EPA’s Use of Science in Benefits Analyses of Air RulesPresentation to APPA Energy & Air Quality Task Force

Washington DC

October 12, 2011

Anne E. Smith

Senior Vice President

2

A Shocking Statement About EGU Air Toxics

“EPA‘s proposed mercury and air toxics standards for power plants that burn coal and oil are projected to save as many as 17,000 American lives every year by 2015. These standards also will prevent up to 11,000 cases of heart attacks, 120,000 cases of asthma attacks, 11,000 cases of acute bronchitis among children, 12,000 emergency room and hospital visits and 850,000 lost work days every year.”

John D. Walke, Natural Resources Defense CouncilTestimony before ECC, April 15, 2011.

3

Outside of its RIAs, EPA Also Overstates its Scientific Evidence

“When these new standards are finalized, they will assist in preventing 11,000 heart attacks, 17,000 premature deaths, 120,000 cases of childhood asthma symptoms and approximately 11,000 fewer cases of acute bronchitis among children each year. Hospital visits will be reduced and nearly 850,000 fewer days of work will be missed due to illness.”

Administrator Lisa Jackson(EPA Air News Release extending public comment period on EGU MACT Rule, June 21, 2011)

4

A Closer Look at the Benefit-Cost Evidence in the EGU MACT RIA

(a)

Benefits from air toxics reduction

(billions/yr)

(b) Co-benefits from non-

toxics (billions/yr)

(c)

Costs (billions/yr)

(d) Net Benefits

without co-benefits (billions/yr)

(e) Net Benefits

including co-benefits (billions/yr)

Mercury MACT

< $0.1 $0.6 to $1.5 $2.3 - $2.3 - $1.7 to -$0.8

Acid Gases MACT

$0 51.7 to 136.9 $5.4 - $5.4 $46.2 to $131.5

Non-Hg Metals MACT

$0 0.7 to 1.6 $3.2 - $3.2 - $2.5 to -$1.6

Organic HAPs Standard

$0 $0 >$0(*) <$0 <$0

Total

< $0.1 $53 to $140 $10.9 - $10.9 $42 to 129.1

Source of Table: from Table 1 in A. Smith, Technical Comments on Utility MACT RIA, available at: http://www.nera.com/67_7412.htm

Almost all EGU MACT benefits are due to acid gases MACT, which forces SO2 reductions that are not needed to meet the PM2.5 NAAQS…but EPA

attributes PM2.5 “benefits” them for all air toxics justification

5

PM2.5 Co-Benefits Play Similar Role in RIA for Ozone NAAQS Reconsideration

Every blue and green bar in the chart includes PM2.5 “co-benefits” from that are significantly larger than the ozone benefits. For the 0.070 ppm alternative on the left:

– PM2.5 co-benefits range from about $3b to $32b in the various bars (they are greater than $10b in 93% of the 168 cases)

– Ozone benefits range from $2b to $10b (all presume causality of ozone for mortality)

Costs range from $19b to $25b (for the .070 ppm alternative standard)

If the PM2.5 co-benefits were removed (leaving only ozone-related benefits and costs) every single bar in the figure would be negative -- see next slide

– The same situation exists for all the other alternative ozone standards, including the current 0.075 ppm standard.

Every blue and green bar in the chart includes PM2.5 “co-benefits” from that are significantly larger than the ozone benefits. For the 0.070 ppm alternative on the left:

– PM2.5 co-benefits range from about $3b to $32b in the various bars (they are greater than $10b in 93% of the 168 cases)

– Ozone benefits range from $2b to $10b (all presume causality of ozone for mortality)

Costs range from $19b to $25b (for the .070 ppm alternative standard)

If the PM2.5 co-benefits were removed (leaving only ozone-related benefits and costs) every single bar in the figure would be negative -- see next slide

– The same situation exists for all the other alternative ozone standards, including the current 0.075 ppm standard.

For example, from Supplemental RIA, p. S1-6:

(This chart shows 168 different estimates of net benefits for the 0.070 ppm alternative, ordered from smallest to largest net benefit, without any probability weights. Each bar contains a different combination of EPA’s ozone mortality, PM2.5 mortality, and cost assumptions. All of the bars include the assumption of ozone-mortality causality.)

6

Net Benefits in O3 Reconsideration RIA Using Only Ozone-Related Benefits

The net benefits range is negative by billions of dollars per year when PM2.5 co-benefits are removed

Source: A. Smith, Summary and Critique of the Benefits Estimates in the RIA for the Ozone NAAQS Reconsideration, prepared for American Petroleum Institute. Available at: http://www.nera.com/67_7390.htm.

7

EPA’s Co-Benefits Habit Has Been Growing

Year

RIAs for Rules Not Targeting Ambient PM2.5

PM2.5 Co-Benefits

Are >50% of Benefits

Co-Benefits Are Only Benefits

Quantified

1996 Ozone NAAQS (.12 1hr=>.08 8hr) X 1997 Pulp & Paper NESHAP 1999 Regional Haze Rule X 1998 NOx SIP Call & Section 126 Petitions X 1999 Final Section 126 Petition Rule X X 2003 Stationary Reciprocating Internal Combustion Engine NESHAP X X 2004 Plywood & Composite Wood Products NESHAP (no health benefits quantified)

2004 Automobile & Light-Duty Vehicle Manufacturing NESHAP (no health benefits quantified)

2004 Industrial Boilers & Process Heaters NESHAP X X 2005 Clean Air Mercury Rule X 2005 Clean Air Visibility Rule/BART Guidelines X 2006 Stationary Compression Ignition Internal Combustion Engine NSPS 2008 Ozone NAAQS (.08 8hr =>.075 8hr) X 2008 Petroleum Refineries NSPS X 2008 Lead (Pb) NAAQS X 2009 Portland Cement Manufacturing NESHAP X X 2010 Ozone Reconsideration X 2010 NO2 NAAQS X X 2010 Existing Stationary Compression Ignition Engine NESHAP X X 2010 Indus'l, Comm'l & Institutional l Boilers & Process Heaters NESHAP X X 2010 Greenhouse Gases PSD and Tailoring Rule (no health benefits quantified)

