Liberal Landscape

I’ve been in London twice this summer, in early July and then just last week.

In London, as in just about every other big city in the Western world, there has been a serious effort over the last fifty or sixty years to reduce the role of the automobile in urban transportation, and I wanted to see some of the results of this effort for myself. I’d been in London numerous times over the years but hadn’t been there (except to change planes) since 2013.

Some of what’s been accomplished has been widely publicized and requires little comment.

[1] Congestion charging. London has had a congestion charge on most vehicles entering central London since 2003. The charge is now a steep £15. Its effect is somewhat uncertain. There has been only a modest reduction in the amount of traffic in central London, and it’s not clear that this is due to the congestion charge.¹ Still, the congestion charge has brought some income to Transport for London (TfL) that’s been used for transit and for various projects that benefit pedestrians and cyclists. More recently, London has instituted an ultra-low emissions zone (the “ULEZ”) in its central city and has begun charging highly-polluting vehicles that enter the area.

[2] New bicycle infrastructure. Government agencies have created a substantial network of bicycle routes throughout the urban area. Some of these were labeled “bicycle superhighways” for a period. This label is no longer used, which is just as well, because the bike routes, impressive as they are, are not all protected and include numerous traffic lights and awkward street crossings; they are not much like superhighways at all. Some of them are quite heavily used nonetheless.²

[3] Constant improvement of rail transit. London, of course, had the world’s first subway lines (1863), and it supplemented the original just-below-the-surface lines from the 1860s and 1870s with new, deep “tube” lines in the 1890s and early in the 20th century. What is most striking to a New Yorker is the energy with which London (like Paris) went back to extending its subway system in the 1960s and continued to add lines in the following decades.³

One of the results is that London’s system, overtaken in length by New York’s in the 1920s, is now, with at least 402 route-kilometers, once again longer than New York’s system and is, in fact, probably the largest system in the Western world.⁴ Ridership before the Pandemic was rising faster than in New York too (although it remains much lower).

Here’s a map of the London Region, showing, among other things, cycleways and TfL rail routes:

And here’s a similar map for central London:

Pedestrian infrastructure improvement. Governments in London have also put a great deal of energy into an activity that’s received much less publicity: the improvement of the city’s pedestrian infrastructure including its recreational trail network. I acknowledge that, by its very nature, improving pedestrian infrastructure does not require the same kind of effort—or the same monetary resources—as some of the other activities noted above. It also may not have the same potential decarbonization consequences. The chief effect of building better pedestrian infrastructure is the creation of a more congenial environment for pedestrians. But there are presumably some decarbonization effects. Every trip made on foot rather than by a vehicle of any sort (including a train or bus) eliminates carbon use almost completely. And it’s a little hard to see how London’s transit systems could operate as they’re supposed to without good pedestrian access. In urban areas (like many in the United States) where walking is difficult, there are firm limits on how much decarbonization is even possible.

Improved pedestrian infrastructure has had the backing of London’s governments for the last several decades. The Walking action plan (2018) from the Mayor’s Office and Transport for London is the most recent of several official statements on the subject. Few other cities have taken pedestrian infrastructure quite so seriously (Singapore is one of them).

London, of course, was an excellent place for walking long before governments started talking about “pedestrian infrastructure.” London has been a walkable city at least since the streets were mostly paved and a proper sewer system was installed late in the 19th century. In most of London, sidewalks these days are in adequate shape and likely to be busy. Density is substantial enough so that walking is practical for many tasks. More often than not, drivers are reasonably respectful of pedestrians (there are exceptions; you’ve got to be careful about turning vehicles). And, in general, in London, there do not seem to be the kinds of cultural barriers to getting places on foot that many Americans face. Walking has been seen in a positive light in Britain among a substantial proportion of the population for many decades.⁵

The problem is that, at least since World War II, London, like most cities, has allowed cars to play a larger and larger role in urban transportation, and this has had all the usual effects. Noise and air pollution can reach astounding levels. Traffic deaths and injuries are common. Large parts of the outer city are so automobile-oriented that walking isn’t easy and doesn’t take you to where you might want to go. Even in the inner city, some arterials (Euston Road, for example⁶) are so choked by traffic that they repel those on foot and make access to certain areas a chore.

Government action to improve pedestrian infrastructure has taken several forms. Some of what’s been done has been easy and cheap. Installing pedestrian-oriented directional signage is tremendously useful and costs practically nothing. Changing traffic signals to reduce pedestrian wait times and to give pedestrians more seconds to cross is a no-brainer (but still needs doing in many places). Adding additional patches of sidewalk and closing streets, at least part-time, is also not difficult. This can engender opposition, but London is now full of newly pedestrianized spaces nonetheless, and there has been a very active discussion in public media about what more could be done in this area.⁷

London has also constructed new pedestrian facilities, some of which are only cheap when compared with the cost of building, say, subways. Among these are two (or maybe three) new pedestrian bridges across the Thames: the Millennium Footbridge (the “wobbly bridge”) between (roughly) the Tate Modern and St. Paul’s (2002) and the twin Golden Jubilee Bridges, which flank the Hungerford Railway Bridge between Charing Cross and Southbank Centre (also 2002).

London has also made an effort in recent years to maintain and improve its long-standing off-road pedestrian facilities on land. There are hundreds of parks in the city that contain walking paths. There are also a small number of longer-distance walking trails, most along water courses, that go back a long time. Examples include the towpaths that line the Regent’s and Grand Union Canals. (The towpath along the Grand Union Canal will take you most of the way to Birmingham.) These older off-road paths have been improved in various ways over the last few decades. Irregular surfaces have in some cases been smoothed or even paved, and the paths have acquired much better signage.

Completely (or mostly) new paths have been built as well. An example is the Lea Valley Walk, a path along the once highly polluted Lea River in East London that was radically upgraded as part of preparations for the 2012 Olympics.

There has also been an effort over the last twenty-five or so years to construct a continuous walking path along the Thames. The Thames Path (as it’s called) is an urban continuation of a path that extends for nearly 300 km from the river’s source to London. The rural portions of this trail (which I’ve never been on) are apparently like Britain’s many other long-distance footpaths, an elaborate concatenation of various rights-of-way that include narrow muddy segments, occasional busy roads, and just about everything in between. Since much of the Upper Thames, like the canals, was used for navigation before the age of steam, the Thames Path often has a towpath to follow. The Thames Path outside London was officially inaugurated as a national trail in 1996, and maps showing a finished route appeared soon afterward (there have been some minor route alterations in the years since).

Most of the Thames Path in London is of necessity quite different. The Thames is tidal below Teddington Lock in western London, and, as it heads to the Channel, it generally gets wider and wider. As a consequence, there never was a towpath along the lower parts of the Thames, east of Putney. In other words, there was often no obvious place for most of the urban continuation of the Thames Path to be sited. An additional complication is that a decision was made to build the Thames Path on both the North and the South Banks of the river between Teddington Lock and the Greenwich Foot Tunnel.⁸ (Upstream from Teddington Lock and downstream from the O2 more or less, only one bank has a path.) The Thames Path in London thus covers a substantial distance, something like 80 km. Including the paths along both banks, the Thames Path in London is approximately 130 km long. (It may be the world’s longest urban recreational path.) I walked very nearly its entire length during the course of my two trips to London this summer, and much of the rest of this post is based on this experience.

The only part of London’s Thames Path that was somewhat straightforward to build was that along the South Bank from Greater London’s western boundary down to the Putney Bridge, which generally just follows the right-of-way of the ancient towpath. Much of this part of the Thames Path has been left unpaved.

In the few places along the river where there was already parkland (for example, Battersea and Wandsworth Parks), park walkways have been declared to be part of the Thames Path.

There are also several stretches where the Thames Path alternates between having its own right-of-way and following lightly-used or pedestrianized roads through old medium-density suburban areas.

Much of the Thames in central London (especially on the North Bank) has long had major roads parallel to the river with a narrow stretch of sidewalk along the river embankment. In these areas, the sidewalk has been declared to be part of the Thames Path. This isn’t ideal, since users are always aware of traffic, but at least there’s often a busy bicycle path between the riverbank sidewalk and the road. The parts of the city where the Thames Path follows a sidewalk along the river are generally high-prestige residential or commercial neighborhoods (like Westminster, Pimlico, and Chelsea), and there are interesting buildings to look at across the highway.

Much of the Thames riverbank in London has, however, been the site well into the 20th century of docks, warehouses, and factories. This is especially true of the lower Thames, downstream from (roughly) the City. But, starting something like sixty years ago, with containerization and an enormous increase in the size of ocean-going ships, major port activities moved eastward, to points many kilometers beyond inner London. At roughly the same time, manufacturing activities in London fell into a radical decline. Land uses in cities don’t change overnight, however, and it’s taken many decades for this process to occur, and, in a few places, it’s still not complete. And there’s an additional complication. Over the last decade, the Thames Tideway Tunnel, a gigantic new sewer line that roughly follows the Thames, has been under construction. Work on this project has required substantial above-ground construction space, which blocks access to the river just as completely as factories and warehouses once did. The Thames Tideway Tunnel is scheduled to be complete, however, in two or three years.

There is now a different competitor for space along the Thames. In the same years that port activities were disappearing, middle- and upper-class residents were finding the idea of inner-city life more and more appealing, especially when it came with views of the cleaned-up river. There has been massive gentrification of much of inner London, including once industrial and/or working-class areas along the river. As in many other cities, this has led both to new construction and to the conversion of older buildings into expensive housing. It turns out that 19th-century warehouses make pretty good apartment buildings. Floors are strong, ceilings are high, and walls are often thick.

Most boroughs have been requiring new construction by the river to include space for the Thames Path. This hasn’t always worked, however, in part because new residents have sometimes been unenthusiastic about having a public right-of-way as a component of their expensive river views. A 2015 article in The Guardian suggested that real estate interests have not complied very well with rules requiring public access.⁹

There has nonetheless been a huge amount of progress. When I tried to follow the then newly declared Thames Path from central London to the still new Canary Wharf along the North Bank (and to came back via the Greenwich Tunnel and the South Bank) in 2001, very little of the path actually existed. It was necessary to use streets parallel to the river most of the way, and signage was not particularly helpful. These days, there are many more sections of continuous path, and the signage is a little better (although still imperfect).

While there are still plenty of places where pedestrians are directed to leave the riverfront, there are now many more areas where a newly built Thames Path passes along the river, often next to brand new developments, which occasionally come with ground-floor amenities.

