits and inspections it 
conducts. In addition, while EPA regions and states have conducted 
limited MS4 audits to date, some states may be planning to increase 
their efforts. For example, EPA Region 5 officials said that the region 
is training state officials to conduct MS4 audits because several 
states in the region have expressed interest in auditing MS4s' permit 
compliance in the future. 

Third, the need to maintain and replace BMPs could raise storm water 
program costs in the future, and some experts indicated that BMP 
maintenance and replacement costs are among the more significant 
determinants of MS4s' overall program costs. In particular, some BMPs 
with relatively low installation costs have relatively high maintenance 
costs. For example, a local storm water official noted that one BMP--an 
underground storm water storage unit--is relatively less expensive to 
install (since it does not require land space), but could be difficult 
and costly to maintain. As a result, the official advocated that MS4s 
consider the life-cycle costs of a BMP; that is, the administrative, 
inspection, and maintenance costs--in addition to the installation 
costs--as they determine which BMPs to use. Furthermore, while private 
developers often bear the costs of initially installing storm water 
BMPs at new or redeveloped sites, MS4s may be responsible for 
maintaining these BMPs over the long term. However, some storm water 
officials we spoke with said that MS4s are not adequately planning to 
maintain storm water BMPs over the long term. Some BMPs, if not 
properly maintained, may need to be replaced, which could lead to even 
greater costs for MS4s. In addition, some BMPs may simply prove 
ineffective at meeting the storm water management goals for which they 
were initially installed. Several storm water experts said that because 
data are not available on the effectiveness of some BMPs, controls are 
being installed that may not achieve their intended purposes over time. 
Should MS4s need to redesign or replace ineffective BMPs, they could 
face additional costs. 

Fourth, the program burdens MS4s face may also increase over time 
because federal funds may not be available for storm water projects-- 
particularly CWSRF loans. Officials in a few states we contacted said 
continued reductions in federal CWSRF allocations could hurt their 
ability to fund storm water projects in the future.[Footnote 26] For 
example, in Massachusetts--which has used more CWSRF funds for storm 
water projects than any other state except Florida--officials told us 
that while current CWSRF funding levels have been sufficient, future 
reductions in the federal allocation to the state's CWSRF fund could 
prevent the state from being able to fund lower priority projects, such 
as storm water projects. Similarly, New Jersey officials said that 
recent decreases in federal funding will require the state to set 
priorities for distributing CWSRF funds for projects, and municipal 
storm water needs would be the first category of projects left off the 
state's priority list if sufficient funding is not available for all 
projects. They also told us that recent reductions in the federal 
allocation to the state's CWSRF fund have reduced New Jersey's ability 
to pay the administrative costs of the CWSRF program. Consequently, New 
Jersey has initiated a fee--2 percent of a project's cost--to help with 
administrative costs. However, such a fee reduces the funds available 
to MS4s to implement their storm water projects. 

Methodological Concerns Raise Questions about the Usefulness of EPA's 
Cost Estimates: 

Because we could not identify the overall burden MS4s face from the 
storm water program, in part due to its early implementation, we could 
not determine whether EPA's Phase I and II analyses over-or 
underestimated actual costs for implementing the storm water program. 
However, the methodologies and data used in these studies raise 
questions about whether the estimates of potential storm water costs 
identified in EPA's analyses are useful as indicators of actual program 
costs. Specifically, (1) the methodology used for the Phase I analysis 
was not designed to estimate national program costs or present ranges 
that reflect actual program costs, (2) concerns about the validity and 
reliability of Phase II data call into question the usefulness of the 
Phase II cost estimates, and (3) EPA's Phase I and II analyses did not 
exclude the costs of storm water activities that MS4s may have been 
conducting prior to the program. 

Phase I Methodology Did Not Realistically Estimate Ranges of National 
Program Costs: 

EPA's 1990 Phase I cost estimate did not provide realistic estimates of 
the range of national program costs. EPA estimated Phase I program 
costs by modeling storm water management programs for eight 
hypothetical cities representing a range of storm sewer system sizes, 
climatic conditions, topographies, and other characteristics relevant 
to storm water pollution.[Footnote 27] EPA modeled the per capita 
capital and operations and maintenance costs of storm water management 
programs for these cities under the following three scenarios: 

* Scenario 1 assumed the cities would implement the program using 
nonstructural BMPs, such as educating the public and cleaning storm 
water catch basins; 

* Scenario 2 assumed the cities would implement both nonstructural BMPs 
and structural controls, such as excavated ditches that hold runoff and 
allow it to settle (known as infiltration trenches) or vegetative 
buffer zones; and: 

* Scenario 3 assumed the cities would install more complex controls, 
such as storm water treatment facilities, in addition to nonstructural 
BMPs. 

Because of multiple limitations in this analysis, however, EPA's Phase 
I costs cannot be statistically projected to identify national program 
costs. First, while the characteristics of the hypothetical cities EPA 
developed were based on actual cities, these cities were not designed 
to be representative of Phase I MS4s nationwide. Second, EPA did not 
model program costs for a Phase I county, even though 47 (or over 21 
percent) of the communities designated by the Phase I rule for 
potential coverage under the program were counties. Third, owing to the 
many assumptions it made about storm water volumes and other local 
characteristics, EPA acknowledged that the costs identified for its 
hypothetical cities should only be viewed as gross indicators of 
compliance costs, not actual compliance costs. Finally, EPA's cost 
estimate did not consider Phase I permit application costs, which 
according to one state official we spoke with, have been one of the 
more expensive aspects of the program.[Footnote 28] For example, we 
found that six Phase I MS4s reported spending (or expecting to spend) 
between $237,000 and almost $1.1 million on preparing their Phase I 
permit applications.[Footnote 29] 

In addition, the three scenarios identified in EPA's Phase I analysis 
did not provide a realistic estimate of the potential range of Phase I 
costs. Specifically, estimated costs for the third scenario in EPA's 
Phase I analysis--implementing complex controls to treat storm water-- 
were so high that the agency considered it unlikely that permit 
authorities would require this level of effort. While presenting cost 
estimates in terms of ranges is generally desirable for capturing 
uncertainty in the data used and assumptions made, EPA's inclusion of 
an unrealistic scenario in this analysis limits the usefulness of its 
estimated cost ranges. Table 2 shows the ranges of per capita capital 
and operations and maintenance costs in each of EPA's three Phase I 
scenarios. 

Table 2: EPA's Estimated Ranges of Phase I Per Capita Costs: 

Scenario: (1); 
Low: Capital: $18; 
Low: Operations and maintenance: $25; 
High: Capital: $318; 
High: Operations and maintenance: $263; 
Weighted average[A]: Capital: $73; 
Weighted average[A]: Operations and maintenance: $61. 

Scenario: (2); 
Low: Capital: 29; 
Low: Operations and maintenance: 30; 
High: Capital: 378; 
High: Operations and maintenance: 372; 
Weighted average[A]: Capital: 91; 
Weighted average[A]: Operations and maintenance: 76. 

Scenario: (3); 
Low: Capital: 54; 
Low: Operations and maintenance: 32; 
High: Capital: 61,279; 
High: Operations and maintenance: 13,456; 
Weighted average[A]: Capital: 7,924; 
Weighted average[A]: Operations and maintenance: 1,762. 

Source: GAO analysis of EPA data. 

Note: Dollars are adjusted to 2006. 

[A] Average Phase I per capita costs weighted by the population of the 
hypothetical cities used in EPA's analysis. 

[End of table] 

Limitations in the Validity and Reliability of the Phase II Data Raise 
Questions about the Cost Estimates: 

EPA's 1999 Phase II cost estimate may not accurately represent program 
costs because of limitations in the validity and reliability of the 
data the agency used. EPA estimated costs using data from a 1998 survey 
distributed to more than 1,600 potential Phase II MS4s by the National 
Association of Flood and Stormwater Management Agencies 
(NAFSMA).[Footnote 30] NAFSMA sought to (1) solicit information about 
the proposed Phase II storm water program, (2) identify current storm 
water spending levels in Phase II MS4s, and (3) identify future needs 
for these communities. It asked the MS4s to provide annual cost data 
for activities they were conducting related to the BMP categories 
identified by EPA's Phase II regulations. 

