A new post on Old Urbanist linking to prior posts about housing affordability, both on his own blog and on New World Economics. The theme is that various design standards – the two sites’ main scourge is streets wider than about 5-10 meters and in general excessive room for parking and front lawns – force the cost of construction up, making housing less affordable.
In reality, the first thing to note about high housing prices is that they exist everywhere: not just in new urbanist towns in the US, the type of development under discussion on the above blogs, but also in New York, and Paris, and Tokyo, and Tel Aviv, and Hong Kong, and London. In my matrix of different types of city planning, every row contains cities whose housing prices stretch the middle class to its limits. Often there’s significant homelessness, but most people have just enough to scrape by. The cities where housing prices are low compensate by either having very poor populations (inner-city Detroit) or requiring people to spend large quantities of money on driving (the Sunbelt): note how across US metro area, the total percentage of household income spent on housing and transportation is essentially constant.
Thus, as a first filter, the cities whose housing prices are low relative to incomes are very spread out and auto-oriented, exactly the opposite of any kind of urbanism other than suburbanism. As a second filter, Ed Glaeser notes that the high cost of housing in coastal cities comes from supply restrictions in the form of zoning, writing about Boston and about Manhattan as case studies.
First, what is clear about situations with unaffordable housing (really, barely-affordable) is that it is not due to high construction costs. Glaeser himself notes that construction of luxury apartments in Manhattan costs about $300 per ft^2, while the sales price per ft^2 is on average $600. In particular, parking requirements and other restrictions that effectively raise construction costs are not the primary agent to blame for high housing costs in general. An extra $20,000 for a parking spot is not going to make housing unaffordable, though it may influence developers’ decisions of what and where to build to maximize profits, in particular by making them abandon urban construction in favor of the suburbs. Glaeser blames persistently high housing prices on a regulatory tax, which forces developers to spend extra money on lobbying and preparing paperwork for permits.
Second, the primary determinant of housing prices is not capital costs, but the cost of the land underneath. An older post on Old Urbanist asks why real housing prices have increased since 1920; the answer is that a house is not a manufactured good, but primarily land, as is especially clear when one considers expensive, desirable cities.
Third, the worth of land is dependent on demand. Land on which a developer can build three apartments is worth three times as much as land on which a developer can build one apartment. That’s why on the level of the individual building, building higher does not reduce rents. Land supply only forms the limiting factor when there’s a regionwide desire to be in an area with a fixed land constraint, such as the national borders of Singapore or Monaco, or the physical extent of the New York City Subway or the walkable radius of Central Tel Aviv. In such cases, it could reduce prices to expand the available space for housing within the fixed constraint, via either increasing density or expanding the desirable area through transportation infrastructure or landfill. But otherwise, there’s not much point.
When high housing prices are genuinely the result of high capital cost, the result is different from that of high demand or a shortage of land. Consider North Tel Aviv, which mandates expensive whitewash on its traditional garden city buildings. When those buildings were first constructed in the 1930s, they were priced too steeply for the working class, leading the rising middle class to move in instead. Since the whitewash is also high-maintenance, apartments deteriorated, and the only buildings that maintain an aesthetic exterior cost much more to maintain and are only affordable to the rich. In effect, the result of high capital cost is worse physical stock, the opposite of what normally happens in Tokyo, New York, and other expensive cities.
Anti-gentrification activists often fight policies that make their areas more desirable; the above three points help explain why. Affordable housing to them is a bargain to richer people, and if they want to move in, they’ll be priced out. The only way to depress housing prices is to depress demand. One activist, a Harlem preacher with extreme right sympathies, even calls for a general economic boycott of his own neighborhood in order to cause an economic collapse and lower rents.
The inevitable conclusion, namely that it’s impossible to make housing persistently cheap without raising other costs or impoverishing people, does not mean that affordable housing issues are moot. First, the equity issue remains; although on average housing is just marginally affordable, to many people it is not affordable, and as a result, expensive cities engage in government intervention to prevent mass homelessness, even ultra-capitalist Singapore.
In addition, although expanding housing supply makes land more valuable and normally prevents prices from falling, it also create better housing in the process. Auto-oriented sprawl in the US has caused dwelling size to increase; upzoning and the construction of better transportation infrastructure in expensive cities would enable people to move from the periphery to the core – or, more precisely, people could stay where they are, but public transit could redefine regions from periphery to core.
For a toy model, suppose there are two kinds of development: regular suburbia and new urbanism, where new urbanism is more expensive. Constructing more new urbanism is going to reduce the price for both kinds of housing (new urbanism has an increase in supply, regular suburbia suffers from a subsequent decline in demand), while also shifting people from regular suburbia to new urbanism. Overall the average price of housing shouldn’t change, but the quality will increase.
