Seattle Bubble, a real-estate blog founded during the inflation of the housing bubble, has an interesting post showing how Seattle proper stacks up to the 100 largest cities in the US. While I don’t think you should really pay attention to the density related information, for reasons explained here, I found the other rankings interesting although not surprising. Out of the 100 largest cities in the US here is how Seattle stacks up.

  • Population: #23
  • Square Miles: #62
  • Density: #15
  • Household Size: #100
  • % with Children: #97
  • Home Ownership Rate: #69

Seattle has long had many demographic factors that favor smaller household sizes and low percentages of households with children, and this information completely jives with the data I reported on earlier showing that a majority of Seattle housing is multifamily already. My take away looking at just these two measures, household size and percent households with children, Seattle has the strongest trends that favor smaller apartment housing than just about every other major city. Go to Seattle Bubble to see all the data.

55 Replies to “Seattle Bubble: How Seattle Stacks Up”

  1. One thing that is unfortunate is the difficulty building apartments with three or four bedrooms due to parking requiements. This tends to shrink household size.

    1. Yeah so this data is just for the city of Seattle. I wonder how the financial costs/limitations of building in Seattle skew demographics and how that skew is different than in other cities. You always get some weirdness with the data because of the impact that city boundaries have on this data.

    2. My understanding is that floorplate is more of a limiting factor in 3+ bedroom apartments (kitchens and bathrooms don’t have to have windows, but bedrooms do) than parking. But I could be wrong.

      Anyway, most townhouses are 3+BR, as are most single-family houses — it’s not clear to me that Seattle has a lot of demand for 3BR apartments at the moment. Much as I’d like to live in one.

    1. Yeah interesting.

      Does anyone know if developers are able to write off their principal interests payments on multifamily housing units like individuals are? I have been thinking about how writing off of debt favors ownership of individuals which favors single family housing because of the scale of debt an individual can take on, while housing built and owned by a developer favors larger multi-familiy housing.

      1. Owning and operating a multifamily (or for that matter a single family rental house) is a business and business expenses, including debt servicing, are deductible.

        On the other side, only single family occupant owners who itemize their deductions write off home loan interest.

        Concentration of residential ownership in the hands of a few may have unintended consequences.

      2. Landlords have a much bigger deduction than home owners. Not only can they write off the interest on debt but they can write off everything related to maintenance. A home owner for instance can’t write off the cost of a new water heater or fixing the roof. They also get to depreciate the structure. Although when sold that’s usually captured in recouped depreciation (i.e. you don’t get around all capital gains).

      3. You’ve correctly identified a distortionary policy in favor of rented apartment building. The mortgage interest deduction, and the massive subsidies to the mortgage market in the form of Fannie and Freddie, have a distortionary effect in the other direction. There are other such policies, too. Which is stronger? It’s an interesting question.

    2. That’s a pretty strong correlation. Part of this must be from demand. Desire => expensive and desire => density.

      1. Sounds right to me. Expensive housing doesn’t always lead to rentals, but I’d buy it’s a strong factor.

        There might also be a multifamily unit effect. It seems like it would be easier to rent out an apartment in a dense area.

    3. I’m curious, what are the two outliers at around 11,000 people/sq-mi? One is around 22% and the other is near 55%. It seems that the linearity really falls off to the right of this point. In fact, just visually it looks like a logarithmic curve might be a better overall fit with the flat portion approaching the 30% line and the 45 degree tangent being about 4,000 per sq-mi and 55% home ownership. For the entire US Home ownership rate, 2006-2010 66.6%. Washington State is 64.8%. A straight line interpolation never gets to 60%.

      1. Philadelphia is the high ownership and Newark is the low ownership at the 11,000 mark.

        I stuck with a linear trend because I don’t have a theory to match the curve to. A 2nd order polonomial curve fits best (R2=.475), though an exponential curve fits well too (.468). But both of those hit at around .6 as well. Logarithmic isn’t bad (.426), but looks a bit goofy at the higher and lower density. I assume there’s more than one effect occuring.

        Keep in mind this is comparing our 100 largest cities, which could act differently than the suburbs/exurbs.

