San Francisco Considers Not One but Two Bills Legalizing 'Missing Middle' Housing Citywide. Both Have Flaws.
The otherwise positive proposals are undermined by affordability requirements and density restrictions.
In what might seem like a Christmas miracle come early, the San Francisco Board of Supervisors is considering not one, but two, bills that would legalize lower density "missing middle" housing across the city.
Competing proposals introduced by Supervisors Gordon Mar and Rafael Mandelson would both allow the construction of four-unit homes (or fourplexes) on all residentially zoned land citywide. Combined with state-level reforms from earlier this year that make it easier to divide residential plots in half, both bills could theoretically allow up to eight primary residences where only one was permitted before.
Unfortunately, the two proposals also include micromanaging regulations that would lead to less missing middle housing being built than a more hands-off free-market approach would produce.
Mar's bill would permit up to four units of housing on all current Residential House zones, which currently allow between one and three homes. That sounds like a pretty sweeping reform. But there's a catch.
Mar's bill would require the new units to be rented out or sold at rates that are affordable to someone making 100 percent of "area median income." The San Francisco Chronicle, which first reported on the bill, notes that the current area median income in the city is $106,550 for a couple or $133,200 for a family of four.
Affordable monthly rent for a family making that amount of money would shake out to be $2,664, according to a press release from Mar's office—$2,000 less than pre-pandemic market-rate rents for a typical two-bedroom apartment.
"While we've overbuilt luxury housing, the housing industry has not been building for our workforce," Mar said in a press release. "I am proposing a new strategy to stabilize, protect and grow our middle class."
Mar's proposal certainly wouldn't produce much "luxury" housing. It also probably wouldn't produce much housing period, says Emily Hamilton, a researcher at George Mason University's Mercatus Center.
Requiring monthly rents be $2,000 below-market-rate rents is "a substantial subsidy," Hamilton tells Reason. "I would imagine units that are delivered under this bill would be either existing housing that could be chopped up into more units or perhaps provided by a nonprofit homebuilder or someone who is motivated by something more than profit."
Consider what happened in Austin, Texas. In 2019, the city technically abolished single-family-only zoning when it created the Affordability Unlocked program, which allows developers to build larger projects with more units and fewer parking spaces in exchange for making the new homes affordable to lower-income people. Specifically, it allows the construction of up to eight units of housing in single-family-zoned areas. But to build those extra units, a developer would have to make as much as 75 percent of the new units affordable to people earning below area median income, include a certain number of two-bedroom units, and adopt a host of tenant protections.
Mar's bill would likely have similarly disappointing results. One could even consider a form of "affordability trolling," whereby new housing is legalized but cynically inserted requirements prevent the units from being built.
Mandelman's fourplex legalization bill looks like laissez faire in comparison. It allows the construction of fourplexes citywide, without any of the affordability requirements in Mar's bill.
Nevertheless, it would require newly legal fourplexes to be built at densities no larger than what the city's current zoning allows for three-unit homes. (Mar's proposal has the same density restrictions.) According to Hamilton, that means Mandelman's bill is more suited for permitting triplexes than the fourplexes it technically allows.
To overcome the limitations of their dueling zoning reform ordinances, both Mandelman and Mar are also proposing companion bills to incentivize the construction of fourplexes.
Mandelman introduced a bill in February that would require people trying to build large single-family homes to get special conditional use permits (a long, expensive process in San Francisco). The idea is to steer people away from building "McMansions" and instead build more affordable multi-developments.
"The way much of San Francisco is zoned today makes it easier to flip existing housing into luxury monster homes than to build small apartment buildings for working people," Mandelman said in a July press release.
Mar has proposed streamlining permitting and providing direct financial assistance to homeowners who add below-market-rate units on their property. "By centering homeowners," said Mar, those incentives will "protect and build homes to meet our actual housing needs and to stabilize our community."
Hamilton thinks policymakers interested in affordability should consider more ambitious proposals that would allow larger multifamily housing in more areas of the city and speed up the process by which new buildings are approved.
Not only would that produce more housing overall, but it would also produce more of the affordable "missing middle" housing.
Without modern zoning restrictions, U.S. cities historically built a lot of housing affordable to working- and middle-class people, from Baltimore's rowhouses to New England's "tripledeckers."
Those were erected "not because of some policy intended to promote missing middle housing but because that was what made sense to build in those locations," Hamilton says.
California has made great headway on zoning reform at the state level this year, passing one bill that legalizes duplexes on almost all residential land in the state and another that makes it easier for cities to zone for smaller apartment buildings without having to go through onerous environmental review.
But both reforms still leave the details of this upzoning to cities to figure out. As we can see from the Mandelman and Mar proposals, that's where the devil is.