In other words, on a national or regional level, affordable housing is never a problem; it may be a problem for poor people, but not in general, on average. Supply restrictions should show in low-quality housing, measured in terms of size, local walkability, aesthetics, and other factors that on the local level determine price.
An often neglected, but very important, unintended consequence of “new urbanism” or “smart growth” development paradigm is that it involves, necessarily, reducing the accessibility of most dwellers, which has immediate repercussion in prices and the overall timeline of housing stock transformation. One of the cornerstones of new urbanism school is to promote the idea that, with a mixed neighborhood (in terms of household income and service/commercial offering) , overall need for mobility would be reduced.
The “you will need to travel less” philosophy means that, by design, it would be much harder (in terms of time-based accessibility) for those living far from a place to reach businesses there for work or shopping or else. The quest for diversity wanes as the design of “new urbanism” neighborhoods reduces the reach of business and residents, further pushing the homogenization it wanted to revert and avert in first place: if it is too difficult (in terms of price and time) to get to/from there from far away locations, there will be an ever-increasing premium embedded on housing prices of the most desirable areas.
As yet another ripple effect, the poorer ones who fill the lower paid jobs are even more penalized, because a “compact-smart” city will likely mean far longer commutes for those who will not be able to afford housing close to work because it is too close to “desirable” areas. Just loot at San Francisco.
On very sprawled areas, auto-centric yet with manageable congestion (like Phoenix, Denver, Houston, Minneapolis), low-income people have greater job accessibility because they can drive (and park for free) to a wider range of jobs in more wealthy areas, which usually pays more than similar low-skilled jobs in the very low-income area where they live. If a city decides to be “non-auto-centric”, it risks, by means of reducing accessibility for the low skilled population, further add to their economic plight by cutting them off a wider access to different workplaces, unless they provide a transit network that provides a similar reach.
This is why “affordable housing” is a much serious and pressing issue in New York metro or Bay Area metro than in Los Angeles metro or Nashville metro: pollution and “wast of land in form of parking lots” aside, it is still easier for those on the bottom 20% income bracket to have relatively easy access to areas full of jobs as they can use a crappy car and drive to their place of employment 30km away, whilst a mere 15km way in a transit-(more)-centered area like NY means that, anyway, if you are on the bottom 20% you will not be able to afford homes close to good transportation (subway, LIRR, PATH) anyway – unless it is a housing project or a derelict building worse than any cheap mass-manufactured prefab suburban house.
I believe the solution to “affordable housing” issues, short of financial regulation that tackles on real estate investment bubbles (another, different world) is to provide, be it by car, train or whatever other mode, far greater accessibility on a regional area at expense of compactness and “designed to be local” approaches. That way, you can have affordable house that is far, far away but still reachable. Cheap gas + plenty of freeways is a proven path to achieve that. Express commuter trains that come from far are other, provided they form a dense network so that there are not enough middle- and upper-class households pricing out the poor ones around all suitable station TOD housing developments.
Tory Grattis, who’s not exactly the world’s most pro-transit person, did a computation about this a few years ago, comparing accessibility in Houston and New York. The end result was that the number of jobs accessible within a certain time from a point in Houston and from one in New York was about the same. The radius is much larger in Houston, but there’s a lot less job density; it’s impossible for a car-oriented area to have even medium density, because then traffic congestion would increase to the levels of Los Angeles (and Houston itself has a lot of congestion – more than New York, less than Los Angeles). It’s the access vs. mobility issue that Jarrett sometimes brings up.
Every region has favored and unfavored quarters, and people who live in the unfavored quarters have a hard time accessing the favored quarters; this is independent of what mode of transportation is dominant. Paris and Houston aren’t really different in that respect.
High housing prices are not due to high costs, but rather due to supply and demand (just like they always are).
Cities have high demand…nothing will ever change that. In fact, demand for cities is increasing steadily and predictably. But that doesn’t condemn us to high prices. We have several historical examples of exponential demand growth with simultaneous price drops. The automobile for one…and its compliment, gasoline. There is a key to price drops in the face of demand increases: increase supply!!! As supply increases people compete with their new found availability of choices. Why buy a Ford, when you can buy a Chevy? Why buy gas from the local refiner when you can buy from Standard Oil? Increased supply naturally competes for existing buyers, and that is what drops prices.
Housing has a natural disadvantage in this area…it is more difficult to build an extra apartment building than it is to build an extra car or an extra gallon of gas. But that doesn’t mean that it is impossible to increase supply. Glaeser hits the key repeatedly: Remove the restrictions to housing development (taxes, permitting, price controls, etc) and you will get more housing development.
