Transport: The Wrong Metric

TORONTO, ONTARIO - Long-term readers of this blog know that the author is very big on setting up proper incentive structures and proper evaluative metrics before letting the invisible hand of the market do its work in a given market. It appears that freight railroading may not be using the proper metrics and thus investors are being grossly mislead about which are the leading freight railroads in North America.

The Association of American Railroads has long published its railroad performance measures which focus on train speed and terminal dwell hours as a metric of how efficiently a railroad is operating. This seemed to make sense; if trains were operating faster and getting through yards faster, then they should be reaching customers faster and thus providing better service. On the financial side, the operating ratio was viewed as a key figure, which seemed to make sense--if a railroad is spending less money to earn more money, then that railroad would seem to be operating efficiently.

Of course, one of the ways to reduce operating costs is to reduce money spent on improving infrastructure, maintaining existing infrastructure, and safety. Neglecting such expenditures hurts the railroad in the long run, so operating ratio seems like a typical Wall Street metric--making money in the short-term at the expense of the long-term. Still, even Wall Street should eventually notice eroding profits from lost business, regardless of operating ratio.

Railfans have been becoming suspicious of these metrics in recent years, as Canadian National (CN), which owns the former Illinois Central, Wisconsin Central, and Grand Trunk Western lines amongst others in the United States as well as its traditional Canadian routes, was cleaning up on the metrics. CN had a 63% operating ratio in the third quarter of 2009, way below the industry average of around 75%. Operating velocity and terminal dwell time were also much lower on the CN than on primary competitor Canadian Pacific. Yet, when railfans observe the two rivals, CP--while far from perfect--seemed to be actually operating like a real railroad trying to deliver freight, and CN has just been bizarre to observe, with schedules rarely held to in many corridors (despite supposedly being a "scheduled railroad"), trains taking traffic well beyond its destinations, and other operating fiascoes. It seemed that CN was mainly making its statistics look good by running exceptionally long trains, at the expense of train frequency and service quality, leading to lost business. Especially suspicious was the fact that CN had re-defined some of the metrics so that the AAR would no longer directly compare the CN's performance with the other "Class I" railroads in North America.

This all came to a head recently in a recent Trains magazine article by noted writer Fred Frailey. According to that article (not available on-line), CN's on-time car delivery to customers is only 60%. This statistic is not easy to find for all railroads. However, buried in this filing related to a proposed merger that never happened, BNSF's figure in 1999--hardly a recession year--was 91%. In intermodal service, BNSF touts 99% on-time delivery, and considering that it offers guaranteed carload service in many service lanes, one has to assume they are achieving at least their 1999 levels, or the guarantee would be too expensive.

Railroads are a service industry. They provide the service of moving freight from one location to another. If they can't do that on their intended schedule, then they aren't doing their job. There's a reason that Warren Buffet chose to purchase BNSF and not Canadian National, and it isn't just that CN is nominally a Canadian company. For a performance metric, on-time delivery seems to be a more appropriate metric than system velocity or terminal dwell time. The AAR should start reporting it publicly and consistently across the major railroads.