Matt Yglesias and Lydia DePillis have been having an interesting discussion about the DC commercial real estate market that I have some thoughts on, so I thought I’d weigh in. I apologize for the length of this post, but I think it’s a really important point that shouldn’t be underestimated.
Matt started by stating the following:
Downtown DC is full. There’s basically no land left to build on, and you’re not allowed to build higher. If you make it a more attractive place to locate jobs, no additional jobs will be created because there’s noplace to put the jobs. The improved quality will show up as higher rent for landlords, and our rents are already the highest in the nation. If you relaxed the height limit, the high rents would spur new construction (=jobs) which would lead to lower rent per square foot which would make downtown, DC a more attractive employment destination.
Lydia agreed that the height restriction should be lifted (I don’t want anyone to think that Lydia is an apologist for this – she’s definitely not, and if given total discretion over DC land use, I think all three of us would implement very similar policies), but argued that Matt is downplaying growth possibilities outside the core of DC’s downtown:
But to say that “there’s noplace left to put jobs” is simplistic. Although many office projects stalled during the recession, they’re starting up again in a big way around the city, from Mount Vernon Square to Anacostia. On the longer term horizon, massive office capacity is planned for McMillan, L’Enfant Plaza, and the Capitol Riverfront. Recent changes in who gets what at Walter Reed – the District may now get all of the Georgia Avenue frontage – has Office of Planning director Harriet Tregoning thinking about “more ambitious uses” like a “major employment center.” The list goes on. So yes, rents are high, but jobs are still coming, and there’s plenty of space to put them–in places that could really use the lift.
I think I’d come down more on Matt’s side. While there is definitely room for growth near the core (for example, L’Enfant Plaza and Mt. Vernon Square, or NoMa), there isn’t that much more room. DC’s office rents have already surpassed New York City’s to be the highest in the nation – clearly DC has a very, very serious problem when it comes to commercial real estate affordability, which has a huge impact on job creation.
Which brings me to my main point: All those outlying districts. The DC metro area has put a lot of stock in density in neighborhoods far from downtown. Lydia cites a few within the District – McMillan, Anacostia, the Capitol Riverfront, Walter Reed – and then there are huge projects in Northern Virginia like Tysons Corner and the Orange Line corridor. But frankly, I’m not convinced that it’s these sorts of “edge downtowns” are realistic long-term solutions.
To see why, you have to understand why downtowns developed in the center of metropolitan areas to begin with. Geometrically, the center of a circle will always be the point that’s closest to the average of all other points within the circle. As a city sprawls outwards, the average difficulty of going from one random point to another grows exponentially, while the difficulty of commuting directly to the center increases linearly. (I’m not the best at explaining mathematical concepts, so if someone could formalize this in the comments, it might help others understand.) Transit entrepreneurs during America’s huge urban boom around the turn of the last century understood this well, as every elevated line or streetcar route was designed to funnel people in and out of downtown. In fact, as I learned from Robert Fogelson’s aptly-titled Downtown, early 20th century city dwellers fully understood and accepted the need for commercial space to be in the center and residential on the outskirts – many people in 1910 believed (and accepted) that by the year 2000, Manhattan would be largely devoid of residential neighborhoods. To use the example of New York City, while many (most?) of the private transit lines, for efficiency reasons, started in one outer borough and ended in another, they all (/almost all?) made a beeline straight for Lower and Midtown Manhattan. Nobody was seriously expected to ride, for example, the 4 train from the Bronx to downtown Brooklyn. Yes, there was a downtown Brooklyn (and a downtown Queens, and a downtown Bronx), but it was a secondary job center at best, and I doubt anybody believed it would ever host the variety of jobs that Lower and Midtown Manhattan did (and still do).
