Connecticut rail policy: the next four years

CT DOT

The first Hartford line train leaving for New Haven.

Slowly but surely, Connecticut is rebuilding its rail system. The newly opened Hartford Line has been a success, to the point that some trains run crowded enough to keep people from boarding them. Metro-North mainline traffic is close to historical highs. The Connecticut Department of Transportation is considering buying new equipment for branch line service and Shoreline East (SLE) trains. There are plans for new stations and expanded service.

The state, of course, remains in its long-held tradition of being stuck in fiscal crisis, thus hampering any new investments. Whatever the new administration does to raise funds for transportation, spending is going to be necessarily limited for the foreseeable future, and money will have to focus on key priorities.

The good news, however, is that even with limited funding, Connecticut can do a lot to improve its rail infrastructure without spending too much money – and can adopt quite a few policies that will create some room for additional investments. To achieve this ConnDOT, Amtrak, and Metro-North will have to adopt some significant changes on how they run their railroads, build infrastructure, and purchase equipment.

Roger Senserrich

Below I offer a list of cost-conscious changes that the state could do to make rail service better — starting with some low- hanging fruit and moving up to more challenging institutional reforms. Most are based on best practices in other places (side note: the U.S. seems to have largely forgotten how to run railroads, so these places are often overseas); several involve actively collaborating between administrations. Both adopting new practices and improving institutional cooperation between agencies are often more challenging to implement, as they involve changing day-to-day work, but they are significantly cheaper than just throwing money at the problem.

Fix the Hartford Line’s schedule of rush hour trains

The Hartford Line was an expensive project, so we certainly should get the most out of it. A quick glance at the scheduling shows quite clearly that at least half of the towns in the line do not, as rail service is not there when it is most needed.

Let’s look at the schedule; For the morning rush hour, New Haven gets three inbound trains between 7  and 9  a.m. Hartford, meanwhile, gets two from the north (both before 8 a.m.), and just one from the south (also before 8 a.m.). The evening rush hour situation is less dire, with five outbound trains from New Haven between 4  and 7  p.m. compared to three northbound and four southbound for Hartford. However, if people could not get into the city by rail in the morning, they will likely not take the train to get back home.

Although there are constraints both regarding equipment availability and track capacity in Hartford itself, these frequencies should be improved. Adding more southbound trains from Springfield is possible (even more so now, with double tracking all the way up to Windsor), and CT Rail should have enough equipment on hand to run those services.

Fix the Hartford Line schedule off-peak

As a cost-saving measure, CT Rail (and SLE) runs on a much lower frequency outside the morning and afternoon peak and on weekends. The thinking goes, trains are mostly empty in the middle of the day so we might as well focus our efforts when people are going to or back from work.

This argument is flawed for two reasons. First, the expensive part of running a commuter railroad is not having a train go up and down a line, but fixed costs: track, signaling, and purchasing/leasing the rolling stock. Trains are expensive machines, so just having them sitting around is expensive. Railroads are capital intensive, and capital depreciation is one of their main costs on their balance sheet. This means that the cost of running a locomotive and a few cars is in practice almost as high as the cost of not running them at all; having a consist sitting idle for three hours at New Haven Union Station is almost as expensive as having it take a trip up and down to Hartford. Alon Levy has an excellent article here going over the math. CT Rail should seriously consider just adding more trains in the middle of the day, as the marginal operating costs are likely lower than the revenue they will generate.

Second, rail lines become exponentially more useful and attractive to riders when their service is frequent and predictable. The Hartford Line, as many rail lines in the U.S., is designed with (white) suburban commuters in mind. They take the train in the morning, go back home in the evening, and have neat and predictable lives. Although commuters might be the single biggest group of riders on the line, rail service truly shines when it is an option for travel that is always there, even for trips that do not follow a commuting pattern.

Ideally, schedules should not even be a factor and frequencies should be high enough that you can just show up at the station. For the Hartford Line, having trains spaced at regular intervals during the day will suffice for now. If you know that there is a northbound train in Meriden at 15 minutes past the hour every hour and southbound trains at 45 minutes past the hour, no exceptions (and with more frequent service during peak, hopefully filling in predictable intervals), users really appreciate it.

Keeping these regular service patterns might be difficult, of course, depending on bottlenecks and track capacity. Which leads me to the next point.

