Subscribe to the Cicero Newsletter Learn More

Solving the Homelessness Crisis in San Francisco

In the past ten years, San Francisco has spent unprecedented time, money, and effort attacking the problem of homelessness. It more than doubled spending on homeless services, from $150 million in 2011 to $360 million today. In 2018 it passed a special business tax, Proposition C, to create another $300 million more a year in homelessness funding. It opened a special Department of Homelessness in 2016 to focus on the issue.¹

The result of this effort? The homeless population has increased from 5,700 to over 8,000, the majority of whom are unsheltered, and there are over 9,700 homeless by the city’s own, more expansive, definition, or about 1 of every 100 residents of the city.² The city’s streets have become world-famous for their dangerous and unsanitary encampments.³ The particularly devastating toll of the coronavirus outbreak on the city’s homeless population demonstrates both the city’s failures and the dangers of continuing to fail to address them.⁴

San Francisco’s tragedy is not due to national trends. Nationwide, despite the Great Recession, and despite increasing house prices, the total U.S. homeless population has been reduced from 650,000 to 550,000 over the past decade, a drop of over 15%, with the greatest drop occurring in the unsheltered population, which has fallen to under 200,000, or about a third of the total.⁵ Whatever San Francisco specifically is doing to solve homelessness, it is not working.

The Cicero Institute proposes three interrelated and evidence-based reforms to solve the homeless crisis in San Francisco, which should also be applicable to other cities with similar problems. First, the city should combine mass-produced temporary shelter for those in immediate need with competitively provided Permanent Supportive Housing and services. Second, the city should broaden mental health conservatorship for truly disabled individuals and also provide pay-for-performance contracts with service providers. Third, the city should empower a Safe Streets Court to both keep public order and to mandate the best treatment for individuals unable to take care of themselves.

These reforms will both alleviate the ills that beset the neediest people in our city, and help restore public faith in our streets. A new compact, that both provides necessary services and mandates their use, will solve San Francisco’s most persistent, and persistently shameful, problem.

Facts about the San Francisco Homeless Population

San Francisco’s homeless population is large, but not unparalleled. As a proportion of population, four cities have higher homelessness (Seattle, Washington D.C., Boston, and New York City, in that order), while three cities (Seattle, Los Angeles, and Berkeley), have a higher percentage of “unsheltered” homeless.⁶ Of large cities, however, San Francisco does have highest percentage of non-family homelessness (91%), meaning our population is largely single men, and, except for Long Beach, CA, it has the highest percentage that are “chronically” homeless, 31%, meaning they suffered over a full year of homelessness, and either the highest or second highest (with San Jose), of youth homelessness.⁷ Our homeless population is thus more vulnerable and isolated than almost any other.

Many public surveys of the homeless conflate the temporarily homeless, most often families, or mothers with children, who tend to be in shelters, with the chronically homeless, who tend to be single-men, and tend to be unsheltered. Again, it is these latter are the most familiar to San Francisco residents, and they of course represent the hardest cases. While 32% of the non-chronically homeless surveyed cited a history of drug or alcohol abuse, 63% of the chronically homeless did, and while 32% of the non-chronically homeless described psychiatric conditions, 53% of the chronically homeless did.⁸

The simple lack of housing is thus only one of many problems confronting homeless individuals. Only 12% of surveyed homeless mentioned “eviction” as the reason for their homelessness, the rest cited alcohol or drug use, arguments with family or friends, lost jobs, or divorce or breakup.⁹ A breakdown in social ties in particular seems to be one important cause of homelessness. About half of all homeless were previously living with a friend or relative before being pushed to the streets. ¹⁰ Of course, all of these issues can be caused by or related to each other, but a mere shortage of rent money for housing does not seem to be the only issue.

Yet San Francisco has focused its anti-homeless effort, and about 85% of its homeless spending, on providing subsidized or city-managed housing.¹¹ Every year in recent times, the San Francisco government permanently “rehouses” about 2,000 homeless people, totaling almost 20,000 over the past decade.¹² San Francisco currently has 2,300 emergency shelter beds and 7450 units of supportive housing for the homeless (9750 beds total), up from 1,400 and 6000 (7,400 total) just since 2016.¹³ San Francisco also has the highest proportion of supportive housing, or about 1 per 100 residents, of any city in the nation.¹⁴ Yet still our homelessness problem grows.

