What City Observatory did this week

1.Measuring Walkability:  Non-car modes of transportation have always been at a disadvantage in policy discussions because of a profound lack of widely available quantitative measures of walkability, and because all of the metrics developed to guide transportation focused on moving cars. That’s begun to change with the advent of Walk Score, which provides a ubiquitous, easy to understand and free ranking of the walkability of every address in the US.  Some have criticized Walk Score’s destination based measures as less-than-perfect, but they provide an invaluable baseline and starting point for measuring—and improving walkability.

2.  Our civic commons infographic.  We’ve summarized some of the key findings from our report Less in Common in the form of a free, downloadable infographic.  We show how a combination of factors have led Americans to live lives more separated from one another, measured by everything from where we live to how frequently we socialize to how we travel.

3.  Polarity Shift:  A key leading indicator of entrepreneurial dynamism and future economic growth is venture capital investment. For decades, Silicon Valley has been the unquestioned leader in VC investment. But in the past five years, its been surpassed by its neighbor to the north, San Francisco, which now accounts for three times as much venture capital investment as does Silicon Valley. This big shift has a lot to do with the growing demand for city living.

4.  Playing Telephone.  There’s no question that on-line shopping is growing fast and driving big changes in the retail environment, as chain store bankruptcies and closed malls attest. But recently repeated claims that Americans now make more purchases on-line than in-person are a wild exaggeration. We show how a classic game of “telephone” translated a narrow research finding into a eye-catching, but simply erroneous claim. For the record, on line purchases accounted for about 8.1 percent of retail sales earlier this year.

This week’s must reads

1.  Two stories on gentrification from the UK.  The term gentrification was coined in the UK a half century ago, so its appropriate that it be hotly discussed there.  Writing in the Guardian, Dave Hill argues that its time to get our gentrification story straight, and to recognize that change has been ongoing in London for many decades, and that it is one key to the cities economic survival. Hill suggests its time to move past protest politics and develop a constructive, practical and flexible response to shaping urban change to its best and most equitable effect. Edward Clarke at the UK Centre for Cities boldly writes “In defence of gentrification.” He points out that The most successful cities in the UK are also the least affordable to live in, primarily because of the shortage of housing, and because the development of new homes comes nowhere close to matching the growing demand for urban living.”  The real problem is not change, but the failure of the planning system to address the UK’s shortage of cities.

2.  Fifty steps toward carbon free transportation.  While we’ve made great progress in reducing emissions from power plants and industry, transportation has emerged as the single largest source of climate pollution in the United States. In a new report published by the Frontier Group, Tony Dutzik and his colleagues detail six common sense principles to guide policy and list 50 specific steps we can take to move toward carbon free transportation.  Most are eminently sensible, but we think #18 is probably the most important:  putting a price on carbon, and a price on road use is key to sending signals to consumers and investors about what they need to do to combat global warming. A price for carbon makes every other one of the 49 strategies easier to accomplish.

New knowledge

1.  The Atlas of Urban Expansion.  New York University’s Marron Center, in cooperation with UN Habitat and the Lincoln Institute of Land Policy has created an Atlas of Urban Expansion.  It’s a truly global look at patterns of urban development, with detailed information from 200 cities on six continents.  Using satellite imagery, they estimate the extent and density of urban development in 1990, 2000 and 2014, and provide estimates of the width and coverage of the road network.  It’s all rendered in colorful, interactive and highly detailed maps.  Fourteen US cities are covered, including Los Angeles, Cleveland, Minneapolis and Portland. Here’s their map of urban expansion in Portland.

2.  Who votes for Mayor?  One of the issues we’ve regularly covered at City Observatory is the “home-voter” hypothesis:  the notion that local land use policies reflect the interests of homeowners because they disproportionately turn out for local elections. Additional evidence for this thesis comes from new research from Portland State University (funded by Knight Foundation) which looks at the turnout patterns in local mayoral elections around the country. They find that these local elections attract abysmally low levels of turnout, and perhaps more importantly are highly skewed toward older voters: On average voters in mayoral elections are roughly a generation older than the voting eligible population. Because homeownership is strongly correlated with age, this finding lends considerable credence to the home voter hypothesis—and of course has broader implications for the policy leanings of local governments.  The report provides detailed data for 30 of the nation’s largest cities.

Last week, we noticed a small item on Streetsblog: “Where Walk Score falls short.” Because we’re keenly interested in to walkability, and routinely use Walk Score to benchmark walkable places, we clicked the link.

It took us to a blog entry from Mariela Alfonzo asking “Does walk score walk the walk?” Dr. Alfonzo has been doing walkability research for a number of years; she collaborated with Chris Leinberger in applying something called the Irvine/Minnesota Inventory to measure walkability in Washington DC, and has since launched a consulting firm – “State of Place — that sells walkability measures.

Her critique of Walk Score is based on an analysis of three instances in Washington DC suburbs where places have high walk scores, but poor walkability, as measured by her firm’s proprietary “state of place” scoring system. She argues that its “irresponsible and potentially discriminating” to use Walk Score “to make planning, private investment or public funding or policy decisions.”

That’s a pretty strong claim. And one frankly that we profoundly disagree with.

That’s not to say that we think Walk Score is perfect. It isn’t. At City Observatory, we’ve never hesitated to critique its limitations–we flagged a problem with its city rankings last year, for example.

To their credit, Matt Lerner and the team at Walk Score have always been utterly open and transparent about the limits of their data and algorithms, and have made changes to address those concerns (substituting street-smart distance measures for straight line calculations, for example). As a result of these steady improvements and despite its limitations, Walk Score has done more to advance interest in and awareness of walkability than any–and perhaps all–of the academic research on the subject. And that’s one of the great things about Redfin, the new owners of Walk Score: despite the fact that they’re a private, profit-making company, they continue to provide Walk Score for free, and make its workings completely transparent to everyone. And they’ve gladly worked with all comers to make it better and to advance the field. They’re real models of how to move markets and do good.

As we’ve noted at City Observatory, for too long our transportation discussions have been subtly but powerfully slanted by the dominance of car-oriented system metrics—average daily traffic, level of service, hours of delay. What’s long been missing from urban planning and transportation investment decisions is clear metrics that characterize the role of walkability in contributing to livability and other public policy goals. Walk Score helps level the quantitative playing field.

Walk Score is simple, transparent, ubiquitous and free. It rates every address in the United States on a scale of 0-100, let’s users drill-down to a neighborhood level map that shows what’s driving the walk score for a particular location, and doesn’t cost a dime to use. Walk Score serves up over 20 million scores a day, and more than 30,000 web sites use its services. Our own research, and that of others, shows that Walk Score has a measurable impact on home values. Walk Score works, it gets used, it has been validated independently, and it has changed real estate markets. Everyone who cares about walkability owes its creators, including Alan Durning at Sightline Institute who brainstormed the idea, and Matt Lerner, Mike Mathieu and the team at Walk Score who turned it into a reality—a huge debt of gratitude for their work.

