Why has the gig economy been a disappointment?
Maybe because traditional companies still have a good reason to exist.
Startup investment tends to go in waves; there’s usually one or two “big things” that a lot of people are rushing to put money in. Right now it’s crypto, a little while ago it was A.I. But back in the misty dawn of time — i.e., about five years ago — it was the gig economy. Uber’s valuation was stratospheric, and everyone was trying to come up with “Uber for X”. Remember those days?
Gig economy mania faded when it became apparent that the business model didn’t make sense for a lot of things — “Uber for X” just didn’t work for most values of X. And thus, an industry that was supposed to transform the face of the U.S. labor market ended up doing very little in terms of changing the way people work.
But there were still those few big successes — Uber, Lyft, DiDi, Airbnb, Instacart, and DoorDash. And most recently, there was OnlyFans, which wasn’t quite “Uber for porn”, but did allow lots of people to make an income freelancing. (It’s not quite clear where the gig economy ends and the “creator economy” begins, of course; OnlyFans has elements of both.)
But now those success stories are also under pressure.
Uber, Lyft, and network effects
First, let’s take Uber — the tentpole company that convinced (almost) everyone that the gig economy was a big deal. If you ever use the service, you’ve noticed that prices have gone way way up, and wait times — which represent an additional hassle cost of using the service — have also risen. A year ago, that could be waved away as an effect of the pandemic; now, with economic activity coming back, it looks like a more long-term change. Cory Doctorow has a long article about the situation, and while it’s too polemic, it does make some good points.
Most importantly, Uber is still losing money despite the price hikes. As Hubert Horan (an analyst who has been gunning for Uber forever) notes, its profit margin was -38% in the first half of this year, even as other Covid-risky activities like restaurant dining have been restored to mostly-normal levels. The company is also cash-flow negative. It releases an “adjusted EBITDA” metric that looks better but really just excludes a lot of expenses that shouldn’t be excluded (like stock-based compensation).
This can’t continue, of course. But what will stanch the bleeding? The idea that self-driving car breakthroughs will save the company was always wrong (autonomous vehicles turn ride-hailing into a commoditized product, since each vehicle can switch seamlessly between any number of apps), but now it’s apparent that Uber is not going to be a leader in self-driving cars anyway. So what’s going to make the company profitable?
One hope was always that investor subsidies would allow Uber to drive competitors like Lyft out of the market, and then jack up prices to extract monopoly profits. But prices have been jacked up, and Lyft is still around (and is even slightly less unprofitable than Uber). Where are those monopoly profits? And even if Uber could drive out Lyft, which looks increasingly unlikely, its network effect isn’t global, meaning that local competitors can easily enter the market in 5 years or 10 years or whenever, and compete effectively in one city or country without attaining anything like Uber’s scale.
A more likely case is that the equilibrium demand for ride-hailing is simply not nearly as big as investors had hoped. Uber got a lot of people to start using the service by subsidizing rides, but that didn’t hook people into becoming dependent on ride-hailing. When prices go up, most people will just take the bus, or an e-bike, or a scooter, or the train, or their own car. Even with no Lyft or other competitors in the market, Uber will simply never be a monopoly, since it’s competing against other kinds of transportation. And getting a human being to chauffeur you around is an inherently expensive form of transportation.
So it seems likely that Uber and Lyft will survive, but not at the scale investors hoped — instead, they’ll mostly be boutique services for the well-heeled. And I expect they’ll probably take a hit to their valuations.
OnlyFans and legal liability risk
OnlyFans, meanwhile, has decided to ban sexually explicit content, which has left a lot of people asking what the service will now be about. When Tumblr banned porn, it basically died — everyone stopped using it, and its valuation dropped from $1.1 billion to $3 million. Why did OnlyFans decide to shoot itself in the fake butt?
One possible reason was that it was having trouble finding investors, even with great financials. Another possible reason — and the excuse the company gave — was that banks didn’t want to process the company’s payments. It could just be that the financial world is naturally conservative and doesn’t like porn.
My bet, though, is that it’s mainly about legal liability. The FOSTA-SESTA anti-trafficking law made internet companies liable for non-consensual material displayed on their platforms, as well as illegal ads for sex work. Though the line is fuzzy and there have been lots of attempts to negotiate the boundaries and carve out exceptions, there is little question that there has been a chilling effect on internet porn businesses. PornHub is among the many companies that have taken a hit.
For a company like OnlyFans, the obvious solution is to stop being just a platform and start being an actual integrated business. Actually check the government-issued IDs of the people on the platform and make sure they’re of legal age and make them sign a consent form (Update: As a commenter points out, the company already requires IDs, but could do more extensive vetting). Of course some creators will be reluctant to divulge their real identities to a company, but that’s basically how you do it. Of course, this would more cost overhead — in other words, making the company less of a gig economy type of thing. Legal liability is one more reason why the business of throwing up a cheap platform and letting third parties go hog-wild is an inherently limited model.
DiDi and regulatory arbitrage
A lot of Uber’s initial success came from its ability to undercut the taxi monopoly in cities like NYC. Users loved Uber so much that the city government was basically forced to accommodate them and legitimize ride-hailing. But that trick didn’t work in China, where the government’s response to consumer demand is notoriously less accommodating. The government is now capping prices and penalizing the company for collecting user data. Its stock has fallen substantially.
