How California can build 3.5 million new homes

California’s new governor, Gavin Newsom, has called for California to build 3.5 million new homes in the next five years as part of a “Marshall Plan for housing” to reduce housing costs and homelessness. For decades, California has added jobs faster than it has built housing, and this is what it’d take to restore the balance and end the shortage.

visualizing 25% more homes by 2025! (and hopefully at least 25% cheaper rent!)

That’s the equivalent of nine new San Franciscos (want to live in SF 10.0?), or an entire new L.A. County. Sounds ambitious.

Let’s build another one of these, what could possibly go wrong?

Not a fan of sprawl? Perhaps you might imagine endless skyscrapers, like in Blade Runner’s Los Angeles of 2019.

It’s not that dramatic. Look at this another way. There are currently 14 million homes in California. Adding another 3.5 million is a 25% increase.

Increasing California’s housing supply by 25% is similar to adding a bedroom to a 4-bedroom house. We can do this.

What does California’s housing currently look like?
- 14 million homes
- 60% single family houses
- 20% townhouses, duplexes, and small apartments up to 9 units
- 20% large buildings with 10 or more units

If we shrink this down to a neighborhood of 1,400 homes (1/10,000th of the state) it would look something like this.

1/10,000th of California’s existing housing

Adding 25% more homes

One way is to evenly distribute it. Accessory dwelling units (in-law flats) would sprout up in backyards. Expanded a house into a fourplex. On a typical block of 20 houses, build one new home each year and you’ll meet your 25% by 2025 goal.

In urban areas it gets more challenging. Densification would mean replacing parking and existing housing. Some cities guarantee tenants relocation benefits and the right to move back into one of the new apartments at the same rent. This is still disruptive and causes concerns about gentrification.

Evenly Distributed 25% Increase

Also, we probably shouldn’t add more homes in areas at risk of wildfire or sea level rise, and there are also plenty of places far from jobs that might not be appealing anyway. The places where the housing shortage is worst is in coastal cities such as Silicon Valley, where the new jobs are.

How about we just put some highrises in Silicon Valley? While there are a handful of proposals that do this, such as the Kylii project in Santa Clara, towers are rare for a number of reasons.

Everyone new lives in that one building

First, there is cost. Highrise construction can cost over twice as much, and the homes produced are priced in the millions. Then, there is the political landscape — height is a sensitive issue. While proposed upzoning legislation Senate Bill 50 calls for more homes in “jobs-rich” areas, it does not intend to change existing height limits in these areas, and even near transit height increases are limited to just a few floors. Finally, not everyone wants to live in a highrise.

How about those six-story apartments that seem to be going up everywhere? Chances are you’ve seen at least one of these block-long “fast-casual” style apartment buildings being built at a transit station near you, replacing a gas station or shopping center.

A couple of yimblocs

These are cheaper to build than highrises, but still costly — in San Francisco, construction costs alone can be over $400,000 per apartment. Add in land, design, financing, school impact fees, permit review costs, etc, and the total cost reaches $700,000 or more — over twice what a median income can afford.

What makes these buildings costly is that a third of the space is taken up by hallways, stairs, lobbies, and garages. Their large size also means lengthy construction times (more interest paid on construction loans), and lots of high-paid executives to manage the project.

While these midrise buildings are a big part of California’s housing pipeline, they can’t do the job alone.

Building 3,500,000 new homes in California will take a mix of approaches. Some places, especially near jobs and transit, will get large buildings. Some will get low-cost accessory dwelling units and duplexes. And some places, such as dense historic neighborhoods or outlying rural places, will see little or no new development.

The All of the Above Approach

Paying for it — the $2 trillion question

At an average cost of half a million each, 3.5 million new homes will cost close to $2 trillion to build. (I’ve rounded up to include additional costs for transportation, utilities, etc.) Most would be privately funded, though state and local governments are also investing billions in affordable housing.

Currently, California builds about 100,000 homes a year. Ramping up to 700,000 homes would represent over $300 billion dollars of new activity a year — a 11% bump to the state’s $2.7 trillion economy. Not quite enough to make us the world’s 4th largest, but still a huge rate for an advanced economy. With the right timing, building affordable housing could ride us out the next recession.

Who Will Build It?

During the Great Recession, most construction workers were out of work, and many found new jobs, moved, or retired. Currently, about 100,000 Californians work in construction. Meeting the 3.5 million new homes by 2025 goal would require a few hundred thousand more — almost enough to give every unemployed Californian a job.

Construction of new homes by raising existing ones.

We’re still going to need rent control

3.5 million new homes may solve the housing problem at the state level, but in high demand neighborhoods such as those near downtown, rents will remain high. Rent control, rent freezes, and other anti-gouging legislation are part of a wider “Protect-Preserve-Produce” strategy that the Bay Area Metropolitan Transportation Commission and other organizations advocate for.

Housing for All!

Let’s do this — we’ve done it before

In the mid-1980s, California built 200,000–300,000 new homes a year, back when the state population was about 2/3rds of what it is today. 3.5 million new homes in the next few years is an attainable goal, and one we ought to work towards.

The SketchUp model used to make the diagrams above can be downloaded here.