The case for banning non-competes
Making the labor market work
As part of a larger raft of competition policy announcements made earlier this week, Joe Biden says he’s going to ask the Federal Trade Commission to try to ban non-compete agreements — clauses of contracts that prevent workers from switching from working for one company to going to work for another company in the same industry.
I think there are some technical questions as to the FTC’s actual ability to do this, but the good news is that there is also some bipartisan support in Congress for a legislative effort to curb non-competes, so I don’t think our hopes rest entirely with the FTC.
The non-compete issue is interesting to me both because it’s important in its own right, especially for lower-wage workers, but also because it implicates a big picture theoretical issue that I don’t think gets discussed enough. The question here is roughly something like “is innovation difficult, or is innovation expensive?” If innovation is expensive then you’ll support all kinds of policies that enrich the owners of the most innovative enterprises, because maximizing the financial returns of innovation is critical to maximizing the quantity of innovation-facilitating financial inputs.
But suppose instead that innovation is difficult. Lukas Walton has a net worth of $22 billion because his grandfather founded Walmart. If he decided he really only needs $5 billion in life and is going to plow the other $17 billion into a massive R&D effort, how much innovation will he really generate? I think likely not that much, because (no offense) he’s just some guy who happens to be rich, not a visionary inventor or a brilliant investor. I don’t think you can just turn on the spigot and innovate. And if the primary impediment to innovation is that it’s objectively difficult, then you’ll be interested not in maximizing the financial rewards but in maximizing the diffusion of ideas.
These priorities point different directions in a bunch of areas, including, notably, patents, but I think they make a crucial difference in the non-compete debate.
But first, some boring labor market considerations
That’s all a bit eccentric, though.
First and foremost, the issue with non-compete agreements is that normally in the United States we give employers and employees broad latitude to agree to whatever they want to agree to. Back 100 years ago, the Supreme Court was really hard-core about this and decided there was a fundamental “freedom of contract” that minimum wage or mandatory overtime laws violated. The contemporary United States isn’t like that, but we do nonetheless retain a less regulated labor market than most countries.
The case for super-strong freedom of contract stems from thinking of the market for labor as a spot market for a commodity — oil or grain or copper ore or whatever — where you have many buyers and many sellers and free entry and exit into the marketplace.
But of course, that’s not the reality. And in recent years, some economists have started talking a lot about the labor market as subject to endemic monopsony issues. There are a lot of ways to think about this, but probably the most intuitive one is that an employer can comfortably let one or two or three vacancies go unfilled for a while. If it’s a big enterprise, you can easily carry a bunch of vacancies. You obviously do want to fill them, but it’s not a crisis if those positions go unfilled. By contrast, for the typical worker, joblessness is a crisis. Unemployment benefits are meager (in normal times) and have a sharp expiration date. A family normally has only one or two workers, so for even one of them to be out of work is a much bigger blow than a vacancy is from an employer’s perspective.
This means an employer has a lot of options — if you want to fill your vacancy really, really fast, you need to offer a high wage. But it is completely viable to just decline to offer a market-clearing compensation package, fully aware that the vacancy will linger for a while, but confident that someone good enough will come along soon enough and say yes. And we clearly see that labor markets typically don’t clear that rapidly, and anyone who’s ever been in the kind of middle management role where the bosses tell you to fill a vacancy but also constrain the pay you can offer can see that it works this way.
Beyond that asymmetry, there are other kinds of sources of monopsony power. Searching for new jobs is time-consuming. In many industries there are a limited number of employers in a given area, and if you churn too quickly among them you’ll burn your reputation. People build whole family schedules around particular commuting patterns, and switching is costly.
In general, there’s a lot of reason to believe that monopsony is a significant factor in labor markets and that therefore regulation can be beneficial. Still “regulation can be beneficial” is a fairly weak claim. We don’t have utopian regulators who can perfectly see the consequences of everything. Where do we get the confidence that a particular rule barring a particular kind of contractual agreement would be a good idea?
Non-compete agreements in particular
The main idea of the non-compete agreement is that employers want to stop people from walking off the job and taking trade secrets to rival companies. If companies weren’t able to secure those protections, they’d need to pay lower salaries, and we’d all be worse off.
In big picture terms, the case against non-competes has gathered steam from two big sources.
One is the observation that California has a lot of policy problems, but does seem like the best place in America for innovative high-growth companies to launch. California is also unusual in that state laws make it impossible to enforce non-compete agreements there. So if you think non-competes help foster a robust innovation economy, it’s hard to understand what Apple and Google and Facebook and Stripe and all these other companies are doing there. California has acute housing scarcity problems, high taxes, and a lot of other issues. So they must be doing something right. And a lot of people think the lack of non-compete agreements is exactly the thing that they are doing right.
The other is that a few years ago, people started noticing that Jimmy John’s was making new hires sign non-compete agreements. And they were not alone. Researchers found that a surprisingly large share of low-income workers was bound by non-compete agreements, which don’t seem to be even notionally directed at the preservation of company secrets. Instead, agreements in this context seemed like a way of translating potential workers’ momentarily weak bargaining power into permanent lack of leverage vis-a-vis their employers.
That spurred a move in some states to start barring non-compete agreements for low-wage workers. But Biden’s proposal, which is more sweeping than that, clearly takes inspiration from the California case as well as the facts about absurd non-competes at sandwich chains.