EPA calculated Phase II costs by dividing reported costs per BMP 
category by the population of that MS4, and then multiplying the result 
by an estimated 2.62 persons per household. In addition to costs for 
implementing storm water BMPs, EPA estimated costs for complying with 
Phase II administrative requirements, such as submitting applications 
and annual reports. These costs were averaged over a 5-year permit term 
to obtain an annual administrative cost per household. EPA calculated 
total Phase II costs by multiplying its per household estimate by the 
number of households (based on total population)[Footnote 31] in the 
5,040 MS4s it anticipated would be regulated by the Phase II 
rule.[Footnote 32] 

While EPA's analysis attempts to provide a national estimate of costs, 
the NAFSMA survey data EPA used are not sufficiently valid or reliable 
for the purpose of making national projections about MS4s' costs for 
implementing Phase II. Of the over 1,600 MS4s surveyed, only 56 (about 
3.5 percent) responded with cost data EPA could use for its analysis. 
Furthermore, the overwhelming majority of the 56 MS4s did not provide 
responses to all of the questions that EPA relied upon to develop its 
estimates. In fact, only 5 MS4s provided data for all of the questions 
that EPA used, while 17 MS4s provided cost data in response to only one 
of the questions EPA used for its cost analysis. Because of the low 
response rate to the NAFSMA survey--either in aggregate or for each of 
the cost data questions--the MS4s that responded to the survey are 
unlikely to be a representative sample of Phase II MS4s 
nationwide.[Footnote 33] 

The NAFSMA survey data are also of limited usefulness because many of 
the key survey questions were extremely complicated and subject to 
multiple interpretations, making it unlikely that MS4s could have 
responded with valid and reliable information. For example, one survey 
question asked whether an MS4 had a map of its storm sewer system. If 
the MS4 answered no, then the survey asked the MS4 to provide an 
estimate of the costs of preparing such a map. However, the survey 
design did not lead the MS4 to provide cost information if it answered 
yes to this question. As a result, an MS4 that had already developed a 
storm sewer map (and should therefore have more reliable data on the 
cost of this effort than an MS4 that had not yet done so) was not 
prompted to provide data on the costs of developing a map in its survey 
response. Moreover, the NAFSMA survey questions did not clarify the 
types of costs that MS4s should report in their responses, leaving 
substantial room for MS4s to interpret questions themselves. For 
example, one question asked MS4s to report the annual costs of their 
program to detect and address illicit discharges, including illegal 
dumping, to the storm sewer system. However, the question provided no 
guidance on what types of activities or costs MS4s should include in 
their responses. 

As we have found in conducting surveys, such survey questions are 
unlikely to yield reliable responses. For example, EPA guidance on 
developing programs to detect and eliminate illicit discharges 
indicates that these programs can have a number of different 
components, including mapping systems, testing for illicit discharges, 
and taking steps to eliminate discharges. However, without providing 
more specific directions on the illicit discharge activities for which 
MS4s were to report cost data, the NASFMA survey cannot ensure that the 
data MS4s reported were comparable. Furthermore, the NAFSMA survey did 
not instruct MS4s on the types of costs, such as capital, 
administrative, and labor, they were to report. The difficulty in 
obtaining reliable data on MS4s' storm water activities is illustrated 
by the experience of one California expert we contacted. This 
researcher attempted to survey six MS4s in California to obtain data on 
the costs of implementing the storm water program. However, owing in 
part to inconsistencies in reporting between the MS4s, he had to 
conduct document reviews as well as in-depth interviews with local 
officials in order to obtain reliable data. 

The problems with the data collected through the NAFSMA survey limit 
the usefulness of EPA's Phase II cost estimate. Based on the responses 
from 56 MS4s, EPA estimated that annual program costs would be $4.30 
per capita, and reported that these costs could range from $0.20 to 
$25.75 per capita.[Footnote 34] However, the size of the sample of MS4s 
that responded with cost data for all of the questions on program 
implementation costs was so small that the survey data cannot be used 
to make precise estimates. As a result, EPA's Phase II cost estimate is 
much less meaningful than EPA suggests. Finally, even if the MS4s that 
provided the data used in this estimate constituted a representative 
sample of Phase II MS4s nationwide, the NAFSMA survey questions were 
not designed to ensure that these data were reliable--raising questions 
about the extent to which the data represent actual program costs. 

EPA's Phase I and II Analyses Did Not Exclude Previous Costs of Storm 
Water Activities: 

The methodologies EPA used in its Phase I and II analyses also raise 
concerns because neither of these analyses excluded costs for storm 
water activities that MS4s were undertaking before implementing the 
storm water program. As a result, EPA cannot determine the costs 
specifically attributable to the program's requirements and could 
overestimate costs, which limits the estimates' usefulness as 
indicators of actual program costs. In its Phase I analysis, EPA 
assumed that none of its hypothetical cities had any storm water 
controls already in place. For Phase II, since the NAFSMA survey was 
distributed in 1998 (prior to the Phase II final regulation in 1999 and 
permit application deadline in 2003), the cost data reported by MS4s 
largely represent costs for activities that they were undertaking prior 
to the program. However, EPA did not exclude costs for such preexisting 
activities from its final cost estimate. 

Several state officials and storm water experts we contacted said that, 
before the program, some states had regulations in place to control 
storm water runoff, and that some MS4s were (1) implementing 
comprehensive storm water management programs or (2) conducting some 
storm water management activities, such as street-sweeping. In 
addition, in publication of both the Phase I and II regulations, EPA 
encouraged MS4s to consider these preexisting activities as they 
developed their storm water management programs. Therefore, to the 
extent that MS4s were already taking steps to control storm water, the 
costs of these activities should not be considered program 
implementation costs. 

Assessment of Program Burden Is Hampered by Limited and Inconsistent 
Data: 

Any assessment of program burden will be hampered by limited and 
inconsistent data on MS4s' storm water activities and their costs. 
Specifically, while MS4s must submit reports of their storm water 
activities to their permitting authority, according to EPA storm water 
officials, there are no national guidelines on what should be included 
in an MS4's annual report, and federal regulatory requirements are weak 
and unspecific. For example, the officials noted that the Phase I rule 
directs MS4s to provide information on expenditures and budgeted 
amounts in their annual reports, but does not include any specific 
direction on what costs should be tracked and how they should be 
reported. Moreover, the Phase II rule did not require MS4s to report 
any cost data. 

Without standard reporting guidelines, we found it difficult to use the 
data in MS4s' reports to assess the costs of the storm water program. 
Specifically, we examined the cost data (either total expenditures or 
unit costs) and output information (such as the number of miles swept 
or the number of catch basins cleaned)[Footnote 35] in the most recent 
annual report provided by the 130 MS4s we sampled.[Footnote 36] We also 
collected data from other studies of BMP costs. Using all these data, 
we tried to identify cost ranges for selected BMPs that we thought 
would be among the more commonly implemented BMPs: public education 
efforts, catch basin/storm drain cleaning, street-sweeping, and illicit 
discharge detection and elimination activities. 

However, our efforts to develop reliable data for all the selected BMPs 
were hampered by the limited and inconsistent data available in MS4s' 
annual reports. Some MS4s' annual reports were hundreds of pages long, 
with detailed data on their activities, while other MS4s' annual 
reports provided little evidence of their storm water activities or 
costs. In one case, a Phase II MS4's annual report was a one-page 
letter which stated that the MS4's estimated implementation status was 
"20-percent," but provided few additional details. As a result, we were 
only able to estimate the per capita costs of street-sweeping 
activities, which we found ranged between $0 and $17.51 for Phase I 
MS4s, and between $0 and $9.61 for Phase II MS4s.[Footnote 37] 

Overall, we found that about 18 percent of the MS4 annual reports we 
reviewed reported some cost data in such a way that these data could be 
used to develop ranges of BMP costs. Some MS4s' annual reports provided 
cost data that were not usable for certain BMPs. In these cases, MS4s' 
data for some BMPs were not useful because MS4s tracked and reported 
costs differently. For example, where they reported cost data, Phase I 
MS4s frequently reported total cost data for their activities. One 
Phase I MS4 estimated spending $770,000 on public education activities, 
including workshops, seminars, and education programs and materials 
related to water quality. However, in order to compare this total cost 
figure with the total cost for other MS4s, we needed sufficiently 
detailed activity data to allow for comparisons among programs with 
similar activities. Because MS4s usually did not report their 
activities and costs consistently, it was not possible to make such 
comparisons. 

In addition, we found that 60 percent of the MS4 annual reports we 
reviewed reported output data in such a way that these data could be 
used to estimate BMP costs. Some reports included output data that 
could not be used to estimate BMP costs. In these instances, MS4s' 
output data for certain BMPs were not usable because MS4s either did 
not report BMP output measures consistently, or did not provide 
sufficiently detailed information to estimate BMP costs. Even for 
street-sweeping, such inconsistencies could cause errors in the 
estimated cost ranges we developed. For instance, some reports provided 
estimates of the number of lane miles swept; others provided estimates 
of the number of curb miles swept; and still others simply provided 
estimates of the miles swept.[Footnote 38] 

The difficulties in assessing program costs due to a lack of detailed 
and consistently reported data are even greater for more complex BMPs, 
which may have different types of costs. For example, the development 
of local ordinances to control storm water pollution entails legal and 
administrative costs, while the installation of a structural storm 
water control (such as a retention pond) entails costs for planning, 
mobilization and demobilization of construction equipment, materials, 
labor, and land. Moreover, the costs of implementing any particular BMP 
are influenced by a number of factors. For example, the costs of street-
sweeping are influenced by fuel and dumping costs, equipment choices, 
and the type of material being swept. Therefore, the more potential 
variation in the scope and nature of a BMP, the greater the need for 
MS4s to report detailed and consistent data on the costs of that BMP in 
order to ensure that the data reported will be comparable to those of 
other MS4s.[Footnote 39] 

Several state officials and storm water experts we contacted cited 
problems with these data as a barrier to assessing program 
implementation, costs, or effectiveness. For example, according to one 
expert who has audited MS4s as an EPA contractor, the quality of annual 
reports varies so greatly that site visits are often necessary to 
obtain accurate and detailed information about program activities. One 
MS4's annual report contained documentation showing that MS4 officials 
conducted two storm water site inspections per week. However, when the 
expert visited the MS4, he found that there were 5,000 sites that 
needed to be inspected, and at the rate of 2 sites per week, it would 
take the MS4 nearly 50 years to visit each of its sites once. In 
addition, EPA storm water officials said that without clear and 
specific reporting guidance, it is difficult to identify the 
incremental activities and costs attributable to the storm water 
program. The officials said that some MS4s may be reporting costs for 
all storm water program activities--including costs for those 
activities that were taking place prior to permit implementation. 

