Souvla, a Greek restaurant with a devoted following, serves spit-fired meat two ways: in a photogenic sandwich, or on a photogenic salad, either available with a glass of Greek wine. The garnishes are thoughtful: pea shoots, harissa-spiked yogurt, mizithra cheese.The story might be one in which restaurant workers see the salaries offered by the information technology companies, which might also be affecting the ability of restaurateurs to hire. We might be seeing induced innovation, as California-Irvine labor economist David Neumark tells the Times.
The small menu is so appealing and the place itself so charming that you almost forget, as a diner, that you have to do much of the work of dining out yourself. You scout your own table. You fetch and fill your own water glass. And if you’d like another glass of wine, you go back to the counter.
Runners will bring your order to the table, but there are no servers to wait on you here, or at the two other San Francisco locations that Souvla has added — or, increasingly, at other popular restaurants that have opened in the last two years: RT Rotisserie, which is roasting cauliflower a few blocks away; Barzotto, a bistro serving hand-rolled pasta in the Mission district; and Media Noche, a Cuban sandwich spot with eye-catching custom tilework.
Inside these restaurants, it’s evident that the forces making this one of the most expensive cities in America are subtly altering the economics of everything. Commercial rents have gone up. Labor costs have soared. And restaurant workers, many of them priced out by the expense of housing, have been moving away.
There might be a lot more at work, though, and I'm keeping what approximates to a working paper below the jump.
San Francisco is a place where luxury apartments coexist with people sleeping in the downtown parks.
“It’s really sad,” said Jennifer Sullivan, who worked for years as a server in the area. Twenty years ago, she moved from Chicago to Oakland, where she rented a $750 studio apartment and waitressed her way through college. She fears that story would not be possible in the Bay Area now.Forgive me the impertinence, but isn't that the purpose of the Bay Area Rapid Transit, and the streetcar lines from the Pacific Coast neighborhoods of the city that dive into the Twin Peaks Tunnel and the light-rail subway under Market Street? Or the Chicago Transit Authority, or New York's famous subways? And why, dear reader, is Google notorious for running its private buses from gathering places in San Francisco to its Silicon Valley headquarters. Has Ms Sullivan never heard of a commuter train?
“I’ve even had dystopian future visions of buses full of labor that come from the outskirts of these really wealthy areas,” she said.
Then, what happens when a prosperous area becomes too expensive for the upscale eateries to do business?
A few blocks from the original Souvla, at the celebrated modern French restaurant Jardinière, the chef Traci Des Jardins said her labor costs, including taxes and health care, now eat up 43 percent of her budget.Where to begin? First, I recently got back from Normandie, where entree refers to the appetizers, and plat refers to the mains. That's a minor quibble. "Rotten economics" about setting up a full-service eatery along Fifth Avenue in New York, or North Michigan in Chicago, or in a loft building in Milwaukee's Third Ward? Not so much.
When she opened Jardinière in 1997, they were 27 percent. (Mr. Bililies said Souvla’s percentage is in the mid-20s now, even with paid vacation and retirement benefits.)
Ms. Des Jardins has experimented with raising her prices, but she said customers simply spent the same amount in different ways, skipping a second glass of wine, or ordering two appetizers instead of an entree.
At one of her other restaurants in town, she now serves lunch as counter service.
“I enjoy doing what I do, and we support a community of people here,” Ms. Des Jardins said. “But the economics are pretty rotten.”
Souvla, on the other hand, is planning to expand beyond the Bay Area, starting with New York. Mr. Bililies said he wanted to occupy “iconic streets in iconic neighborhoods in iconic cities.”
The strategy, in other words, is to go to precisely the places with rotten economics.
What, then, is making San Francisco real estate so expensive that service workers get outbid for the housing, and service businesses get outbid for the commercial space? I've looked, previously, at the broad outlines of this question, and at San Francisco specifically.
Let's consider some fundamental principles of land use.
First, to a first approximation, we can treat the land as inelastically supplied, and its rent as the residual claim to the gains from trade: that's the intellectual underpinning of any single-tax argument in the style of Henry George.
