Access across America!

  Levinson cover

There should be nothing amazing about a new report on how easy it is for Americans to get to work on transit, but there is.   Think about all the arguments we have about transit …

… and ask: Why do we try to discuss these things in the absence of good analysis of the most basic question of all:  Is transit useful?  Does it help people get places in at a time and cost that's a logical choice for them?  Such information is often hiding inside ridership models, but it's rarely revealed in a way that lets people see and discuss it.  Without that, it's hardly surprising that the American transit debate is so confused.   

Access across America: Transit 2014, from the University of Minnesota’s Accessibility Observatory, describes how easy it is to get to jobs in America’s major metro areas by way of transit plus walking.  (For very short trips, it shows what can be reached by walking alone.)   The authors are Andrew Owen and Professor David Levinson.  The report is meant to sit alongside similar studies for the other transportation modes.

So here’s Portland, say, shaded by the number of jobs that can be reached within 30 minutes from each point in the city:

Levinson pdx
The report also offers a ranking of how easy it is to get to jobs for the average residential location in all the major metros.  The ranking looks at where you can get to in a range of travel time budgets, from 10 minutes – basically a measure of walk access – to as much as 60 minutes.  Here's are the top 17:


As the travel time budget rises, the relevance of transit to economic opportunity becomes visible.   Los Angeles, with long access distances but extensive frequent transit, does poorly on 10-minute walkability but climbs in the rankings as you consider longer travel time budgets, thanks to its effective frequent transit grid.  Miami drifts in the other direction, signaling that compared to Los Angeles, transit there is adding relatively little to access to jobs beyond what’s achieved by walking alone.

Now here's what's amazing for a study pubished in 2014:  Owen and Levinson claim (p 6) that this study is unusual in properly accounting for frequency.  Many analysts approach transit from the point of view of the nine-to-five commuter, who was presumed to be largely insensitive to frequency because they have made an appointment with a particular scheduled trip that they take every day.   This sometimes feels right to bureaucrats and civic leaders, many of whom have such commutes themselves, but out there in the larger economy, more and more people work part time, or at irregular hours, or at times outside the standard commute peak.  Increasing numbers of people also value flexibility and spontaneity even in work trips — things that only a robust all-day frequency can provide.

Perusing these maps and rankings, my overwhelming reaction was “what if they’d analyzed it like this, or graphed it like that?”  A huge amount of insight is readily available out of this database if we query it differently.  For example:

  • Instead of ranking cities by the number of jobs reachable on transit in a given time, what if we ranked them by the percentage of jobs accessible?  The current rankings are still, predominantly, just a ranking by total volume of jobs.  Doing it in percentage terms will really pop out the winners in their size class, like Portland.
  • For the same reason, I really want to see Canadian cities ranked in this way.  They tend to have far more transit service per capita than comparable US ones, and higher ridership per capita as a result.  If this shows up in dramatically better economic opportunity and personal liberty, it could create a powerful contrast when cities compare themselves with similar ones across the border.
  • Why confine our attention to 7-9 AM, the classic morning commute peak?  There’s a good argument for starting there: it’s when the maximum number of people are trying to travel.  But the time-of-day dimension is essential to understanding the real lives of the majority of workers who are not peak commuters: those who work part time and in non-standard shifts, like almost everyone in retail, entertainment and manufacturing.
  • Let’s look at access to other things besides jobs.  All transportation studies overemphasize the journey to work because we have better data on it than on anything else.  But with the appropriate layer about locations, we can explore access to retail, access to food, access to education, even access to nightlife.   Regions may not have this data, but many cities do, and much of the interest in this tool will be at the municipal level.

Still, this is a great piece of work.  And Americans should pause over the core of this announcement:  Only now, in 2014, are we starting to study people’s ability to get where they’re going, and their opportunity to access all the opportunity that makes a great city.

11 Responses to Access across America!

  1. Matthew McKibbin October 8, 2014 at 4:19 am #

    Accessibility is a very important metric of transport-land use integration that is also relatively easily understood (how many jobs can I get to?).
    I’ve used accessibility data in a study of the importance of the 5 Ds of the built environment (see Cervero) in Sydney, Australia. This study found that accessibility was the most important built environment factor correlated with public transport use – more important than density, pedestrian oriented design and mixed land uses, for example. For further info see the paper:

  2. Jim October 8, 2014 at 8:09 am #

    Often neglected in talking about transit is land use planning, and often an oddity is the separation of transportation and land use planning. The top rated cities have densities but also residential and commercial zones within walking distance of each other, similar to European cities. If we want to make transit attractive in America, we need to talk about changing zoning laws and land use planning. Also, please stop talking about making things sexy especially transit. A. it dilutes the meaning of sexy and B. transit will never be as sexy as lingerie.

  3. Zoltána October 9, 2014 at 1:28 am #

    I might suggest a cold shower, then getting back to the matter of transit.

