Someone Check The ACS Please

Image result for american community survey data

The 2011-2015 5 year “wave” estimates for the American Community Survey are out.  These are exciting because we now have enough “waves” of the ACS to look at changes in Census tracts over time without having to utilize those painful crosswalks…

My hope is that someone will take a look at the 2006-2010 ACS estimates for tracts against the 2011-2015 estimates and let us know a few things:

  1.  Are the samples robust enough to give us significant differences where we know there are significant differences?
  2. What can they tell us about urban change, if anything?
  3. How has demographic change itself impacted the stability of the estimates in certain places?

Many people use 5 year wave data like the 2011-2015 dataset and, in their manuscripts, refer to it only as the “2015 ACS.”  This is ok if you have acknowledged towards the beginning of the paper that you are referencing what is a five year rolling sample… but many practitioners, media folks and the public at large sees “2015 ACS” and thinks that is what you are talking about.

Failing to appreciate this difference means ignoring what the data is really telling you…

California’s Legislative Analysts’ Office (LAO) twisted the ACS data used by Karen Chapple and Miriam Zuk at Berkeley to try to say, essentially, the opposite of what Zuk and Chapple argued in their own work… to the point where they had to respond.

The biggest critique I have in all of this is much more basic: what does a median rent in a Census Tract really tell you?  When you tell me that adding affordable housing to a census tract over a period of time reduces its median rent change in that period, all you are telling me is that adding numbers to the bottom of a distribution of numbers shifts the distribution’s median downwards.  How much can we really say with the blunt instrument that is the ACS 5 year tract estimates?

Last point on that Berkeley-LAO debate: they relied on the 2009-2013 5 year wave data, but kept constantly referring to their findings as measuring change from the 2000 Census and 2013 ACS.  The 2013 wave includes 2009, when we were in a recession. It also includes 2013, which was undoubtedly a ‘boom’ year for housing in the Bay Area…  I’m wondering if someone ran the Berkeley analysis with the latest ACS wave, what would they find? It’s the first wave we’ve got tract-scale estimates for that is clear of the recession.  Someone please dig in!

A Housing Policy Idea Worthy of A Sanders’ Or Clinton Campaign

Full disclosure: I feel the Bern, especially when I drink too much instant coffee (who can afford beans anymore?).  But I am also a Millennial concerned with income inequality, retirement security for our seniors and climate change, three problems that all surprisingly intersect on an issue that Bernie Sanders and Hillary Clinton have both completely ignored in the campaign so far: Housing.

Here are several of the major housing problems that policymakers must deal with simultaneously:

  1. Providing housing in ‘location efficient‘ communities where residents can save money by walking, biking and taking transit (or at least driving shorter distances because locations are closer).
  2. Assisting boomers who are “aging in place” in homes too big for an empty-nest couple, but who cannot afford assisted living.
  3. Assisting Millennials who are renting longer for a variety of reasons and as such are driving up rents in major cities.
  4.  Building “up” not “out,” or reducing urban sprawl  to protect air and water quality.
  5. And doing all of the above without triggering displacement.

The policy I propose  can assist in making a dent in all of these problems simultaneously.  It can help address these issues without triggering NIMBY (Not In My Backyard) fears about high density development invading neighborhoods, too.

So Here’s The Humble Proposal:

A a grant program which funds home owners to add second units (often called “grandma flats”), or to split their single family homes into duplexes, or to add stories to existing low-density buildings, taking advantage of innovations in ‘stackable’ manufactured multi-family housing.  In exchange, the newly produced units would be rented at affordable rates.  The rental income would be split between the home owners and the government (to help pay off the cost of construction).  Once construction costs were fully paid off, home owners would start receiving the full rental income.

This program would only be available to home owners who were:

  • Seniors or about to be seniors, or
  • Underwater on their mortgage, or
  • Low or moderate income, or
  • Long term residents of a neighborhood being gentrified, or
  • Living in a neighborhood where the average renter is rent burdened as the Department of Housing and Urban Development defines it.

Sound like a funny idea?  If you’ve walked through parts of San Francisco, then you’ve walked through a community where this has already happened naturally without government support.  Many of the city’s large Victorians have been converted into multi-units rentals, have added stories or second units–all while keeping that historic charm and low-density feel NIMBYs love so much.

The Winners: Seniors, Struggling Home Owners, Areas at Risk of Gentrification, Renters, The Environment

This policy could provide aging boomers additional rental income as they age-in-place.  It gives Millennials affordable rental housing, and can increase affordable rental housing in transit-rich neighborhoods (ticking off the environmental box).  It could also help under-water home owners living in areas where the rental market is too hot.  Richmond, CA, has a significant number of underwater homes.  It’s proximity to San Francisco via transit could mean a rapid gentrification of the community.  This program could help those home owners find new rental income, while creating space for newcomers:  a form of development without displacement, creating wealth for existing residents instead of pushing them out.