This transformation is continuing. New segments of riverfront Thames Path are being added every year.

One of the difficulties of building the Thames Path along the river is that many “docks” along the Thames are actually man-made inlets rather than piers. Expensive bridging is thus needed to avoid substantial detours. Older buildings also often require analogous structures, built over the river.

Users of the Thames Path get to see views of a London that’s changed enormously. It’s arguable that the new developments along the Thames make up the largest urban renewal project anywhere in the Global North over the last few decades. It’s no longer correct to view residential London as made up almost entirely of low-rise, “terraced” houses. Many of the new apartment buildings in, for example, Chelsea, the Vauxhall-Nine Elms-Battersea area, and East London are very tall. The view of London from the Thames Path would hardly be recognizable as London to a traveler from, say, the 1980s.

Diversions from the river remain a problem, however. Some of these have a certain charm. There’s no great hardship in having to walk (or run) along a lightly-trafficked road lined with 19th-century warehouses.

Elsewhere, though, you have to walk through areas that only the most dedicated urban pedestrian would feel completely comfortable in on foot. Thames Path diversions in central Brentford, Mortlake, and Woolwich, for example, run on sidewalks right next to crowded arterials. Downstream from Greenwich, the Path often passes between active factories. And detours in Deptford take walkers through decidedly ungentrified neighborhoods.

The South Bank between (roughly) the Tower and Lambeth Bridges (with one gap) is a special case. In the years after World War II, the government began a long-term project to develop the riverfront of the not-so-prestigious South Bank as a cultural center. One consequence of the 1951 Festival of Britain was the construction of the Royal Festival Hall next to the Hungerford Bridge and the development of riverside public space that (radically) ignored the grid of streets. This new space proved enormously popular and was gradually expanded over the next decades. Tourist attractions like the Shakespeare Globe (1997) and the London Eye (2000) enhanced its appeal.¹⁰

One byproduct of the construction of new office buildings for the Greater London Authority headquarters near Potters Fields Park in 2002 was still more South Bank public space. This area has come to function to some extent the way that Times Square functions in New York: it’s a place for tourists to sit and simply enjoy being in a distinctive place. Of course, it’s not very likely that most of the crowds on this part of London’s South Bank realize that they’re on something called the Thames Path. (The government offices recently moved; rent on this high-use space had become too expensive.)

Walkers along the Thames Path will encounter even larger throngs across the river at the Tower of London. And there are more modest crowds in the central parts of Greenwich and (45 km away) Richmond. These business districts, of course, offer places to eat and drink. They also have bathrooms and easy access to public transit.

I acknowledge the risk of overstating the case, but I’m still inclined to argue that the construction of the Thames Path has fundamentally changed the pedestrian geography of London. It’s provided access to the Thames in many places where it simply didn’t exist during the years when most of London’s riverbanks were used for factories and port activities, and it’s provided a new facility for long- and medium-distance urban walking (or running) in parts of the city where no such facility existed. The outer parts of the Thames Path work for cycling too.

Although the Thames Path hasn’t made it yet onto most travelers’ bucket lists of things to see in London, it’s not exactly a secret. There’s an excellent TfL website that provides a detailed guide to the trail. There’s also a published guidebook on the Path’s London portions.¹¹ And the Path now appears on most tourist maps of London.

The Thames Path remains a work in progress, however. Except where the Path follows an old towpath, if you want to go more than a short way, you still have to keep shifting between more or less finished sections and temporary passages along streets, some much more attractive for pedestrians than others.

In this respect, the Thames Path is very much like comparable new facilities in many of the cities with which London competes—and in numerous smaller cities too. Examples include New York’s Hudson River Greenway, Tokyo’s Sumida River Terrace, Hong Kong’s harbourfront “promenades,” and Shanghai’s Huangpu Riverside Greenway. All of these waterfront paths were created partly to provide recreational space for local residents and partly to leverage an urban landscape made newly attractive by the disappearance of industry and docks and the construction of striking new residential and office buildings. These paths have mostly been imposed on complex areas with long-standing if somewhat moribund land uses. Thus, it’s not surprising that their installation has encountered roughly speaking a similar set of problems although to radically different extents. Remaining fragmentary port and industrial establishments often just couldn’t, for the moment, easily be moved. And, just about everywhere, there has been some opposition from local residents. As a result, all of these facilities have been under construction for many years and remain (to different extents) unfinished.¹² But they’re all now usable nonetheless.

I wouldn’t be inclined to argue that encouraging a shift away from automobile use is the chief reason for building new waterfront paths anywhere, but it’s certainly an appreciated byproduct. The addition of the Thames Path (and other new walkways) to London’s repertoire of pedestrian spaces can be seen as one of the results of the city’s turn over the last several decades toward focusing on the creation of alternatives to the automobile.

I spent a few days in Ottawa this month. I’d been in the Ottawa area several times over the years, most recently in 2015. Ottawa is not a huge, complicated metropolis in the way that Toronto, for example, is, but it has seemed to me an exceptionally agreeable place. The aesthetically pleasing and walkable gentrified neighborhoods just south of downtown—for example, Centretown and the Glebe—are particularly noteworthy. A comparison with U.S. urban areas like Raleigh and Oklahoma City that have similar populations (approximately 1.5 million) is revealing. The U.S. places are much more automobile-oriented.

In terms of the themes emphasized in this blog, Ottawa’s most distinctive feature is perhaps its large number of pedestrian and bicycle paths (sometimes collectively called the Capital Pathway). There are several hundred kilometers of these. It’s possible that, in proportion to its population, no substantial urban area in North America has a larger number of such features.

The paths are enormously varied. A few (for example, parts of the path along the Ontario side of the Ottawa River) take users far from roads. Others (for example, the path along the Rideau Canal south of the Queensway) are essentially sidewalks along busy highways. Some paths are wide, others uncomfortably narrow. Most are striped; a few are not. Most are for both pedestrians and cyclists, but some of the green lines on the map below represent protected bicycle lanes where pedestrians would be unwelcome.

Ottawa owes its extensive network of pedestrian and bicycle paths mostly to the efforts of the National Capital Commission (NCC) and its pre-1959 predecessor, the Federal District Commission (FDC), which, in the years after World War II, worked hard to improve Ottawa’s amenities.¹ The goal of the FDC and NCC was to transform Ottawa so that the city’s physical form would reflect its status as the capital city of a major country. The work of these commissions was somewhat handicapped by the fact that little money was put at their disposal. In addition, they had an awkward relationship with the Ottawa area’s city governments. But they did at least have federal backing and ended up accomplishing a great deal. They created parkland along Ottawa’s major watercourses: the Ottawa, Rideau, and Gatineau Rivers and the Rideau Canal. They also established a greenbelt around the older parts of the city of Ottawa and took steps to construct recreational spaces in numerous suburban areas as well. As more and more people took up walking, running, and bicycling in the 1970s, the NCC did what it could to improve the then still rather informal network of pathways by filling gaps and paving trails. There have been numerous additions and improvements in the years since, but the basic network was in the place by the early 1980s.²

Perhaps partly as a result of its pathways, Ottawa had the second highest proportion of workers (8.7%) who got to work by bicycle or on foot among large Canadian urban areas according to the 2016 census (Vancouver was first, with 9.1%). Among large U.S. urban areas, the highest comparable figure was 6.6%, in both New York and San Francisco.³

The open spaces reserved by the FDC and NCC came in handy when a decision was made to improve Ottawa’s transit system in the early 1980s. The Ottawa area by this time had become highly suburbanized and traffic-ridden, but there was still a substantial demand for public transport. The region didn’t seem to have enough people to justify an investment in rail transit. Instead, North America’s largest system of off-road busways (collectively called the Transitway) was created. There were, eventually, east-west corridors in the eastern and western suburbs and north-south corridors on both sides of downtown.⁴ Bus lanes, considered part of the Transitway system, were built in the CBD. I’d been on the Transitway on earlier trips to Ottawa and found it quite impressive. Unlike some other North American BRT lines, Ottawa’s Transitway (except for the downtown bus lanes) was almost entirely grade-separated, and the stops provided some protection from Ottawa’s often harsh weather.

Very likely in part as a result of the pretty good service on the Transitway, public transit has been used by a fairly large proportion of Ottawa’s commuters (18.3% in 2016 in the census metropolitan area). In the second half of the 2010s, among North American urban areas, only New York, Toronto, Montréal, Vancouver, and (barely) San Francisco (all much more populous—and probably denser—places) had a higher proportion of transit users.⁵

A very odd digression in Ottawa’s transit history was the addition of the (original) O-Train (now known as the Trillium Line) in 2001. This was an 8-km, mostly single-track “LRT” line between the southern suburbs and an area approximately two kilometers west of downtown in the middle of the NCC’s greenbelt, but at least on the Transitway. The O-Train’s rolling stock consisted of self-propelled diesel cars. Despite its peculiar geography, the Trillium Line did a respectable amount of business (20,000 passengers a day according to one source),⁶ and it could be said to have whetted the appetite of those in Ottawa who thought that only rail transit would do for the city. (The Trillium Line is now temporarily closed for a major renovation.)

Proponents of a more comprehensive rail line eventually got their way, arguing in part that the downtown bus lanes were not working very well. They were used by too many buses, at least in rush hour, and delays were common. Early in the 2010s, a decision was made to build a more conventional urban rail line, the Confederation Line (now sometimes called Line 1), running east-west.

This new 12.5-km O-Train opened in September 2019. It runs through downtown in a 3-km tunnel and mostly replaces the east-west Transitway elsewhere. The surface portions of the O-Train’s alignment still have a few sharp curves that were easier for buses to manage than they are for a train, but you would otherwise hardly know that the route had once been a busway. The old busway has been almost entirely obliterated, and the busway’s stations have all been replaced by much more elaborate structures.

Bus routes have been rethought completely. Many bus riders who once could ride all the way downtown now must change to the train. They’re able to do so fairly easily since there are no fare gates between the bus stops and the train line at the three main transfer stations. But there has still been some grumbling about the inconvenience of a change of vehicle.

The O-Train has had its share of well-publicized problems.⁸ A downtown cave-in during construction slowed work for months. And the Alstom Citadis rolling stock has not been very reliable. Heating hasn’t always worked; the doors have sometimes gotten stuck; and there have been derailments. But (I’m told) things have settled down over the last year or so.

