A plan to end street homelessness in five years, Part 2: Mitigation

We can get everyone inside for $142 Million

It's possible to end street homelessness in five years. All we need is political will, a little bit of money, and leaders who reject the status quo.

Ending street homelessness requires a three-pronged approach: Mitigation (this essay), Prevention, and Treatment. Mitigation policies will get everyone indoors, prevention policies will keep people in their homes, and treatment policies will help those who can't help themselves.

Before reading the plan below, please read Part 1: Types of Homelessness. And subscribe now so you don’t miss future installments.

Fast mitigation

The number one priority of any plan in San Francisco should be to get all 5,000+ individuals that sleep on the street every night into shelter. At first, we only need to provide the very basics: a roof, a cot, a blanket, a toilet, a sink, a shower, and a locker. Not a single person in San Francisco should ever go without these basic necessities.

But, perhaps surprisingly, new prefab shelters are more economical than the large-tent-style congregate shelters, while also providing better privacy and safety to the residents. So let's skip tents and go straight to cabins.

Small pre-fab shelters

The LifeMoves Mountain View project at 2566 Leghorn St provides low cost and high quality shelter for about 124 people as interim supportive housing on about 42,000 square feet, with about one quarter of that land reserved for parking and another quarter for medical and case management offices, shared bathroom facilities, a dog kennel, and a dining area.

DignityMoves proposed a similar installation at 33 Gough St in San Francisco. It will provide 70 temporary supportive housing units and onsite amenities on two parking lots that are about 19,800 square feet combined. The rough price tag is about $25,000 per unit. (Side-note: this project needs your support! Donate here.)

This shows an average density (excluding the parking) of about 3.5 shelter units per 1000 square feet1, including the space needed for shared amenities, healthcare, and social services.

Getting people into shelter should be considered a public health emergency, so we shouldn't waste time finding the perfect sites or negotiating land purchases or leases -- let's just use parking lots. There are 41 parking lots and garages owned by the city in San Francisco. These range from 10-story garages to small surface parking lots. For this policy proposal, I catalogued the square footage and heights of every city-owned parking lot.

Setting a minimum square footage of 9,000 square feet (about three residential San Francisco lots combined), and excluding underground garages and one farmer's market, we're left with 20 locations suitable for pre-fab shelters. The screenshot below shows their locations, with garages in purple and surface lots in blue.

Applying our average shelter density on the combined ~650,000 square feet at these locations, we have room for 2,275 pre-fab shelters2, at a cost of about $56.9 million3.

We need at least 5,000 shelters, so where else can we get the room?

More floors and congregate shelters

On the multi-level parking garages, we're only taking one of the floors, but we can take more and use some of the space for large congregate shelters. Yes, the congregate shelters cost a bit more but they house more people per square foot.

Let's use 444 Stockton Street and the 5th & Mission garages.

444 Stockton is a 10 story parking garage in the middle of downtown, with the main tower having a floor area of about 40,000 square feet per floor. The top four floors of this garage total 160,000 square feet and provide enough space for about 560 shelter units. 

The 5th & Mission Garage is ideally sized to host multiple congregate shelters. Each floor is 120,000 square feet with no intrusions into the envelope. Using the Navigation Center tent at 600 The Embarcadero as a reference, three tents totaling about 15,000 square feet serve about 200 people on a site that's roughly 43,000 square feet for about $12.5 million (about $62,500 per unit). That lot is irregularly shaped, so the density of 4.6 shelter units per 1000 square feet is below what the 5th & Mission garage could accommodate.

The top floor of this garage could hold roughly 10 large tents plus shared facilities, giving shelter to over 660 people, at a density of 5.5 people per 1000 square feet. We could also use the next level down for pre-fab shelters, adding another 420.

This puts us at a total of about 3,500 shelter units (see spreadsheet for breakdown). Lowering the square footage threshold to include the very small parking lots only brings the total just north of 3,700 units, but I doubt the administrative cost of running very small sites makes sense.

The very rough estimated one-time cost is about $112 million to do this plan.

Housing for humans, not cars

3,500 shelter units is 1,500 units short of the immediate need. Where will that extra space come from?

If we, as a city, decide that we care more about housing people indoors than providing underpriced parking spots for cars, we can house *everyone* in individual pre-fab shelters (no tents!). We just need to take half of every SFMTA parking garage.

Check out the "Maximum prefab" sheet in the sites spreadsheet. Every surface parking lot over 9,000 square feet, and the top half of the floors of every parking garage, are used for prefab homeless shelters.

We can do all of this for only $142 million and provide almost 5,700 homeless shelter units. 

Every one of these locations would have individual and family units, showers, kitchens, social workers, lockers, and pets would be welcome. Every single person sleeping on the streets would have somewhere safe and clean they can sleep every night while they get back on their feet.

What else can we do

If you really love parking and don't want to use so much space for getting people indoors, there is another (much more expensive) option: congregate shelter tents everywhere combined with hotel acquisition.

Using the "Maximum tents" plan, which puts congregate shelters on all of the sites (excluding some floors in 444 Stockton due to ceiling height), we start with about 4,500 units at a cost of $240 million. That's not enough, but it's a great start. The other units could come from extending the very successful hotel shelter model that has operated during Covid-19 by acquiring at least 1000 hotel rooms.

I do not have access to a list of available real estate that the city could purchase, but I know the total hotel capacity in San Francisco is over 34,000 rooms. I can't estimate how much acquiring enough properties will cost, unfortunately, because many hotel operators may be looking to exit the market and might sell at a steep discount, or may be holding out for the economic recovery.

The city maintains an artificial cap on hotel rooms, like it does with almost everything else, and should abandon this backwards policy so the hotel industry can build back capacity as tourism recovers post-Covid. By trading the purchase of hotels for liberalized regulations around new hotel construction, not only will the city add new shelters, but would lower hotel rates long-term while boosting tourism income.

These acquisitions would have to be financed by municipal bonds. Let's be pessimistic and say this would cost $500 million for 1,000 units. San Francisco voters routinely pass bonds for this amount, like this $600 million bond for affordable housing from 2019. But it would also take at least a year to write the bond, get it on the ballot, and have voters approve it (as all municipal bonds must be enacted).

I really don't like this plan because it is the slowest and most expensive option, it's less predictable since it's predicated on the electorate passing a bond and the city finding willing sellers, and it doesn't provide any privacy to the shelter residents.

We can't afford not to

The Maximum Prefab model would cost the city only $142 million, which is less than what the city currently spends per year on much worse facilities. (For reference, San Francisco's yearly city budget is over $13 Billion.) If we declare a homelessness state of emergency, the Federal government would even foot a big chunk of the bill. 

San Francisco desperately needs to increase tourism and business revenue, and our street homelessness crisis stands in the way. Tourists and residents don't feel safe, and don't want to visit or live in a dirty city.

We got a glimpse of our city in decline during Covid-19. We saw people flee our dysfunctional government, crime skyrocket, and businesses get wiped out. That could happen again, this time without a pandemic. We must fix our most shameful and inhumane problem before it's too late.

This is Part 2 of “A plan to end street homelessness in five years”. Read Part 1: Types of Homelessness, and subscribe now so you don’t miss future installments.


(124/(42,000 * 3/4)) + (70/19,800) = 0.0035 shelters per square foot = 3.5 shelters per 1000 square feet


650,000 * (3.5 shelters/1000 sq.ft.)


2,275 * $25,000