Saturday, May 14, 2016
Employment, construction, and the cost of San Francisco apartments
Everyone agrees that housing in San Francisco is expensive, and that the high costs are hurting the city. But there is a lot of disagreement about why the rent is so expensive, and what to do about it.
Sonja Trauss of SFBARF has argued that costs are high because there is not enough housing to go around and that the answer is to build more. Tim Redmond of 48 Hills has argued that building more housing would make the problem worse because the people who would move into it are likely to be wealthy newcomers whose demand for services will increase low-income employment, putting further pressure on older, lower-cost housing.
Who is right? Is anyone?
Housing cost trends over the years
To understand what causes high housing costs, we must first have an idea of how high costs are, and how high they have been over time.
In the 1940s there were very few apartments advertised for rent. It is not clear whether this was because there were extreme shortages of housing or whether people typically found their apartments some other way. Nevertheless, there is at least some rough evidence that prices were dropping every year until they briefly stabilized in 1954.
After this lull, in 1956, apartments began to be listed in increasing numbers, but their prices also began to rise. Overall, they went up 6.6% every year. Today's outrageous prices are exactly in line with the 6.6% trend that began 60 years ago.
There have, however, been a few deviations from this overall trend. In 1978, prices began to rise and did not return to the trend line until they stabilized for a few years in the early 1990s. Then in 1995 they rose sharply again, remaining above the trend line until 2005. Prices rose again in 2007 and dropped in 2009 before returning to the trend line.
Nominal median rent by year
Note that the Y axis here is logarithmic, not linear, so the overall 6.6% trend is shown as a straight line, not as a steep curve.
It is slightly unfair to talk about this rent growth in face value dollars, because purchases in general, not just rent, cost more than they did in 1956 thanks to inflation. After adjusting for the Consumer Price Index, real rents have only gone up 2.5% per year and have only quadrupled in effective cost in 60 years. It is still an alarming increase.
.
Sources of housing cost data
Where do these prices come from? Daniel Hertz at City Observatory warns us to be wary of anyone who says they know what the median rent in a city is, because their claims vary wildly.
Nevertheless there are sources that appear to at least provide an internally-consistent view into the rise and occasional fall of rent over the years. A frequently cited source is the San Francisco Housing DataBook, which tabulates the median rent of two-bedroom apartments advertised in the San Francisco Chronicle on one Sunday per year, usually the first Sunday in April, from 1979 through 2001.
I set out to replicate the DataBook's methodology over a wider range of years, but quickly gave up on including just two-bedroom apartments, because ads in the early 1960s rarely referred to apartment sizes in these terms. Instead, for each first Sunday in April from 1948 through 1979, plus a few other years, I made a list of all the advertised unfurnished apartments, flats, houses, and, later, condos, regardless of size, that were advertised in the Chronicle. Mostly I used the San Francisco Public Library's page scans of the newspaper but resorted to microfilm for the few later years where no page scans are available.
In recent years, another good source is listings from Craigslist that have been preserved by archive.org. The dates that were archived vary considerably from year to year, so I collected all the available listings from each year, regardless of date.
The sources appear to be mutually comparable. There are three years (1979, 1984, and 2001) where I have all the Chronicle prices plus the DataBook two-bedroom figures. Fortunately in each of these years the same simple linear transformation of the DataBook two-bedroom price yields the median price that I found, so I have scaled the other DataBook prices the same way to bring them into my pool. (It would be better not to have to rely on the DataBook, but transcribing prices from microfilm is very tedious and time-consuming.) For the Chronicle and Craigslist, I have listings from both for 2003. The shapes of the distributions are the same and the two medians agree to within $100 (6%) so I have used the Craigslist medians without scaling for years since 2003.
Is median rent the right thing to compare?
Surprisingly, given the anxiety over "luxury" apartments vs regular, non-luxurious ones, there does not appear to be any segmentation in the rental market in any year, even by unit size. In every year with more than a handful of listings, the total pool of prices is a smooth curve (except for the occasional taboo price: landlords were strangely reluctant to offer apartments for exactly $105), approximately lognormally distributed. The 95th percentile apartment always rents for about 2.2 times the price of the median apartment, and the 5th percentile apartment for about 1/2.2 of the price of the median apartment, but see below about variability in this.