      2. Homeownership rate in Pennsylvania is high overall, 70%. The median income to median price is way up there (.27) thanks to a low low median value of owner-occupied housing units, 2006-2010 $135,200! I guess there are lots of affordable old houses in a city founded in 1682 ;-) New Jersey home ownership is pretty good, 66.9% Newark has a similar median household income to Philly but the median housing cost is more than double! Income/cost ration is an aweful .12 which is almost identical to Seattle’s .13 Newark really sucks in per capita income at $17,367. Seattle’s is $40,868 and Philly is $21,117. The mystery for me isn’t so much that home ownership in Newark is so low, they’re to poor to buy a doll house. But why in the heck are prices so damn high? Is it because many of the owners are people commuting to New York? Still odd though because as we see in Seattle you’d expect the affordability index increase in the “suburbs”. The affordability ratio for NYC is just a shade under .1 but the ownership rate is 33%.

    4. Another major factor is the warranty period for condominiums. Unless the law changed drastically in the last five years (happens to me all the time!), condominium builders are on the hook for construction defects on new buildings. They are not on the hook for existing problems in buildings that have been converted from apartments.

      So you see many buildings that are built as apartments, and then flipped over (eg, Empress=>Tobiro on 5th and Washington) once the warranty period expires. Bam, dodged millions of dollars fixing any potentially improperly installed siding, structural supports, or roof!

      Condos have short-term operating costs as well — the builder covers costs of the condo association and property management for some period, I think it’s a year after the last sale perhaps? Might as well just pay for apartment management.

  2. Surely you are going to mention that Seattle is already #15 on the list when it comes to density.

    Given that, is it fair to say that it should already have seen the “benefits” of density long ago…and that pushing for “More density” when it’s already highly ranked is a very specious argument at best? It certainly hasn’t helped Newark (or Jersey City for that matter).

    1. Come on, John, when are you going to quit comparing Seattle to Jersey, for heaven’s sake? Hehe. You must be hot for Snooki.

    2. I’ve said it a million times, and I’ll say it again: density numbers that are based on municipal boundaries, or that are not adjusted for population, are worse than useless.

      Using the metric cited in this article, if everyone in Seattle moved into a single apartment building, its “density” would not change. Doesn’t that tell you something?

      Using population-weighted density of metropolitan areas, Seattle had a density of 4,747 people per square mile in 2000. Portland was 4,383. In contrast, New York was 33,029. Other cities that are denser than us include LA, San Jose, Phoenix, and even Detroit (though I’m sure that last one has changed in the past 10 years :D).

      More up-to-date numbers are available in Ryan Avent’s book, “The Gated City”, but I’m too lazy to look them up right now.

      Using bad metrics is dangerous, because it can lead you to entirely incorrect conclusions. The fact is, the Seattle metro area just isn’t that dense. We have pockets of density, but you can hardly say that we’re a “dense city” when we’re less dense than Phoenix!

      1. Yes, the it seems like the inescapable message from even the numbers you present, and the comparisons with those Dutch “bicycle friendly cities” is that Seattle is already at prime density for a transit oriented city.

        So, constantly chirping about density serves no better purpose.

        To me, it’s just away of moving the goal posts. Transit has been given billions and has to deliver based on what it promised for the given situation…not fail and then ask the whole world to be built around it!

      2. So, constantly chirping about density serves no better purpose.

        Er, then why do you keep on chirping about it?

        No matter what the subject you seem to feel the need to chime in with some bizarre anti-density rant about how great Yakima is or something.

        Give it a rest already.

  3. It looks like these numbers are for the city boundary. In other words, Seattle proper (not Shoreline, etc.). To a certain extent, the numbers become meaningless. If the city founders had drawn the borders like the folks in Anchorage, then the numbers would look really different.

    Also, if you compare inside the city to outside the city, then your argument could easily be flipped on its head. I don’t see the numbers for Shoreline, Bellevue, etc., but my guess is that they have lots more kids than do the people in Seattle. You could argue that this is just personal preference, but you could also say it is because Seattle (proper) doesn’t have enough affordable places for families (I know people who live in the suburbs, but bus their kids into Seattle). Assuming (as you did) that individuals are attracted to apartments while families are attracted to houses, then allowing more densely packed houses might help the situation. More townhouses, smaller lots and smaller houses in the outer fringes of the city (as well as the inner suburbs) would help things immensely. We have already seen plenty of that, but we could see a lot more. In my neighborhood, which is part of that outer fringe (Pinehurst) I’ve seen apartment buildings (where they are allowed) along with big houses go up in the last ten years. I’ve seen very few town houses or skinny houses in my neighborhood. I think this is a shame, as I think it would be a very popular way to increase density and help reduce the cost of housing.