Now of course, this could be unique to China, which is increasingly looking like a command economy, and has its own reasons for crushing various tech companies. But it underscores the fact that gig economy companies tend to be sellers of fairly prosaic services — taxi rides, hotel rooms, sexy photos, etc. If DiDi was creating a complex, technologically novel thing, China’s leaders would probably be more eager to appropriate the tech for themselves and turn it to some nationalistic purpose, rather than just smacking the company around. Instead, it appears that Xi & co. have decided that DiDi is simply using tech as a way to squat on a piece of virtual land, as it were.
The revenge of Ronald Coase
Some other gig economy companies are still doing well — DoorDash and Instacart, for example. Airbnb is bouncing back after the pandemic in a way Uber so far is not (and I’m glad it’s bouncing back, because I depend on Airbnb a lot when I travel). But these are turning out to be the exception rather than the rule. And although it’s possible to find a bunch of idiosyncratic reasons why each of the other big gig companies are struggling, I wonder if there isn’t a more unified explanation here.
That explanation could have something to do with why companies exist in the first place. In 1937, the economist Ronald Coase theorized that companies exist because they minimize transaction costs, like search and information costs. And as the internet has reduced those costs, outsourcing and offshoring have risen, just as Coase would predict. The gig economy was really one more step along this path — the use of technology to expand the set of labor markets that could use temp work and day labor, which are a sort of micro-outsourcing.
But it’s possible that the gains from that sort of micro-outsourcing are simply inherently limited in a lot of markets. Driving yourself somewhere, or riding a bike, is still usually cheaper than hiring a chauffeur. Vetting porn actors to make sure they’re not trafficked or underage cuts down massively on the transaction cost of legal liability. And so on. Gig platforms definitely cut down on transaction costs somewhat, but maybe not enough to justify huge valuations and a massive investment boom in most cases.
If that’s true, it challenges one part of my thesis about the coming internet-driven productivity boom. I theorized that remote work tools would enable more micro-outsourcing between companies — a sort of B2B gig economy:
Right now, outsourcing happens at the level of business functions — you outsource payroll, or IT, or logistics, etc. But a shift to remote work might allow outsourcing at an even more micro level. Need someone to brush up your presentation before the big meeting? Need someone to run a regression or wrangle some data for you? Want someone to pore over a company’s financials and report key metrics? With remote work tools — and more importantly, with everyone getting used to a standardized package of remote work tools! — this sort of micro-outsourcing might become possible, enabling even greater specialization and flexibility in the production of office tasks. (Update: Adam Ozimek of Upwork calls this practice “hybrid teams”.)
But if transaction costs are still a big deal in most parts of the consumer world, maybe this will be a more minor trend than we’d hope.
This is also relevant for DAOs, which basically represent the ultimate endpoint of the outsourcing process. In a world where the internet (including blockchain) basically obviates all search and information costs, DAOs would be the natural way to do everything (in fact, this is sort of the premise of Rainbows End, one of my favorite futuristic sci-fi novels). But the disappointment of the gig economy could be a sign that this corporation-less sci-fi future is still a ways off.
AirBnb is doing well since it provides a way around zoning rules so that one can operate a hotel in a residential neighborhood. Since it can benefit property owners who are also voters and taxpayers, it is hard, but not impossible, to regulate. It drives up local rents and property values, so it can produce a lot of value while strangling a local economy.
AirBnb allows one to maximize the return on one's capital. Uber and the like don't allow one to maximize the return on one's labor. The original Uber ride sharing idea was that Uber would be a spare time operation, but it quickly professionalized when subsidies made it profitable. When the subsidies vanished, a lot of the professionals vanished.
I think Coase was slightly off target. Companies exist because they can cross subsidize operations to support an efficient internal system based on highly optimized and specialized knowledge. A profitable, efficient company is always going to have some components that cannot be justified internally by their own profitability but are critical to their overall success. The necessary transactions could not be justified if outsourced, but can easily be justified if the transaction costs can be hidden. That's hidden, not necessarily minimized. The secret sauce is often that those costs need not be minimized.
One example is Amazon. It internally subsidized the construction of AWS and its logistics systems. It was unable to show a profit for a decade, but as a company it was able to hide the internal cost structure until a dominant position was achieved. Even now, Amazon has a lot of invisible internal cross subsidies. When it starts farming out the relevant operations to Mechanical Turk, it will be time to short the stock.
Another example is Boeing which hid the cost of its engineering based internals. Even assembly line workers were expected to think like and work like engineers. Engineers were expected to have contact with the factory floor and deal with every component of an aircraft. This let them design and build aircraft profitably. When they made the internal subsidies visible, they started outsourcing, relying on contractors and lower cost labor. Now they can barely build aircraft, and it is unclear if they will be able to take the next step in aviation.
In Israel, taxis have always been obtained from many small companies, or even individual drivers who built a customer list. Their response to Uber was swift and effective. An app mysteriously appeared, that all drivers use to respond to your call in the same taxis as before. So they pretty much replaced one dispatcher for each company with the app, and life continues as before.