  4. Michael D. Setty October 9, 2014 at 2:11 am #

    I would ask Matthew how his findings can be translated into a form that elected officials and the general public can understand. Skimming his paper, it appears to me that the importance of “accessibility” is another way of saying that a comprehensive network is needed to maximize transit usage, e.g., one that connects “everywhere to everywhere” like in Switzerland and I think shown clearly by Paul Mees.
    Do you agree with this assessment?

  5. Matthew McKibbin October 9, 2014 at 4:43 am #

    Hi Michael, I think the paper suggests that we should be strategic about where new housing and employment is located, but that we should also be realistic about the potential magnitude of change in travel behaviour resulting from TOD.
    One of the cheapest ways to improve public transport use is to plan for more jobs and housing along the region’s public transport network, particularly in the most accessible areas.
    We also need to be aware that TOD is not likely to be successful in terms of travel behaviour if density is located in inaccessible areas, like a train station with a twice daily commuter service to a city over an hour away.
    And yes, investment in improved or new transit services will also help. These investments should target areas of low accessibility with potential to build demand. This may be achieved though simple bus priority measures right through to new metro/heavy rail lines.
    One issue that I didn’t look at is to what extent decentralised employment would help or hinder accessibility. Jobs closer to the suburbs helps improve access from those areas, but are less accessible to the wider metropolitan area if not located at the core of the transit network. Thoughts welcomed.

  6. Matthew McKibbin October 9, 2014 at 6:57 pm #

    One more comment – the analysis was for the ‘journey to work’ or commute, which in Sydney is around 27% of trips. It is entirely plausible that density, mixed land uses and pedestrian oriented design are more important in influencing other (local) trips.

  7. threestationsquare October 10, 2014 at 1:29 pm #

    Jarrett: “Instead of ranking cities by the number of jobs reachable on transit in a given time, what if we ranked them by the percentage of jobs accessible?”
    The study actually provides all the data necessary to do this, so I’ve gone ahead and done so. Here are the rankings by percentage of jobs accessible by transit with a given time budget, and here are the raw percentage data. Averaging the percentages with the same weighted average formula used by the original study, the top ten are: San Francisco, Salt Lake City, San Jose, Milwaukee, New York, Portland, Buffalo, Denver, Washington, and Seattle.
    Measuring access this way arguably unfairly penalises large multi-polar metro areas, where every resident has access to many jobs but none has access to a large proportion of the region’s total jobs. For example if San Francisco and San Jose were considered as one metro area instead of two it would be much lower in the rankings than either of them is alone.
    An alternative metric is to compare the number of jobs accessible by transit with a given time budget to the number accessible by auto with the same time budget; the ratio essentially measures how competitive transit is with cars in providing its users with access to jobs. Conveniently the Accessibility Observatory published a report last year with auto access data, and I’ve used it to compute rankings (and raw percentages) for this transit competitiveness metric as well. The weighted-average top 10 for this metric are: New York, Boston, Seattle, Philadelphia, Washington, Chicago, San Francisco, Portland, Denver and Milwaukee.

  8. Wanderer October 10, 2014 at 3:21 pm #

    This was a really interesting study.
    It seemed to me that both the absolute number of jobs and the percentage of jobs in a metro area that one could access via transit was important. They each tell you something different.
    So I calculated a “dual” rank. I took the percentage of jobs accessible within 60 minutes, like threestationsquare. Then I added the regions’ ranks for both total and percentage jobs accessible. This weights both scores equally, though one could weight them.
    So the top 10 were San Francisco, New York, San Jose, Washington, Denver, Milwaukee, Portland & Salt Lake City (tied), Seattle, and Los Angeles.

  9. Anne October 11, 2014 at 5:57 pm #

    Just for Jim: some of the problems you get with ‘sexy’ transit: it looks nice, but it’s impractical, it costs a lot to buy, it’s more expensive to take care of, and it’s not very useful for day to day living. Great on a date or special occasion, but not much use when you’re trying to take your RV in for maintenance, or get to work:
    OTOH, I’d call this “useful transit”: (RIP JP AuClair)

  10. Bob Clark October 19, 2014 at 4:39 pm #

    This is the best discussion I have seen in a long time, but I don’t see any reference to my personal demographic which is a retiree who wants to ditch his car and all the expenses that go with it. I also would like to see school children catered to by public transit so that they can be socialized instead of being brought up to regard themselves as a privileged class with little thought but when can I get my license.

  11. Nathanael November 5, 2015 at 3:45 pm #

    Same cities keep showing up in the lists no matter how you measure or weight them, don’t they. It’s the Northeast Corridor cities; the West Coast cities; Chicago, Milwaukee & Minneapolis; Denver; Salt Lake; Miami; and Houston.
    All these cities except Houston have something in common. The attitudes are genuinely regional.