Alexis C. Madrigal is a staff writer at The Atlantic and the author of Powering the Dream: The History and Promise of Green Technology.
A lot of software developers, according to an unprecedented new analysis.
There has never been a town like the one San Francisco is becoming, a place where a single industry composed almost entirely of rich people thoroughly dominates the local economy. Much of the money that’s been squished out of the rest of the world gets funneled by the internet pipes to this little sliver of land on the Pacific Ocean, jutting out into the glory of the bay. The city now sits atop a geyser of cash created from what the scholar Shoshana Zuboff calls“behavioral surplus”—the natural resource created from your behavior, which is to say your mind.
Literal colonies of the working poor now cling to forgotten streets in RV communities. Homeless encampments are stitched onto any liminal plot of land. To lose your apartment doesn’t mean moving one neighborhood over but three cities away, to Antioch or Gilroy or Stockton.
But wait, it gets worse.
This year, eight major tech companies are expected to hold initial public offerings. The first, Lyft, took the public-market plunge last month. Yesterday, Pinterest did. Airbnb, Instacart, Palantir, Postmates, Slack, and Uber remain. Amazingly, all but Palantir are headquartered in San Francisco, currently home to only five other public software companies—Dropbox, Salesforce, Square, Twitter, and Yelp.
Every Uber ride in Minneapolis makes a Bay Area Victorian a smidge more expensive. Every small business running ads in Little Rock, Arkansas, raises a tower a tiny bit higher. Every Pinterest board in Provo, Utah, reshapes this place, where people went to prom and repaired mufflers and dreamed of parrots and poetry. San Francisco is now the town that apps built.
And while digital space is seemingly infinite, San Francisco has an extremely limited housing supply. Only 5,471 properties changed hands last year out of almost 400,000 housing units. So what will happen when billions of dollars in stock options can become cash anytime someone clicks Sell?
The common wisdom is simple: housing armageddon.
But even the end times have a structure. Much of what the world knows about the tech world’s effects on San Francisco’s real-estate market comes from three sources: house-hunting lore (“They bid four hundred grand over asking! All cash!”), realtors talking up their industry’s prospects, and aggregated market data from firms like CoreLogic. The numbers point to crazy market dynamics: The median home price hovered around $1.3 million in 2018.
But precisely because the tech industry has become so ubiquitous, blending in seamlessly with the old-line wealth generated by hometown firms like Bechtel, McKesson, Levi’s, various banks, and more obscure fortunes, it’s been hard to disentangle what all those engineer salaries and options are doing in the world.
At least until Deniz Kahramaner got interested. He’s a 20-something Stanford-trained data scientist turned real-estate agent, and he wanted to understand who was driving the local housing market. When he founded Data Bay Area, a real-estate group affiliated with the unicorn start-up Compass, he came into a common data set of property records. Title companies, which are the internal machinery of the real-estate market, generate business for themselves by giving away the data on who owns all the properties in a city.
“Historically, realtors have used it to spam people,” Kahramaner told me. But as he looked at the records of every property purchase in San Francisco, his data-science background saw not marketing information, but analytical potential. Most realtors think about where property is purchased, not necessarily who is doing the buying. “I thought, Wow, this is an incredibly rich data set. You can see who bought what,” he said. “ Why is no one analyzing this!?”
So he did, creating an unprecedented data set about the nature of San Francisco’s home buyers that allows his analysis of the potential effects of the IPOs on the city to go one layer deeper. His research suggests that the boom is going to be spikier than anticipated, concentrated in just a few neighborhoods, at least at first. It will also proceed more slowly than most people are anticipating. Shares are generally locked up for six months after a company goes public, but the bulk of the money probably won’t enter the market for a year or two, Kahramaner believes.
But don’t get too excited: The money is enormous, and the prospective new buyers dwarf the available inventory, especially at the top and bottom of the financial scales.
Kahramaner is a person matched precisely to his work. Identifying and analyzing data about individual people has been his literal job—first at LinkedIn, then at Accompany, which built software to create profiles of businesspeople through their public internet footprints. The title-company data presented a very similar challenge: He had a name and a city on the title. If he could figure out who they were, he could create a data set of home purchasers in the city, which, to my knowledge, doesn’t exist anywhere else in industry or academia. Then, he could use it to answer questions like: How many Uber employees bought homes in San Francisco in 2018? Or, more structurally, what percentage of San Francisco home buyers come from the tech industry? And maybe, ultimately, how much and where might a flood of new millionaires affect the city most?
So, Kahramaner built software that searched the web for the names of the property owners in the data set, and connected them in big graphs to the entities listed in their LinkedIn, Crunchbase, or other business profiles. Then, he paid humans to go through and match the purchases up with companies and industries. It’s painstaking work, which helps explain why no one else has built such a data set.
It’s also not definitive. First of all, foreign buyers and the very wealthy often use LLCs to purchase homes anonymously. Second, there are buyers with generic names who can’t be definitively connected to a company. Third, there are people with no public profile at all.
But, taking all that into account, Kahramaner’s team ultimately found industries and companies for 55 percent of the home purchasers in San Francisco in 2018.
Fully 51 percent of them worked in software. They bought in specific, desirable neighborhoods closer to San Francisco’s tech companies, as well as the highways and train lines that lead south into Silicon Valley. They were less likely to buy in the foggy Sunset, which is the worst commute to tech businesses.