Shorts: centripetal force, P3s, office park retrofit

Bubble map: growth in 25-34 population, 2010-2012

Growth in 25-34 population for DC-area counties, 2010-2012, with core jurisdictions aggregated for clarity. 42% of the growth is concentrated inside “the diamond,” which is home to 18% of the region’s population. A similar trend was noted in metropolitan NYC.

1a. Another recent report (by James Hughes and Joseph Seneca, reported by my colleague Stephen Miller) confirms that population growth has overwhelmingly shifted towards the core in the New York region since 2010. Third-ring suburban counties like Dutchess and Hunterdon are actually losing population, overall suburban population growth has slowed by 80% — and the core is on pace to recoup its entire 30-year postwar population loss (1950-1980) in just one decade (2010-2020).

Also echoed: a much larger proportion of young people (20-somethings) are heading downtown. “From 1970 to 1980, suburban counties captured 96 percent of the growth in this demographic. From 2010 to 2013, that figure dropped to 56 percent, with the urban core becoming increasingly competitive.” Since the report is from Rutgers, it defines “core jurisdictions” as including Hudson, Bergen, and Essex counties.

We’ve seen these trends reported in a national (PDF) and even DC context, but it’s worth noting since the NYC area is so huge and its economy has done well since 2010.

2. This week, Streetsblog USA will be publishing a three-part article I co-wrote with Angie Schmitt about what befell the Indiana Toll Road P3. Bad timing and bad traffic projections were just the tip of that iceberg.

3. Jonathan O’Connell from the Post reported last week that a new Fairfax County school was repurposed from an abandoned office building.

It’s a clever response to the changing Skyline/Bailey’s Crossroads neighborhood, where an aging housing stock and distance from transit have created a Toronto-esque high-density suburban immigrant enclave. Larger immigrant families have pushed up population density and school enrollment; retail is thriving, judging from the continued traffic jams and low vacancy rates. At the same time, office users have engaged in a “flight to quality,” towards newer and more efficient buildings closer to transit.

What’s more, the cost was very competitive. Fairfax spent $19 million in 2014 to buy and refit the building for 800 students (granted, a few amenities are still forthcoming) — which is right around the national median cost for a greenfield grade school in 2010.

Hong Kong’s revolution is in the streets, not the skyways

Even in “the city without ground,” #UmbrellaRevolution has taken not to the ersatz quasi-public spaces floating above Central, but instead to the ground — or at least to the traffic-sewer highways that fill what little is left of the ground:

(c)2014 NextMedia

It’s an interesting contrast with Occupy Wall Street, which happened to fill privately owned public space even though New York has comparatively more truly public spaces.

Trevor Boddy’s essay about North American “skyway” systems in Variations on a Theme Park seems prescient:

Heretofore streets functioned as periodic reminders and enforcers of the civic domain; the new patterns of city building remove even this remaining vestige of public life, replacing them with an analogue, a surrogate.

Precisely because downtown streets are the last preserve of something approaching a mixing of all sectors of society, their replacement by the sealed realm overhead and underground has enormous implications for all aspects of political life. Constitutional guarantees of free speech and of freedom of association and assembly mean much less if there is literally no peopled public space to serve as a forum in which to act out these rights…

[Protest] activities have been displaced over the past decade from the square and main street to the windswept emptiness of City Hall Mall or Federal Building Plaza. To encounter a ragtag mob of protestors in such places today renders them even more pathetic, their marginality enforced by a physical displacement into so unimportant, uninhabited, and unloved a civic location.

Only a full-scale revolt, involving hundreds of thousands, can be taken seriously under these conditions.

Not that a U.S. Supreme Court case matters much in this context, but Thurgood Marshall’s concurring opinion in the Pruneyard Shopping Center case is also worth remembering (emphasis added):

[S]hopping center owners had opened their centers to the public at large, effectively replacing the State with respect to such traditional First Amendment forums as streets, sidewalks, and parks…. Rights of free expression become illusory when a State has operated in such a way as to shut off effective channels of communication.

