Tuesday, March 31, 2015

Rant: industrial to residential conversion

Okay, so this will rate as a rant. When dealing with urbanists who defend contextual zoning, height limits and the like, once they are confronted with the impossibility of increasing urban population and keeping housing prices down without building higher, they will frequently come up with a "magical solution" to the issue. Converting old industrial neighborhoods to very high-density developments.

Griffintown, an old industrial center near the downtown area of Montréal, now being converted to very high density housing with 10 to 20 stories and a residential density of 40 000 people per square kilometer (over 100 000 per square mile)
The urbanists will then say that we don't need to raise height limits in established neighborhoods, we can keep contextual zoning alive, preserve the "heritage" and "character" of these areas and just concentrate all the construction in these old industrial neighborhoods. There, they have solved the Gordian knot of housing supply and conservation!

There is a big flaw to this, in fact, there are many big flaws.

For one, this is what we call in French "une fuite en avant", literally "flight to the front", meaning that when there are problems, instead of facing them and solving them, people try to flee in front of the problems, to stay ahead of them. In the end, they're just buying time, maybe hoping that someone else will solve the problems or that they will go away on their own.

What I mean by this is that, OK, you have a few disaffected industrial areas that you can repurpose for residential uses. That buys you off a few years, or maybe one or two decades if you're lucky... then what? The problem is still there, we still limit housing supply by an obsession to keep areas as is for the benefit of the lucky duckies who had the chance of getting there when it was affordable. The real solution should be to make sure that it is always possible to have areas densify and build more units, that way, the areas that are most desirable will see the greatest housing growth, and not just the areas that have less NIMBYs (or less effective ones).

But it's not just that, these industrial sectors are TERRIBLE places to live in many cases. They lack public services, lack shops, lack everything to support residential developments, because they were built for industrial uses. Everything has to be built from scratch, soil has to be decontaminated, transit lines have to be built, etc... The areas closest to established areas are a bit better off, but as you go deeper into the old industrial zone, it gets progressively worse.

It's not just that there is a lack of services... sometimes, there's the wrong kind of services. For example, look at this development:
New developments in Pointe-Saint-Charles

It doesn't look terrible, doesn't it? It is quite dense, has plenty of balconies and doors and they even planted trees in front, so that they will grow in the future. It even has a walk score of 91, a transit score of 100 and a bike score of 100. It is near a subway that can get to downtown in 15 minutes.
Walk score report for that locationé
Now, what if I told you that these units are pretty damn cheap, nearly suburban-level cheap, costing 250 000$ for 1000-square foot units? How could that be? Well, look at this other image...
Rail lines within 50 feet of apartments, and guess what?
They're EXTREMELY used by freight trains as there is a railyard a few kilometers to the east and one of the few rail bridges of Montréal
The people who live in the area are now protesting incessantly about noise, as train noises have been measured as high as 75 dB, and they may come and go at every hour of the day... or night. I don't like the freight train companies much, but it's not their fault, they have been in the area for a century, and at the time, a dynamic industrial sector sprouted up to benefit from the rail line. But now, the industrial area is gone and the city's urbanists have encouraged promoters to build housing within mere meters of these heavily-trafficked rail lines, in a bid to find a way to build more housing without modifying existing areas.

Getting rid of industrial areas and other job centers for residential zones also mean contributing to job sprawl. Some of these areas were disaffected of course, but maybe they could have been salvaged with ways to alleviate congestion, like with tolls and congestion charges.

So here ends my rant about how converting industrial sectors to residential sectors in order to avoid having to upzone desirable neighborhoods is, in my opinion, a wrongheaded policy, even a cowardly one.