Academics have suggested backyard units offer great potential in some of most impacted places, like the San Francisco Bay Area.  In-law units cost between $75,000 to $200,000  a piece, much lower than conventional affordable housing.   There are no numbers on the cost of converting single family units to duplexes, and I haven’t seen much.  The program could come with a $15,000 bonus for neighborhood improvements for every unit added by such innovations to help overcome that ever-present NIMBY impulse.

How is this better than existing affordable housing programs?

In the Low Income Housing Tax Credit (LIHTC) program, private developers and their investors (usually banks or investment firms) all take a cut while providing affordable housing.  There’s nothing terribly wrong with that.  But in my proposal, the people ‘taking a cut’ are also people in need of support: home owners who are also struggling or at risk of struggling for reason or another.

This idea may seem unconventional, but is by no means new.  It’s time to get unconventional in approach to creating affordable communities anyways, because the status quo simply isn’t working and increasingly, the rent is too damn high!






Is San Francisco Doing Enough For Affordable Housing Given Its Costs and Its Wealth?

The short answer is

Out of Context: Yes, San Francisco is doing a lot for affordable housing. On a raw dollars per-unit basis, it is spending more money than most other cities in the state on affordable housing development.

In Context: No, not compared to how much money other, less wealthy cities across the State are pouring into affordable housing given their worse budgetary limits and relative to their local building costs.

Here’s The Synopsis: 

When you look at San Francisco’s funding of it’s affordable housing units on a per unit basis, as a percentage of the cost per unit, San Francisco drops from being one of the best performing cities to being somewhere in the middle (but closer to the top) of the pack.  Basically, the per-unit gap in financing necessary to build an affordable unit in San Francisco is so astronomically high compared to the rest of the state that the city’s large raw dollar amounts… well they amount to nowhere near the kind of commitment other, less wealthy cities are putting in to their own affordable housing relative to their costs (and relative to the strength of their economies. San Francisco is doing much better than Bakersfield and Sacramento, yet the latter communities are arguably doing a lot for their affordable housing–read all the way through for more).

(and therefore SF residents should vote yes on the affordable housing bond!)

The Context

Figure One below shows the average per-unit gap financing required to build affordable housing in California.  These numbers are based on data taken from the California Tax Credit Allocation Committee (TCAC) for 4% tax credit projects funded from 2011 to 2015.  I used housing projects’ budgets during that period to calculate the average per-unit gap in financing needing to be filled after a project won its tax credits and identified how much revenue it could pull in directly from rents.  These gaps are what local jurisdictions, foundations, and others must fill to make an affordable project solvent.


Per-Unit Gap Financing Needs for Affordable Housing, Based on 4% Projects from 2011-2015
Per-Unit Gap Financing Needs for Affordable Housing, Based on 4% Projects from 2011-2015

Using this data, we can estimate the per-unit gap cost of funding affordable housing in San Francisco is at least $248,000 per unit (I say at least because in  2011 costs were probably still lower due to the recession, and this is pulling the estimate downward significantly).  This $248,000 per unit compares with a statewide average of $90,000 per unit…

Utilizing the same data, I’ve calculated that about 30.5% of this gap financing need for affordable housing across the state was paid for by local governments during this period (overall).  This compares to 35.7% of gap financing needs paid by local government in San Francisco during the same period.  During this time, San Francisco supported its own affordable housing development at rate 17% higher than the state overall.  So this is not too bad.  But ask yourself: is San Francisco only 17% wealthier than the state as a whole? Using median incomes, the answer is actually yes.  San Francisco’s median income is around $72,000, while for the state it is $61,000 (e.g. it is 18% higher in the City).  But these are medians, and medians can be quite deceptive. As the Chronicle reported, San Francisco’s top 5% are significantly richer than their peers across the country when the top 5% is measured as a ratio against the bottom 20% in the same city.  Putting all of this together: San Francisco’s commitment to funding its affordable housing is merely comparable to the state as a whole given it’s higher costs of construction and higher area incomes.  That said, the enormous new concentration of wealth at the top of San Francisco’s income distribution has not translated into the city providing a disproportionately higher rate of financing for affordable housing.

Simply put: The city is running in the middle of the pack in terms of its affordable housing commitments, maybe in the top third relative to the state overall. But the supposed boom of wealth pouring into town has not added any extra bonus to the city’s efforts (relative to what the rest of the state is doing).