Without specific guidance directing MS4s to provide detailed 
information in their annual reports on storm water activities, costs, 
and results, EPA will encounter difficulties similar to those we and 
others have encountered in using the data in the annual reports to 
assess program implementation, burdens, and effectiveness. For example, 
in January 2007, EPA published guidance that identified a number of 
data elements that permitting authorities may use for evaluating an 
MS4's compliance with program requirements by reviewing its annual 
reports. While the guidance, for example, instructs evaluators to 
identify the total number of sites an MS4 must inspect to determine if 
an inspection schedule is adequate, there are no requirements for 
reporting detailed data on costs and activities in a consistent manner. 
Therefore, the data collected through these evaluations will likely 
vary by MS4, and be of limited use in assessing program implementation 
nationwide. 

Conclusions: 

Because many communities are still implementing their first permits for 
controlling storm water runoff, it is too early to determine the extent 
to which implementation of the program has been a burden. Furthermore, 
the number of factors that can influence a community's storm water 
management activities--such as whether it can take advantage of the 
flexibility provided by EPA's storm water regulations to implement the 
program in a less expensive manner, its current level of water 
pollution, or whether it shares program responsibility with other 
entities--make it difficult to develop a uniform assessment of the 
burden the storm water program may impose nationwide. For some 
communities, challenges to establishing reliable sources of local 
funding for their efforts can also influence the extent to which the 
program is a burden. 

Furthermore, it will be difficult to assess the burden of implementing 
the storm water program, and for EPA to meet its goal of evaluating 
Phase II starting in 2012, without more complete and consistent 
reporting on the scope, costs, and results of communities' storm water 
best management practices. 

Recommendation for Executive Action: 

In order to enable EPA to evaluate the implementation of the storm 
water program, we are recommending that the Administrator, EPA, issue 
additional program guidance and consider regulatory changes to ensure 
that (1) communities report on activities in sufficient detail to 
determine their scope, costs, and results; and (2) communities report 
this information consistently so that it can be analyzed on a national 
basis. 

Agency Comments and Our Evaluation: 

We provided a draft copy of this report to EPA for review and comment. 
EPA's letter is presented as appendix IV. 

In its written comments, EPA stated that it had already taken some 
steps to implement our recommendation, but agreed to investigate ways 
to collect better cost data through communities' annual reports. While 
EPA indicated that it does not believe it appropriate to modify the 
annual reporting regulations at this time, it said it had initiated an 
effort to identify the information communities should submit in their 
annual reports and to develop corresponding guidance for EPA regions, 
states, and local communities. Furthermore, EPA supported the 
development of an annual report template or other information 
collection tool. We believe that this tool, along with EPA guidance, 
could help the agency obtain better data for evaluating program 
implementation. However, some states have already established their own 
annual report formats, and may choose not to follow EPA's guidance or 
use its annual report template. As a result, without strengthening 
federal regulatory reporting requirements--such as requiring Phase II 
MS4s to report cost data--we believe it could be difficult for EPA to 
obtain consistent and reliable data on program implementation 
nationwide. EPA also provided technical comments, which we incorporated 
as appropriate. 

We are sending copies of this report to the congressional committees 
with jurisdiction over EPA and its activities; the Honorable Stephen L. 
Johnson, Administrator, EPA; and the Honorable Rob Portman, Director, 
Office of Management and Budget. In addition, this report will be 
available at no charge on the GAO Web site at http://www.gao.gov. 

If you have any questions about this report or need additional 
information, please contact me at (202) 512-3841 or 
stephensonj@gao.gov. Contact points for our Offices of Congressional 
Relations and of Public Affairs may be found on the last page of this 
report. Key contributors to this report are listed in appendix V. 

Signed by: 

John B. Stephenson: 
Director, Natural Resources and Environment: 

[End of section] 

Appendix I: Briefing Provided to Congressional Requesters: 

To view appendix I, "Briefing Provided to Congressional Requesters," 
click here: http://www.gao.gov/d07479appendix1.pdf. We are providing 
the 51-page briefing as a separate file because the file size is 32 
megabytes. Therefore, appendix II begins on the following page, page 
91. 

[End of section] 

Appendix II: Objectives, Scope, and Methodology: 

In September 2004, Senator Inhofe and Senator Crapo requested that GAO 
provide information on the cumulative costs that communities face from 
implementing drinking water and clean water regulations. In particular, 
Senator Inhofe and Senator Crapo noted that communities faced the costs 
of implementing Phase II of the storm water program at the same time 
they were expected to address a number of other water pollution 
prevention regulations. However, the Environmental Protection Agency's 
(EPA) estimates of the costs of implementing these regulations did not 
consider the burdens communities faced from regulations already in 
place. 

To address these concerns, GAO issued Federal Water Requirements: 
Challenges to Estimating the Cost Impact on Local Communities (GAO-06- 
151R) in November 2005. In this report, we noted that while EPA and 
others developed cost estimates for different regulatory programs, 
these estimates did not provide information on the cumulative costs of 
complying with federal water requirements--primarily because they were 
not intended to do so. Moreover, we found that several methodological 
challenges hinder efforts to develop reliable cumulative cost 
estimates, including obtaining accurate and complete cost data, 
particularly for older requirements; accurately allocating costs (e.g., 
among jurisdictions that share costs); and establishing a causal link 
between community investments and federal water requirements. 

In consideration of Senator Inhofe's and Senator Crapo's continued 
interest in the costs of EPA's storm water program, GAO agreed, in 
December 2005, to examine the implementation of Phases I and II of the 
program. In this context, GAO was asked to (1) identify the progress 
made in implementing the storm water program, (2) determine the extent 
to which the storm water program burdens communities, (3) evaluate the 
accuracy of EPA's cost estimates, and (4) examine the data available 
for future assessment of program burdens. 

To identify the progress made in implementing the storm water program, 
we collected and analyzed storm water permittee data for all 50 states 
between early March and early June 2006. Specifically, from the 3 EPA 
regions and 45 states with permitting authority, we obtained data on 
the number of Phase I and II municipal separate storm sewer systems 
(MS4s) that (1) were required to have a storm water permit, (2) 
obtained a permit, (3) received a waiver from permit coverage, or (4) 
had not obtained either a permit or a waiver.[Footnote 40] Moreover, we 
developed a detailed database of Phase I and II MS4s that included 
information on each MS4's: 

* name, 

* permit number, 

* date of initial permit coverage (by year), 

* type of entity (traditional local government or nontraditional 
entity),[Footnote 41] 

* type of permit (individual or general), 

* source of designation (the federal rule or as a result of additional 
designation criteria set by the EPA region or state), and: 

* co-permittee status (whether the MS4 established a co-permittee 
relationship with another MS4 or other entity to share responsibility 
for implementing permit requirements). 

To ensure a consistent, reliable, and verifiable approach to collecting 
these data, we developed and distributed a standardized data request. 
We sent the data request to the EPA regional and state officials that 
EPA identified as being in charge of storm water permitting for each 
state. Furthermore, to gain a better understanding of the reliability 
of the data we obtained, we conducted structured interviews with each 
of the permitting officials during which we asked about how they 
collected, processed, and maintained their storm water data. 
Specifically, we asked these officials to provide information on: 

* how their data are stored, 

* whether they assign MS4s a unique identifier for tracking purposes, 

* what sources they use to compile their data, 

* how often their data are updated and when this was last done, 

* reasons why their data might not be complete, 

* whether checks are performed on the accuracy of their data or to 
identify duplicate entries, and: 

* their confidence in the accuracy and completeness of their data. 

We also performed additional checks to ensure the reliability of the 
data we collected. First, to ensure the data provided by permitting 
authorities were complete, we compared summary statistics for each 
state with the individual permittee data we received. Second, we 
identified any errors or inconsistencies in the data by, for example, 
sorting the data to distinguish duplicate entries. In cases where we 
identified errors or inconsistencies, we resolved these issues during 
follow-up interviews conducted with the permitting officials. Third, we 
clarified which MS4s should be considered traditional or nontraditional 
MS4s. Fourth, we confirmed the dates of MS4s' initial permit coverage. 
Fifth, in cases where EPA or state officials reported that MS4s 
established co-permittee relationships, we verified, if possible, which 
MS4 was the lead permittee and how many co-permittees were associated 
with the permit. Based on these assessments, we determined that the 
data we collected were sufficiently reliable for the purposes of our 
audit work. 

As many EPA and state permitting authorities indicated that they were 
still developing and finalizing MS4s' initial permits around the time 
of our data collection effort, we conducted a second data collection 
effort between October and December 2006 to obtain updated data. During 
this second data collection effort, we obtained data on the number of 
Phase I and II MS4s that had received permit coverage or a waiver since 
the initial data collection effort. We also obtained data on the number 
of co-permittees associated with these newly issued permits. We did not 
obtain certain data, such as the permit number, type of permit, or 
source of designation for these MS4s. However, we checked all of the 
updated data we received for errors and inconsistencies, and conducted 
follow-up interviews to resolve any issues we identified. In technical 
comments provided on a draft copy of this report, EPA indicated that 
there are some differences between its storm water permittee data and 
the data we present in this report. For example, EPA said that some 
MS4s were required to obtain Phase I permits as a result of additional 
designations made by their permitting authority, but may not have been 
notified that they were required to obtain this coverage until years 
after the program began. We did not obtain information on the dates 
when MS4s were notified that they were required to obtain permit 
coverage; the data we present on the status of storm water program 
implementation include MS4s that received permit coverage in a given 
year, and MS4s that have not yet obtained permit coverage, according to 
EPA regional and state officials. In addition, EPA said that some 
additional MS4s received waivers from Phase I of the program. However, 
the data we obtained from EPA regional and state officials indicated 
that these MS4s were generally either permitted under Phase I, 
permitted under Phase II with no indication that they had been waived, 
or were not yet permitted or waived. 

