


                           Economic Analysis for the
                     Public Notification Requirements for
     Combined Sewer Overflow (CSO) discharges within the Great Lakes Basin





                                     Draft
                               December 1, 2016




                                 Prepared for
                 United States Environmental Protection Agency
                        Office of Wastewater Management
                         1200 Pennsylvania Avenue, NW
                             Washington, DC 20460








                        EPA Contract Number EP-C-11-009
                        EPA Work Assignment Number 5-05

Contents
1	Introduction	1
2	Executive Order 13132: Federalism	1
2.1	Methods	2
2.1.1	Cost Data	2
2.1.2	Revenue Data	3
2.1.3	Federalism 1% Test	4
2.2	Results	5
3	Small Entity Screening Analysis	5
3.1	Qualitative analysis	6
3.2	Quantitative analysis	7
4	Key Findings	8
5	Uncertainties and Limitations	8
6	References	8
Appendix 1. Cash Flows	A-1

	
 Introduction
This document (the Economic Analysis or EA) provides analytical support to the U.S. Environmental Protection Agency's (EPA's) proposed rule Public Notification requirements for Combined Sewer Overflow (CSO) discharges within the Great Lakes Basin (hereafter referred to as the proposed rule). The intention of the proposed rule is to establish public notification requirements for CSOs to the Great Lakes, as required by Section 425 of the Consolidated Appropriations Act (Section 425).
EPA conducted an economic analysis (EA) to fulfill the requirements of Executive Order (EO) 13132 on Federalism and Regulatory Flexibility Act (RFA) as amended by the Small Business and Regulatory Enforcement Fairness Act (SBREFA). 

EO 13132 emphasizes consultation with State and local ("S/L") governments and more sensitivity to their concerns. The EO also specifies a process for agencies to follow as they develop and implement actions that affect S/L governments. EPA's Guidance (2008) on Executive Order 13132, "Federalism", identifies various triggers for determining Federalism implications. 

The Regulatory Flexibility Act (RFA) of 1980 (5 U.S.C. 601 et. seq.) requires Federal agencies to review their proposed rules and regulations to determine if they will have "a significant economic impact on a substantial number" of small entities. Since its enactment in 1980, the RFA has required every federal agency to prepare regulatory flexibility analyses for any rule for which the agency is required to issue a notice of proposed rulemaking under the Administrative Procedure Act or any other statute unless the agency certifies that the rule "will not, if promulgated, have a significant economic impact on a substantial number of small entities." Enacted in 1996, the SBREFA requires formal procedures and analyses for rules that have a significant economic impact on a substantial number of small entities. The Agency refers to the RFA as amended by SBREFA simply as the RFA, but there are other SBREFA provisions as well.

The economic analysis seeks to determine whether the rulemaking has any federalism implications, to fulfil the requirements of EO 13132 or whether the rulemaking presents a significant economic impact on a substantial number of small entities according to the federal guidance for RFA compliance. 

