2020-25212. Metrics and Minimum Standards for Intercity Passenger Rail Service

AGENCY:

Federal Railroad Administration (FRA), Department of Transportation (DOT).

ACTION:

Final rule.

SUMMARY:

This final rule establishes metrics and minimum standards for measuring the performance and service quality of intercity passenger train operations.

DATES:

This final rule is effective on December 16, 2020.

FOR FURTHER INFORMATION CONTACT:

Kristin Ferriter, Transportation Industry Analyst, telephone [phone]; or Zeb Schorr, Assistant Chief Counsel, telephone [phone].

SUPPLEMENTARY INFORMATION:

Table of Contents for Supplementary Information

I. Executive Summary

II. Background

III. Response to Comments on On-Time Performance and Train Delays

IV. FRA Quarterly Reporting

V. Section-by-Section Analysis of Comments and Revisions From the NPRM

VI. Regulatory Impact and Notices

A. Executive Order 12866, Executive Order 13771, and DOT Regulatory Policies and Procedures

B. Regulatory Flexibility Act and Executive Order 13272; Initial Regulatory Flexibility Assessment

C. Paperwork Reduction Act

D. Federalism Implications

E. Environmental Impact

F. Executive Order 12898 (Environmental Justice)

G. Executive Order 13175 (Tribal Consultation)

H. Unfunded Mandates Reform Act of 1995

I. Energy Impact

J. Trade Impact

I. Executive Summary

A. Overview of the Final Rule

This final rule establishes metrics and minimum standards for measuring the performance and service quality of Amtrak's intercity passenger train operations (Metrics and Standards). The Metrics and Standards are organized into four categories: On-time performance (OTP) and train delays, customer service, financial, and public benefits. With respect to on-time performance and train delays, this final rule sets forth a customer on-time performance metric, defined as the percentage of all customers on an intercity passenger rail train who arrive at their detraining point no later than 15 minutes after their published scheduled arrival time, reported by train and by route. This final rule establishes a customer on-time performance minimum standard of 80 percent for any 2 consecutive calendar quarters, and sets forth when the standard begins to apply. In addition, this final rule includes the following related metrics: Ridership data, certified schedule, train delays, train delays per 10,000 train miles, station performance, and host running time.

B. Procedural History

By notice of proposed rulemaking (NPRM) published on March 31, 2020 (85 FR 17835), FRA proposed metrics and minimum standards for measuring the performance and service quality of intercity passenger train operations. FRA held a telephonic public hearing on April 30, 2020. Written comments on the proposed rule were required to be submitted no later than June 1, 2020.

FRA received more than 320 comments, including comments from: Alabama State Port Authority, Alaska Railroad, American Association of State Highway and Transportation Officials, Association of American Railroads, Association of Independent Passenger Rail Operators, BNSF Railway Company, California State Transportation Agency, Canadian National Railway Company, Canadian Pacific, Capitol Corridor Joint Powers Authority, CSX Transportation, Environmental Law and Policy Center, Metropolitan Transportation Authority, Midwest Interstate Passenger Rail Commission, New York State Department of Transportation (DOT), NJ Transit, Norfolk Southern Railway Company, North Carolina DOT, Rail Passengers Association, San Joaquin Regional Rail Commission, Southeastern Pennsylvania Transportation Authority, Southern Rail Commission, States for Passenger Rail Coalition, Surface Transportation Board (STB), Transportation for America, Union Pacific Railroad Company, Utah Rail Passengers Association, Virginia Department of Rail and Public Transportation, Virginia Railway Express, Washington State DOT, the Honorable U.S. Representative Sam Graves, the Honorable U.S. Representative Rick Crawford, and more than 290 other individuals. Comments are addressed in the preamble.

C. Economic Analysis

All costs of this final rule are expected to be incurred during the first year. The following table shows the total 10-year costs of this final rule.

Total 10-Year Costs

Category	Total cost ($)	Annualized, 7 percent ($)	Annualized, 3 percent ($)
Cost of Meetings	473,473	67,412	55,505
Internal Staff Time (Preparation for Meetings)	296,991	42,285	34,816
Monthly Letters	50,328	7,166	5,900
Arbitration	714,030	101,662	83,706
Ridership Data	6,198	882	727
Total	1,541,020	219,407	180,655

This final rule may result in lower operational costs for Amtrak to the extent it results in improved OTP, which may reduce labor costs, fuel costs, and expenses related to passenger inconvenience, and provide benefits to riders from improved travel times and service quality. Due to the difficulty in quantifying future benefits to rail routes from improved OTP, combined with the inability to quantify the potential synergistic effects that improved OTP reliability could have across Amtrak's network, FRA has not quantified any potential benefits from lower operational costs or improved service that may result from the final rule.

II. Background

A. PRIIA

On October 16, 2008, President George W. Bush signed the Passenger Rail Investment and Improvement Act of 2008, Public Law 110-432, 122 Stat. 4907 (PRIIA) into law. Section 207 of PRIIA requires FRA and Amtrak to develop jointly new or improved metrics and minimum standards for measuring the performance and service quality of intercity passenger train operations, including: Cost recovery, on-time performance and minutes of delay, ridership, on-board services, stations, facilities, equipment, and other services.

Section 207 also calls for consultation with STB, rail carriers over whose rail lines Amtrak trains operate, States, Amtrak employees, and groups representing Amtrak passengers, as appropriate.

Section 207 further provides that the metrics, at a minimum, must include: The percentage of avoidable and fully allocated operating costs covered by passenger revenues on each route; ridership per train mile operated; measures of on-time performance and delays incurred by intercity passenger trains on the rail lines of each rail carrier; and, for long-distance routes, measures of connectivity with other routes in all regions currently receiving Amtrak service and the transportation needs of communities and populations that are not well-served by other forms of intercity transportation. Section 207 requires Amtrak to provide reasonable access to FRA to carry out its duty under section 207.

Section 207 provides that the Federal Railroad Administrator must collect the necessary data and publish a quarterly report on the performance and service quality of intercity passenger train operations, including: Amtrak's cost recovery, ridership, on-time performance and minutes of delay, causes of delay, on-board services, stations, facilities, equipment, and other services.

Finally, section 207 provides that, to the extent practicable, Amtrak and its host rail carriers shall incorporate the Metrics and Standards into their access and service agreements (also referred to as operating agreements).

The Metrics and Standards also relate to section 213 of PRIIA (codified at 49 U.S.C. 24308(f)). Section 213 states that if the on-time performance of any intercity passenger train averages less than 80 percent for any 2 consecutive calendar quarters, or the service quality of intercity passenger train operations for which minimum standards are established under section 207 fails to meet those standards for 2 consecutive calendar quarters, STB may initiate an investigation. Under section 213, STB shall also initiate such an investigation upon the filing of a complaint by Amtrak, an intercity passenger rail operator, a host freight railroad over which Amtrak operates, or an entity for which Amtrak operates intercity passenger rail service. Section 213 further describes STB's investigation and STB's related authority to identify reasonable measures and make recommendations to improve the service, quality, and on-time performance of the train and to award damages and prescribe other relief.

B. 2010 Metrics and Standards

In March 2009, FRA published proposed Metrics and Standards, which were jointly developed with Amtrak. After receiving and considering comments, FRA published final Metrics and Standards in May 2010. However, the 2010 Metrics and Standards were subject to a legal challenge on the basis that section 207 of PRIIA was unconstitutional. In 2016, the United States Court of Appeals for the District of Columbia Circuit found that paragraph (d) of section 207 was unconstitutional, and this holding had the effect, in part, of voiding the 2010 Metrics and Standards. Following additional litigation, that Court also found that paragraphs (a) through (c) of section 207 were constitutional and remained in effect (this decision became final upon the U.S. Supreme Court's denial of certiorari on June 3, 2019). As a result, in July 2019, FRA and Amtrak once again began the process of developing joint Metrics and Standards under section 207(a).

C. Stakeholder Consultation

Consistent with section 207(a), FRA and Amtrak consulted with many stakeholders to develop the Metrics and Standards.

Specifically, in August and September, 2019, FRA met individually with representatives of the following Class I railroads that host Amtrak trains: BNSF Railway, Canadian National Railway, Canadian Pacific Railway, CSX Transportation, Norfolk Southern Railway Company, and Union Pacific Railroad. On September 5, 2019, FRA and Amtrak met with representatives of the Rail Passengers Association. On September 10, 2019, FRA and Amtrak met with representatives of the Metro-North Railroad. On September 12, 2019, FRA and Amtrak met with representatives of the Transport Workers Union. On September 13, 2019, FRA and Amtrak met with Surface Transportation Board staff. On September 18, 2019, FRA and Amtrak convened a meeting with members of the State-Amtrak Intercity Passenger Rail Committee, whose members include: Caltrans, Capitol Corridor Joint Powers Authority, Connecticut DOT, Illinois DOT, Los Angeles-San Diego-San Luis Obispo Joint Powers Authority, Massachusetts DOT, Michigan DOT, Missouri DOT, New York State DOT, North Carolina DOT, Northern New England Passenger Rail Authority, Oklahoma DOT, Oregon DOT, Pennsylvania DOT, San Joaquin Joint Powers Authority, Texas DOT, Vermont Agency of Transportation, Virginia Department of Rail and Public Transportation, Washington State DOT, and Wisconsin DOT. On September 20, 2019, Amtrak met separately with representatives of the Union Pacific Railroad. On September 24, 2019, FRA and Amtrak met with representatives of the Vermont Railway. On November 15, 2019, Amtrak met separately with representatives of the BNSF Railway. On November 19, 2019, in two different meetings, FRA met separately with, first, representatives of the International Association of Sheet Metal, Air, Rail, and Transportation Workers, Transportation Division, and, second, with members of the Surface Transportation Board.^[1] FRA and Amtrak also sought input from other potentially interested entities who did not express interest in consulting at that time.^[2]

After publishing the NPRM, FRA invited each of the stakeholders to meet again. As a result of this invitation, on April 23, 2020, FRA met via telephone with representatives of the following Class I railroads that host Amtrak trains: BNSF Railway; Canadian National Railway; Canadian Pacific Railway; CSX Transportation; Norfolk Southern Railway Company; and Union Pacific Railroad. Representatives of the Association of American Railroads and Amtrak also attended this meeting. On May 6, 2020, FRA met via telephone with representatives of the American Association of State Highway Transportation Officials, Capitol Corridor Joint Powers Authority, Connecticut DOT, California DOT, Illinois DOT, Michigan DOT, Missouri DOT, North Carolina DOT, New York State DOT, Northern New England Passenger Rail Authority, Oklahoma DOT, Oregon DOT, San Joaquin Joint Powers Authority, Vermont Agency of Transportation, Virginia Department of Rail and Public Transportation, Washington State DOT, Wisconsin DOT, State Amtrak Intercity Passenger Rail Committee, and States for Passenger Rail Coalition. Representatives of Amtrak also attended this meeting. Lastly, on May 8, 2020, FRA met with representatives of STB. Representatives of Amtrak also attended this meeting. FRA placed summaries of each of these meetings, including the presentation material, in the NPRM's rulemaking docket (FRA-2019-0069-0013, FRA-2019-0069-0022, and FRA-2019-0069-0028).

In addition, on June 17, 2020, FRA met individually via telephone with BNSF Railway, Canadian National Railway, CSX Transportation, Norfolk Southern Railway Company, and Union Pacific Railroad. Representatives of Amtrak attended each of these meetings. On June 19, 2020, FRA met via telephone with Canadian Pacific Railway. Representatives of Amtrak attended this meeting. In these six meetings, FRA sought collaborative commitment to affirm or adjust the intercity passenger train schedules published for stations served across the railroad's network, and continued discipline to maintaining schedules, in order to expand the growing data pool that would support any necessary schedule change. Subsequent FRA letters to these parties summarizing the discussion were placed in the NPRM's rulemaking docket (FRA-2019-0069-0379). On July 31, 2020, FRA met collectively via telephone with Amtrak, BNSF Railway, Canadian National Railway, Canadian Pacific Railway, CSX Transportation, Norfolk Southern Railway Company, and Union Pacific Railroad regarding reaffirmation or reconciliation of Amtrak's published train schedules. FRA's subsequent letter to those parties summarizing the discussion was placed in the NPRM's rulemaking docket (FRA-2019-0069-0382).

D. Amtrak's Role in the Metrics and Standards Rulemaking

Beginning in July 2019, FRA and Amtrak began the process of developing the Metrics and Standards under section 207(a) of PRIIA. FRA and Amtrak held an executive kick-off meeting to initiate the effort, which was followed by a regular cadence of staff level meetings. As described above, FRA and Amtrak then conducted an extensive consultation process with many stakeholders to develop the Metrics and Standards. After the conclusion of the consultation process, FRA worked with Amtrak to develop the Metrics and Standards, which included extensive Amtrak input that was reflected in the Metrics and Standards NPRM. After publication of the NPRM, FRA met with various stakeholders (Class I railroads, States, and the STB) together with Amtrak, as described above. FRA then sought (and received) Amtrak's input on the draft Metrics and Standards final rule, considered Amtrak's input, and then, as the agency with rulemaking authority, FRA ultimately determined the contents of this final rule.

III. Response to Comments on On-Time Performance and Train Delays

A. Customer On-Time Performance

As proposed in the NPRM, this final rule measures the OTP element of intercity passenger train performance using a customer OTP metric, defined as the percentage of all customers on an intercity passenger rail train who arrive at their detraining point no later than 15 minutes after their published scheduled arrival time, reported by train and by route.^[3] The customer OTP metric focuses on intercity passenger train performance as experienced by the customer. Customer OTP measures the on-time arrival of every intercity passenger customer, including those who detrain at intermediate stops along a route and those who ride the entire route.

The customer OTP metric is calculated as follows: The total number of customers on an intercity passenger rail train who arrive at their detraining point no later than 15 minutes after their published scheduled arrival time, divided by the total number of customers on the intercity passenger rail train.^[4] For example:

The following table provides a hypothetical customer OTP calculation for a single train on two separate days. The table provides the minutes late, arrival status (“OT” for on-time, “LT” for late), total number of customer arrivals, and number of on-time customer arrivals, by station, for each day of operation and the two days overall.

In this example, customer OTP is 100% on day 1, 68% on day 2, and 84% for the two days combined. Because the number of customers on this train is different by station and by day, the aggregate customer OTP over the period is not a simple average of the daily numbers.

As also proposed in the NPRM, this final rule establishes a minimum standard for customer OTP of 80 percent for any 2 consecutive calendar quarters. To promote clarity and compliance, the customer OTP standard is the only standard set forth in connection with the OTP and train delays metrics. FRA believes this single standard is the most effective way to achieve dedicated focus on improving on-time performance. FRA emphasizes that 80 percent is a minimum standard, and FRA expects some intercity passenger rail services will reliably achieve a higher standard of performance. The 80 percent customer OTP standard is consistent with the statutory requirement in 49 U.S.C. 24308(f)(1).

Lastly, the final rule includes a provision not proposed in the NPRM, which provides that the customer OTP standard shall apply to a train beginning on the first full calendar quarter after May 17, 2021. For example, if the final rule is published on December 10, 2020, 6 months after that date would be June 10, 2021, and the first full calendar quarter after that would run from July 1, 2021 to October 31, 2021. FRA also understands that in some instances the alignment of a train schedule with the customer OTP metric may require additional time. As such, if Amtrak and a host railroad do not agree on a new train schedule and the schedule is reported as a disputed schedule on or before May 17, 2021, then the customer OTP standard for the disputed schedule shall apply beginning on the second full calendar quarter after May 17, 2021. FRA added these provisions to the final rule to ensure host railroads and Amtrak have sufficient time to align their train schedules before FRA begins reporting the customer OTP metric data.

FRA received hundreds of comments on customer OTP. Some commenters supported the customer OTP metric and standard and some disapproved of it. Many commenters generally supported the use of a single metric to measure OTP and the use of a single OTP standard.