I was quite impressed by the system, especially by the fact that there are trains every five minutes during the day and early evening.  This is more frequent service than is found on most North American heavy-rail lines. One could argue that the O-Train is providing better service than it has to, since the trains are usually not all that crowded, but it’s likely that Covid-19 is a factor here. The system is certainly in most ways state-of-the-art. It’s fully grade-separated (although it’s still called a “light rail” line by its builders). There are accurate countdown clocks in the elaborate stations (but no platform doors). The trains in use when I was there consisted of two very long cars (48 m each), each divided into four modules. They seemed quite comfortable to me. Announcements about coming stations (and mask policy) are provided in English and French, both orally and on digital signs. The large windows offer pleasant views, often of the pathways that run along much of the line.

There are now serious plans (with funding mostly in place) to extend the Confederation Line east and west and the Trillium Line south. The east-west extensions will mostly bypass the existing transitways. The eastern extension will run in part down the middle of a freeway, and the western extension will include a substantial tunnel. A branch of the southern extension will reach the Airport via an elevated segment. An additional extension to Gatineau, across the Ottawa River in Québec, has been discussed, but there are no firm plans actually to build it. A constant lament of some of the literature on Ottawa’s planning efforts has been the lack of cooperation across the provincial boundary.⁹

Even more than elsewhere, ridership has plummeted during the Pandemic. OC Transpo, Ottawa’s transit agency, argues credibly that ridership is down more in Ottawa than in other Canadian urban areas because remote work is more prevalent in an overwhelmingly white-collar city. Passenger loads are recovering. Overall ridership (including buses) reached 3.8 million in April 2022, up 73% from April 2021 totals but still less than half of pre-Pandemic levels. Ottawa’s O-Train is not, of course, the only radial public transit line suffering from diminished ridership these days.

Even if commuting—and rail ridership—remain lower than expected, Ottawa, especially after the extensions are finished, seems set to end up with an impressive rail line to go with its excellent pathway system. It will be able to provide much more non-automobile-oriented transportation than most North American cities.

High-density, pedestrian-oriented residential urban neighborhoods in the United States are rare. If one sets the criteria tightly enough—substantial population density and crowded sidewalks being the most important—the majority of such places are in New York, and even there largely in Manhattan and in certain parts of Brooklyn and the Bronx. There are only a few statistically comparable areas in smallish parts of a few other cities—especially San Francisco, Boston, Philadelphia, Washington, and Chicago—and even these are generally characterized by higher levels of automobile ownership than in New York.

In the great majority of existing high-density, pedestrian-oriented neighborhoods, many aspects of the built environment were created in the 19th century or earlier. The street patterns, the lot-covering building footprints, and arguably cultural expectations about how daily life is to be led often date back a long time.

It’s been argued that the association of high density with older places is basically a function of the fact that developers shifted to building more spread-out environments the minute that the advent of motorized transportation made it possible to do so. They did this because Americans preferred to live at lower densities. No doubt there is much truth in this idea. But it’s pretty clear that, even in the United States, there has often been a substantial minority of people who found the idea of living in compact urban spaces appealing and that this minority has been growing in size over, well, something like the last hundred years (admittedly with a bit of backsliding during certain periods). Interest in high-density urban living seems to have been growing at an especially fast pace over the last forty years or so, as cities have become cleaner and safer, as families have become smaller, and as many people have come to look less favorably on certain aspects of suburban living. This growing interest has run into a supply problem. There isn’t enough high-density housing to meet the demand. Thus, costs have skyrocketed in places that are felt to be desirable. Builders have responded to some extent by building where they could, but it’s not easy to construct high-density housing in the United States.  Zoning codes sometimes forbid it completely, or prohibit the mixing of commerce and residential structures, or insist that new buildings have parking spaces even where there is little demand for them. Then there is NIMBYism. These factors have often been most significant in cities and suburban areas where the majority of growth occurred in the 20th century. There has been serious resistance to adding density to these places.

Recent travels have reminded me that there are actually a number of admittedly relatively small areas where new, high-density neighborhoods have come into being in mostly low-density urban areas during the 20th and 21st centuries. I’m not talking about the inevitably tiny “new urbanist” projects that now dot American suburbs but of actual, complicated urban neighborhoods. The salient characteristic of these areas is that, unlike most American urban spaces of the last century and a quarter, they are to some degree pedestrian-oriented: their sidewalks are often crowded.

It isn’t an accident that many of these neighborhoods function in part as recreational areas. I can’t prove it, but I’m willing to hypothesize an “ideal-typical” history of these places. Because tourists even in the United States often don’t have access to cars when they visit large cities (and in any case may not be in any particular hurry when they’re on vacation), they’re willing to do more walking than when they’re at home. Over time, many visitors to such places find the bustle (along with the tourist attractions) appealing and decide to move in, at least temporarily, perhaps by acquiring a pied-à-terre or even a permanent housing unit in the community. Developers respond by building more residential structures. Because there are few long-term residents (at least at first), there is not the kind of NIMBYism that occurs in established communities, and much of the new housing ends up being high-rise. The fact that the spatial trope of the high-rise urban recreational community (especially along a waterfront) is a positive one even in the United States is a factor here. The result after perhaps several decades is a complicated, dense, walkable, and “vibrant” residential community, with a mix of condominiums, rental apartments, single-family houses, and hotels. There is, obviously, not quite the same mix in such places that you might find in the dense neighborhoods of older cities (in midtown Manhattan, for example). There isn’t likely to be much if any industry. In some cases, there will be few office buildings. In some, but not all, examples, poor people are priced out even more than in older cities, since there is unlikely to be rent control, and there certainly aren’t going to be public housing projects.¹ You do nonetheless have enough density to discourage automobile use—and this in American communities that date from the 20th and 21st centuries.

Consider two quite different examples: Waikiki in Honolulu and Brickell in Miami. I recently spent a week in Waikiki (my only trip there except for an earlier brief visit in 2000 on my way home from China), and I’ve been in Brickell several times over the years, most recently in February 2022. My comments below are based on census data, informal fieldwork, and a modest amount of reading.

Waikiki’s status as a resort may date back to the late 19th century, but it was only after the construction of the Ala Wai Canal on its western and northern edges in the 1920s helped drain the once swampy land that large-scale development began. Tall hotels along the beach began to be added in the 1950s, and, over the decades since, much (but not all) of Waikiki’s once low-rise building stock has been replaced by taller hotels and apartment buildings. The process has been gradual, but the result is quite an impressive skyline despite the fact that the Honolulu urban area has barely a million people.²

Of the tall buildings in Waikiki, the hotels tend to be located along the beach, the apartments (and especially the condos) along the Ala Wai Canal, to the north and west, but there are plenty of exceptions to this generalization.

Brickell has a very different history.³ Its distinctive character dates back to the 1960s and 1970s, as Miami came to take on its modern role as a Latin-America-oriented financial center. Miami’s old, mostly low-rise downtown seemed rather grubby to the sort of people who were making decisions about where to build new skyscrapers for banks, and, Brickell, a high-prestige residential neighborhood across the Miami River from the traditional downtown, was an obvious alternative. (CBDs inevitably move in the direction of wealth when there are no obstacles to doing so.) As elsewhere in North America, many people were increasingly attracted in (roughly) the 1980s to the idea of living in a high-density environment and were willing to pay extra to do so. Thus, developers aiming to build residential buildings began to find that they could profitably outbid developers of office buildings in Brickell (as well as in certain close-to-CBD areas). A factor more important in Miami than anywhere else was the desire of some well-off Latin Americans to own residential real estate in the United States; in no other urban area of the United States can one so easily get by in Spanish alone. The result was that tall apartment buildings began to be built in Brickell (and elsewhere around downtown Miami). Brickell now includes one of the most impressive high-rise residential clusters in the United States. Its tallest apartment building (868 feet/265 m) comes close to being a “super-tall,” and larger buildings are planned (heights until recently were limited by the FAA since Brickell lies close to one of the corridors used by planes landing at Miami International Airport).

Note that, while Brickell has never been a tourist center in the way that Waikiki is, plenty of tourists do come to the Miami area, and some of them have always chosen to stay in or near downtown Miami instead of in Miami Beach or another beach town, so Brickell came to acquire a number of hotels at roughly the same time as office and apartment buildings were being built. The hotels, of course, also house business visitors.

In 2020, Waikiki had a population of 20,470 in an area of 2.00 square kilometers. Its population density was thus 10,235 people per square kilometer (26,452 per square mile). This is high for the United States but way below, say, Manhattan’s population density of 28,873 per square kilometer. Consider, however, that hotels make up a substantial part of Waikiki’s area and that their guests would not have been counted in the census. Note in the map below how much lower in density the hotel district along the beach is, although building density is generally highest there.

Brickell had a population of 42,692 on 2.05 square kilometers in 2020. Its population density was therefore 20,835 per square kilometer, two-thirds of Manhattan’s. Note that, as in Waikiki (and of course Manhattan too), the people staying in hotels weren’t counted in the census figures, so the effective population density at any time is higher than indicated by the Census. Brickell’s still significant office function, of course, makes its high population density even more impressive.

It’s important to note that many apartments in both Waikiki and Brickell are pieds-à-terre only occupied part-time. In 2020, 6,833 of Waikiki’s 18,786 housing units—36.37%—were “vacant.” The Census tries very hard not to double-count, and it considers apartments whose inhabitants’ chief places of residence are elsewhere as “vacant.” Note on the map below how Waikiki’s unoccupied apartments (like its major hotels) tend to be concentrated along its southern, beachfront edge.⁴ In Brickell too (but to a lesser extent) some housing units—17.8%—are occupied only part-time. Their inhabitants’ permanent homes are elsewhere, presumably mostly either in northern United States or in Latin America. (Other parts of the downtown Miami area have an even higher proportion of such units.)⁵

In other words, it’s a safe assumption that, when Waikiki’s and Brickell’s hotels come close to being full and when a substantial number of its temporary inhabitants are using their apartments, the effective population density of these neighborhoods would be much higher than the official figures suggest. Their high density is of course the chief reason that these are bustling places. Everything is close together, and there is not much parking available. It would be ridiculous to try to move around within these areas by automobile, although I don’t doubt that some people do so.

Both of these neighborhoods, reflecting their urban areas’ population makeup, are ethnically diverse places; their diversity is more or less by definition one of the things that makes them seem so intensely urban. Waikiki’s permanent population, when classed into standard Census Bureau racial and ethnic categories, was more non-Hispanic Asian than anything else in 2020⁶; there were also a large number of non-Hispanic white people.

