--------------------------------------------------------------------------------
Analysis of State- and County-Level PM2.5 Benefits

Overview
The Regulatory Impact Analysis (RIA) accompanying the recently proposed Toxics Rule summarized the national and regional PM2.5-related health co-benefits associated with the rule (EPA, 2011). Due to data, time and resource constraints, EPA did not report these co-benefits at a sub-regional level. In this Technical Support Document (TSD) we describe our approach for allocating the national-level PM2.5-related monetized benefits to the state, and the county-level avoided premature deaths to the county. We also summarize the results of this analysis and describe the limitations and uncertainties associated with our approach. A key limitation of our approach is that it does not completely account for PM2.5 transport between states, and so may under- or over-estimate PM-related co-benefits for certain states.

Methodology
As noted in the RIA, the air quality modeling scenario used in developing national benefits estimates for the Toxics Rule reflected an early assessment of the potential utility emissions controls, and as a result, the projected estimates of ambient PM2.5 did not match the expected level and spatial distribution of PM2.5 levels resulting from the specific emissions changes that are associated with implementation of the actual proposed rule; time limitations prevented EPA from modeling the air quality changes associated with emissions controls estimated to result from the proposed rule. For this reason, in estimating the national total PM2.5 related benefits of the proposed rule, EPA applied a benefit per-ton approach, in which it scaled the PM2.5 co-benefits estimated from the air quality modeled scenario based on the relative national levels of SO2 emission reductions; this procedure is described in the PM2.5 co-benefits chapter of the RIA. A key limitation of the benefit-per ton approach is that it generally provides a very limited characterization of the benefits at a sub-regional (i.e. state or county) scale. However, spatially resolved estimates of the co-benefits are clearly very useful to understanding how the expected benefits of this rule are distributed across the U.S.
The spatial resolution of the input data to the benefits calculations affected the scale at which we could report the monetized benefits and premature deaths. For example, estimates of premature death were calculated in BenMAP (prior to the scaling exercise) using a combination of national and county data, which gives us greater confidence in reporting county-level results. The total monetized benefits were calculated using a combination of national, regional and county-level data, giving us greater confidence in reporting state-level results.
For this reason, EPA developed an approximation approach for allocating the national-level PM2.5 co-benefits estimated in the Toxics Rule RIA to the state and county level. Below we first summarize our approach for calculating state-level monetized benefits and then detail our methods for estimating the number of premature deaths avoided at each county. This approach is subject to important uncertainties, which we also describe below.

Allocating the national-level PM2.5 monetized co-benefits to the state-level
EPA followed the sequence of steps below to allocate the national-level PM2.5 monetized co-benefits to each state.
   1. Estimate the state-level monetized co-benefits of the air quality modeled scenario. Using the environmental Benefits Mapping and Analysis Program (Abt, 2008), EPA quantified the total monetized PM2.5-related co-benefits of the air quality modeled scenario for each state. The methods for estimating the monetized benefits of changes in PM2.5 air quality levels are detailed in the co-benefits chapter of the RIA.
   2. Calculate the percentage change in state-level SO2 reductions between the modeled and proposed policy cases. There were significant changes in both the size and distribution of SO2 emission reductions between the modeled and proposed policy case; this step accounts for the percentage change in SO2 reductions for each state.
   3. Scale the state-level monetized co-benefits according to the percentage change in SO2 emissions between the air quality modeled and final policy cases. In this final step, the state-level monetized co-benefits estimated in step 1 is adjusted according to the percentage calculated in step 2.

Allocating the national-level avoided PM2.5 premature deaths to the county-level
EPA followed the sequence of steps below to allocate the national-level avoided PM2.5 premature deaths to each county.
   1. Estimate the number of avoided PM2.5-related premature deaths of the air quality modeled scenario for each county. Using the environmental Benefits Mapping and Analysis Program (Abt, 2008), for each county EPA quantified the number of PM2.5-related premature deaths avoided resulting from the air quality modeled scenario. The methods for estimating the number of avoided PM2.5-related deaths are detailed in the co-benefits chapter of the RIA.
   2. Calculate the percentage change in state-level SO2 reductions between the modeled and proposed policy cases. There were significant changes in both the size and distribution of SO2 emission reductions between the modeled and proposed policy case; this step accounts for the percentage change in SO2 reductions for each state.
   3. Scale the county-level avoided PM2.5 deaths according to the percentage change in SO2 emissions between the air quality modeled and final policy cases for each state. In this final step, the number of PM2.5 related deaths avoided in each county in step 1 is adjusted according to the percentage change in SO2 emissions for that state calculated in step 2.
Limitations and uncertainties
The method described above adds unique uncertainties and limitations beyond those already described in detail in the co-benefits chapter:
   1. This approach does not account fully for the transport of PM2.5 between states. This method assumes that the PM2.5-related monetized benefits and avoided deaths are proportional to the change in PM2.5 precursors in that state. However, emission changes in one state affect the PM2.5 levels in neighboring states. As a result, we may under- or over-estimate the PM2.5 monetized benefits and avoided mortalities in each state.

