Some better data this week suggests that Canadian Pacific Kansas City’s southern U.S. operating problems related to its May 3 IT systems cutover might be starting to correct, which is good news. Let’s walk through what we’re seeing.
Here’s full system terminal dwell (Canada, U.S., Mexico), with a peak of 11.4 hours in the week ending June 6, followed by an improvement to 11.1 hours through June 13. As you can see in the top chart, anything over 11 hours is poor for CPKC, but at least there’s a clear top forming.
If we switch to full system velocity, we actually see a recent deterioration over the past three reported weeks through June 13. In fact, that last week represented a four-month low. Given the cutover problems have been isolated to CPKC’s southern U.S. territory, we need to find data with more granularity. We’ve done so in recent weeks by looking at dwell at individual, impacted yards, as you can see in the bottom updated chart.
In the week ending June 13, the key yard at Shreveport, La., fell from 31 to 28 hours. We’ll take a 10% improvement, but it’s still a far cry from normal, which is about 18 hours for that facility. Similarly, the yards at Jackson, Miss., and Laredo, Tex., improved sequentially, but dwell at both remains well above normal levels.
Regarding adjacent territory we’re monitoring for contagion, to the north the yard at Kansas City remains unaffected, at least in terms of the data; while to the south the key Sanchez Yard in northern Mexico deteriorated for a second week to 37 hours, which is a seven-month high.
Other relevant data includes cars online specific to CPKC’s U.S. territory, which has drifted down (improved) from 61,000 to 59,000 in recent weeks, so again directionally encouraging but underwhelming in terms of magnitude.
Our teaser at the top of this article was new data that gives us some conviction that a rebound is beginning, and we’re specifically referring to this little nugget (top chart) in the CPKC reports to the STB for its U.S. territory. It’s another dwell-related measure that shows the count of cars in the U.S. geography that sit idle for 48-plus hours in any given week. It includes slow-to-move cars anywhere on the U.S. system, rather than just transiting the yards, and is a key measure of fluidity. Lower is obviously better. In this specific case (bottom chart) you can see four weeks of sequential degradation through the end of May, with slow-to-move cars more than doubling, followed by a modest improvement in the first week of June and then a more dramatic step down in the week ending 6/13. It’s this step down in this dataset that’s most encouraging in terms of bringing the cutover damage under control.
The obvious caveat is that it’s just one data point, so let’s see if it proves indicative.
