UP, NS To Surface Transportation Board: ‘Trust Us’

By Jim Mathews / President & CEO

--

Earlier this week, Union Pacific and Norfolk Southern filed their answers to the Surface Transportation Board’s questions revolving around how their proposed merger would affect – or not affect – passenger rail, spending many dozens of pages to argue, unconvincingly, that the answer is “not much.”

The Applicants also declined to answer STB’s requests for modeling of these effects, essentially arguing that to do so would be too hard and take too long. But we’ll discuss that in a later essay…

One of my most enduring criticisms of the entire UPNS application has been that the two railroads have devoted hundreds of pages to explaining why the merger won’t adversely affect today's passenger trains, while devoting zero attention to explaining how their future freight growth plans interact with the future passenger rail network now being planned across the United States.

I’ll give credit where it’s due, and note that some of their positions are substantially more fleshed out than in their earlier submissions. And I don’t want to dismiss their statistical work just because it reaches conclusions that I might not like.

But there’s a “straw man” quality to it all that I think we all need to pay attention to. By zeroing in on the portion of the regulation that requires the Applicants to show that they won’t degrade passenger service, they’ve answered a considerably narrower question than the one that is ultimately before the STB. There’s statutory and regulatory support for including future passenger growth in their evaluation of the proposed takeover, which UP-NS are conveniently sidestepping.

Just as the two previous filings did, this supplemental answers a status quo question rather than a future-network question. Given the amount of federal and state investment currently being directed toward expanding passenger rail, that distinction strikes me as significant.

UP and NS submitted an analysis by a consultant economist, Kristof Zetenyi, meant to show that they can add freight trains to the combined network without making passenger service any worse than it is today. In particular, Zetenyi asserts that the data show no statistically significant relationship between host railroads’ route capacity utilization and Amtrak delays caused by the host railroads, instead claiming that “seasonal and route-specific effects” are mainly driving delays. He uses a fixed-effects regression model to account for these, contending that this makes any apparent relationship between delay and capacity utilization disappear.

But there are two problems here: one, his analysis fails to account for the possibility that capacity utilization is ITSELF also seasonal and route-specific, something he actually acknowledges indirectly in a footnote[1] explaining the decision to remove those factors. But the other, which he also acknowledges, is that he relied only on data UP and NS supplied for the analysis and used selected short segments to analyze the potential effects on the entire system. UP and NS chose a single "Selected Segment" for each passenger service — the growth-affected segment projected to have the highest post-merger capacity utilization — and used that segment as the basis for the analysis.

Does that methodology accurately capture where passenger delay actually occurs? My gut says no. Is one segment truly representative of the passenger service being measured? Could localized conflicts elsewhere along a route be masked by this approach? Is there any disconnect between a route-level performance metric and a single-segment capacity measure?

The kicker? UP and NS actually point to Zetenyi’s results on the Sunset Limited as evidence that the transaction poses no risks to existing passenger rail operations. Yes, THAT Sunset Limited, the one with on-time performance in June of around 47 percent. So they’re measuring whether an already poorly performing passenger service becomes detectably worse, rather than whether passenger service is operating at a level the Board should regard as acceptable in the first place, an especially egregious approach given that UP signed an agreement (so far confidential) promising better Sunset timekeeping in exchange for settling their case before the STB with Amtrak.

In their answer this week, UP and NS also repeatedly cite Amtrak's statutory dispatching preference and existing contractual performance obligations as evidence that passenger operations will be protected. Those protections are certainly important, but by themselves, they’re a pretty slender reed and sure don’t create railroad capacity or even assurance.

Decades of litigation, regulatory proceedings, and negotiated settlements involving passenger-train performance — including proceedings involving both Union Pacific and Norfolk Southern — demonstrate that statutory and contractual protections are far from self-executing. Their existence does not automatically translate into satisfactory passenger operations, especially when the two worst performers, who had to be dragged into litigation over passenger performance, are proposing a marriage.

This filing relies far too heavily on those legal protections in place of demonstrating that sufficient physical capacity actually exists to accommodate projected freight growth while preserving passenger operations. Forgive my snark, but my read is that UP and NS are saying in this section that “We promise that THIS time you won’t have to sue us to make us honor the law or our contracts!”

This filing approaches 500 pages, and our data-science and litigation partners are combing through it for more analysis and insight. There’s much more to say on this filing. But suffice it to say that while there’s more detail today in the docket than there was last week, there remain not a lot of answers or assurances that this transaction would be anything other than an impediment to passenger rail in the United States.