Several commenters stated that section 207 requires the OTP metric to show OTP by host railroad in routes with multiple host railroads. In support, these commenters cited language in section 207(a), which states that the metrics “at a minimum, shall include . . . measures of on-time performance and delays incurred by intercity passenger trains on the rail lines of each rail carrier . . . .” FRA disagrees. As further described below, PRIIA calls for measuring the intercity passenger train's OTP performance, not the host railroad's performance in hosting the intercity passenger train. Section 207, when viewed in its entirety, does not require distinguishing OTP by host railroad. Sec. 207(a) (Requiring the development of metrics and minimum standards “including on-time performance and minutes of delay . . . .”); § 207(b) (Requiring FRA quarterly reporting on intercity passenger train operations, “including . . . on-time performance and minutes of delay . . . .”). Indeed, other sections in PRIIA require an OTP metric that measures a train's performance over an entire route, and not just route segments by host railroad. 49 U.S.C. 24710(a) and (b); see also 49 U.S.C. 24308(f)(1). Furthermore, an OTP metric that measures a host railroad's performance would not depict the customer's experience as passenger trains that arrive late at their destinations may be reported as “on-time.” Lastly, Congress emphasized the importance of measuring delays by host railroad as evidenced in section 213, which requires the STB to investigate whether and to what extent delays are due to causes that could reasonably be addressed by a host railroad. Thus, in compliance with section 207(a), this final rule does include train delay metrics that describe train performance on individual host railroads (e.g., the host running time metric shows train performance over a host railroad as compared to the train's scheduled running time, thereby distinguishing host railroads on multi-host railroad routes).

Regardless of whether the statute requires it, several commenters stated that the final rule should distinguish OTP by host railroad.^[5] In support, these commenters noted that the OTP metric determines when a host railroad may be subjected to an STB investigation (and other delay metrics could not prevent the initiation of an investigation). In other words, these commenters expressed concern that a host railroad could be subject to an STB investigation and/or reputational harm even if its own performance did not cause the train to operate below the standard.^[6] In related comments, commenters stated that the OTP calculation should exclude certain delays for which the host railroad was not responsible (e.g., third party delays or Amtrak-responsible delays) and give host railroads in dense metro territories an “out-of-slot delay tolerance” in connection with the OTP calculation.

In this final rule, FRA's approach to OTP follows the framework Congress set forth in PRIIA. Section 207 calls for measuring the intercity passenger train's OTP performance, not the host railroad's performance in hosting the intercity passenger train.^[7] A host railroad-specific measurement of OTP, accounting for late handoffs, slot time adjustments, and other methods of relief, would result in a system that is misaligned with the customer experience: passenger trains that arrive late at their destinations but are reported as “on-time.” Other sections in PRIIA also require an OTP metric that measures a train's performance over an entire route (that can be compared to other routes), and not just route segments by host railroad.^[8] In addition, Congress specifically identified the OTP metric as a trigger for an STB investigation.^[9] 49 U.S.C. 24308(f)(1).

In any event, the train performance metrics in this final rule do not penalize host railroads for train delays for which they are not responsible. As described below, the final rule's train delays metric and host running time metric speak to the individual host railroad's performance. One commenter stated that the NPRM's train delays metrics are likely to get little attention compared to the customer OTP metric. FRA strongly disagrees. While the customer OTP metric provides a train-level view of actual passenger train performance focused on the customer experience, the train delays metric and the host running time metric can help identify certain categories of delays, their frequency, and their duration, which are central inquiries to understanding and improving passenger train performance, as well as an STB investigation under 49 U.S.C. 24308(f).

In addition, that STB can initiate an investigation certainly does not mean that an investigation will be sought. As acknowledged by several commenters, an STB investigation results in resource expenditures for affected entities, and it has an uncertain outcome. A decision to initiate such an investigation is not made lightly. As a result, it is not reasonable to assume that every train below the minimum OTP standard would be investigated. Furthermore, it is also not reasonable to assume that an STB investigation would be sought against a host railroad where the train delays metric and the host running time metric data do not support an investigation. FRA is confident STB can identify delays for which host railroads are not responsible when armed with data from these metrics.

In lieu of a customer OTP metric, several commenters proposed a key stations OTP metric that would measure train performance at key stations on a host railroad.^[10] The customer OTP metric measures train OTP for every passenger at every station (not just passengers at designated stations), recognizes the relative importance of reliability at stations serving more passengers, and provides flexibility if demand changes. In contrast, a key stations OTP metric fails to recognize the importance of customers who do not use a key station. Such a metric would have additional challenges, including how to identify key stations. For these reasons, FRA determined that the customer OTP metric is superior to a key stations OTP metric. With that said, the customer OTP metric resembles a key stations OTP metric because stations with many detraining passengers have greater influence on the train's customer OTP and serve as de facto key stations.^[11] As discussed elsewhere in this final rule, FRA finds that, aside from predictable and broadly understood seasonal trends and short-term variability, the percentage of a train's detraining passengers at stations on a route is stable for purposes of calculating customer OTP; therefore, host railroads can identify key stations to maximize performance under the customer OTP metric.

Another commenter suggested that the existing, contractually negotiated Amtrak train performance provisions found in the host railroads' operating agreements with Amtrak are preferable to the customer OTP metric because the host railroads often perform well under those contract terms (whereas these same trains don't perform as well when measured by the customer OTP metric). The commenter stated that Amtrak and a host railroad should be allowed to develop and apply alternative OTP standards, such as the existing contractual performance provisions, or use mutually agreed upon times as a baseline to measure OTP. The commenter's proposal is counter to section 207's requirement to establish a metric to measure intercity passenger train performance, as it would result in many different measures of performance that would be, at best, difficult to understand and, at worst, entirely misleading. A single OTP metric and standard allows stakeholders to compare train performance, which may be important to evaluating connectivity information, among other things, and ensures all trains are held to the same standard.

Furthermore, FRA believes the OTP metric should measure train performance from the eyes of the customer. The customer OTP metric is meaningful, precisely because it is reflective of the passenger train's actual performance. The commenter's proposal would routinely produce the anomalous result stated elsewhere in this final rule of a passenger train that arrives late at stations yet has good “OTP.” See Application of the National Railroad Passenger Corporation Under 49 U.S.C. 24308(a)—Canadian National Railway Company, STB Docket No. FD 35743 at 10 (Aug. 9, 2019) (“In general, if an OTP metric only includes checkpoints at the final station and two or three select intermediate points, . . . , the metric does not measure performance in a way that captures whether a significant portion of Amtrak's passengers actually arrived at their selected destinations on time. Such a metric would be an unrepresentative measure of performance.”).

Another commenter stated the final rule should adopt an all-stations OTP metric that would measure train performance at all stations on a route. Like an all-stations OTP metric, the customer OTP metric measures train performance at every station, and it also recognizes the importance of reliability at stations serving more passengers. Customer OTP also offers host railroads more flexibility in adjusting recovery time ^[12] based on passenger load versus recovery needed for every station stop.^[13] For these reasons, FRA determined that the customer OTP metric is preferable to an all-stations OTP metric, and is adopting a customer OTP metric as proposed in the NPRM.

A commenter stated that FRA should have considered the impact of the customer OTP metric and standard on the host railroads' various operating agreements with Amtrak, including the performance incentive payments made under such agreements. FRA is not a party to these agreements, nor does FRA have knowledge of their details, as the parties consider the details of the agreements confidential business information, and have not shared them with FRA. More importantly, this final rule does not require a change to the performance incentive payment provisions in these operating agreements; Amtrak and the host railroads may continue to maintain those provisions as they see fit.

In addition, to the extent a host railroad is concerned with receiving lower performance incentive payments as a result of this final rule, this final rule does not prohibit a host railroad and Amtrak from revising the performance incentive payments to align better with the customer OTP metric and standard.^[14] Indeed, section 207(c) provides that, to the extent practicable, Amtrak and its host rail carriers shall incorporate the metrics and standards into their access and service agreements (the operating agreements). See also Union Pac. R.R. Co. v. Surface Transp. Bd., 863 F.3d at 826 (“The § 207 on-time-performance metric was, to the extent practicable, to be incorporated into Amtrak's contracts with host railroads.”).

A commenter stated that because the customer OTP metric is based on passenger loads it may be an unstable metric (as it may vary on a daily basis). Another commenter stated that this instability would result in lengthening schedules. A commenter also stated that the aggregation of customer OTP data could produce distorted results showing a train service as more reliable or less reliable than is actually the case. And, another commenter stated that the customer OTP metric will likely result in false positives for trains that depart late from congested Amtrak terminals. FRA does not agree with these commenters that customer OTP will be unreliable for two reasons. First, Amtrak has provided some ridership data to host railroads and the ridership data metric in this final rule requires Amtrak to provide additional data to host railroads to allow them to understand and monitor passenger loads.^[15] Second, while the actual number of detraining passengers may change at a station over time, the percentage of passengers detraining at a station is generally stable.^[16] Based on FRA's review of the non-public ridership data Amtrak made available to the host railroads,^[17] FRA found little movement in a station's relative volume of detraining passengers. For example, there were 15,714 total passengers on Amtrak train #391 (on the Illini/Saluki route) in the fourth quarter of 2019, and 10,481 total passengers in the first quarter of 2020, a difference of 5,233 passengers or 33%. Passengers detraining at Champaign-Urbana, IL represented 47.8% of the total passengers on the train in the fourth quarter 2019, and 50.4% of total passengers in the first quarter 2020. Despite this variation in ridership, Champaign-Urbana ranked as the highest volume station for detraining passengers for these two quarters compared to all other stations on the route. Similarly, Carbondale, IL ranks as the second highest volume station for detraining passengers, with 27.1% of the total passengers on the train in the fourth quarter 2019, and 25.6% of total passengers in the first quarter 2020. The relative importance of the station (i.e., the station rank) along the route seldom changes despite fluctuation in the percentage of detraining passengers. As stated above, if carefully analyzed, the ridership data will allow host railroads to identify de facto “key stations” to concentrate performance to ensure most passengers arrive at their destination on-time (thereby meeting the 80% standard).

A commenter stated that host railroads do not have adequate notice of the customer OTP metric because the metric is based on the number of detraining passengers at a station, which the host railroads would receive after the fact. As noted above, there is generally not much change in proportional ridership by station by route (real-time ridership data is of limited utility), and host railroads already received a year of performance data on May 18, 2020. Furthermore, as described below, this final rule includes a ridership data metric that, in part, requires Amtrak to provide ridership data to host railroads. In addition, the final rule provides that the customer OTP standard shall apply to a train beginning, at the earliest, on the first full calendar quarter after May 17, 2021. Amtrak and the host railroads will also have at least a further five months to evaluate two years of relevant ridership data to work towards certifying train schedules, consistent with the data sharing requirement in this final rule. This commenter further suggested an alternative OTP metric that measures OTP by the train's arrival at designated check-points (similar to the approach used in the commenter's operating agreement with Amtrak), which it alleged would provide adequate notice. For the reasons stated above, FRA disagrees with this approach and believes that the OTP standard should be based on the passenger experience.

A commenter stated that a single OTP metric may fail to address certain State-supported trains that have negotiated local expectations of performance with a host railroad and that currently serve passengers reliably above the 80 percent OTP standard. Similarly, another commenter stated that where an existing partnership exists between a State and a railroad, such as a service outcome agreement, the OTP metric and standard should be used to inform and complement that agreement, rather than to supersede it. As stated, the 80 percent customer OTP standard is a minimum standard. FRA expects many services to operate more reliably and this final rule is not intended to obstruct the unique performance arrangements that may exist between host railroads and States.

Some commenters expressed concern that the customer OTP metric would delay commuter rail trains sharing the right-of-way with Amtrak trains due to Amtrak trains “waiting for time” (i.e., when a train arrives early to a station and waits until its scheduled departure time) at intermediate stations. A commenter stated that such an action in high density territory could create a net reduction in rail line capacity. Similarly, other commenters stated that aligning schedules to a customer OTP metric enlarges an Amtrak train's dispatch footprint by redistributing recovery time across intermediate stations, which threatens overall network fluidity, decreases the host railroad's ability to manage slow orders, and will result in longer schedules. FRA disagrees. First, delays waiting for time at intermediate stations can be foreclosed by an accurate schedule. Second, adjusting train schedules to align with the customer OTP standard does not mean that recovery time must be added for each station. Recovery time should, for example, be included across a schedule to protect performance at larger volume stations, locations where passenger trains can wait clear of main tracks, where stations are farther apart, or where trains are more likely to incur operational delays. However, spreading existing recovery time linearly across a schedule would be inefficient and would be more likely to result in trains waiting at stations for departure times if a train performed well on a given segment that included additional, unnecessary recovery time. Furthermore, in the case of capacity impacts great enough to warrant schedule change, reductions of time to remove these waits would be in both parties' favor. Third, Amtrak trains on many routes avoid large numbers of station stops in districts already well served by commuter operations. Lastly, Amtrak trains should not be given more time between stations in commuter train territory than the commuter trains themselves. In these types of territories there should be little slack time written into the schedule, consistent with standard railroad operating best practices. For all these reasons, FRA is confident that the professional railroaders at Amtrak and the host railroads, whose daily job it is to develop train schedules, can account for the issues raised by these commenters.

Another commenter suggested that the customer OTP metric penalizes trains that perform well according to the performance provisions in their Amtrak-host railroad bilateral operating agreement and is not consistent with the intent of section 207. In support, the commenter, a host railroad, stated that it receives payments under its contract with Amtrak for the performance of trains operating on its right-of-way, but is concerned these same trains will not perform well as measured by a customer OTP metric. FRA disagrees. Put simply, a measure that is not focused on when a passenger train arrives at a station is not measuring the on-time performance of the passenger train. FRA encourages Amtrak and the host railroads to work toward aligning the bilateral operating agreements with the customer OTP metric and standard to ensure performance is measured, and appropriately incentivized, in a consistent manner. See PRIIA § 207(c).

A commenter sought clarity regarding whether the customer OTP metric is measured by the actual number of passengers detraining at a station, or by the number of tickets that Amtrak sells to a specific arrival station. Amtrak measures detraining passengers by the number of passengers actually traveling on the train, as determined by conductor ticket collections via electronic ticket scanning for a specific arrival station. Passengers who have reserved a seat, but elect not to travel, are not reflected in passenger counts. Another commenter wondered whether it is possible for Amtrak to calculate customer OTP accurately where Amtrak customers share tickets in metro areas with commuter passenger railroads (e.g., in Los Angeles with Metrolink commuter rail services). Most passengers traveling on Amtrak under a cross-honor arrangement with a commuter rail operator are included in the customer OTP calculation (in most cases, the conductor records the origin and destination station for the cross-honor rider as they board). Amtrak maintains cross-honor agreements with several commuter passenger railroads across the country, and riders traveling under those arrangements represent 2.4% of total Amtrak ridership. Approximately two-thirds of these cross-honor passengers are included in Amtrak detraining counts, including Metrolink and Virginia Railway Express cross-honors.

A commenter stated a concern that, under the customer OTP metric, Amtrak passengers on cancelled trains would be counted as late customer arrivals at their ticketed station if service to their ticketed station is cancelled. In this case, a passenger on a train that has had their ticket scanned and the service to their ticketed station canceled on less than four hours advance notice is counted as a late customer arrival at their ticketed station by design, as it reflects the customer's experience.^[18] In Amtrak fiscal year 2019, the number of passengers impacted by en route cancellations to their detraining stations was 0.04% of Amtrak ridership (14,439 impacted passengers divided by 32,519,241 total passengers).

A commenter stated that the customer OTP metric should be reported by train only, and not by train and by route. However, it is important to maintain route reporting because the customer is less likely to know what train number they are on, and are more likely to know the route they travel.

Lastly, a commenter stated that the customer OTP metric and standard should consider the fluidity of the entire network in determining whether a host railroad has given an Amtrak train preference. Preference under 49 U.S.C. 24308(c) is determined by STB, not FRA. See 49 U.S.C. 24308(c) and (f)(2). The commenter also stated that the customer OTP metric should consider non-Amtrak passengers, in addition to Amtrak passengers. As described further below, FRA developed the metrics for Amtrak intercity passenger train operations, which is consistent with section 207.

