Here in California we are bombarded with news about our “housing crisis.” State politicians have used the housing crisis as justification for removing local control of zoning and handing carrots to developers. We are told that the Bay Area is the “epicenter” of the housing crisis.
Politicians and pundits who use this overblown language should review some of the reports available from state agencies and business sources. Those reports paint a far more nuanced picture.
The reports show:
1) San Francisco is not the epicenter of the affordable housing shortage. The opposite is true.
2) The state does not have a housing crisis. It does have a severe shortage of affordable housing for our lowest-income residents. This is due a combination of a physical housing shortage and simple poverty. There is not a shortage of market-rate housing.
3) 2018 population estimates show that population growth has slowed dramatically statewide, but the decline varies from county to county. Factors including fires, expensive housing, and the search at the urban boundaries for cheaper housing. Housing projections need to take these new figures into account.
4) Hundreds of thousands housing units have been proposed in California—more than enough to meet growth in housing demand statewide since 2010. While some projects are working their way through local government approval processes, most of them have been approved. For most of these projects, the construction phase is the bottleneck, not local government.
5) Growth is not an act of God. The jobs/housing ratio in the Bay Area is out of balance. The trend toward the “Manhattanization” of San Francisco has been fought sporadically for decades, and is now back on the agenda. There needs to be a serious, competent, open and democratic planning process for growth, both at the regional and state level.
This analysis is based on the following four widely available reports. However, the data has been combined to tease out some conclusions that are not well understood:
1) HCD report: California’s Housing Future: Challenges and Opportunities, Final Statewide Housing Assessment 2025. California Department of Housing and Community Development.
2) LAO report: The 2019-20 Budget: Considerations for the Governor’s Housing Plan. Legislative Analyst’s Office.
3) DOF report: E-5 Population and Housing Estimates for Cities, Counties, and the State, 2011-2019 with 2010 Census Benchmark. State of California Department of Finance, Demographics Research Unit.
4) CIRB report: New Development in California 2018. California Homebuilding Foundation, Construction Industry Research Board.
San Francisco is not the epicenter of the housing crisis
Under the standard definition, any household that spends more than 30 percent of its gross income on rent is considered rent-burdened. The chart below displays the percentage of low-income households that are rent-burdened in California. The chart is taken from the LAO report, but the graph has been truncated to save space. The original chart on page 7 of the report lists several more highly cost-burdened California counties.
Source: LAO report, p. 7.
Note that the least burdened county is San Francisco (which is both a city and a county). Low-income San Francisco residents on average have a rent burden that is lower than any other California county, and lower than the rest of the United States. This is probably due to a combination of low-income residents hunkered down in rent-controlled apartments, while highly paid techies are paying market rates.
According to the report, “The figure demonstrates that low-income households experience similar levels of rent burden across communities—suggesting that insufficient income creates housing affordability challenges even in low-cost communities that have a sufficient supply of housing. Because of this, construction of affordable housing has a key role in helping low-income households afford housing.”
The HCD report takes a slightly different approach, but the results are much the same. In the chart below, counties are ranked by the average percentage of income spent on housing and transportation. As in the LAO chart above, this version of the chart has been truncated to save space. Several more highly rent-burdened counties can be found in the original chart on page 34 of the HCD report.
Source: HCD report, p. 34.
The report shows that not only does San Francisco have a low rent burden, it also has a low transportation cost burden due to access to mass transit and jobs that are close to housing. Together these two charts show that despite the myths, San Francisco is not the epicenter of the state’s housing crisis.
California has a severe shortage of affordable housing
Statewide there is a shortage of 1.5 million housing units for very low- and extremely low-income residents. These income categories are defined based on household income with respect the local area median income (AMI). The county is usually the area used in the definition. Extremely low income is defined as less than 30 percent of AMI, while very low income is 30 to 50 percent of AMI.
An example will help make this clear. Let’s assume an AMI for a particular county is $100 thousand annually for a family of four. Then an extremely low income household makes $30 thousand annually or less, a very low income household makes $30 to $50 thousand annually, a low income household makes $50 to $80 thousand annually, a moderate income household makes $80 to $120 thousand annually, and an above moderate household makes more than $120 thousand annually.
A household at the upper boundary of the moderate income category must spend on rent at least 30 percent of annual income, or $3,000 per month, before it is considered rent-stressed. On the other hand, a household at the upper boundary of the extremely low income category must spend on rent more than $750 per month to be considered rent-stressed. In reality, the AMI for Alameda and Contra Costa counties is slightly higher than these figures. This is due in part to high tech salaries in the Bay Area. In at least nine tech companies, the median salary is more than $200 thousand annually.