 Executive Order 13132: Federalism
EPA's Guidance (2008) on Executive Order 13132, "Federalism", identifies the following trigger for determining Federalism implications:
      "...[I]f the impact of your rule on small governments is likely to equal or exceed 1 percent of their revenues, then as a policy matter, EPA will conclude the rule also has Federalism implications..."
This is referred to hereafter as the "1% test." The default definition of small government is a government of a city, county, town, village, school district or special district which serves a population of less than 50,000. This is the same definition used by the Regulatory Flexibility Act, as amended by the Small Business Regulatory Enforcement Fairness Act (RFA/SBREFA), and similar to the definition of small government in the Unfunded Mandates Reform Act (UMRA).
EPA (2008) recommends using an aggregate test for the 1% test, which involves applying total annualized costs as a percent of total revenues for the local governments that must conform to the rule. The EPA guidance also encourages consideration of how the impacts are distributed and how the rule will affect individual entities. While most impacts could be insignificant, a minority may experience significant impacts.  
2.1	Methods
EPA compiled and reviewed the list of Great Lakes CSO communities that would be subject to the proposed rule. There are 182 Great Lakes CSO communities (covered by 190 NPDES permits). EPA included in this analysis the CSO communities discharging to Chicago's Tunnel and Reservoir Plan (TARP). The CSOs and public notification requirements for these 36 communities are managed by the Metropolitan Water Reclamation District of Greater Chicago (MWRDGC). Out of the 182 CSOs, 150 were identified as small entities (defined as a governmental jurisdiction for a population of fewer than 50,000 people). Within the small communities, 32 discharge to TARP. 
Each CSO was matched to a location in the 2015 Census estimated population data (U.S. Census Bureau, 2016). Census location names were matched to either the name of the Great Lakes CSO permittee operator or the name of the city where the CSO discharge is located. 
2.1.1	Cost Data
The cost estimates were based on a breakdown of anticipated activities required by the rule and the estimated burden and cost of each. These costs estimates were derived based on the burden and cost estimates presented in the ICR ("Supporting Statement Information Collection Request: Notification Requirements for CSOs in the Great Lakes Basin" - OMB Control No. 2040-NEW, EPA ICR No. 2562.01). The Economic Analysis assumes that costs will be borne by Great Lakes CSO communities in the form of one-time implementation activities that will occur within one to two years, once per year activities including an annual notice, and ongoing activities that will occur during and after CSO discharges. This proposed rule requires Great Lakes CSO communities to develop a public notification plan and provide an annual notice that includes CSO discharge data. The proposed rule also requires states to ensure that public notification provisions are included in NPDES permits and review public notification plans. 
The three population ranges used for the cost estimates were "very small": 0-10,000 (to identify the smallest CSOs), "small": 10,000-50,000 (to correspond to small entity definition in EA), and large: greater than 50,000. The implementation activities for Great Lakes CSO communities are divided into the following categories:
 Installation of CSO outfall signs 
 Development of a method for collecting data 
 Development of alert system (website, text/email alerts) 
      Includes public outreach
 Develop and submit public notification plan 
 Consultation with local public health department
 Contact with municipalities and other potentially affected public entities

The recurring activities for Great Lakes CSO communities are divided into the following categories:
 Inspect and maintain signs
 Initial and supplemental notification through Public Alerts 
 Initial and supplemental notification of  potentially affected public entities   
 Maintenance/management of alert system 
 Prepare and release Annual Notice 
 Recordkeeping
      
The Implementation activities for State Agencies are further divided into the following categories:
 State Agency review of public notification plan

The Annual activities for State Agencies are divided further into the following categories:
 State Agency review of Annual Notice 
 State Agency revision of CSO NPDES permit conditions (first five years)

The burden hours and cost were calculated by estimating the annual burden, labor cost, and other costs per respondent or state for each activity. The number of Great Lakes CSO communities or State Agencies that will need to conduct each of the activities per year as a result of this regulation were estimated and used to calculate the yearly burden hours and costs. 
Not all Great Lakes CSO communities will need to conduct all the activities in response to these new requirements, and not all activities occur during all years. TARP communities received a reduced burden due to cost leveraging provide by the existing notification system.
The total yearly burden hours and costs are summed and averaged to compute the bottom line average annual burden hours and costs (For detailed year-by-year burden and costs for each activity see Appendix A Tables A.1 and A.2 of the ICR).
2.1.2	Revenue Data
Revenue for each small entity was determined from one of three data sources, in the following order of preference:
 Financial data from the U.S. Census Bureau for the specific government entity.
 Revenue data from internet resources.
 Estimated general revenue value using linear regression of available Census data.

Out of the 150 small entity Great Lakes CSO communities, census revenue data was available for 65 Great Lakes CSO communities. For the remaining CSOs, revenue data from other internet resources was applied if available. Then, a linear regression of available census data was used to estimate the remaining revenues. The three revenue data sources are described below. 


Financial Data from the U.S. Census Bureau for the Specific Government Entity
The U.S. Census Bureau conducts annual surveys on state and local governments, and collects data on revenue, expenditure, debt and assets. The most current state and local finance information available on the U.S. Census Bureau website was for 2013 (U.S. Census Bureau, 2016). 

