In the late 1840s, a physician named Ignaz Semmelweis was working in the maternity ward of Vienna’s general hospital when he noticed that women were dying at an unusually high rate. At that time, people did not have a good understanding of what caused disease, and often attributed illness to “bad humours”, unbalances within bodily elements, or dangerous vapors from swamps and other places in the environment. Seeking to reduce the mortality rate, Semmelweis set about trying to figure out what was causing the abnormally high number of deaths. At first he thought it might be the position in which women were giving birth, so he ordered all would-be mothers to stop giving birth on their side—the mortality rate remained unaffected. Next, he speculated that it might have something to do with a priest’s daily ritual of walking through the hallways ringing a bell in blessing, so he sent the priest home---still no change in mortality. It wasn’t until a fellow doctor died from the infection of a scalpel prick during an autopsy that Semmelweis began to uncover the real cause and effect---right before his death, the other doctor had shown symptoms remarkably similar to those of the women who died in the maternity ward.
As he investigated further, Semmelweis noted that at this hospital, doctors often performed autopsies immediately prior to performing births, and as a result women seemed to be dying more frequently. Women who gave birth with midwives (who did not perform autopsies) seemed to have a better chance of survival. Believing that some sort of agent was being transferred from corpse to living person, Semmelweis created a new regulation that required that all doctors rinse their hands with a chlorine hand wash before performing births, and the mortality rate dropped dramatically. Two centuries later, it is clear enough to see that Dr. Semmelweis had stumbled upon the germ theory of disease. Unfortunately, at that time, other physicians considered these findings distasteful and they were ignored; germ theory was not fully accepted in the field of medicine until several decades later.
This story is a great example of how badly human beings can misunderstand cause and effect, even in the face of what appears self-evident. In the modern day, we probably assume that we have progressed beyond the capacity for this scale of delusion, but unfortunately we haven’t---including in the field of planning. In just about every planning office in the US today, there seems to be an underlying misunderstanding which has reshaped the land use of an entire continent and led to a whole range of negative health, environmental, and economic consequences over several generations. This idea, of course, is that land uses can “generate” motor vehicle trips, referred to as trip generation. Although it isn’t as damaging as the blind faith that bad humours can cause illness or ghosts are the source of disease, it has done significant damage to our entire built environment.
In American planning, trip generation is the idea that land uses magically “generate” motor vehicle trips, and these trips then must be accommodated with more driving infrastructure. Go into any planning department and propose a new development—an office, an apartment building, or maybe a brewery—and one of the first things the planner will tell you is how many trips your proposal is going to “generate”, what traffic impact study you will have to pay for, or how much parking space or road improvements you will have to construct to accommodate it. It doesn’t matter if your future tenants, residents, or patrons don’t drive cars and never will, the actual actions of future individuals who will make the future travel decisions is unimportant. Under the paradigm of trip generation, it is the land use that uniquely determines how many trips will occur, and that is what you have to build to. Belief in trip generation is one of the reasons why so much of our land is sprawling parking lots and fast, expansive roads, and why so much of the United States is simply inaccessible without a car—we mandate that it be designed that way with every new development.
Where Does Trip Generation Come From?
Planning is the management of space, and it’s unclear how trip generation became the standard for managing space related to travel and motor vehicles (although it probably had something to do with trip generation being a fairly straightforward and consistent tool). At least one historical reference suggests that perhaps it could have essentially been a matter of convenience—predicting a person’s viewpoint was difficult, whereas setting a formula based on a land use was not.[1] Today, the gold standard for trip generation comes from the Institute of Transportation Engineers (ITE), which publishes a series of guides on how the concept works. In the third edition of the Trip Generation Handbook, the standard reference, the ITE defines trip generation as follows:
Trip generation is one measure of travel behavior. It is based on the notion that people regularly travel to or from a particular land use and location and that the amount and type of activity at the location—whether retail, office, residential, or service-oriented—uniquely determines the amount, type, and mode of that travel.[2]
The stated causal relationship here is that land use “uniquely determines” travel. Though the modern ITE guides clarify that trips include all modes of travel, when planners and engineers
talk about trip generation in general usage today, they are usually talking solely about motor vehicles. The ubiquity of this idea cannot be understated--look at just about any development code anywhere in the United States and you will find blind faith in trip generation embedded deep in the regulations. Go to any public hearing for a new apartment building and people will be nodding angrily about proposed apartments brimming over with motor vehicles pouring into the streets. No one challenges it, and so the misunderstanding stumbles onward.