Fix the New Haven State Street mess

The state spent a good amount of money adding an additional platform to New Haven State Street Station, the stop closest to downtown that serves all three rail lines (SLE, Hartford, Metro-North). It added a new vestibule, heated platforms, and nicely overkill staircase and awning — the usual massive overbuilding we saw in the rest of stations in the line (seriously, ConnDOT – you did not need to build a Taj Mahal-like structure everywhere. It is a commuter railroad, not a casino). In theory, the additional platform is there to allow more flexibility when moving trains in and out of New Haven. In practice, it is seldom used.

The newly built State Street station entrance.

Last time I checked, only one train uses the new, longer platform on State Street during the p.m. peak; I haven’t looked at the a.m. trains in a while, but I believe no more than one uses it in the morning (and probably none). This means that all trains call on the older, two-track island platform on the south side of the line, leading to pretty silly bouts of congestion in a place where there is a whole platform sitting unused. Making things worse, southbound Hartford Line trains must cross over the main line to get to the platform, having to wait for Amtrak trains to get out of the way.

I am not sure why this happens, frankly; I’ve been looking at the schedules trying to find a good reason for dispatchers to do this, but haven’t found any. The only reason I can think of is that CT Rail wants to use the platforms on the south end of Union Station (close to Metro North) to enable cross-platform transfers, but that rarely happens now. More likely, Amtrak does not want to relinquish using the platforms on the north side, closest to the station building. As of now, it leads to delays, even with the limited amount of SLE trains running.

Fix Shoreline East already

In the past few months ridership on the SLE line has cratered. The reason is the constant bus substitutions that have plagued the line since early summer, as Amtrak does track work at several points along the line.

The thing is, Amtrak is taking absolutely forever for a project that any reasonable railroad would have completed in a couple of months, at night time, without completely sabotaging partner operations for half a year. Ballast renewal and replacement is hard work, but not a six-month project. Amtrak, however, doesn’t mind doing it slowly, as they own the tracks and give priority to their own trains.

This lack of collaboration is not new. Amtrak similarly trolled ConnDOT by slow-walking a fairly trivial project to electrify some sidings along the line to allow the use of electric equipment, and they were not especially accommodating in testing the M8 trains on the line. ConnDOT has been similarly passive-aggressive with Amtrak for years on the other side of the line (from New Haven to the New York border), imposing low-speed limits, often bizarre dispatching practices and generally treating their own Metro-North trains better than the Amtrak’s Acelas or Regionals. Working together to do things better doesn’t cost money, but both agencies have refused to do it so in a productive manner over and over again, even if could improve rail service.

Besides getting Amtrak to play nice, ConnDOT should start using electric trains for SLE services as soon as possible. Electric multiple units (EMUs) accelerate much better, are faster, and cheaper to operate than locomotive-hauled trains. Shifting SLE services to M8 electrics can potentially cut travel time from New London to New Haven by up to 20% (about 16 minutes), making the line much more attractive (calculations here). Using the M8s here would also let CT Rail to use the current SLE equipment in the Hartford Line, allowing for more frequent service.

Unfortunately, Metro-North hurriedly retired the last few M2s last year, instead of keeping a few around so SLE could get some of the new equipment. The railroad is scheduled to receive additional consists this year, however, so they should make electrifying SLE service a priority.

Revise staffing levels

This might raise some mild disagreement (understatement) with the unions, but it makes little to no sense to have a small army of conductors in every single SLE, Hartford Line, or Metro-North train. The vast majority of commuter rail systems in the world run with one-person crews (the engineer) and have random spot checks to validate that passengers have tickets. U.S. rail lines instead use a small horde of conductors that ride in every single train, checking all the tickets.

It is immensely wasteful. The additional revenue collected by raising fare compliance to 100 percent is vastly inferior to the cost of having so many people just collecting tickets. Conductors are not essential for passenger safety either (again – no other country seems to need them), so ConnDOT should seriously consider having less of them on board. That might (will) run against labor contracts, but adopting modern staffing levels is a necessarily step to allow for cost-effective operations.

Revise new equipment purchases

The Federal Railroad Administration (FRA) recently adopted a new set of rules regarding passenger rail equipment in the U.S. that should make every single railroad in the country immediately reconsider any and all of their rolling stock purchases. Up until now, U.S. rail equipment could only be charitably defined as tanks-on-wheels, as FRA crash-worthiness regulations asked for extremely heavy, unwieldy, inefficient railcars. The new rule changes that, allowing railroads to purchase light, modern European-style equipment.