In addition, to assess the status of program implementation we 
collected information on storm water program requirements and reasons 
why MS4s may not yet have permit coverage. Specifically, we examined 
the Clean Water Act and EPA's Phase I and II storm water regulations to 
gain an understanding of permitting processes and program requirements. 
Then, to identify the types of best management practices (BMPs) that 
MS4s can implement to meet program requirements, we reviewed EPA's 
guidance for states and MS4s, and assessed the likelihood that MS4s 
would implement certain BMPs through our interviews with EPA and state 
permitting authorities and other storm water experts. Finally, we 
examined relevant legal decisions concerning storm water permit 
processes and conditions. We also discussed the impact of these 
decisions, including whether they delayed permit issuance, during our 
interviews with storm water officials and experts. 

To determine the extent to which the storm water program has imposed a 
burden on communities, we initially considered estimating the ranges of 
costs for Phase I and II MS4s implementing the program. Using the 
database of permittees we developed through our 50-state data 
collection effort, we selected a random sample of 143 permitted Phase I 
and II MS4s that had been implementing their storm water permits for at 
least 1 year as of June 2006. The sample was designed to allow us to 
generalize any cost data we obtained to the universe of Phase I and II 
MS4s. To this sample, we added a judgmental selection of 7 MS4s that we 
believed had well-established storm water programs and good cost 
information based on our interviews with storm water experts and storm 
water management studies we reviewed. We contacted each of these 150 
MS4s and requested copies of their annual reports, storm water 
management plans, permit applications, and related documents. We 
developed a data collection instrument which we planned to use to 
analyze the information on storm water activities and costs that MS4s 
included in the documents we collected. We intended to supplement the 
cost data obtained through this effort with data on activity costs we 
obtained from our review of national, state, and local studies and 
publications on the costs of storm water BMPs. 

However, during the process of testing our data collection instrument 
we encountered a number of challenges. First, we found that MS4s' 
annual reports provided widely varying detail on their activities, 
which made it hard to compare the data we gathered from different MS4s' 
documents. Second, we found it difficult to determine which activities 
MS4s were conducting before the program. For example, we found evidence 
in MS4s' permit application documents that they were taking some steps 
to manage storm water runoff before the program began. However, in 
reviewing subsequent annual reports we had difficulty determining 
whether these activities had continued at the same level of effort, or, 
if the level of effort had increased, the data for estimating the cost 
impacts of this increase were not provided. Third, we found it 
challenging to identify MS4s' actual program costs due to a lack of 
specific information on (1) how costs were shared in instances where 
MS4s cooperated with other entities to implement program requirements, 
and (2) what efforts were taken because of storm water program 
requirements versus those that were taken because of MS4s' own 
initiative or the requirements of other environmental programs. 

As a result of these and other challenges associated with our initial 
methodology, we revised our approach to evaluating the extent to which 
implementing the storm water program has burdened MS4s by examining the 
factors that influence the extent of program burden instead of 
estimating ranges of program costs.[Footnote 42] To identify the 
factors that influence storm water program burdens and characterize 
which factors may increase or reduce these burdens, we interviewed EPA, 
state, and local officials, and almost 20 storm water experts from 
industry, academia, and relevant state and local government 
associations. Specifically, we interviewed EPA officials with the 
Office of Wastewater Management's Storm Water Program, the Office of 
Science and Technology's Effluents Guidelines Program, the Office of 
Research and Development's Wet-Weather Flow Research Program, and six 
regional offices.[Footnote 43] We asked state permitting officials 
about the factors that could influence program burdens during the 
follow-up interviews we conducted as part of our 50-state data 
collection effort. Further, we interviewed officials from Altamonte 
Springs, Florida; Austin, Texas; Denver, Colorado; and Portland, 
Oregon. Finally, the state and local government associations we 
contacted included the National Association of Flood and Stormwater 
Management Agencies, the Association of State and Interstate Water 
Pollution Control Administrators, and the Environmental Council of the 
States. 

In addition to interviewing regulatory officials and other experts, we 
obtained information on the factors that influence program burdens 
through an extensive literature review of national, state, and local 
storm water BMP studies. We identified roughly 55 studies related to 
storm water management practices based on our expert interviews, and 
internet and database searches. Due to concerns about their 
methodology, data sources, or limits in geographic area or BMP type, we 
excluded 33 of the studies. We then synthesized information from the 
remaining 22 studies based on factors that can influence storm water 
program burdens, such as the costs of implementing particular BMPs. We 
then incorporated this information into our analysis of statements from 
the experts we interviewed. 

As part of our effort to gather information on the extent to which 
program implementation has been a burden on MS4s, we also obtained data 
on federal funding for storm water projects. Specifically, we contacted 
EPA staff, including officials in the Office of Wastewater Management 
and the Office of the Chief Financial Officer, to collect information 
on the availability of federal funds for storm water activities, as 
well as data on the amount of funds provided. After confirming that 
wastewater treatment loans from the Clean Water State Revolving Fund 
(CWSRF) are the primary source of federal storm water funding, we 
conducted targeted follow-up interviews with nine states that we 
selected based on factors including the amount of CWSRF funding they 
used for storm water-related projects and their populations.[Footnote 
44] We asked these officials to provide information on their 
experiences using these loans to fund storm water projects and how they 
set funding priorities. We also obtained information from both EPA and 
state officials we interviewed on the potential barriers that may 
prevent MS4s from applying for or receiving these loans. To assess the 
reliability of the CWSRF funding data we analyzed, we contacted an EPA 
official responsible for managing the data. We obtained information 
describing the CWSRF program, its data management procedures, and the 
reliability of these data. Based on the official's responses, these 
data are sufficiently reliable for the purposes of our report. 

To assess the potential for future changes in program burdens, we asked 
about the factors that could influence future burdens during our 
interviews with regulatory officials and other storm water experts. 
From these interviews, we obtained information on the extent to which 
the burdens MS4s face may increase from more stringent or specific 
permit requirements, as well as increasing BMP maintenance and 
replacement costs. For information on the potential impact of increased 
storm water permit enforcement, we contacted officials from EPA's 
Office of Enforcement and Compliance Assurance. We also discussed this 
issue during our interviews with EPA regional officials and other storm 
water experts. 

To examine the accuracy of EPA's cost estimates, we analyzed the 
methodology and data used by the agency to estimate the costs for 
implementing Phases I and II of the storm water program. Specifically, 
we assessed EPA's use of hypothetical cities as the basis for its 
estimate of Phase I program costs. In addition, we obtained data on 
Phase I permit application costs by reviewing permit application 
documents we collected from the MS4s in our sample. Of the 57 Phase I 
MS4s in our sample of 150, we were able to identify Phase I permit 
application cost data in 10 of these MS4s' permit application 
documents. However, we were not able to identify cost data for both 
Parts I and II of the Phase I permit application for all of these MS4s. 
Consequently, we only reported data for the 6 MS4s for which we were 
able to identify total Phase I permit application costs, or costs for 
both Parts I and II of the application.[Footnote 45] We did not assess 
the reliability of these data. For Phase II, we examined the 
methodology and data EPA used for its analysis of Phase II program 
costs. We also obtained and analyzed the raw survey data collected by 
the National Association of Flood and Stormwater Management Agencies 
and used by EPA to develop the Phase II cost estimate. Finally, we 
interviewed knowledgeable EPA and state officials and other storm water 
experts to obtain their views on the accuracy of EPA's cost estimates, 
and the extent to which MS4s were implementing storm water management 
activities before the program. 

To summarize future challenges associated with evaluating the storm 
water program's burden, we conducted a limited review of the documents 
we received from our nationwide sample of MS4s. Of the 150 MS4s in our 
sample, we received at least one annual report from 130. We reviewed 
the most recent annual report for each of these 130 MS4s to obtain data 
on activities that we expected to be among the more commonly 
implemented, including public education BMPs, catch basin/storm drain 
cleaning, street-sweeping, and illicit discharge detection and 
elimination activities. Through this effort, we collected any cost data 
included in the document for the selected BMPs, either as total 
expenditures or unit costs. We also collected output information to the 
extent it was available, such as the number of miles swept or the 
number of catch basins cleaned. We planned to identify ranges of costs 
for these activities based on the data we found, supplemented with data 
from the storm water BMP studies we reviewed. 

However, we found limited and inconsistent data in MS4s' annual 
reports. As a result, with the exception of street-sweeping, we were 
unable to identify ranges of costs for these activities. The estimated 
range of Phase I street-sweeping costs we present in the report is 
based on data we identified in 10 MS4s' annual reports, as well as data 
for 10 Phase I MS4s that we obtained from other documentation we 
reviewed. The estimated range of Phase II street-sweeping costs is 
based on data from 14 MS4s' annual reports.[Footnote 46] We developed a 
per capita street-sweeping cost for each MS4 by dividing that MS4's 
total street-sweeping cost by its estimated population as reported by 
the U.S. Census Bureau for July 1, 2005. Some MS4s reported only output 
information (e.g., number of miles swept). For those MS4s, we obtained 
total street-sweeping costs using an estimated street-sweeping cost per 
mile that we calculated based on MS4s that reported both total cost and 
output information. We adjusted all cost data to 2006 dollars. 