At her firm, Mariela Alfonzo is pitching its proprietary methodology for assessing walkability. For a fee, they’ll come to your community and gather data on-site, cataloging 280 different aspects of the street scape, from the numbers and types of businesses, to crosswalk markings and materials, building setbacks, the presence of loose dogs, unpleasant smells and graffiti. (You can read the entire list on pages 75 to 84 of this PDF).

It’s fine to suggest that there might be other variables that affect walkability. And yes sidewalks, landscaping, building facades, road widths and traffic levels all matter. None of this is to argue that Walk Score is perfect, or can’t be improved, or that other measures can’t augment its assessment of walkability in particular locations. For example, the kind of work that Jan Gehl has pioneered shows how we can bring meticulous data collection and a design sensibility to the challenge of improving walkability and public spaces. There’s still a lot more to learn in this field.

It’s great that the folks at State of Place have developed their own methodology for assessing walkability. It undoubtedly will be useful in many situations. These metrics, however, are highly complex, extremely labor intensive to gather, and consequently very expensive. And they simply haven’t been gathered in enough places to have the kind of track record that would let an objective third party assess their utility.

Is State of Place Really Better?

Let’s ask a really fundamental question: Does all the added effort of collecting more than 280 data points for each street segment add significantly to our understanding of which places are walkable and which aren’t? For a real world test of that question, I examined a recent report—one of just a handful using the State of Place methodology to score a neighborhood. The data come from a study that State of Place conducted –with help from Portland State University planning students in the suburb of Tigard, Oregon. The study focused on the Tigard Triangle, a neighborhood bounded on each side by three major highways. The summary finding of the SOP report was that the Tigard Triangle had an average State of Place Score of 33, compared to the State of Place score for the city’s old downtown neighborhood of 66.

So how does that compare with what you could learn from Walk Score? While Walk Score’s main feature is the ability to look up walkability for a particular address, one of the great things about Walk Score is that they also use locally designated neighborhood boundaries to map and summarize walkability. So for Tigard, you can easily compare the average walkability of the downtown (Neighborhood Area 5) with that of the Tigard Triangle (Neighborhood Area 9). And Walk Score produces almost identical results to those reported by State of Place: downtown’s walk score is 62; the triangle’s walk score is 38.

Downtown Tigard	Tigard Triangle

In a way, that’s a kind of heartening result: It implies that the definitions of walkability used by the two methodologies are, at some level, broadly congruent. But if you’re being asked to pay for “State of Place” you might want to ask what it is providing that Walk Score doesn’t, and whether it’s worth it.

Different Horses for Different Courses

There’s plenty of room for many different, complementary and multi-scaled measures of walkability. Detailed, site level assessments may make sense for some purposes. Broad system-level measures that encompass entire neighborhoods, cities and metropolitan areas, and which are comparable–and available–for the whole nation also play a role.

Rather than a tendentious critique of Walk Score, this field would be better served by acknowledging that Walk Score is a terrific resource than provides a foundation for understanding and promoting walkability as a public policy objective, and that it creates a market for the more nuanced and frankly much more expensive kinds of tools that State of Place wants communities to buy.

There’s a growing appreciation that walkability is one of the most important and valuable aspects of great urban places. For a long time, we’ve had to make the case for walkability, relying mostly on the lyrical writings of Jane Jacobs and others, who described how a vibrant streetscape makes places livelier, more interesting, safer and more economically productive. While that message resonated with a faithful few, it didn’t carry much weight in the numbers and engineering dominated world of policy.  What’s been missing is the robust, universal and quantitative set of measures that put walking on an equal footing (sorry about the pun) with all of the quantitative measures used to promote auto mobility. If you’re really a passionate supporter of walkability, you ought to celebrate that accomplishment and build on it, rather than taking needless and poorly grounded swipes at its limitations.

Its poor form in any business to get ahead by bad-mouthing your competitors. Its even worse in this situation, when over-the-top claims about Walk Score being “irresponsible” actually undercut a shared common interest in promoting greater awareness of and knowledge about walkability. There’s plenty of work to be done in this field, and room for many different and complementary sets of tools: let’s work together to build up the field, rather than needlessly tearing down the one tool that’s actually worked.

Note:  This was post revised October 29 to correct the spelling of Mariela Alfonzo’s last name; We’ve also referred to her as Dr. Alfonzo, at her request. She has submitted a letter disagreeing with this commentary, which we make available to City Observatory readers here: Dr. Alfonzo’s Reply.

There’s a revolution afoot in transportation.  Transportation network companies, aka “ridesharing” firms, like Uber and Lyft are disrupting both the markets for urban transportation and labor markets.  Their business model–treating drivers as independent contractors, is fueling the so-called “gig economy.”  A new report from the Brookings Institution uses federal tax and administrative records to plot the growth in the number of these independent gig workers in the “rides” and “rooms” sectors (think Uber, Lyft and AirBnB).  Their report provides a useful new metric from understanding the growth and impact of these sectors.  While that’s well worth a read in its own right, the Brookings data also provides a detailed set of metropolitan level statistics that help us judge the places where ridesharing has scaled up the most rapidly.

Released today by the Brookings Institution, this report – Tracking the Gig Economy: New Numbers — estimates the size and growth rate independent employees in two key segments of the gig economy. Mark Muro and Ian Hathaway have exploited a little-used Census data series on “non-employer firms” to look at the growth of independent contracting in the transportation and accommodations industries. (See below for the geeky details of this data source).

The Brookings estimates cover the years through 2014, and give us a pretty good picture of the rate of increase in the number of people reporting taxable income from a non-employing business operation in passenger transportation and accommodations. Their estimates suggest that the employment in ride-sharing increased by about 69 percent between 2010 and 2014. (Data for room-sharing, less well captured due to the nature of tax treatment of rental income, show a smaller, but still substantial increase of 17 percent).

In their analysis, Muro and Hathaway also look at the relative size of the “gig” portion of the accommodations and transportation sectors in each of the 50 largest metropolitan areas compared to the number of wage and salary workers the same industries in those areas. They examine the trend data to see whether the growth of gig workers, say in ridesharing, is cutting into the number of people employed in the taxi business. So far, at least, the results are inconclusive—but another year or two of data may present a different picture, as all indications are that the industry has grown very rapidly in 2015 and 2016.

The Geography of Ridesharing Industry

While much of the Brookings report is focused on the growth of these two key sectors of the gig economy, at City Observatory, we’re always interested in which metropolitan areas are leading the way in the adoption of new ideas. So we set about using the Brookings data to assess the relative size of the ridesharing component of the gig economy in each of the nation’s 50 largest metropolitan areas.