B. Train Schedules

While the NPRM did not propose any metrics related to train schedules, FRA received many comments about train schedules. Some commenters stated that the final rule should require Amtrak and a host railroad to certify that a train's schedule aligns with the customer OTP metric and standard before the customer OTP standard takes effect. STB, for example, supported requiring properly aligned schedules before an OTP standard takes effect. In support, commenters stated that many of Amtrak's existing schedules are not a meaningful benchmark for measuring customer OTP because they were not designed for a customer OTP metric, and they are outdated and unrealistic. As a result, these commenters stated, the use of the customer OTP metric to measure Amtrak schedules would produce misleading train performance data, and may result in unnecessary STB litigation.

Further, some commenters stated that it would be challenging to renegotiate some schedules due to disagreements about train scheduling and challenges with existing schedules, among other reasons. Several commenters stated that the final rule should provide an initial six-month period for Amtrak and the host railroads to certify schedules, and should extend this period for the pendency of any dispute resolution process. Commenters also stated that the final rule should incorporate a dispute resolution process to address schedules in dispute. Several commenters also stated that the dispute resolution process should automatically certify a schedule if the host railroad refused to participate and, conversely, should withhold certification if Amtrak refused to participate. Some commenters stated that the final rule should include a schedule recertification process to ensure ongoing schedule validity.

FRA generally agrees with many of these observations (although not all). FRA agrees that Amtrak and the host railroads should align schedules with the customer OTP metric.^[19] Where a train's OTP is measured against the train schedule provided to the public, the train's schedule should be aligned with the OTP measure used to evaluate the train's performance. Historically,

Amtrak's published train schedules have not been designed with a customer OTP metric in mind. Accordingly, this final rule: Establishes a certified schedule metric that addresses alignment with the customer OTP metric and standard; provides more time for Amtrak and the host railroads to negotiate schedules; and allows for a dispute resolution process if the parties disagree.^[20]

The certified schedule metric first requires Amtrak to report the number of certified schedules, uncertified schedules, and disputed schedules, by train, by route, and by host railroad.^[21] This information is reported monthly for six months, at 12 months, and yearly thereafter. Second, the final rule provides more time to negotiate schedules by delaying application of the customer OTP standard until the first full calendar quarter six months after publication of the final rule. Third, the final rule encourages the parties to certify schedules timely and to resolve disagreements by further delaying application of the OTP standard when a non-binding dispute resolution process is engaged. Specifically, if a train schedule is reported as a disputed schedule during the first six months, then the customer OTP standard does not apply until the second full calendar quarter following those six months.^[22] Fourth, the certified schedule metric further encourages the parties to certify schedules by requiring Amtrak and a host railroad to transmit monthly letters signed by their chief executive officers to Congress (and others) when they have an uncertified schedule after six months.^[23] These letters will make policymakers aware of the status of the train schedule,^[24] and help ensure that a sense of urgency is maintained by the parties to resolve the disagreement. Lastly, the certified schedule metric recognizes that ongoing coordination between Amtrak and a host railroad is needed as certified schedules are impacted by future events.^[25] The graphic below provides an overview of the certified schedule metric process.

A commenter stated that a schedule dispute resolution process should allow for both non-binding and binding dispute resolution (and should not require binding dispute resolution only). Here, the final rule does not require Amtrak or a host railroad to engage in a dispute resolution process, nor does the final rule attempt to prescribe the process the parties use if they do choose to engage a dispute resolution process. However, the final rule only affords delay of the customer OTP standard beyond six months for engagement of a non-binding dispute resolution process.^[26] The resolution of a schedule disagreement must be achieved as quickly as possible. The final rule encourages Amtrak and host railroads who are serious about finding common ground on a schedule to engage in a dispute resolution process if they are unable to reach agreement amongst themselves.^[27] While non-binding, the written decision resulting from a non-binding dispute resolution process may facilitate resolution and may also assist the Surface Transportation Board in a 49 U.S.C. 24308(f) investigation. While parties may seek binding dispute resolution, this final rule does not include that process given the broad array of impacts that may occur from a schedule required by arbitration, such as, among other things, significant additional operating expenses or revenue losses (for Amtrak and its partners), commercially infeasible times of operation or duration, and conflicting schedules on multi-host railroad routes.

Some commenters stated it would be unfair to apply a customer OTP standard to a schedule that is not aligned with the customer OTP metric (because the metric could produce misleading train performance data that could ultimately result in an STB investigation).^[28] A commenter also stated that Amtrak has no incentive to adjust its schedules, and other commenters expressed concern about lengthening schedules. FRA understands that Amtrak and host railroads have some competing interests. This final rule balances those interests consistent with section 207. As explained, the final rule encourages the parties to agree on certified schedules while not explicitly requiring them. In addition, a host railroad or Amtrak may initiate a timely non-binding dispute resolution process (regardless of whether the other party agrees to participate in that process), which would temporarily delay application of the OTP standard to a train. The non-binding dispute resolution process will produce a written decision that will inform Amtrak and a host railroad in aligning the schedule with the customer OTP metric. The final rule empowers Amtrak and the host railroads to resolve schedule disputes without being overly prescriptive (and without government involvement that could hamper the parties' ability to engage in confidential discussions, among other things). Section 207 does not require schedule certification and, indeed, section 213 acknowledges that STB investigations may include STB review of the extent to which scheduling contributed to delay. 49 U.S.C. 24308(f)(1).

Many comments addressed the NPRM's train schedule principles, which recommended, but did not require, alignment of train schedules with the customer OTP metric. Some commenters stated that the principles should be removed, others supported their inclusion, and still others suggested adding to the principles. This final rule does not include the train schedule principles. FRA determined these principles are no longer necessary given the final rule's inclusion of a certified schedule metric; the NPRM's train schedule principles would only serve to complicate the process of determining train schedules for Amtrak and the host railroads.

Several commenters stated that State sponsors of intercity passenger rail should be included in Amtrak and host railroad schedule alignment discussions. FRA agrees that State sponsors are important stakeholders in these discussions. Although the final rule does not require nor prohibit a State sponsor's involvement, FRA expects that a State sponsor may be invited to participate consistent with their existing agreement(s). Based on the comments received, FRA understands that Amtrak and many of the host railroads have existing agreements with State sponsors that relate to schedules. Those agreements remain in place and are not altered or negated by this final rule.

Commenters also stated that Amtrak schedule modifications should not compromise the standardized schedules Amtrak has agreed to with commuter agencies in dense commuting territories, as these existing schedules allow for the optimal use of capacity and ensure reliable operations for both Amtrak and commuter rail operations. Similarly, a commenter stated that Amtrak, host railroads, and commuter services must work cooperatively to update schedules in the interest of providing achievable OTP goals. FRA recognizes the important role commuter rail services play in the passenger rail network. This final rule does not prohibit commuter agency involvement in Amtrak-host railroad schedule discussions, and any Amtrak and/or host railroad agreements with commuter agencies remain in place and are not altered or negated by this final rule.

A commenter stated that there should be a test period for new schedules. With the application provisions for the OTP standard in this final rule, FRA believes Amtrak and the host railroads have sufficient time to test and negotiate train schedules. FRA will not dictate a process for negotiating schedules, but it expects both parties will use data-driven processes, such as modeling, simulation, and real-world testing to validate any proposed schedule changes.

One commenter stated that a new schedule aligned with the customer OTP metric should take into account the existing contractual performance payments that may exist between Amtrak and a host railroad under their operating agreement. It is unnecessary to require new schedules to account for contractual performance payments because any new schedule will be agreed to by Amtrak and the host railroad, and they may consider the implications of the schedule on future performance payments, and can work to adjust those payments to align with the new schedule.

A commenter stated that Amtrak must provide the same consideration to other host railroads that Amtrak grants itself on the Northeast Corridor (NEC) and adjust scheduled running times to accommodate infrastructure work as appropriate. The commenter stated that Amtrak regularly adjusts scheduled running times for its trains on the segments of the NEC that it maintains and dispatches but does not grant similar running-time adjustments to Amtrak trains traversing other host railroad territory on the NEC. Considerations for running time impact are more properly addressed in the operating agreement between the parties.

Lastly, a commenter stated that Amtrak must provide the percentage of recovery time per route segment. FRA sees limited value in this metric and it is not included in this final rule. Together, a host railroad and Amtrak can arrive at an efficient use of recovery time, which is an inherent element in any schedule. Once a schedule is completed, a host railroad will know how much recovery time exists on each line segment for each train and between which stations the recovery time has been placed.

C. Train Delays

FRA recognizes that the customer OTP metric and standard should be accompanied by metrics that provide additional useful information about a train's performance. There are factors that contribute to poor OTP on a route that are not evident from measuring station arrival times alone. For example, an intercity passenger rail train dispatched by multiple hosts may experience delays on one host railroad but not on another host railroad. Because the customer OTP metric does not easily distinguish performance on individual host railroads (including Amtrak), this final rule also establishes metrics to measure train delays, station performance, and host running time, to provide more information about the customer experience, train performance on individual host railroads,^[29] and the minutes and causes of delay.

1. Train Delays

The NPRM proposed to define a train delays metric as the total minutes of delay for all Amtrak-responsible delays, host-responsible delays, and third-party delays, for the host railroad territory within each route.^[30] The NPRM further proposed to define the terms “Amtrak-responsible delays,” “host-responsible delays,” and “third party delays.”

Many commenters stated that the train delays metric should report delays by delay category (i.e., Amtrak-responsible delays, host-responsible delays, and third party delays). Several commenters also stated that the train delays metric should measure Amtrak delays as operator and as host railroad, in total and separately. Some commenters also stated that the final rule should report delays by root cause and that, in instances where Amtrak and the host railroads disagree on the causes of delay, FRA should publish both findings. In addition, several commenters stated that Amtrak and the host railroad should work together on a regular basis to identify and agree on the delay data and the delay causes.

In response to comments on the NPRM, the final rule includes a revised train delays metric. First, the train delays metric in the final rule reports disputed delay minutes, which are those non-Amtrak host responsible delays disputed by the host railroad and not resolved by Amtrak. This additional information captures host-responsible delays disputed by the host railroad pursuant to its operating agreement with Amtrak and not resolved by Amtrak. It is important to note that FRA views the host railroad's National Railroad Passenger Corporation (NRPC) operations officer as a critically important position at the host railroad that demands direct access to the host railroad's chief operations officer and other senior leadership.^[31] In addition to reporting the number of disputed delay minutes, the final rule also provides that the train delays metric is reported by delay code by: Total minutes of delay; Amtrak-responsible delays; Amtrak's host-responsible delays; Amtrak's host-responsible delays and Amtrak-responsible delays, combined; non-Amtrak host-responsible delays; and third party delays. The table below is a sample train delay metric chart to further illustrate the metric.

One commenter stated that all departure and arrival times at each Amtrak station should be automated so that manual data collections by Amtrak conductors are minimized or eliminated. FRA agrees that Amtrak should use automated methods to collect data to the greatest extent practicable. In fact, Amtrak currently uses an automated electronic delay reporting system based primarily on a GPS-based system that automatically logs arrival, departure, and passing times at stations and other locations, and calculates the number of minutes of delay above pure run time within each segment of an Amtrak route. See Application of the National Railroad Passenger Corporation Under 49 U.S.C. 24308(a)—Canadian National Railway Company, STB Docket No. FD 35743 at 23 (Aug. 9, 2019).

Several commenters gave examples of types of delays that should not be designated as host-responsible delays, such as passenger delays to Amtrak trains while at a station, and other commenters expressed concern about Amtrak's identification of root causes of delay. FRA understands that Amtrak and the host railroads may disagree on how to assign responsibility for any particular delay. FRA also understands that some host railroads have processes and data systems in place through which they look closely at delay causes, and that other host railroads do not have such processes or systems and approach the issue in a different way. The train delays metric includes the reporting of disputed delays where Amtrak and the host railroad are unable to agree on a delay category pursuant to the existing process for delay attribution in the Amtrak-host railroad operating agreement.^[32] The metric's reporting of disputed delays ensures transparent reporting, while not prescribing an additional process for the parties to use to reach agreement or inserting FRA in the process to adjudicate disputes. FRA expects that Amtrak and the host railroad's NRPC officer (or equivalent) will be in frequent communication about train delays.

Lastly, one commenter stated that in other FRA and Amtrak reports, delay metrics have not been published for segments that are less than 15 miles in length. The commenter proposed that minutes of delay should be reported for each host railroad territory that exceeds 0.1 miles in length to ensure that delays on short segments (frequently near terminals) are also reflected, as these delays can have an outsized effect on customer OTP. FRA agrees. Amtrak collects delay data on all segments of a route regardless of segment length. The delay data for all segments are available to all host railroad partners via on-line access, and in some cases, automated data feeds. FRA's quarterly reports will include delays for all segments of the route.

2. Station Performance

The NPRM proposed an average minutes late per late customer metric as the average minutes late that late customers arrive at their detraining stations, reported by route (excluding on-time customers that arrive within 15 minutes of their scheduled time). A commenter stated that this metric does not provide information about the location of problems causing the delay or how to fix them, and that it does not differentiate between the performance of individual host railroads. Another commenter proposed that this metric should reflect average minutes late of all customers (not just the late customers).

In response to these comments, FRA is renaming the metric as a station performance metric, and revising it to measure the number of detraining passengers, the number of late passengers, and the average minutes late that late customers arrive at their detraining stations, reported by route, by train, and by station. The average minutes late per late customer calculation excludes on-time customers that arrive not later than 15 minutes after their scheduled time and reflects the severity of the delayed train, as experienced by the customer. To clarify, a customer who arrives at their detraining station 16 minutes late would be included in this calculation and would be recorded as 16 minutes late. The revised metric expands upon the proposed metric by providing information on all passengers, not just late passengers, by route, train, and station. It will offer FRA, hosts, and Amtrak customers more information on the location of performance problems and allow them to calculate the customer OTP metric.

The table below is a sample station performance metric chart to further illustrate the metric.

3. Host Running Time

The final rule establishes a host running time metric to measure the average actual running time and the median actual running time compared with the scheduled running time between the first and final reporting points for a host railroad segment set forth in the Amtrak schedule skeleton,^[33] reported by route, by train, and by host railroad (excluding switching and terminal railroads). For a given host railroad, the scheduled running time is defined as the scheduled duration of a train's travel on a host railroad, as set forth in the Amtrak schedule skeleton, and the actual running time is defined as the actual elapsed travel time of a train's travel on a host railroad, between the departure time at the first reporting point for a host railroad segment and the arrival time at the reporting point at the end of the host railroad segment. As delays may or may not cause a train to be late on its schedule, it is important to measure the performance of host railroads against the scheduled operation. The host running time metric shows the performance of a host railroad against the time allowed for in the schedule and provides more insight into a host railroad's operating impact on OTP. This metric is an indication of which host railroads may be responsible for chronic performance below standard and which ones are not. The metric will not explain the cause of delays, nor will it assign responsibility for them.

The table below is a sample host running time metric chart to illustrate the metric.

Several commenters stated that the NPRM did not distinguish between host railroads on multi-host railroad routes, and that delays on one host railroad can be carried over to a subsequent host railroad. FRA believes the host running time metric specifically addresses this concern by showing train performance over a host railroad as compared to the train's scheduled running time, thereby distinguishing host railroads on multi-host railroad routes.

Lastly, two commenters also stated that a late, out-of-slot Amtrak train can itself cause additional delays on the receiving host railroad.^[34] One commenter stated that the final rule should provide host railroads with an “out-of-slot delay tolerance” in calculating OTP that would account for Amtrak trains that arrive late to the host railroad and miss their scheduled slot. FRA disagrees. Amtrak trains that operate out-of-slot may pose operating issues in certain scheduled network areas where train operation distances are very short, dense, and tightly scheduled (i.e., commuter train territory around major metropolitan areas). However, outside of that situation, effective communication between a host railroad and Amtrak regarding an impending delay is generally the key to mitigate the impact of an out-of-slot Amtrak train. Further, as stated elsewhere in this final rule, FRA believes the most meaningful measurement of OTP is based on the customer experience of actually arriving at their destination on time, not obscured by other tolerance or relief.