The raw financial data for individual government entities is provided as item codes, where each item code represents a financial category (e.g., property taxes, general sales taxes, etc.)(U.S. Census Bureau, 2016). Descriptions, definitions, and additional information for item codes are available in the 2006 Government Finance and Employment Classification Manual (U.S. Census Bureau, 2006).

As defined in the "Methodology for Summary Tabulation  -  State and Local Finance" table (U.S. Census Bureau, 2013), the general revenue for each government entity was calculated as the sum of the following item codes:

      B01, B21, B22, B30, B42, B46, B50, B59, B79, B80, B89, B91, B92, B93, B94, C21, C30, C42, C46, C50, C79, C80, C89, C91, C92, C93, C94, D21, D30, D42, D46, D50, D79, D80, D89, D91, D92, D93, D94, T01, T09, T10, T11, T12, T13, T14, T15, T16, T19, T20, T21, T22, T23, T24, T25, T27, T28, T29, T40, T41, T50, T51, T53, T99, A01, A03, A09, A10, A12, A16, A18, A21, A36, A44, A45, A50, A56, A59, A60, A61, A80, A81, A87, A89, U01, U11, U20, U21, U30, U40, U41, U50, U95, U99

Revenue Data from Internet Resources
Revenue data for several small entities in Indiana, Michigan, New York and Ohio were found at state and community websites (AccessMyGov.com, 2016)(Ohio State Auditor, 2014)( State of Indiana, 2016)(State of Michigan, 2016)(State of New York, 2016).

Calculated General Revenue Value Using Linear Regression of Available Census Data
Data from the U.S. Census of Governments was used to derive the revenue for small governmental jurisdictions which could not be found from other sources. As described earlier, general revenue was calculated as the sum of select item codes.

A regression analysis was performed using population and calculated general revenue ("Type of Government" either "Township", "City" or "County"). The initial regression analysis utilized data for all available local governments (n = 3811) and resulted in a regression equation (linear with set intercept) with an R[2] of 0.3485. The regression equation was "y = 1629.4x" with y as general revenue in dollars and x as population.

The analysis was refined by performing separate regressions for Ohio (n=158) and Illinois (n=151) where the remaining small entities without revenue data were located.  The regression analyses utilized data for all available local governments with a population less than 50,000. The resulting regression equations were "y = 1390x (R[2]=0.525)" for Ohio and y = 1160.6x (R[2] = 0.1773) for Illinois. 

2.1.3	Federalism 1% Test
The annual cost and revenue projections were discounted at rates to reflect a range of potential economic conditions in the future. The Office of Management and Budget (OMB) currently reports a real rate of return of 1.5 percent for projects greater than 30 years. In the past, EPA has used discount rates of 3 and 7 percent which were determined to be appropriate for this EA. Present value (PV) estimates were determined using these rates. 
The time period for the rule is "in perpetuity". The annual cost and revenue streams were discounted across an infinite time period, and summed to estimate total present values. 
An annuity formula was used to calculate the annualized values. Then the annualized costs were divided by the annualized revenue to calculate the final metric: percent of revenue as compliance cost. When the result is less than 1 percent, then the rule does not have Federalism implications. 

2.2	Results
Appendix 1 provides the cash flow used to calculate the annualized cost and revenue values. Table 1 presents the results of the 1% test. The test was conducted using four estimates of revenue data: the average, minimum, and maximum per capita revenue from the data and estimates based on the regression of revenue on population for using available data in states with affected small entities. The results were well below 1% for all four revenue estimate methods. EPA concludes that the impact of the rule is very unlikely to reach or exceed 1% of small local government revenue, and, therefore, the rule does not have federalism implications. 
Table 1. Results of 1% Test (2016 dollars) 
Discount Rate
                                      3%
                                      7%
Annualized Cost for all 150 small CSO communities[1]
                                                                       $163,000
                                                                               
                                                                       $185,000
Annualized Total Revenue
                                                                 $3,038,723,000
                                                                               
