The Crisis Below: An Investigation of the Reliability and Transparency of the MTA’s Subway Performance Reporting

February 8, 2019

Table of Contents

    Executive Summary

    As is well known to every commuter, New York City subway service, including particularly on-time performance, has been steadily declining for years.  Delays[1] reported by the Metropolitan Transportation Authority (“MTA”) more than tripled between 2012 and 2017,[2] the average speed of trains reportedly fell to 1950’s levels,[3] and a lower percentage of trains arrive on time than in any other major subway system.[4]  The subways’ decline has inflicted substantial costs on the City and on New Yorkers personally, wasting an estimated $389 million per year in lost economic activity and wages for workers[5] and subjecting thousands of riders to untold delays getting where they need to go.

    This investigation by New York City Comptroller Scott M. Stringer chronicles something less well-known: for years, the MTA knowingly misled the riding public by reporting information it knew to be inaccurate, thereby obscuring the subways’ decline, misrepresenting the causes of delays, and masking significant operational problems.  Relying on internal MTA analyses never before made public, as well as interviews with key MTA officials, this report makes clear that agency executives continually obscured inconvenient facts and thereby cast the agency in a more positive light.  The net effect of this culture of obfuscation was to hide the truth behind the system’s deterioration, even as MTA executives were repeatedly informed by agency personnel tasked with analyzing system performance that its public disclosures were inaccurate and, in some cases, meaningless.

    In brief, the Comptroller’s Office found:

    • From mid-2015 forward, numerous internal MTA analyses concluded that the MTA’s databases and delay tracking protocols were routinely unable to accurately identify the causes of delays and, in particular, chronically misattributed delays to “Overcrowding.” A July 2015 internal MTA analysis described methodological breakdowns in stark terms, stating “[n]o policy or guidance exists on how dispatchers should properly identify the cause of a particular delay or on how delays should be assigned to incidents,”[6] while a January 2016 internal MTA analysis noted that “much of the delay data is incomplete or unreliable, particularly the classification/categorization of delays and the assignment of delays to particular incidents.”[7] Despite being informed of these deficiencies, MTA officials continued for years to publicly promote inaccurate information and misrepresent what the MTA knew about the causes of delays, casting the agency in a more positive light and shielding it from accountability.
    • Throughout 2016, MTA officials repeatedly asserted that subway service was improving based on reported increases in Wait Assessment scores, a metric intended to approximate the amount of time passengers must wait on platforms and long-touted by the MTA as its most important indicator of subway service. But there again, internal analyses obtained by the Comptroller’s Office show that MTA executives were cautioned that changes in Wait Assessment scores subsequently highlighted to MTA Board members were meaningless and likely the result of sample error. When technological advancements in data collection finally made clear that Wait Assessment scores had actually gotten worse, not better as the MTA had previously reported, the MTA quietly restated its previous inaccurate Wait Assessment results without disclosing that its earlier declarations of progress had been wrong. Five months later, the agency began to emphasize new metrics for reporting subway performance.
    • The MTA distorted its publicly reported statistics on delays by effectively hiding certain delays it internally attributed to “Unknown” causes. Instead of clearly alerting the public that the causes of these delays were unknown, for nearly a decade MTA officials simply apportioned them among the MTA’s fifteen publicly reported categories of delays[8] – obscuring their existence without any public explanation. In this way, in Monthly Operations Reports provided to the public and the MTA Board from 2013 through mid-2018,[9] the MTA hid the truth about 525,710 delays internally grouped as “Unknown” in what was until recently the database MTA used to report delays. This apportionment concealed the fact that the MTA’s delay tracking protocols were unable to identify the causes of a significant number of the delays occurring in the system.
    • Similar to the MTA’s misrepresentative reporting of these “Unknown” delays, the MTA’s recently featured reporting of “Major Incidents” obscures critical information and is also based on unreliable data. Publicly defined by the MTA as any incident that delays 50 or more trains, Major Incident reporting is based on MTA tracking protocols that routinely misidentify the number of delays caused by an incident, such that the MTA cannot reliably determine the number of incidents that cause 50 or more delays. Moreover, the MTA’s Major Incidents reporting methodology excludes significant numbers of Major Incidents the MTA has historically tracked internally – including all incidents charged to “Planned Work,” a large category that regularly bogs down whole subway lines. This exclusion and the MTA’s methodology for identifying Major Incidents has never been clearly explained to the public.

    The MTA has, in fact, acknowledged some of these issues and instituted changes throughout last year. Among other things, New York City Transit (“NYCT”) President Andy Byford has emphasized the need to identify the “root causes” of delays. In recognition of the fact that overcrowding is not the root cause, the MTA removed its “Overcrowding” delay designation from Monthly Operations Reports and re-categorized it as “Operating Environment.” This rebranding, however, has done little to address the underlying inaccuracy of the MTA’s delay data. Rather, as reflected in this report, systemic deficiencies remain embedded in the MTA’s performance reporting and continue to obscure the true causes of delays.  We encourage the MTA to consider the information in this report and use it to improve the MTA’s transparency, accountability, and, ultimately, the overall functioning of the New York City subway system.

    Specifically, we recommend that the MTA:

    1. Structure public reporting of performance information to maximize transparency, reliability, and accountability and, as part of this effort, report all delays on its subway performance Dashboard.
    2. Publish detailed definitions of all delay categories, specifically indicating what each one includes and, as necessary, omits.
    3. Ensure that all procedures relevant to performance reporting are formally codified in official policies and procedures, including establishing written definitions and instructions for all key terms, data categories, and work protocols.
    4. Train all relevant personnel on procedures relevant to performance reporting.
    5. In the context of public reports of Major Incidents, provide the public with information about all categories of service disruptions that cause 50 or more delays tracked as incidents within Subway Incident Reporting System, including specifically Planned Work.
    6. Transparently disclose in each Monthly Operations Report and on the MTA’s subway performance Dashboard the methodologies used to calculate performance metrics, including all exceptions and revisions to those methodologies and methodological weaknesses.
    7. Make available monthly on the MTA’s website or through an Open Data portal all data in the SIRS database and any other databases relied on for public reporting.

    Introduction

    The importance of accurate data cannot be overstated: it is a critical tool for evaluating and managing any organization, especially one as large and complex as the MTA. An organization that does not establish clear performance metrics to track its goals, ensure that data is collected properly and transparently, and make adjustments based on feedback sets itself up for failure.