4. Train Delays per 10,000 Train Miles

The NPRM proposed a train delays per 10,000 train miles metric as the minutes of delay per 10,000 train miles for all Amtrak-responsible and host-responsible delays, for the host railroad territory within each route. Several commenters stated that this metric is not informative as it does not provide data about the location of delays or how to fix them. One commenter stated that the metric can be helpful when comparing delays among different routes. The final rule includes this metric. Minutes of Amtrak-responsible delay and host-responsible delay have historically been normalized by 10,000 train miles to compare performance more easily on routes of varying length. This calculation is helpful when assessing an individual railroad's performance on a route that has more than one host.

D. Ridership Data

Many commenters stated that the final rule must require Amtrak to provide host railroads with sufficient data to calculate and monitor customer OTP. Without this information, these commenters stated, host railroads would not be able to verify the accuracy of customer OTP data, monitor their performance, identify improvement opportunities, or take corrective action. Commenters requested ridership data, such as: Close to real-time access to daily, station-specific Amtrak ridership data, including late arriving customers and the degree of lateness; daily numbers of detraining passengers for each Amtrak train on a station-by-station basis; four years of historical ridership data; the data underlying the customer OTP metric calculation; relevant route data on performance and Amtrak customer travel; and Amtrak's ridership projections.

During the NPRM's comment period, Amtrak agreed to provide some ridership data to the host railroads. See FRA-2019-0069-0295. In response, some commenters stated that this data was not sufficient because it was aggregated and did not show station-specific performance or the number of passengers detraining at each station.

In consideration of these comments, the final rule includes a ridership data metric. The ridership data metric is the number of host railroads to whom Amtrak has provided ridership data, reported by host railroad and by month. In addition, the ridership data metric requires that, not later than December 16, 2020, Amtrak must provide host railroad-specific ridership data to each host railroad for the preceding 24 months. Also, on the 15th day of every month following December 16, 2020, Amtrak must provide host railroad-specific ridership data to each host railroad for the preceding month. The final rule defines the term ridership data to mean, in a machine-readable format: The total number of passengers, by train and by day; the station-specific number of detraining passengers, reported by host railroad whose railroad right-of-way serves the station, by train, and by day; and the station-specific number of on-time passengers reported by host railroad whose railroad right-of-way serves the station, by train, and by day.

A commenter stated that ridership data should be available to the public. FRA's quarterly reports will be publicly available. FRA also recognizes that the ridership data may include information that Amtrak views as confidential/competitively sensitive. Although this final rule requires Amtrak to provide ridership data to host railroads, Amtrak may impose reasonable conditions on the host railroad's use of these data. With that said, at a minimum, the host railroad should be able to use these data in connection with negotiation, review, adjustment, or analysis of relevant Amtrak train schedules, or in connection with an STB proceeding under 49 U.S.C. 24308(f) involving the host railroad.

The tables below are samples of ridership data to illustrate further the format and data that Amtrak will share with host railroads under this metric (however, this supporting data will not be publicly available).

A commenter stated that Amtrak must share the ridership data with its State-supported route partners. FRA encourages Amtrak to share ridership data with its State-supported route partners; however, a requirement to share such data is not directly related to this rulemaking. Amtrak's provision of data to its State partners should be consistent with existing agreements. State entities that provide payments to Amtrak under PRIIA section 209 currently have access to some of Amtrak's online data systems, which include train delay information and ridership information.

Some commenters stated that the host railroad's current lack of access to station-specific ridership data limited their ability to comment on the NPRM, and that the customer OTP metric would not provide host railroads adequate notice. As discussed, above, any OTP standard adopted in this final rule must be relevant to the actual passenger experience; the most relevant of which is whether a passenger arrived at the destination on time. As noted previously, FRA finds that, aside from predictable and broadly understood seasonal trends, the percentage of a train's detraining passengers at stations on a route is stable for purposes of calculating customer OTP. In addition, host railroads have received some additional ridership data and will receive more ridership data under this final rule.

A commenter stated that Amtrak should describe how it collects the ridership data and its passenger-counting methodology. As stated, Amtrak measures detraining passengers by the number of passengers actually traveling on the train, as determined by conductor ticket collections via electronic ticket scanning for a specific arrival station. Passengers who have reserved a seat, but elect not to travel, are not reflected in passenger counts.

Lastly, a commenter stated that host railroads should be able to audit the ridership data provided by Amtrak. FRA determined the ridership data required by this final rule will allow a host railroad to calculate the customer OTP independently. In addition, Amtrak's reported ridership data is subject to verification by Amtrak's Office of the Inspector General.

IV. FRA Quarterly Reporting

Section 207(b) requires FRA to publish a quarterly report on the performance and service quality of intercity passenger train operations, including Amtrak's cost recovery, ridership, on-time performance and minutes of delay, causes of delay, on-board services, stations, facilities, equipment, and other services. FRA's first quarterly report on intercity passenger train performance will cover the first full calendar quarter 3 months after the date of publication of the final rule in the Federal Register. For example, if the final rule is published on December 10, 2020, three months after that date would be March 10, 2021, and the first full calendar quarter after that would run from April 1, 2021 to June 30, 2021.

The first quarterly report will include data on the customer service metrics, the financial metrics, the public benefits metrics, the certified schedule metric, the ridership data metric, the train delays metric, and the train delays per 10,000 train miles metric, but will not include data on the customer OTP metric, the station performance metric, or the host running time metric. Beginning with the second quarterly report, FRA will report data on all of the final rule's metrics, unless a train schedule is a disputed schedule on or before May 17, 2021. In that circumstance, FRA will report customer OTP metric data for that particular train beginning with the second full calendar quarter after May 17, 2021. In addition, in that circumstance, FRA will also not report data for the station performance metric or the host running time metric in connection with the host railroad(s) party to the disputed schedule. Unless otherwise specified, FRA will update metrics on a quarterly basis.

V. Section-by-Section Analysis of Comments and Revisions From the NPRM

This section responds to public comments and identifies any changes made from the provisions as proposed in the NPRM. Provisions that received no comment, and are otherwise being finalized as proposed, are not discussed again here. To review the complete section-by-section analysis in the NPRM, see 85 FR 20466.

Section 273.1 Purpose

This section provides that the final rule establishes metrics and minimum standards for measuring the performance and service quality of intercity passenger train operations.

A commenter sought clarity regarding non-Amtrak operators of intercity passenger rail trains and the metrics (and under what circumstances the STB may initiate an investigation of substandard performance). FRA developed the metrics for Amtrak intercity passenger train operations, which is consistent with section 207's many references to Amtrak, including: The development of the metrics; the entities to consult regarding the development of the metrics; specific metrics; FRA's access to information; and FRA's quarterly reports. This final rule does not apply to non-Amtrak operators of intercity passenger rail trains. Lastly, investigations of substandard performance under 49 U.S.C. 24308(f) are conducted by STB, and as such, STB alone determines when to initiate an investigation.

A commenter stated that FRA should put this rulemaking on hold and, together with the Federal Transit Administration and STB, convene a seminar with freight and passenger stakeholders to address comprehensively issues relating to the shared use of rail right-of-way. FRA appreciates the comment, and while such a meeting is outside the scope of this rulemaking, FRA is always working to advance rail policy and development, both on its own and in partnership with other federal agencies.

A commenter stated that the Metrics and Standards should not create a statutory preference for Amtrak over commuter operations or intercity passenger service operated by non-Amtrak carriers. Amtrak does have certain statutory rights regarding the use of facilities and preference over freight transportation in using a rail line, among other things. See, e.g., 49 U.S.C. 24308. The Metrics and Standards do not create any additional preference in law for Amtrak. Another commenter stated that FRA should identify actions that exhibit preference in the operating environment to facilitate identification of those actions that do not exhibit preference and should be the subject of enforcement. As an initial matter, STB is responsible for investigating substandard train performance under PRIIA section 213. Further, FRA believes the metrics in this final rule provide sufficient information to assist in such an STB investigation.

A commenter also proposed that FRA research the development of an “assignable tax credit” for passenger and highway competitive intermodal freight routes to generate funding for rail infrastructure. FRA appreciates the comment; however, it is outside the scope of this rulemaking.

Lastly, several commenters expressed support for additional rail infrastructure funding. The metrics in this final rule may assist decision makers in identifying rail projects.

Section 273.3 Definitions

This final rule includes several new and revised definitions, which are described here.

This section defines the term “actual running time” to mean the actual elapsed travel time of a train's travel on a host railroad, between the departure time at the first reporting point for a host railroad segment and the arrival time at the reporting point at the end of the host railroad segment. This definition is new to the final rule and supports the host running time metric.

This section defines the term “adjusted operating expenses” to mean Amtrak's operating expenses adjusted to exclude certain Amtrak expenses that are not considered core to operating the business. The major exclusions are depreciation, capital project related expenditures not eligible for capitalization, non-cash portion of pension and post-retirement benefits, and Amtrak's Office of Inspector General expenses (which are separately appropriated). Adjusted operating expenses do not include any operating expenses for State-supported routes that are paid for separately by States. This definition is a revision of the definition proposed in the NPRM to clarify its intent in response to commenters.

This section defines the term “certified schedule” to mean a published train schedule that Amtrak and the host railroad jointly certify is aligned with the customer on-time performance metric and standard in § 273.5(a)(1) and (2). If a published train schedule is reported as a certified schedule under § 273.5(c)(1), then it cannot later be designated as an uncertified schedule. This definition is new to the final rule in support of certified schedule metric.

This section defines the term “disputed schedule” to mean a published train schedule for which a specific change is sought: (i) That is the only subject of a non-binding dispute resolution process led by a neutral third-party and involving Amtrak and one or more host railroads; (ii) that is the only subject of a non-binding dispute resolution process led by a neutral third-party that has been initiated by one or more host railroads and Amtrak has not consented to participate in the process within 30 calendar days; or (iii) that is the only subject of a non-binding dispute resolution process led by a neutral third-party that has been initiated by Amtrak and the host railroad has not consented to participate in the process within 30 calendar days. The written decision resulting from a non-binding dispute resolution process is admissible in Surface Transportation Board investigations under 49 U.S.C. 24308(f). If a published train schedule is reported as a disputed schedule under § 273.5(c)(1), then it remains a disputed schedule until reported as a certified schedule. This definition is new to the final rule and supports the certified schedule metric.

This section defines the term “host railroad” to mean a railroad that is directly accountable to Amtrak by agreement for Amtrak operations over a railroad line segment. Amtrak is a host railroad of Amtrak trains and other trains operating over an Amtrak owned or controlled railroad line segment. For purposes of the certified schedule metric under § 273.5(c), Amtrak is not a host railroad. This definition is new to the final rule and supports several new and revised metrics.

This section defines the term “ridership data” to mean, in a machine-readable format: The total number of passengers, by train and by day; the station-specific number of detraining passengers, reported by host railroad whose railroad right-of-way serves the station, by train, and by day; and the station-specific number of on-time passengers reported by host railroad whose railroad right-of-way serves the station, by train, and by day. This definition is new to the final rule and supports the ridership data metric.

This section defines the term “scheduled running time” to mean the scheduled duration of a train's travel on a host railroad, as set forth in the Amtrak schedule skeleton. This definition is new to the final rule and supports the host running time metric.

This section defines the term “schedule skeleton” to mean a schedule grid used by Amtrak and host railroads to communicate the public schedule of an Amtrak train and the schedule of operations of an Amtrak train on host railroads. This definition is new to the final rule and supports the host running time metric.

This section defines the term “uncertified schedule” to mean a published train schedule that has not been reported as a certified schedule or a disputed schedule under § 273.5(c)(1). This definition is new to the final rule and supports the certified schedule metric.

Section 273.5 On-Time Performance and Train Delays

Paragraph (a)(1) of this section provides that the customer on-time performance metric is the percentage of all customers on an intercity passenger rail train who arrive at their detraining point no later than 15 minutes after their published scheduled arrival time, reported by train and by route.

Paragraph (a)(2) of this section provides a minimum standard for customer on-time performance of 80 percent for any 2 consecutive calendar quarters. This standard is consistent with the statutory requirement in 49 U.S.C. 24308(f)(1).

Paragraph (a)(3)(i) of this section provides that, except as provided in paragraph (a)(3)(ii), the customer on-time performance standard shall apply to a train beginning on the first full calendar quarter after May 17, 2021.

Paragraph (a)(3)(ii) of this section provides that, if a train schedule is a disputed schedule on or before May 17, 2021, then the customer on-time performance standard for the disputed schedule shall apply beginning on the second full calendar quarter after May 17, 2021.

Paragraph (b) of this section provides that the ridership data metric is the number of host railroads to whom Amtrak has provided ridership data consistent with this paragraph (b), reported by host railroad and by month. Not later than December 16, 2020, Amtrak must provide host railroad-specific ridership data to each host railroad for the preceding 24 months. On the 15th day of every month following Decmeber 16, 2020, Amtrak must provide host railroad-specific ridership data to each host railroad for the preceding month.

Paragraph (c)(1) of this section provides that the certified schedule metric is the number of certified schedules, uncertified schedules, and disputed schedules, reported by train, by route, and by host railroad (excluding switching and terminal railroads), identified in a notice to the Federal Railroad Administrator by Amtrak monthly, for the first six months following publication of the final rule, and then annually on the anniversary of the final rule's publication on November 16, 2020.

Paragraph (c)(2) of this section provides that, if a train schedule is reported as an uncertified schedule under paragraph (c)(1)(vi), (vii), or (viii), then Amtrak and the host railroad must transmit a joint letter and status report on the first of each month following the report, signed by their respective chief executive officers to each U.S. Senator and U.S. Representative whose district is served by the train, the Chairman and Ranking Member of the Committee on Transportation and Infrastructure of the House of Representatives, the Chairman and Ranking Member of the Committee on Commerce, Science, and Transportation of the Senate, the Chairman and Ranking Member of the Committee on Appropriations of the House of Representatives, the Chairman and Ranking Member of the Committee on Appropriations of the Senate, the Secretary of Transportation, and the Chairman of the Surface Transportation Board, which states: (i) The Amtrak train schedule(s) at issue; (ii) the specific components of the train schedule(s) on which Amtrak and host railroad cannot reach agreement; (iii) Amtrak's position regarding the disagreed upon components of the train schedule(s); (iv) host railroad's position regarding the disagreed upon components of the train schedule(s); and (v) Amtrak and the host railroad's plan and expectation date to resolve the disagreement(s). The requirement to transmit this joint letter and status report ends for the train schedule at issue when the uncertified schedule becomes a certified schedule.

Paragraph (c)(3) of this section provides that, when conditions have changed that impact a certified schedule, Amtrak or a host railroad may seek to modify the certified schedule. The customer on-time performance standard in subsection (a)(2) remains in effect during the schedule negotiation process.

Paragraph (d) of this section provides that the train delays metric is the minutes of delay for all Amtrak-responsible delays, host-responsible delays, and third party delays, for the host railroad territory within each route. The train delays metric is reported by delay code by: Total minutes of delay; Amtrak-responsible delays; Amtrak's host-responsible delays; Amtrak's host responsible delays and Amtrak-responsible delays, combined; non-Amtrak host-responsible delays; and third party delays. The train delays metric is also reported by the number of non-Amtrak host-responsible delay minutes disputed by host railroad and not resolved by Amtrak.

Paragraph (e) of this section provides that the train delays per 10,000 train miles metric is the minutes of delay per 10,000 train miles for all Amtrak-responsible and host-responsible delays, for the host railroad territory within each route. Paragraph (f) of this section provides that the station performance metric is the number of detraining passengers, the number of late passengers, and the average minutes late that late customers arrive at their detraining stations, reported by route, by train, and by station. The average minutes late per late customer calculation excludes on-time customers that arrive within 15 minutes of their scheduled time. A customer who arrives at their detraining station 16 minutes late would be included in this calculation and would be recorded as 16 minutes late.