While mass transit technology has improved (though not nearly as much as it would have had we not regulated the private companies out of existence and replaced them with sclerotic publicly-managed shitholes), I doubt it will ever get to the point where extra density downtown is not the market equilibrium. For example, in DC, even when the Silver and Purple Lines are completely built out, it would be very difficult, if not impossible, for someone living in Maryland on the Purple Line to commute to work every day in Tysons Corner. The trip, like one to the already built-up Orange Line corridor, would involve transferring to the overcrowded Red Line, which you’d then have to taken into the city and change at Metro Center for the Silver or Orange Line, and then ride that a few stops into Virginia. Without even counting the time it would take to walk to the station from your house or to work from the station, the travel time could easily reach an hour each way – and that’s for peak trips when the trains are running with short headways. For those of you outside the DC bubble, imagine commuting by rail from Bryn Mawr to Cherry Hill in the Philly region, or from North Jersey to Long Island City in the NYC metro area, or from Oakland to San Jose in the Bay Area (California geography isn’t my strong suit – maybe that one’s more doable?).
Now some might respond by saying that such commutes are not necessary because people can work closer to home. Those in Jersey ‘burbs can work in Jersey City, those in Maryland can work in New Carrollton, etc. But then that makes you wonder, what’s the point of everyone living in one metropolitan area anyway? Conservatives and pro-sprawl libertarians might point to the relatively small amount of people who commute to city cores every day – even in very transit-accessible places like Philadelphia and New York – but I think they’re discounting the impact of the people who do. They tend to be very high wage-earners who support a lot of commerce in their local communities. Without access to lucrative jobs in Midtown and Lower Manhattan, would Scarsdale still be able to support all its local jobs? Without the R5 bringing people from Philadelphia’s Main Line into Center City, would all those local jobs still exist? Probably not.
I think it’s also worth pointing out that the vast majority, if not all, cities in capitalist countries have a wedding-cake style design. The ones that don’t are largely cities built by communist dictatorships – cities like Moscow, Beijing, and Pyongyang. For more on this, see this post, which excerpts from an interesting article about land use in socialist economies. Even Paris has a very high core density, which it achieves through large amounts of six-story structures, few parking lots, and narrow streets – all things that DC lacks in key close-in neighborhoods. (Looking out the window from Reason’s office a couple blocks north of Dupont Circle – one of the busiest Metro stops in the region – I don’t see a single building that rises to six stories.)
Now, this isn’t to say that Lydia and DC’s planners are wrong to argue for density where they can – after all, it’s hard to imagine DC actually lifting its height restriction in the near future. Having rail-based edge cities along the Anacostia waterfront or Tysons Corner is definitely better than the status quo – I have no doubt about that. But I don’t think they’ll ever be a very good replacement for extremely high-density, easily-accessible downtowns at the exact center of metro areas.
(…by the way, this post is way too long for me to thoroughly proofread. Apologies for the inevitable typos and bad writing.)
I think this is essentially correct. The outlying downtown areas (NoMA, Capitol Riverfront, R-B corridor, Crystal City, etc) will and should develop, and develop densely. The downtown – the true core – should be even more dense than it is now.
(one quick reminder that density and height are not synonyms – it is only in areas like downtown DC with 100% lot coverage that they are essentially synonymous)
Capping downtown DC only changes the kinds of uses that would otherwise occupy the dense, downtown-adjacent areas. They become extensions of downtown in terms of the kinds of firms that look to locate there (and the kinds of rents charged), instead of their more ‘natural’ role in offering back-office functions for firms, providing cheaper office space for start-ups and whatnot, etc. There’s an economic cost for essentially regulating that mid-level price point out of existence.
Adding more density to downtown DC won’t necessarily bring rents down in the core of the city, but it would better physically match the ‘natural’ economic ‘shape’ of the city. I think it would also open up the market for other uses in the downtown-adjacent areas, such as dense residential (instead of chasing the top dollar of Class-A office space).
See as how the “productive” output of D.C. is Lobbying I want that city to be as inefficient as possible, I say institute a one story height limit on the whole MSA.
For every other city, you are right. The Downtown of normal city are generally located in a place that is most productive due to some type of natural resource/trade connection. Even if that is no longer true for specific cities (alot of cities now that movement of goods is so cheap Gleaser & Kohlhase 2003) agglomeration economies are thought to be non-linear and thus in productivity/output terms it is better to have fewer large employment centers as opposed to more small employment centers. In the Theoretical employment center models the constraint on employment concentration are the non-linear congestion costs->as you have more and more people trying to go to the same place eventually the marginal worker adds more in congestion costs than in agglomeration economies. A binding artificial limit to the growth of an employment center will cause the MSA to not be able to capture all the agglomeration gains to concentration available.