Sunday, March 29, 2015

Urban growth boundaries, mobility and affordability

Urban growth boundaries are a favorite tool among many urbanists concerned with urban sprawl. After all, the basic definition of "sprawl" is that a city grows and occupies more and more space.
A schematic city in grey, with rivers and highways represented
Sprawl is usually defined as when the city's developed area increases, converting natural or agricultural land to residential, commercial or industrial uses
Some regulators looked at this phenomenon and thought that if the problem is that the city is growing, then that could be solved by a simple regulation: simply limit how far the city can grow.
A urban growth boundary is created around the city (the red line)...

...future developments can only occur within the red line, thus limiting the size of the city
This is also often called a "green belt", meaning an area of countryside imposed in regulations to prevent cities from sprawling into one another. The motives for such regulation are varied, most often, is it rooted in a desire to protect fertile lands from urbanization, as most major cities are located in fertile areas (an heritage from earlier cities). Agricultural land protection laws are also, in effect, a form of urban growth boundary.

The UK has allowed town and city planners to implement green belt legislation since the 1940s and the central government actually encourages cities to do just that. Ontario has also adopted more recently a green belt law to restrict how far cities can sprawl. Québec has had a law protecting agricultural lands since the 1970s, the Netherlands has had agricultural zoning since earlier than that.

Oxford in the UK applies a green belt to its development, hence the clear cut limit between urban and rural areas and the presence of undeveloped lands between urbanized neighborhoods

All the green in this map is zoned agricultural and protected from development
This kind of regulation can have a secondary effect wherein each small city is given some place to sprawl, resulting in leapfrogging developments as the central city of a metropolitan area maxes its sprawl allowance and thus developments max out allowances of farther and farther cities. This is what happens in Montréal, which is especially visible on the northern side:
The agricultural zoning of Montréal

The resulting pattern where far-flung developments grow quickly while closer suburbs are stagnant due to their area being zoned for agricultural uses only. In the metro Montréal region, about 60% of population growth occurs in farther suburbs (>20 km from CBD) or exurbs, even if there are undeveloped areas much closer.
There is some leeway in this kind of regulation, some jurisdictions will gradually expand the boundary to allow for gradual sprawl, some will enforce the regulation and only rarely touch the boundary.

Criticism of the urban growth boundary

Some people. notably Demographia, a think tank headed by Wendell Cox, strongly oppose urban growth boundaries on the basis that it makes housing less affordable. Indeed, urban growth boundaries limit greenfield development, which means creating a shortage of land on which developments can occur. As the supply of land is constrained, the value of land is likely to increase significantly, even if zoning mandates low-density developments. This makes developments that require a lot of land, like single-family housing with large front and back yards, much more expensive. For people who believe that this lifestyle should be made affordable to all at any cost, ie promoters of sprawl, that's bad.

Of course, there is a way around the problem (if regulations allow it), and that is to build developments with much higher density, and that often means taller. That way, the value of the land can be diluted among many units and on a greater amount of floor area. However, there are two problems with this:
  1. All else being equal, construction costs tend to increase with the number of floors that are built, so taller buildings cost more per square foot to build than shorter ones, and the cost per square foot of smaller units tends to be higher than that of larger units.
  2. Building higher density can mean infill developments in replacement of current housing, in which case the costs of buying and destroying current buildings are added to construction costs. Infill development is much more expensive than greenfield development.
So in the end, though the shortage of land can be compensated by higher densities, the price of such housing is still going to be much higher per square foot than in low-density, unrestricted greenfield developments. This will result either in higher housing prices or in smaller living spaces.

Critics of green belts therefore argue that they just create more income for speculators and rentiers and make living less affordable. They then call instead for more sprawl in order to keep housing prices down.

In return, defenders of green belts often argue that though house prices may be higher, this reflects the greater value of location as higher-density developments will tend to be closer to services, transit and jobs. This also allows residents to opt out of having cars, saving a lot on transport and so the cost of living can even be lower than in sprawl, as housing costs are only part of the picture of cost of living, transport costs are another important part of it.