   2. This approach does not account for the transport of PM2.5 within states. This method assumes that the number of avoided deaths in each county within each state scales according to the change in emissions between the air quality modeled and policy case in that state. As a result, this method does not account for the fact that the location of emissions within the state are likely to also change, which would affect the number of avoided deaths in each county.

   3. States in which no SO2 emissions are reduced are allocated no benefits. For Kentucky we estimate no change in PM-related deaths or monetized benefits due to the fact that the final policy scenario did not apply any SO2 emission controls to units in Kentucky; this reflects the results of a recent settlement with the Tennessee Valley Authority that went into effect after the completion of the air quality modeling but before the rule was proposed. However, because upwind states have significantly reduced SO2 emissions, we expect Kentucky to receive significant PM2.5-related co-benefits under the proposed rule.

Results
Below we report the PM2.5-related monetized co-benefits in each state (Table 1). We also provide a map of the monetized co-benefits in each state (Figure 1) and the number of PM2.5-related deaths avoided in each county (Figure 2).

Table 1: Estimated PM2.5-related Monetized Co-Benefits for Each State
State
                           Total Monetized Benefits
                     (3% discount rate, millions of 2007$)

                     Pope et al. (2002) mortality estimate
                    Laden et al. (2006) mortality estimate
Alabama
                                    $1,600
                                    $3,900
Arizona
                                     $400
                                    $1,000
Arkansas
                                    $1,200
                                    $2,800
California
                                     $580
                                    $1,400
Colorado
                                     $410
                                     $990
Connecticut
                                    $2,400
                                    $5,900
Delaware
                                     $250
                                     $620
DC
                                     $130
                                     $320
Florida
                                    $4,700
                                    $12,000
Georgia
                                     $880
                                    $2,100
Idaho
                                      $18
                                      $45
Illinois
                                    $1,900
                                    $4,600
Indiana
                                    $1,100
                                    $2,600
Iowa
                                     $550
                                    $1,300
Kansas
                                     $680
                                    $1,700
Kentucky
                                    ---[A]
                                    ---[A]
Louisiana
                                    $1,300
                                    $3,300
Maine
                                      $75
                                     $180
Maryland
                                    $1,300
                                    $3,200
Massachusetts
                                    $2,700
                                    $6,700
Michigan
                                    $2,400
                                    $5,800
Minnesota
                                     $580
                                    $1,400
Mississippi
                                     $480
                                    $1,200
Missouri
                                    $1,400
                                    $3,500
Montana
                                      $75
                                     $190
Nebraska
                                     $320
                                     $790
Nevada
                                      $17
                                      $43
New Hampshire
                                     $200
                                     $490
New Jersey
                                    $6,700
                                    $16,000
New Mexico
                                     $100
                                     $250
New York
                                    $5,100
                                    $12,000
North Carolina
                                    $1,700
                                    $4,100
North Dakota
                                      $79
                                     $190
Ohio
                                    $2,700
                                    $6,700
Oklahoma
                                    $1,100
                                    $2,600
Oregon
                                      $11
                                      $28
Pennsylvania
                                    $3,300
                                    $8,200
Rhode Island
                                    ---[A]
                                    ---[A]
South Carolina
                                    $1,800
                                    $4,400
South Dakota
                                     $130
                                     $320
Tennessee
                                    $2,500
                                    $6,200
Texas
                                    $8,800
                                    $22,000
Utah
                                      $94
                                     $230
Vermont
                                    ---[A]
                                    ---[A]
Virginia
                                    $2,000
                                    $4,900
Washington
                                    ---[A]
                                    ---[A]
West Virginia
                                     $650
                                    $1,600
Wisconsin
                                    $1,300
                                    $3,100
Wyoming
                                      $16
                                      $38
[A] While methodological limitations prevented us from estimating PM2.5-related benefits in these states, emission reductions from neighboring states are likely to provide PM2.5-related benefits.
	

Figure 1: Estimated PM2.5-related Monetized Co-Benefits for Each State



Figure 2: Estimated PM2.5-related Deaths Avoided for Each County