Economics is full of paradoxes, but this isn’t one of them. It is about as intuitive as it gets. Unfortunately in the realm of housing, the people who refuse to understand economics are the ones hit with the paradoxes…as they try to force others to bear the costs of housing the poor, they directly increase the need for public expenditures to house the poor.
Danny, the problem with housing is that it relies on a support asset (land) that gets easily valued as high as it can gets according to local zoning laws. Suppose Washington, DC – a notorious case – suddenly did away with its height and floor limits. The value of apartments in a given area would be as high the potential for revenues of whatever extra floors are possible to be built there. If a neighborhood in expensive Palo Alto sudden got allowed to build 10-story buildings instead of rowhouses, the next day the market value of 2-story houses would skyrocket, because now buying the house conveys the right to build up 8 extra stories (theoretically).
Then, it is just a matter of who captures the value generated by increased legal supply of build-able area: the owners of the low-rise properties, or the government, of the developer, or something shared between them. Certain countries/cities have hefty development taxes meant to capture the value of permission to build up. Others have strong real-estate semi-public agencies that play heavily in the market – and so on.
This is somewhat true, but you have it a little misconstrued.
The value of existing buildings would go up to the value of the land if the value of the land increased to the point where it exceeded the value of the apartments as a whole. Only at that point would the value of the individual apartments increase.
Imagine you are a developer, and with your new found freedom to build 10 story buildings you have decided to go around town looking for land to buy and build on. Since it is a central city, there is no greenfield land to build on. Where do you start first? The one story buildings or the 9 story buildings?
I imagine it would be the ten story building as that would maximize the revenue (since adding a single story to a 9 story building is going to be an expensive endeavor with not much gain, while the one story can be easily torn down and rebuilt with a good deal more revenue potential).
Paulus, I was referring to which existing buildings you should start buying so you could replace them with the new 10 story buildings. A smart developer would never buy the 9 story buildings only to raze them and build 10 story buildings.
One more thing Andre…
While the price of a particular 2 story apartment may rise, due to the increased land value, the price of housing on average (which is what gentrification is about) does not depend on the value of that particular apartment.
And if a 2 story apartment’s price goes up because someone wants to buy it, raze it, and build a 10 story apartment building, then the supply has seen a net increase which, in ceteris paribus terms, lowers prices.
The main difference between cars and houses is that cars are all capital (and labor, but labor gets more efficient over time), while houses are also land. Oil is similar – and indeed, its price trend over the last 100 years of so has not been one of decline, and on the contrary periods of high growth in the world economy, such as the previous decade, were also periods of very large price increases.
And you will never be able to extricate the value of the land from the value of a house. I fully realize that.
But the value of a house/apartment does not boil down to pure land value. It is land + improvements.
A common land/improvement ratio for a suburban home is 20%…meaning that 80% of the value is not subject to the land constraint. A central city apartment building will have land prices several times that of a home…but at least with residential buildings, the land/improvement ratios are typically lower than that of suburban homes. 10% is pretty common, and I have seen some with less than 1%.
Regardless, the value of the improvements strongly overshadow the value of the land.
True. Land/improvement ratios are actually painful to measure, but when you compare buildings with adjacent vacant lots, you can often get a good sense of it. The land is rarely more than 20% of the value.
I think in a place with minimal regulation, one would expect to have low land/improvement ratios (Someone who buys a plot for a million dollars is probably going to want to spend a lot more than that on construction). However, in a place where construction is highly regulated and it is complicated to get approval for anything, I think the ratio can go a lot higher, since the land is still desirable but improvements are limited. I’m thinking of some parts of San Francisco in particular.
Alon: even if we say housing costs are determined primarily by the value of land, not of labor or materials, we’re still left with the puzzle that housing prices rose despite the enormous increase in land “supply” effectively generated by the interstate system over the past 50 years. By contrast, housing prices had remained stable or had fallen during the unprecedented crush of urbanization during the pre-automobile era (1860-1920), when demand was intense and land was very scarce by comparison. The pattern is the opposite of what one would expect from a pure land value-based analysis.
I think that is even more remarkable in light of the fact that in 1920, when housing costs were lower than today, less than 10 percent of Americans owned automobiles, whereas in the study you’ve provided, transportation costs now approach 100% of housing costs in many places, most of which, I presume, is attributable to car use, storage and maintenance. Combined housing and transportation costs have apparently risen dramatically since 1920 in real terms.
If the opening up of thousands of square miles of suburban and exurban land to development in combination with the development of labor-saving construction methods and low-cost materials did not lead to a decrease in home prices, I think it’s worth asking why.