                                                                 $3,156,732,000
                                                                               
Percent Revenue
                                                                         0.005%
                                                                         0.006%
[1]The average annualized costs are about $1,500 for a small CSO community and about $1,200 for a very small CSO community. 
3	Small Entity Screening Analysis
As part of this Economic Analysis, a quantitative analysis was performed to: determine which small entities are subject to the rule's requirements; select appropriate measures for determining economic impacts on these small entities and estimate those impacts; determine whether the rule may be certified as not having a significant impact on a substantial number of small entities (SISNOSE); and document the screening analysis and include the appropriate RFA statements in the preamble. A qualitative analysis was also conducted. 
According to EPA's Guidance for EPA Rulewriters: RFA (2006), it is preferable to start with a relatively simple methodology and add more complexity as needed. In Table 1 of the guidance, the preferred metric for small governmental jurisdictions is the revenue test, which is defined as annualized compliance costs as a percentage of annual government revenues. Thus the revenue test was used for the quantitative analysis. The data from the 1 percent test was used for the revenue test. Discount rates were applied consistent with the 1 percent test. From this analysis, the following information can be determined: the magnitude of economic impact that may be experienced by regulated small entities; total number of regulated small entities that may experience the economic impact; and percentage of regulated small entities that may experience the economic impact.
3.1	Qualitative analysis
The qualitative screening analysis addressed the following questions suggested in the guidance:
 Will small entities incur the same initial compliance costs - such as for buying and installing equipment, establishing new procedures, and/or record keeping - as large entities, or are initial costs likely to be proportionate to the size of the company? 
 Will initial compliance costs be high enough that entities will need to obtain financing and, if so, do small entities have the same access to financing as large entities? 
 Will small entities have to obtain access to professional or administrative services that large entities might maintain in-house?
For question #1, EPA reviewed the individual expenses and assessed whether costs would differ between small and large entities. Many of the required activities related to the number of CSOs in each community. Therefore, the costs would only be disproportionate if certain smaller communities had a disproportionate number of CSOs compared to other communities. 
Table 2 reports the number of outfalls by entity size. Generally, the number of outfalls are proportional to entity size. 
Table 2. Count of CSO Outfalls by Entity Size1 
Entity Size
                                    Average
                                      Min
                                      Max
Very Small
                                                                              4
                                                                              1
                                                                             17
Small
                                                                              9
                                                                              1
                                                                             44
Large
                                                                             25
                                                                              1
                                                                            126
[1]For 23 small entities (small and very small), data on number of outfalls was not available. Their populations ranged from about 1,000 to 40,000. 
Several of the implementation activities are not proportionate to the number of CSOs, and these activities include the development of an alert system (website, text/email alerts) and notification plan. Public outreach as well as communication with the potentially affected public entities may also present greater costs in proportion to revenue for small compared to large CSOs. The recurring activities that could also present slightly higher proportionate costs are maintenance/management of an alert system and preparation of the Annual Notice. 
The majority of the expenses are expected to be proportional to the number of CSOs. A relatively equitable burden is expected across different entity sizes. 
For Question #2, this analysis found that non-TARP cost estimates for very small non-TARP CSO communities range from approximately $1200 to $1300. These relatively low costs would not require financing and could be covered within most current municipal budgets with advanced notice and planning. The smallest non-TARP population among the affected entities is 580. Based on the average estimated costs, the proposed rule would result in an annualized cost per person of about $2.20 to $2.50 or an annualized cost per household of about $5.60 to $6.20 (assuming an average household size of 2.5 persons). A local government would not need financing for an incremental change in budget of this small magnitude. 