We want to know which cities have, proportionately, the largest concentrations of independent contractors involved in providing passenger ground transportation services. To answer this question, we computed the number of non-employers in this category per 100,000 metropolitan area residents. The typical (median) large metropolitan area had about 83 ride sharing drivers per 100,000 population in 2014. Of course, this number varied substantially among different metros.

New York (461 per 100,000), San Francisco (377) and Washington (365) had the highest concentration of transportation non-employers, with about four to five times as many on a population-adjusted basis as the typical city. At the other end of the spectrum, Rochester, Memphis, Hartford and Oklahoma City had the smallest concentrations of transportation non-employers, with fewer than 50 transportation non-employers per 100,000 population.

It’s apparent that the growth of ridesharing has proceeded most rapidly in larger, denser cities, and has lagged in smaller and more sprawling ones. The richest markets for transportation network companies are where there are lots of potential customers, and where private car travel is expensive and inconvenient. In a future post, we’ll take a closer look at the relationship between metro area characteristics and the penetration of ridesharing, as indicated by the Brookings tabulation of transportation non-employer data.

Cross-checking the data

Because this is such a new and poorly measured sector of the economy, we want to be careful in leaning to heavily on any single source of data. Its worth asking how well the Brookings estimates square with other available data the size and growth of the gig economy. While none of these firms has yet gone public and though they carefully guard many of their financial results, there are a few fragments of data in the public realm. While its difficult to directly compare administrative records and the scanty firm level data, the picture painted in the Brookings report fits well with what we know about the size and growth of Uber in the transportation networking market.

In late 2015, Uber reported that it had more than 327,000 active drivers on the road in the US, a figure more than double the 160,000 on the road in 2014. (Active drivers were those who had provided more than four rides in a month. The Brookings report estimates that in 2014, there were about 341,000 non-employer firms engaged in taxi, limousine and ground transportation services, up about 112,000 from two years earlier. Allowing for the fact that many drivers may have worked only part year, or too few hours to be counted by the Census Bureau, that seems like a roughly consistent pattern of growth.

Business Insider reports that Uber had about 5,600 active drivers in San Francisco at the end of 2013. Brookings reports that there were about 7,000 non-employers in the transportation services categories in 2012 and about 17,000 in 2014.   This suggests that Uber’s active drivers were equal to about half of the reported number of non-employers (assuming steady growth between 2012 and 2014).

Geeky data details

The most commonly reported data series for the economy report non-farm wage and salary employment—the number of people on the payroll of governments and private businesses. Using administrative records—tax returns filed with the Internal Revenue Service—the Census Bureau computes data on people working for themselves. Most of these firms are in the so-called “1099 economy” – people who instead of receiving a W-2 from an employer receive a form 1099 indicating that they’ve received a payment from a firm but that they are not an employee of that firm. The 1099 form also picks up income from investments–like rents and royalties–and some non-Internet businesses would clearly be captured by this same data source (say an independent, sole-proprietor taxi driver, or someone who ran an actual bed-and-breakfast). But the growth in this category in the past few years is likely to be very indicative of the growth of these businesses.

The Census Bureau tabulates these records according to industry (using the North American Industry Classification System – NAICS) and by county. Muro and Hathaway aggregated the county level data up to metropolitan areas, and looked at the statistics for two particular sets of NAICS codes (roughly corresponding to taxi and limousine services and other ground transportation providers, and rooming houses and other accommodation.

As regular readers of City Observatory already know, the use, misuse and abuse of real estate price indices is one of our pet peeves. We’ve repeatedly excoriated Abodo, Zumper and others for mis-representing median values calculated from their apartment listings, as rental inflation gauges, because they work more like random number generators than measures of actual market activity.

It’s all well and good that so many more journalists are paying attention to data, but in our view, they need to be a lot more thoughtful about presenting and interpreting it. This month we turn to a story from the Guardian, which reports on a new 15 percent tax the British Columbia Government imposed on foreign purchases of residential real estate.

In an article entitled “Has Vancouver found the solution to a super-heated housing market?”* As the Guardian’s Patrick Colison hails the tax as having produces a nearly immediate one-sixth decline in real estate prices.

The impact has, by some measures, been more startling than campaigners could have hoped for. The price of the average detached home reportedly slumped by an astonishing 16.7% in August alone to C$1.47m (£856,000), according to the Real Estate Board of Greater Vancouver.

“Startling.”

“Astonishing.”

But in fact, is there any evidence that prices are down, or if the new tax had anything to do with it? Actually, no. The dramatic data Colison cites comes from an article published by Huffington Post BC, which quotes the original source of the data, the Vancouver real estate board.  If you click through and read the Huffington Post article (alarmingly titled “Vancouver average detached home prices show worst slide in 39 years” ) it does present the 16.7 percent figure but then also goes on to point out why that reported number is likely a highly inaccurate indicator of real changes in home prices.

As it turns out, the Real Estate Board’s figure is the simple arithmetic mean of the value of home sales in a given month. As a result, it’s highly subject to short term variation due to composition effects. In one month, the average price of homes sold fell by 16.7%, but not because the price of a typical home declined, but because fewer very high end homes sold that month. There are lots of reasons why the composition of sales might change, summer is a slow season, and perhaps more importantly, some people may have made sure their sales went through before the tax went into effect.

Here’s how composition effects work: Imagine a community where 10 houses sold each month: nine $500,000 houses and one $5 million house. The arithmetic mean of the sales prices is $950,000 ($9.5 million in total value divided by 10 houses). Now suppose in the next month, ten houses are sold, but they are all of the $500,000 variety. Then, the average price of home is $500,000 ($5 million divided by 10 houses). If you looked just at the overall mean, you might conclude that prices had fallen by more than 40 percent. Of course, the price of the typical house didn’t change.

The gold standard for judging house price changes is a repeat sales index, which looks at the same house each time its sold and looks to see how much it appreciated since its last sale.  You compute an annualized growth rate (every house has a different period since it last sold) and then use these changes in value to judge how much housing prices increased or decreased.  You can also use a hedonic or quality adjusted model, so that you factor in the differences in the size, location and other aspects of a house.  Vancouver’s quality adjusted “benchmark” estimates are that prices were basically flat in the last month or so, and are up about 30 percent over the last year.

If one looks at sales of like homes, there’s no evidence of a decline in Vancouver home prices, at least not yet, and certainly not a startling, astonishing 16.7 percent.

None of this is to say that a foreign buyer tax isn’t a good idea—and it might actually have an impact on the rate of home price inflation at the high end of the market or that it might turn out to be a good way to help finance more affordable housing. But cherry-picking a pretty obviously unreliable one-month indicator to make such a claim its already solving the problem isn’t responsible journalism.