This analysis has a number of limitations. First, the sample size is 
extremely small, which could impact the reliability of our estimated 
cost ranges. Second, the MS4s for which we were able to obtain street- 
sweeping data did not always report these data in consistent units. For 
instance, some reports included estimates of the number of lane miles 
swept, while others reported estimates of the number of curb miles 
swept, and still others simply reported estimates of the miles swept. 
While it might be reasonable to assume that curb and lane miles are 
approximately equal measures of distance, additional follow-up would be 
needed to confirm the data for MS4s that simply reported the number of 
miles swept. Third, we did not assess the reliability of the data 
included in these annual reports. 

We conducted our work between January 2006 and April 2007 in accordance 
with generally accepted government auditing standards. 

[End of section] 

Appendix III: Detailed Analysis of Storm Water Program Implementation 
Data: 

Table 3: Phase I Storm Water Program Implementation Data, by State as 
of Fall 2006: 

AK; 
Number of lead permittees[A]: 1; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

AL; 
Number of lead permittees[A]: 5; 
Number of co-permittees[B]: 41; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

AR; 
Number of lead permittees[A]: 1; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

AZ; 
Number of lead permittees[A]: 7; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

CA; 
Number of lead permittees[A]: 22[F]; 
Number of co-permittees[B]: 260[F]; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

CO; 
Number of lead permittees[A]: 4; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

CT; 
Number of lead permittees[A]: 1; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

DE; 
Number of lead permittees[A]: 1; 
Number of co-permittees[B]: 13[G]; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

FL; 
Number of lead permittees[A]: 27; 
Number of co-permittees[B]: 161; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

GA; 
Number of lead permittees[A]: 58; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

HI; 
Number of lead permittees[A]: 1; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

IA; 
Number of lead permittees[A]: 2; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 1; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

ID; 
Number of lead permittees[A]: 1; 
Number of co-permittees[B]: 3; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

IL; 
Number of lead permittees[A]: 1; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 2; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

IN; 
Number of lead permittees[A]: 1; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 1; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

KS; 
Number of lead permittees[A]: 3; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

KY; 
Number of lead permittees[A]: 2; 
Number of co-permittees[B]: 100; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

LA; 
Number of lead permittees[A]: 4; 
Number of co-permittees[B]: 8; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

MA; 
Number of lead permittees[A]: 2; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

MD; 
Number of lead permittees[A]: 10; 
Number of co-permittees[B]: 6; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

ME[E]; 
Number of lead permittees[A]: 0; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

MI; 
Number of lead permittees[A]: 5; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 3; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

MN; 
Number of lead permittees[A]: 2; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

MO; 
Number of lead permittees[A]: 3; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 1; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

MS; 
Number of lead permittees[A]: 1; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

MT[E]; 
Number of lead permittees[A]: 0; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

NC; 
Number of lead permittees[A]: 6; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

ND[E]; 
Number of lead permittees[A]: 0; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

NE; 
Number of lead permittees[A]: 2; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

NH[E]; 
Number of lead permittees[A]: 0; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

NJ; 
Number of lead permittees[A]: 0; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 4; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

NM; 
Number of lead permittees[A]: 1; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

NV; 
Number of lead permittees[A]: 2; 
Number of co-permittees[B]: 6; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

NY; 
Number of lead permittees[A]: 14; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 5; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

OH; 
Number of lead permittees[A]: 4; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 2; 
Number of MS4s that have not received a permit or a waiver[D]: 1. 

OK; 
Number of lead permittees[A]: 2; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

OR; 
Number of lead permittees[A]: 6; 
Number of co-permittees[B]: 12[F]; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

PA; 
Number of lead permittees[A]: 2; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 2; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

RI[E]; 
Number of lead permittees[A]: 0; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

SC; 
Number of lead permittees[A]: 2; 
Number of co-permittees[B]: 2; 
Number of MS4s that received waivers[C]: 1; 
Number of MS4s that have not received a permit or a waiver[D]: 1. 

SD; 
Number of lead permittees[A]: 1; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

TN; 
Number of lead permittees[A]: 4; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

TX; 
Number of lead permittees[A]: 19; 
Number of co-permittees[B]: 1; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

UT; 
Number of lead permittees[A]: 2; 
Number of co-permittees[B]: 14; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

VA; 
Number of lead permittees[A]: 11; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 1; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

VT[E]; 
Number of lead permittees[A]: 0; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

WA; 
Number of lead permittees[A]: 6[F]; 
Number of co-permittees[B]: 0[F]; 
Number of MS4s that received waivers[C]: 2; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

WI; 
Number of lead permittees[A]: 46[H]; 
Number of co-permittees[B]: 17[H]; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 9. 

WV[E]; 
Number of lead permittees[A]: 0; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

WY[E]; 
Number of lead permittees[A]: 0; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

Total; 
Number of lead permittees[A]: 295; 
Number of co-permittees[B]: 644; 
Number of MS4s that received waivers[C]: 25; 
Number of MS4s that have not received a permit or a waiver[D]: 11. 

Source: GAO analysis of EPA and state data. 

Note: The Phase I statistics include only traditional local government 
entities, such as counties, cities, towns, and boroughs. Nontraditional 
Phase I MS4s, such as universities, military bases, state departments 
of transportation, and Indian Nations were excluded. Also, we did not 
assess whether all of the potentially regulated communities listed in 
EPA's Phase I regulations received permit coverage. 

[A] Lead permittees include MS4s reported by EPA and state permitting 
authorities as being a co-permittee with one or more MS4s and having 
lead responsibility for coordinating or implementing activities for 
those co-permittees. Some MS4s cooperate with other MS4s without 
becoming co-permittees; however, the number of MS4s with these 
cooperative relationships are not represented in the table. 

[B] Co-permittees do not include MS4s counted as lead permittees. These 
data do include some Phase II MS4s that became co-permittees with Phase 
I MS4s prior to, or after the Phase II regulations became effective. 
For example, the number of Phase I co-permittees in Kentucky includes 
94 MS4s that are co-permittees with one lead Phase I MS4. A state 
official said these MS4s would have been required to obtain Phase II 
permit coverage had they not obtained coverage under Phase I. We 
counted such MS4s as Phase I co-permittees because they are 
implementing the program under the conditions of a Phase I permit. 

[C] Waived MS4s include those issued official waivers from Phase I 
permit requirements by permitting authorities because they met certain 
conditions, such as having a combined sewer system that served enough 
of their population to bring them below the Phase I population 
threshold. Many of these MS4s were subsequently permitted under Phase 
II of the storm water program, and we included them in the data 
provided in table 4. In some cases, Phase I MS4s were exempted, but not 
officially waived, from Phase I requirements. For consistency, we 
counted these MS4s as being waived. 

[D] South Carolina and Wisconsin Phase I MS4s that had not obtained 
either a permit or a waiver include MS4s for whom these states were in 
the process of developing final permits. In the case of the one Ohio 
Phase I MS4 that had not obtained either a permit or a waiver, the MS4 
has a combined sewer system, and according to an Ohio official, the 
state and the MS4 had not reached agreement on which type of permit the 
MS4 should be required to obtain. 

[E] These states did not have any MS4s that were required to obtain 
Phase I permits. 

[F] In these states, some MS4s are covered under multiple permits: 

˛ In California, three MS4s are the lead permittees for multiple 
permits; however, we counted each of these permittees only once in 
table 3. Also, one MS4 received a Phase I permit and later obtained 
coverage under a Phase II permit, while another MS4 is the lead 
permittee under both a Phase I and a Phase II permit. We included these 
MS4s in table 3 rather than table 4 because they first received 
coverage under a Phase I permit. In addition, four MS4s received 
coverage as both Phase I co-permittees and Phase II lead permittees. We 
included the data for these four MS4s in table 4 since that is the 
phase of the program under which they are lead permittees. 

˛ One Oregon MS4 has received coverage as a co-permittee under two 
different permits; however, we counted this MS4 only once in table 3. 

˛ Two of Washington's Phase I MS4s are lead permittees that received 
coverage under multiple permits; however, these permittees are counted 
only once. Also, one Washington MS4 received coverage as both a lead 
permittee, and a co-permittee with a different MS4. We counted this MS4 
only once in table 3. 

[G] Delaware's Phase I co-permittees include four MS4s that volunteered 
to implement the program, but which were never officially required to 
obtain permit coverage under Phase I or II. 

[H] A state official indicated some Wisconsin MS4s have established co- 
permittee relationships, but there is no lead permittee. In these 
instances, we counted all of the co-permittees as lead permittees. 

[End of table] 

Table 4: Phase II Storm Water Program Implementation Data, by State as 
of Fall 2006: 

AK; 
Number of lead permittees[A]: 2; 
Number of co-permittees[B]: 1; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

AL; 
Number of lead permittees[A]: 27; 
Number of co-permittees[B]: 4; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 1. 

AR; 
Number of lead permittees[A]: 39; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: N/A[F]; 
Number of MS4s that have not received a permit or a waiver[D]: 8[E]. 

AZ; 
Number of lead permittees[A]: 32; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 1; 
Number of MS4s that have not received a permit or a waiver[D]: 1. 