Under this regime, proposals for new developments are often required to provide traffic impact studies which in turn predict huge traffic flows and vehicle storage needs—based on the trip generation rates published by ITE. New residences are always required to provide a certain number of parking spaces—also based on the idea that each residential unit “generates” a certain number of trips or “needs” a certain number of vehicles. As a result of this foundation, our built environment is designed from the start as if motor vehicles are constantly being generated out of every land use; a quick glance at the monotonous arterial thoroughfare of any suburb in America shows how this has played out on the ground.
Why Trip Generation Doesn’t Make Sense
Trip generation is so deeply ingrained in our thinking that most US planners may have never even considered that it is worth a second look. However, once we give ourselves permission to doubt the entire paradigm, it doesn’t take too much scrutiny of the concept to see that the logic behind it is flawed. Trip generation says that land uses “uniquely determine” or “generate” travel—yet it is obvious that land uses don’t make trips, people do. A motor vehicle trip doesn’t occur because a land use exists, but because a person chooses to make that trip. And how do people decide to make travel decisions? They decide where to go and what mode they will take based on the infrastructural context available to them. And who or what determines what infrastructural context will be available? Municipalities, states, and the federal government: the entities that determine the infrastructural context when they build or require infrastructure. These governments mandate how many parking spaces must be set aside on private property, how fast road speeds should be, and how much land should be dedicated to roads. If we’re looking for the reasons people drive so much, we shouldn’t blame land uses—we should blame the governments that require the infrastructure that makes driving convenient.
If we are trying to predict whether a person is likely to travel by car, foot, bike, horse, or helicopter, we would be much better served to base our estimation on the surrounding infrastructural context rather than the land use at their origin or destination. Imagine three identical pharmacies—one on a rural Alaskan island, one in suburban Oklahoma, and one in Manhattan. Let’s assume that the size, layout, product mix, and use of these three places can be exactly the same—according to trip generation, all of these places should see the same number of cars driving up on a regular basis. Of course, this notion is absurd. If we were to predict how people would travel to these pharmacies, we might say a person in rural Alaska might take a boat, plane, or ATV; a person in suburban Oklahoma will likely drive; and a person in Manhattan will likely walk or take the bus. The different mode choices have nothing to do with the land use, but the infrastructural context. Despite the insistence of the trip generation theory, there is no connection between the land use and the mode of travel. (There is one caveat to this—travel infrastructure, like roads and parking lots, is a land use, which facilitates and thus encourages a specific type of activity—travel by motor vehicles. The more of this land use you have, the more motor vehicle travel you are likely to see. The same would go with canals, railways, monorails, or airports for their corresponding modes of travel.)
Even if someone refuses to acknowledge that trip generation fundamentally doesn’t make much sense, there are other reasons to do away with it altogether. Primary among these are the flawed data the ITE uses to justify their trip generation numbers. The 3rd Edition of the Trip Generation Handbook states:
Nearly all data presented in the current Trip Generation Manual data volumes have been collected at low-density, single-use homogenous, general urban or suburban developments with little or no public transit service and little or no convenient pedestrian access. These proxy sites are called baseline sites in this Handbook because they are the starting points for most vehicle trip generation estimation recommended in the following chapters.[3]
Let’s consider that for a second. Although trip generation is based on the idea that there is a causal link between land uses and motor vehicle trips, the ITE is acknowledging here that most of their data, which suggests that people drive a lot, is from environments where people have been encouraged to drive a lot. The baseline is a “pre-programmed” environment that doesn’t really tell us anything useful; observing how drivers drive in a place meant for driving is like observing how fish swim in a place meant for swimming. Where you observe, and what you choose to observe, makes all the difference.
Unfortunately, City planning departments regularly use ITE data to require more parking, wider roads, and greater speeds, all of which endlessly reproduce the suburban context that will later serve again as the baseline environment for ITE data in the future. The baseline is, and always has been, a contaminated sample and a positive feedback loop propelling us constantly forward towards a universal suburbia. Of course, these are not original or novel observations, observed by academics such as Donald Shoup and Jonathan Levine long ago. What is surprising, however, is how prevalent the use of ITE materials remains in modern planning, and how we continue to reproduce the same faulty thinking every time a new development comes across our desks. Even though “progressive” places might concede that land uses might “need” less parking or have “lower” traffic impacts, they seem to overlook that the land uses themselves were never the source of any need or impact at all.
At this point the defenders of the status quo will say: “But you quoted the trip generation manual for suburban sites. The individual context-specific ITE manuals cover the urban and mixed use sites. ITE says trip generation data should be from comparable sites, and your previous example mentioned three sites that are not comparable. Manhattan is not rural Alaska.” This is completely correct, and it further underlines the fundamental flaw in the concept. The ITE Tip Generation Manual 10th Edition Volume 1 Desk Reference states:
For the purpose of estimating vehicle trips in a suburban setting, these characteristics are typically sufficient to enable an accurate estimate. However, in an urban setting where there are opportunities for walk, bike, and transit trips, the number of vehicle trips may be affected.