This is easier to understand when mapped.  The map below shows, for each county, the percentage of affordable housing gap financing covered by local jurisdictions from the dataset I referenced above. Look at who is giving real commitment:

Share of Affordable Housing Gap Financing Needs Funded by Local Jurisdictions--Blue means no data
Percent of Affordable Housing Gap Financing Needs Funded by Local Jurisdictions–Blue means no data, 4% Projects 2011-2015

The counties where local governments are doing the most to support affordable housing are mostly in the Central Valley.  At the level of the city itself, the results are even more interesting.  Below are the top 20 cities ranked in terms of the percentage of their affordable housing gap financing need they paid for through various local sources.

Top 20 Cities Ranked by The Percentage of Their Gap Financing Needs Were Met Through Local Sources
Top 20 Cities Ranked by The Percentage of Their Gap Financing Needs Met Through Local Sources, 4% Projects 2011-2015

First off, San Francisco isn’t even on the list.  In fact, most major cities in California aren’t.  The only two of cities 10 most populous cities on the list are Sacramento and Bakersfield.  Here is the top ten most populous cities arranged in order of the percentage of affordable housing gap financing they managed to cover through local sources for 4% projects during this period:

  1.  Bakersfield (72%)
  2. Sacramento (67%)
  3. Oakland (48%)
  4. San Francisco (36%)
  5. Los Angeles (28%)
  6. San Diego (25%)
  7. Anaheim (17.7%)
  8. San Jose (13.3%)
  9. Long Beach (8.2%)
  10. Fresno (1%)

It is worth exploring why Sacramento and Bakersfield–which are significantly less wealthy than San Jose, San Francisco and others on this list–are doing so much more for affordable housing in their areas given relative to the costs they face.

Here’s the challenge for San Franciscans looking to improve the housing crisis in the state: Increasing affordable housing funding the city to levels comparable to Oakland, Sacramento or Bakersfield will certainly yield a lot more units, but it will be nowhere near enough.

The real challenge is: how do you get San Jose to increase its affordable housing financing commitments to comparable levels as well???

The Stranger Implications Of Berlin’s New Rent Control Law

To deal with gentrification and speculation the people of Berlin passed a rent control law that doesn’t allow landlords to rents higher than a geographically-constrained average.  The threshold is set on a per-square-meter rate in each city district, and is derived based on a biennial state census of rents (the district average, I think… But I am having trouble finding formulas in English).  New rent contracts can only be 10% more than this average rent–anything more than that is illegal until a new Census shows the average has risen.  Then the max is 10% higher than the new average.  Basically then, landlords’ ability to raise rents is tied to the decisions of other landlords—if everyone raises rents by the maximum amount then the maximum amount set by the next census will be that much higher.

Berlin’s districts. The scale of policy implementation, which holds significant implications for the long run effects of this policy.

This changes the game, dramatically.  In this post I want to walk through some theoretical implications. This is a theoretical exercise. I don’t necessarily see all of these implications unfolding, but they are worth thinking about to truly grasp the implications of this sea-change policy.  Here are some of the stranger possible implications:

  1. Collective Action on the Part of Landlords?

If I’m a landlord in the beautiful and charming Litchterfelde West, then my ability to raise rents is based on what everyone else in that neighborhood’s administrative district decides to do. Hypothetically, if every landlord in the area who can afford to in fact raises rents up to that 10% threshold, then next year I can raise my rents more than if every able landlord in the area ends up raising rents by 5% or 3% of that average.  Could this lead to a situation where landlord associations actively push members to work collectively to push rents upwards?  If my neighbor might get picked for the Census, then I want him to have high rents so that this average and the limit it creates is higher. This enables me to raise rents higher in the future.

Obviously not everyone could participate in that type of strategy.  If my unit is in disrepair I can’t help other landlords out by raising rents unless demand is so pent up people will accept a sub-par unit at higher costs.  Do other landlords thus want to ensure I am more likely to “keep up” unit quality then?

  1. New Development Can Now Directly Impact Existing Rents

If new developments add a combined 5,000 new, upscale units to my area—that could push a district of even 50,000 households’ median a good deal higher in the next census than if the new developments weren’t coming in.  That means the mere addition of these units to the area increases what I can charge next time around.

Two implications in this:

A: Landlords might have a new incentive to support expensive new development coming into areas where they own property.  Might they thus want to oppose developments for the poor coming in–and thus pulling the average down?

B: the geographic scale at which local rates are set mitigates/exacerbates the impact of new developments on this policy. If districts have 300,000 households—and only adds a few thousand new units every year—then this impact might be marginal.  But if my district has only 70,000 households, then a few thousand matters.