Hispanic                                         1479       (7.2%)
Non-Hispanic white                     7690     (37.6%)
Non-Hispanic Black                       526       (2.6%)
Non-Hispanic Asian                      7727     (37.7%)
Non-Hispanic Pacific Islander      641       (3.1%)

Brickell’s 2020 population, just like the Miami area’s, had a large proportion of people of Hispanic descent:

Hispanic                                          24106   (56.5%)
Non-Hispanic white                      13846   (32.4%)
Non-Hispanic Black                          950     (2.2%)
Non-Hispanic Asian                        1304      (3.1%)
Non-Hispanic Pacific Islander             3     (0.0%)

Note how, in both neighborhoods, the locally dominant “racial” category makes up the largest share of the population while non-Hispanic whites are second. The latter, compared to their proportion of the general population, are overrepresented in these generally prosperous neighborhoods, while minority groups that are on average poorer (Pacific Islanders in Waikiki and non-Hispanic Blacks in both neighborhoods) are underrepresented.

The small representation of some minority groups in Waikiki and Brickell reflects the fact that these are expensive neighborhoods. When you consider, however, that condo and rental prices in both Waikiki and Brickell are high, Waikiki’s permanent residents are (somewhat surprisingly) not uniformly well-off. Per capita income was $45,392 in 2014/2018 (as reported in the 2015/2019 American Community Survey). Income was highest along the beachfront where “vacancy” levels were highest and much lower (although hardly low) on its northern edge. I suspect that many of the poorer inhabitants of Waikiki are living in one of the several dozen lower-rise buildings with external corridors that still make up a substantial part of the neighborhood’s housing stock. It’s possible that many of these people work in the tourist industry.⁷

Per capita income in Brickell is consistently higher than in Waikiki (and in most other parts of Miami). But, because the census tract boundaries changed in 2020, it’s impossible to calculate an income figure for the area defined as Brickell on the maps. The per capita income of central Brickell in the 2014/2018 period (from the 2015/2019 American Community Survey) was $86,742; the per capita for what might be called Greater Brickell (including two tracts that spill over the borders of 2020 Brickell) was $68,931.

Car ownership among permanent residents in Waikiki and Brickell is not the 99% typical of American suburbs, but it’s still fairly high. In Waikiki, 16.6% of occupied housing units were carfree in 2015/2019. Still, plenty of residents did walk or take transit to work. (Honolulu has a pretty good bus system and, one of these years, the Honolulu Area Transit Authority may start operating trains to Ala Moana.⁸) Car ownership among permanent residents in Brickell is higher than in Waikiki but, again, not universal. Occupied units in the central tracts of Brickell were 9.0% carfree in 2015/2019. The figure for a slightly broader definition of Brickell, including two tracts that extend beyond the borders of Brickell as defined on the maps, was 12.5%. But here too many people did walk or take transit to work. Note though that more than 80% of the inhabitants of Brickell Key—the triangular island that is one of Brickell’s densest tracts—drove to work, and (somewhat unbelievably) no one used public transit (the recreational trail around the perimeter of Brickell Key is full of people during both day and evening, however). Their substantial level of car ownership is one of the ways that these neighborhoods differ most obviously from neighborhoods in Manhattan.

Both neighborhoods have an oddly unbalanced pattern of retailing. In Waikiki, there are numerous upscale stores (Prada, Gucci, and the like) on Kalakaua Avenue, and it almost seems as though there’s a tourist-oriented ABC Store selling sunglasses, bathing suits, packaged sandwiches, and the like on every block, but there are no full-service supermarkets. To find a supermarket you have to go to Ala Moana Center, said to be the world’s largest open-air shopping center, which lies just northwest of Waikiki, and what you get there is a branch of the upscale Foodland Farms. At least Ala Moana is easily reached by foot, bus, or car. Brickell has fewer street-level shops than Waikiki, but it’s arguable that, perhaps because tourism doesn’t color the retail sector quite so much, it has a less peculiar pattern of retailing than Waikiki but not by much. It’s easier to find an upscale restaurant or a gym on one of Brickell’s few streets zoned for commerce than a grocery store (and then you’d be stuck with a 7-Eleven). But there’s an enormous shopping center in the middle of Brickell (called, not surprisingly, Brickell City Centre) that sells all sorts of mostly expensive things. And there’s a large Publix supermarket just across the Metrorail tracks from southwestern Brickell and a Whole Foods just across the Miami River. I’d estimate that in the early evening maybe a quarter of the many pedestrians crossing the Brickell Avenue Bridge in a southerly direction are carrying Whole Foods bags. (My New Yorker’s instinct is that you can define a pedestrian-oriented neighborhood as one in which most people shlep groceries home by hand.)

Waikiki and Brickell are most certainly not the only places in Honolulu and Miami with newish big apartment buildings. It’s possible to argue that the success—along with the limited availability of spaces to build—in these neighborhoods has, in fact, led to the proliferation of high-rise apartment buildings elsewhere. Developers have constructed high-rise residential buildings all around All Moana Center, as well as in Kaka’ako, the neighborhood just to the northwest of Ala Moana, and even around Honolulu’s somewhat forlorn downtown.⁹ The view of these buildings from, say, Ala Moana Beach is quite impressive.

But I was struck in walking around these often huge buildings at how few pedestrians they generate. One project quite close to Ala Moana—Ward Village—has released a great deal of publicity about its walkability and has included some pedestrian facilities in its plans, but, so far (the development is far from finished), sidewalks in Ward Village are fairly (although not completely) empty. Could it be that carless tourists are absolutely necessary for high building density to generate high pedestrian density in American urban neighborhoods of the 20th and 21st centuries?

Maybe even more than in Waikiki, Brickell’s apparent success has encouraged the construction of high-rise apartment towers in other parts of Miami. It generally hasn’t been possible to build such structures just south(west) of Brickell, since the area is largely zoned for single-family housing (and its well-off inhabitants have mostly been uninterested in selling), but new and sometimes huge high-end apartment buildings have been built north of Brickell along Biscayne Bay from downtown almost to the city limits. Just as is true of high-rise buildings built in Honolulu outside Waikiki, I’ve been struck in walking among these enormous buildings by how few pedestrians they generate. Again, one can hypothesize that perhaps you need Brickell’s office workers and tourists to have enough pedestrians so that high-rise apartment dwellers feel comfortable about leaving their buildings on foot. There is also the fact that, while Brickell is a completely respectable place and not bordered by any dubious neighborhoods, numerous poor people frequent parts of downtown and live in neighborhoods of North Miami just inland from Biscayne Bay. I’m not justifying this, but I do wonder whether the presence of relatively poor people in these areas may make relatively well-off apartment dwellers a bit nervous about going out on foot.

Waikiki and Brickell are distinctive neighborhoods—that’s part of their appeal—but there are still some roughly comparable places.

There are, for example, several other high-density urban beachfront neighborhoods in the U.S. that welcome numerous tourists and that also serve as residential areas for substantial numbers of both permanent and temporary residents—and that are comfortable places for pedestrians. Obvious examples are the first few blocks in from the beach in Santa Monica and the southern half or so of Miami Beach (up through Mid-Beach).  Santa Monica beachfront neighborhoods may have more pedestrians than any other well-off part of the Los Angeles area, and Miami Beach is even more pedestrian-oriented than Brickell (it’s larger too). Note that neither beachfront Santa Monica nor Miami Beach is quite as high-rise as Waikiki. NIMBYism and zoning restrictions make it hard to build high in Santa Monica, and Miami Beach has (wisely) chosen to protect its art deco inheritance rather than to allow developers to replace it, but, despite the often modest height of buildings, density is pretty high by U.S. standards, and pedestrians in these places are numerous.

It’s a little harder to identify places completely comparable to Brickell, but all the many new high-rise apartment districts on the edge of the CBDs of even low-density U.S. cities are obvious candidates. The area north/northwest of downtown Denver including much of LoDo as well as Lower Highland is a good example. Residential structures in this area are mostly either completely new or else carved out of industrial buildings (the Highland end of the area is also the site of traditional gentrification of older residential structures). The area’s sidewalks are full of people, some of them presumably tourists (including many local tourists). There are also office workers in the area. There’s even a Whole Foods that attracts numerous walk-in customers. The substantial amount of new high-rise residential construction in the southwestern part of downtown Los Angeles also seems to be on its way to becoming a healthy, well-off, walkable neighborhood (it has many other functions too, and its pedestrians include a shocking number of homeless people). Belltown, northwest of downtown Seattle, is another close-to-CBD area that has acquired pedestrians as apartment buildings have been added. And, of course, all the newly residential, at least partly high-rise neighborhoods in once-industrial districts near the CBDs of Chicago (the West Loop), Boston (the Seaport area), and New York (Hudson Yards and Long Island City) are at least vaguely comparable too, although not as surprising as the new neighborhoods in once completely car-oriented places.

There is a general consensus that the United States could mitigate both its contribution to global warming and its lack of affordable-housing by building large amounts of high-density housing, especially in areas where housing is particularly expensive. There is also a widespread sense that the obstacles to moving in this direction are so overwhelming that not much has been or maybe ever could be accomplished. It looks to me as though, in fact, there are some places where there are relatively recent, successful, high-density neighborhoods. They’re not very numerous; they’re not very big; they’re expensive places to live; they tend to be rather distinct places, attractive both to tourists and to potential well-off residents; and some of them (for example, Waikiki) are more oriented to tourists than permanent residents might want. Because these neighborhoods are so special, it might be difficult to increase their size and number, and it must be admitted that the diffusion of high-rise apartment buildings away from the distinctive core neighborhoods hasn’t always resulted in places with large numbers of pedestrians. But it’s certainly worth remembering that relatively recent high-density, pedestrian-oriented places in the United States are at least possible.

I spent some time on Miami Beach Walk this week for the first time in a couple of years. It was interesting to see that the replacement of the old wooden boardwalk sections with pavers that I mentioned in an earlier post has now been completed.

I was especially delighted by the new signs forbidding motor vehicles on the path (see above). One could quibble with the English (what are “electric rides”?; the editor in me wants to substitute “vehicles” for “rides”). But the meaning of the signs is clear, and they appear to have been somewhat effective. In the course of a six-mile walk, I saw only three or four violators.

