A Brief Rant on Cost-Effectiveness Analysis

Posted on October 2, 2015 by Seth Jaffe

On September 29, 2015, the 3rd Circuit Court of Appeals remanded EPA’s approval of Pennsylvania’s regional haze SIP.  Although I think that the decision was important and largely unobjectionable, it did get one issue wrong, and it happens to be an issue near and dear to my heart – cost-effectiveness analysis.  I am regularly surprised by the number of people who oppose its use and the number of people who just plain don’t get it.  The 3rd Circuit (and EPA and the plaintiffs) fall into the latter category in this case.

The issue here was what metric to use in measuring the cost-effectiveness of technologies intended to reduce regional haze.  Since this case involves compliance with a rule intended to reduce haze, I would have thought it self-evident that cost-effectiveness would be measured by the dollars spent divided by the amount of haze reduced.  Silly me.

For those of you who don’t know, there is a measure of visibility; it is known as a “deciview.”  Pennsylvania, to its credit, indeed measured cost-effectiveness on a dollars/deciview metric.  The plaintiffs argued that cost-effectiveness should be measured on the basis of dollars/ton of pollutant removed.  EPA waffled, first agreeing with the plaintiffs, but then concluding that, while its guidelines call for $/ton, it is acceptable to use $/dv.  The Court, following the rule of decision that it must evaluate EPA’s decision based on the reasoning used in the rule, rather than on a rationale first provided in litigation, concluded that, because EPA’s own guidelines found the $/dv metric to be “flawed”, EPA’s approval of the $/dv metric was unjustified and must be remanded.

I am not sure I can count the ways this was screwed up, but let me put it simply.  If we’re assessing cost effectiveness, and we have a measure of the outcomes we care about, we should use it.  To use a proxy – emissions – instead of the actual outcome the rule is intended to affect – visibility – is just plain nuts.

And I have to add that, not only did the Court get this 180 degrees wrong, but it did not even seem to be aware of just how bizarre it is to reject the metric that actually measures the outcomes the rule is trying to achieve.

Comments (2) -

Eric Groten United States
10/2/2015 6:51:58 PM #

Just wait until the cost per future degree celsius saved is argued in the CPP litigation....uh, wait,  EPA doesn't even try to show any change in the metric that it uses to justify that entire enterprise.

John Hall United States
10/3/2015 11:33:15 AM #

Certainly agree; I have had more than a few cases where courts did not understand basic mathematical and statistical concepts, failing to get simple division correct is no shock either.  What annoys me more is that EPA allows this to occur so long as it supports their case, even when they know it is not correct.  That is not the Agency I worked for in the 80s.

Add comment

  Country flag
  • Comment
  • Preview