    As this report makes clear, time and again the MTA has failed to adopt necessary controls to ensure the reliability and integrity of its public disclosures and misrepresented subway performance information in ways that cast its operations in a more favorable light than was reflected in the information it had internally. Worse, even after MTA staff repeatedly flagged weaknesses in data reliability, top agency officials continued to communicate information to the public that they knew misrepresented internal data and failed to take sufficient action to remedy the true causes of declining service.

    Internal records show that, as far back as 2015, analysts in the Department of Subways’ PAU, the unit specifically tasked with briefing agency executives on subway performance information, deemed both of the MTA’s delay tracking databases – the Terminal Delay Database (“TDD”) and the Subway Incident Reporting System (“SIRS”) – to be fundamentally unreliable. A July 2015 analysis described the MTA’s databases as being critically undermined by control weaknesses and said that employees blamed door holding (one of several circumstances the MTA publicly reported under the category “Overcrowding”) for delays so excessively that the resulting data was nearly useless.  According to that analysis:

    No policy or guidance exists on how dispatchers should properly identify the cause of a particular delay or on how delays should be assigned to incidents. . . .   Dispatchers rely on train crews to report the cause of delays, and these explanations are suspect.  The root cause of a delay often may not be apparent to a train crew.  In addition, a train might be delayed by more than one cause or incident. ‘Door holding’ is the most frequently used incident code (over 20% of all incidents), but it is used both inconsistently and incorrectly.  Therefore it provides almost no useful information.[10]

    To his credit, NYCT President Andy Byford has driven a reappraisal of the MTA’s subway performance data, heeded analysts’ findings, and begun remedying substantial operational deficiencies that previous administrations did not sufficiently disclose or address. Early in his tenure, President Byford acknowledged that the MTA’s previous “Overcrowding” reporting category was “not particularly meaningful”[11] and “a misrepresentation” because it did not communicate the “underlying root cause” of delays.[12] While these acknowledgements and the changes he has implemented are laudable, it nonetheless remains important to understand the MTA’s chronic failures to ensure the accuracy and transparency of its data reporting going forward. Such an understanding is critical to improve current practices and prevent a culture of obfuscation from again undermining the integrity of the MTA’s performance disclosures and misleading the public about the true causes of problems and the path to fix them.

    I. Investigative Findings

    A. The MTA Blamed Overcrowding as the Cause of Delays Where Its Own Records Did Not Support That Conclusion

    Until recently, the MTA employed two systems to record delays, the TDD, used to publicly report the causes of delays until July 2018, and SIRS, used for internal tracking and analysis of the causes of delays. In July 2018 SIRS replaced the TDD as the database used for public delay reporting.[13] As far back as 2015, MTA analysts determined that the TDD and SIRS were both generally unreliable and that the MTA’s official protocols could not accurately identify the causes of delays it publicly attributed to “Overcrowding.” Accordingly, for years, while the MTA continued to blame crowding for the system’s woes, top agency officials received monthly performance briefings showing that the MTA’s delay data was unreliable and that its public delay reporting misrepresented the causes of delays attributed to “Overcrowding.”

    Understanding the depths of the disconnect between the MTA’s public delay reporting and the internal information provided to senior officials requires some knowledge of how TDD and SIRS operate.

    • In the TDD, dispatchers wrote brief remarks summarizing the cause of delays based on oral explanations provided by the crews of delayed trains after they reached their final destinations. Other MTA employees then individually reviewed the remarks and, based on their reviews, tagged each delay with one of ninety-nine possible TDD “Reason Codes” that the reviewers determined most closely reflected the cause of each delay. Every month until July 2018, when the MTA began using SIRS to report delays, these TDD Reason Codes were mapped to the fifteen delay categories previously listed in Monthly Operations Reports, such as “Track Gangs,” “Car Equipment,” and “Overcrowding,” the last of which consisted mostly of delays tagged with Reason Codes respectively titled “Customer Holding Doors” and “Insufficient Capacity.”
    • In SIRS, which predated the TDD and was used for internal delay analysis before also becoming the database used for public delay reporting in July 2018, dispatchers stationed in the MTA’s Rail Control Center record the causes of delays in “Incident Letters” based on calls to the center from train crews as they encounter “incidents.” The MTA has no official definition of what an “incident” is, but agency officials have publicly described an “incident” to mean any interruption of service.[14] Following the creation of the Incident Letters, RCC personnel compare the time and location of incidents with logs of delayed trains’ travel history. If a delay appears to have clearly resulted from an incident, the delay is tagged with the Trouble Code previously applied to that incident. Where delays are not clearly attributable to previously identified incidents, RCC employees create new incidents in SIRS to account for the delays, tagging those incidents and delays with Trouble Codes corresponding to a category inaccurately titled “crowding” within SIRS. These delays are reported in MTA Monthly Operations Reports under a category titled “Operating Environment,” one of several revised delay categories the MTA adopted in July 2018 when it began using SIRS to report delays.

    Internal MTA records show that, by mid-2015, MTA analysts had determined that delay cause attributions in the TDD and SIRS were generally unreliable and that, in particular, the protocols for creating this data were unable to correctly identify the causes of delays publicly attributed to “Overcrowding.” An analysis drafted July 2015 described these databases as being critically undermined by control weaknesses and said that employees blamed door holding for delays so excessively that the resulting data was nearly useless.[15]  According to that analysis:

    No policy or guidance exists on how dispatchers should properly identify the cause of a particular delay or on how delays should be assigned to incidents. . . .   Dispatchers rely on train crews to report the cause of delays, and these explanations are suspect.  The root cause of a delay often may not be apparent to a train crew.  In addition, a train might be delayed by more than one cause or incident. ‘Door holding’ is the most frequently used incident code (over 20% of all incidents), but it is used both inconsistently and incorrectly.  Therefore it provides almost no useful information.[16]

    Analyses prepared soon thereafter again questioned the reliability of crowding attributions in the MTA’s databases, noting that the growth in the number of delays attributed to “Overcrowding” had dramatically outpaced contemporaneous ridership increases.  One such report dated as drafted in January 2016 stated, “Much of the delay data is incomplete or unreliable, particularly the classification/categorization of delays and the assignment of delays to particular incidents.”[17] The report further remarked on inconsistencies inherent in the data:

    From 2003 to 2013, weekday ridership increased 21% but total weekday delays increased nearly 400% … Yet delays fell from 1994 to 2003, simultaneously with a large increase in ridership, so ridership is clearly not the only cause… Indeed the share of delays in the morning peak has declined, despite being the time of day with the heaviest ridership and worst crowding.[18]

    Similarly, an analysis dated as drafted in February 2016 stated that:

    Although ridership has grown, and some relationship has been established between ridership and delays, the increase in delays attributed to crowding have significantly outpaced the increase in ridership. While most of the crowding delays occur during the peak periods, which see the highest concentration of ridership, again, there doesn’t seem to be a proportional relationship between the increase in delays and ridership. . .   As such, Crowding is now the single largest category of delays . . . and yet there is no clear explanation of what is happening, nor is it clear how accurate the attribution of delays to Crowding is.  In order to help address the underlying issues, it must first be understood what is happening.[19]

    Consistent with the concerns expressed above, by February 2016, analyses provided to senior MTA officials in advance of monthly meetings of the MTA Board’s Transit Committee indicated that the MTA’s official delay tracking protocols could not identify the causes of a significant portion of all delays, and particularly of those delays which the MTA publicly attributed to “Overcrowding.”[20] For analytical purposes, these analyses[21] described the growing category of delays aggregated under “Overcrowding” with various terms including “Unknown/Other,”[22] “No Capacity, Crowding, Excess Dwell, Unknown,”[23] and “Insufficient Capacity/Excess Dwell/Unknown.”[24]

    At the core of the MTA’s misattribution of the causes of delays was its decision to group delays tagged with the TDD Reason Codes “Insufficient Capacity”[25] and “Customer Holding Doors”[26] under the umbrella of “Overcrowding.”[27] The Comptroller’s Office’s review of TDD records confirmed that neither of these two Reason Codes reliably indicated that crowding had caused the delays to which either was applied.

    The MTA has never formally defined the meaning of “Insufficient Capacity” and could not provide any official policy or procedure governing the circumstances under which the term should be used describe the causes of delays. In interviews with the Comptroller’s Office, MTA employees with responsibility for generating TDD data were unable to concretely define “Insufficient Capacity” and said it applied to any circumstance where trains are delayed and the train’s crew cannot point to a specific incident or circumstance that caused the delay.[28]

    Similarly, “Customer Holding Doors” also proved to be an unreliable indicator of crowding.  As recited in the July 2015 analysis mentioned above, MTA employees blamed door holding so “inconsistently and incorrectly” that the explanation was characterized by MTA analysts as providing “almost no useful information.”[29]

    To gain a more complete picture of the how the MTA misrepresented crowding as the greatest cause of delays, the Comptroller’s Office reviewed TDD records for delays that the MTA publicly attributed to “Overcrowding” from 2016 through 2017, during which “Insufficient Capacity” and “Customer Holding Doors” delays accounted for 87 percent (570,093) of reported “Overcrowding” delays.[30] Our review found that the relevant TDD data did not provide detailed, reliable information about the causes of those delays sufficient to support their attribution to “Overcrowding,”[31] and in thousands of cases, explicitly pointed to other causes. For more than 140,000 delays associated with these two Reason Codes, the spaces intended for explanatory remarks were left entirely blank, merely referred to the fact that the train was late without indicating why, or only referred to unspecified system congestion.[32] For example, for thousands of remarks, the only description of the cause of delay was the phrase “Excess Dwell,” sometimes with the name of a particular subway station. According to documents provided by the MTA, “Excess Dwell” refers to any circumstance where a delayed train spent a greater than usual time at a station without a clear cause for why it did so.[33] Thousands of other remarks simply read “late arrival,”[34] “late arr,”[35] or other descriptions that similarly indicate that a train was late but do not indicate the cause.

    Moreover, our sample of such delays identified hundreds of instances where TDD remarks explicitly indicated that delays resulted from operational failures rather than from crowding. For example, in over 450 instances, remarks indicated that delays primarily resulted from track or train inspections, with remarks reading “Inspection,” “MONDAY INSPECTION,” “!!FRI INSP!!,” or other similar phrases.

    When the MTA developed revised delay categories in July 2018, it adopted the term “Operating Environment” in place of its historic “Overcrowding” category.[36] In documents prepared in response to the Comptroller’s information requests, the MTA explained that the term “Operating Environment” reflects its view that the delays in this category “are due to the operating environment rather than specific events that create delays.”[37] The MTA’s public Monthly Operations Reports do not list any subcategories under this heading or otherwise disclose a fuller explanation of the causes of the delays identified as caused by “Operating Environment.” As discussed below in Section II, though the MTA’s official delay tracking protocols are unable to formally identify their causes, analyses provided to MTA executives in 2017 concluded that “most ‘crowding’ delay charges . . . are largely the result of operating environment issues other than ridership/crowding.” (Emphasis in original.)[38]

    B. Wait Assessment: MTA Executives Repeatedly Claimed That Service Had Improved Based on Information Known to Be Meaningless

     

    For years, the MTA designated Wait Assessment[39] as the agency’s most important metric for gauging the quality of subway service.  It accounted for 60 percent – more than all other metrics combined – of the subway “Service Key Performance Indicator” the MTA published in its Monthly Operations Reports. It listed Wait Assessment statistics first in those reports, and almost always highlighted Wait Assessment trends at the beginning of the Department of Subways’ monthly oral performance reports to the Transit Committee.[40]

    Until April 2017, the MTA reported Wait Assessment statistics using a mixture of fully accurate, electronic data for A-Division subway lines (numbered lines) and manually-collected sample data for B-Division subway lines (lettered lines), which necessarily had a margin of error.

    Records obtained during the investigation establish that during at least five different meetings of the Transit Committee in 2016 and 2017, MTA officials prominently highlighted what they cast as improved or unchanged Wait Assessment scores, purporting to show that subway service was getting better or remaining stable. In fact, internal pre-meeting analyses presented to these officials beforehand stated that the results the officials subsequently highlighted during Transit Committee Meetings were statistically insignificant. Specifically, these pre-meeting analyses stated that certain increases in by-month and by-year Wait Assessment scores did not evidence “real” service improvements because they fell within or did not exceed the margin of error inherent to their calculation.[41]

    For instance, an internal analysis circulated to MTA executives before the Transit Committee’s September 2016 meeting stated “Improvement in [12-month average Wait Assessment scores] is due to B-Division and not statistically significant.” Nevertheless, in the Transit Committee’s meeting that month, MTA’s Acting-Vice President of Subways positively described the Authority’s progress over the previous year as measured by Wait Assessment, beginning his remarks by stating “Good morning. The 12-month system-wide Wait Assessment ending in July was 78.5 percent, which is .4 percent higher than last year.”[42]

    Although the Comptroller’s Office identified certain instances where disclosures concerning margins of error were included in pre-meeting draft scripts of MTA executives’ comments,[43] none of these disclosures were recited in the relevant oral reports to the Transit Committee; and MTA personnel interviewed in connection with this investigation could not identify any point where the MTA otherwise disclosed this information.