That counter-argument assumes that space-limited developments will always evolve a good urban design, which isn't correct. You can certainly have dense areas with bad urban planning that results in high transport needs despite higher densities. For example, LA's metropolitan area is edged in by the coast, mountains and deserts, a form of natural urban growth boundary that forced LA to be relatively dense, with LA's metro being denser than New York's metro area, yet any claim of LA being more walkable and less car-dependent than New York would be rightfully mocked.

It is important to point out that too often, pro-urbanism people make the huge mistake of both supporting urban growth boundaries AND "contextual" zoning that limits the density of built areas to what is currently there to avoid any redevelopment that would change an area's "character". These two policies combined create a perfect storm for sky-high housing costs as you forbid any new supply of housing, which in a situation with increasing population, and so demand for housing, is going to lead to prices spiraling upwards. It is clearly a choice: either one supports urban growth boundaries or one supports contextual zoning. You can't have both at the same time, not if you want housing to stay affordable.

Mobility... the unspoken factor

What seems to be largely ignored by both sides is the issue of mobility. If people choose to live in suburbs rather than in the middle of nowhere, it's because they still want to access the jobs and services of the metropolitan area the suburb is part of. So suburbs are only possible if there exists a transport network that allows high enough travel speed to connect suburbs to the central city and other suburbs. In North America, this usually takes the form of highways..

Distance is measured in minutes, not miles. The faster people can travel, the vaster the area that is likely to be developed. Therefore, it is a delusion to believe that urban growth boundaries are only a result of regulation, they always exist either as a result of geographical features or as a result of the limitations of transport networks. That is why old medieval cities were always compact, because the transport speed was 5 km/h or so, ie walking speed, and this created a kind of limit to how much these cities could sprawl before the periphery was too far to be connected with the economic whole of the city. So sprawl is always limited in some way, it is not the result of laissez-faire, it is created by investments in transport infrastructure.

One consequence of this is that it must be understood that high-speed transport networks inherently subsidize sprawl, they make it possible for sprawl to spread so far, so these cheap lots available in sprawl actually bear a high cost in terms of transport infrastructure spending that the buyers of these lots rarely have to pay. It is very rare in North America that the people who use the infrastructure are asked to pay for it directly through fees. Largely, highways are funded through gas taxes and other taxes that are levied on everyone regardless of whether they use the highways or not. This forms a kind of subsidy to suburban housing prices and to suburban cost of living.

So sprawl does make housing cheaper, but it's largely a result of distortion of price signals as much of the infrastructure that supports it is financially supported by everyone, and not just those that use it most.

So in the end, urban growth boundaries are largely unnecessary, the best way to control sprawl is through the collective choices we make on the kind of transport infrastructure that will be built and how it will be funded.

A schizophrenic approach to urban planning: free high-speed mobility and green belts

I must point out the inherent stupidity of many North American urban centers, the schizophrenic nature of public policy on urban developments.

On one hand, you have DOTs and Transport Ministries that seek to enable fast, uncongested travel in metropolitan areas. They will build new roads and widen existing ones to make sure that people can keep traveling around quickly, increasing mobility. They will do so with tax money and not with fees, because the unspoken idea in North America is that people are entitled to highways, to fast roads to get them where they want to go, and all of that without having to pay for it directly. Try denying someone highways or asking highways to be tolled and they will cry out that they are being treated as "second class citizens" (I know, one of my old friends used that line on me when I told him I opposed highway construction to the suburb he recently moved to).

On the other, you have State/provincial governments or city governments adopting green belt regulations to restrict sprawl and prevent it.

Do you see the problem? We are creating sprawl with our transport policy and investments, with public money, but then we turn around and say "sprawl is bad" and try to limit it. We are doing one thing and its opposite, spending a lot of money on both, not realizing that we are largely fighting against ourselves and wasting billions of dollars doing it! Using regulation to fight something that we created by other regulations and government intervention.


