The Sharing Economy and the Future of Digital Governance


DAVID: Arun Sundararajan is a professor at New York University’s Stern School of Business and an affiliated faculty member at NYU’s Center for Data Science and NYU’s Center for Urban Science and Progress. He is here today to discuss the sharing economy and the future of digital governance. Please join me in welcoming Arun Sundararajan. [APPLAUSE] ARUN SUNDARARAJAN: Thank you, David. And thank you, guys, for having me. Thank you for taking time out of your lunch break to learn more about the sharing economy. That’s a picture of me at Google the last time I visited four years ago. I was trying out your bicycles. So this is the book that you’ve been given. I have been a professor at NYU for more than 15 years I was never inclined to write a book until I wrote this one.

It wasn’t part of the plan. The reason why I was motivated enough to write this book was that I started getting interested in the sharing economy through some examples. They seemed similar but in different industries. You had Airbnb. You had Uber and Lyft. You had food-sharing sites. You had crowdfunding sites. And over time I started to see a pattern that they shared. And I started to see them grow at the pace that they did. And I began to conjecture that we were seeing the emergence of a new way of organizing economic activity. And so I wrote the book to try and unpack what that new way of organizing economic activity is, what its economic impacts are going to be, how it’s going to change what it means to have a job, and what it means for the future of how we regulate business in the consumer interest and in the employee interest. The subtitle of the book has what I think is the most accurate label for this new way of organizing economic activity. I call it crowd-based capitalism. That’s not the title of the book because “The Sharing Economy” maximizes the number of people who kind of know what I’m talking about.

And that was my publisher’s decision rather than mine But let me explain to you what crowd-based capitalism is by giving you some examples, many of which will be familiar. So when I was a kid, I would watch television. I’d watch television with my friends. The television programs that I watched, the content itself, tended to be created by a large corporation, typically a television studio or a television network. And then that content was delivered to me through the infrastructure of another large corporation, either a government-owned television station or a privately-owned television station. I have a 12-year-old daughter now.

She watches a little television, but she watches YouTube. She watches YouTube the way that I would watch television. She watches YouTube with her friends the way I watched TV with my friends. And if you think about YouTube, it’s a platform owned by a large company. But the content that is being delivered on YouTube is pretty diverse. Some of it is created by an individual amateurs. Some of it is created by a group of people. Some of it is created by small businesses. Some of it is created by large studios. And so we’ve replaced the model, or we’re starting to replace the model, of corporate-produced content– video content– with a model of corporate-delivered video content but where the content is created by a distributed crowd of people of varying scale and varying levels of professionalism. And so this, to me, is sort of a microcosm of what’s happening in a lot of other industries. Philanthropy now has a crowd-based alternative.

If you want to be a philanthropist, you can on Kickstarter. Again, you’ve got a platform that is aggregating requests for philanthropy, but the philanthropists themselves vary in their scale and it’s typically a distributed group of people who are collectively crowd-funding the cause that they believe in. Traditional venture capital is being challenged by a crowd-based venture capital where a platform aggregates promising companies, but the funding is actually provided by a distributed group of people. Some of them are professional investors. Some of them are amateur investors. If you want to loan, a small business who wants a loan can go to a bank. But they can also go to a growing variety of crowd-based lending platforms. Let’s take Funding Circle as an example. It’s based in the United Kingdom. The typical loan ask on Funding Circle is 50,000 pounds. The difference between the 50,000 pounds that you get from Funding Circle and 50,000 pounds you get from a bank is that you’re providing the same kind of documentation on Funding Circle as a small business, but the typical loan is funded by over 200 lenders, some lending as little as 20 pounds.

So we’re replacing the model of institution-based small business lending with a crowd-based model of small business lending. Many of you have used to Getaround, a peer-to-peer car rental company. There’s a competitor, Turo, which is also a peer-to-peer car rental. The platform aggregates demand, but the supply of the cars come from individuals who want to monetize their cars when they’re not driving them. So, again, we’re replacing the Hertz and Europcar model with a crowd-based model, but a platform aggregates micro-entrepreneurs of sorts who are running their own tiny car rental businesses through the platform. We’re used to getting our groceries from a grocery store, like big food. There is a small, but growing number of peer-to-peer alternatives.

