Federal Railroad Agency Way Off-Track in Approach to Safety

The Federal Railroad Administration (FRA) recently announced that Amtrak and select railroads around the country would be awarded $25 million to help implement positive train control technology. Positive train control (PTC) — mandated by the Railroad Safety Improvement Act of 2008 and subsequent regulations issued by the FRA — consists of adding GPS, computers and software to trains and railroad systems that automatically intervene to stop unsafe train movements. In theory, operator errors that lead to train collisions, derailments due to unsafe speeds or railway worker injuries would be a thing of the past, thanks to the automation of safety rule enforcement.

The idea of using automation to reduce accidents caused by human error has a proven track record across all modes of transportation. Nevertheless, the FRA recently proposed a rule requiring that train crews consist of at least two people, which largely ignores the potential safety gains from automation as well as investments in capital and maintenance. While it may be tempting to assume that two is always better than one, a more careful analysis of the historical causes of safety improvements in rail transportation indicates that track and equipment expenditures are much more important to safety than crew size.

The FRA's misguided and disjointed regulatory policy not only risks hindering further development of innovative technologies that can deliver greater safety, but it could also create a less-safe operating environment by deflecting scarce resources from known safety-enhancing uses like track maintenance to unproven uses.

Although one-person crews are currently relatively rare, the FRA believes that one-person crews are inherently more dangerous than two-person crews. In fact, the FRA believes this so fervently that it used the phrase "FRA believes" or "FRA further believes" at least 67 times in its documents proposing the new regulation. The problem, of course, is that regulating based on a belief does not necessarily lead to positive outcomes, such as improved safety, and the FRA offers no evidence to substantiate this belief.

Even worse, the proposed rule may not simply fail to achieve any improvements in safety; it may actually reduce safety and hinder the development of better, safer technology and equipment.

Still, it's easy to understand some of the FRA's logic. After all, why wouldn't two people create a safer train operation, when compared to just one? Why examine statistical evidence, when that would just slow down the implementation of a rule that is a self-evident win for safety?

Indeed, the FRA didn't let a little thing like data get in the way. As it noted in its proposed rule, the FRA "does not currently collect sufficient data related to the size of a train crew nor do accident reports investigations generally address the size of a crew in order for [the] FRA or any entity to definitively compare one-person operations to multiple person operations." The agency proposed the rule anyway, heedless of whether there was an issue in the first place or whether the new regulation might create unintended consequences that are even worse than the "solution" brought by the proposed rule.

There are several flaws with the FRA's approach: For one, consider the role of the human operator in other modes of transportation. Many decades ago, the introduction and widespread adoption of autopilot precipitated a 90 percent reduction in the pilot-attributable crash rate, according to a recent report. Automobile safety has likewise improved because of the intelligent-driver technology introduced in recent years, such as lane-departure warnings, pedestrian detection and automatic emergency braking.

The advent of completely driverless cars offers even more promise, primarily by reducing human error — which, according to the U.S. Department of Transportation, accounts for 94 percent of crashes. By cutting back on operator error, driverless cars could save as many as 21,700 lives and reduce the number of crashed by over 4 million annually, according to a study from the Eno Center for Transportation.

Similarly, railroad safety has dramatically improved over the past few decades. The total number of train accidents on the systems of the major freight railroads fell from over 11,000 in 1978 to 1,867 in 2013, even while revenue ton-miles — a metric of train usage — doubled. In a study that I recently published with my colleague Jerry Ellig, we found that most of this tremendous gain in safety is attributable to the removal of economic regulations that deterred investment in equipment and track and improvements in operational practices.

The demonstrably positive relationship between safety and investment in equipment and track maintenance raises the issue of unintended consequences. Any regulation — even a safety regulation — that deters investment in those areas that have driven the improvements in safety for decades could have the perverse effect of increasing the accident rate. By requiring a greater expenditure on labor (despite lacking any evidence that such expenditures would improve safety), the proposed rule could financially constrain some railroads from making other safety-improving investments.

As I wrote in a public interest comment filed with the agency:

"Any credible estimation of the net effect of the proposed rule would need to consider losses to safety caused by an induced diminution of track and equipment maintenance or other safety-enhancing investments. Given the proven record of maintenance and infrastructure investments on safety rates ... this proposed rule may not only be ineffective in reducing accident rates, but it may also actually increase the net accident rate."

The FRA's approach in proposing the train crew size rule exactly contradicts the logic of the most significant regulation issued by the FRA in decades: PTC. While PTC represents an effort to reduce human error by automating the enforcement of speed limits and other operational rules, the FRA's crew size rule could deter further automation.

Like auto accidents, most railroad accidents that occur these days are primarily caused by operators' mistakes or oversight. A regulation that permanently requires a minimum crew size of two — especially when there is no evidence that one-person crews are less safe — can only stand in the way of further reductions in accidents caused by human error.