Technology-based companies such as Uber and Airbnb have disrupted more than just their business sectors. They’ve raised complicated questions about how they should be regulated and by whom. New research from Wharton could help decision-makers sort out the answers. Sarah Light, Wharton professor of legal studies and business ethics; Eric Biber, law professor at UC Berkeley; J.B. Ruhl, chair of the Vanderbilt Law School; and James Salzman, environmental law professor at UCLA, have written a paper titled, “Regulating Business Innovation as Policy Disruption: From Model T to Airbnb.” It takes a historical look at the challenges and offers an analytical framework for regulators to respond. Light, who recently conducted a seminar on Capitol Hill about the paper for the Penn Wharton Public Policy Initiative, recently joined the Knowledge at Wharton radio show on SiriusXM to explain. (Listen to the full podcast above.)
An edited transcript of the conversation follows.
Knowledge at Wharton: What is the current state of regulation of companies that exist on a digital platform, like Facebook and Twitter? It feels like we’re just starting what could be a very long process.
Sarah Light: Each of these business and technological innovations has been prompting a lot of thought in the regulatory space about whether something like the internet can be regulated the same way as telephones, whether Uber and Lyft should be regulated the same as taxis, whether Airbnb should be subject to the same regulations as hotels. This raises a whole host of questions, not just for the entrepreneurs who are building these new forms of business based on new technologies, but also for regulators in terms of who should respond and how they should respond. Should the regulation be at the local level, at the national level? There’s a whole host of questions that all of this innovation is raising.
Knowledge at Wharton: It feels like technology is about 10 to 15 years ahead of where the regulators are right now.
Light: The time sequence is a really key point here. When we think about existing laws and regulations, often they are designed with a kind of specific technology in mind or a specific vision of the economy in mind. Then new business arises — be it in the sharing economy, be it the internet, be it autonomous vehicles — that challenges the existing regulatory schema in a way that creates what my coauthors and I in this paper call a policy disruption.
There’s a technical term in management literature coined by Clayton Christensen about disruptive innovation, and that refers to an innovative business that is somehow undercutting a major, established firm by drawing away customers. The term gets thrown around a lot. Whenever I give this talk, I ask everyone in my audience, “How many of you have heard the term disruptive innovation?” Every single hand goes up. But most people don’t realize that it has a technical meaning in the business and management literature.
“Uber is the classic example of what we call an end-run.”
The challenge for someone like me, who works on law and regulation, is what is the regulatory response? Is there a relationship between disruptive business innovation and policy disruption? That’s what my coauthors and I have explored in this paper and we’ve come up with our own schema of different types of policy disruption.
What we’re trying to do is make clear that this isn’t just about the sharing economy. The sharing economy has created a lot of policy disruption, but it’s not the first time that a new form of business or technological innovation has created a policy disruption. We try to abstract it and say that we need to be historically minded about this. The same issues arose when the internet was developing, raising questions for regulators as to whether this should be regulated more like a telephone or like something else. The issues keep coming up again and again, so I think it’s important to keep that historical perspective in mind.
Knowledge at Wharton: Uber really disrupted the traditional taxi industry. From a policy perspective, it has been a very important point of contention in a lot of cities because ridesharing companies are not subject to the same regulations as cabs.
Light: Depending upon your perspective as to whether Uber is good or bad, you might call what they’re doing regulatory entrepreneurship or regulatory arbitrage. In cities around the country, they have taken the position that they are not the same as a taxi, so taxi laws don’t apply to them. They have lobbied very actively to get laws passed at the state level that would preempt local governments from being able to have their own regulatory rules.
In our schema in this paper, we come up with four different types of policy disruption. Uber is the classic example of what we call an end-run. An end-run is the idea of a policy disruption where, notwithstanding similarities to the incumbent industry, the innovative business argues that it shouldn’t be or isn’t subject to the regulations governing the incumbent. This is Uber saying, “We should not be subject to taxi rules. We shouldn’t have to purchase medallions. We shouldn’t be subject to supply caps. We shouldn’t be subject to rate regulation. Our drivers shouldn’t have to have fingerprinting and the same kind of criminal background checks, etc.” That’s a very clear example of an end-run. But that’s not the only kind of policy disruption.
