Appendix A

Lessons Learned from International Best Practices

The following examples of regional rail systems across the globe provide important lessons that inform our thinking. While there is no precisely exact analogue to Metro Boston, these examples demonstrate the effectiveness and importance of electrification, high-level platforms and connecting stub-end terminals as approaches to providing riders with a rail service that serves their mobility needs.


Note: All dollar figures cited are in US dollars, adjusted for exchange rates and inflation.


Paris: A Best Practices Model

With a combined 1.1 billion riders a year, Paris’s RER (“Réseau Express Régional” or Regional Express Network) and Transilien regional rail networks carry more passengers than nearly any other city in the world. The only busier systems are in Japan and South Korea, where there is little technological distinction between Commuter Rail and the subway, and in India, where the Mumbai Suburban Railway substitutes for a subway and is infamously dangerous and overcrowded. As the busiest proper regional rail system running alongside a separate metro
system, the RER is a useful case study for Boston.

Paris historically had six intercity rail terminals and three smaller terminals used only by commuter lines. It had some frequent service starting in the 1920s, but no real regional integration. In 1937-8, Metro operator RATP bought one isolated commuter line, the Ligne de Sceaux, and electrified it and began running frequent service, treating it as a regional Metro line. The diagrams to the right show the evolution of the Paris RER network.

Paris’ model has been copied worldwide with Commuter Rail transformations in Toronto (RER) and Brussels (RER). In the last two decades, London has also invested in two RER-like projects, both of which run through central London: Thameslink connecting the Brighton Main Line and the Midland Main Line with an unused short rail link, and Crossrail which connects the Great Western Main Line to the Great Eastern Main Line and the North London Line with 13 mi of deep bore tunnels.

Predating Paris’s transformation, Philadelphia connected its two legacy Commuter Rail networks, the Reading and Pennsylvania Railroads. Similar to our preferred version of the NSRL, the connecting tunnel is short and includes 4 tracks. Unfortunately, SEPTA did not complete relatively inexpensive complementary capital projects to support more efficient use. The only through-running American rail doesn’t come close to its full potential. They have also begun terminating many trains downtown. Thus, our plan for regional rail and the NSRL would make Boston a pioneer among American transit agencies.



Frequent commuter rail in Paris, 1970


In the postwar era, the lack of through-service and the distance of most rail terminals from the emerging central business district became serious enough problems that RATP and national rail operator SNCF began planning for multiple rail links.

The Ligne de Sceaux is in blue pointing southwest. The southeast pointing red line is the Ligne de Vincennes, bought by RATP in the 1960s and electrified in 1969, connecting to a new terminal in Paris,
Nation, a major Metro transfer station.

Frequent regional rail in Paris, 1985


RATP and SNCF cooperated in planning. RATP would route its new tunnels to serve intercity rail stations, and there would be through-service between RATP- and SNCF-owned lines.

The RER A, run by RATP, is in red; the east branch is new construction. The RER B is in blue; the Ligne de Sceaux is operated by RATP whereas the northern side, with a new branch going north to the airport, is operated by SNCF. The RER C is in yellow, operated by SNCF. Other lines, still stub-ending at the terminals, are called Transilien.

Frequent Regional Rail in Paris, today


The RER A has a long tunnel going west, serving La Defense, a suburban high-rise business district. Unlike Paris, Boston’s office density in the South Station area enables similar functionality with a shorter tunnel.

The green line is RER D, and the magenta RER E, both run by SNCF. The RER E is being extended to the east, taking over some additional Transilien lines and relieving the overburdened RER A. On the RER A, the new east branch now serves suburban jobs in Marne-la-Vallée and Disneyland Paris; a new west branch serves jobs in Cergy.

Munich: Connecting Two Stub-End Train Terminals

Like Boston, Munich historically had two main stub-end train terminals: Hauptbahnhof (Main Station) and Ostbahnhof (East Station). Passengers transferring between these stations had to use the city’s overcrowded streetcar network.

Since 1930 the German term for regional rail, including high frequency and through-running, has been S-Bahn, where S stands for Stadtschnell (city-fast). But whereas Berlin and Hamburg had such systems for decades, Munich would have to wait. There were recurrent plans for underground rail tunnels; a plan for an S-Bahn network from the Nazi era began construction in 1938, with an east-west and north-south tunnel, but the works were suspended in the war and little had been built. After the war, West Germany recovered quickly, and Munich grew especially fast, but infrastructure was inadequate. Plans for both a municipal subway and an S-Bahn run by national railway DB resumed, reducing the S-Bahn to just one east-west tunnel connecting Hauptbahnhof and Ostbahnhof.

Both the subway and the S-Bahn began construction in 1965. In the next year, Munich was chosen as the host city for the 1972 Summer Olympics, forcing both projects to accelerate. In the next six years, not only did the region build the east-west tunnel and the first subway line, but also it electrified the commuter lines that hadn’t yet been wired, and lengthened and raised the platforms to allow for long trains, comparable to 8.5-car American trains.

To reduce costs, the S-Bahn was not built to the same standards on the outer branches as in the city. Some branches are single-track, and some require S-Bahn trains to share track with intercity passenger and freight trains. Since the system’s initial opening there has been additional construction, but these design compromises still impose constraints on the system’s timetable. To resolve them, Munich invented the clockface schedule, in which timetables repeat regularly all day on a prescribed interval, such as every 20 minutes. Riders began crowding the system. Daily ridership had crept up from 114,000 in 1961 to 160,000 when the system opened, but then reached 250,000 by the end of 1972, 430,000 at the end of the 1970s, and 640,000 in 1987. Today there are 840,000 passengers per day.27 Some lines today run every 10 minutes, combining to a peak frequency of a train every 2 minutes in the central tunnel.