For Question #3, the activities necessary to comply with the proposed rule require administrative or professional services that a typical local government of any population should have available in-house. Most of the activities require word processing, electronic or paper communication, and basic web applications  -  any local government is likely to have to these skills internally. The installation and maintenance of signs could be handled by a typical local government maintenance crew.  The activity that is most likely to require external professional services is the reporting of discharge volume information for the initial notifications and the annual notice.  NPDES permits for CSO discharges in the Great Lakes basin currently require the reporting of CSO volumes in discharge monitoring reports (DMRs).  Typically, Great Lakes CSO communities use a combination of monitoring and modelling to estimate CSO volume.  EPA anticipates that most small CSO communities will use volume information developed for DMR reporting to comply with public notification requirements.  If the NPDES authority requires additional volume estimates for public notification for some communities, those communities who would rely on external contractors are probably already using contractors for similar work, and the additional monitoring, modeling, or other estimation methods would represent a small increase in their external work assignments. 
Most small governments have at least one professional on staff who is knowledgeable about water quality. The smallest communities may need to hire outside of their own staff to cover these needs. To account for this, communities in the very small population range were assigned costs for contractor services with the cost equal to the assumed burden hours multiplied by a labor rate equal to 150 percent of the community labor rate (see the ICR for more details).
3.2	Quantitative analysis
The cost impact ratios for small governmental jurisdictions were estimated by dividing the total annualized compliance cost for that entity by the annual revenue of the local government. Based on this calculation, at a 3% discount rate, all 150 of the small local governments affected by the proposed rule are expected to incur a cost impact of less than 1% of annual revenue (Table 3). Similarly, at a 7% discount rate, all 150 of the small local governments affected by the proposed rule are expected to incur a cost impact of less than 1% of annual revenue (Table 3). Cost impacts of less than 1% are considered to be minimal, and no further action is required. 

Table 3. Quantitative Screening Analysis

                         Cost-Impact Ratio Percentage
                         Small Local Government Count
                         % of Small Local Governments
                                    >3%
                                       0
                                       0
                                     1-3%
                                       0
                                       0
                                    <1%
                                      150
                                     100%
                                     Total
                                      150
                                     100%

4	Key Findings
The economic analysis determined that the rulemaking does not have any federalism implications, fulfilling the requirements of EO 13132. The percentage of total annualized revenue represented by total annualized cost for small CSO communities was less than 0.01%. The screening analysis also determined that the rulemaking does not present a significant economic impact on a substantial number of small entities according to the federal guidance for RFA compliance. The screening analysis showed that the rule would produce a Cost-Impact Ratio of less than <1% for all affected small entities. 
 5	Uncertainties and Limitations
The cost estimates were based on available data for number of CSO outfalls and assumptions regarding the frequency that CSO communities would need to conduct required activities. The assumptions were developed to represent typical local government operations and were tailored to the three population categories of very small (<10,000), small (<50,000), and larger. While the assumptions were developed to account for major differences between the size categories, the accuracy of the analysis is limited to representing a typical local government within each size category. None of the assumptions are expected to exactly reflect an individual local government, and some variability is expected. 
The revenue estimates were based on data available through internet sources as described above. For the local governments where revenue data were not available, use of a regression to estimate their revenue introduces additional uncertainty. The regression provides an estimate of the average estimated annual revenue for a local government with a given population. The actual revenues could vary widely. In the available revenue dataset for OH entities used for the regression, the very small entities (<10,000) ranged in annual revenue from about $400,000 to $31 million with a mean and standard deviation of $10 and $6 million respectively. 
The cost and revenue estimates do not account for changes in values over time. The analysis assumes that the proportion of costs to revenue will remain constant in perpetuity. While some variation in the costs and revenue is expected over time and across entities, this variation is not expected to affect the conclusions of the analysis. As the local governments grow in population and build more infrastructure, the costs for monitoring and public notification are expected to become a smaller fraction of the total revenue. Since this change was not accounted for in the analysis, the results provide a conservative estimate of the rule's impact on CSO communities. 