A second type of policy disruption, which is exemplified by Airbnb, is what we call an exemption. An exemption is a case in which the new business fits into an explicit legal exception in the existing regulatory framework, but the way that the new business is scaling up is creating the same kind of policy problem that the original regulation was designed to address. Here’s the example with Airbnb: There are fair housing laws in the United States with respect to home sales, home rentals and hotels. Under civil rights laws, you’re not allowed to discriminate in renting out a hotel room.
There are, however, exceptions under these laws for owner-occupied residences. If I choose to rent my couch to a roommate because I need the money, I can choose whoever I want, for whatever reason I want. There is an express exemption because people should have more freedom in their homes. As a woman, maybe I have safety concerns about renting out my couch to someone. The problem is that Airbnb has scaled up so massively that they are renting out millions of rooms annually and effectively competing with the hotel industry.
When you aggregate the individual decisions not to rent to people of a particular gender or race, it has the same impact as a function of discrimination as if the exemption weren’t there for the hotel industry. So, that’s what we call an exemption, and this has obviously raised really serious issues. There was a study published about a year ago finding that people who had African American-sounding names were 16% less likely to have an Airbnb rental made to them. Airbnb hired the former U.S. Attorney General Eric Holder to advise them on anti-discrimination policies.
Knowledge at Wharton: There are two other types of policy disruption that you bring up: gap and solution. Can you explain those?
Light: Gap is the situation in which there is no regulatory regime in place. The thing itself is so new that we don’t know what to do with it. Maybe there are background judicial doctrines, like contract law or property law or tort law, that would deal with some of the effects of this new form of business, but there’s no regulatory regime. A good example would be automobiles at the turn of the century. We just didn’t have rules.
“It seems like the states should be allowed to step in and write their own rules until the federal government has some uniform standard.”
Another example that has come up in the Tesla context is that many states have laws that protect the franchise relationship. When an auto manufacturer wants to sell its vehicles, it often does so through a vehicle showroom that isn’t owned by Ford or GM. It’s owned by John Smith in New York. For many years, that relationship was governed by ordinary contract law. The problem was that the Fords and the GMs had too much power over the franchises with whom they contracted, so states created new laws designed to protect these franchises in that relationship of unequal bargaining power.
The final category of policy disruption is what we call a solution. A solution is where the innovative business solves a regulatory problem but may be blocked by over-inclusive legal rules. The example that we give is something like distributed solar generation. The idea that you could put solar panels on your roof is solving a problem. You’re reducing greenhouse gas emissions, so that’s a good thing. The problem is that many states have rules that say that if you want to connect to the electric power grid, you need to apply for a permit through the state public utility commission, because the assumption behind the regulatory system is that anyone who wants to connect to the grid is an investor-owned utility like a large coal or natural gas-fired power plant.
Imagine a situation where the existing legal rules would say, “Oh, you want to connect to the power grid with your solar panels? You need to file an application with the state Public Utility Commission,” which is expensive and time-consuming. How likely are you to put the solar panels on your roof? Probably not so likely. The over-inclusive existing legal regime might block what it is that you want to do, and you have this solution that could potentially solve regulatory problems.
Knowledge at Wharton: Where do autonomous vehicles fit into this?
Light: Autonomous vehicles are really interesting. It’s important to understand the background of how ordinary vehicles are regulated in the United States. In 1966, the Congress passed the Motor Vehicle Safety Law, which is the primary statute that governs vehicle safety in the United States. It basically says the federal government is responsible, through the Department of Transportation, for regulating motor vehicles and motor vehicle equipment. The states get to regulate everything else, meaning the driver. When you get your driver’s license, when you get insurance, that is governed by state rules. Traffic laws are state and sometimes local. That’s the division of labor — there’s the car on the one side and the driver on the other.
Autonomous vehicles raise this interesting policy disruption question of, “Are they the car, or are they the driver?” They’re clearly creating a policy disruption. Whether you think that they’re a good thing or a bad thing is the question. Whether you think that the existing legal rules are under-inclusive or over-inclusive determines whether they fall into the end-run. They’re trying to get around existing rules versus they’re a solution.
They could potentially save many lives by avoiding the problem of distracted driving, and they’re potentially blocked by over-inclusive legal rules. So, they could potentially fit into either of these two categories. But what is happening right now is that Congress is attempting to write new legal rules. The House has passed a bill, and the Senate has been debating a similar bill, that would say autonomous vehicles are vehicles, and the federal government gets to regulate them.