Second, to a first approximation, we can treat a city as a relatively small space supplied from a surrounding countryside. The model I'm alluding to is two centuries old, and it has limitations, including taking the location of the city as exogenous: you might consider Athens, Greece as an ancient example, the Acropolis being a high, and thus defensible, place; Jerusalem also comes to mind for the same reason; then San Francisco atop hills at the north end of a peninsula has possibilities. The model I'm alluding to also focuses on agricultural land use, although the argument it makes, an arbitrage argument, applies to any possible use of a piece of land. The value V of a parcel located at distance d from the city center in use by industry i will be one at which a type i user is indifferent between locating there, or locating somewhere else:
Vis = Yi(Pi - ci - tid),
where Yi is the output a type i industry produces, Pi the price industry i realizes in the city, ci the cost of producing that type of good, which we'll treat as the same at any location (I grappled with location-dependent costs this summer, forty years ago, those introduce complications but do not vitiate the arbitrage argument), and ti the transportation costs on good i.
Two principles of land use immediately emerge. First, activities that command the highest price Pi have a ceteris paribus advantage in bidding for any parcel of land. Second, activities that encounter higher transportation costs lose advantage at greater distances. Thus, in the two-century old model, dairy products command a high price and spoil (we're before mechanical refrigeration or railroads, recall) quickly, thus those will be closest to the city, but dairies will be out-bid by timbering or cornfields or ranching.
Introduce mechanical refrigeration and railroads and motor trucks, and the highest bidders for land might be factories or restaurants or households. The principle is the same, though, the value of a parcel of land will be one at which a householder is indifferent between locating there, or locating somewhere else. The model as modified still predicts falling land values at distances from the city center. Thus, we'll see additional margins along which people might substitute, such as suffering the transportation costs but being able to live in a bigger house: easier for a faculty member at Cornell than it might be at Northwestern or at Stanford, where faculty qualify for housing subsidies. The underlying principle is the same: in equilibrium a householder will be indifferent among the bundles of location and housing type offered. Because that equilibrating tendency leads to similar houses clustered together, providing temptations for builders to put down cookie-cutter developments, Wise Experts think they can dictate neighborhood form with zoning codes. That works until new price incentives emerge.
But will we ever see a central business district that generates so much in the form of locational rents, in the presence of so small a transportation cost, for knowledge or financial industries (and there are reasons to suspect that it will be knowledge or financial industries, not manufacturing) that the service businesses get priced out, and the only land devoted to housing of any kind is at a distance, as appears to be happening in San Francisco, but not yet in Chicago?
Perhaps the location rents are a consequence of political decisions, such as devoting space to expressways and to parking, these have the effect of taking land out of use for locational-rent-generating activities at the same time that they make working together possible. Yes, even San Francisco, according to Devon Zuegel.
Walkable, transit-friendly cities like New York, Chicago, and San Francisco are infamous for their high housing prices. However, when housing and transportation costs are considered together, these places are far more affordable than sprawling cities. It is true that rents and property values decrease as you move farther from the city center, but the added costs of owning and operating a car quickly burns through whatever savings came from moving away.That's the indifference principle, again: it doesn't matter whether it's the condo a streetcar ride from the office or restaurant, or the tract house a rapid transit ride or an aggravating commute away, if others are outbidding you, or former server Jennifer Sullivan, or Michigan economist Hal Varian, latterly at Google, or Steph Curry, to build office towers and research parks instead. Then there are those people who thought they could move to San Francisco from Manhattan in order to be able to park!
It does matter, though, if zoning codes are requiring inefficiently many parking lots or otherwise making the provision of housing ... are you telling me there are no parcels that might be more profitably put to that use ... so difficult that nobody bothers?
I know that many of our civic and political leaders are proposing ways to address the crisis. But so many of the ideas (Mandatory solar! Higher affordable housing requirements! Rent control!) would only add to the costs of housing. Your local leaders don’t approve housing because politics and financial incentives run against it. And instead of changing the calculus by encouraging them to do the right thing, your state legislators suggest new laws to coerce them.In the course of preparing this post, I discovered that the popular Sim City series of urban simulation games assume away the transport-and-land-cost increasing effects of devoting space to parking by assuming all the vehicle parking is underground. Real life isn't that simple: each additional parking space takes land away from residential, commercial, industrial, agricultural, or cultural uses, accordingly affecting the incentives to bid for the land remaining for those other purposes.
When listening to those lawmakers, I thought of an old line of mine: “It’s only because of their stupidity that they’re able to be so sure of themselves.”
Do I have any suggestions for you? Nothing that isn’t obvious. As I once wrote, “start with what is right rather than what is acceptable.” Creating enough housing, whether it’s in Prague in the 1910s or Pleasanton in the 2010s, requires creating enough housing.
Get the incentives right, and the restaurant staffing will take care of itself.