So urban growth boundaries are a bad reaction to a problem that we create with our choices of transport policy and investments. We don't need them, what we need is to stop treating access to high-speed, toll-free roads as a human right and stop spending so much tax money on it. Make highway users bear the cost of highway construction and maintenance through tolls, impose congestion charges on vehicles accessing congested areas and do not build roads if they wouldn't be profitable, and you won't need urban growth boundaries as it will create incentives to avoid traveling high distances regularly. Of course, we also need to accept the fact that infill developments need to happen and that areas need to be able to change to accommodate higher demand.

Ultimately, yes, the result is higher housing costs per square foot. However, the cheap housing prices of sprawl is largely illusory, the result of hidden costs through higher transport costs for residents and higher infrastructure costs for governments. People should be free to live in the middle of nowhere if they want, to benefit from low land costs to have a big house, I am fine with that... as long as they don't ask the rest of society to pick up the tab and build the infrastructure required to allow them to live in that way without all the inconveniences of distance.

Saturday, March 14, 2015

Cities Skylines: A new tool for understanding urban and transport issues

So, all my free time is now gone since March 10. Why is that? Because Cities Skylines has been released. Not one week and I've already played 18 hours.

Cities Skylines is a city-builder in the vein of SimCity and Cities XL, but it stands out among the competition for having:

1- An agent-based engine that has every citizen in the city separately modeled, with a home, a job/school and some trips to leisure or commercial spots. Of course, the scale is quite off as a result, a 60 000 city in Skylines would probably have one or two million residents in real-life.
2- Transit options including bus, metro and train right out the door (not surprising considering this game is made by the makers of Cities in Motion 1 and 2, a transit-simulator game from which Cities Skylines is adapted)

The game uses zoning to shape the form of cities, with the player having to build public services like water mains, sewers and electricity generation. If you're at all interested in city simulation games, go get it, now.

More than just a game, Cities Skylines can help people understand urban and transport issues thanks to its basic characteristics. One of the main criticism leveled at SimCity was that it was a great game to build North American sprawl, and nothing else. Cities Skylines allows for the building of non-car-dependent cities.

It's not perfect, it doesn't have mixed used zoning, it doesn't have trams (present in Cities in Motion 2), it doesn't have bikes, it doesn't have transit-only lanes and the road builder is limited (you can't synchronize traffic lights, can't tell the game to have two-lane turning except in very specific situations, etc...).

Anyway, I've built two cities up to now, that I may call Transit City and Sprawlville.

Transit City: a bit of streetcar suburbs and of Japanese train station towns

The first city I built was intended from the get-go to be a dense, urban city, built around transit. To do so, I sought inspiration from streetcar suburbs and Japanese TOD towns built around train station.
Basic design of a streetcar suburb, grid of arterial (black lines), commercial/office is in blue, residential is in green, industrial is in yellow
Train station town, inspired by examples in Japan
The game also allows to zone for high-density housing or low-density housing, and low-density housing is needed to attract families. So I zoned the residential areas closest to commercial areas as high-density and kept low-density residential areas a bit farther out.

Here are some images of the city:
This view from the sky shows the different zones, unfortunately, deep green high-density residential and deep blue high-density commercial are hard to tell apart. The blue-green zones are offices. You can see the streetcar-style design still.
One of the difficulties of the game is managing traffic, but as you can see in this image down the main arterials, traffic is quite under control (industrial areas are another thing)
These are two recent developments I've built around train stations outside the main city, who take after Japanese suburban cities built around train stations:
Garden City 1: train station is circled in red