    The illusion of improvement unraveled in early 2017 when the MTA completed a multi-year process to generate electronic train location information for lettered lines and thus no longer needed to rely on manually-collected sample data for the B-Division. This allowed the MTA for the first time to retroactively calculate Wait Assessments using a complete population of statistics with no margin of error. Once it did, however, it was confronted with an uncomfortable truth – the new, fully accurate figures showed that the Wait Assessment statistics previously highlighted as improved had actually worsened over the previous year.

    Although this new, more accurate data became available in February 2017, MTA officials did not mention its availability during the Transit Committee’s March 2017 meeting and did not report the new figures until the Committee’s April 2017 meeting.[44] At that April 2017 meeting, the MTA quietly restated its previously-published 2016 Wait Assessment statistics without disclosing that the newly-available data contradicted the MTA’s many prior declarations that subway service had improved.[45]

    In September 2017, five months after the fully accurate statistics revealed a decline in performance rather than improvement, the MTA declared that it no longer considered Wait Assessment to be a relevant performance indicator and announced new performance metrics, particularly emphasizing Major Incidents.[46]

    C. The MTA Buried Certain Delays Internally Recorded as Resulting from “Unknown” Factors by Apportioning Them to Other Reported Causes without Explanation

    Another way the MTA misrepresented the causes of delays concerns  its treatment of delays tagged with TDD Reason Codes titled “Illegible”[47] and “No Reason,”[48] both of which it rolled up into an undisclosed internal TDD delay tracking category titled “Unknown.”[49]  Specifically, from February 2009 through April 2018, pursuant to agency policy, MTA officials took all the “Unknown” cause delays and simply apportioned them across the MTA’s fifteen publicly reported categories of delay causes.[50] So, if a specific cause accounted for 10 percent of all delays, then 10 percent of the “Unknown” delays were added to that cause’s numbers. This apportionment had the effect of hiding the “Unknowns” from view and of concealing the fact that the MTA’s delay reporting was substantially less precise than its published reports suggested.

    Accordingly, in Monthly Operations Reports provided to the public and the MTA Board from 2013 through mid-2018, the MTA apportioned 525,710 delays internally grouped under this “Unknown” cause category (13.4 percent of all delays reported during that period) to the MTA’s fifteen reported delay categories.

    In addition to masking the fact that more than 10 percent of the causes of delays were categorized by the MTA as “Unknown,” this apportionment most substantially increased the number of “Overcrowding” delays reported because that category already contained the largest number of delays. As such, from 2013 through mid-2018, the “Overcrowding” category in Monthly Operations Reports received 29.3 percent of all TDD “Unknown” cause delays, inflating the total number of delays attributed to “Overcrowding” by 154,256.  This percentage was even higher in 2016 and 2017, during which 36 percent of these internally tracked “Unknown” cause delays were publicly attributed to “Overcrowding” (82,868 additional delay attributions) and 64 percent were attributed to the other reported delay categories (collectively inflating these categories by 144,581 delays).

    In interviews with the Comptroller’s Office, MTA officials were unable to identify any instance where the practice of apportioning these “Unknown” cause delays to other categories was disclosed to the public or the MTA Board. However, shortly after the Comptroller’s Office learned of this practice through an interview of an MTA employee, the MTA briefly reported these “Unknown” cause delays under a separate category titled “Unassigned.” Making that change, the MTA’s Monthly Operations Reports for April through June 2018 included a note stating that “[h]istorically, unassigned delays have been proportionately distributed across delay categories in Board reporting materials.  This month they are shown separately as unassigned.” From the first inclusion of the “Unassigned” delay category in Monthly Operations Reports in April 2018 through its last appearance in June of the same year, none of the Department of Subways’ oral performance reports mentioned the MTA’s former policy of apportioning those delays and the impact that policy had on the MTA’s official delay statistics over the previous decade.

    D. The MTA’s New Public Reporting of “Major Incidents” Suffers from Flaws Similar to Those Found in Its Reporting of Delays

    In recent months, the MTA has prominently positioned Major Incidents in communications with its Board and the broader public as the metric most indicative of the customer experience.[51] The agency publicly defines Major Incidents as incidents that delay 50 or more trains. [52] According to the MTA Subway Performance Dashboard, “such events cause the most disruption to customers.” (Emphasis added.)[53]

    Despite the attention the MTA has given them, Major Incidents have not proven to be a transparent or reliable indicator of overall service quality[54] or of the MTA’s success at reversing the subways’ long-term decline. To begin with, the MTA cannot reliably calculate the number of Major Incidents that occur because it is unable to reliably determine the number of delays caused by each incident. As stated in an internal analysis from October 2017, existing protocols for identifying Major Incidents result in MTA employees “arbitrarily grouping delays into incident letters,” and “most incidents are merely groups of delays.”[55] Incident Letters obtained by the Comptroller’s Office included numerous instances where hundreds of delays were attributed to incidents with “incident durations” (i.e., initial train stoppage/blockage time) of only a few minutes, without detailed explanations of why the delays were attributed to the corresponding incidents or even specific identification of the trains deemed to have been delayed.

    More importantly, our investigation found that the agency does not publicly report certain Major Incidents that it has historically tracked in briefings for agency executives. Specifically, the MTA excludes from its public reports all incidents attributed in SIRS to both “Planned Work” – such as incidents caused by track and signal work, both huge sources of disruption throughout the system – and “Other Operating Environment” causes. MTA officials explained these exclusions by asserting that “Planned Work” and “Other Operating Environment” are not true incidents but rather are conditions (i.e., ever-present characteristics of the environment in which subways operate). However, the investigation found that internal MTA briefing materials for agency executives historically included both “incidents” and “Major Incidents” attributed to “Planned Work” and “Other Operating Environment” causes. MTA executives’ briefing materials obtained by the Comptroller’s Office show that this practice continued into 2018, though portions of these briefing documents sometimes described such service disruptions as “Other Significant Events” or “50+ Train Delay Letters Not Reported As Major Incidents” when including them in conjunction with lists of reported Major Incidents.[56]

    The omission of internally-tracked “Planned Work” Major Incidents from the MTA’s published statistics significantly lowers the total number of publicly reported Major Incidents. “Planned Work” accounted for roughly 8 percent of internally-tracked Major Incidents prior to the implementation of the Subway Action Plan[57] but rose in frequency to as much as 28 percent after it was implemented.[58]  Between October 2017 (the first month MTA began publishing Major Incidents statistics) to August 2018 (the latest month for which the Comptroller’s Office obtained SIRS data), the MTA reported 860 Major Incidents but omitted 322 Major Incidents caused by Planned Work.  By doing so, the MTA lowered the number of publicly reported Major Incidents by 37 percent.

