EPA’s Wood Stove Analysis is Smoke & Mirrors

May 05, 2014

EPA recently requested comments on proposed New Source Performance Standards (NSPS) for residential wood stoves. It projects that the benefits of the proposal would outweigh the costs by a factor of more than 100. However, in a comment filed on the rulemaking docket, we point out that EPA’s analysis justifying the proposal is flawed in ways that not only make these net benefit estimates suspect but violate Presidential requirements, and may undermine EPA’s determination of what reflects the “best system of emissions reduction.”

Two elements essential to a sound ex ante regulatory analysis are 1) consideration of alternatives, and 2) analysis of how sensitive outcomes are to influential assumptions and uncertainties. Further, to improve future analysis, the President has asked agencies to prepare for ex post analysis to measure how well their regulations are working. EPA’s analysis fails all these conditions.

EPA’s regulatory docket for this rule includes numerous technical support documents and memos, as well as a 222-page regulatory impact analysis, yet none of them present information on alternatives to the performance standard EPA proposes to establish. This is contrary to OMB guidance, which directs agencies to “consider analyzing at least three options: the preferred option; a more stringent option; and a less stringent one.” This is particularly important since EPA presents widely varying cost-effectiveness for the different types of wood stoves covered; the total cost-per-ton of fine particulate matter (PM2.5) emissions reduced ranges from $37 for forced air furnaces to $1,513 for pellet stoves. Incremental analysis at different emission limits is essential to inform decision makers who wish to meet statutory requirements and ensure the greatest environmental protection at the lowest cost.

The majority (98%) of EPA’s estimated benefits derive from avoided premature mortality that studies have associated with reductions in fine particles. These figures depend on projections of the emission reductions that would be achieved by the regulation, and a dollar benefit-per-ton that would be generated by those reductions. EPA’s estimates of emissions reductions are based on the unrealistic assumption that higher consumer prices for wood stoves will not affect demand, nor lead consumers to keep their old units longer, nor switch to other (possibly more polluting) appliances for heating their homes. This may be a particularly appropriate assumption for forced-air furnaces for which EPA projects consumer prices will more than quadruple to over $4,000 per unit, and hydronic heating systems, for which it projects price increases of over $6,000 per unit.

Rather than relying on air quality modeling to estimate the health effects of the proposal (as EPA has done elsewhere), EPA estimates total benefits by multiplying a benefit-per-ton value to estimated emissions reductions. This “benefits transfer” approach relies on modeling conducted for other purposes to estimate the value of human health benefits associated with reducing one ton of PM2.5. Yet, several factors suggest that these national benefit-per-ton estimates will overstate benefits in this case. First, it is very likely that population density in the more remote areas where wood heaters are a common source of heat is significantly less than that of the cities on which the benefit-per-ton estimates are derived. Second, this approach does not take into account background air quality, which matters here because wood stoves tend to be used in remote areas where concentrations are often below the national ambient air quality standards. If the benefit-per-ton estimates are overstated because they don’t reflect the air quality conditions and populations relevant for the wood heater market, benefits are likely significantly overstated. Further, if EPA’s assumptions of a causal, linear, no-threshold relationship between PM2.5 exposure and premature mortality are inaccurate, as some have suggested, even the national average benefit-per-ton attributable to reducing PM2.5 would be less than estimated, and possibly as low as zero.

EPA estimates that total average annualized costs will be $15,688,471 between 2014 and 2022. Given the number of assumptions and degree of uncertainty involved in this estimate, the level of precision presented is misleading. EPA should conduct sensitivity analysis on these assumptions and generate a plausible range of effects (health effects, benefits, and costs) before proceeding with this rule. Further, particularly given all the assumptions and uncertainties in the ex-ante analysis, if EPA proceeds to issue these NSPS, it should commit to gathering information on actual outcomes, including emission reductions, air quality concentrations, population health, sales and prices of new residential wood heaters by type, and turnover of old units.

Without this, EPA’s regulatory analysis is little more than smoke and mirrors.