Garden City 2: again, the train station is circled in red

Sprawlville: low-density, use separation, how developments occur nowadays

The second city I built was as a challenge. The city would be typical sprawl with extreme use separation. The main issue with something like this is traffic, how to manage it. The basic design here is to build each neighborhood in cul-de-sac, connecting to huge arterial streets and highways. Each neighborhood has a different use zoned for it, and connections between them outside of arterials are rare if not non-existent. That way, I can control traffic coming out of neighborhoods, if it's too much, I can remove some of it, and build another neighborhood farther away. In order to maintain fluidity on the arterials, there must be as few connections as possible.
Basic Sprawlville design, self-contained single-use neighborhoods connected to arterials
Here is an example of this kind of design, to show that I'm not just "strawmanning" sprawl.
Sprawl near Austin, TX. See how the main road a bit to the right of center on the image has no development on it, but has residential and industrial neighborhoods connected to it
A few down the arterial street on the image above to prove that there is no development alongside it, or very, very few
 Now, let's see what it looks like from up high:
Sprawlville from the air, showing the different zones and the massive roads that connect them

A zoom on some of the neighborhoods


Both of these cities are thriving according to the game. The Sprawlville design "works" in the game, as it does in reality. This design helps to dilute car traffic and to easily expand cities. Need more housing? Just build a new residential neighborhood. More industries? Open up a new industrial "park" somewhere. Let people find their way with cars between the neighborhoods.

The residents are "happy" in the game, in that they live in areas with no noise or air pollution. The game doesn't care about the health effects of such a design and the human and social consequences of it on individuals and communities. The point is: it works to absorb more residents and more jobs. Which is the entire point of the city-building game: to grow a city.

I've managed to grow the Transit City design also by planning the street grid and by building train stations in undeveloped areas to build "garden cities" around them. But this goes against everything we're used to doing in North America. Imagine, spending millions on transit lines to "nowhere"! Except that this "nowhere" soon becomes a thriving city thanks to an influx of people, shops and industries, and a city connected to an existing city with a high-capacity, high-speed transit line from the get-go.

Still, the game provides good lessons on this issue. Cities are growing all across the world, we have to find how to accommodate the population and the jobs inside the urbanized areas. But with the habits of "Development-Oriented Transit", of only building transit when density requires it, it is impossible to allow a city to grow around transit. Sure, you can densify older areas, but at one point, you reach a point when you no longer can, you need to build new areas outside the old city.

Expanding a city around cars and highways is easy. Sprawlville is an easy model to have, especially as you already have highways built, connecting regions together. Which provides a backbone for future developments. It also avoids having to review older neighborhoods. Once built, you leave them alone and move somewhere else. For urban planners, Sprawlville is exceedingly easy to plan, Transit City requires facing down opposition to densification or to massive public investment in transit.

Finally, an interesting point. If you don't know about Strong Towns, you're missing out. It's an American organization opposing sprawl on the grounds that sprawl is financially unsustainable and exceedingly wasteful. The organization is coming from a right-wing, fiscally conservative perspective, not really my cup of tea as a left-wing social-democrat, but let it not be said I'm not ready to listen to the other side.

Anyways, since the game has you planning and funding public services, it provides a demonstration of Strong Towns' main point about financial sustainability.

Look at the budgets for the two cities:
Transit City budget (60 000 population)
Sprawlville budget (35 000 population)
Transit City has an expenses budget of 167 000 C-dollars per game week, including nearly 25 000 in transit services and 44 000 in various optional policies (like free smoke detectors and subsidies).

Sprawlville has an expenses budget of 100 000 C-dollars per game week, with no transit nor optional policy in place.

By person, Transit City actually has 6% less expenses than Sprawlville, but if you take out the optional policies, Transit City residents actually incur 26% less expenses than Sprawlville residents... and if you remove transit expenses (which actually save residents money as they would need to buy cars otherwise, a factor not accounted for in the game), then Transit City's mode of development is 41% less expensive than Sprawlville's mode of development in terms of maintenance of public infrastructure AND less public services.

Of course, it's just a game, but a game based on micro models to simulate reality. This game demonstrates how the traditional mode of development, one based on density and public transit, is much more efficient than the sprawl. However, it also demonstrates why sprawl is winning in North America: sprawl is easy to plan for and makes it easier to grow cities without running in political conflicts. In other words, lazy urban planners, not economics, explain why sprawl is so dominant.