I think the one that I’m most familiar with is called La Ruche Qui Dit Oui. It’s based in France. It has 700 different locations now. And it’s a platform that connects the growers of food, the producers of meat, the fisher, the people who are actually catching the fish with the people who want to buy their produce. And so the orders are taken electronically, and then sellers and the buyers are connected. So it’s converting someone who might have otherwise supplied a large grocery chain with the ability to actually sell to customers directly and run their own small micro food distribution business. Hotels would either franchise their brand to large companies, which would then build buildings and dedicate them for short-term accommodation.

That’s being replaced by a wide array of crowd-based alternatives where the demand is still aggregated by a platform. Airbnb is clearly the most visible example here. But if you think about who is providing the actual short-term accommodation on Airbnb, it’s a couple million hosts offering 3 million listings, each of whom, for all practical purposes, is running their own tiny business. They are managing their brand through their reputation. They’re setting prices. They’re managing inventory. They are providing customer service. They are providing whatever labor inputs are needed to provide the short-term accommodation. The world’s largest hotel chain, which is the merger of Starwood and Marriott, has million rooms.

Airbnb has 3 million listings. By early next year, they will have over a million instant book listings. And so by any measure, by next year, they will be the world’s largest provider of short-term accommodation. And I say that not to point out how successful Airbnb has become, but to underscore the fact that this isn’t fringe activity anymore. A lot of these are familiar things– giving someone a ride, accommodating someone, lending someone money, supporting a cause that you believe in, maybe preparing food for someone. But they’re starting to come out of the fringes, and we’re starting to see platform-based mainstream, world-class, world-leading center of the economy alternatives to the traditional corporate model of aggregate resources and provide goods and services. How many of you have stayed in an Airbnb? How many of you have hosted on Airbnb? How many of you have taken a Lyft? How many of you have you taken an Uber? How many of you have tried the carpooling feature that Waze has? Not bad.

It’s still early days. What I’m expecting to see tip to crowd-based capitalism in the future is energy production and distribution. Today, energy is largely produced by large companies that accumulate resources and produce and distribute this energy. Once the battery technology for storing solar power is good enough to be able to reliably store it, you will see local networks that are platform-based that will emerge where people who have excess power can then redistribute it to someone who needs it, either through a transmission line or by actually transporting the battery from one place to another. There’s a company in China called Trina Solar that is thinking about a pilot along those lines, where the power is transmitted by actually physically moving the battery. So the sharing economy is much more than the companies that you hear about in the news, like Lyft and Airbnb and Uber. It spans a wide range of industries. If you look at all these industries and put them together, they constitute over half the GDP of a number of advanced economies. It’s like the digital transformations that we saw in advertising, that we saw in music, that we saw in video entertainment, that we saw in publishing are starting to come to the physical world.

We’re starting to see digitally-enabled ways of organizing economic activity that are entering the very physical world industries of accommodation, of health care, of retailing, of hotels, of transportation. So I call this new way of organizing economic activity crowd-based capitalism. I see it as the third phase in– you know, I’m a business school professor, so everything has to follow three phases. The first phase was pre-18th century when most economies that were market-based or that were capitalist were organized through markets– individuals producing for other individuals. How many of you have taken an economics course? So in your textbook, you would’ve learned about the economy that was the entrepreneur selling to the customer. There isn’t so much of a mention of large companies. It’s more like the producer and the consumer. And that was what the economy looked like when Adam Smith wrote “The Wealth of Nations” in the 18th century.