The challenge with these two proposed bills are, in my view, that they are insufficiently protective of consumer safety because the laws themselves don’t actually write safety rules. They don’t say in the statute itself, “This is how quickly an autonomous vehicle must be able to stop.” Those rules get written by the Department of Transportation. But there are no such rules yet for autonomous vehicles. So, the laws would essentially say, “The federal government gets to regulate, the states can’t regulate and, by the way, the federal government hasn’t written any regulations yet.” From my perspective, it seems like the states should be allowed to step in and write their own rules until the federal government has some uniform standard. That strikes me as more protective of safety.
Knowledge at Wharton: With all of this uncertainty, especially using the example of autonomous vehicles, how will regulators be ready when autonomous vehicles get approved, as opposed to three or four years after these vehicles are on the roads?
Light: There’s a lot of dialogue going on between auto manufacturers and regulators about what the best standards ought to be. There is an industry organization called the Society of Automotive Engineers, SAE, which sets best practices standards. The current state of play is that more than 20 states have written laws that allow autonomous vehicles on the roads. In most cases, a human driver must be present and monitoring the vehicle. A couple of states have said that a human driver need not be physically present in the vehicle. The federal government has taken a somewhat advisory role up to now.
Knowledge at Wharton: Going back to Uber, it considers itself a technology company and not an automaker or a taxi service.
Light: That’s been their argument. That argument has not been super successful. At a certain point, they had to concede that they are going to subject to some kind of regulation, and that’s why they lobbied for laws at the state level so that they wouldn’t have to go to individual municipalities where the interest group pressure in favor of taxis is stronger. They concede that they’re required to provide insurance for their drivers. They have been willing to put their drivers through certain kinds of background checks, although not fingerprinting. That was a big issue in the city of Austin. Ultimately, Uber pulled out when the city of Austin insisted that the drivers be fingerprinted. There are slightly different rules for taxis. A new identity called a transportation network company is what they are called under most of these laws.
Knowledge at Wharton: You also talk about these policy responses that are out there to block a free pass, which a lot of companies would love to have. How does that play in with the four that you were mentioning before?
“These firms don’t want to have to deal with a patchwork of different local rules.”
Light: Just in the same way that we categorize the types of policy disruption as falling into four different buckets, we argue that there are four primary policy responses.
When confronted with some kind of innovative business that’s creating a policy disruption, a regulator can choose to block. You are not allowed to enter this market. The example would be if the city of Philadelphia, as it did several years ago, impounded vehicles that were being used for Uber rides. The second would be a free pass, which basically says, “We’re going to let you in. We’re not going to regulate you at all.” This is what a lot of entrepreneurs would love. Uber’s initial argument was that they shouldn’t be subject to regulation at all.
The third type is what we call old reg — the idea that even if the existing legal rules are an imperfect fit, you do your best to apply them to the new form of business. Rather than crafting a new set of legal rules for Uber, Lyft or Airbnb, you say, “You know what? They’re close to a taxi. We’re going to make them follow the taxi rules.”
Finally, new reg is the idea that you have to write new legal rules. This would be the example of states over the past several years that have passed a law that governs what they call transportation network companies. They create a new legal category for Uber, Lyft and their ilk. They create different legal rules for the new form of business, or they scrap the old legal regime for taxis and say, “Everybody’s going to be regulated as a transportation network company. It doesn’t matter if you’re organized as a taxi fleet or if you’re organized as Uber and Lyft. All the drivers need to do X, all the vehicles need to have this amount of insurance.”
Knowledge at Wharton: As technologies develop, we will have innovations that will cross over into a couple of different regulatory territories. Who becomes responsible?
Light: This has been a really important issue, particularly in the sharing economy context but also in the autonomous vehicle context. In the sharing economy context, it’s been important because a lot of the existing businesses, like hotels and taxis, are subject to local regulation. But these firms don’t want to have to deal with a patchwork of different local rules. They are national firms, so they want to, in some cases, be subject to national standards.
Just in the same way that there can be a policy disruption, there can also be what I call a federalism disruption. I think the best and most clear example of that is in the autonomous vehicle context, where I mentioned earlier that the current law says the federal government regulates the vehicle and the equipment, whereas states regulate the driver. But it’s not clear whether the autonomous vehicle is the vehicle or the driver. The federal government says it’s the vehicle. But I think there’s a fair argument that at least the software is the driver. Computer programmers are perfectly capable of programming different speed limits for different states, so why shouldn’t they be programming different ethics for different states or different behavior for different states?