  1. Un-Even Capital Investment?

Now wealthy districts can continue to raise rents faster than non-wealthy ones.  Would induce investors to only finance new developments in already wealthy areas where the rents (eg returns) can rise faster?  Rather, could this produce arbitrarily un-even investment patterns? If your neighborhood’s primary concern is gentrification then this could be a good thing. A gentrifying neighborhood in a district that is mostly poor could see this put a break on investment in further gentrification.  The other poor neighborhoods of the district “pull” the average max rental change downward, lowering profitability of new development in said gentrifying neighborhood.  In contrast, if you are in a gentrifying neighborhood in a district that is much more expensive than your neighborhood, then the gentrification is still readily facilitated by rich areas pulling the average upwards.  So… Could this contribute to income segregation by district? (again—just thinking out loud here—not offering super robust hypothesis).

  1. Edge Effects

Suppose I live at the poorer edge of a district which is otherwise quite wealthy.  This enables landlords on my (poorer) end to raise rents faster than landlords across the street from me in a district that is poorer overall.  If there’s something on my street that attracts residents: a park, a rail stop, a large transit center…. Then there’s going to be a lot more competition for units on the other side of my street.  This is because over time my side of the street is rising in rents faster than the side in the poorer district.

What we could see, then, are a lot of arbitrary breaks in the rent surface or rent gradient across the city—with these breaks taking place at the edges of districts.

  1. The Politicization of the Data and Data Gathering Process

Bloomberg basically suggests this is already happening.  People are very upset about the methodology.  If Bloomberg is right and a sample of 20,000 households decides these rules for the entire city–then all of the sudden what my tenants report as their rents has very immediate, real impacts on my ability as a landlord to profit in the near future!

These are opening thought.  I hope serious research gets conducted on this topic so we can evaluate if this is the type of solution that we need.

Is AirBnb Globalizing Gentrification?

What I mean by the title of this piece is not that gentrification is suddenly going global, as it’s a local phenomenon that has been happening in cities around the planet for decades.  What I mean is that whereas in the past gentrification was about rich residents from other parts of town moving into working class communities, it’s now also a process of well off people from anywhere in the world contributing to mass evictions just by going on vacation.  Without a doubt, AirBnb is “disrupting” traditional housing markets in places like San Francisco, but not in a good way.

Consider the Mission in San Francisco.  The 24% of units taken offline as a result of AirBnb have not been bought up by eager new tech money that’s just moved to the city.  In The Mission that number is closer to 30%.  They are being rented out to people from who knows where who are drawn to the area for who knows what reason (probably all the new fancy restaurants, coffee shops, and flocks of disturbingly identical-looking hipsters-bros).  The strange irony of all of this is that the benefits of tourism traditionally touted by the City’s tourism industries are now being cancelled out by a new cost: tourism creates tens of thousands of low-skill jobs for people who, thanks to tourism + AirBnb, are less likely to be able to afford to live in the city anyways! 

Chew on that for a moment.  Is San Francisco going to end up like Santa Barbara or many of those beach towns in Florida where only the super-rich actually live there and the janitors, dishwashers and cooks mostly commute in from smoggy inland areas?  The city has two geographic bottlenecks on its north and east side—pushing more workers over those bridges isn’t going to help with meeting greenhouse gas emissions goals (not to mention the developmental and educational damage to children when their parents must spend 2 hours a day commuting).

An average hotel room in the city in 2014 was roughly $187.  Putz around on AirBinb for a while looking at sites in the city, and you’ll find the distribution may be only slightly lower in places like the mission (although I’m not being scientific here).  Are we creating an economy where tourists can spend $20-$30 less a night for a hotel at the expense of the existence of neighborhoods affordable to working families?  Is that what “disruptive technology” really means?

I don’t mean to sound so extreme, I have in fact used AirBnB once to stay in a rather nice part of northern Chicago (everyone looks so miserable in that city for some reason).    But what is the “disruptive” crowd offering up that actually improves the lives of working families?  Micro-Units?  That’s great for a single person like me, but you can’t raise a family in a closet.

Is It Time To Abolish The City?

Today I want to share with you two tales about three cities. After reading this, you may wonder whether or not it is time to abolish the city/town concept and replace it with more seamlessly integrated regional government. The more time I spend researching equity, social justice and housing and transportation policy, the more I strongly I feel that this needs to happen. The first tale takes us to the north end of the island of Alameda.

Alameda Point: Lawsuits, Failed Referendums, Millions of Dollars Spent and No New Housing to Show For It


Alameda Point plan: 1600 new homes, lots of retail space and a wildlife refuge. All 19 years in the making.