I know my thoughts on this subject will seem crabby to some, but it’s my sense that it’s quite dangerous to allow personal mobility devices with electric motors to be used in spaces that are supposed to be exclusively for people moving under their own power. Scooter and e-bicycle riders typically move at a higher speed than all but the fastest traditional cyclists, and riders of motorized unicycles and even many scooters can’t stop quickly. It’s also aesthetically distasteful when motorized vehicles enter pedestrian/bicycling spaces—especially crowded ones like Miami Beach Walk. When you’re in one of the few places in an American city where motor vehicles are banned, who wants to have to dodge scooters? They just degrade the environment. From my perhaps eccentric point-of-view, it would be wonderful if all the world’s recreational paths installed signs like those on Miami Beach Walk—and enforced them.

Note added 23 December 2022. I found myself on Miami Beach Walk this week and was, well, amused to discover that the authorities have been adding new signs with language that’s perhaps a bit less snappy than on the old signs but a little more likely to please a fussy editor. The prohibition on using non-human-powered transport modes is currently phrased: “No motorized means of transportation.”

Unfortunately, there seemed to be more violators than there were ten months ago. Scooter and e-bike riders were using the Walk with impunity, typically riding as fast as their vehicles would let them and endangering and annoying pedestrians and traditional cyclists. I acknowledge that my concern about this issue will seem uncalled-for to many.

I visited the new Underline trail when I was in Miami last week.

The Underline is supposed to replace and to be a big improvement over the M-Path, the simple trail that was created under or next to the southern branch of Metrorail when the train line was built in the early 1980s. The plan is that the new trail will separate pedestrians and cyclists and be surrounded by high-quality recreational spaces. It’s being paid for by both government agencies and private funds.

Only the first four-tenths of a mile (700 m) of the Underline trail has been completed. This segment lies entirely in Brickell, a high-density, generally upscale residential and commercial area just south of Miami’s traditional downtown. Brickell has a surprisingly large number of pedestrians on just about all of its sidewalks, and the trail here seems in some ways like a continuation of the surrounding bustling neighborhood. It’s often full of people, although I believe it’s reasonable to hypothesize that probably most users of the Underline’s initial segment are barely aware that they’re on a recreational trail.

One of the characteristics of the trail in Brickell is that the rail line is raised quite a bit above the street level in order not to impede boat traffic on the Miami River.

The enormous Metrorail columns provide plenty of room for huge distance markers and identifying and directional texts. Markers showing northbound distances, however, are aspirational. The finished Underline Trail ends just south of the point shown in the photograph below. Note that one of the dogwalkers in this photo is on the bicycle path. When I was there, no one seemed to be paying any attention at all to the distinction between walking and cycling paths.

South of Southwest 13th Street, the old M-Path has been left in place for approximately six-tenths of a mile (1 km). It runs under the tracks for a short distance and then alongside them. There is no separation of pedestrians and cyclists, but there really doesn’t need to be, since there are few users.

Between the first two stations south of Brickell—Vizcaya and Coconut Grove—the old M-Path is closed and is being turned into a continuation of the Underline (phase 2). An alternate route on local streets through low-density, more or less middle-class neighborhoods has been marked, but I saw very few people using it.

South of Coconut Grove, the old M-Path has for the moment been left in place. It continues to the end of the Metrorail line at downtown Dadeland (at the Dadeland South station), a distance of about six miles (10 km).

I walked this path when I was in Miami. It’s probably fair to say that, in its current state, the M-Path is not a very enticing trail. The trains overhead, despite being supported by concrete columns, are noisy, and (far worse) users are almost always right next (or very close) to South Dixie Highway (U.S. 1), a major arterial (with Ponce de Leon, another heavily-trafficked road, on the other side of the train line for much of the way).

There are numerous street crossings. In some cases, pedestrians and cyclists are aided only by a crosswalk; in others, there’s a traffic light to help, but users must sometimes wait a couple of minutes for a walk light to appear.

Since Miami-area drivers can’t be trusted to yield to pedestrians in crosswalks or when they’re making a turn, none of the street crossings seems very safe, except for the one where Snapper Creek Expressway joins the South Dixie Highway, where a bright red bridge has been installed.

The trail for the most part passes through what might be called an ordinary American suburban arterial commercial landscape. Automobile-oriented businesses and occasional residential structures line the parallel streets. Near certain stations, the trail consists of a narrow corridor through linear parking lots. The overhead Metrorail line is the trail’s chief distinguishing feature. There are also some new, tall apartment buildings in several places that are sometimes considered to be TOD projects.

For the moment, the M-Path seems to attract very few users. I saw perhaps five or six people during the couple of hours I was on the trail.

This remaining segment of the M-Path is supposed to be transformed into a continuation of the Underline trail in phase 3. It’s a little hard to see how the trail could be improved enough to make it a really attractive place, but perhaps I’m wrong. The literature on the Underline has alluring images of large groups of people walking, cycling, sitting, or playing in the midst of native vegetation, but of course no one has plans to get rid of the Metrorail tracks or the South Dixie Highway or all the level crossings.

Still, recreational trails in many car-oriented parts of the United States (for example, Dallas) attract a surprisingly large number of users. The city of Miami really has no other long-distance hiking/running/biking trail. The fact that the short paths near the junction of the Miami River and Biscayne Bay get quite a lot of use suggests that there is a market for a facility like the Underline. But the trails near downtown don’t border highways; they don’t have level street crossings; and they provide stunning views of Miami’s impressive new skyscrapers and of traffic on its busy waterways.

The views from most of the right-of-way of the projected Underline are of an ordinary American carscape. The designers of the phases 2 and 3 of the Underline have a difficult assignment.

Map revised 23 February 2022.

I made two short trips to Paris this fall.

I particularly wanted to take a look at some of the changes in Paris introduced by the administration of Mayor Anne Hidalgo, which, since 2014, has garnered a huge amount of publicity for its moves to reduce the role of the automobile in the city.

I ended up being quite impressed, although I was constantly aware that most of the recent developments in Paris are part of a long-term trend. Paris has been trying to tame the automobile since at least the 1980s.

The arguments favoring this effort are essentially the same in Paris as they are in the Western world’s other major cities. Automobiles cause vast amounts of air and noise pollution (there have been some periods when Paris has had the world’s worst air) and are responsible for a substantial number of deaths and injuries (there were more than 250 automobile-related deaths and 18,000 automobile-related injuries between January and November of 2021 in Île-de-France). Accommodating the automobile requires a huge amount of urban space (roughly 50% of the surface area of Paris), while in fact trips by automobile account for only a small percentage of all urban trips (approximately 13% in Paris, where 60% of all urban journeys are made on foot).¹ Most of the inhabitants of the majority of the largest Western cities (including Paris) do not even own an automobile. It makes no sense for urban planning to continue to focus above all on obliging automobile drivers. The end result of continuing this approach would be the destruction of the traditional city (and hardly anyone since Le Corbusier back in the 1920s has wanted to destroy the old city of Paris).

In this post I share observations not just from my most recent trips to Paris but from my many other visits there over the years. I also refer to some of the materials I’ve read. I acknowledge that a short essay by an occasional visitor on a huge and complicated place like Paris can only scratch the surface.

Recent developments fall, roughly, into three categories: [1] facilities for cyclists; [2] facilities for pedestrians; and [3] restrictions on driving.

[1] Facilities for cyclists. The Hidalgo administration has been boasting—accurately—of the facilities it’s built for cyclists. There are now said to be 1000 km of bicycle paths of one sort or another in Paris. They are, in fact, quite impressive.

Many of Paris’s new bicycle lanes are fully protected. Protected lanes between parking spaces and main roads are particularly common along some of the boulevards that were created in the 19th century under the direction of Baron Haussmann.

Where parking lanes are absent, protected lanes are typically much more modest, but there are usually still barriers of one sort or another between cyclists and automobile traffic.

On narrower streets, bicycle lanes have often been painted on sidewalks, or indicated by a different kind of pavement (in a style that’s common in the German-speaking countries):

There are also numerous contraflow lanes on narrow one-way streets.

And in one case—the Rue de Rivoli—half the roadway has been given over to bicycle traffic, and, for most of the day, private cars are no longer allowed on the lanes that are still open to automobiles.

Many of the city’s protected bicycle paths are shown on the map above.

But I wonder whether newspaper stories haven’t to some degree overstated the Hidalgo regime’s contribution to Paris’s bicycle facilities.

Paris has, in fact, been trying to improve its bicycling infrastructure for several decades. I am pretty sure that I saw my first painted bicycle lane on the Rue de Rivoli in 1969. During the 1990s, the city set up several bus-and-bicycle lanes; they turned out (predictably) to work awkwardly—buses and bicycles move at different speeds and have different stop-and-go patterns—but they were an attempt to make more room for bicycles. Early in the 21st century, Paris also began to set up protected lanes for bicycles.² I took this photo from what was then called the Promenade plantée in 2004.

An important component of bicycle infrastructure in Paris has been its bike-share system. Paris was probably the first really large city to set up such a service, the Vélib’ system in 2007. The Hidalgo administration’s contribution to Vélib’ has chiefly been allowing competition, and this change has caused a great many problems. The system is now run by a different company than it was originally and goes by a new name, Vélib’ Métropole, but it’s still functioning more or less as it has for fourteen years.

Let me add that, in general, I suspect that the role of the bicycle in contemporary Paris has sometimes been exaggerated. My sense in the course of my recent trips was that the city’s bicycle lanes are not particularly crowded. When I was taking photos of them, I often had to wait a long time for a cyclist to show up. Paris’s bicycle lanes may be a bit less crowded than New York’s. Despite some newspaper stories suggesting that the city was on its way to being “Copenhagenized,” cyclists in Paris, unlike those in Copenhagen or Amsterdam, usually make up only a small percentage of those moving on its streets.³ A story in the New York Times⁴ suggested that the tendency of French cyclists not to obey traffic lights has made Paris a dangerous place for pedestrians. It was a great story, but I’m not sure that it was particularly accurate. I walked approximately 200 km on my trips to Paris this fall, crossing hundreds of streets with bicycle lanes, and I was never in any way threatened by a lawless cyclist. The New York Times story’s source was perhaps a driver annoyed by some of the new restrictions on driving (of which more below). I don’t doubt, however, that cyclists in Paris (like those elsewhere in the world) pay less attention to traffic lights than they should.

It’s pretty clear that, even if Paris’s adoption of the bicycle can be exaggerated, the city has many more cyclists than it did a few years ago. It’s said that the number of cyclists in Paris rose 79% between 2019 and 2021. There are now supposed to be a million bicycle rides a day in Paris. Some of the rise is presumably due to the fact that Covid-19 has discouraged people from using public transit, but surely part of it is also due to the very real improvements in the city’s bicycle infrastructure.