    The MTA’s omission of Planned Work Major Incidents in its publicly reported Major Incidents statistics continues its practice of presenting performance data in ways that cast the subways in the best light without disclosing what its internal statistics truly reflect.  Just as the MTA arbitrarily apportioned certain “Unknown” delays tracked in the TDD to “Overcrowding” and other categories without telling the public, it excludes “Planned Work” from its publicly reported Major Incidents statistics without clearly disclosing that it does so, thereby obscuring the total number of service disruptions experienced by the public that result in 50 or more delays.

     

    II. The Path Forward

    As previously noted, New York City Transit President Andy Byford has driven a reappraisal of the MTA’s protocols for tracking and reporting delay data. In early 2018, President Byford acknowledged that the MTA’s “Overcrowding” delay classification was “not particularly meaningful” and stated that it reveals nothing about the “underlying root cause” of delays.[59] In June 2018, the MTA debuted revised reporting categories, and replaced the term “Overcrowding” with “Operating Environment.”[60] In addition, over the last few months of 2018, the MTA achieved an 11 percent reduction in delays system-wide.[61] At the Transit Committee’s December 2018 meeting, MTA executives credited this reduction in delays to having developed “a far greater understanding of root cause” during the previous year, which they said had allowed them to assign “meaningful programs and actions to address those root causes of delay.”[62]

    The reforms instituted in 2018 address some of the issues highlighted in this report and certain causes of the steady rise in delays. Although the MTA’s removal of “Overcrowding” from its categories of delays is a long-overdue step towards transparency, to date, the MTA has yet to acknowledge that its prior use of “Overcrowding” was internally known for years to be a mischaracterization before it ceased publicly using that term.

    “Operating Environment,” the term adopted in its place, still perpetuates misleading aspects of the MTA’s previous reporting by obscuring the MTA’s full understanding of the causes behind such delays.  Just as the MTA reported delays under “Overcrowding” in the TDD which it internally determined were not caused by crowding, now using SIRS, its use of the term “Operating Environment” obscures its determination that delays reported under that category are primarily the result of avoidable operational failures. For example, a “Preliminary Format” of the MTA’s revised delay category descriptions provided in advance of the June 2018 Transit Committee meeting included a proposed category described as “Operating Environment Non-Incidents (e.g., schedule misalignment, insufficient capacity, operator variability).”[63]

    Describing the root causes the MTA has now taken action on to achieve the recent reduction in delays, at the December 2018 meeting of the Transit Committee, MTA officials specifically singled out improperly calibrated track signals and unnecessarily slow speed limits as significant contributing factors to a widespread reduction in train speeds throughout the subway system.[64]  However, they did not make clear when these problems first came to the MTA’s attention, which our investigation found dated back to at least early 2017. A March 2017 presentation provided to MTA executives stated that signal modifications to reduce speed limits had resulted in “a reduction in throughput beyond any projected impact,” and that, as a particular result of faulty signal modifications, train crews “tend to operate significantly below posted [speed limits], further reducing capacity and lengthening running times.”[65]  Echoing this analysis, briefing materials prepared for MTA executives the following month stated:

    Temporary Disruption/Incident delays are relatively easy to explain via SIRS data, but delays due to changes in the operating environment are not. “Crowding” delays behave consistent with changes in operation environment (signal modifications, flagging rules/practices, more cautious train operation, etc.).[66]

    Another analysis for senior MTA officials in October 2017 stated that, while the MTA had not yet performed sufficient research to rule out all other causes, “The great majority of evidence to date points towards signal modifications and slower crews as the (proximate) cause of declining reliability” and the “primary culprit behind the [subways’] gradual, long term decline.”[67] This analysis further stated that “most ‘crowding’ delay charges . . . are largely the result of operating environment issues other than ridership/crowding.” (Emphasis in original.)[68]

    At the December 2018 Transit Committee meeting, NYCT President Andy Byford conceded that the MTA’s recent success reducing delays was the result of “things that could and should have probably been done a long time ago.  No brainers.  Things that actually don’t necessarily cost very much but just I hope demonstrate greater attention to detail.”[69]  As the MTA has attributed approximately 30 percent of all delays in December 2018 to “Operating Environment,” work to address these delays must continue.[70]

    III. Recommendations

    Based on the findings set forth in this report, the Comptroller’s Office recommends that the MTA:

    1. Structure public reporting of performance information to maximize transparency, reliability, and accountability and, as part of this effort, report all delays on its subway performance Dashboard.
    2. Publish detailed definitions of all delay categories, specifically indicating what each one includes and, as necessary, omits.
    3. Ensure that all procedures relevant to performance reporting are formally codified in official policies and procedures, including establishing written definitions and instructions for all key terms, data categories, and work protocols.
    4. Train all relevant personnel on procedures relevant to performance reporting.
    5. In the context of public reports of Major Incidents, provide the public with information about all categories of service disruptions that cause 50 or more delays tracked as incidents within SIRS, including specifically Planned Work.
    6. Transparently disclose in each Monthly Operations Report and on the MTA’s subway performance Dashboard the methodologies used to calculate performance metrics, including all exceptions and revisions to those methodologies and methodological weaknesses.
    7. Make available each month on the MTA’s website or through an Open Data portal all data in the SIRS database and any other databases relied on for public reporting.

    Appendices

    View the appendices here.

    Endnotes

    [1] Under MTA Operational Directive 1.303 (issued October 1, 2014), section 4.2, the MTA deems a “Delay” to have occurred when a scheduled train: (1) reaches its final destination more than five minutes behind schedule; (2) fails to make any scheduled stops (an “Enroute Abandonment” or “EABD”); or (3) fails to depart from its originating terminal within certain time limits (a “Terminal Abandonment” or “TABD”). For purposes of official delay reporting, unscheduled trains cannot be “delayed.”

    [2] Monthly delays totaled 27,682 in January 2012 and 83,167 in December 2017, a 200.4 percent increase. See New York City Transit (“NYCT”) Committee Monthly Operations Reports for meetings respectively held March 26, 2012, and February 20, 2018. Monthly Operations Reports are provided to the Board and Board Committees and posted on the MTA’s website in advance of the monthly meetings.  See http://web.mta.info/mta/news/books/.

    [3] Mahler, Jonathan, “The Case for the Subway,” The New York Times, January 3, 2018. https://www.nytimes.com/2018/01/03/magazine/subway-new-york-city-public-transportation-wealth-inequality.html (last visited February 6, 2019).

