Yes, this study found that new housing drives down nearby rents

Does supply and demand work for housing the way it’s supposed to? Does San Francisco, in particular, represent a weird freak of market rule that reverses the traditional effects? These questions drive both our housing policy and the general public’s reaction to it.

Traditional thinking holds that housing shortages drive up prices, and thus creating new housing relieves demand and pushes prices down—easy.

But in the Bay Area, especially in SF and on the Peninsula, contrary thought holds that factors like inflation and appreciation will always work faster than we can build. Indeed, neighborhood activists often insist that new housing drives rents up by gentrifying new areas and making them attractive to even more high-priced development, at the expense (literally) of existing residents.

In December, the Upjohn Institute, a Michigan-based employment research firm, published findings about the effect of new housing on housing prices in large U.S. cities, including in San Francisco.

The results are not likely to resolve local arguments—in fact, they will leave few people happy. Nevertheless, here are some key highlights:

• Researchers created an initial sample pool of 1,483 buildings with 50 units or more built between 2010 and 2019 in 11 cities, San Francisco among them. Then, they created a baseline effect by tracking average rent prices in similar units within 250 meters of the new building after construction, and compared this to buildings within a larger 600 meter area to act as a control.
• Generally speaking, the results were that for three or so years after a building’s completion, the adjusted effects on rents in the surrounding neighborhood “hover[ed] around zero,” per Zillow data, and then afterward declined, by an average of about five percent.
• Extending the radius further out to 400 meters from the subject building found a similar but smaller phenomena, with prices dropping about 2.8 percent. Although the paper acknowledges that the further away you get the less ideal the experiment was, the results “still strongly points to a negative rent effect of the new building.”
• The institute’s writers note that fear of gentrification from new construction is widespread and meaningful, “[playing] a large role in the defeat of major housing reforms such as California’s Senate Bill 50” (which is not quite dead yet) but conclude that the concerns are “mostly unfounded,” at least according to their models.

So at first glance this seems like an absolute win for the supply-side argument.

However.

The big catch—and the reason why these numbers are unlikely to resolve the arguments in SF—comes in the form of the most common bugbear for any statistical outing: sample size.

The team started with nearly 1,500 buildings to study. However, after narrowing that pool down with a variety of additional standards, they ended up with just 92.

While these 92 buildings are supposed to be meaningful examples—hence the rigorous standards that disqualified the hundreds of others—that small sample leaves critics with contrary opinions a lot of room to criticize.

And how many of those final buildings were in San Francisco? By the time all was said and done, just one.

So while this research does suggest that most housing markets react to traditional supply and demand incentives the way one would expect, the possibility that San Francisco (for whatever reason) is a special case remains frustratingly unresolved.

As usual, the cry of the responsible researcher carries the day: Further research is required.