I think you are essentially correct, but there’s one historical fact you get wrong: when the NYC subway was built, Downtown Brooklyn was a very large CBD, in fact larger than today relative to Manhattan. Nobody seriously expected people to live in the Bronx and work in Brooklyn, but people did expect Brooklynites to work in Downtown Brooklyn, taking advantage of the large network of trolleys and els.
The subway was built around Downtown Brooklyn and not just Manhattan. The lines all go through Downtown Brooklyn, with the exception of the L. In addition, the BRT built a few loops going from Downtown Brooklyn to Lower Manhattan and then back to Brooklyn, which have since been rerouted as Midtown eclipsed Brooklyn as a job center.
Nowadays, the problem of traveling from a community to one side of the primary CBD to one on the opposite side is acute, on both transit and highways. As a result, the poorest slice of suburbs will usually be the one on the opposite side of the favored quarter and the dominant edge cities; in Washington, this means PG County, which is opposite Tyson’s Corner and poorer than the white-majority DC suburbs.
“I think it’s also worth pointing out that the vast majority, if not all, cities in capitalist countries have a wedding-cake style design.”
For the long history of cities, they’re been characterized by high-value, high-densities cores surrounded by progressively less dense and cheaper suburbs. (In older days, “suburb” was a synonym for slum, and this is still true in most of the world’s cities.) There’s always a certain cachet with keeping a villa or estate in the exurban “country,” but this is usually well away from both the city center and the residential suburbs, and understood to be a place of leisure not production.
In a certain sense, this is true even for central economies — after all, most of North Korea’s resources are poured into the maintenance of their “model” city, where most of the military elite work and perhaps live (yes they’re less dense and underbuilt, but it doesn’t follow that their “value” is in their underbuiltness). This was also true for ancient non-commercial or ritual cities like Tenochtitlan: the elite classes all lived near the city center.
It’s only in the last century and in a few parts of the world that we’ve dared to buck that trend, and I wouldn’t bet long odds on the stability of the alternative.
Rush hour traffic in Northern Virginia is a clear example of the exponential increases in commuting times when job centers are located away from urban cores. Fairfax County’s tax dollars would be better spent lobbying to abolish DC’s height limit than redeveloping Tysons Corner.
There is only no place left to put jobs in downtown Washington if one assumes that the existing zoning designations in neighborhoods such as DuPont, Capitol Hill and others remain in place indefinitely. There are large swaths of high-density residential areas off limits to commercial use in numerous locations adjacent to the central business district. Or put another way: is it more likely that these neighborhoods, or parts of them, would be rezoned for mixed use, or that the height act would be substantially amended for the CBD?
I’d actually say the height act, to be honest. NIMBY power tends to be strongest closest to residential neighborhoods, and Downtown is a lot farther from residential neighborhoods like Dupont and Capitol Hill than…Dupont and Capitol Hill.
That might be the best Market Urbanism comment ever!
Rezoning for commercial in an already-dense urban area doesn’t necessarily raise the same NIMBY issues as, say, a rezoning for multifamily housing in a suburb. If anything, it would provide a considerable windfall to property owners, whose property immediately jumps in value. And at least it’s in the power of the city to change.
Rezoning for commercial in an already-dense urban area doesn’t necessarily raise the same NIMBY issues as, say, a rezoning for multifamily housing in a suburb. If anything, it would provide a considerable windfall to property owners, whose property immediately jumps in value. And at least it’s in the power of the city to change.
There is no extent mass transit system in the world that could move a metro area the size of Washington DC that is private. I doubt a case could even be made for one, please elucidate your contention,
“While mass transit technology has improved (though not nearly as much as it would have had we not regulated the private companies out of existence and replaced them with sclerotic publicly-managed shitholes).”