Military personnel called the drab northern corner of Alameda Island a home until a wave of federally driven base closures gave the City of Alameda a chance to convert the dirt and scrub weeds into housing and commerce.  We are 19 years out from the publication of that first, fateful planning document outlining the City of Alameda’s vision for the parcel.  And yet, several lawsuits and elections later, the area known as Alameda Point (referenced herein as AP) remains largely desolate.  This is the sage of Alameda Point:

  • 1996Alameda releases a reuse plan for Alameda Point Naval Air Station
  • 1997Alameda Point Naval Air Station closes
  • 19972001: Plans solidify, city hires “Alameda Point Community Partners” as master developer.
    • Plan calls for 1600 new housing units and 4.6 million square feet of commercial development
  • 2003: Residents in Oakland’s Chinatown organize and sue over concerns about traffic impacts (the road connecting AP to the highway system runs right through the heart of Oakland’s Chinatown)
  • 2004: As Asian Health Services vs. City of Alameda heats up, the city backs down and agrees to go back to the drawing board to come up with a better plan
  • 2005: City of Alaemda unveils revised “AP Station Area Transportation Plan”
  • 2006: City releases 5-year plan to finish project by 2010
  • AUG 2006: Navy finalizes price for land: $108 million.  “Alameda Point Community Partners” balks at the cost and backs out.
  • 2006: City begins search for new developer, finds SunCal
  • 2010: SunCal proposes “Measure B” which would reform existing land use laws to enable building 3,000 units on the site (much higher than the previous plan for 1600)
  • 2010: Measure B fails in Alameda (85% vote NO)
  • 2010: A councilmember  violates open meeting laws by leaking planning documents to the public; SunCal publicly attacks/blames city manager for fallout
  • 2010: A newly-elected mayor gives SunCal the boot.
  • 2011: The military decides to give the land away for free
  • 20112014: Alameda returns to original plan
  • 2014: An Oakland Chinatown coalition threatens another lawsuit.
  • 2014: Alameda’s city manager, an attorney who once helped Oakland sue Alameda over AP the first time, now defends the project and signals he will continue to move forward

The irony is a bit obvious: an attorney who fought to kill the project earlier is suddenly a champion of the project as of this writing (Feb. 2014). It gets to me. So here’s my challenge for you:  What went wrong? I’ve been documenting quite a few case studies of failed housing projects for my research. What stands out to me about this project is the roll of inter-jurisdictional conflict and how it all but ensured that once this plan finally got off the drawing board (again) and into the public eye (again), it would get dragged down (you know . . .).

Something to consider:  if Alameda and Oakland were one city, would the lawsuit(s) have occurred?  If they were one city, then this unified city government’s elected leaders would be required to balance the needs of ALL their existing constituents (Oakland and Chinatown) with the pros and cons of generating more tax revenue (from the development).  Would government, working as one jurisdiction and within one process, be better able to balance competing pressures and achieve a more broadly amenable plan?

Because right now, Oakland’s leaders really have no incentive to allow this to go through–they won’t benefit.  And Chinatown’s residents won’t benefit either, as the tax revenue will stay in Alameda (beyond getting funding for traffic improvements in Chinatown paid for by the development).  Sure it’s possible that if Alameda were a part of Oakland a lawsuit could have still taken place, but I think it would have been a lot less likely, as the incentive structure for Oakland’s city and community leaders would be quite different.

Mountain View: Their Housing Development Plan is Called “San Francisco”

Someone recently wrote a great piece on why San Francisco should sue Mountain View.  Apparently the sleepy bedroom community/Tech Mecca wants to add space for up to 20,000 more jobs, whilst only providing up to 5,000 new housing units.  So commute times are going to go up and up (even though they have already been going up and up) because those other 15,000-19,999 workers are going to have to drive in from somewhere else… (Let’s face it, most of those 5,000 units will never be built–so we’re talking about up to another 19,999 cars driving into that area). Meanwhile many of the nearby cities have been fighting the current regional government tooth and nail over increasing their housing supply, meaning these new workers are going to be driving in from quite far away.

So who are these city council members in Mountain View?  And how big are the political coalitions of the illustrious statesmen advancing this plan that will impact several million people?  The highest vote-getter in Mountain View’s latest city council election won only 6,181 votes!  People running for undergraduate student government here at U.C. Davis routinely win more votes than that, but  do not then go on to pass development plans with region-wide implications for congestion and housing costs. (Of course they are just students, but you get what I mean and frankly.. after having a few of them intern for me I wonder if they might make better decisions for the long haul.)

I spent a good half hour trying to find a simple map of Silicon Valley showing the cities with labels and this is what I could find.  Can you point out Mountain View on a map?