    Oh, and changing the stupid laws against mother-in-law apartments would also go a really long way towards achieving those goals as well.

    1. Yes completely agree about the boundary issues. It’s always an issue with these kinds of ratios. While I have seen alternative approaches to measuring density that get around this issue, I haven’t seen alternative approaches to measuring demographic data like this.

      1. Kirkland has had them for years and they’re fairly successful. Lots of new houses, even in the upscale neighborhood West of Market had a large number of houses with these. Great for providing affordable housing for younger people and provide income to the owner of the house.

      2. Yeah, good point. But notice that I didn’t say “make mother-in-law apartments legal.” Those restrictions look awfully onerous. The lot has to be big, and the unit has to be small. It looks like it requires off street parking for the unit unless it is “in designated urban villages and urban centers”. Entrances have to be by a street or alley, etc.

        Basically, they make it difficult to appease the folks that don’t want to see lots of new neighbors, with their cars clogging the streets, etc. This is a reasonable request, but it ends up backfiring. Demand for housing increases, and more big apartment buildings are built (even if they aren’t next door to the house). People rent out their house, which can lead to lots and lots of cars (five adults in the house instead of a couple in the main house and one tenant). As with a lot of regulation, it only makes it worse for everyone.

        The whole process should be streamlined and encouraged. This is a win-win situation for both folks. The owner of the house can more easily make ends meet, while the tenant gets a cheaper place to live.

      3. Thank god for the tough codes, or you’d have shacks in too many places.
        Another factor is that many people with a single family home aren’t interested in sharing their home or lot with strangers. It isn’t all about the extra income they provide. ADU’s and DADU’s are surprisingly easy. You just have to set your mind to the process.

      4. Backyard cottages certainly increase density, but I wonder how many “cottagers” are drivers, given that cottages must be built in low-medium density residential areas. Not excatly TOD, I suspect.

      5. Exactly, it’s the peanut butter approach to density. Allowing DADUs city wide is an abdication of the responsibility of government to manage growth efficiently. The good news is that hardly any have been built. The bad news is that it can really suck for the immediate neighbors.

      6. “…I wonder how many “cottagers” are drivers, given that cottages must be built in low-medium density residential areas. Not excatly TOD, I suspect.”

        Of course cottages are built in low and medium density areas; back yards in high-density areas aren’t large enough!

    2. Shoreline, Bellevue, etc., but my guess is that they have lots more kids than do the people in Seattle.Don’t know about children but I believe the numbers for household size are 2.0 for Seattle and 2.4 for Bellevue. Persons under 18, 21.2% for Bellevue and 15.4% for Seattle. Shoreline 19.1% under 18 and 2.4 persons per household. Home ownership rate vs median price is a direct correlation. Seattle 48.9% @ $456,200, Shoreline 68.2% @ $370,400, Belleuve 58% @ $556,500. Household income/home price; Seattle .13, Shoreline .18, Bellevue .14 so I guess that partially explains it.

    3. They bus their kids into Seattle Public Schools? Plenty of folks in Seattle who private school their kids due to the terrible shape of Seattle Schools – particularly around new transit.

      1. While the Seattle Schools aren’t THAT bad, I believe sending kids from Seattle to districts outside the city is more common.

        The only two guesses I have is perhaps one of the districts to the south is even worse than Seattle or there is some unique program in Seattle that attracts a few students from outside the district.

      2. It’s difficult to send your kid to a different public school inside your own district. Virtually impossible to go out of district to a standard public school. The exception being something like Aviation High School. Seattle Public Schools are decided average. They look bad because several of the surrounding districts like Northshore, Lake Washington, Bellevue, Bainbridge Island, Mercer Island are tops in the State. However, Seattle has a number of very good, and very expensive private schools; Bush School, Blanchet, O’Dea, Lakeside etc. Yes, eastside parents do send kids into Seattle Schools but not usually on a bus ;-)

      3. I grew up in Lake Forest Park and went to O’Dea. Took the old Metro 307 every day for four years of school.

        SPS is growing, however, despite their challenges; the latest numbers for BEX IV, which will be the follow-on capital levy for the schools, sees them needing significant increase in capacity.