(Image of Occupy Central on Monday, 29 September 2014 from Apple Daily/NextMedia.)

My five days on the GAP + C&O trails

GAP ride

Overlook just east of the continental divide.

Three general observations:

  • Six days would be perfect; the five-day schedule offered insufficient recovery/slack time and left me feeling rushed, despite good stretches at 15+ MPH. In my case, a rainy morning and the Fallingwater side trip ended up taking up a good chunk of day two, which left me with almost a century to accomplish on day three. Plus, the towns were more interesting than I expected.
  • For the most part, you actually might not want to stay overnight in the most interesting towns (IMO: Ohiopyle, Cumberland, Harpers Ferry). Their attractions are mostly open during the day, whereas for an overnight location you mostly just want to eat dinner and crash. Specifically, I’d recommend stopping at Harpers Ferry during the day when the historic park’s attractions are open, then overnighting at Lockhouse 28, and then a leisurely reintroduction to metropolitan civilization the next day.
  • The C&O’s surface varies tremendously. Some of the sections are, like the parts closest to DC, very rocky and tough to take at speed, but others are much smoother (and often muddier). I did appreciate taking breaks from it, though.

The daily itinerary, as it played out:

  1. Pittsburgh to Ohiopyle, Penna. 77 mi. slightly uphill via McKeesport, West Newton, Connellsville. Stopped at Target in Homestead to buy a $30 tent, which did in fact pay off. Some of the Steel Valley towns would have been worth a side trip, but the main goal was to get out of town. Highlight: Appalachian Juice Company in Connellsville.
  2. Ohiopyle to Meyersdale, Penna. 42 mi. uphill via Confluence, Rockwood. The morning trip to Fallingwater was very steep: it’s 780′ of vertical over less than five miles. By comparison, the GAP’s first hundred miles rise by about the same amount. I’d planned to make it over the divide to Frostburg, but had to stop 15 miles short once night fell — it’s really, truly dark. That said, Fallingwater is truly transcendent.
  3. Meyersdale to Hancock, Md. 93 mi. mostly downhill via eastern continental divide, Cumberland, Paw Paw. Detour onto Western Maryland Rail-Trail. Highlight: dinner at Buddy Lou’s in Hancock.
  4. Hancock to Sandy Hook, Md. (Harpers Ferry) 48 mi. flat via Williamsport, Sharpsburg. Detoured over land from Williamsport through Antietam Battlefield. This cut out about 16 miles by keeping us away from a particularly windy stretch of the Potomac; plus, the main uphill was from the valley up to Williamsport, where we were stopping for lunch anyways. Highlights: lunch at Desert Rose Cafe in Williamsport and Nutter’s Ice Cream in Sharpsburg. However, we did end up bypassing Sheperdstown, W.V., a town that other riders commended.
  5. Sandy Hook to Reston, Va. 45 mi. via Leesburg, switching to W&OD via White’s Ferry. Took the Silver Line to Washington Union Station from there. Highlight: ultra-smooth cold-brewed Hopscotch Coffee. Although this was the only day that passed by any breweries (Crooked Run, Old Ox, Lost Rhino, Mad Fox, Bluejacket, etc.), we were in too much of a hurry for any stops.

Parallel trends: bigger houses, tinier apartments

size of new houses

It’s lately become very fashionable to contrast the concurrent trend towards ever-larger new houses and ever-smaller new apartments. Recently, Richard Florida,* Kaid Benfield, and even the Streetsblog podcast have all mentioned this peculiar contradiction.

But scratch a little bit at the surface and you’ll find that it’s an artifact of the Great Recession, not an existential conundrum. Consider that the Census’ Survey of Construction, the source of said statistic, only measures a universe consisting of new single family (usually for-sale) houses. However, these share of total new housing construction that is in single-family houses continues to fall. Just from 2011 to 2013, multi-family houses went from 29% to 33% of all new housing completions. Meanwhile, their size hasn’t budged: it’s consistently been around 1,100 square feet (+/- 50) since 1999.