It’s the orange-ish colored polygon to the upper-left of San Jose. (A question for you to pocket for later: do any of the jurisdictional boundaries in that map make any sense from a planning, topographical or ecological standpoint?)

So why is Mountain View doing this?  Its simple.  Tax revenue.  But that revenue won’t be shared with the communities that will have to absorb all those new workers into their housing markets.

So suppose all the cities in Silicon Valley were merged into a more comprehensive regional government:  wouldn’t a regional government, if it wants to get re-elected, approve business expansion where there is also adequate housing available, or barring that, where there is at least room to build nearby housing?  It wouldn’t impact who got to share in the revenue (although the revenue would be more  equitably distributed than if it just stayed in Mountain View).  Yes, Linkedin and Google might prefer to have their workers crammed into one area, but at the same time, their employees are filling office buildings up in San Francisco while still maintaining their Silicon Valley presence. So they can clearly manage with campuses spread across the region.

It’s time to seriously consider abolishing the municipality in its current form here in California and replacing it with something larger and more regionally driven.

References on Alameda Point

Alameda Point Community Partners, “Alameda Point Conditional Acquisition Agreement” Alameda, CA, September 20, 2006,

CBS San Francisco. “Alameda regains control of former navy base, plan redevelopment,” CBS San Francsico (San Franciso, CA), June 4th, 2013,

Community Improvement Commission of the City of Alameda.  Adopting the five-year implementation plan: fiscal years 2005/06-2009/10 for the Alameda Point Improvement Project. Alameda, CA: June 20th, 2006

Dieter, Irene. “Case of the disappearing park strikes here again,” Alameda Sun (Alameda, CA), Nov. 8 2013,

EDAW, Inc. NAS Alameda Community Reuse Plan prepared for Alameda Reuse and Redevelopment Authority. Alameda, CA: 1996.

Ellson, Michele.  “Alameda, Oakland in fresh fight over traffic,” The Alamedan (Alameda, CA). January 7, 2014

Hegarty, Peter, “SunCal attorney says Alameda could face lawsuit,” Inside Bay Area (San Francisco, CA), July 15, 2010,

Lee, Henry. “Alameda, SunCal settle for $4.1 million,” San Francisco Chronicle  (San Franciso, CA), Dec. 19 2012,

Mara, Janis, “Housing element adopted after passionate debate,” Marin Independent Journal (San Rafael, CA), Nov 20, 2013,

McDonough, Susan, “New lawsuit aimed at Alameda Point,” Oakland Tribune (Oakland, CA), Jan. 14 2004.

San Francisco Business Times, “Alameda Point Community Partners wins 775-acre development project,” San Francisco Business Times (San Francisco, CA), Aug. 10, 2001

Bianca, “SunCal threatens to sue Alameda city manager”, San Francisco Business Times (San Francisco, CA), July 14, 2010,

White, John Knox, “SunCal submits a new plan for Alameda Point,” San Francisco Chronicle (San Francisco, CA), March 23, 2010,

Who Really Are Beneficiaries of Rent Control in San Francisco?

48 Hills Online covered a map produced by the Anti-Eviction Mapping Project about how renters have dramatically lower incomes in San Francisco than home owners.  The results are pretty stunning: watch the changes of the census tracts in the Mission when you swipe from home owner to renter.

While useful in illustrating where and how the non-rich manage to afford to live in the City, the jump made by 48 Hills Online and AEMP that this suggests rent control isn’t being taken advantage of by high income earners it is a bit simplistic.

This post is devoted to more deeply understanding who is really benefiting from rent control in San Francisco.  To do this, I pull data from the 2000 Long form Census and 2007-2011 American Community Survey microdata (PUMS).  It’s the actual survey responses used to make all kinds of Census economic and demographic estimates, but detached from small geographic tags (like tracts, blocks, block groups) to preserve anonymity.  The richness of the dataset and ability to isolate responses just living in San Francisco proper makes it the ideal dataset for this task.

A.  Income Differences

Figure One below charts density plots of the income distributions of households in rent controlled units, non-rent controlled rental units, homes completely owned/paid off by their owners, and homes owned by residents with mortgages using the 2000 Census data:

Income Distributions
Click on image to see full size

In this dataset, there’s no systematic difference in incomes between those in rent controlled and non-rent controlled units according to both a KS Test (p value: .11) and T-Test (p value: .66).  Given what the AEMP has shown, the enormous income differential between those still paying off mortgages (in blue) and those renting shouldn’t be much of a surprise.  The last group, those who paid off their homes, have lower incomes–similar to renters (we’re talking about a lot of retired families who may have bought before the market got real hot).

