Part 4 in a continuing series on CBA -- First half of CBA chapter in study Now that the preliminaries are over, we can look at the detailed results from the "Placing SDWA Costs In Context" chapter of the CBO study. Before I present their results in detail, let's review a few quotes from the summary and present a few of the questions that you might want to think of when reading it: > ... the average cost per cancer case avoided (averaged for all system >sizes) is estimated at $500,000 for regulating the pesticide ethylene >dibromide and its co-contaminants compared with more than $4 billion for >regulating the pesticides atrazine and alachlor. ... Sounds like a lot for atrazine and alachlor, doesn't it? After all, no one would go into a hospital and spend $4 billion to save a single person from cancer. Keep this factoid in mind, we'll be back to it soon. For the moment, though, did anyone wonder what happened to the non-cancer hazards that SDWA is also supposed to prevent? Why did CBO only look at cancer cases? More on this in a moment. >reduce the risk of cancer is extremely uncertain. The EPA estimates that >the average cost per cancer case avoided ranges between $867,000 and >$8.7 billion in the initial stage of the rule and between $840,000 and $19 >billion in the extended stage. Now, $867,000 to $8,700,000,000 per case is a tremendous range of values. Where do people suppose this uncertaincy came from? Certainly not in the cost figures: people may not know the cost of building something to a factor of 2 or 3, but that's about it. And did any scientists out there wonder at the precision of three significant digits in that "$867,000" number, given that it's part of a four order of magnitude range? Keep an eye out for this pseudo-precision later. > The second most costly rule currently proposed is the Enhanced Surface >Water Treatment (ESWT) Rule, which is designed to prevent the outbreak of >waterborne gastrointestinal diseases. Using an estimate of medical costs >and lost wages as a measure of benefit, the benefits are expected to exceed >the average cost per case of waterborne disease avoided as a result of >complying with the proposed initial phase of the ESWT rule (averaged for >all systems). Strangely enough, though, when it comes to a case where there is an instant, noticeable effect -- like the onset of gastrointestinal disease -- our data seem to show that benefits exceed costs. We'll be getting back to this later, too. Below will be most of the "Placing SDWA Cost Estimates in Context" chapter of the CBO study. Vebatim quotes will be in double quotes; my own summary and paraphrase will be in brackets []. My own comments will be in a later post. The next post after this will have the second half of this section of the chapter. ------------------------------------------------------------------------- [The chapter starts with an introductory paragraph that again mentions the atrazine/alachlor number. The chapter itself is divided into three sections; cost/benefit, cost relative to other costs of supplying water, and cost comparing to local measures of fiscal capacity. I will only type in the first section.] "Costs Relative to Benefits Whether costs of a regulation are 'too large' or not depend, of course, on the benefits that result from the regulation. Regulations are generally thought to be too costly when the cost of complying with them exceeds the value of the benefits received. However, measuring costs and benefits can be difficult. Information on the benefits associated with the SDWA is limited. CBO used information available in EPA documents and applied a consistent method to calculate the cost per cancer case avoided from several carcinogens that are regulated, or proposed to be regulated, under the SDWA (see Table 4). The cost per cancer case varies enormously among contaminants." [Examples are then given.] [At this point, I'm going to quote part of Table 4, giving only the column that averages results for all sizes of water systems -- the chart in the report has multiple columns for each contaminant, showing values for various sizes of systems.] "Table 4: Cost per Health effect Avoided for Selected Rules and Contaminants (in millions of 1992 dollars) Carcinogens ----------- Radionuclides (proposed) Radon: MCL = 300 