The DOT 111 as a Best Technology Problem: RIAs, NEPA & Risk

NEPA's approach to risk is much more public than the average Regulatory Impact Analysis.

Moving crude or ethanol by freight train is dangerous.  Trains that are bigger or faster are more dangerous.  And bigger, faster trains in populous areas are the most dangerous.  That much can be said with confidence.  Beyond these generalities, things get contentious.

In an earlier post we noted that the Department of Transportation Notice of Proposed Rulemaking (NPRM) released in August proposing to phase out and/or retrofit the DOT 111—the standard tank car carrying so much of the Nation’s Bakken crude oil and refined ethanol today—had an after-thought of a NEPA document.  Media attention to this problem has kept up.  {See, e.g., this Pro Publica piece.}  But the problem is complex and it’s currently rutted in an opaque rulemaking procedure.  In this post we argue that DOT could’ve invested more in its NEPA analysis and that, if it had, we might have drawn some better conclusions about this problem.

A Tank Car Problem?

A series of derailments and fires since the oil-by-rail boom began in 2009 have shown the DOT 111 to be a substandard technology given its susceptibility to failure in the event of a crash.  DOT’s current rulemaking, because it’s governed by Executive Orders 12866 and 13563, included the preparation of a “regulatory impact analysis” (RIA) consistent with OMB’s standards thereon {See Circular A-4}.  The RIA was extensive and, as OMB requires, it sought to monetize expected benefits and costs from the options being weighed.  See NPRM, 79 Fed. Reg. at 45063-68.

Monetizing costs and benefits for RIAs has a long and sordid history which mostly demonstrates just how hard it is to monetize most of the things that really matter.  {See this excellent long-term assessment of RIA methodology from Resources For the Future.}  The Office of Information and Regulatory Affairs (OIRA) at OMB takes the lead here, a role for which it’s often vilified.  {See this commentary by a recent OIRA chief.}

It doesn’t help that DOT got a late start.  For one thing, the National Transportation Safety Board (NTSB) had found back in 2009 that the DOT 111 “can almost always be expected to breach in the event of a train accident resulting in car-to-car impacts or pileups.”  See 79 Fed. Reg. at 45026.  So NTSB recommended that the DOT 111 be phased out of service for the hauling of crude oil, ethanol, and other like flammables.

Secondly, the rail industry—well positioned to give plausible cost/benefit estimates—has been working to scuttle the rulemaking.  An Association of American Railroads (AAR) task force—ominously named “T87.6”— recommended some retrofitting of the DOT 111s (but very little) and a market-driven phase-out.  But those recommendations came in a report arguing that failing tracks and other infrastructure were the real causes.

Causes, Costs & Benefits: What are the Alternatives in Crude-by-Rail?

What is the best way to tackle this problem?  Upgrading the train cars hauling these dangerous payloads?  Improving our rail infrastructure?  A law forcing the DOT 111 out of service quickly would be costly.  And continuing its use indefinitely seems like inviting more tragic accidents: trains derail, often when they are travelling at high speeds with dangerous cargoes.

This is a dilemma that agencies like EPA confront all the time.  From EPA’s perspective, it is a “transitions” question at heart: how best to balance distinct factors or goals like fairness to those heavily invested in current technology, fairness to those bearing the risks of that technology, overall “social” costs, etc.?  DOT identified 27 such factors in its NPRM.  {See NPRM, 79 Fed. Reg. at 45029 (Table 10).}

Notably missing from DOT’s NPRM is the agency’s statutory mandate.  Careful thought often goes into a statute’s selection of factors, the agency’s interpretation thereof and the technical analyses balancing competing goals, etc. (inconclusive as all of those things often remain).  See, e.g., Entergy Corp. v. Riverkeeper, Inc., 129 S. Ct. 1428 (2009).

But for this problem the Department’s authority is incredibly broad.  It must/may “prescribe regulations for the safe transportation, including security, of hazardous material in intrastate, interstate, and foreign commerce” and say what constitutes a “hazardous material.”  {49 U.S.C. § 5103}  Nothing else constrains or guides the agency statutorily.

Like most risk problems, hazard definition is much of the challenge here.  Depending on one’s perspective, the “threat” behind the tragic recent derailments (and, thus, projected future derailments) include train speed, human error, faulty tracks, faulty cars, faulty brakes, urbanized settings, flawed emergency responses, America’s addiction to oil—or all of the above (and more).