Did transportation costs really increase since 1920? What we know is that more people own cars now, but transportation is all else being equal cheaper: cars are manufactured goods. The Model T’s price was trending down, and so was the real fare on public transportation. I don’t have figures for passenger transportation, but Glaeser claims that the real cost of freight transportation went down 95% in the 20th century. Nor do I know much about the breakdown of the price index by region, though I presume that prices in New York, at the time still more or less zoning-free (the 1916 resolution provided room for 55 million people), were not yet out of line, that is beyond the higher wages in the city.
I think you’re overinterpreting 1920 as a watershed point. What I see in the data is different: real home prices were stable from the late 1800s to about 1919, crashed during the post-WW1 bout of inflation, stayed low, and then rose again during the WW2-era boom. My guess based on the graph is that there were no price controls on owner-occupied housing in WW2, and as a result housing prices rose faster than the general rate of inflation.
I’ve seen it frequently cited that transportation costs, as a percentage of the household budget, have increased from 3 percent in 1920 to 18 percent today (roughly where they’ve been since the oil price shocks of the 1970s). That does not mean that the price of cars did not decrease over that time period, but that the increase in car ownership and in total driving over that period overwhelmed the cost savings of cheaper or more reliable cars. Savings on public transit fares probably did not affect this trend much, given the steady decline in transit use since the 1920s.
I think you’re right that I may be reading a bit too much into the chart. Still, the building cost trendline also shows a low point around 1920. Despite some wild swings the overall trend since that time has been up.
I find it hard to compare transportation costs as a percentage of household budget between 1911 and 2011 because, indisputably, people “roam around” much more now than they did in 1911, in only a part of that could be attributed to the introduction, in large scale, of the automobile.
Americans living in farms at the time essentially didn’t “commute” for most of the week altogether. People living close to bustling factories limited their lives to extremely long work hours, the occasional and limited shop for clothing or something, basic grocery shopping and church attendance.
We have dozens of normal, run-of-the-mill “activities” that simple didn’t exist in 1911, like the concept of tourism as something for the masses, gyms, spas, electronic appliance shopping.
Mobility increased much, so expenses in transportation are not suitable for direct comparison. It would be the same to compare expenses in – say – telecommunications in 1911 x 2011 and conclude the Internet, cable/satellite TV, phones (land/cell) just ate a much larger share of the domestic budget pie than the occasional snail mail and telegraph message in 1900.
Charlie, Andre: okay, let me rephrase – the question is what happened to urban transportation costs, excluding vacation trips. Right now the vast majority of household transportation spending in the US is not vacation trips: it goes primarily toward buying and maintaining cars, and secondarily toward the marginal cost of driving to work and for shopping. Of the 18% of the household budget that goes to transportation, 1% covers intercity modes plus rental cars, and an unknown amount covers fuel for intercity driving (link). The unknown amount can’t be too large – I’ll be surprised if it’s more than 1%, since 5.2% is the total amount of fuel.
If it was actually 3% in 1920, and it was not an artifact of transit fares that hadn’t been increased to match the postwar inflation, then it’s not enough to close the gap, no. But if it was an artifact of inflation, then maybe. To pull a calculation ex recto, 3% then would want to be 6%, which would be blown up to 12% when only considering urban residents.
Another question: has either of you been to the NYC Transit Museum, or another place that mentions skilled worker wages in the 1910s? I have, but I no longer remember the numbers. We could try to compute the cost directly, by comparing the wages to known transit fares from the era.
Alon: the history of NYC subway fares does suggest a sharp inflationary impact leading up to 1920 (and seeming to support your intuition that a 6% was a more accurate number than 3%, based on a comparison with where fares had been in 1910):
Overall, fares have slightly less than doubled in inflation-adjusted terms since 1910, which would put the increase behind the rate of increase in real wages over that span, but then again, in 1910 there was no income tax …
I’ve never been to the Transit Museum, but that’s a good idea for a weekend field trip.
If you have a country with a declining population, the price of housing will drop. As long as the population is rising, the price of housing will generally increase — obviously the supply will increase, bringing the price down, but the supply is limited by the supply of land, which is *permanently* limited.
So we’re back to “stabilize the population if you want housing to get cheaper”.
But… US population
growthhas always been growing. And nationally there’s no limit to the supply of land; even Marx noted that and refused to list the US with Britain and Germany as one of the countries where revolution is imminent. Of course some parts have since filled up – e.g. the Northeast – but those parts no longer have high population growth.
Update: the crossed out word should be deleted from the sentence.
There may have been a huge increase in the land “supply”, but it was matched by an increase in land demand. Not only does every household demand its own 6000 square feet, but the parking area required for every place of business, plus the road surface to connect the two, probably doubles or triples the area that a business in the pre-automotive era would have used.
Thought execise: Zoning levels in NYC (proper) multiplied by two overnight. What would the short, medium and long-term effects be?