It may well be that in a handful of global markets (London, Manhattan, Sydney, Vancouver) foreign buyers are driving up prices. But that’s unlikely to be the case in most places. As we’ve seen in most cities around the US, the cause of rising rents is not rich foreigners, but the combination of a growing demand for urban living running headlong into significant constraints on building housing in the kind of city neighborhoods people increasing want to live in. Whether a foreign-buyer tax will actually dampen rent or house price inflation is still far from clear—and it certainly can’t be discerned from a hasty and inaccurate reading of one-month’s data.

Notes

* – This is a classic example of what at City Observatory we have called “Hertz’s Law” which was propounded by our colleague Daniel Kay Hertz. It holds that if a blog post headline is phrased as a question the answer is almost always: “no.” Framing a tenuous proposition as a question allows the clever journalist the ability to simultaneously imply that the proposition is in fact true, while maintaining plausible deniability that the journalist has actually reached that conclusion (or in fact, any conclusion).

If you are looking for an apartment to rent, you definitely might want to check out listings on Abodo and Zumper; just please don’t use their deeply flawed measures to judge rent changes.

Every few months, the national prize pool in the multi-state Powerball lottery piles up to tens or even hundreds of millions of dollars. Early this year, three lucky winners split a prize with a total value of $1.6 billion dollars. Bucking odds of about one in 262 million, Marvin and Mae Acosta’s purchase of a Powerball ticket at a Chino Hills, California convenience store changed their lives. Overnight, this couple of modest means became very, very wealthy.

For the rest of us, it’s a pleasant diversion to daydream about what it must be like to win the lottery. And while it’s an entertaining scenario, given its magnitude, it’s not like anyone is suggesting that the Powerball is a solution to our problems of poverty or inequality. And yet, when it comes to housing, one widely touted policy amounts to holding highly visible lotteries that help a relative handful of people, but really don’t do much to address the problem’s scale.

For many affordable housing advocates, inclusionary zoning is all the rage. The idea behind inclusionary zoning  that we can solve our housing affordability problems by requiring developers to build a certain percentage of affordable units as part of any new housing development and then sell or lease them at below-market prices to households with low or moderate incomes.

We and others have pointed out that inclusionary zoning requirements tend to drive up the cost of the market rate units that do get built, because the cost of the affordable units has to be made up somewhere. Additionally, inclusionary zoning requirements tend to reduce the number of new units that get built. This adds to the upward pressure on prices and rents, lowering affordability for those not fortunate enough to land subsidized housing.

But another key problem with inclusionary zoning is scale: The number of affordable units that get built is tiny, compared to the demand for them. From a policy perspective, that means that IZ hardly makes a dent in the size of the affordability problem. And for cities it creates the very practical problem of deciding who, among tens of thousands of potentially eligible low- and moderate-income households, will get one of the coveted affordable units.

Sally French, a reporter for Dow Jones Market Wire, gives a real-world dispatch on this process in  “How I bought a condo in San Francisco for $268,000.”  French explains how she applied, entered the housing lottery, and was chosen to buy one of a few affordable condominiums available in San Francisco. By her estimate, she paid about one-third the going price for unit. Her story illuminates two key features of the program: how few units are actually available and what it takes to get selected to buy one.

In a good year, San Francisco has about 200 to 300 below market rate units to allocate. But based on the program’s income limits, something like 60 percent of the city’s 350,000 households qualify to participate in the lottery: nearly 200,000 potential applicants, or several hundred per available unit. And that understates the ratio considerably, because people living outside San Francisco can also apply. French was renting a room across the bay in Berkeley when she applied.

Holding a lottery to award deeply discounted condominiums to a lucky few is no more a way to solve our housing problems than encouraging poor people to buy lotto tickets is a viable solution for poverty and inequality. Every week, the state lotteries and the big megabucks turn an infinitesimal fraction of average citizens into newly minted millionaires or multi-millionaires. No one pretends big-prize lotteries reduce inequality. Yet that’s pretty much what cities are doing when they offer up big prizes for token numbers of households.

With the actual lottery, you get the disclaimer that “these are games of chance and should not be played for investment purposes.” The same kind of warning ought to apply to inclusionary zoning and attendant housing lotteries: This is a policy that is more about symbolism than large scale results, and shouldn’t be regarded as a solution to housing affordability.

Its very clear that IZ housing lotteries are too small to meaningfully address the affordability problem. Whether the lottery process actually gets homes to the neediest households is the second question. French’s story makes it clear that you have to have the acumen and tenacity to compile a complex application and supporting documents (three years of tax returns and bank statements) as well as dedicating time to attending the lottery itself and homeownership classes. Little surprise that energetic younger, well-educated applicants, like French, would have a leg up in negotiating the very bureaucratic process. And her co-winners include a teacher, an architect and an urban planner.

Additionally, the program’s eligibility requirements create some perverse incentives. French describes how other applicants have turned down raises or passed up overtime to keep their incomes below the level that would qualify them for lottery housing. As she details how the below-market rent program works, including the trials and tribulations of applying and coping with bureaucracy, and the excitement of becoming a homeowner, it’s hard not to be happy for French. And as envious as we might be, we certainly don’t begrudge her winning; and its very much a public service that she has candidly related her personal story of navigating the lottery. But at the same time, we have to recognize that the process of just applying for an IZ unit leaves dozens or hundreds of equally deserving households with nothing.

The approach to allocating below-market rate housing in San Francisco mirrors the situation with housing vouchers, the nation’s largest program to provide rental assistance for the poor. Of those eligible, fewer than a quarter receive any assistance. It too is effectively a lottery; recipients of vouchers call them “the golden ticket.” This frequently makes housing assistance more like a lottery game than a social service, as when 58,000 people applied for 105 homes in New York, or nearly 300,000 people were placed on the waitlist for a Chicago Housing Authority unit. In contrast, the nation’s homeownership programs are generously funded ($250 billion annually through the tax code) and are automatically and universally available to everyone who qualifies. When affordable housing is so scarce that it has to be allocated to a relatively fortunate few via a lottery, it’s a good indication our policies aren’t up to the challenge. Good luck solving our housing affordability problems with this approach.

One trend we’ve tracked at City Observatory has been the movement of jobs back to city centers. While there are an increasing number of examples of prominent firms moving downtown — GE abandoning its suburban campus for a location in Boston’s Seaport district, McDonalds moving from Oak Brook to a site near Chicago’s Loop — we’re interested in broader economic data that shed light on the strength and persistence of this trend.

In our report, Surging City Center Job Growth, we’ve used data from the Census Bureau’s Local Employment and Housing Dynamics (LEHD) dataset to track neighborhood level changes in employment patterns in large US cities.  These data confirm, that for most large cities for which reliable data is available, that job growth has slowed on the urban periphery and accelerated in urban centers.