The chart above also comes from Census data, and tells a more complete story. The average size of new houses built in the USA, both single- and multi-family, declined from 2011 to 2012, then ticked up to a record in 2013. It hasn’t been on an upward march forever. Another reason why this data point isn’t quite as important as it might seem at first glance: many fewer Americans are in new houses, period.

houses creeping back up

Construction may be up slightly, but it’s still well below its peak. Completions fell 70% from the 2004-2006 rate, and now we’re back up to 60% below that (completely unsustainable) rate. From builders I’ve heard at ULI panels and the like, new move-up product is still selling in the suburbs, but first-time buyers are staying in rentals longer, largely due to tighter financing. Hence, fewer small homes are getting built, as younger households stay in existing stock for much longer, and the average home size is increasing. (Oh, and remember: we’re talking about fewer smaller single-family houses. The number of multi-family houses, small and medium-sized, is increasing.)

Thus, the real story is that the for-sale single-family starter home, as a product type, is dying off. For example, when my brother recently purchased his first new house (at 35, considerably older than the typical new buyer at, say, Levittown), he didn’t get a starter home. He went directly from an existing multi-family unit to a move-up sized house.

* Making the additional methodological error of taking the definition of “central cities” seriously.

LocationAffordability’s denominator problem

Location Affordability: numerator problem

HUD’s LocationAffordability.info, which maps CNT’s H+T Index ([housing + transportation costs]/income), shows a sharp affordability divide slicing across Maryland’s suburban heartland, from the northwest to the southeast. This particular example emerged while I was researching yesterday’s Streetsblog USA post about the Citizens Budget Commission’s report citing DC, SF, and NYC as being rather affordable from an H+T perspective — but using some slightly odd figures.

Here, I’ve highlighted Crofton, Maryland, just inside Anne Arundel County, with apparently more affordable Bowie just inside Prince George’s County next door. Curiously, though, costs are $700 higher in Bowie, but it’s still deemed “more affordable.” How? Prince George’s is within metropolitan Washington, where median households have incomes $10,000 higher than in metropolitan Baltimore, which includes Anne Arundel.

The added irony: median household incomes in Anne Arundel (again, metro Baltimore) are quite high: $14,000 higher than in Prince George’s. Similarly, Howard County (Clarksville, Columbia) is Maryland’s wealthiest, with median household incomes $12,000 higher than in neighboring Montgomery (Olney, Gaithersburg). Although Anne Arundel and Howard derive much of their income from their legions of Washington commuters, they do border Baltimore County and the Maryland state legislature has designated them as part of metro Baltimore. Thus, for the H+T Index’s purpose, they’re comparatively deprived.

The H+T Index is a useful tool, but it does rely on a lot of moving data points. In this case, it’s a bit strange that the numerator (H+T costs) are defined by a very small geography (census tracts in the map, cities in the CBC report), but the denominator (income) is related to entire metro areas. While it’s true that labor markets are metropolitan in scale, it’s also true that incomes vary locally just as much as housing and transportation costs do.

Building type survey: stacked flats in Norfolk

A quick trip to Norfolk last weekend turned up at least one pleasant surprise: a tradition of Chicago-esque stacked flats apartment buildings, here in the rowhouse-heavy Mid-Atlantic. Many were around the Ghent neighborhood, primarily along higher-traffic (probably former streetcar route) streets like Colonial Ave. and Hampton Blvd.

The neighborhood’s lots are a fairly generous 30′ wide, so a double lot can easily fit narrow courtyards between three-story stacked flats:

Ghent, Norfolk

More common, though, were double-lot six-flats — with deep neoclassical porches, reflecting the fact that they are, after all, in the South:

Ghent, Norfolk

Just as surprising was this liner apartment building, apparently built in 2006 at the corner of Colonial and Princess Anne to mask a 1970s-era serrated senior housing high-rise. The porch detailing is a bit clumsily done — the building is too wide and shallow to match to the six-flats’ columns across the street — but the building holds a busy corner much better than whatever parking lot or open lawn that preceded it. Similar proposals have been controversial even in New York City, so it’s heartening to see a fairly good example.

