Using its regulatory review authority, the White House Office of Management and Budget (OMB) substantially weakened an EPA proposal to protect the trillions of fish and aquatic organisms that are sucked up and killed each year by power plants that use rivers, estuaries, and oceans to cool their systems. As originally prepared, EPA’s proposed rule sought to require the 59 largest plants in the most ecologically sensitive areas to meet the performance achievable by a closed-cycle cooling system, which reduces fish kills by up to 98 percent by recirculating or reusing water, withdrawing only 2 percent to 28 percent of the water used by older systems. This would have honored the Clean Water Act’s requirement to set the standard based on the “best technology available” -- which clearly is a closed-cycle system, currently in use by 69 facilities. EPA sought less stringent requirements for the roughly 500 remaining plants subject to the rule. Yet citing costs, and ignoring the requirements of the law, OMB’s Office of Information and Regulatory Affairs (OIRA) -- which must give its consent to major rules -- stripped out the closed-cycle standard for the most harmful plants, according to EPA documentation required by Executive Order 12866. Instead, OIRA embraced alternative, less protective measures urged by energy companies -- including Cinergy, Edison Electric, and Public Service Electric & Gas (PSE&G), among others -- which EPA published as a proposed rule for public comment on April 9. Specifically, OIRA substantially lowered the performance standard to require a 60 percent reduction in entrainment (from a baseline of no controls) -- where fish are sucked into the plant, resulting in the vast majority of kills -- and an 80 percent reduction in impingement -- where fish are caught and stranded on a plant’s intake screen. Moreover, under further OIRA changes, facilities would be allowed to plead special circumstances with state permit authorities, and avoid meeting even this watered-down standard. “This weak mandate would allow existing plants to kill 20 to 1000 times more fish per megawatt than [plants with closed-cycle systems], and continue to decimate aquatic life in U.S. waterways indefinitely,” according to Riverkeeper Inc., which brought suit against EPA resulting in a 2000 consent decree requiring a standard. The reduction targets pushed by OIRA are based on the performance achievable by lesser technologies, including better intake screens and fish return systems. In addition, OIRA altered EPA’s original proposal to allow plants to forgo even these minimal upgrades by instead employing “restoration measures,” such as fish restocking programs and the restoration of wetlands and aquatic habitat “that will result in increases in fish and shellfish in the watershed.” Yet such measures have proven highly unreliable, and don’t actually replace the aquatic life killed -- there is no requirement to put back the same amount, kind, age, etc., of fish killed, and indeed, this would be impossible. It appears that OIRA’s changes were driven by cost-benefit analysis, and its desire to maximize “net benefits” -- costs minus monetized benefits. In writing other standards based on the “best technology available,” EPA has never employed such a cost-benefit test. Indeed, Congress gave EPA no instruction or authority to base its decision on cost-benefit analysis, as it clearly was more concerned with solving what it saw as a significant problem. Rather, since the 1970s, EPA has sought to ensure that the costs of the technology are not “wholly disproportionate” with the benefits, imposing an “impracticable and unbearable burden.” Clearly, a standard based on a closed-cycle cooling system passes such a test. In fact, EPA found “net benefits” of $65 million for a closed-cycle standard based on location and the amount of water withdrawn. Nor are the costs “impracticable and unbearable.” For an option similar to its original proposal, EPA estimates $595 million in compliance costs (and $848.86 million in benefits). This sounds expensive to be sure. But when spread out over more than 500 facilities in a multi-billion dollar industry, the costs don’t seem so scary. For instance, “Of the 131 unique parent entities that own the facilities subject to this rule, 108 entities would incur compliance costs of less than .5 percent of revenues,” according to EPA. Only three would incur costs greater than 3 percent of revenue. These costs would likely be passed onto customers, raising electric bills by about 28 cents a month where closed-cycle systems are installed, based on a recent study done in developing the rule. Nonetheless, OIRA sided with the less protective option that showed fewer benefits, but greater “net benefits” by EPA’s estimates. Yet as EPA points out, there are significant benefits it was unable to monetize, and thus are not captured by a net benefits calculation. For instance, EPA’s benefit estimate does not incorporate losses of invertebrate species, such as lobsters, crabs, and shrimp, as well as endangered or threatened species, nor does it consider the interrelationships of the species affected. Rather, EPA’s estimate is based exclusively on the commercial value of the fish that would have been caught in the absence of impingement or entrainment. This accounts for less than 20 percent of the total fish killed by cooling systems. In comments on the proposed rule, Frank Ackerman, a Tufts University economist, pointed out the absurdity of basing decision-making on these monetized estimates, as OIRA has insisted upon. “In the private sector, a balance sheet that weighs all of the