    4. Don’t confuse percentage of population with kids with absolute numbers of kids (i.e., density of kids). When I looked at density of kids per census tract, I found that Seattle actually has more higher kid density areas:

      “Say a cul-de-sac with 10 houses has 8 families with children. That’s an impressive percentage! However, on Capitol Hill that same amount of land area can contain several apartment buildings with hundreds of residents, including more total children than the cul-de-sac.”

      1. The number of children enrolled in public schools has been on the decline. But, when you factor in the private school population it’s been basically flat. What you have in Seattle is a serious split of “haves” and “have nots”. Diversity doesn’t account for much with this type of barrier. Seattle is an isolationist alcove as represented in the number of people per household. It doesn’t really matter how many kids are in the building if parents choose to isolate them rather than let them roam the neighborhood as in suburbia where everyone is just as different as everyone else.

      2. Actually, Seattle School District’s enrollment has been rising, not declining, over the past few years. They hit bottom for enrollment in the 07-08 school year, and enrollment has been increasing since then. Article:

        I haven’t seen any data on private school enrollment over the same time period, so I’m not sure how much is demographic changes in the city, and how much is the recession means people can’t afford private school anymore, would love to see some data on that. I think there’s a sense among parents in the district that there are more families living in the city than there had been, and the district is now predicting enrollment increases for the next several years. Link:

  4. The cities with the largest average household sizes are almost all in the Sunbelt. The only execeptions above 2.66 are Newark (dirt poor) and Chesapeake, VA (huge military population).

    1. The central argument of this book is that density is not destiny.

      I guess that explains Portland having a density of 4,347/sq mi, not much different than Bellevue’s 3,950/sq mi, and Seattle being at 7,255.

      1. See above. When you look at population-weighted density of the metropolitan area, Seattle and Portland are about the same, and both are way worse than any of the big East Coast cities (and even some of the newer West Coast ones, like Phoenix).

      2. Please explain population weighted density. Seattle and Portland are almost equal in population. Honestly, I’m not understanding this metric.

      3. Standard density is land-weighted, meaning that it tells you the expected number of people who live in a given land area. Population-weighted density, by contrast, tells you the expected density of the nearby environment (let’s say census tract) of the average person.

        I’m not great at explaining it abstractly, but it’s easy to understand the difference by example. Suppose that you have two census tracts of a square mile each, each with 1,000 people. The combined density of those tracts is 1,000 people per square mile. Now, suppose that everyone from one of the tracts moves into the other. Using standard metrics, the combined population density of the two has not changed. But using population-weighted density, the density of the two tracts has doubled, because the average person (in fact, any person) is now living in a neighborhood that’s twice as dense. The empty land does not factor into the calculation.

      4. OK, thanks. I sort of get it. A simple example would be including or not including the area of lakes in your density calculation (7,200 does not include water area which accounts for 40% of Seattle city limits). Although people might be out there on a boat it’s not part of the daily living space. Large parks might also be excluded. And golf courses? Small “pocket parks” on the other hand would not be so it gets complicated. What do you do with something like Husky Stadium? Once renovated it will only be used by the public about ten days a year. Century Link Field OTOH gets used what, maybe 30 times a year? Then you’ve got acres and acres of container storage at the Port and runways at Boeing field.

      5. In the abstract, the question you’re asking is, “What is the density of each person’s living area?”, and then averaging it all together. But that’s pretty much impossible. Thus, any number is an estimate.

        Still, the estimates are pretty good. Census data will tell you the number of people who live in a given block, as well as its size. So for each person who lives in a block, you say that their personal density is the density of their block. Therefore, any block where no one lives — like the stadiums and the docks — doesn’t factor into the calculation at all.

        Of course, now you run into issues of residential versus nonresidential density. Somewhere like downtown has much less residential density than you might think based on all the towers. So it’s good to measure commercial density as well. But combining that with residential density in a way that makes sense can be a very difficult problem.