The chart below compares incomes and housing availability statewide:
Source: HCD report, p. 30.
There is a lot going on in this graph, so it is useful to work through it carefully. On the left is the bar that shows renter households by comparison to AMI. On the right the bar displays rental units that are affordable for renters in the various income categories.
As the legend on the right shows, in every group except the above moderate group, some households must pay more for housing than they can afford. For example, note the upper boundaries of the light blue bars. For low-income groups and below, there is a shortage of 960 thousand units statewide—the difference between the top of the light blue bars on the left and right. For moderate income households (green bars) and below, there is still a smaller shortage of 61 thousand rental units statewide.
The gap of 1.5 million housing units is shown by the blue arrow. Note that for low, moderate and above-moderate income households, the bars on the right (light blue, green and yellow) are thicker on the right than on the left—within those groups, there is enough housing. If there was enough affordable housing for very low and extremely low income households, all income categories would have sufficient housing (these numbers are for the state as a whole and figures can vary by region).
Taking a broader view, this chart details the mismatch between renter incomes and rental housing supply. One solution is to expand the bottom of the right hand bar—increase the amount of affordable housing. An alternative would be to shrink the bottom of the left hand bar—increase incomes to push renters into higher income categories. As the LAO report stated, the real culprit here is poverty (see here, here and here). But in either case, California doesn’t need to subsidize more market-rate housing. It needs to concentrate on housing for its lowest-income residents.
California’s population growth has slowed dramatically
Housing estimates have to account for population growth. In recent years, this issue has grown more complicated. Population growth has slowed tremendously in the last few years and is falling well behind projections, but in very uneven ways.
Source: DOF E-5 Population and Housing Estimates.
The table above is derived from the population estimates of the California Department of Finance demographic unit. It shows the figures for the state, selected large counties and all nine Bay-Area counties. Population for the state and various counties is shown for the years 2010, 2018 and 2019 (in blue on the left). On the right (in green) we see the total percent change from 2010-19, the annual average for the nine-year period, and the annual percent change for 2018-19.
Let’s start with the far right hand column. Last year’s statewide population growth figure of 0.47 percent is the lowest figure ever recorded. The state population growth rate is about half what had been forecast in long-range estimates. Only two counties, Riverside and Sacramento, had population growth rates of greater than one percent. Also note that three counties with negative growth last year all suffered from devastating fires.
In addition to the three counties with fires, four counties had slower growth than the statewide average—Marin County and the three Silicon Valley counties of Santa Clara, San Francisco and San Mateo. In Silicon Valley, the lack of population growth is partly due to people leaving because they can no longer afford to live there.
Next, it is useful to compare the last two columns. In the state as a whole and in every county except one, last year’s growth rate is below the 9-year average. Population growth has slowed down. The only exception is Sacramento County. The region has become a popular destination for people exiting the Bay Area to look for cheaper housing. Estimates of future housing needs must be adjusted for these slower population growth trends.
Proposed housing is sufficient to meet statewide growth since 2010
A comparison of population growth data, along with proposed housing projects, tells another interesting and complex story. The table below is derived from the Dept. of Finance population figures combined with estimates from the Construction Industry Research Board (CIRM) report of housing projects in the pipeline.
When reviewing these numbers, it’s important to keep in mind two things. First 2010 is the base year for the DOF population reports. California’s housing shortage existed even then. Renter’s income began to lag behind rising rents at least a decade before 2010 (HCD report, p. 29, fig. 1.23). Second, the number of proposed housing projects exceeds population growth in some Silicon Valley counties. If built, this new housing would help restore the jobs/housing balance in those counties, reducing the need for commuting.
Sources: DOF E-5 Population and Housing Estimates and CIRB report.
In the left hand column are the number of persons in 2019 living in households. Note that these numbers are slightly lower than the previous population statistics because they exclude people living in dormitories, assisted living centers, prisons and other group settings. The figures in the green section show the number of housing units in 2019, and how many housing units we would need to keep up with population growth from 2010-19. In the state as a whole, and in every county in the table, there is a shortage—since 2010, housing growth has not kept up with population growth.
The first column in the blue area shows the size of the shortfall, and compared it to the number of housing units that are in the approval and construction pipeline. In some cases, there is more than enough housing in the pipeline to meet population growth, and in some cases it is the opposite. In Los Angeles and Sacramento counties far more housing has been planned than is needed to meet population growth since 2010. The East Bay counties of Contra Costa and Alameda are a little behind. Interestingly, the Silicon Valley counties of Santa Clara, San Mateo and San Francisco all have far more housing in the pipeline that they need to accommodate population growth.