Ghent, Norfolk

The Chicago six-flat is a particular adaptation to several factors: fairly wide (25′-28′) lots, readily available brick, and a fire code that both largely banned party walls and required two exit stairs. They’re readily identifiable from a front stair off to the side, leaving space for a spacious “front room,” and exposed “back porches.” True to form, the Norfolk houses also had exposed rear exit stairs, even in the absence of alleys.

Alas, the city’s pattern books don’t have much to say about the type.

Why inclusionary zoning has a cash-out provision

Daniel Kay Hertz has a recent post about how Chicago’s inclusionary zoning (IZ) policy is insufficient at creating enough units to meet Chicago’s affordable housing needs.

Montgomery Ward Complex

Some of the loft condominiums within the former Montgomery Ward Catalog House, where one penthouse unit sold last October for $2.95 million, were set aside as public housing replacement units.

When I was working for the Chicago Rehab Network 11 years ago, I wrote up the broad outlines of what was eventually adopted as Chicago’s IZ policy. I certainly concur that it is not going to solve the affordability crisis in Chicago anytime soon, but I still think it’s a reasonable approach to providing workforce-level affordable housing within the context of how Chicago builds housing — and once it was implemented, IZ multiplied the number of affordable units that Chicago’s Department of Housing could take credit for (primarily through LIHTC).

During the process of drafting this policy, we anticipated and understood that IZ would absolutely not be a cure-all, regardless of how future politicians would try and take credit for it. Furthermore, as Alex Block points out in a comment to the post, IZ absolutely does attempt to do two, contradictory things: (1) integrate gentrifying neighborhoods by creating new, permanently affordable units and (2) creating as many units as possible.

Since CRN is a coalition of CDCs, almost all of whom work exclusively in poor neighborhoods, the CDCs stood to benefit more from approach #2, and so the law probably errs in that favor. Even CRN’s members who worked in fast-gentrifying neighborhoods, though, would rather have served two families in Oakland than one in the South Loop, and the cash-out provision allows them to do so. I certainly don’t blame them, even if the net result does to a small extent perpetuate socioeconomic segregation.

As part of the process of creating this legislation, we conferred with developers of both low-rise and high-rise units, who shared their pro formas with us, and with very extensive research done by groups like MPC and BPI, mostly relying on established policies in primarily low-rise places like Montgomery County, Md. and Burlington, Vt. We saw very few examples of successful policies that worked in a high-rise context. And since a large share of the development in Chicago, then as now, was in downtown high-rises, we needed to find some way to get buy-in from high-rises.

In short, affordable units within high-rises turn out to be very difficult to create and administer. High-rises are costly to build per square foot, and there isn’t much latitude to trim the costs through things like unit sizes and finishes. Most crucially, high-rises are subject to numerous cost thresholds, beyond which the primary incentive of IZ (“free land” in the form of higher density) can become worthless — e.g., a 7-story building is actually far less profitable than a 6-story building. And once a high-rise is completed, it’s difficult to balance the operating costs of luxury amenities (concierge, pool, etc.) across market and affordable units, which has recently been in the news with the “poor door” controversy. (This is somewhat less of a problem in MoCo, since the Washington area’s very high AMI allows for luxury studio apartments to be counted as “moderately priced dwelling units.”)

So, given these difficulties — and given the CDCs’ thirst to capitalize a housing trust fund that could significantly expand their efforts at helping low-income families in neighborhoods (rather than moderate-income singles downtown), we went with the “cash-out” provision that pretty much exempts downtown high-rises.

As for exempting small developments, that’s solely related to the fact that the requirement kicks in based on the number of units, and it’s impossible to deliver a fraction of a housing unit.