Paragraph (g) of this section provides that the host running time metric is the average actual running time and the median actual running time compared with the scheduled running time between the first and final reporting points for a host railroad set forth in the Amtrak schedule skeleton, reported by route, by train, and by host railroad (excluding switching and terminal railroads).

Section 273.7 Customer Service

Paragraph (a) of this section provides that the customer satisfaction metric is the percent of respondents to Amtrak's customer satisfaction survey who provided a score of 70 percent or greater for their “overall satisfaction” on a 100 point scale for their most recent trip, by route, shown both adjusted for performance and unadjusted. Amtrak's customer satisfaction survey is a market-research survey that measures more than fifty specific service attributes that cover the entire customer journey. It should be noted that Amtrak can change the customer satisfaction survey, and such changes could in turn impact the information reported for the customer service metrics. However, in the event Amtrak changes the survey, the new survey would continue to seek information in connection with the customer satisfaction metrics required in this final rule (a survey change would just modify how the survey solicits this information). FRA will publish information about Amtrak's survey (including the survey questions and methodology) annually as an appendix to the quarterly report.

Several commenters provided feedback on Amtrak's customer satisfaction survey, including stating that the survey: Does not address accessibility concerns for disabled or elderly passengers (e.g., at the boarding station, on board the train, and at the destination station); and does not address ticket-purchase methods (e.g., phone, in-person agent, or website). First, as discussed above, Amtrak may change the customer satisfaction survey in the future. FRA understands that Amtrak is evaluating these suggestions and is committed to working with stakeholders to address these comments in future survey updates and/or by regularly providing related information on accessibility for disabled and elderly passengers that it collects already. A commenter also stated that Amtrak should offer additional contact methods for passengers to complete the customer satisfaction survey, such as postal mail and telephone. However, most Americans have access to the internet and there would be a substantial additional cost to providing surveys by postal mail or telephone with a corresponding limited benefit to the statistical sample of respondents.^[35]

A commenter stated that the survey should directly ask whether the customer was satisfied with the train's on-time performance. The Amtrak CSI Survey, which FRA included in docket number FRA-2019-0069-0004 for reference, does have a question asking respondents to rate their satisfaction with the reliability or on-time performance of the train on which they traveled. A commenter stated that the survey should include questions about customer/passenger interactions with Amtrak customer relations to evaluate this customer-facing service. FRA understands that Amtrak is evaluating this suggestion.

A commenter stated that a net promoter score or a median survey response should be used instead of the customer satisfaction survey. As noted, Amtrak may change the customer satisfaction survey. With that said, FRA considered several approaches to measuring customer service, including the net promoter score, but determined that the customer satisfaction survey offers an accurate assessment of the customer experience. Specifically, the customer satisfaction metric measures the percentage of respondents who provided a score of 70 percent or greater for their overall satisfaction. The use of 70 percent as the threshold is based on Amtrak's analysis of the relationship between customer satisfaction and the likelihood of future travel. As reported by Amtrak, the historical data suggests that customers who rate their overall satisfaction as 70 percent or greater are likely to travel on Amtrak again. In addition, Amtrak reports it adheres to industry best practices and solicits feedback from a random selection of riders, with a sample size of survey responses far greater than industry minimum standards. Lastly, FRA further understands that Amtrak distributes email surveys from a centralized database to ensure that employees are unable to provide surveys to targeted customers.

Amtrak adjusts overall satisfaction score performance by removing passengers who arrive at their destinations on State-supported and long-distance routes excessively late (30 minutes late for State-supported routes and 120 minutes for long-distance routes) from the system-wide calculation. Typically, on these routes, many of the major causes of passenger lateness are beyond Amtrak's control. By removing these customer responses from the calculations, most of the impact from these significantly late customers (whose responses may be overly influenced by the train's late arrival) is removed. Both the performance adjusted and non-performance adjusted overall satisfaction scores will be reported under this final rule to reflect the responses of all Amtrak customers.

A commenter stated that there should be a performance adjusted customer service metric and a separate non-performance adjusted customer service metric. FRA revised the final rule to clearly state that the customer satisfaction metric will be shown both adjusted for performance and not adjusted for performance. A commenter stated that the customer satisfaction metric should also be adjusted to show customer satisfaction surveys in which the excessive delays are Amtrak-related. FRA does not believe this would provide useful information. The intent of the customer satisfaction metric is to understand the experience of customers and measure “overall satisfaction,” not to determine the impacts of delay responsibility. Information on minutes of delay by category, responsible party, route and host territory, including Amtrak-responsible delays, are reported by other metrics in this final rule.

A commenter stated that the definition of excessively late should be changed to match the definition of late used in the customer OTP metric. However, aligning these two definitions would render the customer service metric less meaningful by significantly decreasing the number of survey responses included in the performance adjusted customer service score (on some routes, more than 70 percent of current customers would be excluded). FRA determined reporting both performance adjusted and non-performance adjusted customer service scores best provides a full and accurate view of customer satisfaction while also accounting for the impact of poor performance on customers' scores.

Several commenters stated that there should be additional customer service metrics with quantitative measurements not based on a survey score regarding: Mishandled bags; denied boardings; consumer complaints; riders needing assistance; riders using mobility-enhancing devices; and riders who paid for their tickets in cash. As a counterpoint, one commenter noted that including customer service metrics with quantitative measurements may require significant time and cost to build specific monitoring systems. FRA agrees that the cost to implement these metrics is unduly burdensome in cases where Amtrak does not already collect the data. In addition, FRA did not include a mishandled bags metric in the final rule because, unlike air and bus travel, Amtrak reported that the majority of intercity rail passengers handle their own bags. FRA believes the additional cost to collect this information is not warranted as Amtrak does not already collect the data on a routine basis. FRA did not include a denied boardings metric because the final rule's missed connections metric offers a broader measurement of customers who do not travel on their originally ticketed itinerary. FRA did not include a consumer complaints metric in the final rule because the customer satisfaction survey offers a more comprehensive quantitative measurement of customer satisfaction for the overall trip, as well as specific attributes of the experience, as compared to the number of complaints received. FRA did not include metrics about riders needing assistance, riders using mobility-enhancing devices, and riders who paid for their tickets in cash because, while these metrics may provide information about the customers Amtrak serves, these metrics do not measure the quality of service provided.

Finally, a commenter stated that all customer service metrics should be reported on a quarterly basis. FRA agrees and the final rule establishes quarterly reporting of all customer service metrics.

Paragraph (b) of this section provides that the Amtrak personnel metric is the average score from respondents to the Amtrak customer satisfaction survey for their overall review of Amtrak personnel on their most recent trip, by route.

Paragraph (c) of this section provides that the information given metric is the average score from respondents to the Amtrak customer satisfaction survey for their overall review of information provided by Amtrak on their most recent trip, by route.

Paragraph (d) of this section provides that the on-board comfort metric is the average score from respondents to the Amtrak customer satisfaction survey for their overall review of on-board comfort on their most recent trip, by route.

Paragraph (e) of this section provides that the on-board cleanliness metric is the average score from respondents to the Amtrak customer satisfaction survey for their overall review of on-board cleanliness on their most recent trip, by route.

Paragraph (f) of this section provides that the on-board food service metric is the average score from respondents to the Amtrak customer satisfaction survey for their review of on-board food service on their most recent trip, by route.

Section 273.9 Financial

Paragraph (a) of this section provides that the cost recovery metric is Amtrak's adjusted operating revenue divided by Amtrak's adjusted operating expense. This metric is reported at the corporate level/system-wide and for each route and is reported in constant dollars of the reporting year based on the Office of Management and Budget's gross domestic product chain deflator.

A commenter stated that the definition of the cost recovery metric presumes that Amtrak is responsible for all operating expenses over State-supported routes, which does not accurately represent the cost of service delivery routes where States cover the cost of some of the component services. FRA acknowledges that some States have separate arrangements to pay for operating expenses that are not reflected in Amtrak's adjusted operating expenses. Section 273.3 of the final rule includes a revised definition of the term “adjusted operating expenses” to clarify that the cost recovery metric does not include operating expenses for State-supported routes paid for separately by States.

Paragraph (b) of this section provides that the avoidable operating costs covered by passenger revenue metric is the percent of avoidable operating costs divided by passenger revenue for each route, shown with and without State operating payments. Each route's operating costs can be separated into three components: Frequency variable costs, route variable costs, and system/fixed costs. Avoidable operating costs are the sum of frequency and route variable costs. Frequency variable costs are costs that vary based on short-term decisions to adjust a route's schedule or frequency, not as a result of long-term decisions to add or eliminate a service permanently. Frequency variable costs typically occur directly and immediately with the service change. Frequency variable costs may include train and engine crew labor, on-board service labor, fuel and power, commissary provisions, specific yard operations, connecting motor coaches, and station staffing expenses.

Route variable costs are costs that vary based on long-term decisions to add or eliminate service and have a broader impact. Route variable costs typically require a separate management action to achieve a change in cost. Route variable costs may include car and locomotive maintenance turnaround, on-board passenger technology, commissary operations, direct advertising, specific reservations and call centers costs, station facility operations, station technology, maintenance of way, block and tower operations, regional/local police, and insurance expenses. These costs do not vary with individual train frequencies but may vary if service is increased or reduced on a larger scale. For example, costs for food and beverages stocked on a train would be avoidable if a single train were cancelled, but the commissary supporting the route would continue operations if other trains remained. Route variable costs attempt to capture the potential costs that would vary if the entire route were suspended or eliminated and the commissary supporting it no longer operated. Over time, or with a large enough expansion or reduction in service, the shared costs would be expected to change.

System/fixed costs are not likely to vary with smaller service changes and would not change if a single route were added or eliminated. System/fixed costs may include marketing and distribution, national police, environmental and safety, and general and administrative expenses.

Adding frequency variable and route variable costs to calculate avoidable operating costs does not make any distinction between short- and long-term avoidable costs, but results in a single avoidable cost figure for a single route at a future time. This approach represents a maximum saving, or cost avoided, and may be lower depending on the specific context of each individual route. The results of this approach are limited to the costs avoided if a single service is permanently eliminated. If multiple routes are eliminated, it is likely that some fixed costs will also decrease. Corporate-wide costs such as general and administrative expenses may shrink to reflect the size of the smaller business. In the event an actual elimination in service is contemplated, a detailed planning analysis would be required, considering the location of the route and the facilities that serve it, to determine the cost impacts.

The metric reflects avoidable operating costs as a percentage of passenger revenue, which, when shown at the route level, provides information about cost recovery, or the ability of the route to cover avoidable operating costs with revenue generated. States or other sponsoring entities also provide operating payments to Amtrak to provide service for trains on State-supported routes, which is classified as passenger revenue. To understand better the impact of these State payments, the metric avoidable operating costs covered by passenger revenue is calculated in two ways: First, as a percent dividing avoidable operating costs by passenger revenue, and second, as a percent dividing avoidable operating costs by passenger revenue without State operating payments.

One commenter stated general support for segregating State operating payments from passenger revenue for this metric (and for the fully allocated core operating costs covered by the passenger revenue metric). Another commenter stated that the avoidable operating costs and the fully allocated core operating costs covered by the passenger revenue metric should be reported by the specific sub-categories listed in the definition of passenger revenue. FRA disagrees. The final rule establishes metrics that report passenger revenue as a percent of avoidable costs and, separately, as a percent of fully allocated costs per route. Consistent with section 207, these metrics do not show the actual amount of revenue generated, but rather set forth a ratio of revenue to cost. In addition, the purpose of representing passenger revenue with and without State operating payments is to understand better the impact of State payments on route financial performance.

A commenter stated that the proposed avoidable cost metric is deficient and that the final rule should instead include a short-term avoidable cost metric, a long-term avoidable cost metric, and a long-term average infrastructure cost metric. FRA believes the avoidable cost metric is appropriate. Section 207 requires a metric that measures “the percentage of avoidable and fully allocated operating costs covered by passenger revenues on each route . . . .” The statute does not specify the time horizon of the metric or differentiate between short-term and long-term avoidable costs. The commenter also asserted that the proposed definition of avoidable costs includes some costs that may not be fully avoidable for a single route because they are shared among multiple routes. Although some costs are shared, FRA believes that these costs are avoidable, as over time they will scale to the size of the service provided. The commenter also proposed definitions of long-term avoidable costs and long-term average infrastructure costs that equate them with above-the-rail costs and below-the-rail costs, respectively. However, these proposed definitions do not align with the way Amtrak is organized as a business or the way that it allocates costs across its service lines and routes. In addition, the commenter proposed that the long-term avoidable cost definition include off-book equipment interest and depreciation expenses, but as equipment is shared across Amtrak's network, these costs likely are not avoidable because equipment may be used on other routes.

Paragraph (c) of this section provides that the fully allocated core operating costs covered by the passenger revenue metric is the percent of fully allocated core operating costs divided by passenger revenue for each route, shown with and without State operating payments. Fully allocated core operating costs include the fully-loaded share of overhead-type costs that pertain to more than one route or to the company as a whole. Costs are limited to “core” expenses (i.e., related to the provision of intercity passenger trains) to match expenses with passenger revenue. Several commenters stated general support for this metric, especially when reported alongside the avoidable operating costs covered by the passenger revenue metric.

Paragraph (d) of this section provides that the average ridership metric is the number of passenger-miles divided by train-miles for each route. This metric measures the average number of passengers on each of the route's trains. One commenter proposed that FRA also report an additional ridership metric to reflect total passengers by route alongside the passenger-miles per train-miles metric for convenience in comparing ridership data in FRA's quarterly report. FRA agrees, and the final rule includes such an additional metric in paragraph (e).

Paragraph (e) of this section provides that the total ridership metric is the total number of passengers on Amtrak trains, reported by route.

The definitions of terms in section 273.9 are only intended to apply to this final rule and the Amtrak financial reporting herein.

Section 273.11 Public Benefits

Paragraph (a) of this section provides that the connectivity metric is the percent of passengers connecting to and from other Amtrak routes, updated on an annual basis. The metric reports passengers making connections between the Northeast Corridor, State-supported, and long distances routes, or any combination thereof. Under this metric, a connection means a passenger arriving on one train and connecting to a departing train within 23 hours. Section 207 of PRIIA specifies that the metrics shall include “measures of connectivity with other routes in all regions currently receiving Amtrak service” for long distance routes. The connectivity metric provides connectivity information for the entire Amtrak network, including by route for long distance routes. One commenter expressed support for the connectivity metric, stating that it would give States more granular data with which to adjust schedules and build more regional-scale service.

Paragraph (b) of this section provides that the missed connections metric is the percent of passengers connecting to/from other Amtrak routes who missed connections due to a late arrival from another Amtrak train, reported by route and updated on an annual basis. A missed connection, particularly in a location with only one daily train, can result in a significant impact to the customer. A commenter stated that FRA should revise the missed connections metric to include the financial impact of missed connections and to report the results more frequently than once per year. FRA does not have the economic data to quantify the total financial impact of missed connections, and acquiring such data and methodologies would be challenging and burdensome, as FRA does not believe these data are readily available.

Paragraph (c) of this section provides that the community access metric is the percent of Amtrak passenger-trips to and from not well-served communities, updated on an annual basis. While one commenter expressed general support for this metric, another commenter stated that the community access metric does not adequately measure transportation needs because it does not identify communities that do not have access to intercity passenger rail or airports, nor does it address the convenience of train arrival times at rural stations. However, section 207(a) requires “measures of . . . the transportation needs of communities and populations that are not well-served by other forms of intercity transportation.” The final rule's definition of not well-served communities identifies rural communities that are not well-served by other intercity transportation modes (air and bus), but that do have regularly scheduled intercity passenger rail service, using distance from airports or station stops as a proxy for access. FRA recognizes the importance of understanding how to improve intercity passenger rail service to these communities, and views the current metric as an initial step in identifying the communities and analyzing their current use of Amtrak service. In addition, Amtrak is required to consider the transportation needs of not well-served communities in their route and service planning decisions. Fixing America's Surface Transportation Act, Public Law 114-94, 11206 (2015); 49 U.S.C. 24101, note.