This is a very big deal: it means that ConnDOT can buy off-the-shelf European-style diesel or electric multiple units (DMU and EMUs) instead of loco-hauled trains. Modern EMU/DMUs are much cheaper and perform much better than the current FRA-compliant equipment. They accelerate faster, use less fuel, and can accommodate more passengers in less space. The M8s (Metro North’s newest EMUs) are 50 percent heavier than comparable European EMUs and accelerate much worse. Any light European DMU will run circles around the heavy, unwieldy, ancient SLE and Hartford Line trains.

ConnDOT is currently planning a significant equipment purchase for diesel trains for the Waterbury and Danbury branches, as well as new trains to replace the leased (and frankly, pretty crummy) MBB coaches in the Hartford Line. Getting competent, modern DMUs will save the state a lot of money, and allow for much faster, more reliable service.

Fix construction costs

This is a bigger picture item, but it is equally important: ConnDOT rail infrastructure costs are way too high. I already mentioned the extravagant overbuilding on Hartford Line stations, but frankly, the infrastructure costs of many projects are out of hand.

Norwalk’s problematic railroad bridge.

The Walk Bridge replacement in Norwalk, for instance, is currently budgeted at over one billion. This is, frankly, nuts; a 562-foot rail bridge, even with a movable span, should not be that expensive. More complex, longer, taller bridges in other wealthy countries cost far less than that; Spain just opened a 787 feet, two-track rail bridge on a high-speed line for less than 150 million. The Hartford Line itself was extremely expensive on a per-mile basis, even more so considering that it was an upgrade, not a new line.

The new administration should have a good, hard look at why building costs for rail infrastructure are so high compared to other countries. This issue, incidentally, is not unique to Connecticut, but something that plagues U.S. infrastructure projects nationwide.

Pay way more attention to buses

Not exactly rail policy, but this is probably the most important point of the whole list. As much as (white) people like trains, fixing Connecticut’s often shambolic bus services is probably the best way to improve transportation to the largest amount of people in the state. Fixing CT Transit might not be glamorous, attract many headlines, or have a random blogger write 2,000+ words of passionate advocacy about rail policies and practices, but will get a lot more people to work.

Case in point: the Hartford Line has an average daily ridership of 1,860 people. This is a decent (albeit not spectacular) number, but it is not much. Seven New Haven city bus lines had a higher ridership than the Hartford Line. The Dixwell corridor itself moves more than 6,000 riders every weekday.

The thing is, most CT Transit systems are pretty terrible: slow, unreliable, unpractical, often following routes that have remained unchanged since the time they were trolley lines. Fixing them to make them faster, reliable, and practical, however, is actually cheap, as it requires changes that are essentially cost-free following some careful analysis on usage patterns. These include longer spacing between stops, less branching, eliminating low-ridership lines, and so on.

Luckily, CT Transit knows this, so they are in the process of radically updating the bus service in places like New Haven. Fixing the similarly complex, and similarly broken Hartford, Waterbury, and Stamford bus systems can potentially make them much more useful and attractive to riders, for little to no cost.

For another article: Metro North, zoning

The New Haven line is the busiest commuter line in the country, and far more complex than the other two rail lines in the state. Although it is not an entirely terrible rail line, there is much room for improvement; travel times between New Haven and New York, for instance, are slower now that they were in 1949 (the Merchant’s Limited was scheduled at 83 minutes; the fastest Acela now takes 92), even if the equipment and infrastructure are (at least in theory) much better.

Even commuter trains are significantly slower now that they were six decades ago. Although the line is close to capacity-constrained during rush hour due to limited track capacity in Grand Central (although there is room for improvement there, as well), increasing operating speeds should be a priority. The fixes are quite a bit wonkier (and some of them cost money) so they deserve a separate article.

These investments, however, need to be paired with new zoning practices and transit-oriented development. What we build around rail stations matters (hint: parking lots are not a good idea) and we need ways to ensure towns don’t mess it up (note: former Gov. Dannel Malloy had the right idea), but that also deserves a longer, separate discussion.

Roger Senserrich lives in East Haven.

CTViewpoints welcomes rebuttal or opposing views to this and all its commentaries. Read our guidelines and submit your commentary here.