If DC can preserve its somewhat unique character – and maintain a high density in the center – by filling in the gaps you allude to, why not encourage that over the pattern of alternating towers and gaps seen in every other American downtown? Yeah, such an approach is not exactly “market-oriented” but DC is one of the few American cities that can make a plausible case for interfering in the market in this manner, largely because it was designed as a ceremonial government center, not a center of commerce.
The trend now almost everywhere except the US is privatization, at least of the operation of the vehicles. The main benefit is to lower the cost of such operations by removing the power that public-sector unions have amassed for themselves over the years, which has resulted in grossly outsized salaries and benefits in comparison to the private sector – costs which are simply not sustainable in the long run as we are beginning to see here in America as cities and states across the country are slowly admitting how broke they are.
For mainline rail, sure. For urban transit, not really. Tokyo Metro and the MTR are being successfully privatized, but Japanese subways other than Tokyo Metro and a few lines owned by the private railroads remain municipal; in Europe, the last attempt to privatize, in London, ended in disaster and the public sector had to take over again at a loss.
(On another note, privatization means two separate things in Europe and most of Asia. In Europe, rail privatization means open access. In Japan, it means going back to the private systems used in pre-Amtrak America – i.e. the tracks are privatized and companies run trains on their own tracks, with through-service to other companies’ tracks baesd on bilateral trackage rights agreements. In Asia ex-Japan, mainline rail is public, but subways can be private or public; if a subway is operated by multiple operators, like in Singapore or Shanghai, transferring from one company to another will be seamless, unlike in Japan.)
How can it be exponentially harder to go from one random point to another but only linearly harder to go to the centre?
If it’s only getting linearly harder to go to the centre then going to the centre then back out will also be linear?
Counterpoint to a city with its CBD in the centre: New York. Manhattan isn’t remotely in the centre of the the city.
But it is at the center of its transportation network, which is what really matters here.
I think what you mean to say is that if an urban area spreads like a circle, than the area (= pi * r^2) will be the square of the radius. So, if you’re traveling randomly in that area, your trip will grow with the square of the radius. But if you are going to the center of the area, then your trip will be along the radius.
But these days, there are two sorts of “downtown” jobs. One is the “old” downtown, which is typically comprised of government workers, non-profits, and the like. This downtown is what used to be the main downtown, which the high-value “Midtown” jobs have long-deserted. In fact, it’s curious that you use Lower Manhattan as an example of a place with “lucrative jobs”. They’re lucrative if you count the DMV, the DOT, the MTA, and so forth, Otherwise, the financial firms have largely left; a walk down Wall Street will pass more residential conversions than anything else.
Try this: Go to Google Maps and a do search for “financial” in the Wall St zip code, 10004. Yes, you will see many results. Now zoom out a bit and scroll to Midtown, and you’ll see how tiny the downtown “financial” district is. A strong possibility for this is that the people in finance are pretty likely going to live in Westchester, New Jersey, and Long Island — for which Midtown is much more convenient.
New York is an exception, though. In other cities, these sort of businesses did not leave downtown for midtown; they entirely left the city itself. But they left for the same reasons that created the exodus in Manhattan.
That is: In almost every city, the new downtowns have shifted their locations (generally to the north and the west) and offer jobs that are not intended for people living everywhere. Instead (as in Chicago, Dallas, Atlanta, Detroit, Memphis, etc., etc., etc.) these new downtowns are quite convenient to their workers — who generally live nearby. Yes, the janitors have to sometimes have to travel a bit — but a good deal of the janitors’ neighbors (except for the civil servants) aren’t working in any downtown anyway. The construction workers, home health attendants, supermarket employees, etc. work where people live — so their “downtown commute” is not relevant.
Perhaps the “wedding cake” still exists, but it’s no longer in the geographic center — especially if you examine the density of high-value jobs as opposed to population densities. (And of course, in cities like Detroit and Cleveland, the “business district” is neither the center of population, high-value jobs, low-value-jobs, or even the geographic center.)
“I think it’s also worth pointing out that the vast majority, if not all,
cities in capitalist countries have a wedding-cake style design. The
ones that don’t are largely cities built by communist dictatorships –
cities like Moscow, Beijing, and Pyongyang.”
This is really offensive. You should have included DC and Arlington in that list.