Paragraph (d) of this section provides that the service availability metric is the total number of daily Amtrak trains per 100,000 residents in a metropolitan statistical area (MSA) for each of the top 100 MSAs in the United States, shown in total and adjusted for time of day, updated on an annual basis. Many MSAs are served regularly by Amtrak trains, but during inconvenient travel times. The metric, as adjusted for time of day, shows only those trains that arrive or depart between 5:00 a.m. and 11:00 p.m.

A commenter stated that there should be two economic and station development metrics to measure the annual total economic value to communities served by the intercity passenger rail service, accounting for factors such as labor, value-added benefits, and increased tax revenue, and to report that value as a ratio to the investment made in a route. The commenter also stated that these metrics should be based on an economic model developed by the Rail Passengers Association for such a purpose. FRA declines to include these metrics in this final rule. The final rule addresses service quality metrics that measure the actual provision of rail service. Although important, economic and station development metrics are indirectly related to intercity passenger rail service. In addition, measures of economic and development activity often require detailed information on local market conditions, and as such, are not well-suited for national metrics and may rely too heavily on general assumptions. Finally, these metrics would impose a significant burden on FRA to identify the appropriate data, obtain and track the detailed economic data, as well as to develop modeling capabilities.

A commenter stated that there should be an overlapping corridors metric to measure the number and economic value of passenger trips dependent upon intermediate connections on long-distance corridors. The commenter stated that the data for this metric could be gathered using the commenter's proposed economic and station development metric, with underlying community economic data updated annually, as well as the connections data from the final rule's missed connections metric. FRA declines to include this metric in the final rule. The missed connections metric is the percent of passengers connecting to/from other Amtrak routes who missed connections due to a late arrival from another Amtrak train, reported by route and updated on an annual basis. The reported data from the missed connections metric would not comprehensively identify intermediate connections on long-distance corridors. FRA selected metrics to measure the public benefit of intercity rail across all services and routes for the entire nation; this commenter's proposed metric would focus exclusively on long-distance routes. In addition, and as noted above, the proposed economic and station development metric would impose a significant burden on FRA to identify the appropriate data, obtain and track the detailed economic data, as well as to develop modeling capabilities.

A commenter stated that there should be a normalized route performance metric, reported quarterly, which would measure route performance for all routes on a per-passenger-mile basis and on a passengers-per-departure from each originating station basis. FRA declines to include this metric in the final rule and believes presenting the route-level information without any normalization is the most straight-forward method. The final rule does include a route-level ridership metric (the number of passenger miles divided by train-miles), which is consistent with section 207. Parties seeking additional information about Amtrak's operating statistics may also view Amtrak's monthly performance report, which includes seat miles and passenger miles by route.

Several commenters expressed general support for metrics that would measure the public benefit of passenger rail service. One commenter stated that the public benefits metrics listed in paragraphs (a) through (d) should be reported by route and updated quarterly, on a rolling previous 12-month basis. FRA recognizes the value of providing data more frequently to measure performance and to identify trends; however, the metrics listed in paragraphs (a) through (d) require significant effort to compile and calculate, and as such, the final rule provides that these metrics will be updated annually.

VI. Regulatory Impact and Notices

A. Executive Order (E.O.) 12866, E.O. 13771, and DOT Regulatory Policies and Procedures

This final rule is a significant regulatory action within the meaning of Executive Order 12866 and DOT regulatory policies and procedures.^[36] Although the economic effects of this regulatory action would not exceed the $100 million annual threshold defined by Executive Order 12866, the rule is significant because of the substantial public interest in this rulemaking. Pursuant to the Congressional Review Act (5 U.S.C. 801 et seq.), the Office of Information and Regulatory Affairs designated this rule as not a 'major rule', as defined by 5 U.S.C. 804(2). Additionally, this final rule is considered an E.O. 13771 regulatory action. FRA has provided an assessment of the costs and cost savings expected to result from implementation of this final rule.

The Metrics and Standards measure the performance and service quality of intercity passenger train operations as required by section 207 of PRIIA. The Metrics and Standards are generally organized into four categories: On-time performance and train delays, customer service, financial, and public benefits.

Other than the OTP and train delays metrics, the Metrics and Standards in this final rule will not pose an additional burden on Amtrak or host railroads. Data such as customer satisfaction and financial information are currently collected by Amtrak and submitted to FRA on a quarterly basis. Other data, such as train delays and on-time performance, are already shared between Amtrak, host railroads, and State partners under their various agreements, and the parties have established protocols for data collection, distribution, and reconciliation. While the final rule establishes a new data-sharing requirement to assist with calculating the customer OTP metric (specifically, ridership data), this information is already collected by Amtrak. FRA expects that Amtrak will develop additional procedures for sharing the data, but once established, this data sharing will not burden Amtrak's routine operations. Lastly, as a result of the final rule's customer OTP metric and certified schedule metric, Amtrak and host railroads may adjust Amtrak's published train schedules to align them with the customer OTP metric. As part of that effort, Amtrak and host railroads may meet to discuss and agree upon schedule modifications to the published train schedules.

FRA received several comments addressing the NPRM's cost estimates. A commenter stated that the NPRM did not consider the impacts on commerce and a host railroad's operations and network fluidity. A commenter stated that a customer OTP metric enlarges an Amtrak train's dispatch footprint (i.e., it would cause the Amtrak train to take up additional capacity on the rail line) by redistributing recovery time across intermediate stations, which threatens overall network fluidity, among other things. A commenter also stated that FRA did not consider payments made under the Amtrak-host railroad operating agreement (stating that the host railroad would receive less performance payments under the existing operating agreement).

With respect to operational impacts, as discussed above, delays waiting for time at intermediate stations can be foreclosed by an accurate schedule, and adjusting train schedules to align with the customer OTP metric does not mean that recovery time will be added for each station. In the case of capacity impacts great enough to warrant schedule change, reductions of time to remove these waits would be in both parties' interests. In addition, with respect to impacts on commerce specifically, Congress has accounted for such impacts by providing that STB's enforcement of the preference requirement not “materially lessen the quality of freight transportation provided to shippers.” 49 U.S.C. 24308(c).

With respect to operating agreement payments, as noted previously, FRA is not a party to these agreements, nor does FRA have knowledge of their details. More importantly, this final rule does not require a change to the performance payment provisions in these operating agreements; Amtrak and the host railroads may continue to maintain those provisions as they see fit. In addition, to the extent a host railroad is concerned with receiving lower performance payments as a result of this final rule, this final rule likewise does not prohibit a host railroad and Amtrak from revising the performance payments to align better with the customer OTP metric and standard. In fact, section 207(c) provides that, to the extent practicable, Amtrak and its host rail carriers shall incorporate the metrics and standards into their operating agreements. Also, performance payments, even if they change as a result of the final rule, would not change the estimate of costs due to the rule. Such payments represent transfers rather than economic costs or benefits.

One Class I host railroad stated that the NPRM's costs are too low and their railroad alone would require more than 10 hours of meetings to discuss schedule revisions. Another commenter stated that the NPRM substantially underestimates the cost of attempting to negotiate schedule adjustments. Based on both comments, FRA has increased the estimate of meeting time and number of employees present at those meetings. Additionally, FRA has substantially increased the estimated time spent on preparations for those meetings.

For purposes of this analysis, FRA assumed that Amtrak and each of the host railroads will meet five times during the first year to discuss revising Amtrak's published train schedules. Amtrak currently has agreements with 31 host railroads. However, eight of these railroads are switching and terminal railroads that will not likely be involved in revising schedules, as Amtrak only operates over those railroads for short distances with very few, if any, stops. If there were discussions between Amtrak and any switching and terminal railroads, then it would be expected to occur during regularly scheduled meetings and would not add any additional burden.

For the other 23 host railroads, schedule discussions will add time to the current regular meetings held with Amtrak. FRA estimates that such schedule alignment discussions will require 40 hours of additional meeting time between Amtrak and each host railroad. FRA estimates that Amtrak and the host railroad will each have approximately three to six employees at the meetings. The following table shows the total cost of additional meetings between Amtrak and host railroads. Wage rates for this analysis are from the Surface Transportation Board.^[37] Over the course of the first year, the total cost of all additional meetings is estimated to be $473,473.

Further, to prepare for these meetings, Amtrak and the 23 host railroads will need to perform the necessary groundwork, such as historical data analysis of schedules and train performance, as well as analysis of current and future operations, to determine how train schedules should be adjusted.

The cost for host railroads preparing for meetings will vary depending on the complexity of the route. FRA estimates that Class I host railroads will have more extensive discussions than non-Class I host railroads, based largely on the greater amount of route miles hosted. The following table shows the estimated costs of preparing for meetings. Amtrak and host railroads will spend $296,991 over the first year to prepare for meetings.

In addition, this final rule requires Amtrak and a host railroad to transmit a monthly joint letter and status report, signed by their respective chief executive officers, to certain members of Congress and other Federal Agencies, in the event a published train schedule is not certified or disputed by May 17, 2021. Preparing a letter will require staff time by Amtrak and a host railroad, as well as briefings with the chief executive officers. Each letter is estimated to require $656 in labor on Amtrak's part and $1,022 on the host railroad's part. FRA estimates that five routes will be uncertified in the first year; each of which will require six letters. The following table shows the cost of the monthly letters. The total estimated cost to Amtrak and host railroads for the monthly letters will be $50,328.

Due to this final rule, some railroads will likely initiate a non-binding dispute resolution process to resolve scheduling disputes. Based on an analysis by FRA subject matter experts, FRA estimates that approximately eight routes will be the subject of such a non-binding dispute resolution process. The total cost of such a non-binding dispute resolution process per route is approximately $52,200, and includes arbitration fees and compensation for the arbitrators. The arbitration fees include administrative fees,^[38] arbitrator travel fees, and the rental fee for the hearing room. The table below shows the estimated costs for arbitration fees.

The compensation paid to the arbitrator includes time spent by each arbitrator to prepare for the hearing, attend the hearing, and review the hearing after completion. The table below shows the costs for arbitrator compensation.

The cost paid to the arbitrator for their fees would likely be split between Amtrak and the host railroad. The total estimated cost paid for the non-binding dispute resolution process for all eight routes will be $417,600, which includes arbitrator fees and compensation.

In addition to the cost of the non-binding dispute resolution process, Amtrak and a host railroad will need to spend time: Preparing documents in connection with the non-binding dispute resolution process; briefing within their organization; and attending the hearing. The table below shows the total cost of staff time for Amtrak and host railroads.

FRA assumes that employees from the host railroads and Amtrak will incur some travel costs associated with the hearing. The table below shows the expected cost of travel related to the hearing.

The table below shows all estimated arbitration costs, including: Arbitration fees, arbitrator compensation, and Amtrak and the host railroad's staff compensation and travel costs. The total cost of arbitration will be $714,030.

This final rule also requires Amtrak to share ridership data with each host railroad. Although systems are already in place for sharing of data, it will require additional time from an Amtrak employee to process the data and share it in a usable format. The following table shows the estimated cost to prepare the ridership reports.

All costs of this final rule are expected to be incurred during the first year, though FRA acknowledges that conditions regarding a certified schedule may change. The following table shows the total 10-year estimated costs of this final rule.

This final rule may result in lower operational costs for Amtrak, to the extent it results in improved OTP, which may reduce labor costs, fuel costs, and expenses related to passenger inconvenience, and provide benefits to riders from improved travel times and service quality. A commenter stated that improved OTP should have a significant effect on ridership, and would make a significant improvement on operational costs. Due to the difficulty in precisely quantifying future benefits to rail routes from improved OTP, combined with the inability to quantify the potential synergistic effects that improved OTP reliability could have across Amtrak's network, FRA has not quantified any potential benefits from lower operational costs or improved service that may result from the final rule. FRA expects Amtrak and host railroads to structure schedules to achieve performance that meets this rule's OTP standard, thus avoiding the expense and uncertainty of an STB investigation under section 213.

B. Regulatory Flexibility Act and Executive Order 13272

The Regulatory Flexibility Act of 1980 (RFA) (5 U.S.C. 601 et seq.) and Executive Order 13272 (67 FR 53461, Aug. 16, 2002) require agency review of proposed and final rules to assess their impacts on small entities. When an agency issues a rulemaking proposal, the RFA requires the agency to “prepare and make available for public comment an initial regulatory flexibility analysis” which will “describe the impact of the proposed rule on small entities.” (5 U.S.C. 603(a)).

Section 605 of the RFA allows an agency to certify a rule, in lieu of preparing an analysis, if the proposed rulemaking is not expected to have a significant economic impact on a substantial number of small entities. Out of an abundance of caution, FRA prepared an initial regulatory flexibility analysis to accompany the NPRM, which noted no expected significant economic impact on a substantial number of small entities. FRA is now certifying that this final rule will not have a significant economic impact on a substantial number of small entities.

Description of Small Entities Impacted by the Final Rule

In consultation with the SBA, FRA has published a final statement of agency policy that formally establishes “small entities” or “small businesses” as railroads, contractors, and hazardous materials shippers that meet the revenue requirements of a Class III railroad as set forth in 49 CFR 1201.1-1, which is $20 million or less in inflation-adjusted annual revenues, and commuter railroads or small governmental jurisdictions that serve populations of 50,000 or less. See 68 FR 24891 (May 9, 2003) (codified at appendix C to 49 CFR part 209). FRA is using this definition for the final rule.

This final rule impacts Amtrak and Amtrak's host railroads. This rule establishes a customer OTP metric and a certified schedule metric, which will likely result in modifications to some of Amtrak's published train schedules. Amtrak is not a small entity and the majority of the host railroads are Class I railroads or State Departments of Transportation, none of which are small entities. There are currently 12 host railroads that are small entities, including approximately 8 switching and terminal railroads and 4 short line or regional railroads.^[39] There are approximately 695 class III railroads on the general system. Therefore, the 12 small entities potentially affected by this final rule are not considered a substantial number of small entities.

Economic Impact on Small Entities

FRA has determined that the economic impact on small entities will not be significant. This final rule does not require published train schedule modifications. However, FRA assumes that, as a result of the Metrics and Standards, Amtrak will engage with many host railroads to discuss modifications to the published train schedule to align the schedules with the customer OTP metric.

There are currently twelve host railroads that are small entities, including approximately eight switching and terminal railroads and four short line and regional railroads. The impact on those small entities are very minimal. The switching and terminal railroads are not likely burdened by this final rule because Amtrak only operates over those routes for short distances and has very few stops along those sections of track. Those railroads already meet with Amtrak on a periodic basis, so any discussions regarding their schedule will take place at that time. It is likely that no schedule adjustments are required along those routes.

Amtrak has limited stops along the routes of the four short line and regional railroads; therefore, published train schedule adjustments would be brief. Those railroads also already meet with Amtrak on a periodic basis and discussions regarding schedules can take place at that time. Such discussions may add a minimal amount of time to those meetings. However, published train schedule adjustments may not even be necessary for these railroads.

Other than the customer OTP metric, the final rule does not provide an additional burden on Amtrak or the host railroads. Amtrak already collects the data to support these new metrics; therefore, there is no additional burden.

Certification

Consistent with the findings in FRA's initial regulatory flexibility analysis, the FRA Administrator hereby certifies that this final rule will not have a significant economic impact on a substantial number of small entities.

C. Paperwork Reduction Act

FRA is publishing a new information collection request in connection with this final rule in a separate notice. For information or a copy of the paperwork package submitted to OMB, contact Ms. Kim Toone, at 202-493-6132, or [e-mail].