One reason the current housing program have not solved our problem is that expansive housing and service spending can indeed pull some individuals either into the homeless system or from other cities, even if such “pull” factors explain only a small part of our overall homeless population. About 55% of our homeless have lived in San Francisco for at least 10 years. Nonetheless, the 15% who have moved to the city in the past decade, but did not move here while homeless, could at least partially be responding to high levels of services her than elsewhere. About 30% of our homeless population, however, moved to San Francisco while homeless, most likely for because of both our climate and services.¹⁵ Some surveys of the homeless or near homeless do show they are attracted, understandably, to high service provision, including during the coronavirus pandemic.¹⁶ Although San Francisco does not separate out the origins of chronically homeless and unsheltered individuals, it is likely a higher proportion of the chronically homeless came from outside the city.¹⁷

The Need for More and Competitively Provided Housing

Although individual units of subsidizing housing on their own have not cured homelessness, there is a strong correlation between lack of affordable market-rate housing in a city and homelessness.¹⁸ The fact that San Francisco has the most unaffordable housing in the United States certainly contributes to our homelessness problem. A recent analysis estimated that merely deregulating housing in the San Francisco area, and bringing housing costs down to a national level, would lower homelessness by over half, allowing even many of those with mental health problems or substance abuse issues to rent low-cost Single Resident Occupancy units, by using already provided federal social security or disability support payments.¹⁹

The current “Housing First” consensus, however, assumes that permanent and publicly subsidized housing can be provided to any homeless individual in just about any city. One major problem with this strategy is that most homeless individuals are only temporarily homeless, and it is hard to target permanent housing at the truly needy.²⁰ A recent study in Science, for instance, demonstrated that while Chicago housing voucher provided real benefits to their recipients, only 2% of those targeted would actually stay homeless without them.²¹ Researchers today estimate that it takes at least 10 Permanent Supportive Housing (PSH) units to take one homeless individual off the streets.²² The fact that PSH in much of California now costs over $500,000 a unit means it cannot solve our problems.²³ In San Francisco, we have more than doubled our supportive housing, which would have been enough to house every chronically homeless individual back in 2010, but the total number of chronically homeless has also almost doubled in the last five years, to about 3,000.²⁴

Instead of housing any individual that comes into the system, we need two types of essential housing services for the homeless. One is mass-produced temporary shelter. Although homeless shelters today can cost hundreds of thousands of dollars per bed, Oakland has had recent success with their “tough cabins” program, using off-the-shelf sheds or cabins at low cost in a camp-like setting, for only about $20,000 a cabin, each of which houses two individuals. Increases in 3D printing and modular housing means further cheap shelter can be provided for those in temporary need.²⁵

For long-term permanent supportive housing, or PSH, we need a model that helps some people get back on their feet, and opens up beds for the truly needy. The only way to do this is through a pay-for-performance housing support system. Economist Kevin Corinth has suggested using such a model, where housing and service providers are rewarded for every assigned person who emerges to an independent life.²⁶

The model would provide a base housing cost for any PSH provider, say $4,000 a month. If a PSH provider moved an individual to independent housing, they would get $2,000 for every month that individual did not use homeless services, up to $36,000. To ensure that service providers are not penalized for reducing service use, those that move individuals to independent living would be guaranteed to receive other homeless individuals in the vacated unit, or the equivalent value, for at least the estimated time of the housing use. Those organizations who do not move homeless individuals to independence would not receive more clients. Thus, instead of paying housing providers hundreds of thousands of dollars to keep a single person in a room for decades, the service providers can compete to provide the services people need, and can open up beds for those with truly no prospects for independent living. We need to enforce accountability on current providers, instead of ending the relationship with a check and some summary inspections.

The combination of a generally deregulated housing market, mass-produced shelter, and competitive long-term housing and service provision would do much towards alleviating San Francisco’s homelessness.

The Need for Mandated Mental Health Treatment

Some of San Francisco’s homeless do have little prospect of living independently, and need the government to help them just sustain their life and well-being. As mentioned above, of San Francisco’s homeless, 42% report alcohol and drug use, while 39% report psychiatric conditions. Those with these two problems are more often chronically homeless and more often unsheltered.²⁷

Unfortunately, the current climate in California makes it difficult to get the mentally disabled or severely addicted into treatment. “Conservatorship” the main route to mandated treatment, is only reserved for cases of people that pose imminent threat to themselves or to others and who are “gravely disabled’ beyond a reasonable doubt.²⁸ There are numerous reports of how difficult it is to secure conservatorship even for the most difficult and dangerous cases.²⁹ A very minor expansion of conservatorship on the state level in 2018, and in San Francisco the following year, which allows conservatorship for those who have been detained at least eight times in the previous year, does not do enough, with some estimates that it would just allow about 6 more people to be put into conservatorship.³⁰