The California dream
In the early days of the computer industry, Silicon Valley was rivaled by the Route 128 corridor (this is a road through the Boston suburbs) in Massachusetts as a hub of innovation. But California won out decisively here, and many people think the lack of non-compete agreements in the Golden State is part of the reason.
To see why, think about why Mark Zuckerberg moved Facebook to the Bay Area when he decided to get serious about it. He and his investors believed in Facebook’s promise. They thought it was possible that it would become a huge, super-successful company. For that to happen, you need the ability to literally grow the enterprise at a rapid clip — hiring lots of technology workers, hiring lawyers who are familiar with tech sector legal issues, hiring HR people who are used to dealing with computer guys, the whole deal. And the best place in the world to do that is the Bay Area, because there are already a lot of tech companies there.
But also, if you are graduating college and starting out your career in the tech industry, the best place to work is in the Bay Area because there are lots of startups launching there all the time, so you’ll have plenty of opportunities to hitch a ride to a potential rocketship. But also if your rocketship fails, there are plenty of big corporate tech employers there you can go work for.
To a considerable extent, this hinges on poaching. Silicon Valley was a good place to grow Facebook because fast-growing Facebook could poach from older tech companies. And Silicon Valley is a good place to launch a startup today because you can poach from Facebook.
Innovating is hard
This is a good story, but note that it doesn’t necessarily establish that federal action is a good idea. It could be the case that California is essentially doing a beggar-thy-neighbor policy. By implementing pro-poaching regulations, they themselves poached the startup economy from the rest of the country. But the poaching itself is bad because it reduces the financial returns to innovation by making it hard to safeguard trade secrets.
That just seems wrong to me. Was Steve Jobs sitting on stealth innovations that he could have unleashed if only the personal financial returns to him were a little larger?
I doubt it. In fact, Jobs got surprisingly little financial upside from the iPhone compared to the founders of other giant tech companies since he sold his shares during his absence from the company. His widow Laurene Jobs Powell’s fortune is mostly Disney stock obtained from the sale of Pixar rather than Apple stock. And Apple itself generated its most successful innovation at a time when its financial resources were meager compared to what it has available today. The new, richer company isn’t rolling out innovations at two times the pace of Jobs-era Apple because that’s just really hard. You can’t will innovations into existence with resources and incentives; you need ideas and vision, and they are just objectively in short supply.
In other words, while letting companies better defend their business methods might let them capture a larger share of the upside of innovation, it seems unlikely to me that it would meaningfully increase the amount of innovation. On the contrary, the diffusion of ideas out from innovative firms could have significant social benefits.
The spread of ideas
There are a lot of reasons why the New York Times is at the top of the news game.
But one of them is that when challenged by startups, the Times refused to be disrupted. Of the original team at Vox, there are now six — Max Fisher, Amanda Taub, Brad Plumer, Eleanor Barkhorn, Sarah Kliff, and Ezra Klein — who have gone to work at the NYT across a bunch of different sections. And they’ve also been joined by several later-stage Vox hires. But it’s not like Vox is unique. The Times has hired a lot of stars from Politico over the years. They hired the former editor-in-chief of Buzzfeed. That hasn’t just given them great journalistic talent — it’s allowed them to do institutional learning from successful digital-native startups and incorporate a lot of those ideas into their practice.
That’s taken courage and vision from Times executives and top editors, who could easily have been complacent instead. But it’s also required them to be able to poach. If the digital news startups had locked their workers down, the Times wouldn’t just be lacking individual journalists — they’d be lacking certain elements of digital-savvy that they imported directly from the innovators.
And in my view, that’s good.
One of the big, overarching problems of our time is that productivity growth has slowed down sharply from where it was during the post-WWII decades. But researchers are increasingly finding that this isn’t necessarily because the pace of innovation per se has slowed down. Instead, the future is unevenly distributed, and a minority of companies have seen their productivity surge forward without that productivity diffusing to a majority of firms.
I wish I could tell you that banning non-competes would single-handedly solve the non-diffusion problem, but I don’t see any reason to think that’s true. What is true is that if banning non-competes makes it harder for innovative companies to keep their ideas locked up, that would be a good thing.
We need a world where the incremental gains made by Jimmy John’s in efficient sandwich-assembly spread rapidly to Subway and Jersey Mike’s.
Make labor markets great again
I started with a discussion of old-fashioned freedom of contract ideas versus newer economic research on labor markets as shot-through with monopsony issues.
Some people really like those monopsony findings because they serve as a license for all kinds of regulatory interventions. But the reality is that even when you have the economic criteria in place for “regulatory interventions may be welfare-enhancing,” there are still a lot of practical barriers to creating regulations that are actually welfare-enhancing instead of just rent-seeking.
So I think it’s good to actually do what we can to tackle monopsony issues, and that’s why rules against non-competes are an especially good form of regulatory intervention. Monopsony is part of what makes intervention warranted, but the intervention also directly reduces monopsony power. Full employment monetary policy also helps here by forcing employers to behave more like price-takers than like price-makers. But things like zoning rules that restrict geographic mobility and occupational licensing laws (also targeted by Biden’s order) also play a role here by weakening people’s ability to move around and try something new.
In the right circumstances, even investments in transportation infrastructure can cut labor market monopsony. And my guess is that so can investments in a robust but sensibly designed welfare state that preserves a strong financial incentive to work without threatening the jobless with total immiseration.
Some amount of monopsony seems like a feature of the human condition, and there will always be a need for some form of socially embedded labor market. But we ought to try to create a situation where we have more rather than less competition here, and banning non-compete agreements is a good step down that path.