company’s income against some of its expenditures does not provide a useful picture of the company’s true financial condition,” he wrote. “Likewise, in the public sector, a comparison of complete costs and incomplete benefits does not provide an accurate picture of the net benefits to society.” Accordingly, he attempted to incorporate missing benefits into EPA’s figures, first factoring in “non-use” ecological benefits based on available literature, and second by accounting for those fish not expected to be caught, based on the premise they are needed to perpetuate future commercially valuable fish. Incorporating these adjustments yields “net benefits” drastically different from those found by EPA. Indeed, of the six options considered by EPA, its preferred option -- the least protective among them -- scores the worst with Ackerman’s adjustments (where before it scored the best). Options incorporating a closed-cycle system all score better, and an option based on dry cooling systems (where no water is used) best of all. Unfortunately, it seems unlikely that OIRA will listen to these findings. Contrary to the Clean Water Act, its clear concern is cost to energy companies. Further details are available below under the following categories:

• The Problem
• EPA’s Original Proposal
• OMB Changes
• Cost Considerations Preeminent
• Undervaluing Benefits
• Improving EPA’s Estimates Changes the Conclusion
• Background

The Problem Each year, electric generating plants withdraw more than 70 trillion gallons of water for cooling purposes, killing trillions of aquatic organisms in the process, including plankton, crustaceans, shellfish, sea turtles, marine mammals, as well as fish and their eggs. Larger fish and shellfish are often trapped on a plant’s intake screen, and die there from lack of oxygen and movement, in what’s known as impingement. Smaller fish or eggs that make it into the cooling structure are usually killed in the cooling process, called entrainment, which generally results in much higher losses than impingement. In the Delaware Bay, for instance, EPA estimates staggering annual losses of 359.4 million fish from the Salem Nuclear Generating Station (which withdraws more cooling water than any plant in the country), 356.3 million of these from entrainment. At the Big Bend facility in Tampa Bay, Florida, an estimated 8.13 million fish are lost annually. And in the San Francisco Bay Area, more than 400,000 threatened or endangered fish are thought to be lost annually from two plants, Pittsburg and Contra Costa. EPA’s Original Proposal The good news is that technology exists that could cut fish losses by more than 90 percent. A closed-cycle cooling system, which is already in use by at least 69 facilities, recirculates or reuses water, withdrawing only 2 percent to 28 percent of the water used by older systems. Almost all new plants built over the last decade use closed-cycle cooling systems, though EPA only made this a requirement for new plants this past year. Moreover, a number of older facilities of various size and fuel type have retrofit closed-cycle technology. EPA sought to require this same sort of upgrade at 59 large plants in the most ecologically sensitive areas to meet the performance achievable by a closed-cycle cooling system. This would have represented a disappointment for Riverkeeper, which urged a closed-cycle standard for all of the 539 plants that withdraw water for cooling, but it was far preferable than the final result. OMB Changes In December of last year, OIRA received EPA’s proposed rule for review. Citing costs, OIRA objected to requiring a closed-cycle cooling system for any plant, no matter how many fish are being killed, according to EPA’s publicly available documentation of OIRA changes. Instead, OIRA embraced alternative, less protective measures favored by energy companies, including Cinergy, Edison Electric, and Public Service Electric & Gas (PSE&G), among others, which met with OIRA officials on Feb. 8 to press their case (a day after OIRA met with Riverkeeper). Most significant, under changes forced by OIRA -- which conspicuously mirror recommendations submitted to EPA by PSE&G -- the proposed rule:

• Sets weak performance standards. As stated above, EPA originally sought to require upgrades at 59 of the largest, most harmful plants to meet the performance achievable by a closed-cycle cooling system, which EPA says can reduce fish kills by up to 98 percent. Indeed, the Clean Water Act’s Sec. 316(b) requires that “cooling water intake structures reflect the best technology available for minimizing adverse environmental impact,” which would seem to dictate a closed-cycle cooling system -- by any measure, the “best technology available.” Yet OIRA vetoed this requirement, and instead lowered the performance standard to require only a 60 percent reduction (from a baseline of no controls) for entrainment at any facility operating at more than 15 percent capacity or that impacts more than 5 percent of the aquatic life in the area -- with no entrainment standard for those with smaller impact -- and an 80 percent reduction from impingement at all plants. In comments on EPA’s proposed rule, Riverkeeper argues, “This weak mandate would allow existing plants to kill 20 to 1000 times more fish per megawatt than [plants with closed-cycle systems], and continue to decimate aquatic life in U.S. waterways indefinitely.” These standards would allow for less reliable measures, such as better intake screens and fish return systems. Nothing would have to be done about the amount of water withdrawn, which is the critical factor in entrainment.