Munich rail network trains are very crowded, and as a result, there have been plans to build a second east-west tunnel to add capacity. However, with more underground infrastructure in Munich today than there was in the 1960s, construction has run into schedule and budget overruns. The second tunnel, under construction since last year, is now projected to open in 2026 and cost $1.1 billion per underground mile, the second highest figure outside the English-speaking world (the RER A, at $1.2 billion, is first). At this cost, the NSRL tunnel could be built for $6 billion.

Boston had 80 years of head start on Munich in building underground urban rail. It has an opportunity to use NSRL to catch up. Munich’s innovations in scheduling show the MBTA how to run a fast, reliable schedule even with grade crossings and single-track segments on some branches. Conversely, Munich’s one failure, the mounting costs of the second S-Bahn tunnel, presents a strong argument for building a four-track NSRL from the start. It might look cheaper to start with two tracks and then expand to four later, but the lifetime costs would be much higher.

Philadelphia: A Tunnel Needs Trains

Like Boston and Munich, Philadelphia historically had two separate stub-end train terminals: Suburban Station and Reading Terminal. Suburban Station was built by the Pennsylvania Railroad; intercity trains stopped just outside Center City, at 30th Street Station, which offered through-service from New York to Washington. Reading Terminal served the trains of the Reading Railroad as well as other railroads that competed with the Pennsylvania, such as the Baltimore and Ohio. Each station was about a quarter mile from City Hall, where the city’s two main subway lines intersect, but connecting between the two stations required walking the half mile.

In the 1970s, with both the Pennsylvania and Reading bankrupt, SEPTA took over both of their commuter rail networks, and began to plan to unify them. Both networks were already electrified, having been wired between the 1910s and 1930s. SEPTA designed the Center City Commuter Connection, a 1.8-mile tunnel connecting the two networks, extending four tracks of the already-underground Suburban Station to the east, with a new Market East station replacing the above-ground Reading Terminal. In today’s money it cost $1.24 billion; if the North-South Rail Link could built at the same cost per mile, it would cost $2 billion.

The effect was not as transformational as hoped. The through-service patterns originally proposed were based on European best industry practices, with high frequency on every branch, going up to a train every ten minutes. Unfortunately SEPTA ran trains infrequently off-peak, and fares were set at a premium over bus and subway fare even within the city. This is especially harmful to in-city ridership: several SEPTA Regional Rail branches run entirely within Philadelphia’s city limits, and all have weak ridership, even ones running parallel to overcrowded bus lines. This pattern of stronger ridership in the suburbs than in the city should not surprise Boston-based readers, since the Fairmount Line has relatively low ridership, and had the lowest ridership per mile when it charged premium fares.

Thus Philadelphia completed the expensive part of regional rail - the tunnel and electrification - but has refrained from increasing service and integrating fares with city transit. Nonetheless, with through-running offering suburbanites access to several Center City stops, SEPTA Regional Rail has slightly higher ridership than the MBTA commuter rail network, on barely half the route-length. It also has the lowest operating costs of all major American commuter rail systems: $14 per car-mile and $310 per car-hour, compared with $18.5 and $540 respectively on the MBTA.28
In terms of size, historic urban form, and transit usage, Philadelphia and Boston are similar. That SEPTA’s partial modernization (through-running and electrification, but no high frequency or fare integration) led to partial success (somewhat lower operating costs and ridership per route-mile) should make Boston more confident that full modernization of MBTA service would lead to full success.

Auckland: Electrify First

With 1.7 million people, Auckland is the largest metro area in New Zealand, and the fastest-growing. Provoked by increasing traffic and the unreliability of aging equipment and infrastructure, it has spent the last decade modernizing its formerly decrepit regional rail system, converting from a mixture of locomotive-hauled trains and secondhand DMUs to modern EMUs, with the full electrified system opening in 2015.29 The region wired about 60 miles of track for NZ$80 million, about $60 million in 2018 US dollars; if the MBTA could electrify for the same per-mile cost, it would be able to wire the entire network for $400 million.

Along with electrification and modern rolling stock, Auckland introduced much more frequent service. Ridership has responded accordingly, rising from under 6 million per month in June 2011 to almost 20 million in June 2017 (the first electrified services began in April 2014), with no end in sight.30



At the same time that it modernized the regional rail system, Auckland also implemented a complete revision of the bus system. By using the added capacity to shift passengers onto the more efficient rail services, Auckland demonstrated a willingness to coordinate between modes that has often eluded Boston. Even before the opening of electric service, discussions about building a downtown tunnel to turn Auckland’s stub-end Britomart terminal into one stop on a through route had begun. Known as City Rail Link, that project is now underway.31 Auckland’s planners and advocates used the momentum generated by the activity around and success of rail modernization to build momentum for the tunnel, a much larger capital project. The tunnel project has begun to suffer from cost increases, and as of September 2016 its cost is estimated at $2.8-$3.4 billion NZ ($1.85-$2.25 billion US) for 2.2 miles of double-track tunnel. 32 Per track-mile, this is similar to the Harvard study’s estimate of $4-6 billion for NSRL, a four-track tunnel.33

Auckland’s experience offers three major takeaways for Boston: 1) Riders respond to provision of frequent, modern service even if there’s no precedent for it, 2) integration of modes—such as Auckland’s commitment to using buses and trains together—is crucial, and 3) The incremental strategy of building ridership through modernization and then immediately committing to the heavier infrastructure lift, the downtown tunnel, is viable; though it will not be ideal, or recommended, in every scenario. For those wishing to read more about Auckland’s experience, the City Rail Link website provides an excellent breakdown and history.

33: We recognize that NSRL faces unique potential cost pressures associated with the highway and transit tunnel infrastructure already in place along the route.