CA; 
Number of lead permittees[A]: 77[H,I]; 
Number of co-permittees[B]: 21[H]; 
Number of MS4s that received waivers[C]: 8; 
Number of MS4s that have not received a permit or a waiver[D]: 61. 

CO; 
Number of lead permittees[A]: 53; 
Number of co-permittees[B]: 1; 
Number of MS4s that received waivers[C]: 7; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

CT; 
Number of lead permittees[A]: 113[I]; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 19; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

DE; 
Number of lead permittees[A]: 2; 
Number of co-permittees[B]: 1; 
Number of MS4s that received waivers[C]: 2; 
Number of MS4s that have not received a permit or a waiver[D]: 1. 

FL; 
Number of lead permittees[A]: 94; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

GA; 
Number of lead permittees[A]: 84; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

HI; 
Number of lead permittees[A]: 0; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 2. 

IA; 
Number of lead permittees[A]: 41; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 13; 
Number of MS4s that have not received a permit or a waiver[D]: 9. 

ID; 
Number of lead permittees[A]: 0; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 28[E]. 

IL; 
Number of lead permittees[A]: 426; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 195; 
Number of MS4s that have not received a permit or a waiver[D]: 18. 

IN; 
Number of lead permittees[A]: 125; 
Number of co-permittees[B]: 18; 
Number of MS4s that received waivers[C]: 20; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

KS; 
Number of lead permittees[A]: 52; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 27; 
Number of MS4s that have not received a permit or a waiver[D]: 1. 

KY; 
Number of lead permittees[A]: 43; 
Number of co-permittees[B]: 54[J]; 
Number of MS4s that received waivers[C]: 1; 
Number of MS4s that have not received a permit or a waiver[D]: 1. 

LA; 
Number of lead permittees[A]: 39; 
Number of co-permittees[B]: 10; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

MA; 
Number of lead permittees[A]: 237; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 13; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

MD; 
Number of lead permittees[A]: 52; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 15; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

ME; 
Number of lead permittees[A]: 28; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 4; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

MI; 
Number of lead permittees[A]: 355[G,I]; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 1. 

MN; 
Number of lead permittees[A]: 168; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 1; 
Number of MS4s that have not received a permit or a waiver[D]: 44. 

MO; 
Number of lead permittees[A]: 64; 
Number of co-permittees[B]: 67; 
Number of MS4s that received waivers[C]: N/A[F]; 
Number of MS4s that have not received a permit or a waiver[D]: 24[E]. 

MS; 
Number of lead permittees[A]: 32; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

MT; 
Number of lead permittees[A]: 8; 
Number of co-permittees[B]: 2; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

NC; 
Number of lead permittees[A]: 81; 
Number of co-permittees[B]: 12; 
Number of MS4s that received waivers[C]: 40; 
Number of MS4s that have not received a permit or a waiver[D]: 20. 

ND; 
Number of lead permittees[A]: 14; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 4; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

NE; 
Number of lead permittees[A]: 19; 
Number of co-permittees[B]: 2; 
Number of MS4s that received waivers[C]: N/A[F]; 
Number of MS4s that have not received a permit or a waiver[D]: 1. 

NH; 
Number of lead permittees[A]: 38; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 7; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

NJ; 
Number of lead permittees[A]: 462[K]; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 31[K]; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

NM; 
Number of lead permittees[A]: 0; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 28[E]. 

NV; 
Number of lead permittees[A]: 5; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

NY; 
Number of lead permittees[A]: 448; 
Number of co-permittees[B]: 2; 
Number of MS4s that received waivers[C]: 17; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

OH; 
Number of lead permittees[A]: 252; 
Number of co-permittees[B]: 231; 
Number of MS4s that received waivers[C]: 42; 
Number of MS4s that have not received a permit or a waiver[D]: 14. 

OK; 
Number of lead permittees[A]: 38; 
Number of co-permittees[B]: 1; 
Number of MS4s that received waivers[C]: 14; 
Number of MS4s that have not received a permit or a waiver[D]: 1. 

OR; 
Number of lead permittees[A]: 0; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 8; 
Number of MS4s that have not received a permit or a waiver[D]: 18. 

PA; 
Number of lead permittees[A]: 722; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 216; 
Number of MS4s that have not received a permit or a waiver[D]: 4. 

RI; 
Number of lead permittees[A]: 32; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 1. 

SC; 
Number of lead permittees[A]: 0; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 4; 
Number of MS4s that have not received a permit or a waiver[D]: 64. 

SD; 
Number of lead permittees[A]: 13[G]; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

TN; 
Number of lead permittees[A]: 72; 
Number of co-permittees[B]: 10; 
Number of MS4s that received waivers[C]: 2; 
Number of MS4s that have not received a permit or a waiver[D]: 2. 

TX; 
Number of lead permittees[A]: 0; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 300[E]. 

UT; 
Number of lead permittees[A]: 56; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 3; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

VA; 
Number of lead permittees[A]: 39; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 5; 
Number of MS4s that have not received a permit or a waiver[D]: 3. 

VT; 
Number of lead permittees[A]: 9; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 2; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

WA; 
Number of lead permittees[A]: 0; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 111[E]. 

WI; 
Number of lead permittees[A]: 102; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 28; 
Number of MS4s that have not received a permit or a waiver[D]: 36. 

WV; 
Number of lead permittees[A]: 28; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 9; 
Number of MS4s that have not received a permit or a waiver[D]: 0. 

WY; 
Number of lead permittees[A]: 0; 
Number of co-permittees[B]: 0; 
Number of MS4s that received waivers[C]: 0; 
Number of MS4s that have not received a permit or a waiver[D]: 6. 

Total; 
Number of lead permittees[A]: 4623; 
Number of co-permittees[B]: 437; 
Number of MS4s that received waivers[C]: 758; 
Number of MS4s that have not received a permit or a waiver[D]: 809. 

Source: GAO analysis of EPA and state data. 

Note: The Phase II statistics include only traditional local government 
entities, such as cities, towns, and villages. Nontraditional Phase II 
MS4s, such as universities, military bases, state departments of 
transportation, and Indian Nations were excluded. Also, we did not 
assess whether all of the potentially regulated communities listed in 
EPA's Phase II regulations received permit coverage. 

[A] Lead permittees include MS4s reported by EPA and state permitting 
authorities as being a co-permittee with one or more MS4s and having 
lead responsibility for coordinating or implementing activities for 
those co-permittees. Some MS4s cooperate with other MS4s without 
becoming co-permittees; however, with some exceptions, the number of 
MS4s with these cooperative relationships are not represented in the 
table. 

[B] Co-permittees do not include MS4s counted as lead permittees. 

[C] Waived MS4s include those issued official waivers from Phase II 
permit requirements by EPA and state permitting authorities because 
they met certain conditions, such as they did not own or operate any 
roads or drains that discharged into local waterways. In some cases, 
Phase II MS4s were exempted, but not officially waived, from Phase II 
requirements. For consistency, we counted these MS4s as being waived. 

[D] Phase II MS4s that have not obtained either a permit or a waiver 
include MS4s whose permit authorities have not yet (1) determined 
whether the MS4s need to obtain permit coverage, or (2) issued a final 
permit. 

[E] For these states, EPA and state permitting authorities could not 
provide the exact number of Phase II MS4s that may ultimately be 
required to obtain permit coverage. The totals shown are estimates 
provided by these officials. 

[F] Officials in these states reported that they had waived or exempted 
some Phase II MS4s, but could not provide an estimate of how many. 

[G] Michigan and South Dakota officials reported four MS4s and two 
MS4s, respectively, that volunteered to obtain permit coverage. 

[H] A state official indicated some California MS4s have established co-
permittee relationships, but there is no lead permittee. In these 
instances, we counted all of the co-permittees as lead permittees. 

[I] In these states, some MS4s are covered under multiple permits: 

˛ One California MS4 received coverage under the state's Phase II 
general permit twice; however, we counted this MS4 only once in table 
4. Also, one MS4 received a Phase I permit and later obtained coverage 
under a Phase II permit, while another MS4 is the lead permittee under 
both a Phase I and a Phase II permit. We included these MS4s in table 3 
rather than table 4 because they first received coverage under a Phase 
I permit. In addition, four MS4s received coverage as both Phase I co- 
permittees and Phase II lead permittees. We included the data for these 
four MS4s in table 4 since that is the phase of the program under which 
they are lead permittees. 

˛ One Connecticut MS4 received two Phase II permits; however, we 
counted this MS4 only once in table 4. 

˛ Thirteen Michigan MS4s received multiple Phase II permits; however, 
we counted these MS4s only once in table 4. 

[J] The number of co-permittees in Kentucky includes five 
unincorporated communities, which according to a state official, have 
not received either a permit or a waiver. These MS4s' storm water 
management responsibilities are being covered by their respective 
counties. Consequently, although these MS4s have not entered into co- 
permittee relationships, Kentucky considers them to be in compliance 
with program requirements. Therefore, we included these MS4s among 
those that have received coverage as co-permittees. 

[K] New Jersey issued two different types of storm water permits for 
traditional MS4s. One type of permit meets federal storm water 
requirements, while the other meets the requirements of the state's 
pollution prevention permitting system. The data we present for New 
Jersey include only MS4s covered under the permits meeting federal 
requirements. Regarding the number of waivers issued by New Jersey, 
some of the MS4s were waived from meeting federal storm water 
requirements, but were covered under permits issued under the state 
program. We did not obtain data on the number of MS4s for which this 
occurred. 

