skip to main content

→ Top Stories:
Clean Power plan
Safe Chemicals

Kaid Benfield’s Blog

Estimating the impacts of smart growth: the importance of forecasting traffic correctly

Kaid Benfield

Posted July 1, 2010

, , , ,
Share | | |

The metropolitan planning organization for the region surrounding San Diego has adopted a new methodology for predicting the impacts of development.  The more sophisticated model that the metro Association of Governments will now use more accurately reflects the differences between walkable neighborhoods and conventional suburban development.   (This is another seriously wonky post, but it’s an important one.)

Before considering new development for approval, regulating authorities generally require an assessment of the likely transportation impacts, particularly on traffic.  If new housing, shops and/or jobs are likely to create additional driving trips, walkable, transit-accessible, mixed-use in San Diego's Little Italy (by: LA Wad/Chris, creative commons license)it is important to know that in advance, both to assess the merits of the project and the potential need for additional infrastructure.

The problem is that almost all jurisdictions in the US employ a forecasting methodology (“model”) that has not changed much since the 1970s and that fails to account for the considerable effects that differing particular characteristics of development can have on the outcome.   For example, research reported here in 2008 from metropolitan Philadelphia, Washington, Portland, and San Francisco determined that, on average, transit-oriented development reduces car trips by 49 percent in the morning peak period and 48 percent in the evening peak, compared to what would be expected from standard engineering estimates typically used by municipalities.  Ewing and Cervero’s painstaking review of 50 published studies found that increasing any of five factors associated with smart growth – destination accessibility, street connectivity, transit access, mix of uses, and neighborhood density – can reduce driving per capita compared to conventional suburban development.

But if the travel forecasting model does not reflect the extent to which walkable, well located neighborhood development reduces driving, the result could be inflated traffic predictions that disfavor good proposals, walkable, mixed-use in Hercules, CA (by: Greenbelt Alliance, creative commons license)leading to disproportionate fears on the part of nearby residents or regulatory disadvantages.  These could take the form of higher impact fees, exactions and negotiated payments, or even disapproval altogether.

To address this problem, last year the federal EPA commissioned a study to more properly account for the impacts of mixed use and the other smart neighborhood characteristics noted above on traffic.  The transportation consultancy Fehr & Peers, which led the study, found the following:

“The study evaluated household travel surveys from 239 mixed-use developments in Seattle, Portland, Sacramento, Houston, Atlanta and Boston and found statistical relationships between site characteristics and the amount of vehicle travel generated in and out of the site. MXDs were found to reduce traffic impacts relative to single-use suburban development, due to the following key factors such as diverse on-site activities that capture a large share of trips internally, placement within walkable areas with good transit access that generate high shares of walk and transit trips, and central locations that reduce trip lengths.”

The result was the construction of a complex trip generation equation, called ‘MXD’ (for mixed-use development) and explained in part herewalkable La Jolla, near San Diego (courtesy of Eric Fredericks, equation more accurately predicts the amount of driving that a development will create and corrects the deficiencies of outmoded models.  The development of MXD was led by Reid Ewing at the University of Utah and Jerry Walters at Fehr & Peers.

Last week, the Board of the San Diego Association of Governments approved the MXD method as the preferred means of adjusting trip generation estimates to account for the effects of smart growth.   Fehr & Peers reports that the method is also under review by the Institute of Transportation Engineers for wider adoption, and that it is undergoing evaluation by panels of experts and practitioners in California as part of a study to assess its acceptability for use in development reviews required under state law.  This is good news for walkable neighborhoods and accurate environmental assessments.

Move your cursor over the images for credit information.

Kaid Benfield writes (almost) daily about community, development, and the environment.  For more posts, see his blog's home page