2010 SO2 NAAQS (==> 1-hr 75 ppb) X >99% 2010 Portland Cement Manuf'g NSPS & NESHAP Amendment X X 2011 Sewage Sludge Incineration Units NSPS & Emission Guidelines X X 2011 Comm'l & Indus'l Solid Waste Incin Units NSPS & Emission Guidelines X X 2011 Utility Boiler MACT NESHAP X >99.99% 2011 Mercury Cell Chlor Alkali Plant Mercury Emissions NESHAP X X 2011 Oil & Natural Gas Industry NSPS & NESHAP Amendment (no health benefits quantified)

8

Distribution of PM2.5 Mortality Risk in 2005

Source: Figure C-2, EGU MACT RIA

Read the Fine Print: This figure is

consistent with the 6,800 deaths estimate

in the EGU MACT

9

Distribution of PM2.5 Mortality Risk in 2005

This is the figure that is consistent with the

17,000 deaths estimate in the EGU MACT(not provided by EPA)

2% to 7%7% to 10%

10% to 13%13% to 16%16% to 22%

320,000 deathsin 2005 (13%)

10

Implications of EPA’s High End Risk Estimates: Is this Credible?

Using PM2.5 concentrations in 1979-1983:

25% of all deaths nationwide average

~40% of deaths in highest PM cities

Numbers would probably have been even higher ca. 1970 (earliest national SO2 emissions data)

11

How Did EPA Get to These High Mortality Risk Estimates?

RIAs 2006-2009: 68,000 – 88,000 (Pope 2002) (Laden

2006)

PM2.5 NAAQS QRA (2010) 63,000 (Krewski 2009)

RIAs since 2009: 120,000 – 320,000 (Pope 2002) Laden

2006)

EPA RIA’s started extrapolating epidemiological estimatesto PM2.5 levels as low as modeled (< 4 ug/m3)

(counting below the “lowest measured level” in the studies)

3.6x

12

EPA’s 2009 Methodology Change Feeds its Co-Benefits Habit

The upper bound of 320,000 deaths in 2005= $3 trillion of WTP to reduce ALL PM2.5 in 2005

Post-CSAPR, ~250,000 deaths “still to be saved”

– That leaves about $2.3 trillion more dollars of benefits to be had by driving PM2.5 down further

– Almost all of that remaining $2.3 trillion is associated with PM2.5 < 11 μg/m3

EPA is not prepared to set the PM2.5 NAAQS to protect these lives …

but is prepared to use them as “co-benefits” to justify regulations on other pollutants that lack their own quantifiable benefits

13

90% of CSAPR’s 13,000-34,000 “lives saved” are from PM2.5 exposures already below 13 μg/m3 (NAAQS is 15 μg/m3 )

(Figure 5-19, copied from CSAPR RIA)

90% at about 13 μg/m3

Median at about11 μg/m3

14

90% of Utility MACT’s 6,800 to 17,000 “lives saved” (after CSAPR) due to PM2.5 co-benefits at PM2.5 exposures below 11.5 μg/m3

Median atabout 9.3 μg/m3

90% at about 11.5 μg/m3

(Figure 6-15, copied from Utility MACT RIA)

This means that almost all the Utility MACT’s co-benefits are from reductions in PM2.5 that occur below the tightest level

recommended by EPA staff and CASAC for a revised PM2.5

NAAQS

15

Illustration of the “Wealth Sharing Method”: 3 Sets of Benefits & Costs before Inclusion of Co-Benefits

PM2.5 Ozone UtilityMACT

Benefit Benefit BenefitCostCost Cost

Note: While EPA estimates such large PM2.5 benefits, many consider them speculative because, as shown on earlier slides, they are mainly attributable to reductions in PM2.5 well below current and potential future “safe” NAAQS levels.

16

Same 3 Rules After Including PM2.5

Co-Benefits

PM2.5 Ozone UtilityMACT

Benefit Benefit BenefitCostCost Cost

Portions of PM2.5’s benefits could be shared out to every justify every regulation on the RIA list and still have enough benefits left over to justify tightening the PM2.5 NAAQS.

17

The Already-Regulated PM2.5 Should Not Serve as Justification for Regulating Other Pollutants

Scares the public into believing that new rules are absolutely essential for their health.

Gives EPA a shield to build a complex web of many different rules, when EPA could provide almost all of those purported health-protective benefits with just a single rule: the PM2.5 NAAQS.

– That EPA does not take this simple, streamlined approach hints at the degree to which it realizes that its co-benefits calculations do not reflect true public health risks.

Is just plain bad policy: This approach cannot possibly result in a cost-effective path to addressing a nation’s clean air needs.

The Problems with EPA’s Use of Co-Benefits:

Contact UsAnne E. Smith

Senior Vice PresidentNERA—Washington DC+1 202 466 9229Anne.Smith@NERA.com

© Copyright 2011National Economic Research Associates, Inc.

All rights reserved.

19

Our Global Presence

Our global team of more than 500 professionals operates in more than 25 offices across North America, Europe, and Asia Pacific.

Sydney

Tokyo

Melbourne

Shanghai

Auckland

Wellington

BeijingSan FranciscoChicago

DenverLos Angeles

Boston

New York CityPhiladelphia

Washington, DC

White Plains

IthacaToronto

Madrid

Rome

Frankfurt

London Brussels

Paris Geneva