6	References
AccessMyGov.com. 2016. Municipal Directory. https://accessmygov.com/MunicipalDirectory 
Ohio State Auditor. 2014. Village of Dunkirk, Hardin County, Regular Audit for the Year Ended December 31, 2013, Single Audit for the Year Ended December 31, 2012. http://villageofdunkirk.com/yahoo_site_admin/assets/docs/12-13_AUDIT.2683615.pdf 
State of Indiana. 2016. Indiana Gateway for Government Units, Fiscal Health Indicators. https://gateway.ifionline.org/public/FISCAL/download.aspx 
State of Michigan Department of Treasury. 2016. Data Summary Report. https://f65.mitreasury.msu.edu/Reports/DataSummaryReport.aspx 
State of New York. 2016. Financial Data for Local Governments. New York State Office of the State Comptroller. https://www.osc.state.ny.us/localgov/datanstat/findata/index_choice.htm 
U.S. Census Bureau. 2006. Government Finance and Employment Classification Manual. http://www2.census.gov/govs/pubs/classification/2006_classification_manual.pdf 
U.S. Census Bureau. 2013. Methodology for Summary Tabulation  -  State and Local Finance. http://www2.census.gov/govs/local/methodology_for_summary_tabulations_2013.xls 
U.S. Census Bureau. 2016. Annual Estimates of the Resident Population for Incorporated Places: April 1, 2010 to July 1, 2016. https://www.census.gov/popest/data/cities/totals/2015/SUB-EST2015-3.html.  Accessed October 9, 2016
U.S. Environmental Protection Agency (EPA). 2006. EPA's Action Development Process Final Guidance for EPA Rulewriters: Regulatory Flexibility Act. Amended by the Small Business Regulatory Enforcement Fairness Act. https://www.epa.gov/sites/production/files/2015-06/documents/guidance-regflexact.pdf 
U.S. Environmental Protection Agency (EPA). 2008. EPA's Action Development Process. Guidance on Executive Order 13132: Federalism. http://www.govexec.com/pdfs/111908rb1.pdf 
U.S. Environmental Protection Agency (EPA). 2016. Report to Congress: Combined Sewer Overflows into the Great Lakes Basin. EPA 833-R-16-0060. https://www.epa.gov/sites/production/files/2016-05/documents/gls_cso_report_to_congress_-_4-12-2016.pdf 

Appendix 1. Cash Flows 

Discount Rate
                                                                             7%
 
 
 
Period
                                                                  In perpetuity
 
 
 
 
 
                                                                         Year 1
                                                                         Year 2
                                                                    Years 3+[1]
One-time costs
Present Value
                                                                       $416,000
                                                                        $20,000
                                                                             $0
Recurring costs
Present Value
                                                                       $103,000
                                                                       $136,000
                                                                     $1,967,000
Total costs by year
Present Value
                                                                       $519,000
                                                                       $156,000
                                                                     $1,967,000

 
 
 
 
Costs, Total Present Value (PV)
                                                                     $2,642,000
 
 
 
 
 
 
 
 
Recurring revenue
Present Value
                                                                 $2,950,217,000
                                                                 $2,757,212,000
                                                                $39,388,743,000
 
 
 
 
 
Revenue, Total PV
                                                                $45,096,171,000
 
 
 
 
 
 
 
 
Annualized Total Costs 
                                                                       $185,000
 
 
Annualized Total Revenue
                                                                 $3,156,732,000
 

Percent of Revenue Represented by Costs
                                                                         0.006%
 
 

[1]Values represent total present value from Year 3 in perpetuity. 


Discount Rate
                                                                             3%
 
 
 
Period
In perpetuity
 
 
 
 
 
                                                                              1
                                                                              2
                                                                    Years 3+[1]
One-time costs
PV
                                                                       $416,000
                                                                        $21,000
                                                                             $0
Recurring costs
PV
                                                                       $103,000
                                                                       $141,000
                                                                     $4,769,000
Total costs by year
PV
                                                                       $519,000
                                                                       $162,000
                                                                     $4,769,000
 
 
 
 
 
Costs, Total PV
                                                                     $5,449,000
 
 
 
 
 
 
 
 
Recurring revenue
PV
                                                                 $2,950,217,000
                                                                 $2,864,288,000
                                                                $95,476,272,000
 
 
 
 
 
Revenue, Total PV
                                                               $101,290,777,000
 
 
 
 
 
 
 
 
Annualized Total Costs 
                                                                       $163,000
 
 
Annualized Total Revenue
                                                                 $3,038,723,000

 
Percent of Revenue Represented by Costs 
                                                                         0.005%
 
 

[1]Values represent total PV from Year 3 in perpetuity. 