• Allows plants to avoid compliance by pleading special circumstances. Under changes pushed by OIRA, a plant can avoid the rule’s performance standards if it can “demonstrate” that compliance costs are “significantly greater” than EPA’s projections, or they significantly exceed benefit projections. If such a demonstration is made, state permit writers (who are chiefly responsible for enforcing the Clean Water Act) are to make a “site-specific determination” to devise a plan based on less costly measures. EPA originally considered this approach, which was pushed by energy companies, but rejected it as too unruly and burdensome on states, a conclusion overruled by OIRA. “Since even environmentally sympathetic regulators lack the resources needed to rebut, or in most cases fully evaluate, these arguments, [EPA’s proposed rule] will allow applicants to continue to obstruct and delay needed technology upgrades,” Riverkeeper argues. Moreover, this formulation turns the Clean Water Act on its head. Rather than focus on minimizing the impact on the environment through a uniform, national standard, as the law calls for, EPA’s proposed rule is largely concerned with minimizing impacts on power plants. EPA’s original submission to OIRA allowed facilities to appeal for alternative, less stringent measures, but made a point to clearly state, “The burden is on the person requesting the alternative requirement to demonstrate that alternative requirements should be imposed.” Moreover, EPA did not present this as a primary way to achieve compliance, as the rule approved by OIRA now does. With OIRA’s change, EPA’s proposed rule appears to invite a return to past (and current) practice, in which, absent a national standard, regulation of cooling water intake structures is done on an ad hoc basis by state permit writers. Yet the whole reason EPA is doing a standard is that this approach hasn’t worked.
• Allows “restoration measures” in place of preventing fish kills. Pressed by OIRA, EPA’s proposed rule allows facilities to forgo upgrades that would prevent fish kills in the first place by instead employing “restoration measures,” such as fish restocking programs and the restoration of wetlands and aquatic habitat “that will result in increases in fish and shellfish in the watershed.” While restoring habitat destroyed by human activity is worthwhile, these measures don’t actually replace the aquatic life killed by power plants -- there is no requirement to put back the same amount, kind, age, etc., of fish killed, and indeed, this would be impossible. Moreover, where it’s been tried, restoration measures have proven extremely unreliable, subject to high failure rates. For instance, the Salem Nuclear Generating Station, which is owned by PSE&G, has used habitat restoration measures most prominently. Yet according to Riverkeeper, “PSE&G has been unable to demonstrate that its restoration effort for the fish kills at its Salem plant is providing any benefit whatsoever to the fish populations of the Delaware estuary.” In fact, EPA has found no evidence where restoration measures have successfully replaced fish killed by a power plant. Yet under the Clean Water Act, such measures must be successfully employed by at least one facility before being incorporated into an EPA standard. Again, clean water standards are to be based on the “best technology available,” and cannot be based on experimental technology, which absent any evidence, clearly applies to restoration measures. EPA’s original proposal permitted facilities to “supplement” technology and operational upgrades with restoration measures only if the plant could demonstrate environmental improvements “to a level commensurate with that which can be attained by a closed-cycle recirculating system.” In other words, EPA originally did not permit restoration measures as a replacement to upgrades, as reflected in OIRA’s changes, and restoration measures could only be considered in the context of meeting the highest possible performance standard.

