In Peak California, I argued that California was no longer the best place in the world for startups, but that the state’s economic ecosystem and tax ecosystem rely on a steady flow of new companies.
I got some feedback.
In this post, I’ll address some of the questions and critiques, and talk a little more about what to do in a world where California’s future as a startup hub is uncertain.
It takes two sides to make a market; the price of any asset, including California real estate or early-stage startup equity, is the point at which a buyer and a seller agree to transact. Since dollars change hand, these prices are weighted by both the confidence of people on either side of the deal and whatever track record led them to have dollars.
So, for any asset, at any price, if you think it’s expensive there’s a good argument that you’re wrong.
Here are some of the best I got:
California’s still attracting a disproportionate share of VC funding, which counteracts the effects of higher costs of living. Absolutely true, and this is why the peak is hard to time. If the fixed cost of three cofounders working hard on a problem for a year doubles, but the capital available to back them also doubles, it sounds like a wash.
The question is not about aggregate capital, though: it’s about which kinds of companies are relatively more sensitive to higher fixed costs; if the growth in available funding matches the growth in aggregate cost of living, there are still relative winners and losers.
I’d argue that the biggest losers are the companies whose model either a) relies on uncertain research, or b) relies on network effects. Breakthroughs happen stochastically. The right process might take a month or a decade, and there’s no good way to estimate it in advance. You can set up milestones and benchmarks, but the more fundamental the problem, the more bogus those are. So higher fixed costs select for companies that, intentionally or not, either focus on solving easy problems or target fake milestones.
Network effects are great once they work, but they’re a headwind when they don’t. Right now, Airbnb has an enormous supply of guests and an abundant inventory of rooms; they can compound both numbers with good unit economics. But getting to that point was a painful slog. When they started, every city had zero available rooms, and zero guest demand. Getting both sides of the market started required a lot of door-to-door work and no fewer than three official launches. Or, to put it in mathematical terms: if your company is on an exponential rather than linear growth path, there’s a period at the beginning where progress as a function of effort is lower than if the growth path were linear.
So those are the companies at risk: the two kinds of companies that are likely to change the world.
And anyway, funding-offset thesis begs an important question: are VCs behaving rationally, or not? If they’re being rational, there has to be some reason that higher fixed costs are a relative benefit to startups over established companies; if they’re acting irrationally, you’re betting on a trend you expect to mean-revert. Neither seems defensible to me.
That said, VC capital can keep things going for a while. VCs have lots of dry powder, and if the research is to be trusted, most of them don’t own cars and are deathly afraid of flying, so investments tend to be close by. Further research indicates that they are willing to talk to prospective LPs in Boston, New York, Dubai, and even the distant land of New Canaan.
Overall, the argument that VC funding will offset high real estate costs hinges on the theory that high real estate affects all startups equally. That’s hard to argue — high costs straightforwardly privilege lower-risk companies, and a startup ecosystem where everyone’s afraid to take risks is a startup ecosystem that won’t produce big successes.
Founders will find a way — they don’t want to work at big companies even if the cash comp is higher. That may be true, but we already have a case study in how founders respond when they want to start companies and their geographical constraints make it harder. They move.
You’re just jealous/this is a lame breakup letter/California has dealt with crises before. Maybe! I find the breaking-up-with-a-city genre boring, because a lot of them boil down to someone realizing a given location was a bad fit. I once read a long, long article about how bad it is to live in Japan, by someone who moved there to work as a salaryman and said it was a not-fun place to be a teetotaling vegan. True, but maybe something to figure out before you get there.
My point in Peak California was that many of California’s plans are the non-obvious side effects of self-inflicted problems. I’m not saying California is bad — ”peak” means it’s never been better — just that it’s poised to start getting worse. It’s always important to know what derivative you’re looking at, though: if California is peaking due to fixable problems, that means fixing the problems will make the state more prosperous than ever.
What To Do Next
The key problem is housing supply. Other California issues are derivative. More housing reduces burn rates at startups; it reduces pressure to raise pension fund payouts; it even solves some of SF’s quality-of-life issues, by giving poor people an option between sleeping rough and paying $2,500 a month for a studio apartment.
S.B. 50 is a proposal to allow higher-density housing near transit hubs. This is a start, but it’s a great start — public transportation tends to serve lower-income people, who are the highest risk of getting priced out altogether.
The details are complicated, but I found a nice explainer here, on a site called Livable California. The site appears to be one of those niche parody news sites — what Duffel Blog is to the military and The Hard Times is to punk music, Livable California is to NIMBYs. The jokes are a little over-the-top, but you get the idea. Here are some of the key points:
- “SB 50 encourages 75-ft and 85-ft-tall luxury towers in single-family areas that are either too close to transit or too close to jobs and good schools.” This is great; schools and jobs are convenient places to live next to.