There is increasing potential for walk trips as development densities increase in the proximity of the site and as activity at the nearby development complements a particular site. The pedestrian environment (e.g. sidewalk continuity, slow vehicle speeds, accessibility, protected street crossings) and bicyclist environment (e.g. bicycle lanes) enable and encourage walk and bike trips that otherwise would be made by vehicle or not at all.[4]
But again, let’s look at this in the context of the larger paradigm, and get back to the bigger question of what causes (or influences) travel. In this paragraph, ITE is explaining that different infrastructural contexts appear to result in different types of travel behavior. They are acknowledging that when you build more infrastructure to favor different modes of travel, people seem to use other modes of travel. They are stating that a major influence in travel might not be land use, but the context around it—so why wouldn’t we focus on that as the primary influential relationship? Infrastructural contexts such as roads, sidewalks, transit lines, and others are completely in control of public entities. This means that the biggest factor in what mode people use is essentially how much of this infrastructure governments provide or force others to provide. Traditionally, governments in the US have focused on infrastructure for driving. As UCLA’s Mike Manville stated so clearly in his 2017 article “Travel and the Built Environment: Time for Change”:
Governments give drivers free land; people as a result drive more than they otherwise would.
That’s it. The rest is commentary. [5]
That is it. Government decisions on what infrastructure gets built are the reason people drive so much, not land uses. While land use might matter slightly, infrastructural context matters more. The more of one context we build, the more people are likely to use it. To anticipate future motor vehicle trips based on land uses completely confuses cause and effect, and to require driving infrastructure based on anticipated trips only creates a self-fulfilling prophecy. Why is so much of planning still unable to see this clearly?
Finally, it should be added here that a few thousand feet of disconnected sidewalks, some painted sharrows, or a disconnected light rail line or two are not a significant change to a car-centered context—the existence of these things are a trifle in comparison to existing vehicle infrastructure, and they don’t make driving less convenient. If anyone is seriously interested in seeing less driving, alternatives need to be the most convenient option. But once we stop assuming that driving is intrinsic to land use and leave trip generation behind, that type of environment will be a lot easier to create naturally.
Moving Forward
People scratch their heads about increasing traffic, vehicle emissions, and why people keep driving and mass transit just can’t seem to compete. At the same time, planners, academics, and even APA seem unwilling to scrutinize what it is that really encourages people to travel. How can this be? The history of Ignaz Semmelweis in 19th century Vienna might offer some hints—we have spent so long focusing on the wrong things and the wrong relationships that we have blinded ourselves to what is right in front of us: we, as planners, play a fundamental role in facilitating how people travel.
We shouldn’t focus on how many trips might occur, but rather where we would like to see the trips that might occur. It will always be difficult for a municipality to predict when or how people will make decisions to travel, but it will not be difficult to manage those trips once the decisions are made. If we want more people to drive, let’s keep building more roads and parking lots. If we want people to move in a different way, let’s build something else. Whichever we decide, let’s stop clinging to the fiction that land uses can magically “generate” motor vehicle trips. That way, we can escape the framework within which new buildings always “cause” new traffic, and minimum parking requirements are considered a reasonable regulation. Let’s accept that it is germs that cause disease, and not the mal aria from swampy environments. Whether it is because we care about emissions, better built environments, more affordable housing, or simply a respect for the logical connection between cause and effect, it is time to discard the backwards idea of trip generation, and let our built environment begin to repair itself.
yorro is a planner working in municipal housing.
[1]From the 1968 ICMA “Green Book”: The purpose of a trip…can be looked at in two ways: from the viewpoint of the type of activity involved, or from the viewpoint of the person making the trip. For transportation planning, the most accurate description of the activity is land use, and the most accurate description of the person’s viewpoint is purpose. Land use is perhaps the more convenient measure of the two, because it enables travel to be related to a tangible, and relatively predictable quality. However, it must be thoroughly understood that people make trips, and that land use is only a convenient indirect measure of the type and geographic location of trips made. Today, this “convenient indirect measure” is often treated as a primary direct measure. Goodman, William I., and Eric C. Freund. "Principles and Practice of Urban Planning." (1968). 138 [2] Trip Generation Handbook, 3rd Edition. Institute of Transportation Engineers. September 2017. 5 [3] Trip Generation Handbook, 3rd Edition. Institute of Transportation Engineers. September 2017. 6 [4] Trip Generation Manual, 10th Edition, Volume 1 Desk Reference. Institute of Transportation Engineers. September 2017. 21 [5] Manville, Michael. "Travel and the Built Environment: Time for Change." Journal of the American Planning Association 83.1 (2017): 29-32.