        At the end of the day, all the numbers in the world will only tell you so much. I don’t need a report to tell me that downtown’s ratio of daytime to night-time uses is way out of whack; I can see it for myself, just by spending five minutes on 3rd Ave in daylight, and another 5 minutes at night. I’m more interested in making sure that people don’t draw false conclusions from misleading numbers.

  5. I’m surprised that Vegas and Phoenix have such high percentages of households with children. I’d assumed the building boom there, especially Vegas was driven by retirees. The rest that are above 30% I can understand. Especially Anaheim with Disneyland at 38.9% := OK, Santa Ana is a weird one:

    Density 11,887
    HH size 4.37
    w/ kids 46.5%
    Ownership 47.5%

    Large population of Hispanic Catholic immigrants?

    1. I would have assumed that the growth in Vegas was driven by the casinos and convention business. Lots of dealers, hotel clerks, maids, waitresses, showgirls, etc.

      For the Phoenix area, compare Phoenix, Mesa, Chandler and Glendale with Scottsdale.

      1. Nah, the gaming industry is virtually recession proof and hasn’t grown anywhere close to the exploding population. Plus, the famous Las Vegas Strip is actual in unincorporated Clark County.

        1980 164,674 30.9%
        1990 258,295 56.9%
        2000 478,434 85.2%
        2010 583,756 22.0%

        Vegas was one of areas if not the hardest hit by the real estate bust.

        Sin City’s metro area led the nation in mortgage defaults for 22 straight months through August and home prices plunged a whopping 60% from their 2006 peak,… Las Vegas’ low cost of living, warm, dry climate and its attractions, make it a destination, especially for retirees and people looking to buy vacation homes.

        I’m not sure how the census accounts for snowbirds that live half the year in places like AZ, NV and FL.

  6. “Square Miles: #62”

    They obviously didn’t factor in the disconnected part of Seattle up in the Cascades in the Cedar River Watershed. I think we’re top-10 with that. Is it part of the city? No… but are there pieces of the “larger” cities that should be discounted for the same reason? Obviously this has a direct impact on “Density”.

  7. KIRO News radio this morning ran a teaser piece about US growth of distant suburbs falls to historic low. The author completely skews the facts to try and make her case for “the end of an era.” She points to Kendall County in Chicago but the census shows this area grew 110% since the previous census. Illinios grew by only 3.3%, a third of the national average, and Chicago shrank by 6.9%. And the Windy City isn’t a rust belt town of shuttered up factories. It’s a global finacial center and has attracted international companies… like Boeing. In real numbers, Kendall grew by 60,000 and 200,000 left Chicago for greener pastures.

    There is no universal trend based on population or density. What the census numbers do show is that on whole the majority of the growth based on percentage of total growth is in the inner ring suburbs. The places like Bellevue and Redmond that were exurbs in the 50’s. The idea that people are now moving to cities in large numbers has to be tempered by the realization that in many cases the city is moving to the people through annexation and incorporation. Lakewood is the seventeenth largest city in Washington. It “didn’t exist” until 1996. That’s 60,000 people that suddenly “moved” to the city.

    One thing the article alluded to but didn’t really cover is Ghost Towns in America. Maybe because not all of these are abandoned McMansions in the exurbs:

    In South Florida, high-rise ghostly towers filled with empty condos speckle the skyline. Miami alone produced 23,000 new condo units from 2003 to 2007 and is now experiencing one of the worst condo busts of any city in the country… One condo-owner in downtown Fort Myers, FL, refused to leave the 32-story building as other owners took buy-outs, and he and his family are now the only residents living in the entire building.

    1. All the more reason to build up the inner suburbs with rapid transit and walkable centers. If they’re heading in the direction of city neighborhoods, we should push it along. The suburbanites have already gotten used to the idea that “urban features” like crime and multiethnism are inevitable because they’ve existed in the burbs for well over fifteen years now.

  8. this information completely jives with the data I reported on earlier showing that a majority of Seattle housing is multifamily already

    The majority of Seattle voters live in single-family homes. Most apartment dwellers are transient, and many are poor and marginal. We house dwellers run the show, and we’ll keep running it. I know you hate that, but it’s the fact.

Comments are closed.