That was the American economy. That was the British economy. But over the 19th and 20th century, the world’s economy in many countries reorganized in a way where hierarchies of increasing complexity were created and started to dominate how people made a living and how the world’s goods and services were produced. So by the end of the 20th century, five out of six people who made a living in the United States did so by working as a full-time employee for a company. That became the dominant model of how you earned a living. It became the dominant model of how the world’s goods and services were produced. What we’re transitioning to is a third alternative that’s going to set aside managerial capitalism as a mainstream alternative to organizing the world’s economic activities where instead of an institution accumulating the resources and hiring the people and producing the goods and services itself, that institution will instead aggregate production by a distributed crowd of people, whether it’s lending money or finding– offering accommodation, offering transportation, food services, energy, health care– all of these are going to transition in many sectors away from the managerial capitalism model and towards the crowd-based capitalism model.

So what I’m going to do in the rest of the talk is just give you a flavor for some of the features of crowd-based capitalism I find interesting, but focus mostly on digital governance and regulation, and how I think the way that we regulate the world’s economic activity is going to have to change, and what’s actually happening on that front. So there are many aspects of crowd-based capitalism that I found appealing. I think it redefines our notion of perfection. We are used to expecting perfectly folded towels that are all the same color neatly arranged in our hotel rooms. Once we start staying in Airbnbs, we realize that that feature of a hotel may not be as important as we expected it to be.

We feel impatient when our room service orders are late and then you start to wonder, does this really matter? So there are aspects of perfection that have been over-invested in by the industrial economy that we’re going to start to realize don’t matter so much. But I think that a more compelling change is the blurring of lines in the supply system between the personal and the professional. What do I mean by this? Well, all of you probably own cars. How many of you own a car? I don’t own a car. So I haven’t owned a car for a long time. So if you own a car, you’ve probably picked someone up from somewhere, from the airport, from the train station. You’ve probably given someone a ride at some point. You’ve had people stay over in your apartment or in your house at some point. Maybe you’ve lent a friend money.

It’s quite likely that if you cook, you’ve prepared a meal for someone. All of these are personal activities. They fell under the personal umbrella. You didn’t have to go and apply for a government permit before you gave your friend a ride. It was just assumed that it’s between you and the friend. And on the other side, you had professional providers of the same services– taxi drivers, restaurants, bed and breakfasts, money lenders. Now what the sharing economy is doing is it’s blurring the lines between personal and professional in the provision of these commercial services. It’s allowing people who don’t do something or who aren’t trained to do something as their full-time profession to participate as part of the supply infrastructure, whether it’s lending money, investing in companies, providing accommodation, providing transportation. And the example for transportation that reflects this change that– the two examples that I find most compelling because they are crowd-based capitalism in the purest form, where you’re taking a distributed group of people, tapping into the resources that they have, and collectively using those resources to create something that resembles a billion dollar infrastructure– the newly announced Waze carpool where people can carpool with other people through the Waze app, and the network that has been created by BlaBlaCar.

How many of you have heard of BlaBlaCar? So for those of you who haven’t, it’s based out of Paris. It’s got a funny name, a carefully chosen funny name that was chosen because it had extremely high recall. If I asked you to name five companies that I talked about a week from now, I’m sure you’re going to remember BlaBlaCar. What BlaBlaCar does is that it makes it possible for you to sell seats in your car if you’re driving from one city to another. So if you’re driving from Paris to Leon, and you’re driving by yourself, and you have three seats, you can advertise them. You can say it’s 10 euros a seat. There are leather seats in my Mercedes. I play soft jazz. And I’m a quiet person. Or you can say I’m driving my Volkswagon. I play heavy metal. And I talk a lot. You can even rate yourself as either being blah, blah blah, or blah blah blah, depending on how talkative you are.

But here’s 12 hours in the life of BlaBlaCar. They operate across 22 countries. The blue dots are people traveling in an empty seat in someone else’s car from one city to another. You count the number of people who travel on BlaBlaCar in France every day, it’s five times the number of people who travel on the Eurostar train network. You add up all the people who travel on BlaBlaCar across their countires, it’s way more than the number of people who travel on Amtrak every day. So they’ve created the capacity of billion dollar infrastructure without spending any money on new physical investments. This is pure crowd-based capitalism. So this promises a new model for building infrastructure, which is crowd-based, where you’d stop thinking of infrastructure as being an investment with high capital expenditure and start thinking of it as, well, there’s some underlying infrastructure– the roads– but how do we layer on top of that enough so that people will start sharing commercially the resources they have to create new collective capacity? You can think of this applying to parking spots.