D. Federalism Implications

Executive Order 13132, “Federalism” (64 FR 43255, Aug. 10, 1999), requires FRA to develop an accountable process to ensure “meaningful and timely input by State and local officials in the development of regulatory policies that have federalism implications.” “Policies that have federalism implications” are defined in the Executive Order to include regulations that have “substantial direct effects on the States, on the relationship between the national government and the States, or on the distribution of power and responsibilities among the various levels of government.” Under Executive Order 13132, the agency may not issue a regulation with federalism implications that imposes substantial direct compliance costs and that is not required by statute, unless the Federal Government provides the funds necessary to pay the direct compliance costs incurred by State and local governments, or the agency consults with State and local government officials early in the process of developing the regulation. Where a regulation has federalism implications and preempts State law, the agency seeks to consult with State and local officials in the process of developing the regulation.

FRA has analyzed this final rule under the principles and criteria contained in Executive Order 13132. This final rule could affect State and local governments to the extent that they sponsor, or exercise oversight of, intercity passenger rail service. Because this final rule is required by Federal statute, the consultation and funding requirements of Executive Order 13132 do not apply.

In sum, FRA has analyzed this final rule under the principles and criteria in Executive Order 13132. As explained above, FRA has determined this final rule has no federalism implications. Therefore, preparation of a federalism summary impact statement for this final rule is not required.

E. Environmental Impact

FRA has evaluated this final rule consistent with the National Environmental Policy Act (NEPA; 42 U.S.C. 4321 et seq.), other environmental statutes, related regulatory requirements, and its NEPA implementing regulations at 23 CFR part 771. Under NEPA, categorical exclusions (CEs) are actions identified in an agency's NEPA implementing regulations that do not normally have a significant impact on the environment and therefore do not require either an environmental assessment (EA) or environmental impact statement (EIS). See 40 CFR 1508.4. FRA has determined that this final rule is categorically excluded from detailed environmental review pursuant to 23 CFR 771.116(c)(15), “Promulgation of rules, the issuance of policy statements, the waiver or modification of existing regulatory requirements, or discretionary approvals that do not result in significantly increased emissions of air or water pollutants or noise.”

In analyzing the applicability of a CE, FRA must also consider whether unusual circumstances are present that would warrant a more detailed environmental review through the preparation of an EA or EIS. See 23 CFR 771.116(b). FRA has concluded that no unusual circumstances exist with respect to this regulation that would trigger the need for a more detailed environmental review. The purpose of this rulemaking is to establish metrics and standards to measure the performance and service quality of intercity passenger train operations. FRA does not anticipate any environmental impacts from this final rule and finds there are no unusual circumstances present in connection with this final rule.

A commenter stated that FRA should consider whether the rulemaking meets the requirements of a categorical exclusion under NEPA given the operational impacts on the host railroads. As discussed elsewhere in this final rule, any such operational impacts relate to, and should be resolved by, the development of new schedules. FRA expects Amtrak and the host railroads to account for these issues when they develop new schedules. Therefore, FRA finds that a categorical exclusion is appropriate here.

Pursuant to Section 106 of the National Historic Preservation Act and its implementing regulations, FRA has determined this undertaking has no potential to affect historic properties. See 16 U.S.C. 470. FRA has also determined that this rulemaking does not approve a project resulting in a use of a resource protected by Section 4(f). See Department of Transportation Act of 1966, as amended (Pub. L. 89-670, 80 Stat. 931); 49 U.S.C. 303.

F. Executive Order 12898 (Environmental Justice)

Executive Order 12898, Federal Actions to Address Environmental Justice in Minority Populations and Low-Income Populations, and DOT Order 5610.2(a) (91 FR 27534 May 10, 2012) require DOT agencies to achieve environmental justice as part of their mission by identifying and addressing, as appropriate, disproportionately high and adverse human health or environmental effects, including interrelated social and economic effects, of their programs, policies, and activities on minority populations and low-income populations. The DOT Order instructs DOT agencies to address compliance with Executive Order 12898 and requirements within the DOT Order in rulemaking activities, as appropriate. FRA has evaluated this final rule under Executive Order 12898 and the DOT Order and has determined it would not cause disproportionately high and adverse human health and environmental effects on minority populations or low-income populations.

G. Executive Order 13175 (Tribal Consultation)

FRA has evaluated this final rule under the principles and criteria in Executive Order 13175, Consultation and Coordination with Indian Tribal Governments, dated November 6, 2000. The final rule will not have a substantial direct effect on one or more Indian tribes, will not impose substantial direct compliance costs on Indian tribal governments, and will not preempt tribal laws. Therefore, the funding and consultation requirements of Executive Order 13175 do not apply, and a tribal summary impact statement is not required.

H. Unfunded Mandates Reform Act of 1995

Under Section 201 of the Unfunded Mandates Reform Act of 1995 (Pub. L. 104-4, 2 U.S.C. 1531), each Federal agency “shall, unless otherwise prohibited by law, assess the effects of Federal regulatory actions on State, local, and tribal governments, and the private sector (other than to the extent that such regulations incorporate requirements specifically set forth in law).” Section 202 of the Unfunded Mandates Reform Act (2 U.S.C. 1532) further requires that before promulgating any general notice of proposed rulemaking that is likely to result in the promulgation of any rule that includes any Federal mandate that may result in expenditure by State, local, and tribal governments, in the aggregate, or by the private sector, of $100,000,000 or more (adjusted annually for inflation) in any 1 year, and before promulgating any final rule for which a general notice of proposed rulemaking was published, the agency shall prepare a written statement detailing the effect on State, local, and tribal governments and the private sector. This final rule will not result in the expenditure, in the aggregate, of $100,000,000 or more (as adjusted annually for inflation) in any one year, and thus preparation of such a statement is not required.

I. Energy Impact

Executive Order 13211 requires Federal agencies to prepare a Statement of Energy Effects for any “significant energy action.” 66 FR 28355 (May 22, 2001). Under the Executive Order, a “significant energy action” is defined as any action by an agency (normally published in the Federal Register) that promulgates or is expected to lead to the promulgation of a final rule or regulation, including notices of inquiry, advance notices of proposed rulemaking, and notices of proposed rulemaking: (1)(i) That is a significant regulatory action under Executive Order 12866 or any successor order, and (ii) is likely to have a significant adverse effect on the supply, distribution, or use of energy; or (2) that is designated by the Administrator of the Office of Information and Regulatory Affairs as a significant energy action. FRA has evaluated this final rule in accordance with Executive Order 13211. FRA has determined that this rule is not likely to have a significant adverse effect on the supply, distribution, or use of energy. Consequently, FRA has determined that this final rule is not a “significant energy action” within the meaning of Executive Order 13211.

Executive Order 13783, “Promoting Energy Independence and Economic Growth,” requires Federal agencies to review regulations to determine whether they potentially burden the development or use of domestically produced energy resources, with attention to oil, natural gas, coal, and nuclear energy resources. 82 FR 16093 (March 31, 2017). Executive Order 13783 defines “burden” to mean unnecessarily obstruct, delay, curtail, or otherwise impose significant costs on the siting, permitting, production, utilization, transmission, or delivery of energy resources. FRA has determined this final rule will not potentially burden the development or use of domestically produced energy resources.

J. Trade Impact

The Trade Agreements Act of 1979 (Pub. L. 96-39, 19 U.S.C. 2501 et seq.) prohibits Federal agencies from engaging in any standards setting or related activities that create unnecessary obstacles to the foreign commerce of the United States. Legitimate domestic objectives, such as safety, are not considered unnecessary obstacles. The statute also requires consideration of international standards and, where appropriate, that they be the basis for U.S. standards. FRA has assessed the potential effect of this final rule on foreign commerce and believes that its requirements are consistent with the Trade Agreements Act of 1979.

List of Subjects in 49 CFR Part 273

• Railroads
• Transportation

The Rule

For the reasons discussed in the preamble, FRA amends chapter II, subtitle B of title 49, Code of Federal Regulations, by adding part 273 to read as follows:

PART 273—METRICS AND MINIMUM STANDARDS FOR INTERCITY PASSENGER TRAIN OPERATIONS

273.1

Purpose.

273.3

Definitions.

273.5

On-time performance and train delays.

273.7

Customer service.

273.9

Financial.

273.11

Public benefits.

Authority: Sec. 207, Div. B, Pub. L. 110-432; 49 U.S.C. 24101, note; 49 U.S.C. 103(j); 49 CFR 1.81; 49 CFR 1.88; and 49 CFR 1.89.

§ 273.1

Purpose.

The purpose of this part is to establish metrics and minimum standards for measuring the performance and service quality of intercity passenger train operations.

§ 273.3

Definitions.

As used in this part—

Actual running time means the actual elapsed travel time of a train's travel on a host railroad, between the departure time at the first reporting point for a host railroad segment and the arrival time at the reporting point at the end of the host railroad segment.

Adjusted operating expenses means Amtrak's operating expenses adjusted to exclude certain Amtrak expenses that are not considered core to operating the business. The major exclusions are depreciation, capital project related expenditures not eligible for capitalization, non-cash portion of pension and post-retirement benefits, and Amtrak's Office of Inspector General expenses. Adjusted operating expenses do not include any operating expenses for State-supported routes that are paid for separately by States.

Adjusted operating revenue means Amtrak's operating revenue adjusted to exclude certain revenue that is associated with capital projects. The major exclusions are the amortization of State capital payments and capital project revenue related to expenses not eligible for capitalization.

Amtrak means the National Railroad Passenger Corporation.

Amtrak's customer satisfaction survey means a market-research survey that measures Amtrak's satisfaction score as measured by specific service attributes that cover the entire customer journey.

Amtrak-responsible delays means delays recorded by Amtrak, in accordance with Amtrak procedures, as Amtrak-responsible delays, including passenger-related delays at stations, Amtrak equipment failures, holding for connections, injuries, initial terminal delays, servicing delays, crew and system delays, and other miscellaneous Amtrak-responsible delays.

Avoidable operating costs means costs incurred by Amtrak to operate train service along a route that would no longer be incurred if the route were no longer operated.

Certified schedule means a published train schedule that Amtrak and the host railroad jointly certify is aligned with the customer on-time performance metric and standard in § 273.5(a)(1) and (2). If a published train schedule is reported as a certified schedule under § 273.5(c)(1), then it cannot later be designated as an uncertified schedule.

Disputed schedule means:

(1) A published train schedule for which a specific change is sought:

(i) That is the only subject of a non-binding dispute resolution process led by a neutral third-party and involving Amtrak and one or more host railroads;

(ii) That is the only subject of a non-binding dispute resolution process led by a neutral third-party that has been initiated by one or more host railroads and Amtrak has not consented to participate in the process within 30 calendar days; or

(iii) That is the only subject of a non-binding dispute resolution process led by a neutral third-party that has been initiated by Amtrak and the host railroad has not consented to participate in the process within 30 calendar days.

(2) The written decision resulting from a non-binding dispute resolution process is admissible in Surface Transportation Board investigations under 49 U.S.C. 24308(f). If a published train schedule is reported as a disputed schedule under § 273.5(c)(1), then it remains a disputed schedule until reported as a certified schedule.

Fully allocated core operating costs means Amtrak's total costs associated with operating an Amtrak route, including direct operating expenses, a portion of shared expenses, and a portion of corporate overhead expenses. Fully allocated core operating costs exclude ancillary and other expenses that are not directly reimbursed by passenger revenue to match revenues with expenses.

Host railroad means a railroad that is directly accountable to Amtrak by agreement for Amtrak operations over a railroad line segment. Amtrak is a host railroad of Amtrak trains and other trains operating over an Amtrak owned or controlled railroad line segment. For purposes of the certified schedule metric under § 273.5(c), Amtrak is not a host railroad.

Host-responsible delays means delays recorded by Amtrak, in accordance with Amtrak procedures, as host-responsible delays, including freight train interference, slow orders, signals, routing, maintenance of way, commuter train interference, passenger train interference, catenary or wayside power system failure, and detours.

Not well-served communities means those rural communities: Within 25 miles of an intercity passenger rail station; more than 75 miles from a large airport; and more than 25 miles from any other airport with scheduled commercial service or an intercity bus stop.

Passenger revenue means intercity passenger rail revenue generated from passenger train operations, including ticket revenue, food and beverage sales, operating payments collected from States or other sponsoring entities, special trains, and private car operations.

Ridership data means, in a machine-readable format: The total number of passengers, by train and by day; the station-specific number of detraining passengers, reported by host railroad whose railroad right-of-way serves the station, by train, and by day; and the station-specific number of on-time passengers reported by host railroad whose railroad right-of-way serves the station, by train, and by day.

Scheduled running time means the scheduled duration of a train's travel on a host railroad, as set forth in the Amtrak schedule skeleton.

Schedule skeleton means a schedule grid used by Amtrak and host railroads to communicate the public schedule of an Amtrak train and the schedule of operations of an Amtrak train on host railroads.

Third party delays means delays recorded by Amtrak, in accordance with Amtrak procedures, as third party delays, including bridge strikes, debris strikes, customs, drawbridge openings, police-related delays, trespassers, vehicle strikes, utility company delays, weather-related delays (including heat or cold orders, storms, floods/washouts, earthquake-related delays, slippery rail due to leaves, flash-flood warnings, wayside defect detector actuations caused by ice, and high-wind restrictions), acts of God, or waiting for scheduled departure time.

Uncertified schedule means a published train schedule that has not been reported as a certified schedule or a disputed schedule under § 273.5(c)(1).

§ 273.5

On-time performance and train delays.

(a) Customer on-time performance—(1) Metric. The customer on-time performance metric is the percentage of all customers on an intercity passenger rail train who arrive at their detraining point no later than 15 minutes after their published scheduled arrival time, reported by train and by route.

(2) Standard. The customer on-time performance minimum standard is 80 percent for any 2 consecutive calendar quarters.

(3) Application. (i) Except as provided in paragraph (a)(3)(ii) of this section, the customer on-time performance standard shall apply to a train beginning on the first full calendar quarter after May 17, 2021.

(ii) If a train schedule is a disputed schedule on or before May 17, 2021, then the customer on-time performance standard for the disputed schedule shall apply beginning on the second full calendar quarter after May 17, 2021.

(b) Ridership data. The ridership data metric is the number of host railroads to whom Amtrak has provided ridership data consistent with this paragraph (b), reported by host railroad and by month. Not later than December 16, 2020, Amtrak must provide host railroad-specific ridership data to each host railroad for the preceding 24 months. On the 15th day of every month following December 16, 2020, Amtrak must provide host railroad-specific ridership data to each host railroad for the preceding month.

(c) Certified schedule— (1) Metric. The certified schedule metric is the number of certified schedules, uncertified schedules, and disputed schedules, reported by train, by route, and by host railroad (excluding switching and terminal railroads), identified in a notice to the Federal Railroad Administrator by Amtrak:

(i) On December 16, 2020;

(ii) On January 19, 2021;

(iii) On February 16, 2021;

(iv) On March 16, 2021;

(v) On April 16, 2021;

(vi) On May 17, 2021;

(vii) On November 16, 2021; and

(viii) Every 12 months after November 16, 2021.

(2) Reporting. If a train schedule is reported as a an uncertified schedule under paragraph (c)(1)(vi), (vii), or (viii) of this section, then Amtrak and the host railroad must transmit a joint letter and status report on the first of each month following the report, signed by their respective chief executive officers to each U.S. Senator and U.S. Representative whose district is served by the train, the Chairman and Ranking Member of the Committee on Transportation and Infrastructure of the House of Representatives, the Chairman and Ranking Member of the Committee on Commerce, Science, and Transportation of the Senate, the Chairman and Ranking Member of the Committee on Appropriations of the House of Representatives, the Chairman and Ranking Member of the Committee on Appropriations of the Senate, the Secretary of Transportation, and the Chairman of the Surface Transportation Board, which states:

(i) The Amtrak train schedule(s) at issue;

(ii) The specific components of the train schedule(s) on which Amtrak and host railroad cannot reach agreement;

(iii) Amtrak's position regarding the disagreed upon components of the train schedule(s);

(iv) Host railroad's position regarding the disagreed upon components of the train schedule(s); and

(v) Amtrak and the host railroad's plan and expectation date to resolve the disagreement(s). The requirement to transmit this joint letter and status report ends for the train schedule at issue when the uncertified schedule becomes a certified schedule.

