The FTA just released 2020 breakdowns data.1 Before the sensationalists dub New Jersey Transit the worst in the nation again, let’s analyze the data better.2
I’ll use everyone’s favorite statistical manipulation: Long division. Specifically, I argue that it is wrong to declare, say, Metro-North–a rail network comprised of hundreds of passenger cars running millions of miles per year–worse than Tennessee’s tiny Regional Transportation Authority, simply because the RTA had only 1 breakdown while Metro-North had 50. Instead, I divide mechanical failures by vehicle miles traveled to get the failure rate–breakdowns per 1 million miles–and I judge transit agencies by this metric.
In doing so, I hope to counter the sloppy analysis that makes headlines every year. Lawmakers heed these articles. Laws are drafted and funds allocated based on their findings. It’s unfortunate, then, that those articles are misleading and missing out on important insights.
These tables are sorted by the blue column, Failure Rate Rank. The higher an agency is on the rankings (that is, the closer it is to #1), the more often its vehicles break down.
Once again, NJ Transit’s commuter rail isn’t actually the worst in the nation. It ranks 18th out of 29 for mechanical failure frequency. 5 breakdowns per 1 million VRM–vehicle and passenger car miles traveled in revenue service.
It’s better to compare a big operation like NJT to agencies of comparable size. Here are the 10 largest commuter rail networks by passenger car miles:
Once again, NJT rail falls exactly in the middle. Barely better than SEPTA but worse than LIRR and Metro-North. It doesn’t sell newspapers or rack up retweets, but it’s our mediocre reality. We should strive to be like Metro-North.
Note the change in NJT’s rank since 2019. Compared to all agencies, NJT commuter rail improved; it was 16th worst last year, now it’s 18th worst. Yet its rank among comparably sized agencies didn’t move. So hold your applause; NJ Transit commuter rail isn’t improving.
But–ignoring its rank–NJT’s failure rate did improve a tiny bit. There are now 5 mechanical failures per 1 million passenger car revenue miles, down from 6 last year. It seems most agencies’ failure rates fell in 2020, perhaps due to widespread service reductions during the COVID-19 pandemic.
The directly operated3 portion of NJT’s bus network has the 6th most breakdowns. That looks quite bad if you ignore a relevant factor like its size. But with proper context, they’re actually 255th worst out of 463. Like its commuter rail, NJT’s buses fall in the middle. They also improved in the last year, falling 18 spots in the rankings, getting even further from “worst” in the nation. It’s not all good news. NJT’s purchased buses got slightly worse and 15 spots closer to the “worst”.
But it’s unfair to compare behemoths like NJT or the MTA to Maple Grove, Minnesota. Here are the 10 largest bus networks:
That’s right, NJ Transit’s buses are actually 2nd best among its peers, right behind the MTA. And in the last year, NJT’s bus failure rate fell from 116 to about 108 breakdowns per 1 million miles. This underscores the irresponsibility of how media outlets portray this data. They fabricate a mechanical failure crisis, when we should actually hope NJT’s trains one day get as good as its buses.
2019 repeats itself yet again. The directly operated portion of NJT light rail–the River Line and Newark Light Rail–is truly bad: 3rd worst in the nation (out of 23). But you wouldn’t know it if you only paid attention to total breakdowns. My lovely statistical manipulation uncovered a problem that others missed.
Meanwhile, the Hudson-Bergen Line–the purchased portion of NJT’s light rail–is 2nd best in the nation (22nd worst out of 23).
Here are the 10 largest light rail networks, though NJT isn’t among them:
And the Amazon gift card again goes to San Francisco.
I relied on the Northeast Corridor and North Jersey Coast lines to get to work in midtown Manhattan daily. It was fine most days, especially in the morning. It could get truly awful after work. I was never on a train as it broke down, but I remember cancellations due to equipment failure. So it’s easy for me to believe the numbers when they say NJ Transit commuter rail isn’t great but isn’t the worst.
New Jersey’s trains–especially light rail–need work. That’ll require careful analysis and competent leadership. It’s a shame, then, that the media won’t get the metrics right.
Mechanical failure rate is just one metric by which we should judge our transit. The buses don’t break down the most often, but are they on time? Are they clean? Do they run often enough? Are bus stops accessible for people experiencing disability, thoughtfully located, protected from extreme weather and car traffic? It doesn’t take much searching to find flaws.
We need more and better public transit in New Jersey (build the damn GCL!). We need trains and buses everywhere there are masses of people. We need less driving. Let’s fix transit in New Jersey. Let’s be smarter about measuring its performance.
Check out my blog.
This article is licensed under the Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0).
Here’s the code I used to clean the FTA breakdowns data.↩︎
You’ll notice there are 2 NJ Transit rows in the bus and light rail tables. Note the “Type of Service” column. One row is the “Directly Operated” portion of, say, NJT’s bus network, while the other row is the “Purchased Transit” portion. In the light rail table, DO means the River Line and Newark Light Rail combined, while PT means the Hudson-Bergen Line; they’re identified as such here.↩︎