Construction booms
What happened in 1954 to stop prices from dropping? The most obvious explanation is that that was when San Francisco ran out of large tracts of vacant land. It took a while before anyone noticed that anything was wrong, but by 1966 there was talk of a housing crisis, and the Planning Department began maintaining an annual report, the Housing Inventory, to keep track of the rate of housing construction and demolition. Recent issues of the Housing Inventory are online; older ones are in the Berkeley Environmental Design Library.
San Francisco's post-earthquake housing was built in a series of booms. The biggest were the immediate post-earthquake rebuilding from 1906 through 1918, when essentially all of the densest areas of the city were built, and then the transportation-led boom from 1919 through 1934, when the Marina, the Outer Richmond, West Portal, the Parkside, and the Outer Mission were built. From 1935 to 1943, the Central Sunset and Parkmerced filled in. From 1944 to 1954, the Outer Sunset and Ocean View were built. And that was essentially the end of the easily developed greenfield housing.
The next boom, from 1955 through 1967, still could fill in the hillsides of Twin Peaks and above O'Shaughnessy Boulevard. But the boom from 1968 through 1982 had no genuine vacant tracts to fill at all. It was the Redevelopment Agency's boom, rebuilding the Western Addition and the Golden Gateway and Diamond Heights. (The demolition of the Western Addition is subtracted from the net new unit count in earlier years.)
The first private infill boom was from 1983 to 1993. It was entirely on scattered sites, with no large tracts available. A long, double-peaked boom ran from 1994 through 2011, with a dip in the middle for the dot-com crash. It too was a scattered infill boom, aided by the South of Market sites made available by the former Embarcadero Freeway and I-280 extension. Since 2012 we have been in the ninth building boom, focused on sites on or near Market Street. Few sites are involved, but the numbers are the largest since the early 1960s because the buildings are large. I don't know how long it will continue.
(The Housing Inventory data only goes back to 1960, so information about earlier construction rates comes from the Planning Department's 2016 Land Use map. It tracks the year of construction of each building, not the year of first occupancy as the Housing Inventory does, and only of buildings that stil exist, but the numbers are generally comparable.)
invite response
Saturday, May 14, 2016
Employment, construction, and the cost of San Francisco apartments
Everyone agrees that housing in San Francisco is expensive, and that the high costs are hurting the city. But there is a lot of disagreement about why the rent is so expensive, and what to do about it.
Sonja Trauss of SFBARF has argued that costs are high because there is not enough housing to go around and that the answer is to build more. Tim Redmond of 48 Hills has argued that building more housing would make the problem worse because the people who would move into it are likely to be wealthy newcomers whose demand for services will increase low-income employment, putting further pressure on older, lower-cost housing.
Who is right? Is anyone?
Housing cost trends over the years
To understand what causes high housing costs, we must first have an idea of how high costs are, and how high they have been over time.
In the 1940s there were very few apartments advertised for rent. It is not clear whether this was because there were extreme shortages of housing or whether people typically found their apartments some other way. Nevertheless, there is at least some rough evidence that prices were dropping every year until they briefly stabilized in 1954.
After this lull, in 1956, apartments began to be listed in increasing numbers, but their prices also began to rise. Overall, they went up 6.6% every year. Today's outrageous prices are exactly in line with the 6.6% trend that began 60 years ago.
There have, however, been a few deviations from this overall trend. In 1978, prices began to rise and did not return to the trend line until they stabilized for a few years in the early 1990s. Then in 1995 they rose sharply again, remaining above the trend line until 2005. Prices rose again in 2007 and dropped in 2009 before returning to the trend line.
Nominal median rent by year
Note that the Y axis here is logarithmic, not linear, so the overall 6.6% trend is shown as a straight line, not as a steep curve.
It is slightly unfair to talk about this rent growth in face value dollars, because purchases in general, not just rent, cost more than they did in 1956 thanks to inflation. After adjusting for the Consumer Price Index, real rents have only gone up 2.5% per year and have only quadrupled in effective cost in 60 years. It is still an alarming increase.
.
Sources of housing cost data
Where do these prices come from? Daniel Hertz at City Observatory warns us to be wary of anyone who says they know what the median rent in a city is, because their claims vary wildly.