So you’re wondering what the big deal is.  I have re-run these distributions for 2000 and 2007-2011 ACS wave (results are dotted lines).  Note whose incomes move up and whose do not:

Click on image to see full size

Those in non-rent controlled units saw their incomes rise faster (or wealthier people started moving in) compared to those in rent controlled units.  At this point in time, a statistically significant gap in incomes emerges between households with rent control and those without–according to both the KS (Kolmogorov-Smirnov) Test (p-score .053) and T-Test (p-score .06).  I re-ran this comparison using the 2009-2013 wave of the ACS, the latest version, and found the difference vanishes.

What explains the swing?  When cohorts of renters started renting:

Click on image to see full size

It’s among long-time residents of San Francisco where the difference in income between rent-control and non-rent control households is most apparent.  Rent control is extremely beneficial to older residents who started renting before San Francisco became way to expensive for the middle class. There are a lot of elderly and retired people in this category, an issue that AEMP has brought up repeatedly in protests against speculative land lords.

B. Searching For Other Differences–Focusing on Race

I spent a solid couple of hour looking for other differences between residents of rent control and non-rent control units, looking for anything significant along lines of poverty status, Household Structure, Location of Employment, Sector of Employment, etc.  The only significant story is along lines of race.  The racial breakdown within each category of tenure is presented below:

Click on image to see full size

My apologies for granularity. To see a better version click this to see in pdf form: RaceAndTenure.

People of Color disproportionately benefit from rent control, particularly the African American and Latino communities.  This is why future Ellis Act evictions could be so devastating if investors continue eye Bayview-Hunters Point as the next site for property flipping.  This is on top of the pressures these communities face from how gentrification influences policing behavior ( learn about Alex Nieto’s case to learn more in the SF context).


In some respects, the households in rent controlled units represent an older San Francisco: more diverse, not as wealthy, a little older.  Otherwise, they are relatively indistinguishable to renters on the open market.  What does this mean for the debate on rent control?  It means that the abolition of rent control would be disproportionately destructive to communities of color and to older residents who are not making as much money in this new economy as their younger counterparts.

Choice Vouchers on San Francisco’s Open Rental Markets: Nowhere Near Enough?

Why Talk About Vouchers?  They’re one of the largest supply-side housing programs in the US, and they may not go far enough in San Francisco’s hot market.

Choice Vouchers are given to low income families to spend on the open rental market.  When recipients secure a unit using a voucher, they pay up to 30% of their income in rent, the rest is paid for by the voucher… unless the rent is higher than the maximum limit for a voucher payout.  Every dollar above that maximum is theoretically paid for by the renter.  The maximums are the Fair Market Rents (FMRs) established by Housing and Urban Development (HUD) using census rental data.  The idea behind vouchers is to help low income families choose where to live instead of ‘stack and packing’ people in housing projects in “bad” neighborhoods.  As I will show below, vouchers do not go very far in San Francisco’s hot market.

San Francisco has a dataset of for-rent postings from June to December of 2012, freely available on their website.  For each rental offer, I’ve coded if it is below the FMR threshold for San Francisco using HUD’s reference (pg 4).

The Map below shows which units from that dataset are voucher-eligible (in red) and those that would require a voucher recipient to pay more than the max rent charged (basically, in-eligible in blue), but limited (for simplicity purposes) to single bedroom units.

Rental Listings Available By Eligibility With Vouchers June-Dec 2012. Note: the underlying map taken from Google is not perfectly aligned with the listings, hence a few units appear to be in Lake Merced and the Bay.


1.  Only 18.5% of the units could be paid for via voucher, even though the fair market rent which sets thresholds is based on average rents.

How did that happen?  The 2012 voucher maximum was based on 2005-2009 American Community Survey (ACS) Data average rent for a two bedroom (which gets adjusted up and down for units with more or less bedrooms). That was initially ~$1,477. But HUD adjusted it upwards to the mean rent of the newcomers in the dataset (moved in within the last year) at two bedrooms which was ~$1,800.  They then add a few cost-increase adjustments and set the two-bedroom FMR threshold at ~$1900.  However, according to the data gathered by the ACS in 2012, rents in the San Francisco HUD Market Area (which includes Marin and San Mateo County), two bedrooms ran, on average, at ~$1700 with a median of $1600.  Among new renters, it was even higher at $2150 on average with a median at $2000.  Table One details the difference between the data HUD had to work with when setting the FMR and what ended up being actual average rents:

Table 1: Mean Rents Used for 2012 FMR Limits Versus Actual 2012 Rents
Dataset Two Bedroom Average Rent Two Bedroom Average Rent (New Renters)
05-09 ACS (used to make 2012 Limits) $1,447 $1,800
2012 ACS* $1,700* $2,152*
2012 ACS-Only San Francisco County* $1,750* $2,375*
June-Dec 2012 Rent Listings (* $2,388**

*Authors calculations using PUMS of Datasf data, R-scripts available upon request.  I should admit I did not bring these estimates up to 2013 dollars but kept them in the amounts reported for each year, so all my calculations are already under-estimates.

