Blog

Can Anybody Tell Us How Many People Are Riding Transit?

Posted by: Tony Dutzik on

Tags: transportation , MBTA

In late 2015, the MBTA’s Fiscal and Management Control Board, in its first annual report (PDF), made the startling claim that MBTA ridership – which grew nearly twice as fast as the region’s population since 2000 and faster than that of transit agencies in nearly every major city since 2008 – was “basically flat.”  

The claim was supported by the following chart (which, apologies, may be hard to read), which drew instant ridicule from local transit advocates (check out those y-axes and the lack of inflation adjustment) and even led to the creation of a parody Twitter account.

 

This week, the MBTA resurrected the “basically flat” chart, but with an important difference:

Note that in the earlier version of the chart, ridership for FY 2015 was estimated at 394 million, whereas in the 2017 version, ridership had been revised down to 385 million. That kind of revision is not usually a big deal. But the backstory behind it reveals a big problem with our ability to understand precisely what is happening in our transit agencies and our transportation system as a whole – a problem that I fear makes it far more difficult than it should be to make intelligent transportation policy choices.

You see, the MBTA’s ridership numbers for both FY 2015 and the preceding fiscal year have ping-ponged around over the last 15 months in ways that lead to dramatically different conclusions:

  • In early 2016, we wrote a blog post contesting the MBTA’s “basically flat” assertion. That post led to a quibble over numbers with the MBTA that wound up being covered by Commonwealth magazine. At that time, the MBTA reported that ridership in FY 2014 was 398 million and ridership in FY 2015 was 394 million, for a year-over-year decline of 0.9 percent. Keep in mind that 2015 was the year of “Snowmageddon,” when a series of major snowstorms brought the MBTA to a screeching halt. A decline in ridership was not unexpected.
  • In our blog post, we cited figures from the National Transit Database (NTD) showing that MBTA ridership was 409 million in FY 2014 – much higher than the figure then being circulated by the MBTA. We did not have NTD figures for FY 2015 at the time, but if you go into the NTD today, you would find a ridership figure for FY 2015 of 406 million. The year-over-year decline is very similar to the MBTA’s figures (0.8 percent versus 0.9 percent), but total ridership for both years was much higher - a difference that means a great deal in assessing long-run changes in transit use.
  • The new, 2017 version of the “basically flat” chart, however, uses the higher NTD ridership figure for FY 2014 (409 million) and a revised lower estimate of ridership in FY 2015 of 385 million. The MBTA’s excellent Data Blog explains that the FY 2015 data the agency originally reported to the NTD included an error in bus ridership, which was presumably the reason for the downward revision in that year. The new data, however, imply that year-over-year ridership fell in 2015 by a whopping 5.9 percent – a much greater rate of decline than was previously understood to be the case.

Again, as someone who lived through the winter of 2015, the idea that ridership would have fallen to that degree is not surprising. But you see the problem here: there are three sources/vintages of information – all produced by the same agency – that lead to three very different conclusions about short-term and long-term ridership trends. Assuming that the data now being circulated by the MBTA are correct, anyone who thought they knew what was going on with MBTA ridership based on what the agency told them previously likely had it wrong.

And that’s not all: as MassInc Polling Group’s Steve Koczela has pointed out for years, no one knows with any great accuracy how many people ride MBTA commuter rail, or whether ridership has been trending up, down, or been “basically flat.”

Nor is transit the only mode of transportation that suffers from inconsistent or contradictory data. We’ve noted before the significant revisions that occur between the U.S. Department of Transportation’s monthly and annual estimates of vehicle travel. A person picking up U.S. DOT’s January 2017 Traffic Volume Trends report would have learned that Americans drove 50 billion fewer miles in 2015 than had been estimated in the December 2015 version of the same report. That’s a difference of 1.7 percent – not a big deal in the grand scheme of things, but very important if you’re trying to get a handle on trends.

Here’s why this matters: transportation behaviors are changing rapidly. As transit consultant Jarrett Walker and others have noted, official data suggest that many of the nation’s transit agencies have experienced dramatic recent declines in bus ridership, and no one knows precisely why. Similarly, the number of vehicle-miles traveled in the U.S. has seemingly shot up in the last couple of years. Reporters, transportation analysts and policy-makers – all of whom understandably want to know what’s going on – keep close watch over data presented in federal reports, often without much of an idea as to how squirrelly those data sets can be.

It is not just that transportation analysts and policy-makers are flying blind – that would almost be an improvement. Rather, the problem is that we often provide interpretations on the basis of “best available data” that is either of sketchy provenance or subject to significant later revision. That breeds snap judgments and bad decisions.

In 21st century America, there is little excuse for this kind of thing. If we want to use all the powers at our disposal to manage the transportation system for the maximum benefit of its users and society, we need better, faster and more consistent data – along with people capable of interpreting those data with intelligence, honesty and a solid understanding of their limitations. We have a long way to go.