Nevertheless there are sources that appear to at least provide an internally-consistent view into the rise and occasional fall of rent over the years. A frequently cited source is the San Francisco Housing DataBook, which tabulates the median rent of two-bedroom apartments advertised in the San Francisco Chronicle on one Sunday per year, usually the first Sunday in April, from 1979 through 2001.
I set out to replicate the DataBook's methodology over a wider range of years, but quickly gave up on including just two-bedroom apartments, because ads in the early 1960s rarely referred to apartment sizes in these terms. Instead, for each first Sunday in April from 1948 through 1979, plus a few other years, I made a list of all the advertised unfurnished apartments, flats, houses, and, later, condos, regardless of size, that were advertised in the Chronicle. Mostly I used the San Francisco Public Library's page scans of the newspaper but resorted to microfilm for the few later years where no page scans are available.
In recent years, another good source is listings from Craigslist that have been preserved by archive.org. The dates that were archived vary considerably from year to year, so I collected all the available listings from each year, regardless of date.
The sources appear to be mutually comparable. There are three years (1979, 1984, and 2001) where I have all the Chronicle prices plus the DataBook two-bedroom figures. Fortunately in each of these years the same simple linear transformation of the DataBook two-bedroom price yields the median price that I found, so I have scaled the other DataBook prices the same way to bring them into my pool. (It would be better not to have to rely on the DataBook, but transcribing prices from microfilm is very tedious and time-consuming.) For the Chronicle and Craigslist, I have listings from both for 2003. The shapes of the distributions are the same and the two medians agree to within $100 (6%) so I have used the Craigslist medians without scaling for years since 2003.
Is median rent the right thing to compare?
Surprisingly, given the anxiety over "luxury" apartments vs regular, non-luxurious ones, there does not appear to be any segmentation in the rental market in any year, even by unit size. In every year with more than a handful of listings, the total pool of prices is a smooth curve (except for the occasional taboo price: landlords were strangely reluctant to offer apartments for exactly $105), approximately lognormally distributed. The 95th percentile apartment always rents for about 2.2 times the price of the median apartment, and the 5th percentile apartment for about 1/2.2 of the price of the median apartment, but see below about variability in this.
Construction booms
What happened in 1954 to stop prices from dropping? The most obvious explanation is that that was when San Francisco ran out of large tracts of vacant land. It took a while before anyone noticed that anything was wrong, but by 1966 there was talk of a housing crisis, and the Planning Department began maintaining an annual report, the Housing Inventory, to keep track of the rate of housing construction and demolition. Recent issues of the Housing Inventory are online; older ones are in the Berkeley Environmental Design Library.
San Francisco's post-earthquake housing was built in a series of booms. The biggest were the immediate post-earthquake rebuilding from 1906 through 1918, when essentially all of the densest areas of the city were built, and then the transportation-led boom from 1919 through 1934, when the Marina, the Outer Richmond, West Portal, the Parkside, and the Outer Mission were built. From 1935 to 1943, the Central Sunset and Parkmerced filled in. From 1944 to 1954, the Outer Sunset and Ocean View were built. And that was essentially the end of the easily developed greenfield housing.
The next boom, from 1955 through 1967, still could fill in the hillsides of Twin Peaks and above O'Shaughnessy Boulevard. But the boom from 1968 through 1982 had no genuine vacant tracts to fill at all. It was the Redevelopment Agency's boom, rebuilding the Western Addition and the Golden Gateway and Diamond Heights. (The demolition of the Western Addition is subtracted from the net new unit count in earlier years.)
The first private infill boom was from 1983 to 1993. It was entirely on scattered sites, with no large tracts available. A long, double-peaked boom ran from 1994 through 2011, with a dip in the middle for the dot-com crash. It too was a scattered infill boom, aided by the South of Market sites made available by the former Embarcadero Freeway and I-280 extension. Since 2012 we have been in the ninth building boom, focused on sites on or near Market Street. Few sites are involved, but the numbers are the largest since the early 1960s because the buildings are large. I don't know how long it will continue.
(The Housing Inventory data only goes back to 1960, so information about earlier construction rates comes from the Planning Department's 2016 Land Use map. It tracks the year of construction of each building, not the year of first occupancy as the Housing Inventory does, and only of buildings that stil exist, but the numbers are generally comparable.)
There is more here:
http://experimental-geography.blogspot.com/2016/05/employment-construction-and-cost-of-san.html
#housing #politics