Hazard definition decides which costs and benefits seem relevant.  For example, capping train speeds impacts railroads more than forcing new cars into service: shippers own most of the cars and railroads own most of the tracks.  Most importantly, the frequency and magnitude of expected losses/harms—where they remain uncertain—spur methodological disputes that often cannot be resolved before decisions come due.

NEPA v. OIRA: Alternatives for Whom?

DOT’s RIA aligned the significant choices in terms of car design options, perhaps because speed limits, routing, and braking technologies leave open too many possibilities of failure.  The risk analysis done to estimate the puncture resistance of different car designs was grounded in frequentist statistical methods.  It is open to reasonable disagreement, which industry has pressed to the limit at OIRA since the rule was proposed.  To many, meetings like this are the antithesis of truly public-regarding risk regulation because they happen off-the-record and behind closed doors.  As Cass Sunstein has put it,

[f]or those who express such concerns, the essential problem is that businesses and others subject to regulation arrange a strong majority of meetings . . . . With regard to many regulatory actions, those who are in opposition, or seek to scale them back, meet with OIRA far more often than do those who support such actions and seek to make them more protective.

Sunstein acknowledges risks of “epistemic capture” within this imbalance.  Those who benefit from the status quo distribution of risks provide a disproportionate share of the relevant information (which they’ve already filtered).  Thus, a regulator with the best of intentions might still unwittingly cater to a thin slice of concerns instead of balancing everything in the public interest.

Take the uncertainty surrounding what’s being shipped.  Packing “groups” are used in DOT’s Hazardous Materials Regulations to denote the risks different cargoes pose.  Packing Groups I and II are supposed to be high and medium hazard substances, respectively—and whether “crude oil” belongs therein has suddenly become a hard question.  It probably depends on the crude and in a very particularized way.  {See this report that helpfully breaks out how different crude types have complicated the rulemaking.}  The oil industry argues that crude oil is “safe,” that it belongs in Packing Group III, and that overly cautious characterizations cost consumers needlessly.  If Bakken or Alberta crudes were so “safe,” though, we wouldn’t have seen the tragedies we have, especially at Lac Mégantic (which the Canadian Transportation Board found was as flammable as gasoline).

This is where a NEPA-driven exploration of impacts and alternatives might have been a big improvement over the RIA-driven analysis DOT did.  Had there been a real NEPA procedure instead of an afterthought, DOT would’ve begun with public participation in “scoping,” a public review of the regulatory possibilities.  Most importantly, the selection of options (alternatives) ultimately analyzed would’ve been done anticipating their scrutiny by a reviewing court—not OIRA.  That alternatives analysis would’ve confronted a statute, 49 U.S.C. § 5101 et seq., which delegates massive discretion to DOT.  In short, that exercise could’ve facilitated a broader review of risks and options, one that didn’t naturally bottleneck at OIRA or limit our horizons to what affected industries prefer/would tolerate.


Public or private?

NEPA’s approach to alternatives is one rooted in public, transparent information collection and sorting—at least to the extent that CEQ’s ideals for the scoping process are followed.  Keeping the public engaged in DOT’s crude-by-rail rulemaking is virtually impossible as it has been managed.   As the NEPA Lab has written about before, even the flawed EISs for the KXL pipeline show that its approval would offer benefits to few, environmental risks to many, and a further sunk investment in a socially costly technology (fossil fuels).  In short, that EIS made approving the pipeline a rather difficult proposition.

OIRA’s approach to alternatives is one rooted in a corporatist advantaging of organized groups able to locate the levers of power in Washington.  If a recent OIRA Administrator, Cass Sunstein, is to be believed, in fact, OIRA views its role under EOs 12866/13563 as facilitating meetings and information exchange for these groups.  That can give regulated parties undue influence, especially in technology selection choices.

In a final post, we will offer some conjectures on this rule’s future, especially given the decision to find that DOT’s preferred technology will have insignificant environmental impacts.

I teach environmental, natural resources, and administrative law at Penn State Law. Before teaching I was an enforcement lawyer at U.S. EPA. Along the way I've done work for environmental nonprofits and written a fair bit about NEPA.
No Comment

Leave a Reply