Cost Considerations Preeminent According to EPA, the proposed rule “would establish a cost-benefit test that is different from the ‘wholly disproportionate’ cost-benefit test that has been in use since the 1970s.” EPA’s statement is surprising in that it elevates cost-benefit analysis above its statutorily set requirement. As stated above, the Clean Water Act demands standards based on the “best technology available.” Its provision on cooling water intake structures gives EPA no instruction or authority to base its decision on cost-benefit analysis. Indeed, a review of the legislative history shows that Congress clearly understood such a standard could impose burdens, even resulting in the closure of plants, but reasoned that solving the problem was more important than the costs it might impose. In writing other standards based on the “best technology available,” EPA has never employed a straight “cost-benefit test” in deciding how to proceed. Rather, the agency has sought to ensure that the costs of the technology are not “wholly disproportionate” with the benefits, imposing an “impracticable and unbearable burden.” Clearly, a standard based on a closed-cycle cooling system -- the best technology available in this case -- passes such a test. As part of its cost-benefit analysis, EPA presented estimates for various regulatory options that would incorporate a closed-cycle standard. EPA found “net benefits” -- which excludes qualitative benefits -- of $65 million for a closed-cycle standard based on location and the amount of water withdrawn. For a two-track standard requiring a closed-cycle system for some facilities and allowing others to appeal for site-specific approaches, EPA found “net benefits” of $280 million. Thus, by EPA’s monetized estimates, the benefits of a closed-cycle standard are in proportion with -- and in fact significantly exceed -- costs. Nor are the costs “impracticable and unbearable.” EPA estimates $595 million in compliance costs (and $848.86 million in benefits) for a standard similar to its original submission to OIRA, setting a closed-cycle cooling standard the largest, most harmful plants located on oceans, as well as less stringent requirements for the roughly 500 smaller facilities. This sounds expensive to be sure. But when spread out over more than 500 facilities in a multi-billion dollar industry, the costs don’t seem so scary. In fact, EPA estimates that compliance costs would be low compared to total revenue: “Of the 131 unique parent entities that own the facilities subject to this rule, 108 entities would incur compliance costs of less than .5 percent of revenues; 12 entities would incur compliance costs of between .5 and 1 percent of revenues; 6 entities would incur compliance costs of between 1 and 3 percent of revenues; and three entities would incur compliance costs of greater than 3 percent of revenues.” Nonetheless, for industry and OIRA, this was too expensive. Instead, OIRA insisted that EPA propose a less protective, less costly standard, which it contended would produce greater “net benefits” -- monetized benefits minus monetized costs. This emphasis on “net benefits” ostensibly puts costs on an even par with benefits, which is not what Congress had in mind with its “best technology available” standard, nor is it how EPA has done its water standards in the past. Yet given severe limitations in EPA’s benefits estimates, as explained below, costs are in fact given considerably more weight. Undervaluing Benefits For the proposed rule, as shaped by OIRA, EPA estimates costs of $283 million and benefits of $735 million for “net benefits” of $452 million, exceeding its estimated “net benefits” for all other approaches. Again, Congress directed EPA to adopt the “best technology available” given the seriousness of the problem, which may not necessarily be the most “cost-effective” option. Yet leaving aside the issue of congressional intent, there are fundamental problems with OIRA’s “net benefits” decision-making. Most glaringly, EPA’s monetized benefits were based exclusively on the commercial value of the fish that would have been caught in the absence of impingement or entrainment. This accounts for less than 20 percent of the total fish killed by cooling systems. Yet this is only one small benefit of the proposed rule, and itself is understated, as “facility-provided monitoring data (the basis for EPA’s analysis) typically focus on only a subset of the species impacted…” As EPA forthrightly admits, it did not take into account losses of invertebrate species, such as lobsters, mussels, crabs, and shrimp, even though these are commercially valuable. Nor did the agency factor in effects on fish-eating birds, or impacts on marinas, bait sales, or property values. Moreover, EPA’s estimates do not incorporate the value of threatened or endangered fish -- which may not be fished commercially -- or the interrelationships of the species affected. Thus, on a general level, EPA’s benefit analysis completely misses the mark. The proposed rule is not about preserving commercially valuable