    [4] Santora, Marc, “Failing Subway Threatens New York’s Financial Future, M.T.A. Chief Says,” The New York Times, November 20, 2017.  https://www.nytimes.com/2017/11/20/nyregion/subways-new-york-lhota-mta.html (last visited February 6, 2019).

    [5] Office of New York City Comptroller Scott M. Stringer, “The Economic Cost of Subway Delays,” October 1, 2017, at p. 2-3.

    [6] MTA Performance Analysis Unit (“PAU”) internal “Project Note” analysis, “Delays – Attributing Incidents and Causes,” July 2, 2015, at p. 1. The MTA created PAU in late 2013 because it could not explain the causes of rising system-wide delays. Among other things, MTA executives relied on analyses composed by PAU when preparing for monthly meetings of the MTA Board’s Transit and Bus Committee (“Transit Committee”).  Based on the MTA’s responses to the Comptroller’s information requests, from at least June 2017 forward, PAU authored virtually all of MTA analyses of the causes of improvement or worsening subway service.

    [7] PAU internal “Project Note” analysis, “Delay Study Project Notes,” January 13, 2016, at p. 1.

    [8] The number of delay categories in MTA’s Monthly Operations Reports has varied slightly over time, such that the MTA has not reported precisely fifteen different delay categories over the entirety of the time that MTA apportioned TDD “Unknown” cause delays.

    [9] Monthly Operations Reports are provided to the Board and Board Committees and posted on the MTA’s website in advance of the monthly meetings.  See http://web.mta.info/mta/news/books/.

    [10] PAU internal “Project Note” analysis, “Delays – Attributing Incidents and Causes,” July 2, 2015, at p. 1.

    [11] Nir, Sarah M., and Rosentha, Brian M., “‘Overcrowding’ Is Not at the Root of Delays, Subway Chief Says,” The New York Times, February 20, 2018. https://www.nytimes.com/2018/02/20/nyregion/subway-delays-overcrowding.html (last visited February 6, 2019).

    [12] Rivoli, Dan, “NYC Transit will stop blaming train delays on ‘overcrowding,’” New York Daily News, June 17, 2018. https://www.nydailynews.com/new-york/ny-metro-mta-subway-delays-andy-byford-20180616-story.html (last visited February 6, 2019).

    [13] The MTA first used SIRS for public performance reporting in October 2017, when it began reporting Major Incidents.

    [14] See description of “incidents” articulated to MTA Transit Committee by former Vice President of Subways during the Transit Committee’s May 2015 monthly meeting, held May 18, 2015, at minute 1:02:15-30 of meeting video, available at https://www.youtube.com/watch?v=gHYk0qUnmqI&feature=youtu.be&t=3735. The MTA posts videos of all Board committee meetings on its official Youtube.com channel, available at https://www.youtube.com/user/mtainfo.

    [15] The July 2015 internal analysis relied on above was labeled a “draft” when provided to the Comptroller’s Office, as were most of the PAU analysis relied on and quoted in this report. Except as relates to the creation of SIRS Incident Letters, the MTA does not have any policies or procedures requiring that subway performance analyses be officially finalized or approved. As a result, almost all of the analyses composed by PAU were perpetually labeled as drafts.  In addition, some contain minimally conflicting date information, and none were formally certified as representing the official opinion of the MTA. However, as noted, PAU was created for the purpose of performing these analyses and the PAU findings that are relied on and quoted in this report are consistent with multiple years of briefing materials provided monthly to MTA executives, with public comments by MTA officials about historic deficiencies in the MTA’s performance reporting, and with recent MTA disclosures concerning how the MTA succeeded in reducing delays in late 2018.

    [16] PAU internal “Project Note” analysis, “Delays – Attributing Incidents and Causes,” July 2, 2015, at p. 1. The Comptroller’s Office first requested that MTA provide all policies, procedures, or guidance relevant to MTA’s reporting of Delays and Major Incidents in January 2018. Thereafter, consistent with the comments reflected in MTA’s internal analyses, long-tenured employees of the respective work units responsible for recording the causes of delays in the TDD and SIRS interviewed by the Comptroller’s Office said that they could not recall any policies, procedures, guidance, or other documents providing instruction on how employees should determine the root causes of delays. However, at the close of this investigation, MTA provided the Comptroller’s Office with several documents which contained partial instructions related to identifying the causes of delays, including two draft Microsoft PowerPoint presentations and an untitled, undated, one-page document concerning the differing ways that certain SIRS Incident codes should be used. None of the employees interviewed recalled any such documents and we did not identify any references to these documents in the MTA’s numerous internal analyses concerning the identification of causes of delays.  The documents themselves were not captioned as policies, procedures, or otherwise as controlling documents governing MTA employees, and were not provided to our Office until almost a year after they were first requested.

    [17] PAU internal “Project Note” analysis, “Delay Study Project Notes,” January 13, 2016, at p. 1.

    [18] PAU internal “Project Note” analysis, “Delay Study Project Notes,” January 13, 2016, at p. 1.

    [19] PAU internal “Project Note” analysis, “Effects of Crowding on Service,” February 29, 2016, at p. 1.

    [20] “TAC Prep” analysis, “Performance Variance Explanations – Estimated Quantification of Causes of Change in Performance – December 2015,” February 5, 2016 (estimating that “Other/Unknown” factors accounted for 29 percent of the increase in delays as tracked in the TDD and 30 percent of the increase in Delays as tracked in SIRS from January through December 2015, and further that “JZ Line Unknown Issues (mostly in crowding and planned ROW work categories)” accounted for 14 percent of the increase in TDD-tracked delays and 13 percent of the increase in SIRS-tracked delays during that time.).  With NYCT’s Department of Subways, the term “TAC Prep” was used to refer to meetings and briefing materials related to preparing MTA executives for meetings of the Transit Committee. Based on interviews and documents obtained from MTA, in some instances “TAC Prep” materials were distributed to meeting attendees electronically and in other cases hardcopy “TAC Prep” materials were brought to these meetings.

    [21] See Appendix I for an example of a “Heat Maps,” a type of document regularly included in “TAC Prep” materials prepared for MTA executives in advance of the Transit Committee’s monthly meetings.