pCi/L Cases avoided (per year) 79.46 Cost per case avoided 3.6 Radium-226: MCL = 20 pCi/L Cases avoided (per year) 3.1 Cost per case avoided 14.2 Radium-228: MCL = 20 pCi/L Cases avoided (per year) 0.210914 Cost per case avoided 45.5 Adjusted gross alpha emitters: MCL = 15 pCi/L Cases avoided (per year) 1.4 Cost per case avoided 39.5 Volatile Organic Compounds (all combined) Cases avoided (per year) 27.99 Cost per case avoided 2.4 Synthetic Organic Compounds EDB and co-contaminants Cases avoided (per year) 71.833 Cost per case avoided 0.5 1,2 dichloropropane Cases avoided (per year) 0.1676 Cost per case avoided 24.6 Atrazine and alachlor Cases avoided (per year) 0.0024 Cost per case avoided 4,258.0 Noncarcinogens (acute effects) Surface Water Treatment Rule Cases avoided (per year) 83,194 Cost per case avoided 7.1" [For comparison, here is a chart from an earlier part of the CBO report, giving total costs for each rule under the SDWA:] Table 1: Annual Cost of Treatment According to Standards Specified by SDWA (pg.11) (in millions of 1992 dollars) [I used EPA estimates only; an industry group called AWWA also provided estimates, which were about 50% larger.] Existing Rules Proposed Rules --------------- -------------- Fluoride 7.5 Radionuclides Phase I VOCs 63.4 Radon 280 Surface Water Treatment Rule 549.1 Radium-226 48.6 Total Chloroform Monitoring 139.5 Radium-228 8.7 Phase II SOCs 106.4 Adjusted gross alpha 53.4 Phase II IOCs 15.2 Uranium 80.7 Lead and Copper 503.9 Disinfectants/Disinfection By-Products Phase V SOCs and IOCs 46.1 Stage I 1,064 ------- Stage II (lg. sys) 1,820 All Existing Rules 1,431.1 Stage II (all sys) 2,631 Enhanced Surface Water Treatment Interim Rule (lg. sys only) Stage I D/DBP 402 Stage II D/DBP 746 Long-term rule (all systems) Stage I D/DBP 519 Stage II D/DBP 927 Sulfate 80 [Now, back to the CBO chapter text:] " In addition, the cost per cancer case avoided generally declines (sometimes drastically) as the size of the system increases. For example, the cost per cancer case avoided as a result of regulating 1,2 dichloropropane falls from $135 million for the category for the smallest-sized systems to $13.2 million for systems serving between 10,000 and 25,000 people. That decline primarily takes place because unit treatment costs decrease as system size increases. Two of the proposed regulations are expected to reduce the risk of cancer: the Disinfectants/Disinfection By-Products Rule and the Radionuclides Rule. However, the EPA is uncertain about how much the risk of cancer would be reduced by the proposed D/DBP rule. Based on differing estimates about the baseline cancer risk associated with disinfection by-products, the EPA estimates that the average cost per cancer case avoided ranges between $867,000 and $8.67 billion in the initial stage of the rule. In the extended stage, the incremental cost per cancer case avoided is expected to be between $840,000 and $19 billion." [Footnote: EPA, _Regulatory Impact Analysis of Proposed National Primary Drinking Water Regulations: D/DBP Rule_ (May 1994), p. 5-7]. " In some cases, the proposed standards for radionuclides may result in extremely high costs per cancer case avoided. For example, the standard proposed for adjusted gross alpha emitters (which is primarily designed to reduce the risk of exposure to the radionuclide polonium-210) would cost more than $1 billion per cancer case avoided for systems in the smallest-sized category. Although considering the costs per cancer case avoided is useful, it is important to realize that those estimates are highly uncertain. They are based on the best available data, but those data are limited. For example, the occurance data used in estimating the cost per cancer case avoided for the SOCs are not based on a nationwide survey. Those data indicates where and at what levels a contaminant is expected to be found. A nationwide estimate of occurence was obtained by piecing together many sources of information (none of which was designed to be representative at the national level) and by using considerable judgement." [Footnote: EPA, Regulatory Impact Analysis...Regulations for SOCs (April 1989), p. 1-2.] "Given that uncertaincy, the actual costs per cancer case avoided could either exceed or fall below the estimates provided."