- “SB 50 openly threatens “sensitive communities” — low-income, diverse areas. It requires them to upzone their Community Plans in 5 years to conform to SB 50, annihilating their homeowner areas, a direct attack on starter homes.” Also good! Since homes are so expensive in CA, a “starter” home is out of reach unless you have an extremely good job, so the existence of such homes is just a nice perk to the already well-off. If we can reduce rich-person benefits while giving the middle class a lot more access to affordable housing, that’s a win.
- “Developers can build apartment towers with NO PARKING.” Given that these are buildings near public transportation, that seems like an obvious approach. The developers of the Sim City series have said in interviews that they chose to omit parking because a true top-down view of a city full of parking lots is both boring and depressing.
If S.B. 50 passes, we’ll be able to see if additional supply reduces the market-clearing price for a given level of demand, or if Tokyo is, as NIMBYs believe, some sort of bizarre anomaly that’s somehow able to supply apartments at one third of SF’s prices despite a growing population.
Startups can open satellite offices earlier and accept more remote work. Have a headquarters where the check-writers and dealmakers live, but move as much of your staff as you can to China (like Zoom — which had the advantage of dogfooding their own product) or the Netherlands. Devs are cheaper. Time zones are tough, but note that the longer your working hours, the more your work hours overlap with everyone else’s. And high real estate is a head tax. Your cost of spending a month in SF is fixed by your rent, which means your marginal cost of more SF working hours is zero — and that’s a sweet deal. Also, thanks to the popularity of Jocko Willinck, it’s trendy to get up at 4:30am. The perfect time for a conference call with Mumbai or Warsaw.
GET SOME calls in with the engineering team before you 5am kettlebell sesh
Silicon Valley real estate might actually be a fantastic deal. Let’s turn to the dark side: suppose housing supply-restrictionists win, and the Bay continues to suffer from a housing shortage. And let’s further suppose that the tech industry keeps growing. What’s the optimal move?
It’s probably to own Bay Area real estate. Think about it: VC funding mostly pays for operational expenditures, and the single biggest piece of opex is salary. Higher salaries raise rents. Other opex lines also flow to rent, albeit indirectly: many startups have to pay for advertising, or cloud services, and both of those represent revenues for big tech companies that allow them to expand their operations.
So, as tech companies increasingly determine demand for housing, and supply doesn’t change in response, every Bay Area landlord becomes a levered, zero-carry investor who owns a tranche of senior preferred stock in a portfolio of tech companies.
It’s the deal of the decade. You don’t have to beg for an LP allocation at a hot fund, or settle for one at a less marquee name. The minimum investment is lower. You don’t even have to be accredited. You do miss out on the peak upside — once a company is mature enough to generate free cash flow, opex growth slows relative to revenues, by definition. But that also makes the cash flow safer. You won’t get the upside from a spectacular investment in an amazing company, but you’ll get a piece of all the action.
In the long term, it’s a bad bet. Peak California implies that eventually, after-tax incomes for Californians will start to look pretty bad. But if other real estate owners are looking at rents and cap rates, and you’re the one tracking Preqin stats on fundraising and scraping CrunchBase to see where the funding goes, you can be the first to know it’s time to sell.
But that long-term is very long indeed. Right now, programming is a young industry. But industries age. As the median age of tech employees creeps up, they’ll be hitting their peak earning years and starting to form families. San Francisco doesn’t have many kids now — the joke is that on Halloween, the only people in SF who go trick-or-treating are the Thiel Fellows — but delayed family formation is still family formation, and family formation makes people a) more willing to buy a house, and b) less tolerant of high risk and all-nighters. If we model techies as choosing between high-base and high-upside jobs, with base salaries determining real estate demand, a gradually aging tech workforce is yet another tailwind to real estate demand.
Betting against bubbles is spiritually rewarding but financially painful. Someone who shorts a bubble without a catalyst is like a matador without a cape or a lance. Riding bubbles out when you know what drives them can be immensely profitable.
There might be more real estate than we think. One-bedroom apartments in SF seem to sell for about a million dollars apiece. But a friend pointed out an opportunity to buy a big block of one-bedroom units for $99k apiece. Check it out. Sure, they’re small apartments. Yes, you might get cabin fever. But you’re saving 90% on your biggest, dumbest expense. What a deal! It’s not even the first time this has been tried.
This idea is obviously absurd, but it’s not precisely obvious why it’s absurd. When the status quo makes silly thought experiments hard to argue with, there’s something wrong with the status quo.
I remain pessimistic about California, but I want the state to succeed. The tech industry does a lot of good for the country, and is the single best way to get rich. It would be a disaster to throw that away because we’re afraid of shadows and want to get a little bit more return on our disastrous bet on housing.