[2] Facilities for pedestrians. The Hidalgo administration has also done a great deal to increase the space allotted to pedestrians in Central Paris. It’s eliminated traffic lanes and enlarged sidewalk space in several key locations, for example around the Place de la Nation. When I was there, the new pedestrian space hadn’t been raised above street level, but bollards had been installed so that cars couldn’t enter.

The Hidalgo administration has also established the Promenade sportive végétale Nation-Stalingrad.

This is a pedestrian path now approximately 4 km long through the linear middle-of-the-street park that was constructed in eastern Paris when Métro line 2 (underground here) was built in 1903. It runs from a block north of the Place de la Nation almost to the Jaurès Métro station (despite its name, it doesn’t quite get to the Place de la Bataille de Stalingrad). The park had not been well-maintained, and barriers in many places made walking along its full length difficult. Under the Hidalgo administration, the park was renovated, traffic lights for pedestrians were added when they were lacking, some athletic equipment was installed, and signage was added.

The areas along the path include high bourgeois districts, the Père Lachaise Cemetery, and much more modest, predominantly Maghrebi areas like Belleville; it takes users on a real tour of northeastern Paris’s social geography. And, while it can’t be claimed that this park is new, it has certainly been revitalized.

The Hidalgo administration has also presided over what might be termed the recategorization of central Paris’s splendid pedestrian and bicycling paths along the Seine that replaced two one-way freeways, one on each bank. I believe this transformation is one of the few instances in Europe where a freeway was turned into pedestrian space.

These paths originated in what turned out to be exceptionally popular Sunday closings of the freeways in the 1990s (long before Hidalgo became mayor). The approximately 4-km right-bank segment, once labeled Paris-Plage (even though swimming in the Seine, a bad idea, has never been allowed), was established when Bernard Delanoë was mayor in 2007. The original Paris-Plage was joined by an additional 2.5-km segment on the left bank of the Seine called the Promenade des Berges de la Seine in 2013, also while Bernard Delanoë was mayor. The two segments were joined administratively (although not geographically) and made permanent in 2017 under the name Parc Rives-de-Seine.

Somewhat confusingly, the Parc Rives-de-Seine, together with an improved path along the Bassin de la Villette in northeast Paris, are now collectively known as Paris-Plages (plural).

There has also been a less formal and still incomplete pedestrianization of the roads along the Canal St-Martin in northeastern Paris. The above-ground portions of this canal⁵ were bordered to a considerable extent by low-key industrial and warehousing facilities and very modest housing as late as the 1980s. The area became increasingly fashionable as it was gradually gentrified, and the banks were more or less cleaned up late in the 20th century.

The Hidalgo administration has encouraged a continuation of this process, making many of the streets along the canal zones of “pedestrian priority” (and closing them completely on some Sundays) so that the area has become an excellent pedestrian corridor.

There have also been some systematic street closings throughout the city. The Champs-Élysées, for example, has been closed to motorized traffic on the first Sunday of each month as well as on other occasions. (The photo at the beginning of this post was taken on the day of a road race that started on the Champs-Élysées.) There are also many street closings for weekend markets, many of which go back some decades.

Paris has constructed many other large-scale pedestrian facilities over the years, particularly since the 1990s. The extensive pedestrianization around the Centre Pompidou and the Forum des Halles has been in place for something like thirty years.

The Coulée verte René-Dumont, a 4.7-km rail trail for pedestrians between a spot just east of the Bastille Opéra through to the Bois de Vincennes, has been open for the most part nearly as long. It was established in the late 1980s, although it wasn’t fully open until the next decade. Originally called the Promenade plantée, the western half of this corridor (the Viaduc des Arts) runs high above street level.

The eastern portion runs in a culvert.

The western half of the Coulée verte appears to have been the model for the High Line in New York (and therefore of the 606 Trail in Chicago).⁶

The Hidalgo administration’s most widely-noted goal has been to turn Paris into a “fifteen-minute city” (ville du quart d’heure), in which a fifteen-minute walk or bike ride would provide access to essential services in all parts of the city. This is a wonderfully succinct way to describe a dense, pedestrian-oriented place, but, well, I’ve known Paris fairly well since the late 1960s, and it’s been a little hard for me to believe that Paris hasn’t come close to being a fifteen-minute city for all this time. For as long as I’ve been visiting the city (but probably much longer), there have been grocery stores every few blocks in nearly every residential district in the city, and schools and clinics are widely distributed too. Paris has been a dense place for several centuries and one of the best places in the world for urban walking, and it’s, of course, the city where flânerie was first identified as a distinct activity. I acknowledge that, by keeping the fifteen-minute city concept an important goal, the Hidalgo administration may have been helping to assure that rare gaps are filled. But it was continuing a pedestrian-first policy that goes back for many decades.

[3] Limitations on traffic. The Hidalgo administration has taken several additional steps to reduce the role of the automobile in the city, some of which aren’t in any obvious way based on the work of earlier administrations. For example, as of August 30, 2021, it imposed a speed limit of 30 kph in most of the city (the Boulevard périphérique and several major arterials are exempted). This seems like an important step to all those of us who’ve wondered why drivers were allowed to go so fast on the Haussmann-era boulevards. I was under the impression in walking around Paris this fall that most drivers were obeying the new speed limits (although, since average driving speed in Paris is said to be approximately 12 kph, the speed limit may not make a huge amount of difference). The speed limit is apparently popular with most Parisians.⁷

There are also plans to ban vehicles with diesel engines from the center of Paris from 2022 onwards, and to ban vehicles with gasoline engines from 2030.

The Hidalgo administration has also tried to tame scooters by limiting their speed in many places to 10 kph on November 15, 2021 (but, as in many other cities in recent years, scooters are allowed on bicycle paths, something that strikes me as rather unfortunate). And it’s also tried to limit street parking.

Statistical evidence⁸ supports the notion that automobile use in Paris has declined considerably, even before the Covid-19 Pandemic caused it to plunge still further. This is a real victory in the many-decades-long battle against automobile dominance in large Western cities.

[4] Beyond the city of Paris. Virtually all the steps described above concern the city of Paris alone and not the city’s suburbs. Paris these days has a population of something like 2.2 million. The Paris metropolitan area has a population of between ten and thirteen million depending on where its boundary is put. (Just as is true of American urban areas, the Paris region’s limits are impossible to define with certainty.) What’s clear is that, however you define the Paris area’s limits, the great majority of its population lives in the city’s suburbs.

Even more than is the case with American urban areas, Paris’s suburbs are often considered to be the place where the region’s major problems lie. Some of these are familiar ecological problems. Even though Paris’s suburbs are generally denser and have better public transportation than American suburbs, there are still enormous areas that can only be accessed efficiently by automobile. These days, no one thinks that this is a good idea. There is also the issue that a large proportion of the region’s poorer and/or immigrant population lives in the suburbs, sometimes in modest older housing and sometimes in one of the housing projects (HLMs, habitations à loyer modéré) that were built between the 1950s and 1980s. Zones with large populations of immigrants often have problems that are comparable to those in American ghettos (although there are fewer guns and murders). There is a feeling that the inhabitants of these areas are alienated in part because they find it difficult to participate fully in the modern economy. There are many reasons for this. One of them, it is often said, is that they don’t find it easy to move freely around the region.

But middle-class inhabitants of the suburbs have transportation problems too. Car dependence for many is as complete as it is for the majority of the inhabitants of American cities. Traffic jams and the need to travel vast distances make moving around Paris’s suburbs a tiresome and inefficient activity.

These problems have led to hugely ambitious plans to change the character of Paris’s suburban areas. These are subsumed under the label “Grand Paris” (“greater Paris”). Grand Paris, the subject of many books and a huge amount of short-form writing as well, involves many things, but perhaps the most important are attempts [1] to create stronger, less automobile-oriented nodes in selected parts of the Paris suburbs and [2] to join the nodes with a series of exceptionally fast, automatic Métro lines.⁹ The longest of these is a circumferential line two to seven kilometers beyond Paris’s city limits.¹⁰ Two semi-circumferential lines further out are also planned as are a branch line to Charles de Gaulle Airport and extensions of several existing Métro lines out to meet the Grand Paris Express.

The ambitious goal is not only to improve public transportation between Paris’s sprawling suburban areas and to improve life for the suburbs’ inhabitants but also to reduce the role of the automobile in the Paris region. Because the project more or less by definition lies entirely outside the city of Paris, the office of the mayor of Paris has had little to do with it.

Most of the work of creating Grand Paris lies in the future, but it’s the very near future. The nodes have been chosen, and construction of some of the new Métro lines has begun. And, as a kind of preview of what’s to come, the suburbs have been the scene of a massive infusion of tram lines over (roughly) the last twenty years, some of which are circumferential and some of which take passengers from Métro termini further into the suburbs. The only tram lines entirely in the city of Paris—lines 3a and 3b–follow the Boulevards des Maréchaux, a series of non-freeway arterials that run for the most part just inside the city limits.

Just about all aspects of Grand Paris have been the source of debate, and plans have changed a bewildering number of times. The number of government agencies that have been involved in the project is huge, and the cost of constructing everything that’s planned is enormous.

Cynics should remember that Paris has a history dating back several centuries of coming close to finishing its grand projects. Think, for example, of Baron Haussmann’s activities in the 19th century. The RER is another, near-contemporary example of a complex project that has mostly actually been built. Its key component is a group of fantastically expensive deep tunnels under the center of Paris that join suburban railroad lines. One of these tunnels (the RER A line) traverses nearly the entire city, east to west. The RER was planned in the 1960s, and its first line opened in stages between 1969 and 1977. That is to say, it was begun during a period sometimes characterized in French planning history as an era of “tout automobile,” when urban development in France chiefly involved figuring out ways to accommodate the automobile. The RER was developed despite the emphasis elsewhere in France on planning for cars, and it’s continued to grow—slowly!—in the decades since. The Grand Paris Express is in many ways a continuation of the same large-scale planning process out into the suburbs.

Whether the Grand Paris Express will accomplish its goal of reducing automobile usage remains to be seen. It’s pretty easy to be cynical. There really aren’t many (or perhaps any) cases where an automobile-oriented area has been transformed into one that’s genuinely pedestrian- and bicycle-friendly and significantly less automobile-dependent. But Paris may be in a better position to move in this direction than anyplace else in the world.