[End of table] 

[End of section] 

Appendix IV: Comments from EPA: 

United States Environmental Protection Agency: 
OFFICE OF WATER: 
Washington, D.C. 20460: 

May 2 222007: 

Mr. John B. Stephenson, Director:
Natural Resources and Environment: 
United States Government Accountability Office: 
Washington, DC 20548: 

Dear Mr. Stephenson: 

Thank you for the opportunity to review your draft report entitled 
"Further Implementation and Better Cost Data Needed to Determine Impact 
of EPA's Storm Water Program on Communities." The U.S. Environmental 
Protection Agency (EPA) believes that stormwater runoff is a 
significant threat to water quality and is actively working with our 
State and municipal partners to control stormwater discharges to 
protect and improve water quality. Although the stormwater program has 
been in place since the early 1990s, the complexity of the water 
quality problem and the myriad of stakeholders who must contribute to 
its successful implementation make it imperative that an iterative 
approach be employed. Your report concludes that: 

* It too early to determine if program implementation has been a 
burden: 

* It is difficult to develop a uniform assessment of the burden of the 
program nationally. 

* A community's ability to establish reliable sources of funding 
influences whether the program is a burden: 

* More complete and consistent reporting on the scope, costs, and 
results of communities' stormwater best management practices is needed 
to accurately assess burden. 

Your report recommends that EPA: 

"issue additional program guidance and consider regulatory changes to 
ensure that (1) communities report on activities in sufficient detail 
to determine their scope, costs, and results, and (2) communities 
report this information consistently so that it can be analyzed on a 
national basis." 

EPA has already taken steps to address your recommendation. Existing 
stormwater regulations establish annual reporting requirements for 
communities, at 40 CFR § 122.42(c) for medium and large communities 
(i.e., Phase I) and §122.34(g)(3) for smaller communities (i.e., Phase 
II). Also, § 122.37 specifies that EPA will evaluate the stormwater 
regulations for Phase II communities after December 10, 2012, and make 
any necessary revisions. EPA acknowledges the importance of obtaining 
robust data on which to base decisions. The framework of EPA's 
stormwater permitting program for local communities, however, does not 
lend itself to reporting data in a form that can be easily compiled and 
compared with other communities. 

The stormwater program provides local flexibility to implement programs 
tailored to best address the varying array of complex environmental 
issues associated with stormwater runoff in different locations. EPA 
expects local stormwater program activities to reflect local problems 
and locally derived solutions. In addition, many communities have been 
implementing stormwater control practices for many years, well before 
the advent of EPA's stormwater program. Separating burden associated 
with EPA's stormwater program from that incurred irrespective of the 
program has proven challenging. Accounting for stormwater control 
activities in all the different departments and divisions within these 
local governments is quite complex. As a result, gauging the true 
burden associated with EPA's stormwater program has been and will 
continue to be a challenge. 

EPA does recognize the importance of being able to assess stormwater 
program performance and sees the annual report as one of the best ways 
to gather information necessary to perform this evaluation. While EPA 
does not believe it is appropriate to modify the annual reporting 
regulations at this time, EPA initiated an effort to identify necessary 
information to be able to track and assess local stormwater program 
performance for both environmental effectiveness and compliance. The 
purpose is to identify the information that communities should submit 
in their annual reports and then to develop corresponding guidance for 
EPA Regions, states, and local communities. 

While the actual reporting format has not been determined, EPA supports 
developing an annual report template or form for use by communities to 
provide a consistent level of program information. Based on GAO's 
recommendation, EPA will investigate ways to incorporate better cost 
information into the annual report. 

Enclosed are additional comments that we provided on the draft report 
that addressed the most substantive issues. Again, we appreciate the 
opportunity to review and comment on the draft report. If you have any 
questions about these comments or would like to discuss our national 
stormwater program, please contact me or call Linda Boornazian, 
Director of the Water Permits Division, at (202) 564-9545. 

Sincerely, 

Signed by:  

Benjamin H. Grumbles: 
Assistant Administrator: 

Enclosure: 

[End of section] 

Appendix V: Key Contributors to This Report: 

Contact: 

John B. Stephenson, (202) 512-3841: 

Acknowledgments: 

In addition to the individual named above, Diane B. Raynes, Assistant 
Director; Tim Bazzle; Kevin Bray; Mark Braza; Ellen M. Crocker; 
Kathleen M. Drennan; Brian Hanley; Benjamin Howe; Krista Loose; 
Christopher Murray; Mehrzad Nadji; Ellen Phelps Ranen; Carol Herrnstadt 
Shulman; and Winnie Tsen made key contributions to this report. 

[End of section] 

Related GAO Products: 

Clean Water: How States Allocate Revolving Loan Funds and Measure Their 
Benefits. GAO-06-579. Washington, D.C.: June 5, 2006. 

Federal Water Requirements: Challenges to Estimating the Cost Impact on 
Local Communities. GAO-06-151R. Washington, D.C.: November 30, 2005. 

Storm Water Pollution: Information Needed on the Implications of 
Permitting Oil and Gas Construction Activities. GAO-05-240. Washington, 
D.C.: February 9, 2005. 

Water Quality: Better Data and Evaluation of Urban Runoff Programs 
Needed to Assess Effectiveness. GAO-01-679. Washington, D.C.: June 29, 
2001. 

FOOTNOTES 

[1] EPA's storm water regulations apply to communities with separate 
storm water and sanitary sewer systems. In some communities, 
particularly older communities, the sanitary and storm sewer systems 
are combined. Combined sewer systems are regulated under different 
NPDES permits. 

[2] EPA's Phase II regulations also listed 590 communities outside 
urban areas that met certain population criteria and could potentially 
be required to obtain permit coverage. Furthermore, the regulations 
listed a number of communities in Puerto Rico as potential Phase II 
communities. 

[3] We limited our data collection effort to communities in the 50 
states. As a result, we did not include the District of Columbia or 
communities in the U.S. territories in our analysis. 

[4] Section 402(a) of the Clean Water Act. 33 U.S.C. §1342(a). 

[5] Section 402(p) of the Clean Water Act. 33 U.S.C. §1342(p). 

[6] While we generally use "MS4" to refer to the local government 
implementing the program, an MS4 is technically a system of storm water 
conveyances (including roads with drainage systems and municipal 
streets) that is owned or operated by a state, city, town, borough, 
county, parish, district, association, or other public body and that is 
not a combined sewer or part of a publicly owned treatment works. 

[7] Phases I and II of the storm water program also included 
requirements relating to storm water runoff from industrial facilities 
and construction sites; however, we did not examine the implementation 
of these requirements as part of our work. 

[8] EPA, Report to Congress on the Phase I Storm Water Regulations, EPA 
833-R-00-001 (Washington, D.C.: February 2000). 

[9] Five states (Alaska, Idaho, Massachusetts, New Hampshire, and New 
Mexico) do not have authority to issue storm water permits, and in 
these states EPA administers the storm water program. Arizona may lose 
NPDES permitting authority as a result of pending litigation. 

[10] EPA and state permitting authorities are also authorized to waive 
MS4s from permit requirements if they meet certain conditions. 

[11] Permit authorities can deny coverage under a general permit and 
require an MS4 to obtain an individual permit for its storm water 
discharges. 

[12] In addition to the individual and general permitting processes 
outlined in figure 1, EPA must have an opportunity to review each 
permit issued and the agency may object to elements that conflict with 
federal requirements. If the permit authority does not address these 
concerns, EPA may issue the permit itself. 

[13] EPA's regulations for Phases I and II of the storm water program 
were set up differently. Specifically, the Phase II regulations 
directed MS4s to develop storm water management programs that included 
actions in six BMP areas, while the Phase I regulations did not 
specifically mention all six areas. However, in practice, many of the 
BMPs that MS4s are implementing under Phase I or II are similar. 

[14] 33 U.S.C. §1313(d). Territories and authorized tribes are also 
required to establish TMDLs where necessary. TMDL requirements 
specifically apply to waterbodies that have been designated as 
impaired, or not meeting water quality standards. 

[15] Endangered Species Act of 1973, Pub. L. No. 93-205, §9(a), 87 
Stat. 884, 893 (codified at 16 U.S.C. §1538(a)). 

[16] EPA's regulations for Phase II required that MS4s submit reports 
annually during the first permit cycle, and then in years 2 and 4 of 
subsequent permit cycles, unless otherwise specified by the permitting 
authority. 

[17] Through the CWSRF, EPA provides annual grants to the states to 
capitalize state-level CWSRFs. States must match these EPA grants with 
a minimum of 20 percent of their own contributions. States loan their 
CWSRF dollars to local governments and other entities for various water 
quality projects, and loan repayments are cycled back into the state- 
level programs to fund additional projects. Since 1987, states have 
used 96 percent of their CWSRF dollars to build, upgrade, or enlarge 
conventional wastewater treatment facilities and conveyances. 

[18] Prior to the implementation of Phases I and II of the storm water 
program, storm water control projects in Phase I and II MS4s could have 
been funded using CWSRF loans; however, according to EPA, these loans 
would have been tracked as nonpoint source (not wastewater) 
expenditures. Through June 2006, EPA data indicated that $2.4 billion 
in CWSRF funds had been spent on nonpoint source projects, such as 
pollution control from sanitary landfills. Because, prior to the storm 
water program, nonpoint source loans were provided for a variety of 
purposes in addition to storm water projects, EPA could not determine 
the extent to which these loans were provided for projects that would 
now be part of the storm water program. 

