Saturday, May 06, 2006

Accessibility vs Mobility: A Resolution but Not Quite the Last Word

For what I trust are obvious reasons, transportation planners traditionally position "mobility" as a key performance objective. In the past decade or so, an influential group of reformers have favored supplanting that with "accessibility."

Mobility is all about getting from A to B easily, which usually means quickly. The counter-argument is that what the trip actually accomplishes should matter more than how fast you got there. So rather than emphasize how a new road, or an additional lane, or a new connector, will increase traffic flows between two points, system planners should ask what that really amounts to for the individual traveler. They should emphasize the role of destinations.

That is, what is the value of getting from point A to B? This includes the question of what travelers can do once they get to B, or whether they want to go at all.

Accessibility is usually advanced as a concept that captures this. At that level of generality and intuition, so far so good. As I heard Mel Webber say a few years ago, "Accessibility is a simple, intuitive concept ... so some smart assistant professor should be able to explain it to me clearly." The problem is that fractures appear in the story as we add more detail.

More pointedly, the reform argument often implicitly or explicitly implies that mobility should be reduced in order to improve accessibility. This is no trifling technical debate: There will be winners and losers.

Not so subtle subtext
Mobility goals favor the car, while accessibility goals are often characterized as favoring walking, biking & transit, on the one hand, and compact mixed land use, on the other.

That is, the promotion of accessibility -- used this way -- is consistent with the new urbanism and smart growth, while mobility per se is not.

Examples of confusion
Kevin Krizek and David Levinson organized a great University of Minnesota conference on this topic in November 2004. Titled, Access to Destinations and soon available in book form, it brought together a number of researchers to discuss these points. There was common agreement over what is meant by mobility. Accessibility, as the more complex concept, was the tougher nut to crack by far.

I haven't seen the organizers' final spin on things in print form yet, but I can say that as a participant the picture was possibly less clear at the close of the conference than at the opening. Part of the confusion is the subtext, as the merits of some modes and land use features were often part of the conversation. For example, one paper talked about improving accessibility by increasing bicycle use. Alan Pisarski commented, "Oh, I see. Accessibility is the mobility that you like." The reply didn't clarify things one bit.

What to do
First, read the excellent chapter on "Access" in Site Planning by Kevin Lynch and Gary Hack. Physical planning considerations are prominent, with extremely erudite discussions of shapes, forms and functions (even parking).

Second, since you've read this far already, take the plunge with my take below. It's short.

A resolution
Accessibility is confusing because it is two completely different things. It refers to both (a) the travel characteristics of the built environment and (b) the demand for travel. Any discussion that does not fully separate these two elements out will only serve to cloud things. And yet they are almost never identified as distinct, virtually unrelated components.

I gave a talk at ACSP about ten years ago on this, "A note on supply versus demand in travel access." Subsequently, both Dru van Hengel and Lisa Schweitzer helped with an article version, but it remains sort of incomplete for reasons I can't cogently articulate.

Here's the bottom line:

Traditional accessibility measures are gravity-models, or gravity-type formulations, of the characteristics of the network (as in the London axials map above), together with the density of this or that, such as Portland food stores in Marc Schlossberg's map to the right.

Together, they measure how to get (mobility) where things are (destinations). The importance of the latter, however, depends on the demand for travel, which as we know is mainly a derived demand for what travel obtains. So one can't really measure accessibility in the abstract; the accessibility of Starbucks or a manufacturing job or a Oaxacan restaurant will vary with the tastes and circumstances of each traveler.

Language to this effect can be found in Lynch and Hack, as well as here and there in passing in more recent work by people such as geographer Harvey Miller. But these do not really spell out the fundamental supply/demand dichotomy in a way that lays out the terms on which this debate tends to founder and flounder.

Not the Last Word
While often cast as an innocent mechanism for costlessly increasing travelers' choices (e.g., see Susan Handy's presentation at the December 2004 University of Michigan conference, From Mobility to Accessibility -- scroll to bottom of page), the take-away message here is that improving accessibility is no cure-all and definitely not a no-brainer.

It will involve tradeoffs, both between access and mobility (where any change would have major implications for those who rely on cheap mobility) and among those who value different travel purposes differently (such as to work vs. to funky restaurants).

The trick is thus to evaluate the tradeoffs transparently before making them.


David said...

Response posted by David Levinson here .

randall crane said...

David, if you are proposing that we restrict access measures to supply metrics only (and I am not sure you are), that's ok with me. But I don't believe that is the case with the modern literature on this issue.

Implicitly including demand (i.e., the value of access) but not always recognizing this, explains much of the considerable confusion over these measures and their usefulness for policy. I suggest we say so plainly.


David said...

I propose that there be a set of access measures (the traditional Hansen measure, cumulative opportunities, etc.) which are measures of the supply of desinations. (Ahmed El-Geneidy and I have also developed some new measures that take actual O-D demand into account as a better way of valuing opportunities, called PlaceRank, which are I think interesting at a technical level, but still pretty difficult to communicate).

The supply of destinations has value which is capitalized in land. There is a value to opportunities or accessibility, which should also be plain.

The demand for travel to a place is of course a function of opportunities there, elsewhere, and the respective costs, as well as characteristics of the individual. The demand for development in a place also strongly depends on existing accessiblity.