Conservatorship should be a viable possibility for more mentally ill individuals and their families. Although studies are sparse, some show that mandated conservatorship can indeed improve outcomes for the mentally ill.³¹ Thus the law should expand the standard of “gravely disabled,” to include those who cannot recognize their own mental disability and who pose a threat to themselves or others, and such decision should be by a “clear and convincing” standard, which is slightly more lenient, instead of beyond a reasonable doubt.³²

Under California’s 2002 “Laura’s Law,” a family member can also ask for Assisted Outpatient Treatment (AOT) of a mentally ill relative, which mandates they follow a mental health treatment regime. Although less expansive than conservatorship, it is also difficult to secure, and the individual must again be shown to be an imminent threat to themselves or others.³³ Still, studies prove it can be effective. A survey of almost 3,000 New Yorkers with mandated AOT found that while 19% experienced homeless in the three years prior to treatment, that dropped to 5% afterwards.³⁴ Other studies show that AOT reduces the likelihood for getting arrested for all crimes by over 50%, and violent crimes by 80%.³⁵

AOT should be expanded to include not just those whose threat is “imminent,” but those for whom treatment could substantially reduce future incidents of violent or destructive behavior, even if an individual has stabilized after current treatment.³⁶ If services are provided by independent organizations, they too should be paid on a pay-for-performance standard, with rewards for those who move into independent living situations or prevent continued descents back into mental illness, and a similar pay-scale as that described above. Mandating more conservatorship and AOT should help those most in need of it, instead of, as today, allowing the mentally ill to fall apart on our streets.

Safe Streets Court

San Franciscans cite both homelessness and public order as two of the city’s three biggest problems, along with housing affordability.³⁷ Every San Franciscan knows these problems are related, and in particular that untreated and uncared for homeless individuals pose a threat to both themselves and others on city streets, or simply drive many people away from them.

San Francisco’s homeless already have extensive contact with our criminal justice system. About 25% of San Francisco’s homeless reported spending at least one night in jail within the past twelve months. Many cycle in and out of jail numerous times over the course of a year.³⁸

Yet the police can help the homeless while maintaining public safety. Ever since the first Drug Court opened, in Miami in 1989, criminal justice reformers have demonstrated the importance of specialized courts that focus on non-punitive ways to return people to productive lives. San Francisco has embraced this movement, and today has eight high-quality “collaborative courts,” from a Drug Court to a Behavioral Health Court to a Veterans Justice Court, all of which address the specialized needs of certain groups of offenders.

The general tendency of these courts is the combination of “graduated incentives and sanctions” along with frequent “status” hearings, to keep tabs on their “clients.” Research has shown that regular mandated treatment through such courts can do much to move people out of cycles of law-breaking and homelessness. Research has also shown that repeated interactions with a judge are one of the most important indicators of success in such programs.³⁹ The Columbia University economist Brendan O’Flaherty argues that “aggressive but nonpunitive referral is a positive strategy” for helping homeless people with serious disabilities or barriers to employment get back on their feet.⁴⁰

Today, there is an attenuated version of a San Francisco court for the homeless. The “Community Justice Center” focuses on individuals with mental or behavioral problems in the in downtown and the South of Market Area (SOMA). Almost half of all homeless people in San Francisco live in these areas, thus the Community Justice Center serves many homeless individuals.⁴¹ Today, it has more “attendees” than all other cooperative courts in San Francisco combined.⁴²

The Community Justice Center program should be expanded into a general “Safe Streets Court” that focuses on the needs of the homeless and others mainly living on streets across the city. It should follow the practices of other collaborative courts and use graduated sanctions and rewards to improve treatment and help people adhere to the rules of San Francisco’s public spaces.