fish, but protecting our aquatic ecosystems. Of course, measuring the benefits of this societal goal in terms of dollars and cents is not only impossible, but also absurd on its face, exposing the weakness of cost-benefit decision-making. Conversely while benefits are difficult to measure, costs are all about dollars and cents, and easily incorporated into such an analysis. In this case, they involve expenditures on better intake screens, fish return systems, and other upgrades in equipment. This leaves benefits drastically undervalued in comparison to costs, which unfortunately didn’t stop OIRA from forcing a decision based on these numbers. Improving EPA’s Estimates Changes the Conclusion “In the private sector, a balance sheet that weighs all of the company’s income against some of its expenditures does not provide a useful picture of the company’s true financial condition,” according to Tufts University economist Frank Ackerman in comments to EPA on the proposed rule. “Likewise, in the public sector, a comparison of complete costs and incomplete benefits does not provide an accurate picture of the net benefits to society.” Accordingly, Ackerman attempts to incorporate missing benefits into EPA’s figures, performing two adjustments:

• First, based on the available literature, he assumes that “non-use” benefits are two times greater than benefits derived from recreational or commercial uses, which EPA’s estimates are based on. “Environmental economics increasingly recognizes the importance of non-use values: people place a substantial value on the mere existence of animals, ecosystems, wildernesses, and unique natural locations, quite apart from any past, present, or planned future use of those aspects of nature,” Ackerman writes.
• Second, Ackerman attempts to value total fish mortality. As stated above, EPA does not account for more than 80 percent of the fish killed by impingement or entrainment, because they would otherwise not be caught for commercial use. “The nonlanded fraction of these species -- the ones that survive uncaught -- have an obvious ecological value,” Ackerman writes. “If nothing else, their reproduction is the source of the catch in future years; that is, they are essential to the creation of future recreational and commercial values... I have conservatively assumed that nonlanded fish have a value equal to .25 times the value of landed fish of the same species.”

Incorporating these adjustments yields “net benefits” drastically different from those found by EPA. Indeed, of the six options considered by EPA, its preferred option -- the least protective among them -- scores the worst in terms of “net benefits” with Ackerman’s adjustments (where before it scored the best). Options incorporating a closed-cycle system all score better, and an option based on dry cooling systems (where no water is used) best of all. In the end, Ackerman suggests that EPA avoid getting bogged down in these valuation questions and cost-benefit decision-making. Instead, he recommends that EPA return to the intent of the Clean Water Act and examine the issue on its most basic level. “Any regulation will impose some costs; power plant operators will undoubtedly pass those costs on to their customers,” Ackerman writes. “So the question that ultimately matters is, are ratepayers willing to pay the increased costs imposed by regulation? As Synapse Energy Economics has estimated in their comments on this rule, if all costs were passed on to the consumer, an all [closed-cycle] cooling tower rule would cost each ratepayer 28 cents per month on their electric bills. If the public was asked, 'Are you willing to pay 28 cents more per month on your electric bill to avoid massive fish mortality and other underwater environmental damages caused by power plants?', I feel confident the answer would be 'yes.'" Unfortunately, OIRA says no. Background Under section 316(b) of the Clean Water Act, EPA must set standards for steam electricity producing power plants that operate cooling water intake structures, which suck in water for cooling purposes. These standards are to be based on the "best technology available for minimizing adverse environmental impact." In 1976, EPA published a rule to address section 316(b) and cooling water intake structures. In 1977, 58 electric utility companies challenged the regulations on the basis that EPA had failed to comply with the Administrative Procedure Act, which governs regulatory procedure, including public notice and comment. The court sided with the power companies, throwing out the rule for these process considerations, but made no determination on the merits of the rule. In 1977, EPA issued non-binding guidance for limiting the adverse impact of cooling water intake structures on the aquatic environment, which remains in effect to this day. After a series of negotiations and subsequent failures to promulgate adequate regulations in the 1990s, Riverkeeper Inc. sued EPA, and the two parties eventually signed a consent decree in 2000, with EPA agreeing to complete final regulations by the end of 2004.