** This is in actual 2012 dollars, from

I want to note how strikingly similar the rent listings are to estimated new-renters rents in the ACS, the ACS is truly a powerful data-tool at scales of this size and greater.  So the ACS does in fact keep up with the market in San Francisco, but the market has moved much faster than the FMR.

2.  The Eligible Units are Concentrated in Census Tracts with Higher Commute Times, More Poverty and Lower Incomes

Table 2: Tract-Level Correlations Between DataSF Rent Listings Percent Eligible Vs. Census Demographic Variables*
PCT Voucher Eligible (1BED) Mean Commute Time PCT Families in Poverty PCT Families on Food Stamps Mean Income
PCT Voucher Eligible (1BED) 1
Mean Commute Time 0.275 1
PCT Families in Poverty 0.233 0.063 1
PCT Families on Food Stamps 0.248 0.064 0.7 1
Mean Income -0.202 -0.15 -0.46 -0.51 1

*Data is from the 2007-2013 wave of the ACS.

See anything interesting?  Where are the eligible units disproportionately concentrated?  I’ll tell you where: away from transit access, away from jobs, and in parts of the city that haven’t clearly benefited from the Tech Boom (or suffered through it, depending on who you ask).

3. Many Areas With Ample Voucher-Eligible Units Are Some of the “Next” Hottest Neighborhoods; Slated For Redevelopment

Map 2 Below Shows The Share of One Bedroom Apartments in the DataSF database that were Voucher Eligible at the Census Tract Level Across the City:

Share of 1-Bedroom Listings Eligible for Vouchers, By Tract
Share of 1-Bedroom Listings Eligible for Vouchers, By Tract

Notice the areas with few voucher eligible units include SOMA, CASTRO, and parts of the Mission.  There’s nothing on Twin Peaks (but there were very few listings there anyways) and not a whole lot north of Market except for parts of the Tenderloin.

I chose to overlay this map with the City’s Priority Development Areas (PDAs), areas which may see concentrated infrastructure and housing development under Plan Bay Area.  Some of the neighborhoods with the highest concentration of voucher-eligible one bedrooms are in areas slated for redevelopment: Bayview, Outer Mission, Visitacion Valley, Balboa Park, Bernal Heights, Park Merced, etc.)   If the concentrated development in these areas contributes to a decrease in the amount of voucher eligible units available, you have a case for compensation to the community for transit oriented gentrification.  But to some extent, these neighborhoods are already the next front lines of gentrification as buyers on the private market are already driving prices up (The people in the photograph for this article are a great exampleHere’s some more stats).


HUD’s policies may need to be reformed or agencies may need greater power to “flex” the FMRs to ensure residents can find decent units.  There is also a growing body of research on the possible benefits of re-scaling the FMR to the zip code level.  There are both scale and temporal issues in HUD FMR thresholds that should be re-evaluated for cities like San Francisco.  Perhaps the city could provide an additional “bonus” to vouchers so that renters can still access neighborhoods that they are currently priced out of with or without vouchers?  Maybe AirBnB could donate that kind of money instead of giving people $10 on the street randomly?  I’m grasping at solutions a bit here–but really we need to get more creative and move beyond the basic formula system if we are serious about vouchers enabling low and middle income families affording quality units in jobs-richaccessible communities.


Hello World!

My name is Matt.  I am a doctoral student who has been published and has presented in multiple conferences on housing and transportation issues in urban areas, focused particularly on the inter-relationship between transportation infrastructure and housing markets.

Current projects:

–  Mapping affordable housing production in the San Francisco Bay Area, San Diego and Sacramento regions.

–  Evaluating the benefits of folding bikes loaned to working families in Woodland, CA

–  Understanding the drivers of public opinion on ride shares and transportation and housing ballot measures in San Francisco, CA.

-Understanding how federal affordable housing policy scales and thresholds limit the effectiveness of demand-side affordable housing programs (like Sec 8 and Choice Vouchers).

I have a Masters in Public Policy from Oregon State University, and am currently funded for my PhD at U.C. Davis by an Eisenhower Graduate Fellowship from the U.S. Department of Transportation.

The first and last of these will form the basis of my dissertation research.

I can be contacted at