Friday’s City Comptroller report attempts to tell the story of how MTA management misled the public for years on the nature and severity of deteriorating subway service. The report, which the office was working on for about a year, mostly rehashes revelations made by reporters like myself, the Daily News’s Dan Rivoli, The New York Times’s Metro investigative team, and others. But there is a tidbit buried deep in the appendix that isn’t necessarily a revelation, but does more to encapsulate what has happened to the subway than the rest of the report combined.
In 2016, subway ridership was at its modern-day peak, with more daily riders than the system had had since 1948. But thanks in large part to the widespread installation of signal timers and other operational inefficiencies slowing trains down, delays were mounting. The official “delay” statistic is, technically speaking, any train that arrives at its terminal more than five minutes after its scheduled arrival time. In fact, every single subway performance metric is pegged to the schedule. So, in order to improve the stats, NYCT took a very simple step: they changed the schedules.
In March 2016, NYCT issued a “supplement” schedule—meaning a schedule change that doesn’t require board approval—for the 6 train. As an internal NYCT report at the time noted, 6 trains were running slower because of “changed circumstances” on the 6 line—“signal modifications and a fleet with narrower doors,” two factors entirely within NYCT’s control—so the division they padded the schedules during both rush hours, including reducing scheduled AM peak service from 24 trains per hour to 22. By giving trains more time to make their runs, it meant trains were more likely to arrive at the terminal “on time.”
As I have noted before, there are good reasons to adjust schedules to reflect reality. It allows dispatchers to manage complicated merges more efficiently and facilitates getting the trains back out on the next run as quickly as possible, among other benefits. In general, it’s good to know how long things take and plan accordingly. But there is a subtle difference between adjusting schedules to allow for more efficient operations—which should improve the rider experience—and merely shifting the goalposts.
The 6 train schedule adjustment was much more the latter. The 6 is one of the simpler lines on the system. It never crosses tracks with another line. The only merges to manage are the switches between the 6 and 6 express up in the Bronx. The report included in the Comptroller appendix, which the Comptroller says was reviewed by NYCT management, explicitly notes which type of schedule adjustment this was. “The actual running times and throughput, however, changed little,” it says, after the schedule changes were implemented [emphasis in original].
So what did change? The schedule adjustments resulted in “improved WA [Wait Assessment] and OTP [On-Time Performance],” NYCT’s two primary publicly-available statistics at the time.
There was even a bit of a natural experiment to test this. During a six-week period over that summer, the supplement schedule was temporarily suspended—meaning NYCT went back to the official schedule—and delays once again increased, which of course meant WA and OTP declined as well. In fact, 50 percent of the increased delays attributed to “crowding” across the entire system during those weeks came from the 6 line, simply because this supplement schedule had been suspended. This increase in “crowding” delays occurred despite the fact that, as is typical most summers, ridership over the summer was down from prior months.
The 6 train schedule adjustment is the subway crisis in a nutshell. With trains running slower due to the over-implementation of signal timers and other operating inefficiencies, management concealed that service was getting worse by using measures that only exacerbated those inefficiencies. Rather than trying to correct for the fact that they could no longer run the 6 on schedule, they simply moved the goalposts and then blamed most of the delays on “crowding,” doing little to address the fundamental problems causing the delays in the first place.
This is the institutional mindset that resulted in delays with no obvious cause gradually increasing by a whopping 1,190 percent from January 2012 to December 2017. With no incentive to call attention to the slow and steady deterioration of subway service, management rotated through an array of different reporting statistics based on which made them look best. When one statistic got discredited or started delivering bad news, they moved on to a new one.
We can see this dynamic very clearly just by observing the metric-juggling over the past several years. When On-Time Performance was high, NYCT reported that first. When OTP began falling approximately five years ago, Wait Assessment—a convoluted calculation meant to approximate how long people are waiting for trains that the state Comptroller raised serious concerns about back in 2016, and which fundamentally works to obscure the raw number of delays—became the statistic of preference despite repeated warnings from NYCT’s internal analysis unit that it was a bad stat. (My favorite problem with Wait Assessment: to create the system-wide statistic, they averaged the numbers for each individual line, meaning the Wait Assessment for the Franklin Ave shuttle was weighted just as heavily as the Lexington Ave express lines.) When Wait Assessment finally got electronic reporting on all lines in 2017, resulting in better data, it turned out much of the “improved” or steady performance of the past had been an illusion—a revelation that was never reported to the MTA board, much less to the public.
Even though Wait Assessment had at that point become a more reliable and better metric than it ever had been before, this was the same time NYCT management “no longer considered Wait Assessment to be a relevant performance indicator,” the Comptroller report notes. At this juncture, NYCT began to emphasize “Major Incidents,” yet another deeply flawed metric, which it still prominently reports today.
On Friday, the MTA pushed back on the broad narrative of the Comptroller’s report and said they never recategorized delays to mislead the public. Both the report and an MTA statement noted the numerous reforms put into place by Andy Byford since his tenure began in January 2018. In fact, one of Byford's very first major pronouncements was to dismiss “overcrowding” as a delay cause, something NYCT had vigorously defended for years prior. As late as the summer of 2017, MTA officials were still calling overcrowding the “main issue” facing the subways. Thankfully, nobody is saying that anymore.
Indeed, the MTA criticized the Comptroller's report as “more history and politics than news, focusing on rejected practices of the past while glossing over recent reforms” as NYCT pursues “additional transparency and accountability.”
Which brings me to a key point of not just the Comptroller report, but this entire era of subway operations: transparency. The statistics and heat maps included in the appendix—which require no more than five minutes of examination to see something was deeply wrong with the subway’s performance—have never been shared with the public until now. Had they been sooner, it is quite possible, even likely, that board members, journalists, or attentive subway riders with a few minutes to kill while waiting for a train would have noticed the subway’s rapidly declining performance and how the causes deviated so wildly from NYCT’s preferred narrative.
In fact, releasing this data on a monthly basis was one of the Comptroller’s key recommendations, a suggestion that Byford told me he is “receptive” to. I view this as nothing short of a necessity. It’s the only way to possibly prevent this series of events from happening again, particularly with one of Governor Cuomo’s closest allies on the MTA Board talking about pegging fare revenue to subway performance statistics. The MTA says it’s now all about transparency and accountability. This would be an easy way to prove it.