[19] As of fall 2006, Washington had not yet issued its Phase II 
general permit. However, on January 17, 2007, the Washington Department 
of Ecology issued two Phase II permits, one for eastern Washington 
(covering 20 cities and 8 counties), and one for western Washington 
(covering at least 80 cities and 5 counties). These permits became 
effective on February 16, 2007. 

[20] EPA Region 6 issued a general permit for Phase II MS4s in New 
Mexico on September 29, 2006; however, this permit did not become 
effective until January 1, 2007. Phase II MS4s in New Mexico had until 
April 1, 2007, to submit their notices of intent for coverage under 
this permit. 

[21] EPA's Phase II regulations allowed permit authorities to phase in 
permit coverage for MS4s with a population under 10,000. Under this 
option, these MS4s were required to have permit coverage no later than 
March 8, 2007. Permit authorities were required to obtain EPA approval 
to exercise this option; however, we did not obtain information on how 
many permit authorities used this approach. 

[22] 344 F.3d 832 (9TH Cir. 2003). 

[23] Since fall 2006, EPA Region 10 issued permits to 3 Idaho Phase II 
MS4s, leaving 25 MS4s without permits. 

[24] As of the end of fiscal year 2006, federal allocations to state 
CWSRF grants declined by about another $181 million. In 2007, 
congressional appropriations for CWSRF grants increased by about $183 
million to over $1.08 billion. However, this increase was largely due 
to the inclusion of formerly earmarked funds. Comparable data on fiscal 
year 2007 federal allocations to state CWSRF grants were not available 
at the time of our report. 

[25] EPA said that limited resources impacted the amount of outreach it 
could conduct on the eligibility of storm water projects for CWSRF 
loans. However, the agency also noted that, due to the early stage of 
implementation of Phase II of the storm water program, it is not yet 
clear whether there will be a significant demand from Phase II MS4s for 
CWSRF loans to help fund storm water projects. 

[26] However, officials in a few other states we contacted said that 
the federal allocations to their funds make up such a small percentage 
of their overall funds that recent reductions in the federal 
allocations have had limited impact. 

[27] Data for the characteristics of these 8 hypothetical cities were 
taken from 18 actual Phase I cities, including Birmingham, Alabama; 
Cleveland, Ohio; Des Moines, Iowa; Detroit, Michigan; Durham, North 
Carolina; Hialeah, Florida; Honolulu, Hawaii; Lexington, Kentucky; Los 
Angeles, California; Nashville, Tennessee; New Orleans, Louisiana; 
Oklahoma City, Oklahoma; Pittsburgh, Pennsylvania; Providence, Rhode 
Island; St. Petersburg, Florida; Salt Lake City, Utah; South Bend, 
Indiana; and Tempe, Arizona. 

[28] EPA decided not to include permit application costs in its 
analysis--reasoning that these costs would be incurred before storm 
water permits are issued, and should therefore not be considered as 
costs of complying with permit requirements. 

[29] These 6 were the only MS4s, out of the 57 Phase I MS4s we 
contacted, for which we could identify data on the costs of their Part 
I and II permit applications from the documents they provided. These 
estimates represent costs reported between 1991 and 1994. Dollar values 
have not been adjusted for inflation. 

[30] The Phase II analysis did not indicate whether NAFSMA selected the 
1,600 MS4s it surveyed randomly or by some other method. 

[31] EPA estimated the total population affected by Phase II would be 
85 million on the basis of the 5,040 MS4s it identified as potentially 
regulated under the Phase II rule. 

[32] To check these results, EPA analyzed cost data from the annual 
reports of 26 Phase I MS4s to obtain information on incremental Phase 
II costs. EPA selected these MS4s because they had been implementing 
the Phase I program for one permit term, were smaller cities that 
closely reflected the population of Phase II MS4s, and had detailed 
data reflecting actual implementation costs for program elements 
similar to Phase II. While this analysis yielded results similar to 
EPA's analysis of the NAFSMA survey data, we have concerns about the 
sufficiency and comparability of data reported by MS4s in their annual 
reports on the basis of our attempts to conduct a similar type of 
analysis. 

[33] Survey response rate is generally considered to be one indication 
of the quality of survey data because a survey with a low response rate 
may report only extreme views--potentially biasing the results. For 
this reason, current Office of Management and Budget guidance suggests 
that agencies aim for a response rate of 80 percent or higher. 

[34] EPA reported that the per household costs for these 56 MS4s ranged 
from $0.42 to $54.91. We adjusted these costs to arrive at per capita 
costs in 2006 dollars. 

[35] We intended to use this output information, in conjunction with 
cost data obtained from other MS4s and studies, to estimate costs for 
MS4s that did not report detailed cost data for the selected BMPs. 

[36] After collecting data on which MS4s nationwide have received storm 
water permits, we took a sample of 150 MS4s and requested copies of 
annual reports, among other program documents. Sampled MS4s were 
selected largely at random; however, we added 7 MS4s that could serve 
as potential case study locations because, on the basis of our 
discussions with storm water experts, we believed they had well- 
established storm water programs and could provide good cost data. Of 
the 150 MS4s for which we requested program documents, we received at 
least one annual report from 130. 

[37] These estimates are presented in 2006 dollars. 

[38] While it might be reasonable to assume that curb and lane miles 
are approximately equal measures of distance, additional follow-up 
would be needed to confirm the data for MS4s that simply reported the 
number of miles swept. 

[39] In a June 2001 report on storm water runoff, we also found that 
any cost information reported by MS4s would be difficult to analyze 
unless EPA and the states set guidelines to elicit better and more 
standardized data. As a result, we recommended that EPA, among other 
things, establish guidelines for obtaining consistent and reliable data 
from local governments with Phase I permits, including data on the 
effects of the program and the costs to these governments. GAO, Water 
Quality: Better Data and Evaluation of Urban Runoff Programs Needed to 
Assess Effectiveness, GAO-01-679 (Washington, D.C.: June 29, 2001). 

[40] We limited our data collection effort to MS4s in the 50 states. 
For the states of Massachusetts and New Hampshire, data were provided 
by an EPA Region 1 official; New Mexico data were provided by EPA 
Region 6 officials; and Alaska and Idaho data were provided by an EPA 
Region 10 official. We did not include the District of Columbia or MS4s 
in the U.S. territories in our analysis. 

[41] Nontraditional MS4s include prisons, universities, military bases, 
or other entities that are not considered traditional local 
governments. These entities, and Indian Nations, were excluded from our 
review. 

[42] For purposes of this report, we use burden to mean additional 
costs for implementing storm water control measures, increased 
administrative activities, reduced budget flexibility because of the 
need to divert resources from other governmental activities, actions 
related to litigation, and the influences of other regulatory programs 
on how storm water runoff is managed. 

[43] Both the Office of Wastewater Management and the Office of Science 
and Technology are located in EPA's Office of Water. 

[44] We based this analysis on loans made under authority provided by 
the Clean Water Act for funding of wastewater treatment projects. 
According to an official with EPA's State Revolving Fund Branch, the 
majority of CWSRF loans made for storm water projects would be made 
under this authority. 

[45] We did not adjust these data to 2006 dollars. 

[46] For one Phase II MS4, we excluded street-sweeping data included in 
its annual report, because upon calculation of per capita cost, the 
data appeared to be a significant outlier. We tried to contact the MS4 
to confirm its data, but we were unsuccessful. 

GAO's Mission: 

The Government Accountability Office, the audit, evaluation and 
investigative arm of Congress, exists to support Congress in meeting 
its constitutional responsibilities and to help improve the performance 
and accountability of the federal government for the American people. 
GAO examines the use of public funds; evaluates federal programs and 
policies; and provides analyses, recommendations, and other assistance 
to help Congress make informed oversight, policy, and funding 
decisions. GAO's commitment to good government is reflected in its core 
values of accountability, integrity, and reliability. 

Obtaining Copies of GAO Reports and Testimony: 

The fastest and easiest way to obtain copies of GAO documents at no 
cost is through GAO's Web site (www.gao.gov). Each weekday, GAO posts 
newly released reports, testimony, and correspondence on its Web site. 
To have GAO e-mail you a list of newly posted products every afternoon, 
go to www.gao.gov and select "Subscribe to Updates." 

Order by Mail or Phone: 

The first copy of each printed report is free. Additional copies are $2 
each. A check or money order should be made out to the Superintendent 
of Documents. GAO also accepts VISA and Mastercard. Orders for 100 or 
more copies mailed to a single address are discounted 25 percent. 
Orders should be sent to: 

U.S. Government Accountability Office 441 G Street NW, Room LM 
Washington, D.C. 20548: 

To order by Phone: Voice: (202) 512-6000 TDD: (202) 512-2537 Fax: (202) 
512-6061: 

To Report Fraud, Waste, and Abuse in Federal Programs: 

Contact: 

Web site: www.gao.gov/fraudnet/fraudnet.htm E-mail: fraudnet@gao.gov 
Automated answering system: (800) 424-5454 or (202) 512-7470: 

Congressional Relations: 

Gloria Jarmon, Managing Director, JarmonG@gao.gov (202) 512-4400 U.S. 
Government Accountability Office, 441 G Street NW, Room 7125 
Washington, D.C. 20548: 

Public Affairs: 

Paul Anderson, Managing Director, AndersonP1@gao.gov (202) 512-4800 
U.S. Government Accountability Office, 441 G Street NW, Room 7149 
Washington, D.C. 20548: