Looking Back at Fifty Years of the Clean Air Act - After major expansion in 1970, the Clean Air Act led to substantial emissions reductions and health improvements—as well as some unintended consequences.

Abstract

Since 1970, transportation, power generation, and manufacturing have dramatically transformed as air pollutant emissions fell significantly. To evaluate the causal impacts of the Clean Air Act on these changes, we synthesize and review retrospective analyses of air quality regulations. The geographic heterogeneity in regulatory stringency common to many regulations has important implications for emissions, public health, compliance costs, and employment. Cap-and-trade programs have delivered greater emission reductions at lower cost than conventional regulatory mandates, but policy practice has fallen short of the cost-effective ideal. Implementing regulations in imperfectly competitive markets have also influenced the Clean Air Act’s benefits and costs.

NYC Smog 1966 https://en.wikipedia.org/wiki/1966_New_York_City_smog

Key Findings

• Spatially varying regulations can impose substantial costs on local economies.
• Current applications of market-based mechanisms may fall short of cost-saving expectations.
• Varying fuel content regulations across the United States may impose unnecessary costs on consumers in separated markets.
• Regulatory flexibility for fuel content rules doesn’t always yield cost-effective results.
• Unanticipated costs arising from overly optimistic technology projections are an important issue in the design of renewable fuel requirements.

The SO2 program has been subject to extensive research, with a number of papers focusing on the early years (such as Carlson et al. 2000 and Ellerman et al. 2000) and some recent synthesis and review papers which combine ex-ante and ex-post papers (such as Schmalensee and Stavins 2013). The ex-ante analyses all suggest large cost savings based on a comparison of the least cost solution of achieving the cap to the command-and-control uniform performance standard case. Carlson et al. (2000) note that this cost reduction reflected dramatic declines in their estimated marginal abatement cost functions for sulfur dioxide emissions resulting from changes in technology and low-sulfur coal prices over 1985-1995.

The only true ex post study of the program’s benefits and costs is by Chan et al. (2018), which finds much smaller cost savings than predicted ex ante. In part, this is the result of decisions of several power plants—in concert with their state public utility commissions—to install scrubbers rather than comply by purchasing allowances and/or using low sulfur coal, a decision that Chan et al. estimate increased annual compliance costs by nearly $100 million. Focusing on 2002 as a Phase II year before the transition to a period of regulatory uncertainty and using a mixed logit model of the firm’s compliance decision, the authors find that the SO2 program reduced compliance costs by about $200 million (1995$) and increased public health benefits by roughly $170 million. Chan et al. examine a performance standard that delivers the same aggregate emission outcome as the Acid Rain Program in 2002, which had much higher emissions than the cap due to use of banked allowances. Thus, the cost-savings of the two instruments may be smaller than they would have been under the statutory cap for 2002. Chan et al. also find that the prevailing pattern of allowance trading— from western generating units in sparsely populated areas to eastern generating units in more densely populated areas—increases public health damages by about $2 billion relative to a no-trade counterfactual.

...

The Chan et al. paper builds on the insights in Muller and Mendelsohn (2009), which illustrated through an integrated assessment model how the location of an emission source relative to a downwind population could dramatically affect the monetized damages of a ton of sulfur dioxide emitted at that source. In their counterfactual analyses, Muller and Mendelsohn estimated that trading ratios, based on the relative damages associated with a ton of emissions for a pair of locations, could improve social welfare by nearly $1 billion per year compared to the ton-for-ton trading in the SO2 program as implemented. However, such differentiation in cap-and-trade implementation raises questions about administrative feasibility and accuracy in estimating ratios, especially in the presence of a complicated atmospheric chemistry that could induce negative ratios for NOx (Fraas and Lutter 2012). 

...

While overall coal prices fell during the latter half of the 1990’s, Busse and Keohane found that delivered prices rose for plants covered by Phase I of the SO2 cap-and-trade program relative to those still operating under command-and-control regulation, and prices rose more at plants near a low-sulfur coal source. Overall, they estimate that railroads enjoyed an increase in annual producer surplus of more than $40 million, which represented about 15 percent of the economic surplus created by the cap-andtrade program....