    [22] “TAC Prep” analysis, “Estimated Quantification of Causes of Change in Performance – February 2017,”   April 21, 2017 (with regards to delays tracked in the TDD, estimating that, from March 2016 through February 2017, declining ridership prevented 30 delays per weekday; “TABD-induced crowding” accounted for an increase of 35 delays per weekday (and stating “Increase in delays charged to dwell/capacity correlated with TABDs, which are increasing”); and that “Unknown/Other” factors accounted for an increase of 162 delays per weekday (and stating “Temporary Disruption/Incident delays are relatively easy to explain via SIRS data, but delays due to changes in the operating environment are not. “Crowding” delays behave consistent with changes in operating environment (signal modifications, flagging rules/practices, more cautious train operation, etc.).”). See Appendix II for this variance analysis, a type of document regularly included in “TAC Prep” materials prepared for MTA executives in advance of the Transit Committee’s monthly meetings.

    [23] “TAC Prep” analysis, “SIRS Heat Map: Delays Per Weekday, System,” April 11, 2017, at p. 5 of April 12, 2017, “TAC Prep” packet prepared in advance of the Transit Committee’s April 2017 meeting.

    [24] “TAC Prep” analysis, “SIRS Heat Map: Delays per Weekday, System, Peak & Off-Peak,” September 7, 2017, at p. 74 of September 2017 “TAC Prep” packet prepared in advance of the Transit Committee’s September 2017 meeting (listing the “Insufficient Capacity/Excess Dwell/Unknown” delay category as accounting for 39 percent of the overall growth in average weekday delays tracked in SIRS from August 2015 through August 2017).

    [25] From 2013 through April 2018, the TDD Reason Code “Insufficient Capacity” was applied to 346,102 delays tracked in the TDD.

    [26] From 2013 through April 2018, the TDD Reason Code “Customer Holding Doors” was applied to 733,173 delays tracked in the TDD.

    [27] In materials prepared in response to Comptroller information requests, the MTA stated that it was unable to identify the individuals responsible for this decision or when the decision was made. Accordingly, the MTA is unable to identify how long it practiced the misrepresentative TDD delay reporting practices detailed in this report.

    [28] This description is consistent with a 2016 training presentation obtained from the MTA which said that Insufficient Capacity referred to circumstances where a train becomes delayed gradually along its route rather than at any one particular location. See “Stringlines Training – Identifying Incidents and Service Management Actions,” drafted April 2016, and revised August 2016, at p. 25.

    [29] PAU internal “Project Note” analysis, “Delays – Attributing Incidents and Causes,” July 2, 2015, at p. 1.

    [30] The remaining 13 percent were tracked in the TDD as “Unknown” cause delays that were apportioned into the MTA’s reported “Overcrowding” delay category, as discussed in Section I(C) below.

    [31] See Appendix III for selected illustrative TDD “remarks” included in the sample of reported “Overcrowding” delays from 2016-2017.

    [32] For example, 2016-2017 TDD Delay data included at least 387 delays reported as “Overcrowding” for which the associated TDD remarks read “Plugged by Leader” (251 coded as “Insufficient Capacity” and 136 coded as “Customer Holding Doors”). The almost identical remark “plug by leader” appeared at least 173 times (114 delays coded as “Insufficient Capacity” and 59 coded as “Customer Holding Doors”).

    [33] See “Stringlines Training – Identifying Incidents and Service Management Actions,” drafted April 2016, and revised August 2016, at p. 25 (stating that “Excess Dwell” should be used to describe the cause of a delay where a “Train experiences above normal (atypical) dwell times” and there are “No other contributing causes.” While there was no separate TDD Reason Code titled “Excess Dwell,” this phrase appeared in thousands of TDD Delay remarks coded “Customer Holding Doors” and “Insufficient Capacity.” In SIRS, the MTA has a Trouble Code titled “Excess Dwell Time.”

    [34] The Comptroller’s review of a sample of TDD Delays reported under “Overcrowding” from 2016-2017 identified 863 delays for which remarks read only “late arrival.”

    [35] The Comptroller’s review of a sample of TDD Delays reported under “Overcrowding” from 2016-2017 identified 438 delays for which remarks read only “late arr.”

    [36] From October 2017 to the MTA’s July 2017 transition to using SIRS to publicly report delays, MTA Monthly Operations Reports used the term “Overcrowding/Insufficient Capacity/Other” in place of the former “Overcrowding.” Though MTA officials had repeatedly declared crowding to be the greatest cause of delays throughout the preceding years, no mention of this terminology change was made in the Department of Subways’ October 2017 monthly oral performance report to the Transit Committee.

    [37] See Appendix IV, containing untitled, undated document prepared by MTA in response to the Comptroller’s information requests, describing the categories of Delays reported and tracked internally by the MTA in connection with TDD delay reporting.

    [38] October 10, 2017 Memoranda titled “Variance Analysis Supplementary Information,” included in October 2017 “TAC Prep Packet.” Consistent with this analysis, a document prepared by the MTA in response to the Comptroller’s information requests stated that delays reported under the MTA’s “Overcrowding” reporting category were “typically” related to factors other than crowding.

    [39] Wait Assessment is a calculation intended to quantify the evenness of subway service by measuring the percentage of intervals –the time that passes between consecutive trains on the same line– that exceed the scheduled interval time.  As explained in MTA Monthly Operations Reports, “Wait Assessment (WA), is measured as the percentage of intervals between trains that are no more than the scheduled interval plus 25%. Minor gaps are more than 25% to 50% over the scheduled headway, medium gaps are more than 50% to 100% over the scheduled headway, and major gaps are more than 100% over the scheduled headway, or missed intervals.”

    [40] NYCT executives make oral presentations to the MTA Transit Committee during the Transit Committee’s monthly meetings. The MTA has no rules or procedures regarding the content of these oral performance reports. In an interview, a former MTA Acting-Vice President of Subways said that the content of the oral reports the Acting-Vice President provided to the Transit Committee reflected his “final opinion on what [information was] important” to convey to the Transit Committee.

    [41] See Appendix V for an example of a “One Pager,” a type of analysis regularly included in “TAC Prep” materials.

    [42] See video of the Transit Committee’s September 2016 meeting, held September 26, 2016, at minute 11:23 (available at https://www.youtube.com/watch?v=ewjCWGbXGoQ&feature=youtu.be&t=683).

    [43] See October 31, 2016, draft script for November 2016 Transit Committee meeting and December 2, 2016, draft script for December 2016 Transit Committee meeting, which included language reading “but statistically unchanged” after the scripted references to improved Wait Assessment scores. MTA records do not clearly establish if the draft scripts provided to the Comptroller’s Office were the final versions used by Department of Subways’ leadership. Accordingly, it is unclear if these disclaimers regarding margins of error were removed during the drafting process or were included but not read aloud to the Transit Committee.