A new paper from Nathaniel Baum-Snow and Michael Hartley provides some interesting data on the geographic patterns of employment change within U.S. cities. While the paper is primarily focused on the demographics of neighborhood change, it looks in part at how changing job growth patterns may have influenced population location. Analyzing data for 120 large US metropolitan areas for the two decades 1990 to 2010, Baum-Snow and Hartley compute the rate of employment growth by distance to the center of the central business district.

The following chart shows employment growth rates for two decade-long periods, 1990 to 2000 and 2000 to 2010.  The vertical axis shows the percentage increase in employment over the decade (shown as a decimal, so .10 = 10%).  The horizontal axis shows the distance from the center of the central business district in kilometers (0 kilometers is the center of the CBD).  The values reported are the median percentage increase in employment for the 120 metropolitan areas covered, and so represent the experience of the “middle” metropolitan area in this group (not a population or employment-weighted median).

There’s a stark difference between the two decades.  At first, the suburbanizing, decentralizing pattern of employment growth was very much in evidence. In the 1990s, the further you went from the urban core, the faster jobs were growing. Within 2 kilometers (1.2 miles of the center), job growth was, in the typical metropolitan area declining.  Beyond 15 kilometers (about 9 miles) the total number of jobs more than doubled during the decade (i.e. and increase of more than 100 percent over the decade).

More recently, that decentralizing trend is disappeared. Between 2000 and 2010, job growth was nearly the same regardless of distance to the central business district.  Within four kilometers of the CBD, employment growth improved compared to the 1990s; beyond four kilometers, employment growth was much slower than in the 1990s.

The dramatic flattening of the employment growth curve over the last decade is evidence that the historic pattern of steady employment decentralization has, at least for the time being, ground to a halt.  These data confirm that over the last decade, city centers have dramatically improved their economic performance relative to suburbs. We’ll keep following this and other data, to see where this trend leads.

We shouldn’t expect the return of the trade-up buyer anytime soon.

Is the American homebuyer increasingly stuck in a starter home? That’s the premise of a recent commentary from the Urban Institute “Do we have a generation stuck in starter homes?” Looking at data on the share of home mortgages going to first time home buyers as opposed to repeat buyers, the Urban Institute’s Laurie Goodman, Sheryl Pardo and Bing Bai show, quite persuasively, that since 2008, a smaller and smaller share of home sales are accounted for by repeat purchasers (i.e. a household that already owned a home, and is buying a new home.) In 2015, there were only about 900,000 repeat homebuyers, compared to more than 1.8 million in 2001.

While the facts presented here strike us as exactly correct, we’re having a hard time signing on to the interpretation that’s offered. Its far from clear that we should want or expect their to be lots more repeat or trade-up home buyers. For the most part, the dominance of trade-up buying was a symptom of last decade’s tragic and un-sustainable housing bubble.

For a moment, let’s step back and think about the logic that’s implied here: the notion is that first time homebuyers buy a small home that they can afford and that over some period of time it appreciates – i.e. in real terms, over and above the rate of inflation – so that the homeowners can now take their increased equity (and perhaps add it to a higher level of household income) and “trade up” to a larger home or a nicer neighborhood. As Goodman and her co-authors explain:

The basic strategy for moving up from a starter home to a bigger home, however, has been to accumulate equity in the home through consistent price appreciation, and convert that equity into the down payment for a more expensive home.

That was certainly a big explanation for the housing market dynamics of the early 2000s. Thanks to rapidly rising home prices, homeowners found themselves awash in equity. And there were two other factors at work as well: First interest rates were dropping. From 2000 to 2006 the average interest rate on a thirty year mortgage fell from more than 8 percent to less than 6 percent. Lower rates mean—in a straightforward way—that one can buy more house for the same payment. At the same time, as is well-known, private sector lenders offered a range of “sub-prime” loans with lower but floating interest rates, and non-amortizing features.

The other thing to keep in mind, (silly as it seems now, in retrospect) was the nearly universal belief among consumers that home prices could only go up. So, in effect, existing homeowners had a bonus in terms of rising equity, more purchasing power thanks to declining mortgage rates, and the expectation that the financially smart thing to do was double down in the housing market. All this was amplified by reckless lending policies by many financial institutions. Nobel Laureate Robert Shiller has underscored the importance of price expectations in driving the bubble—and they were certainly a key factor in enabling and incentivizing homeowners to “trade up” in the hot housing market. At the height of the housing boom, home price expectations greatly exceeded mortgage interest rates: so to buyers homes looked “free”: appreciation more than offset the cost of interest payments.

Little wonder that so much of the market activity in the bubble years was propelled by repeat borrowers. But actual and expected home price inflation came crashing earthward in 2006 and 2007 and suddenly the demand for homes collapsed. Karl Case, Robert Shiller and Ann Thompson illustrated this in their paper “What have they been thinking? Home buyer behavior in hot and cold markets.

There’s little sign that anyone expects home price appreciation to accelerate anytime soon. The Federal Reserve Bank of New York annually surveys a representative sample of consumers nationally about their five-year expectations for home price inflation; that number has declined by a third over the past two years from 3.01 per cent in 2014 to 1.89 percent in 2016

But there’s little reason that we should regard sustained home price appreciation that leads large numbers of homeowners to repeatedly “trade up” as a normal condition of housing markets, or, given changed circumstances any thing that is likely to recur any time soon. Nor should we wish or expect that to happen. As the collapse of the last bubble showed, that kind of growth isn’t sustainable. And as we’ve pointed out at City Observatory, sustained house price appreciation runs counter to our interests in promoting housing affordability.

Fretting that today’s generation of first time home-buyers—who are generally wealthier and have higher incomes than their home-renting peers of the same age—are somehow stuck in “starter homes” because they’re not “trading up” at the same pace as their predecessors were at the height of the housing bubble seems like one of the smaller problems facing U.S. housing policy, if it is a problem at all. Our attention and policy concern might better be directed to the situation of millions of renters who are facing steadily higher housing costs.

.

What City Observatory did this week

1.  More evidence job growth is shifting to city centers.  A recent paper by Nathaniel Baum-Snow and Daniel Hartley has some interesting data on the pattern of job growth in the nation’s largest metropolitan areas. They find that while suburban area job growth greatly outpaced that of the central city in the 1990s – basically, the further away from the core you were, the faster you grew – that this pattern totally changed in the decade 2000 to 2010. (Chart shows job growth rate by distance to the CBD, in kilometers).

2.  Vancouver’s foreign-buyer tax didn’t slash home prices. This summer, the Province of British Columbia imposed a 15 percent tax on the sale of residences to foreign buyers. Some press accounts described the tax increase as an instant cure for home price inflation, claiming it resulted in an overnight 16 percent decline in home prices. In fact, Vancouver home prices were basically flat; we explain how journalists made a fundamental error in examining the data; one that is unfortunately all too common in reporting on real estate trends.