NOX Budget Program

The efforts to employ a cap-and-trade program to reduce nitrogen oxide (NOx) pollution emerged over two phases in the eastern United States. 

 Deschênes et al. (2017) exploit ... design characteristics in a triple-differencing empirical strategy to estimate a reduction in NOx emissions of about 40 percent in the summer months for sources in the states covered by the program after it started. This translated into about a 6 percent reduction in mean ozone concentrations and a 35 percent reduction in the number of high-ozone

days during the summer months.

LA Smog 1968 https://www.kcet.org/history-society/how-los-angeles-began-to-put-its-smoggy-days-behind

The significant reductions in emissions and ozone concentrations contributed to substantial public health benefits. Deschênes et al. employed the same strategy to estimate a reduction in premature mortality of about 2,000 individuals annually, primarily among the 75 and older population. They monetized these mortality risk reduction benefits at about $1.3 billion. A novel element of the Deschênes et al. analysis focuses on how regulations improving air quality can reduce the demand for

and expenditures on pharmaceuticals, medical care, and related defensive activities.

With high-frequency, spatially disaggregated proprietary data on health insurance-related pharmaceutical spending, they estimate large reductions in such defensive expenditures, on the order of about $800 million per year.

To characterize the welfare impacts of the NOx Budget Program, Deschênes et al compile their monetized estimates of the benefits and compare them to a back-of-the-envelope estimate of the costs of the program. For the latter, they assume that the allowance price clearing in the market (on average about $2,500 per ton of NOx) can serve as the upper bound on the abatement costs. The product of the average allowance price and their estimated NOx emission reductions to produce an upper bound cost estimate of about $1.1 billion annually. Based on medication expenditure cost-savings and reduced premature mortality, they estimate annual social benefits ranging from about $1.5 to $2.1 billion (2015$). Overall, they conclude that the net social benefits of the NOx Budget Program are positive. 

...

In contrast to the capital investment compliance strategies studied in Fowlie (2010), Linn (2008) focuses on those facilities that opted against making major capital investments in abatement technologies, such as selective catalytic reduction, and instead pursued temporary boiler modifications as a way to reduce NOx emissions. These modifications are considered relatively low-cost and can be reversed during the winter months when the NOx cap-and-trade program does not operate.

...

Linn finds that such modifications reduce NOx emissions by 10 to 15 percent, at costs likely less than $2,000 per ton. He also notes how the cap-and-trade policy delivers incentives for emissions abatement through fairly modest process changes that would not likely occur under more prescriptive command-and-control regulations.

...

 Air Toxic Regulations Under the 1990 CAAA

...

The intent of MACT standards was to raise the laggards to the level of the best performers in the industry [defined as the level of control achieved by the best performing 12 percent of plants in the industry], rather than force the adoption of exotic and unproven technologies.31 Between 1994 and 1998 the EPA issued 21 sets of MACT standards, including standards for 13 manufacturing industries. In its second report to Congress on the Benefits and Costs of the CAA, EPA (1999) reported that these standards would impose annual costs of $480 million in 2000.

...

Effectiveness and Health Benefits of RVP and RFG Regulations

...

Marcus finds that asthma hospitalizations decrease by 4.5 per 10,000 children, an 8 percent reduction relative to the group’s pre-policy level. Treatment effects are not different for crosswind vs. downwind zip codes, but both these groups do have larger negative effects than the upwind zip codes. Impacts are also greater for high traffic zip codes. In sum, she finds that the policy reduced asthma hospitalizations by 1,449 per year, resulting in $13.2 million in avoided health expenditures. This suggests that more stringent regulations on gasoline had a significant impact on child health, as well as reducing asthma treatment costs.