I would hardly presume to speak for all researchers, but in Minnesota we have been trying to be very clear that accessibility is a measure of ease of movement to opportunities (which is far better than mobility, which is simply a measure of ease of movement without any relationship to where that movement is ... movement to opportunities is more important than movement between two vacant parcels).

The transportation profession (engineers at DOTs) has been moving very much towards performance measures or performance indicators, as they are solving a multi-objective problem (welfare is far too complex to define, much less optimize, when you are dealing both with clearing snow, filling potholes, straightening roads, and investing in new highways). Books like The Balanced Scorecard have been influential. Unfortunately, mobility has been used as a measure to the exclusion of accessibility.

Accessibility can be an important performance measure/indicator, and an important output to compare alternative investment scenarios (does investment A or B improve accessibility (to jobs, to shops, etc.) more?).

Ceteris paribus, more access is better than less (arguments like Schwartz's Paradox of Choice aside.

-- dml

Rob Dawg said...

The word "accessibility" was co-opted by the Smart Growther New Urbanists because the usual word; "cachement" was already taken and no amount of data manipulation of cachement could get them to their desired conclusion by analyzing cachement metrics. It is really simple; POV/roads mobility is unarguably cheaper and faster with perceived acceptable external costs such as health, safety and sense of community.

The real problem is that the SmUGLERs employ dishonest assertions. They conflate the fact that the current POV/roads model is not perfect as being one of fatally flawed. At the same time they promote alternatives as if they carried no externalities whatsoever.

Jonathan Levine said...

“An experienced Australian traveler once said that on business trips to Australian cities he could reckon to make four meetings in a day,” writes Thomson (1977:48). “In Europe he could manage five; in the United States he could manage only three.” The reason behind the variations in this traveler’s itineraries was not an American propensity for long meetings, or the speed of travel in American cities, which is in any case faster than in Western Europe or Australia (Kenworthy and Laube 2002). Instead, the travelers’ business contacts were spread so far apart in American cities that he could manage fewer meetings in a day. His travel speed didn’t especially matter to him; what mattered was the amount of business that he could accomplish for a given investment of time and money. The traveler was expressing an accessibility perspective on transportation.

The perspective is a necessary outcome of the consensus view that the demand for transportation is derived, meaning that people do not usually travel for the pleasure of motion per se, but in order to reach destinations. The derived nature of travel demand implies that accessibility is the service that is provided by mobility; that is, access is the ends and mobility is a means. Yet there are other routes to access as well, including proximity and remote connectivity.

The question is how we evaluate transportation-policy outcomes. If we evaluate them in terms of either mobility, proximity, or remote connectivity, we are measuring means, or inputs. It would be like developing a standard for illumination in watts rather than lumens. “Derived demand” implies that we ought to be measuring accessibility—the desired service provided by transportation—to gauge transportation outcomes. To see why this matters, imagine a hypothetical set of transportation policies and investments that, in aggregate, improve mobility but degrade accessibility. (This would happen if they induce land uses to move farther and farther apart, and the added distance isn’t fully compensated with increased speed). These policies would leave people with less time and money with which to interact with their destinations. If the demand for transportation is derived from the desire to reach destinations, this must be seen as an undesirable transportation-policy outcome.

So taking “derived demand” seriously implies reordering the goals hierarchy and putting accessibility on the top. The question then becomes, “accessibility to what,” which I take to be at the heart of Randy’s concerns. If we knew individuals’ wants, needs, and preferences, we could presumably estimate their accessibility in a given land-use and transportation environment. In absence of that knowledge, we can use access to a range of choices as an approximator. You may like Oaxacan restaurants and hate Canadian restaurants (lucky you live in LA and not Ann Arbor), but we don’t know that, and as a consequence measure access to retail and services generally. There’s a reason behind it: the larger the number of goods and services you can reach with a given investment of time and money, the more likely it is that you can get what you want.

One hopes that we’ll get better at measuring accessibility, based on improved data and methods. Accessibility measurements are surely an approximation, as opposed to mobility measurements, which can be very precise. But the derived nature of transportation demand means that as a policy yardstick, accessibility is approximately right, and mobility is precisely wrong.

Jonathan Levine


Kenworthy, Jeff and Felix Laube (2002). Urban Transport Patterns in a Global Sample of Cites and Their Linkages to Transport Infrastructure, Land Use, Economics, and Environment. World Transport Policy and Practice, 8(3), 5-19.

Thomson, J. Michael. (1977) Great Cities and their Traffic. London: Gollancz.

Michael said...

It seems to me that (at least in the context of intracity travel) terms such as "accessibility" and "mobility" create more confusion than they are worth.

Maybe a better way of phrasing the issue would be "mobility for cars alone" vs. "mobility for everyone."

Imagine a hypothetical auto-utopia (or maybe dystopia), where people could get anywhere within half an hour by driving 70 mph but where there were no sidewalks or public transit, and every street was twelve lanes wide. Drivers would have high levels of mobility. But people who (for one reason or another) did not drive would have zero mobility.

Query: is this the society we want? And if not, what are we going to do about it?

Anonymous said...

Good article and good point!

Dennis from Axistive.comod point.

David said...

The first publication from the Access to Destinations study has now been posted at (click to the Final Report, which is a pdf). Comments are welcome.

-- David Levinson


My blog has moved!

You should be automatically redirected in 6 seconds. If not, visit
and update your bookmarks.