In elementary school, it wasn’t enough – and often not even the point – to get the right answer. Unless the teacher saw slash marks and numbers borrowed or carried over in your arithmetic homework, she couldn’t be sure that you knew what you were doing.
Apparently, the Federal Motor Carrier Safety Administration also needs to prove it knows what it is doing. During arguments over the latest hours-of-service regulations last month, Chief Judge Douglas Ginsburg of the U.S. Court of Appeals for the District of Columbia Circuit faulted the agency’s sloppiness in its regulatory analysis.
“Any number of times we have said ‘show your work,’ ” Ginsburg told government attorney Matthew Colette. People can’t really respond to conclusions if the agency doesn’t show how it arrived at them, Ginsburg said. Unsatisfied with Collette’s responses regarding crash risk after extended driving hours, Judge Merrick Garland declared, “You’re not a statistician. This is a debate among statisticians.”
That’s the problem. All the analyses FMCSA generates to support its decisions on regulations might as well be written in ancient Mayan as far as the average person is concerned. Only highly motivated parties – the American Trucking Associations or Public Citizen, for example – would even try to understand the crucial finer points and methodologies. And those parties already have prejudged the analysis as good or bad based on the rulemaking’s outcome.
Rarely is there a single, authoritative scientific consensus on anything. As Colette said at one point, “The studies are all over the map.” In drafting a rule, a government agency assigns more credibility to some scientific studies than to others with opposing conclusions. The agency hopes that its arguments are strong enough to satisfy most people and, especially, the federal appeals court.
FMCSA already failed to do that once in the hours-of-service rulemaking. The agency then had to consider driver health, and it relied heavily on research emphasizing the need for adequate, uninterrupted sleep on a daily basis. That’s why the August 2005 revision sharply limits split rest in sleepers and leaves controversial items like the 34-hour restart and the 11 hours of driving in place. Public Citizen, which deplores the latter provisions, counters with research emphasizing the degradation of safety as “time on task” – hours driving per day or week – grows. The fate of the current hours-of-service regulations may rest on which fundamental view the appeals court finds more persuasive.
Competing research isn’t the only complication. Sometimes bureaucrats engage in statistical gymnastics, bending and stretching the numbers so far that the result is truly incredible – meaning not at all believable. For example, the Clinton administration’s analysis of a proposed hours-of-service regulation in 2000 concluded that mandatory electronic onboard recorders would produce $3.2 billion in benefits over 10 years because drivers would save the estimated 1.5 minutes per day it takes to complete a paper log. Who knew a driver could accomplish so much in 90 seconds? The analysis was far murkier, however, on the costs to buy, install and maintain the devices.
It’s impossible to remove subjectivity totally, but perhaps regulatory analyses would be more consistent and believable if a competent and reasonably neutral party – maybe the Government Accountability Office – conducted or at least carefully reviewed them. Ongoing monitoring of FMCSA’s performance in both promotion of safety and holding down regulatory costs might be even more important. Does it make sense to fret over how regulations might affect safety and industry costs and then ignore the impact once the regulations take effect?
Presumably, regulators aren’t children. Perhaps we should worry a little less about them showing their work and a little more about whether they got the answer right.