3.  The most interesting neighborhood in the world.  Google Maps has added a new feature, areas of interest, which show up as peach-colored blotches. Its not entirely clear how Google selects these areas, and some observers are concerned that there may be hidden biases at work. Using City Observatory’s Storefront Index, we’ve looked to see how the clustering of consumer-facing retail and service businesses corresponds to the places Google characterizes as interesting.  Storefront clusters closely correspond to the areas google flags as “interesting.”

4.  Where is ridesharing growing fastest?  A new report from the Brookings Institution uses federal data on so-called non-employer businesses to track the growth of independent contractors in the “rides and rooms” segment of the gig economy. We use the Brookings data as a proxy for the penetration of transportation network companies like Lyft and Uber in the top 50 metropolitan markets. The leaders are New York, San Francisco and Washington.

This week’s must reads

1.  The under-appreciated benefits of public housing. Public housing has long been associated with negative consequences. It’s emblematic of concentrated poverty, and it’s widely believed that large-scale, high density public housing projects breed social pathologies.  But a new paper from John Haltiwanger and colleagues challenges that view.  Part of the problem of assessing the impact of public housing is the problem of “selection effects” – people who are chosen to live in public housing, by definition, are low income, and suffer from many challenges.  The key question is, once chosen to be in public housing, how do children do? Haltiwanger, et al overcome this selection effect problem by looking at the different outcomes for different children within a single household based on the amount of time they live in public housing. Comparing siblings allows the researchers to net out the effects of families (parental education, income, employment, etc), and judge the separate impact of housing itself.

While the impacts are different for boys and girls, the study shows that public housing actually benefits children of poor families over their lifetimes, as measured by income and likelihood of incarceration, (girls see larger gains in income and lower rates of incarceration than boys). This holds for both traditional public housing and housing vouchers. In large part this seems likely to be an income effect: families getting vouchers or living in public housing spend less on housing than they would if they rented in the private market, without subsidies, and therefore have more resources for all their other expenses, which may translate into better outcomes for their children. As we’ve noted at City Observatory, the big problem with our housing programs for the poor is their lack of scale: fewer than one-in-four households that technically qualify for housing assistance get anything. And this study suggests that there could be substantial benefits from expanding the reach of these programs.  The full report is available from the National Bureau of Economic Research (NBER): Childhood Housing and Adult Earnings: A Between-Siblings Analysis of Housing Vouchers and Public Housing.

2.  Tales of Rent Control  John McNellis tells us what rent control looks like from the perspective of a small landlord. His family inherited a small 5-unit San Francisco apartment building which fell under the city’s rent control program. Two of the apartments turned over after they inherited the building, but three tenants were ones who were in place. Rent control created a strong adversarial relationship between the landlord and tenants; one tenant even insisted all communication be in writing. Eventually, McNellis’s family tired of the burden of being landlords, and opted to sell . After the sale, McNellis reports finding out that three of his nominal tenants had actually sublet their apartments — at market rates, “Our apartments were thus rented at market rates: just not by us.” Tens of thousands of apartments, particularly smaller complexes that make up the “missing middle” housing are owned not by big corporations, but by smaller part-time and amateur landlords. Rent control creates strong incentives for them to take their capital and invest elsewhere, with the paradoxical result that the landlord business is further dominated by more ruthless and efficient corporate owners.

New knowledge

1.  Housing, Interest Rates and Inequality:  Writing at VoxEU, economist Gianni LaCava describes the results of his research looking at the relationship between home prices, interest rates and economic inequality. Digging in to state level data, LaCava confirms what Thomas Piketty and Matthew Rognlie have argued:  that the distribution of income and wealth in the United States has become more unequal, and that it has been the growth of housing wealth, in particular, that has been the big driver of inequality.  His work also shows that housing prices have accelerated most in those states with the most constrained (i.e. least elastic) housing supply, and that inequality accelerated as real interest rates declined steadily from the 1990s onward. If we’re concerned about inequality, LaCava argues, we need to pay more attention to the distribution of housing wealth, and the “imputed” income that homeowners receive.

2.  Home prices and economic mobility in California. California’s Legislative Analyst has produced some of the most insightful examinations of the connections between housing markets and the economy.  A new paper by LAO looks at the connection between high home prices in coastal counties (around San Francisco and Los Angeles) and the inland counties of the state.  It finds that historically, incomes were converging between the lower income inland counties and the higher income coast counties. But that pattern has reversed in recent years, and that’s attributable to higher housing prices on the coast. The lack of housing has forced more people to live in the interior counties, where economic opportunities and productivity are lower. And significantly, as measured by residual income—the amount of income left after paying housing and transportation costs—incomes are actually lower in these “cheaper” counties, meaning that these families would be economically better off if they could live in one of the coastal counties.

What City Observatory did this week

1.  Cities for Everyone:  Our Birthday Wish.  October 15 marked City Observatory’s second birthday. We reviewed some of the highlights of the past year, focusing on the growing evidence of the economic resurgence building around the nation’s cities. For the coming year, we’re planning on focusing on what it takes to build and maintain diverse inclusive communities:  cities for everyone.

2.  The Price of Parking.  Using data from the website Parkme, we’ve constructed an index of typical monthly parking costs in the nation’s largest metropolitan areas. While the median price in large cities is around $200 per month, there’s huge variation. Prices range from more than $700 per month in New York to less than $30 monthly in Oklahoma City. See how your city’s parking prices compare to others.

3.  Cities and the Price of Parking. Using our city parking price index, we look at the relationship between parking prices and transportation behavior in different metropolitan areas.  Our analysis shows a strong correlation between parking prices and transit use: people are much more likely to take transit in cities with expensive parking. Parking prices also correlate closely to the penetration of ride-hailing businesses like Lyft and Uber. The population-adjusted number of transportation service non-employers (a proxy for ride sharing businesses) is highest in cities with the most expensive parking. Parking, as it turns out, is a surrogate form of road pricing, with parking charges discouraging peak period car trips to urban centers and shifting travel to other modes. The more widespread deployment of ride-hailing may transform the role that parking prices play.

4.  The Myth Rich Cities/Poor Suburbs. There’s a new narrative about cities that claims that we’ve already experienced a great inversion, with poverty in the suburbs and wealth in the cities. Despite the resurgence of city populations in the past two decades, however, it’s still the case that poverty, especially concentrated poverty, is disproportionately found in cities, and that suburbs, especially newer and more distant ones, have much lower rates of poverty. At current rates of change it will be many decades before city and suburban poverty rates are equal. Rather than assuming that any movement of better educated and higher income people into cities makes their problems worse, we ought to look to see how we can leverage the re-investment in cities in away that increases opportunity and maintains diversity.