Here are four census-tract-level maps showing the “modal split” of journeys to work by workers 16 and over during the 2015/2019 period in the central parts of the New York, Chicago, Los Angeles, and San Francisco areas. All these maps use the same symbols, have a nominal scale of 1:100,000, and, if printed at 300 dpi, would occupy the same percentage of a 17 x 17 inch sheet of paper. (They will appear on a smaller scale on most computer screens unless you zoom in.) The maps use pie charts placed at tract center points to show modal split.

These maps are based on data from the U.S. Census Bureau’s American Community Survey (ACS).¹ The results of this survey for larger areas are probably quite reliable, but, at the census-tract level, the margins of error can be substantial, especially for smaller numbers. The narrower slivers on the pie charts, in other words, are less likely to be accurate than the thicker slivers. For most places, this means that the data on journeys to work by bicycle are less reliable than the data on journeys to work by car or (depending on the area) transit.

The numbers used to make the maps come from answers to the census question: “How did this person usually get to work last week?” Respondents were not allowed to check more than one box. It’s not clear how someone who biked to work on Monday, walked on Tuesday, took a bus on Wednesday, carpooled on Thursday, and drove him- or herself on Friday would have been most likely to answer the question.² Hopefully, things balance out. There really is no other national data set at the tract level that gives any kind of consistent information on choice of travel mode. And I don’t know of any national data at all on non-work trips.

Census tracts (for those unfamiliar with the term) are small, supposedly homogeneous areas devised by the Census Bureau. Certain large cities were first divided into census tracts early in the 20th century, and the entire country was “tracted” in 2000. The United States is now divided into nearly 74,000 census tracts. The average census tract has a population of something like 4,500, but in fact, if only because tract boundaries are changed reluctantly, census-tract population varies enormously, from zero to many tens of thousands in a few cases. For the 2015/2019 ACS, the middle 80% of the tracts varied in population from 2041 up to 7344. This seems like a big range, but it’s still possible to say that census tracts almost all have populations on the same order of magnitude. They most definitely do not, however, occupy the same amount of physical space. Census tracts in the denser regions of cities (for example, in much of Manhattan) are tiny; those in parts of Alaska are bigger than some states. The proximity of the pie charts on the maps below is a rough indicator of population density. There is a substantial cartographic problem here. The Lower East Side and the central Bronx, for example, end up being so crowded that it’s hard to tell what’s going on, while the San Gabriel Mountains and Berkeley Hills, at the other extreme, are so empty that it seems as though valuable map space is being wasted. The New York map, in other words, would be easier to read if the nominal scale had been something like 1:75,000; the other maps would mostly have looked better with nominal scales of approximately 1:150,000. Another problem I faced was choosing what to include. I went back and forth an embarrassing number of times trying to decide whether to include tract boundaries. The problem is that they disappear completely in much of the New York map, while they’re all too visible in the least populated areas.

The maps are clickable and downloadable, but the files are large, so that redrawing the maps can take a little while. Zooming in makes many of the details clearer; it also reveals minor faults in the data.

The message of the maps is the unsurprising one that large parts of New York and much smaller areas in Chicago and San Francisco have fundamentally different travel habits than does most of the rest of the United States. Transit use and walking are common modes for the journey to work, and automobile use is rare. This is true at all income levels.³

Another message of the maps is that walking is a more frequent way of getting to work than many would think. It’s most common around central business districts (even in Los Angeles) and near major residential universities, like Columbia, the University of Chicago, UC Berkeley, and even UCLA and USC. Bicycling to work occurs on a large scale around certain big universities too.

The maps also demonstrate that Los Angeles, despite its enormous investment in rail transit, remains quite different in its travel habits from large parts of New York, Chicago, and San Francisco. Except in the dense, socially complicated but (on the whole) relatively impoverished neighborhoods west of downtown, transit users in most of Los Angeles are uncommon, even along rail lines. Because the Los Angeles area has such a large number of people, the total number of transit riders is substantial, but they are a minority in most places. The same thing, of course, is true in suburban areas in New York, Chicago, and San Francisco, although suburban rail lines in all these cities do attract quite a number of users. I acknowledge that none of this will surprise anyone.

The 2015/2019 ACS is the latest to be released by the Census Bureau, but it already feels like historical data. Because of the Pandemic, many people have been working at home, at least part of the time, and transit use remains depressed everywhere. Bicycling advocates claim that there has been a huge rise in urban cycling and cite some scattered data to support their belief. They may be right, but an increase from, say, 2% to 4% doesn’t change the big picture very much.

No one knows whether the changes that have occurred during the Pandemic will last. We will have to wait until something like 2026 for the Census Bureau to produce ACS data for the Pandemic and post-Pandemic period.

Over the last thirty or so years, most of the urban areas of the Mountain West and Sunbelt have been taking some tentative steps to mitigate the less attractive aspects of their dependence on automobiles. They’ve built hiking and biking trails; they’ve encouraged and even subsidized downtown redevelopment; they’ve worked to create walkable neighborhoods; and they’ve added rail transit lines. The statistics suggest that none has succeeded to any great extent. The percentage of people using automobiles to commute to work has barely budged, and most downtowns are still not very healthy. But some urban areas have done better than others. It’s arguable that, among big cities, Denver has come a little closer to succeeding than its competitors.

Denver is in many ways typical of urban places in the country’s Mountain States and South. It’s big and it’s been growing quickly. The urban area has a population of something like three million, up approximately 15% from ten years earlier, and it sprawls enormously. While the boundaries of urban areas in the United States are always vague, continuous settlement in the Denver area runs at least sixty miles (100 km) north-south along the Front Range and at least thirty miles (50 km) east-west, from high up in the Rockies to far out in the Great Plains. Most travel in this huge region is by automobile. Very little of the Denver area would be considered “walkable” by any measure. And there’s a tremendous pollution problem.

As it happens, I’ve been in Denver at least every few years since the 1980s, and I went through a period in the early 1990s when I was there several times a year. I feel I know at least the central city moderately well and have a good sense of how the place has changed. I recently spent a few days in Denver and made an effort to take a close look at some of the ways that the Denver area has changed since I started visiting regularly.

From the perhaps peculiar point-of-view of this blog, Denver’s greatest claim to fame may be that it has such a complete network of off-street recreational paths. There are at least 250 miles (400 km) of such paths in the Denver area, and the major components of the system intersect with each other in a complicated enough way as to constitute a kind of network.

Most of the basic trail network dates from at least the 1980s. I say “at least” because the parks through which most of the trails run date back much further, in some cases to something like the 1890s (and perhaps beyond). These parks were established as Denverites discovered that the smallish streams flowing through the city out of the nearby Rockies could flood during spring thaws and after summer thunderstorms. Areas along the streams were gradually turned into parkland, and the streams themselves were tamed to some degree. Numerous dams were built, and, in some cases, streambeds were made less irregular and acquired concrete walls.

Paths through the linear parks along the streams naturally followed, sometimes as a result of work by government agencies, sometimes just because people liked to walk in the parks and eventually eroded pathways. In the 1980s, as more and more people took up bicycling, running, and walking, governments turned these paths into formal “bicycle trails,” and that’s what they’re still usually called, although they’re used by plenty of pedestrians. Since the 1980s these paths have slowly been improved in many small ways. They’ve mostly been paved. They’ve acquired stripes and mileage markers.

Gaps have been filled in. New side trails have been added. Bicycle and pedestrian traffic has been separated on the lower mile or so of Cherry Creek Trail, where there are paths on both sides of the creek.

And Cherry Creek Trail now ends at Confluence Park, which was built in part to celebrate the new millennium. Confluence Park incorporates the cleaned-up junction of Cherry Creek and the South Platte River and, with neighboring Commons Park, it provides impressive views of Denver’s renovated inner city.

I certainly wouldn’t claim that Denver’s bicycle trails are perfect. Because the trails follow streams that settlers avoided, they don’t always go where people might have found them most useful (but the two main trails, along Cherry Creek and the South Platte River, do both pass along the edge of downtown). Another problem: In a few places, busy freeways, also built away from settlements, abut the trails.

In addition, on pleasant weekend afternoons there can be so many cyclists that it’s hard to walk on the trails. Also, in at least one place a couple of miles from downtown, where the Denver Country Club insisted on controlling the banks of Cherry Creek, the Cherry Creek Trail is forced to divert onto a narrow sidewalk segment along a busy arterial. Of course, the sheer crowdedness of this narrow and unappealing segment is a sign of the trails’ success in attracting users.

Other cities have been building recreational paths too, but Denver, thanks to its geography, established them earlier than most cities, and it still has more of them for its size than just about any other U.S. urban area with the likely exception of Washington. Its competitors in the Mountain West and Sunbelt are far behind. Atlanta has struggled for decades to build its single long recreational trail, the Beltway. Dallas is only now trying to connect its scattered trails. Houston is just beginning to construct paths along its bayous. Austin’s excellent trail network is minute compared to Denver’s. New Orleans has had a hard time connecting its riverside levees near its central business district. And Phoenix has done little to join its dispersed trails, although it does have plans to do so.

Denver’s comparative success in building bicycle trails has increased the number of bicycle commuters only a little according to the Census Bureau’s 2015/2019 statistics on commuting to work by those aged 16 and over.¹ Only 2.2% of the city of Denver’s commuters got to work by bicycle while the comparable figure for the Denver-Aurora urban area was 0.9%. Several American cities (mostly but not all college towns) did better.² But, with the exception of New Orleans, most big Sunbelt and Mountain West cities and urban areas did much worse. Here are some city and urban area figures: Atlanta 1.1% and 0.2%; Dallas 0.2% and 0.1%; Houston 0.4% and 0.2%; New Orleans 3.1% and 1.4%; Phoenix 0.6% and 0.8%.

Denver has also been more successful than its competitors at reviving its downtown. In the early 1980s, Denver’s CBD, stretching something like 2 km (1.3 miles) between Union Station and the State Capitol, had all the usual problems of American downtowns. Its department stores were on their last legs. The inhabitants of close-in neighborhoods tended to be poor. There was a sense that downtown Denver wasn’t quite safe. It’s possible that the establishment of the 16th Street Mall in 1982 was a key catalyst for change. The sidewalks and the roadway along 16th Street were paved with ornamental tile. A system of free mall buses was instituted, and these buses have been the only vehicles allowed along most of 16th Street for virtually all the time since then. Even during the Pandemic, they’ve been running every couple of minutes. There may be no more frequent bus service in the United States.