(3) Ongoing coordination between Amtrak and host railroads. When conditions have changed that impact a certified schedule, Amtrak or a host railroad may seek to modify the certified schedule. The customer on-time performance standard in paragraph (a)(2) of this section remains in effect for the existing certified schedule, until a modified schedule is jointly certified.

(d) Train delays. The train delays metric is the minutes of delay for all Amtrak-responsible delays, host-responsible delays, and third party delays, for the host railroad territory within each route. The train delays metric is reported by delay code by: total minutes of delay; Amtrak-responsible delays; Amtrak's host-responsible delays; Amtrak's host responsible delays and Amtrak-responsible delays, combined; non-Amtrak host-responsible delays; and third party delays. The train delays metric is also reported by the number of non-Amtrak host-responsible delay minutes disputed by host railroad and not resolved by Amtrak.

(e) Train delays per 10,000 train miles. The train delays per 10,000 train miles metric is the minutes of delay per 10,000 train miles for all Amtrak-responsible and host-responsible delays, for the host railroad territory within each route.

(f) Station performance. The station performance metric is the number of detraining passengers, the number of late passengers, and the average minutes late that late customers arrive at their detraining stations, reported by route, by train, and by station. The average minutes late per late customer calculation excludes on-time customers that arrive no later than 15 minutes after their scheduled time.

(g) Host running time. The host running time metric is the average actual running time and the median actual running time compared with the scheduled running time between the first and final reporting points for a host railroad set forth in the Amtrak schedule skeleton, reported by route, by train, and by host railroad (excluding switching and terminal railroads).

§ 273.7

Customer service.

(a) Customer satisfaction. The customer satisfaction metric is the percent of respondents to the Amtrak customer satisfaction survey who provided a score of 70 percent or greater for their “overall satisfaction” on a 100 point scale for their most recent trip, by route, shown both adjusted for performance and unadjusted.

(b) Amtrak personnel. The Amtrak personnel metric is the average score from respondents to the Amtrak customer satisfaction survey for their overall review of Amtrak personnel on their most recent trip, by route.

(c) Information given. The information given metric is the average score from respondents to the Amtrak customer satisfaction survey for their overall review of information provided by Amtrak on their most recent trip, by route.

(d) On-board comfort. The on-board comfort metric is the average score from respondents to the Amtrak customer satisfaction survey for their overall review of on-board comfort on their most recent trip, by route.

(e) On-board cleanliness. The on-board cleanliness metric is the average score from respondents to the Amtrak customer satisfaction survey for their overall review of on-board cleanliness on their most recent trip, by route.

(f) On-board food service. The on-board food service metric is the average score from respondents to the Amtrak customer satisfaction survey for their overall review of on-board food service on their most recent trip, by route.

§ 273.9

Financial.

(a) Cost recovery. The cost recovery metric is Amtrak's adjusted operating revenue divided by Amtrak's adjusted operating expense. This metric is reported at the corporate level/system-wide and for each route and is reported in constant dollars of the reporting year based on the Office of Management and Budget's gross domestic product chain deflator.

(b) Avoidable operating costs covered by passenger revenue. The avoidable operating costs covered by passenger revenue metric is the percent of avoidable operating costs divided by passenger revenue for each route, shown with and without State operating payments.

(c) Fully allocated core operating costs covered by passenger revenue. The fully allocated core operating costs covered by passenger revenue metric is the percent of fully allocated core operating costs divided by passenger revenue for each route, shown with and without State operating payments.

(d) Average ridership. The average ridership metric is the number of passenger-miles divided by train-mile for each route.

(e) Total ridership. The total ridership metric is the total number of passengers on Amtrak trains, reported by route.

§ 273.11

Public benefits.

(a) Connectivity. The connectivity metric is the percent of passengers connecting to and from other Amtrak routes, updated on an annual basis.

(b) Missed connections. The missed connections metric is the percent of passengers connecting to/from other Amtrak routes who missed connections due to a late arrival from another Amtrak train, reported by route and updated on an annual basis.

(c) Community access. The community access metric is the percent of Amtrak passenger-trips to and from not well-served communities, updated on an annual basis.

(d) Service availability. The service availability metric is the total number of daily Amtrak trains per 100,000 residents in a metropolitan statistical area (MSA) for each of the top 100 MSAs in the United States, shown in total and adjusted for time of day, updated on an annual basis.

Issued in Washington, DC.

Gerald A. Reynolds,

Chief Counsel.

Footnotes

1.  One commenter stated that FRA should have also consulted with heavy tonnage seaports with terminal and switching railroads. FRA notes that, while such specific consultation was not required by the statute, FRA had many in-depth meetings with Class I railroads who are well-versed in the issues related to providing rail service to seaports; indeed Class I railroad comments mirrored those from this commenter.

2.  FRA sought input from certain rail labor groups that did not express interest in consulting at the time.

3.  This definition reflects a minor revision to the NPRM's definition of customer OTP, which clarifies that early trains are counted as on-time. FRA made this revision in response to a comment seeking this clarification.

4.  There are several uncommon situations that can affect the calculation of customer OTP. Customers on canceled trains (less than 4 hours advance notice) are counted as late customer arrivals at their ticketed station if service to their ticketed station is canceled. Customers that are carried beyond their ticketed off-point are included in the customer arrival count at their ticketed off-points. Re-accommodated customers not due to the suspension of a train are excluded from the calculation for their original trip but would be counted for customer OTP for the rescheduled trip. Customers on bus bridges (transportation on buses for a portion of a regularly scheduled train route) are excluded from the calculation.

5.  For example, one commenter stated that OTP on multi-host routes should be measured against the run time for each host railroad line segment (and not against the scheduled departure and arrival time at each station).

6.  One commenter also stated that the customer OTP metric would harm the morale of the host railroad's employees who take pride in achieving good OTP. FRA appreciates the commitment of all employees, at Amtrak and the host railroads, and understand they work hard in support of Amtrak trains.

7.  FRA's quarterly reports do not exist solely to serve as a trigger for an STB investigation. These reports also provide information for policymakers and the public, consistent with the data reporting for other modes of transportation, such as air travel. See https://www.transportation.gov/​individuals/​aviation-consumer-protection/​air-travel-consumer-reports.

8.  See 49 U.S.C. 24710(a) (Requiring Amtrak to use the section 207 performance metrics to evaluate annually the operating performance of each long-distance train); 49 U.S.C. 24710(b) (Requiring Amtrak to develop a performance improvement plan for its long-distance routes based on the data collected from the section 207 performance metrics, to include OTP); 49 U.S.C. 24308(f)(1) (Referring to the on-time performance of an “intercity passenger train”); see also Union Pac. R.R. Co. v. Surface Transp. Bd., 863 F.3d 816, 826 (8th Cir. 2017).

9.  FRA's quarterly reports showing Amtrak's performance under the OTP metric are relied upon to determine whether a train is below the standard. See Union Pac. R.R. Co. v. Surface Transp. Bd., 863 F.3d 816, 826 (8th Cir. 2017). Congress also assigned STB with the responsibility to determine whether and to what extent delays . . . are due to causes that could reasonably be addressed” by the host railroad or by Amtrak. 49 U.S.C. 24308(f)(1).

10.  Another commenter suggested a key stations OTP metric combined with changes to the Amtrak-host railroad operating agreement to preserve a similar contractual performance payment regime. As stated elsewhere in this final rule, this final rule does not prohibit Amtrak and a host railroad from revising their operating agreement.

11.  See Application of the National Railroad Passenger Corporation Under 49 U.S.C. 24308(a)—Canadian National Railway Company, STB Docket No. FD 35743 at 11, FN 25 (Aug. 9, 2019) (“An OTP metric that measures the percentage of passengers that arrive at their destination stations on time could—in some circumstances—allow for greater host railroad operational flexibility and create an incentive structure more closely tied to the service delivery to the end consumer, the passenger.”).

12.  Recovery time means time added to a schedule to help a train “recover” to published schedule on-time operation in the event that it encounters delays.

13.  One commenter stated that under a customer OTP metric it is not reasonable to believe a host railroad would agree to a schedule that did not achieve OTP at all stations. Although Amtrak and a host railroad may agree on a schedule that reliably achieves OTP at all stations, the customer OTP metric provides greater flexibility to the parties by allowing them to focus on those stations with greater numbers of detraining passengers.

14.  As STB stated, “[i]t is not reasonable for an incentives and penalties system to have at its foundation a performance metric that fails to account for the OTP at stations central to the passenger experience for a significant portion of Amtrak passengers.” Application of the National Railroad Passenger Corporation Under 49 U.S.C. 24308(a)—Canadian National Railway Company, STB Docket No. FD 35743 at 10 (Aug. 9, 2019).

15.  The percentage of detraining passengers to each station on a route can be calculated from the information Amtrak is currently providing to host railroads for their internal use. See FRA-2019-0069-0295. This data provides quarterly detraining totals by station by train.

16.  Station rank in absolute terms may also be a helpful tool for schedule planning in connection with the customer OTP metric.

17.  While Amtrak does not make this ridership data publicly available, Amtrak shared this data with relevant host railroads. See FRA-2019-0069-0295. Amtrak also consented to this minimal public disclosure of ridership data to provide this illustrative example.

18.  In Amtrak's system, a cancellation with less than four hours advance notice represents an unplanned en route event. Amtrak established the four-hour benchmark to recognize that a cancellation with less than four hours advance notice would not give the customer sufficient time to make alternative travel arrangements. The four-hour benchmark is the same used for several other measures of Amtrak performance. The cancellation need not include the entire train or trip such as in an emergency detour situation, where selected stations may be bypassed (and passengers bussed to their original detraining location) but the train continues to its final destination. Passengers who are required to take a bus bridge to their final destination as a result of an unplanned cancellation are counted as late. Amtrak makes every effort to get these passengers to their desired destination, typically by bus or by re-accommodation on another train. Implementing these alternative travel plans due to an en route event nearly always results in passengers arriving late to their final destination. They are therefore counted as late to their detraining station and are included as such in customer OTP calculations.

19.  An OTP metric, in part, can inform the formulation of a train schedule. For example, a customer OTP metric may encourage a schedule with more recovery time at those stations with more de-boarding passengers, while an endpoint OTP metric may encourage a schedule with more recovery time at the endpoints of a line segment.

20.  A certified schedule metric is consistent with section 207's direction to measure on-time performance, as the schedule is a benchmark of train performance.

21.  Although the certified schedule metric is reported by host railroad (excluding switching and terminal railroads), FRA encourages all the host railroads for a route to work together in aligning the train schedule.

22.  The final rule defines the term disputed schedule to mean a published train schedule for which a specific change is sought: (1) That is the only subject of a non-binding dispute resolution process led by a neutral third-party and involving Amtrak and one or more host railroads; (2) that is the only subject of a non-binding dispute resolution process led by a neutral third-party that has been initiated by one or more host railroads and Amtrak has not consented to participate in the process within 30 calendar days; or (3) that is the only subject of a non-binding dispute resolution process led by a neutral third-party that has been initiated by Amtrak and the host railroad has not consented to participate in the process within 30 calendar days. The written decision resulting from a non-binding dispute resolution process is admissible in Surface Transportation Board investigations under 49 U.S.C. 24308(f). If a published train schedule is reported as a disputed schedule under subsection (c)(1), then it remains a disputed schedule until designated as a certified schedule.

23.  If a train schedule is reported as an uncertified schedule at six months, twelve months, or yearly thereafter, then Amtrak and the host railroad must transmit a joint letter and status update, signed by their respective chief executive officers, to each U.S. Senator and U.S. Representative whose district is served by the train, in addition to several other government offices. This joint letter and status update must identify the Amtrak published train schedule(s) at issue and the plan and expectation date to resolve the disagreement(s), among other details.

24.  In addition, FRA will post such joint letters on its website.

25.  FRA recognizes the importance of reviewing schedules periodically to ensure their integrity. However, the customer OTP standard would continue to apply during a schedule review period. In addition, the customer OTP standard will apply to any new Amtrak train service initiated after application of the customer OTP standard (and that train will be subject to the certified schedule metric).

26.  The final rule only affords delay of the customer OTP standard beyond six months for disputed schedules. After the six-month period, the customer OTP standard applies to both certified schedules and uncertified schedules. There may be a scenario where one host railroad for a train has a disputed schedule (to which the customer OTP standard is not yet applied) and another host railroad for that train has either a certified schedule or an uncertified schedule. As the customer OTP metric is reported by train (and by route), in this situation, FRA will not include customer OTP metric data in the quarterly report for that train during the time when there is a disputed schedule (to which the customer OTP standard is not yet applied) for some portion of the train's route. FRA encourages Amtrak and all of the host railroads of a train to work together when evaluating the published train schedules.

27.  The final rule does not dictate a specific process beyond that it is a non-binding dispute resolution process led by a neutral third-party. For example, the final rule does not address how the parties pay the fees and costs associated with such a process (although an equal share of such costs would be one reasonable approach), nor does the final rule address the number of arbitrators (although the associated costs for an arbitration in the final rule's section regarding economic impacts are based on a panel of three arbitrators).

28.  In a related comment, a commenter stated that Congress only intended for a limited number of Amtrak trains to be subject to an STB investigation. FRA is not aware of any language in section 207, or PRIIA, to support this interpretation.

29.  To the customer, there may be no discernable difference as to whether they are on one host railroad's territory or another's while traveling on a route. However, most intercity passenger rail routes involve one or more host railroads. This final rule establishes metrics that measure route-level performance reflecting the customer experience and that measure aspects of performance of the individual host railroads within the route segments that they control.

30.  In response to a comment seeking clarification, the train delays metric measures the minutes of delay for each individual host railroad territory within a route.

31.  If the host railroad does not have an NRPC officer, then another officer with the appropriate expertise and authority at the host railroad would fulfill this responsibility.

32.  See Application of the National Railroad Passenger Corporation Under 49 U.S.C. 24308(a)—Canadian National Railway Company, STB Docket No. FD 35743 at 23-24 (Aug. 9, 2019) (Describing the delay cause identification process under an existing operating agreement).

33.  The final rule defines schedule skeleton to mean a schedule grid used by Amtrak and host railroads to communicate the public schedule of an Amtrak train and the schedule of operations of an Amtrak train on host railroads. Schedule skeletons indicate, for each train, the: (a) Time of arrival at the point of entry to the rail lines of a host railroad, and time of departure from the point of exit from the rail lines of a host railroad; (b) dwell time at each station and servicing location on the rail lines of a host railroad; and (c) pure running time, recovery time, and miscellaneous time within a segment.

34.  FRA understands an out-of-slot train to be a train that arrives after the time the host railroad anticipated and planned for the train in its operating plan.

35.  In 2016, the U.S. Census reported that eighty-one percent of American households had a broadband internet subscription. See https://www.census.gov/​content/​dam/​Census/​library/​publications/​2018/​acs/​ACS-39.pdf.

36.  See 5 CFR part 5.

37.  2019 STB wage rates: Group #100 (Executives, Officials, & Staff Assistants) Wage Rate: $68.81 or $120.42 with a 75% burden factor. Group #200 (Professional & Administrative) Wage Rate: $44.27 or $77.47 with a 75% burden factor. Group #500 (Transportation (Other than Train & Engine)) Wage Rate: $40.27 or $70.47 with a 75% burden factor.

38.  Source: American Arbitration Association. See “Undetermined Monetary Claims” Standard Fee Schedule at https://www.adr.org/​sites/​default/​files/​Commercial_​Arbitration_​Fee_​Schedule_​1.pdf

39.  FRA received one comment from a Class III terminal railroad operating on track controlled by another railroad, expressing concern about being the subject of an STB investigation. However, it is FRA's understanding that Amtrak does not currently operate over the right-of-way in question (and although the possibility of future Amtrak service may exist, such future service would be subject to the certified schedule metric in this final rule).