Market Impacts of RVP and RFG Regulations and Oxygenated Fuel Regulations

Using data from 1994 to 2003, Sweeney estimates refinery-level costs to produce gasoline subject to applicable fuel content regulations. He finds that gasoline prices in RFG regions are about 7 cents per gallon higher than they otherwise would be. This price change translates to a $25 billion reduction in consumer surplus from 1995 to 2003 ($2.85 billion per year). Since some refiners opt against producing RFG and reallocate supply toward unregulated regions, the gasoline price in unregulated regions decreases by about 2 cents per gallon, resulting in an $11 billion increase in consumer surplus in these regions. This also illustrates how structural models permit an explicit examination of the interplay of market structure and regulatory implementation, which can be important in a variety of air quality contexts in which the regulated sources operate in imperfectly competitive markets.

...

Renewable Fuel Standards (RFS)

...

Lade, et al. (2018b) estimate a ... specification with the logarithm of first-differenced stock market prices of biofuel firms as the dependent variable. They show that RIN prices increased in 2013 as mandates forced ethanol consumption closer to the blend wall. In August 2013, the EPA’s 2013 final rule hinted that the 2014 total mandate would be reduced because of the market’s limited capacity to consume gasoline containing more than 10 percent ethanol. This announcement reduced RIN prices by about 30 percent over the next three days, which translates to a $7 billion reduction in the value of the 2013 RIN market. Two subsequent events—a leak of the 2014 proposed rule and the official release of the 2014 proposed rule—are associated with smaller decreases in RIN prices. There are small changes (1-2 percent) in commodity futures prices coincident with some of the three events.

...

Impact of Attainment Status on Employment and Earnings

...

Greenstone ...asks whether the 1970 and 1977 CAAs affected manufacturing activity and employment both for new and existing plants. Under the 1970 and 1977 CAAs, SIPs were to control existing sources in NA areas; hence the analysis captures the impact of controls on existing plants, and on both plant births and deaths.48 Each model controls for the impact of NA status for a particular criteria pollutant, holding constant NA status for other pollutants. The results are most pronounced for CO and ozone non-attainment status. For plants in high-emitting industries in NA counties, NA for CO is associated with statistically significant declines in employment (16 percent) and the value of shipments (15 percent), both measured over a five-year period. Ozone NA status is associated with a statistically significant decline in employment of 4.9 percent for plants in high-emitting industries. Employment effects for ozone NA are largest in the pulp and paper, iron and steel, printing and plastics industries, as well as stone, clay and glass, ranging 48 Of the 1,737,753 plant observations across 4 periods, 29% represent births, 27% deaths and 44% stayers. Looking Back at Fifty Years of the Clean Air Act 34 from decreases of 7 percent to 11 percent over a five-year period. CO regulation effects on employment are largest in iron and steel (-18 percent) and petroleum refining (-13 percent). The implications of these estimates for the number of jobs lost are that environmental regulations resulted in a loss of 591,000 jobs over the period 1972-89 at high-emitting plants in NA counties (39,000 jobs per year). To put this in perspective, total annual manufacturing employment was 17.4 million during the 1967-72 period. As Greenstone acknowledges, it is not possible to say whether some of the jobs lost in NA counties went to attainment counties. The corresponding figures for the declines in at high emitting plants in NA counties are $37 billion and $75 billion (1987$), for the capital stock and the value of shipments, respectively. Both figures represent declines, over the 1972-87 period, relative to plants in attainment counties, but are not significantly different from zero.

...

Walker (2013) uses a triple-difference estimator to capture the impact of NA status on employment and earnings.... In Walker’s preferred specification, the average worker in a newly regulated plant experiences a present discounted earnings loss equal to 20 percent of annual preregulatory earnings (over a 9-year period, using a 4 percent discount rate). In the aggregate this loss is $5.4 billion, although there is great heterogeneity in the pattern of losses. Workers who remain with their pre-regulation firms suffer essentially no losses. Losses are born by workers who change firms—especially older, higher-paid workers. On average, workers who change firms suffer an earnings loss equal in present value to 120 percent of their pre-regulation annual earnings. Within this group, workers who change industries suffer larger losses than those who remain in the same industry. 

...

by Joseph E. Aldy, Maximillian Auffhammer, Maureen L. Cropper, Arthur G. Fraas, and Richard D. Morgenstern

Resources for the Future (RFF) www.RFF.org 

Working Paper (20-01); January 6, 2020