This week’s must reads

1.  Implementing a carbon tax: politics makes strange enemies. In less than three weeks, Washington State voters will decide the fate of a proposed carbon tax. I-732 would impose a $25 per ton carbon tax, and use the proceeds to reduce the state’s sales taxes and pay rebates to low income families, as well as cutting some business taxes. A key aspect of the plan is that it is revenue-neutral—with the funds raised by the tax entirely returned in the form of tax cuts. While that’s a feature for some advocates, it’s a fatal bug to other interests, which is why, surprisingly, some of the strongest opposition comes from a number of environmental and social justice groups. Writing at Vox, David Roberts tells the story of a how a progressive economist’s market-based pricing solution for climate change has run into a buzz-saw of political opposition from groups who’d like to see carbon tax revenue used to fund a slew of clean energy and transition support programs.  The practical difficulties of engineering an agreed-upon policy approach to climate change – among groups that agree something needs to be done –likely foreshadows the conflicts that will play out when this issue finally reaches the national stage, which one hopes will be sooner rather than later.

2.  Achieving economic integration:  Tales from Chappaqua.  Bill and Hilary Clinton live in (for the next little while at least) in Chappaqua, New York, an up-scale suburb in Westchester County. In addition to its famous residents, Chappaqua is also ground zero for the battle to build more affordable housing in an area that has traditionally been zoned almost exclusively for relatively expensive single-family homes.  Politco relates the history of the city’s exclusionary policies, and describes the present day conflicts in trying to site affordable multi-family housing.  As President Obama’s recent endorsement of “YIMBY”-Yes in my backyard zoning reforms indicates, this may become more of a national issue in the months ahead.

3.  A road much less traveled. Between Austin and San Antonio, there’s a brand new freeway where you can cruise along—legally—at 85 miles per hour. And you’ll find that it’s nearly free of traffic. And that’s the problem: the SH 130 toll-road, built by one of those vaunted “public-private partnerships” we hear so much about, is careening into bankruptcy. The San Antonio Express-News calls the project “a monument to failure” and tells the story of how wildly optimistic traffic estimates and federal loan guarantees led to the construction of a billion dollar highway, that nobody seems to think is worth paying to drive on.

New knowledge

Historical Maps of Redlining.  And this week, we have some “old knowledge” in a new form.  A team of researchers from three universities, including the University of Richmond, the University of Maryland, Virginia Tech and Johns Hopkins, has digitized the maps and neighborhood descriptions compiled by the Home Owners Loan Corporation (HOLC) in the 1930s.  These maps show the redlining of many urban neighborhoods that pre-figured decades of disinvestment and decline. In addition to the maps, you can also read individual, type-written descriptions of particular neighborhoods.

Although the HOLC maps are often themselves blamed for redlining and disinvestment, its actually more likely that they mostly codified widespread community attitudes about older, poorer and minority neighborhoods.  The maps and descriptions were compiled by mortgage lenders, developers and real estate appraisers.  And some academic evidence suggestions that HOLC and others still made loans in the redlined areas—although at higher interest rates.  The maps and the narrative are well-worth a read:  they provide a real historical context for thinking about the way “neighborhood stigma” can become the kind of self-fulfilling prophecy that has long term economic consequences.

What City Observatory did this week

1.  Bubble Logic.  A major and persistent change in the housing market from a decade ago has been the decline in the number of “trade-up” home-buyers. While some fret that recent first-time homebuyers have become locked in to so-called starter homes, we point out that in many ways, trade-up demand was a product of the unsustainable housing bubble of the last decade. The lingering effects of the bubble’s collapse, and an enduring change in expectations about home price inflation strongly suggest we won’t see a resurgence of trade-up demand anytime soon.

2.  Why a housing lottery won’t solve our affordability problems.  Inclusionary zoning programs require developers to set aside some units in new developments for low and moderate income households. Because tens of thousands of households are potentially eligible for a few hundred units, cities face the practical problem of choosing who gets to buy or rent these cut-price homes. Sally French tells the story of how she entered, and won, San Francisco’s housing lottery, and was able to buy a new condo for about one-third the going rate. That’s great for her—and a relative handful of others—but is hardly a scalable solution to our housing affordability problems. Her experience shows that you also need a good deal of savvy and persistence to negotiate the lottery process.

3.  Are integrated neighborhoods stable? Its long been popular to think of the process of neighborhood change as being characterized by “tipping points.” Once the demographics of a neighborhood start changing, from say mostly white, to more mixed race, it tends to “tip” to being entirely a community of color, because many whites may not feel comfortable if they’re not a majority.  New evidence of neighborhood change shows that once established, mixed race neighborhoods are in fact stable.

4.  A memo for Stockholm.  Next Monday, we’ll learn the name of the latest Nobel laureate in economics. We think a good case could be made that the award should go to Paul Romer, recently named as the new chief economist for the World Bank. Romer’s seminal contributions to New Growth Theory have long been recognized by his academic peers—and have important implications for urban economic policies. And in recent days, Romer has been making a strong case that the profession’s approach to macroeconomic policy has become profoundly unscientific and needs to be fundamentally re-thought.  This is the kind of thoughtful and provocative speaking truth to power that the Nobel prize ought to recognize.

This week’s must reads

1. How we talk about pedestrian deaths.  More than 4,000 pedestrians are killed in car crashes each year, and the way their deaths are reported in the media obscures the systemic nature of this problem.  In an essay at Streetsblog, Angie Schmidt points out that press stories routinely call deaths “accidents,” tend to blame the victims, describe the car, rather than its driver as the cause of the crash, and fail treat design of streets as a factor in the deaths. We know that multi-lane arterials and streets that encourage high travel speeds are responsible for a disproportionate share of pedestrian deaths.

2. Thinking hard about infrastructure investment. The two major party Presidential candidates may agree in some general way about infrastructure, but urban economist Ed Glaeser does not.  In an Interview with Vox, Glaeser points out we need to re-think the way we invest in infrastructure. Current approaches tend to systematically neglect maintenance in favor of shiny new projects, infrastructure investments are seldom subjected to serious cost-benefit analysis, and actual users rarely pay for the costs of projects they benefit from—which in some cases amplifies inequality. And in the case of roads, building more capacity without implementing some form of road pricing simply stimulates more demand—the fundamental law of road congestion.

3. Making mixed use developments legal.  President Obama got a lot of press in the urbanist world last week with the release of the White House’s housing toolkit—essentially a list of recommended policy changes that states and cities could undertake to allow more density. Writing this week in the Washington Post, Jonathan Coppage points out that the federal government could play a key role here as well, by changing its guidelines on residential mortgages to make it easier to include commercial space in residential buildings – think ground floor shops with apartments above.  Currently, federally purchased or guaranteed loans can only go to projects with no more than 15 to 25 percent non-residential uses, effectively precluding this important form of financing for developments that are less than four stories in height.