    [44] The version of the MTA’s March 2017 Monthly Operations Report included in the MTA’s “TAC Prep” materials (dated March 9, 2017) prepared in advance of the Transit Committee’s March 2017 meeting included the accurate Wait Assessment data, but the final version provided to the public and the Transit Committee eleven days later did not. In response to Comptroller’s Office inquiries, the MTA stated that it was unable to identify the officials responsible for this decision.

    [45] The MTA’s April 2017 Monthly Operations Report included a brief note stating that prior period Wait Assessment data had been restated. While the draft script included in “TAC Prep” materials prepared in advance of the Transit Committee’s April 2017 meeting included the language “currently reported prior period Wait Assessment figures that were derived from sample data have been restated with fully electronic data,” these comments were not included in the Department of Subways’ April 2017 oral performance report to the Transit Committee.

    [46] See September 27, 2017, MTA Press Release, “MTA Launches New Customer-Focused Subway Performance Dashboard Providing Metrics Surpassing Global Standards for Transit Systems,” (stating that the MTA’s “Legacy” performance metrics are “not considered relevant indicators of customer experience). Available at http://www.mta.info/press-release/nyc-transit/mta-launches-new-customer-focused-subway-performance-dashboard-providing (last visited February 7, 2019).

    [47] From 2013 through April 2018, the TDD Reason Code “Illegible” was applied to 126 Delays tracked in the TDD.

    [48] From 2013 through April 2018, the TDD Reason Code “No Reason” was applied to 525,584 Delays tracked in the TDD.

    [49] The “Insufficient Capacity” and “Customer Holding Doors” TDD delays discussed earlier were also grouped under a different internal tracking category titled, in part, “Unknown” (“Unknown/Insufficient Capacity/Crowding/Door Holding”). Though for sake of clarity this report discusses those delays separately from the “Unknown” cause delays discussed here, in substance, the MTA misrepresented the causes of all or almost all delays tagged as “Customer Holding Doors,” “Insufficient Capacity,” “No Reason,” and “Illegible.” At present, the MTA reports all delays tagged with SIRS Trouble Codes that are analogous to these TDD Reason Codes under the same category, “Operating Environment.” For December 2018, the MTA’s reporting of Delays based on SIRS data attributed 30.6 percent of delays to “Operating Environment.”

    [50] See Appendix VI, a February 20, 2009 email and attachment, the only document identified by the MTA as authorizing and providing instructions for the practice.

    [51] Major Incidents were listed first in the MTA’s September 2017 announcement of its New Metrics. See September 27, 2017, MTA Press Release, “MTA Launches New Customer-Focused Subway Performance Dashboard Providing Metrics Surpassing Global Standards for Transit Systems,” available at http://www.mta.info/press-release/nyc-transit/mta-launches-new-customer-focused-subway-performance-dashboard-providing (last visited February 7, 2019). Major Incidents have since been consistently listed first in the MTA’s Monthly Operations Reports. http://web.mta.info/mta/boardmaterials.html. Major Incidents are currently listed first on the MTA’s Dashboard and automatically open upon arrival on the site, such that they serve as the Dashboard’s homepage.  See http://dashboard.mta.info/ (last visited February 7, 2019).

    [52] After initially introducing Major Incidents as “the number of incidents each month that delay 50 or more trains,” the MTA now describes them as the number of “unplanned incidents that delay 50 or more trains” in the Monthly Operations Reports.  (Emphasis added.)  The Dashboard still omits the word “unplanned.”  See http://dashboard.mta.info/ (last visited February 7, 2019).

    [53] http://dashboard.mta.info/ (last visited February 7, 2019).

    [54] The MTA has historically attributed less than 16 percent of reported delays to Incidents tracked within SIRS as causing over 50 delays.

    [55] PAU internal “Project Note” analysis, “Delay Data Improvement Project,” October 30, 2017, at p. 1.

    [56] See Appendix VII for an example of a document illustrating such practice.

    [57] In the six months before the Subway Action Plan was announced (January through June 2017), 552 incidents were recorded in SIRS as having caused 50 or more delays, of which 44 were attributed to Planned Work.

    [58] In the six months after the Subway Action Plan was announced (August 2017 through January 2018), 593 incidents were recorded in SIRS as having caused 50 or more delays, of which 133 were attributed to Planned Work (22 percent). During the following six months (February 2018 through July 2018), 668 incidents were recorded in SIRS as having caused 50 or more delays, of which 189 were attributed to Planned Work (28 percent).

    [59] Nir, Sarah Maslin and Brian M. Rosenthal. “‘Overcrowding’ Is Not at the Root of Delays, Subway Chief Says,” New York Times, February 20, 2018.

    [60] See Appendix VIII, containing the presentation on revised delay reporting categories presented at the Transit Committee’s June 2018 meeting.

    [61] The MTA reported 67,952 delays for December 2018, 9,045 fewer than the 76,997 delays the MTA reported for July 2018.

    [62] See video of the Transit Committee’s December 2018 meeting, at minutes 32:05-36:27.

    [63] See “PRELIMINARY FORMAT” revised delay reporting categories and associated descriptions, at p. 66 of June 2018 “TAC Prep” materials prepared in advance of the Transit Committee’s June 2016 meeting, attached at Appendix IX. This additional detail was not included in the MTA’s presentation of revised delay categories later that month, and since that time, the MTA’s Monthly Operations Reports have not listed any subcategories explaining the causes of delays reported under “Operating Environment.”

    [64] See video of the Transit Committee’s December 2018 meeting, at minutes 40:00-42:42.

    [65] Email titled “RE: Questions for presentation,” March 11, 2017, attaching presentation document with file name “Subway Performance Challenges for President 2017-02-13.”

    [66] “TAC Prep” analysis, “Performance Variance Analysis – Estimated Quantification of Causes of Change in Performance,” April 11, 2017, included materials prepared in advance of the Transit Committee’s April 2017 meeting.

    [67] “TAC Prep” analysis, “Variance Analysis Supplementary Information,” October 10, 2017, included in materials prepared in advance of the Transit Committee’s October 2017 meeting.

    [68] “TAC Prep” analysis, “Variance Analysis Supplementary Information,” October 10, 2017, included in materials prepared in advance of the Transit Committee’s October 2017 meeting.

    [69] See NYCT President’s Oral Remarks after approval of minutes during December 2018 Transit Committee Meeting. Included in these remarks was President Andy Byford’s thanks to the PAU employees for playing an important role in performing the analysis behind the MTA’s “Save Safe Seconds,” which President Byford has credited for significant delay reductions achieved in late 2018.

    [70] MTA delay statistics reported for December 2018 attributed 16,523 delays to “Operating Environment” out of a total of 67,952 delays reported for that month, equaling 30.6 percent.

    $242 billion
    Aug
    2022