There were periods as retailers closed and parking lots spread along 16th Street when the Mall seemed threatened, but restaurants have always done an adequate business. The addition of a nearby baseball stadium (1995) and several performance venues helped enormously. These days, despite the continued scarcity of office workers and tourists because of the Pandemic, downtown Denver is certainly the liveliest big-city downtown between Chicago and San Francisco.

A factor in the success of downtown Denver is that inner-city Denver has come to be seen as an attractive place to live. Capitol Hill, the neighborhood just to the south and east of the State Capitol, began to gentrify as early as the 1950s. In the 1980s it was a congenial, vaguely bohemian place, and, actually, it still is despite continued slow gentrification.

The next neighborhood out, Cheesman Park, east of Capitol Hill, seems actually to have never had any problems at all; it’s always been a prosperous place, and, with its apartment buildings (including a few high rises) a surprisingly urban one.

Although automobile ownership in these neighborhoods is high and the sidewalks are not exactly crowded, Capitol Hill and Cheesman Park are some of the only residential neighborhoods in the Mountain West and Sunbelt where I’ve had a considerable amount of company as I walked about.

A little more startling: the once rather gritty northern half of downtown Denver has also become a prestigious residential area, “LoDo” (or, more formally, Lower Downtown). A few old warehouses and industrial buildings were converted into hotels in the 1980s, some others (generally a bit later) into apartment buildings. New apartment buildings (some very tall) have been added to the mix in the years since. And, at the north end of 16th Street, three elegant pedestrian bridges over active railroad tracks, the South Platte River, and Interstate 25 stimulated the creation of additional housing in the early 21st century, partly built from scratch and partly as a result of renovations.

The formerly grim areas north of Union Station and between the South Platte and I-25 are now high-prestige, walkable residential areas. These areas are not only walkable; unlike some theoretically “walkable” areas in American cities, they actually have quite a few pedestrians. In how many other places in North America has a clever (and modestly priced) city planning project created a successful new urban place?

Gentrification has not been limited to the neighborhoods mentioned above. It’s also spread into Five Points and North Capitol Hill, where, I acknowledge, there has probably been a certain amount of slow displacement. It’s a little hard to see how this could have been avoided. Much further out, Central Park, a very large neighborhood that’s been built where the old Stapleton Airport was, has theoretically been planned on the basis of “new urbanist” principles, although it looks to me to consist almost entirely of single-family homes on wide streets, and its sidewalks are pretty empty. At least it does have sidewalks (a great deal of outer and suburban Denver doesn’t).

Most of Denver’s competitors have also tried to encourage the construction of housing in and near their central business districts, and some have had some success. Austin now has the tallest residential buildings west of the Mississippi; Atlanta’s downtown and Midtown have numerous new or newish middle-class apartment buildings too; and Dallas has tried to turn its downtown and several nearby neighborhoods into walkable places. But—and I acknowledge that this isn’t the only criterion that should be applied—none of these other cities has managed to create places that attract anything like as many pedestrians as Denver’s LoDo. And, with the spectacular exception of many parts of New Orleans, no other older gentrified neighborhood in the Mountain States or South has retained as much late 19th- and early 20th-century housing as Denver’s Capitol Hill.

Denver has also added more than its share of rail transit, and it’s done so with the help of locally raised taxes under the FasTracks program. Starting in 1994, it opened several light-rail lines. Electrified suburban rail lines were added starting in 2016; the line to the airport, which operates every fifteen minutes during the day, arguably provides better service than any suburban railroad in North America. The rail lines reach out along eight corridors from downtown; there’s also a line through central Aurora that avoids downtown completely.

The light-rail lines were mostly built in places where they were relatively cheap to build—along rail or freeway rights-of-way—and it’s arguable that, as a consequence, they don’t go where they would have been most useful (the inner city lost out most dramatically), but, as elsewhere in the United States, funding to build the lines was not very generous.  Denver’s lines did get built, and a couple of light-rail extensions and an additional railroad line or two are planned.

Like new rail lines elsewhere in the Mountain States and Sunbelt, Denver’s lines have not attracted as many riders as their builders had expected, but at least Denver has done a little better than most of its competitors. Pre-Pandemic, Denver’s light-rail lines had somewhat under 100,000 riders a day, its suburban lines just under 30,000, in other words, approximately as many light-rail riders as Dallas’s much longer system and far more suburban-rail riders than in Dallas. According to the Census Bureau, 4.4% of the Denver-Aurora urbanized area’s workers 16 and over took public transit to work in the 2015/2019 period. The comparable figure for the city of Denver was 7.6%. These numbers would seem pretty pitiful for any European urban area, and even for American cities in the Northeast and on the West Coast, but they’re actually high for the Mountain West and Sunbelt. Comparable figures for Dallas, for example, were 1.3 and 3.5; for Houston 2.0 and 3.8; for New Orleans 3.3 and 6.8; and for Phoenix 1.8 and 2.9 (but the city of Atlanta, with its heavy-rail lines, beat the city Denver; its figures were 2.8 and 10.0).

It’s impossible to be sure of why Denver has become a slightly less automobile-oriented place than its competitors, but it seems worthwhile to speculate. Size could have something to do with it. Denver is smaller (and perhaps more manageable?) than Atlanta, Dallas, Houston, and Phoenix, but it’s larger than Albuquerque, Austin, New Orleans, and Salt Lake. Denver also has the advantage of having been a reasonably big place earlier than most of its competitors (exception: New Orleans).³ Thus, it had a great many more preservable late 19th-/early 20th-century buildings in neighborhoods around downtown than its competitors. This probably helped make gentrification both easier and more gentle (Dallas’ attempt to build walkable neighborhoods involved bulldozing some of the old ones). In addition, Denver’s network of watercourses provided a relatively easy opportunity to build recreational trails. The fact that Denver rarely gets extraordinarily hot and/or humid in the summer may be an advantage too (but it can also get bitterly cold in the winter). Denver has on the whole probably attracted a larger proportion of highly educated, younger people than most of its competitors (but surely not more than Austin). These immigrants may have pushed governments to make positive decisions about building, for example, recreational paths and public transportation; governments in, say, the Houston or Phoenix regions have probably faced much less pressure in this area. Something else that has made Denver different is that Colorado’s government, unlike the state governments in most other Mountain West and Sunbelt states, has often been in the hands of Democrats, who have generally been more likely than Republicans to vote to contribute to non-automotive infrastructure projects.

Denver, like other urban areas of the Mountain West and Sunbelt, remains for the most part a sprawling, automobile-oriented, and energy-inefficient place. But it really has managed in some small ways to reduce some of the more disagreeable aspects of automobile dependence. As a result, you could probably live a little more comfortably without a car (or with very little car-use) in central Denver than in the central parts of, say, Atlanta, Dallas, Houston, or Phoenix. This isn’t, I acknowledge. saying much.

A map of the United States by census tract suggests that–except in a few remote and nearly roadless parts of Alaska–few households are carfree:

In fact, a closer look reveals a different story. In 351 (out of 74,002) tracts, 75 or more percent of occupied households were carfree according to the 2015/2019 American Community Survey.¹  Of these, only four were in Alaska. 312 were in New York, distributed as follows: Manhattan (New York County): 163; the Bronx: 68; Brooklyn (Kings County): 74; Queens: 6; and Staten Island (Richmond County): 1. Other U.S. cities had only a scattering of tracts where carfree levels were as high: San Francisco: 11; Boston: 3; Baltimore: 3; Philadelphia: 3; Washington: 1; Chicago: 1; Los Angeles: 1. High carfree levels in New York (and elsewhere as well) are highly correlated with density.² Here’s a map:

Other older cities in the United States all do have substantial areas where carfree levels are extraordinarily high by American standards, but none of these areas is quite as free of automobiles as large parts of Manhattan, Brooklyn, and the Bronx. There were 1309 tracts that were more than 50% but less than 75% carfree in 2015/2019. They were widely scattered across the United States. Several of the tracts were in Indian reservations; a few were in pockets of poverty in the South or contained substantial Amish settlements; and a smaller number were in Alaska (where tracts are huge). But most were in large cities. Here’s a list, except for New York essentially by county (cities are not indicated in the data):

New York: 785 [well over half the total!]
Cook County, Ill. (Chicago): 65
Philadelphia: 55
Washington, D.C.: 44
Suffolk County, Mass. (mostly Boston): 37
Baltimore (city): 27
Wayne County, Mich. (mostly Detroit): 19
San Francisco: 18
Cuyahoga County, Ohio (mostly Cleveland): 16
Allegheny County, Pa. (mostly Pittsburgh): 16
Hudson County, N.J. (mostly Jersey City and Hoboken): 16
Essex County, N.J. (mostly Newark): 15
San Juan, P.R.: 14
New Orleans: 9
Los Angeles County, Calif. (mostly Los Angeles): 9
Hamilton County, Ohio (mostly Cincinnati): 9
King County, Wash. (mostly Seattle): 8
Onondaga County, N.Y. (mostly Syracuse): 8
Westchester County, N.Y.: 7
Alameda County, Calif. (mostly Oakland):  3
Saint Louis (city): 2

Here are maps of parts of a few large cities, again showing the distribution of carfree households, on the same nominal scale and with the same class intervals as on the map of New York above.³ Note that, just as in New York, density seems to be the major factor in determining the geography of these areas:

There were a great many more tracts—4304—in the third category identified on the maps, those in which between 25 and 50% of households had no vehicle available. In 2015/2019 such tracts made up approximately 5.8% of American census tracts. Like the tracts where more than half the households were carfree, they were disproportionately located in America’s denser large cities.⁴ But even many smaller and less dense cities—including most cities of the Sunbelt—had such tracts.⁵ Since most U.S. cities (unlike, say, New York) do not have much in the way of dense, high-prestige neighborhoods, the majority of relatively carfree tracts in smaller cities and in those of the Sunbelt are located in less well-off areas, but population density still appears to be a critical factor in determining their geography. Dense, high-prestige areas like Oakland in Pittsburgh, the Central West End in Saint Louis, South Beach in Miami Beach, and the French Quarter in New Orleans are nearly as carfree as nearby less privileged neighborhoods.⁶

It almost goes without saying that an obvious explanation for the inability of the United States government to do anything much to change car dependence is the country’s high level of car ownership. There are, however, several million households in large cities in which a choice has been made not to acquire an automobile. New York has many more such households than any other urban area. It’s the one large place in the United States where only a minority of households have a vehicle available.

Note added 15 October 2021. I inserted the second paragraph from the end (including footnotes 4, 5, and 6) in response to some questions posed by a couple of readers.