Share | | |


Josh MartinJul 1 2010 12:47 PM

Kaid: As usual, fabulous post--my thoughts: At the very basic level, transportation studies are submitted with major development projects of a certain size in the City of Charleston. The very nature of transportation or traffic impact studies are based upon suburban conditions and have been used in the past few decades to justify suburban sprawl with its list of recommendations stated in order to remedy or resolve the impacts of the proposed development around its environs—which typically included widening of highways, roadways, access management plans which in return have led to the complete degradation/bastardidation of the public realm in suburbia. The City also requires these studies for the urban areas of the City—i.e. Downtown. So, what will this study tell us? Hmm, developer, you need to widen Calhoun Street to seven lanes, or you need to build a new interchange at the end of Montagu Street? See, the traffic impact studies DO NOT work as they are currently calibrated as they base factors upon which models are run are suburban factors (i.e. Models are not set up to model network of streets which are inherent in urban areas of the City)---thus an incompatible/useless/shelf-serving study results. So, what is the alternative. While at the City, I suggested that instead of having the developers spending all this money on these useless studies, why not be a bit more progressive and have the City conduct its own master transportation study that would include the following: model of build out scenario under current zoning/height, MULTI MODES OF TRANSPORTATION NEEDS—mass transit, car, foot, bike, water taxi, etc, mass transportation infrastructure such as shelters/transit stations if light rail comes, water taxi platforms, opportunities for CARTA improvement, trolleys, bike racks, parking, economics, retrofit of sprawl thoroughfares downtown, etc. Essentially, the plan would provide a “master plan” for transportation planning within the study area. The City would need to issue an RFP that would be composed at a very highly sophisticated level by someone who understands the aforementioned concepts. The RFP could perhaps be funded by one of the TIF districts downtown and the master plan would also include a Capital Improvements Plan for the aforementioned identified needs. Then, as developers come to develop downtown, they would pay a fee-in-lieu of the transportation study requirement based upon the size of their requirement into the pot that would fund the improvements.

Charles MarohnJul 1 2010 03:54 PM

Important topic - great post.

One could go on forever about this, but I thought I would respectfully add that there is also the impact of induced demand. The traditional model's oversizing of auto infrastructure does create this hostile development environment, as you describe, especially for high-density projects.

But that oversizing, if built, also induces traffic that would otherwise not exist. The excess capacity (high supply) allows people to drive more, so they take lots of extra trips.

In a free market that may be their choice, but in this case the regulation has artificially created an excess supply of transportation infrastructure, the costs of which are not paid by those that use it. The excess supply creates artificial demand, not just locally but throughout the system.

Just another way in which we are our own worst enemy.

Kaid @ NRDCJul 1 2010 05:17 PM

Josh and Chuck, great additions to the conversation.

Josh, with luck, the kind of master planning you suggested for Charleston will be among the things produced out of the SB 375 process in California, with the land use appropriately coordinated and vice versa.

Chuck, there is lots of evidence now on induced demand, as I am sure you know. And, as you correctly point out, the costs are borne by the public at large. Eventually, the induced demand leads to induced development and within a decade or less the corridor is even more congested than before.

Ann DaigleJul 1 2010 06:40 PM

Great discussion - and appropriate nationwide, even for small and mid-size cities.

Kaid, this context-sensitive approach is the very one detailed in Transect planning. Transportation sections, parking requirements and urban design details are "coded" to their appropriate level of intensity. The T6 Urban Core or T5 Urban Center (transit rich mixed use that is highly pedestrian friendly) or T4 General Urban fabric (walkable but with a predominance of more residential building types) are different - and have different needs. The Transect offers us an operating system within which to measure and compare different scenarios.

I am interested in how this approach is impacted by (or impacts) CA planning legislation. General plans require traffic analysis for transportation elements, and must meet CEQA. Can we hope that these outdated requirements might also change?


Patrick KennedyJul 1 2010 07:21 PM

Hi Ann,
Nothing like interacting with someone you know on a national website...

A transect-based approach might ultimately be what is called for here as this feels like a first step to me. My initial reaction was that there are more "tiers" in trip reduction than a simple dichotomy of urban and suburban.

For example, knowing residential/mixed-use developers here in the Dallas area and having discussed with them some of their internal studies, they've found that walkable Urban leads to approx 40% reduction in trips whereas walkable SUBurban leads to approx 20% reduction in car trips (both in relation to conventional drive-only suburban). The walkable urban is from uptown Dallas and the walkable suburban data is from Legacy Town Center in west Plano (no transit).

Kaid @ NRDCJul 1 2010 10:14 PM

My understanding of the F&P model (MXD) is that it is much more sophisticated than simply dividing developments into 'urban' or 'suburban' categories. Instead, it accounts for all five factors measured by Ewing and Cervero, with the traffic estimate derived from how the development will perform given its combined characteristics considering data on all five variables. In a sense, it measures an infinite number of possibilities, not just two or a specified number of transect zones.

Watch this space tomorrow for a graph depicting the relative effects of each factor on vehicle miles traveled, and see the E&C study for the influence of the factors also on walking and transit usage.

Alvin SarmientoJul 7 2010 07:42 PM

The MXD model is an excellent new method that could be implemented across the United States. With increased congestion and reduced transportation funds, MXD is a good start to the industry as we need to find ways to reduce errors.

Comments are closed for this post.


Switchboard is the staff blog of the Natural Resources Defense Council, the nation’s most effective environmental group. For more about our work, including in-depth policy documents, action alerts and ways you can contribute, visit

Feeds: Stay Plugged In