Descend from 40,000 feet into just about any major metropolitan airport in the United States, and patterns of the trajectory of American life over the last century become clearly visible. Old urban cores are etched out in tight grids modeled off a sheet of graph paper. Further out, all those neat lines and right angles begin their curling meander into suburbia. Sparsely populated roads loop through the countryside in an odd geometry designed around the residential real estate dream of post-war America: a cul-de-sac for every family.
“I think it’s a missed opportunity when I’m flying and I can’t look out the window and see what the patterns are,” says Norman Garrick, an associate professor of civil and environmental engineering at the University of Connecticut.
This is where it’s most apparent – from an airplane window – that American ideas about how to live and build communities have changed dramatically over time. For decades, families fled the dense urban grid for newer types of neighborhoods that felt safer, more private, even pastoral. Through their research, Garrick and colleague Wesley Marshall are now making the argument that we got it all wrong: We’ve really been designing communities that make us drive more, make us less safe, keep us disconnected from one another, and that may even make us less healthy.
“What I understand now is that the patterns of places matter enormously,” Garrick says. “Even from 40,000 feet, you can tell the difference between places. It’s not going to give you all the answers, but it’s going to tell you an awful lot about how people live in different places, just by looking at these patterns.”
• • • • •
Most of the oldest cities in America – not to mention the oldest capitals in Europe, or in the Roman Empire, for that matter – were laid out in neat, densely interconnected grids that enabled people to get around before cars came along. Manhattan looks like this. So does Savannah, New Haven and Washington, D.C.
These communities had what Scott Bernstein, president of the Center for Neighborhood Technology, calls “location efficiency,” a rough analogue to the idea of energy efficiency that captures the extent to which your job, your grocery store and your favorite pub are all convenient to you. Around the turn of the century, U.S. cities of all sizes built thousands of miles of railway for streetcars that made the urban grid even more efficient.
“It happened everywhere, it happened brilliantly,” Bernstein says, “and we threw it away.”
Americans lost sight of that tightly knit model when we got into cars and began to envision something else: the Garden City. In the early 20th century, modernists decried overcrowded cities that were synonymous with pollution, slums, and poverty. They wanted to do away with unnecessary streets and give each factory worker and company man his own slice of the country. He would drive there, of course, first along a large arterial highway, then down a main thoroughfare, then a collector road, then a local street pulling into his own private driveway at the end of a cul-de-sac.
There would be no such thing as noisy and menacing through traffic because no one else had any reason to drive through. The traditional street grid, with its busy intersections and jumble of apartments, shops and restaurants, was out.
“It was addressing real problems, but it went overboard,” Garrick says of the suburban model. “It took real problems and then made caricatures for solutions.”
The Federal Housing Authority embraced the cul-de-sac and published technical bulletins in the 1930s that painted the urban street grid as monotonous, unsafe, and characterless. Government pamphlets literally showed illustrations of the two neighborhood designs with the words “bad” and “good” printed alongside them.
The FHA had a hand in developing tens of millions of new properties and mortgages, and its idiosyncratic design preferences evolved into regulation. From the 1950s until the late 1980s, there were almost no new housing developments in the U.S. built on a simple grid.
“You hear the idea that a lot of it was just the free market, that’s what people wanted at the time,” says Marshall, an assistant professor of civil engineering at the University of Colorado.
“At the same time, a lot of it was that we were starting to require these types of places along the way. It wasn’t just that people wanted to live in these types of communities. It was hard for a developer to come in and build anything different from what had been done.”
The FHA never put it quite this way, but what we were really doing was building communities for cars, not people. Earlier neighborhoods were literally built on a scale for the human body, with architectural embellishments at eye level and blocks and sidewalks designed for foot travel. The human measuring stick hasn’t changed much over the last 200 years, and so, in theory, that model should still apply.
“That is the fundamental connection between looking back toward older methods of design,” Garrick says. “We need to remember when we’re designing that we’re designing for humans, not for objects, and not for the movement of these objects. It’s about human beings, about humans being able to get from one place to the other.”
• • • • •
Garrick and Marshall’s research into street network patterns began in Davis, California. Often cited as the most bike-friendly city in America, it has the country’s highest rate – more than 16 percent – of people commuting to work on two wheels. It turns out, though, that Davis also has one of the lowest traffic fatality rates in the country, a counterintuitive discovery for traffic engineers who consider biking a riskier alternative to driving.
Inspired by Davis, Garrick and Marshall compiled data on 230,000 crashes spanning 11 years in 24 medium-sized California cities. And they began to parse and classify street patterns in a kind of taxonomy. There are networks that look like square grids and others that resemble trees, with one trunk, many branches. There are networks that have tributaries, like a river, and others that have main roads radiating out from a central hub. There are hybrids of all these, and street blocks of different lengths, and networks that have 45 intersections per square mile (like Salt Lake City) and others that have as many as 550 (Portland, Ore).
The historical trend, through, has generally moved in one direction – toward ever more creative works of geometric art.
In their California study, Garrick and Marshall eventually realized the safest cities had an element in common: They were all incorporated before 1930. Something about the way they were designed made them safer. The key wasn’t necessarily that large numbers of bikers produced safer cities, but that the design elements of cities that encouraged people to bike in places like Davis were the same ones that were yielding fewer traffic fatalities.
These cities were built the old way: along those monotonous grids. In general, they didn’t have fewer accidents overall, but they had far fewer deadly ones. Marshall and Garrick figured that cars (and cars with bikes) must be colliding at lower speeds on these types of street networks. At first glance such tightly interconnected communities might appear more dangerous, with cars traveling from all directions and constantly intersecting with each other. But what if such patterns actually force people to drive slower and pay more attention?
“A lot of people feel that they want to live in a cul-de-sac, they feel like it’s a safer place to be,” Marshall says. “The reality is yes, you’re safer – if you never leave your cul-de-sac. But if you actually move around town like a normal person, your town as a whole is much more dangerous.”
This is the opposite of what traffic engineers (and home buyers) have thought for decades. And it’s just the beginning of what we’re now starting to understand about the relative advantages of going back to the way we designed communities a century ago.
Marshall and Garrick took the same group of California cities and also examined all their minutely classified street networks for the amount of driving associated with them. On average, they found, people who live in more sparse, tree-like communities drive about 18 percent more than people who live in dense grids. And that’s a conservative calculation.
This undoubtedly has to do with the fact that the grocery store, your house and your office are probably farther apart – and with less direct connections between them – if you live in a subdivision. But the difference may also have to do with the fact that people who live in gridded urban networks get in their cars less in the first place. They’re able, instead, to walk, bike, or ride transit. And if they’re doing that, Marshall wonders, does that mean people who live in grids are also healthier? His next study will examine that question.
What is harder to measure is the value of simply being connected – to where we want to go, but also to each other. Bernstein’s location efficiency data speaks to some of this. He’s even found that foreclosure hotspots tend to be focused in places with the least location efficiency – in spread-out subdivisions, where a family already stretched to the limit can go broke driving 10 miles each way for a gallon of milk.
“You make a terrible mistake if you plan a city in terms of buildings and facilities and parks,” Bernstein says, “and don’t look at the space that those things occupy.”
In other words, at the underlying patterns: the way streets and people connect with each other. Now, what intuitively made sense to us a hundred years ago can be justified and measured in foreclosure rates, vehicle miles traveled, and traffic fatalities.
“It’s ironic,” Garrick says, “but the thing is the patterns that we used to use in American cities are patterns that were built over thousands of years. And there’s a reason they were built that way.”