Statewide, there are 801,300 housing units in the pipeline. Of that number, according to the CIRM report, 451,000 are either under construction or are awaiting developers to pull the permits and begin work. In other words, cities have done their jobs and have signed off on 451,000 new housing units statewide—more than enough to meet population growth since 2010. The bottleneck for those units is the construction industry, not local governments. The HCD report states California needs to build 180,000 housing units annually. The number of units the in construction backlog (downstream of local governments) is 2.5 years of needed housing production. In the business press, the problems with the construction industry are well known. For more on the construction industry, see these two articles in the San Francisco Business Times (here and here).
The need for democratic growth planning
Building more housing and increasing housing density in the Bay Area are often justified because they will reduce rents. But if adding density will reduce rents, why does Manhattan, the most densely populated place in the United States, still have high rents? That is the question that the Bay Area’s pro-growth advocates have yet to answer. The relationship between increasing urban population density and lower rents is not straightforward.
The housing crisis has been talked about almost exclusively in terms of supply. Sooner or later, anyone who has taken an economics course will began to wonder about the demand side of the housing market, or put another way, the job/housing balance. Growth is not an act of God. Do we accept growth at any cost? Do we want to turn the San Francisco into Manhattan and the rest of the Bay Area into Brooklyn, Bronx and Queens? Who gets to decide?
To help clarify these issues, the chart above demonstrates a stylized version of San Francisco’s office and housing markets. On the vertical axis are office units. On the horizontal axis are housing units. We define these such that one more office unit requires one more unit of housing for balance. Each office unit provides space for one more person to work, and each housing unit provides space for one more person to live. The thick grey line is the balanced-growth expansion path. Along this line rents will stay reasonable. The points on the upper right indicate places where there are more office units than housing units, and rents are high. To the lower left, there is a surplus of housing units, and rents are low.
Let’s start on the lower left, at point (0,0). Assume some time in the halcyon past San Francisco had job/housing balance. But with a sudden burst of commercial development, 4,000 office units were developed, but only 1,500 housing units. We find ourselves at point A. Rents have risen, low-income residents are being displaced, artists are moving away, etc.
From here there are two possible courses of action. The first would be to get back on the balanced growth path quickly. Commercial growth could be discouraged and held to 1,000 new office units. Meanwhile to preserve housing possibilities for current residents, 3,500 affordable units would be produced. Balanced growth and reasonable rents would be reestablished at point B with a new total of 5,000 office units and 5,000 housing units.
Another alternative would to add 3,500 more market rate units while adding 6,000 more office units. The end result would be 10,000 office units and 5,000 housing units. This is point C. Note that at point C there are double the number of office units with respect to point B. This is the road to Manhattan. Point C has moves us even further away from the balanced growth line than point A. Yet getting to point C requires building just as many housing units as getting to point B.
For developers, point C is advantageous. First, because of the influx of well-paid tech workers, new housing can be market rate. Second, with the job/housing ratio even more out of balance, rents will continue to rise, allowing developers and their allies to continue to decry NIMBYism, impact fees, high construction costs, etc. The reality is that as long as the tech boom continues, developers can maintain the housing “crisis” simply by expanding office construction faster than housing construction can keep up.
But balanced grow per se is not the complete answer. Even if we can stay on the balanced grow path, there remains a question—how much balanced growth do we want? In the short term even balanced growth puts stresses on the kind of infrastructure local governments provide. Think of it this way: When you buy a house, you don’t stop there. You have to turn your house into a home. You furnish it inside with carpets, furniture, beds, kitchenware and eating utensils.
When a developer builds a new neighborhood, all those houses have to be furnished on the outside to turn that collection of houses into a community—infrastructure like water and electricity, sewer systems, police and fire services, park, libraries and schools. Building communities is expensive and takes more time than building offices.
In the long run, too much growth, even balanced growth, leads to crowding. Californians don’t just live to work. They want to get outdoors and enjoy weekends and vacations. Many people already feel that the Bay Area is too crowded (here and here). In addition, even balanced growth can start bumping into environmental constraints—sea level rise, droughts, fire, earthquakes. None of these are made easier to mitigate by more growth and more population.
To focus exclusively on California’s “housing crisis” obscures as much as it illuminates. The issues are far broader, and require a more informed and democratic discussion of how much California should grow, and how it needs to change to adapt to its future. Those are the issues our legislators and all Californians should be discussing.