New knowledge

1.  Concentrated Poverty in the Wake of the Great Recession.  The Brookings Institution’s Elizabeth Kneebone and Natalie Holmes have sifted through the 2010-2014 five-year American Community Survey to track the growth of concentrated poverty in the US. Concentrated poverty is defined in their work as neighborhoods with a poverty rate of 40 percent or higher.  Their key finding:  since 2009, the number of people living in these extremely poor neighborhoods has increased 34 percent, from 8.7 million to 13.7 million.  That comes on top of a big increase in concentrated poverty since 2000; the number of people living in these high poverty neighborhoods in the US has more than doubled since 2000.  Concentrated poverty disproportionately affects people of color: blacks are nearly five times likelier than whites to live in neighborhoods of concentrated poverty. And despite the much noted increase in the number of people living in poverty in the suburbs, concentrated poverty is much more common in cities: about one in four poor persons in cities lives in a neighborhood of concentrated poverty, compared to about 1 in fourteen poor persons living in the suburbs. The Brookings report has detailed data on the top 100 metropolitan areas.

2.  US consumers pay some of the highest real estate commissions in the world.  A new survey of real estate broker commissions in 17 countries around the world shows that the US pays an average commission of about 5.5 percent, compared with about 1.5 to 3.0 percent in other high income countries.  While commissions in the US have declined slightly from an average of about 6.0 percent in 2002, the declines in other nations have on average been much sharper. Commissions in Canada have fallen from about 4.5 percent a decade ago, to about 3.0 percent per day.  Lower real estate commissions would make houses more affordable for everyone and lower transaction costs for real estate sales would make it easier for households to move to new homes and new neighborhoods.

If so many people live in suburbs, it must be because that’s what they prefer, right? But the evidence is to the contrary.

One of the chief arguments in favor of the suburbs is simply that that is where millions and millions of people actually live. If so many Americans live in suburbs, this must be proof that they actually prefer suburban locations to urban ones. The counterargument, of course, is that people can only choose from among the options presented to them. And the options for most people are not evenly split between cities and suburbs, for a variety of reasons, including the subsidization of highways and parking, school policies, and the continuing legacies of racism, redlining, and segregation. One of the biggest reasons, of course, is restrictive zoning, which prohibits the construction of new urban neighborhoods all over the country.

But does zoning really act as a constraint on more compact, urban housing? Sure, some skeptics might say, it appears that local zoning laws prohibit denser housing and walkable retail districts. But in fact, city governments pass such strict laws because that’s what their constituents want. Especially within a metropolitan region with many different suburban municipalities, these governments are essentially competing for residents and businesses. If there were real demand for denser, walkable neighborhoods, wouldn’t some municipalities figure out that they could attract those people by allowing that type of development?

A 2005 study by Jonathan Levine—and explored further in Levine’s 2006 book, Zoned Out—seeks to answer this question. Are local governments just responding to “market” demand in ensuring that new development is low-density and auto-oriented? Or is there really pent-up demand for more urban neighborhoods that can’t be satisfied because of zoning?

Levine looks for the answer in two contrasting metropolitan areas: Boston and Atlanta. Boston, as a much older region, has a relatively higher number of dense, walkable neighborhoods, while in Atlanta, which mostly boomed after World War Two, urban neighborhoods are much more scarce. Levine hypothesizes that if dense housing is adequately supplied to match people’s preferences, you should find a pretty good match between the kinds of places people say they’d like to live, and the kinds of places they actually do live. But if zoning really creates a “shortage of cities,” then the greater the shortfall of urban neighborhoods, the worse the matchup between stated preferences and actual living arrangements.

This is an important wrinkle to the “revealed preference” arguments of many defenders of the suburban status quo. Recent Census population figures sparked what were only the latest of a long line of scuffles over whether, or to what extent, the “back to the city” movement is real. But if Levine’s argument is correct, measuring demand for urban areas simply by how many people end up living there is flawed, because some people who would like to live in more compact neighborhoods can’t do so because there aren’t enough to go around.

To begin his analysis, Levine classified neighborhoods in both the Boston and Atlanta metro areas according to their level of “urban-ness” on a five-point scale, with “A” neighborhoods being the densest and most urban, and “E” being the most sprawling and exurban. Levine and his researchers then conducted a survey of residents in each of the zones, asking about their housing preferences and satisfaction with their current housing situation.

In Boston, about 40 percent of respondents said they preferred denser, more pedestrian-friendly neighborhoods, while in Atlanta, just under 30 percent of respondents did so. (Auto-oriented neighborhoods were preferred by 29 percent of people in Boston and 41 percent of people in Atlanta, with remaining respondents neutral.)

And how well did these preferences match actual behavior? Well, in Boston—where neighborhoods in the three most urban categories made up over half of all housing—83 percent of people with strong preferences for urban neighborhoods lived in one of these three urban zones. In Atlanta—where the same top three urban categories make up barely over 10 percent of all housing—just 48 percent of people with strong preferences for urban neighborhoods lived in an urban zone.

In fact, all down the line, people whose stated preferences were more urban were much more likely to actually live in an urban neighborhood in the Boston area than in the Atlanta area—suggesting that in Atlanta something might be preventing them from satisfying their preferences. At the same time, people who expressed preferences for the most auto-oriented neighborhoods were able to satisfy that demand the vast majority of the time in both regions—about 95 percent of those in Atlanta, and 80-90 percent of those in Boston. More rigorous tests prove that this difference is statistically significant.

This seems like strong evidence that there is a “shortage of cities” in Atlanta. Why, otherwise, would there be such a gap between the number of people who satisfy their preferences for urban neighborhoods in the Boston and Atlanta metro areas—and much smaller gaps between people who can satisfy their preferences for more car-oriented areas?

If this is correct, it helps explain a other issues we see. If urban neighborhoods are undersupplied compared to demand for them, we would expect to see urban housing go to the people willing to outbid other households, increasing prices relative to auto-oriented neighborhoods, which are more plentiful. In a place like Atlanta, lots of urban housing would have to be built before this bidding war could be ended, returning prices to a “normal” market level.

It’s also notable that this kind of “shortage of cities” can occur even where there is no overall housing shortage. Atlanta, for example, is not a particularly high-cost region, but it has mostly added new housing on the suburban periphery. So while there’s no bidding war for housing in the metropolitan area as a whole, there is a bidding war for more urban housing, making walkable neighborhoods more expensive than they would have to be. Boston is almost the opposite: walkable neighborhoods appear to be less undersupplied relative to auto-oriented neighborhoods, but the region as a whole has very expensive housing, suggesting that the total supply of housing is too low. Boston could help bring down housing prices by building any housing at all—auto-oriented or more walkable. (Though walkable housing would have lower total location costs.)

Levine’s study ought to be known by anyone who works in urban planning or housing. It’s one of the strongest pieces of evidence that “revealed preferences”—the choices that people actually make about where to live—actually reveal the limited choices that people are given as a result of restrictive land use laws.