Yet the trend in California right now has made it harder to use courts to mandate services. The San Diego Tribune reported a drop in drug court treatments after Prop. 47 reduced penalties on many lower level crimes. California journalists have also reported about one homeless man who was arrested almost a dozen times in a year for trespassing, without ever spending more than a night in jail.⁴³ Instead of creating consequences for ubiquitous defecation on the city’s streets, the city continues to set up “poop patrols” to clean them, at a cost of almost $200,000 per employee. Once again restoring penalties, and enforcing them, for low-level public drug dealing, shoplifting, theft, and disorderly behavior is necessary to get people back into treatment and make an expanded Safe Streets Court effective.⁴⁴

History shows that policing can indeed make the streets safer, even while it improves the lives of the homeless. For instance, L.A. Safer Cities Initiative, pioneered by Los Angeles Mayor Antonio Villaraigosa and LAPD Police Chief William Bratton (who was later tapped as New York Mayor Bill de Blasio’s police chief), helped clean up L.A.’s Skid Row just over a decade ago. L.A. City Council Member Jan Perry said before the program’s inauguration in 2006, the area used to be like “’Dante’s Inferno.’ Now it’s like a neighborhood.”⁴⁵ In just the first year of the initiative, which demanded adherence to public order laws and mandated treatment for those who refused it, property crimes dropped by 25% and violent crimes by 33%.⁴⁶ Counter to some expectations the decline in crime did not involve pushing it to other areas, since adjacent areas actually experienced further reductions in crime. At the same time, the homeless population in the area dropped by half.⁴⁷ Yet, after the withdrawal of police in 2014, the homeless population quadrupled, and crime returned.⁴⁸ The use of Safer City type initiatives, along with a Safe Streets Court mandating treatment, can restore order on San Francisco’s streets and help the homeless at the same time.


There was a time not too long ago when San Francisco, like many large cities, had a small homeless population with reasonable social support.⁴⁹ Although the system was not perfect, the combination of a housing market with affordable rents, accompanied by temporary shelter, ensured mental health care, and police and social workers willing to mandate treatment and enforce public order, kept the city safe and the homeless housed. Such a situation is not beyond our reach. With a combination of the right services, incentives, and mandates, we can ensure that San Francisco becomes again a livable, vibrant, and caring city, for both its neediest citizens and the wider population.





[5] Although there remain some questions about the national decline, see:

About 25% of the nation’s homeless population is in California, but the total California homeless population has been relatively stable over the past decade.





[10] Nationally, the number is closer to 75%.






[16] Nationally, about 1/3 of homeless individuals who moved to a different city cite shelter or service availability as a reason for the move. For particular attraction to San Francisco during the coronavirus pandemic because of service provisions, see



[19], 2. See also impact of housing costs in homelessness in Brendan O’Flaherty, Making Room: The Economics of Homelessness (Cambridge, Mass.: Cambridge University Press, 1996), 96–164.



[22] For instance, the “chronically homeless,” those homeless for more than a year have dropped by 15% in the US, from about 105,000 too 90,000, over the past decade. But in this same time, the amount of permanent supportive housing (PSH), long-term housing specifically for the chronically homeless, has increased from 55,000 to 110,000, almost a doubling, Over 55,000 new homes have decreased chronic homelessness by less than 15,000 people. And an analysis of individual cities, shows there is almost no correlation between increases in PSH and decreases in homelessness., p. 4.


[24] chrome-extension://oemmndcbldboiebfnladdacbdfmadadm/, 34.



[27] See research on “mental health courts,” which shows that most crimes committed by mentally ill are not during psychotic states, but simply because mentally ill tend to share characteristics (anti-social history and peers, substance abuse, family discord), with other offenders. Mental illness does make it harder for the ill to readjust behavior after offenses, however.

[28] E. Fuller Torrey, American Psychosis: How the Federal Government Destroyed the Mental Illness Treatment System (Oxford: Oxford University, 2013), 86, 96–98.



[31] Such conservatorships should be “limited conservatorships,” only dealing with a person, and not their estate or property.

[32] This would be similar to AB2156 or AB 1971, both introduced last year The change in “reasonable doubt” to “clear and convincing” however, might require a new ruling of the State Supreme Court, which back in 1979, in Conservatorship of Roulet, mandated a criminal law standard for conservatorship, although a 1980 case does allow some leeway for probate conservators.–Poster_292971.pdf



[35] Torrey, American Psychosis, 147.

[36] See discussion here:


[38], p. 38.

[39], p. 12.

[40] O’Flaherty, Making Room, 280.

[41], p. 11.






[47] See even positive portrayals of improvement such as


[49] See discussion of relatively small homeless populations in major cities like San Francisco in the 1960s, and massive increases in the later 1970s and early 1980s in Kenneth Kusmer, Down and Out, on the Road: The Homeless in American History (Oxford: Oxford University Press, 2002), 239–247.