It’s already applying to demand spikes during a sporting event, like Airbnb picking up a whole bunch of the excess demand during the Olympics or during the World Cup in Rio. And I’ve always understood why people ride on BlaBlaCar. I mean, that was like I understood that instantly. It’s cheap. It’s cheaper than a train ticket. The schedule is far more flexible. There are people leaving at all hours of the day. If your driver is nice, he or she will come and pick you up. And so it could be more convenient as well.

But I never fully understood the logic of why someone would drive for BlaBlaCar. I mean, you’ve got a nice $30,000 car. Do you really want strangers you don’t know sitting in your car? And what do you make at the end of the day? If you’ve got three passengers, 30 euros? It’s decent, but it’s not life-changing wealth. So I kept asking the founders, why is it that people drive for BlaBlaCar? And they’d keep sending me these charts that say, look, we’ve launched in Poland– hockey stick growth. And I’d ask again. And they’d say, well, consider the following survey that we’ve done. Look at our trust framework. Finally, after asking over and over again, they sent me a picture. And they said, look at the picture. These are two young French guys, like the picture explains why people drive for BlaBlaCar. And this is the picture that they sent me. I looked at the picture. And I said, aha. I get it. But it made me think. And what it made me think about was– and this is central to understanding one of the fundamental shifts that is being caused by the sharing economy, which is that we have been accustomed to a narrative over the last 15 years about how technology is isolating.

It takes us further apart from each other. It creates digital tethers where they used to be high bandwidth, emotive exchanges that were face to face. And this is not a new narrative. It’s something that Nisbet was talking about in the ’50s when he talked about the fragmentation of community. And Nisbet quotes Durkheim from his 1897 book when he talks about how technological progress– fundamentally, the narrative is that it takes us further apart. Now when I talk to people who are participating in this new world of crowd-based capitalism, typically they’re drawn to the service because it’s cheaper, because it’s more convenient, because it’s higher quality. It’s more flexible. But what keeps them coming back, what keeps them using the service is that feeling that they are forming a genuine connection with someone else as part of an everyday economic activity. It’s not so much about I want to meet friends and so let me go on this friend-making app.

It’s more about I need a ride or I need to get to work or I need a place to stay or I need a meal, and we’ve taken out all of the community. We’ve stripped out of all of the connectedness from our everyday economic activities over the 20th century as industrial capitalism reached its pinnacle. And I feel like crowd-based capitalism is re-integrating into our everyday economic activities some of what we yearn for, the simple connection with someone else, whether it’s sitting next to them when you get a ride or that feeling of intimacy you have when you stay in someone’s Airbnb property even if they’re not there because you’re seeing little slivers of their life, or like the nicer experience of sitting at someone’s dining table and sharing a meal with a bunch of people rather than sitting in a restaurant.

Now the other point that I want to make with the BlaBlaCar example is that we are completely rebuilding our trust infrastructure. What do I mean by that? Well, I’ve been surveying BlaBlaCar uses for a little while. And this is one of the results that we got out of a survey that we ran last year. It’s across 11 countries. And the question had to do with, give us the level of trust that you place in the following people. This was surprising not just because the result was so striking to us, but because it persisted across. It’s the same in Russia as it is in France and Germany. It’s the same in Poland as it is in Belgium. And what seems to be the reported level of trust that these BlaBlaCar users have is that they trust someone who they don’t know who has a full profile on the platform at levels that are comparable to the trust that they place in friends and family, higher than the level of trust that they place in a colleague or neighbor, and much, much higher than the level of trust that you place in a random social media contact.

So this either means that BlaBlaCar users have highly undesirable neighbors and colleagues, or it means that we’ve reached this point where there’s a segment of the population who, while they stopped trusting– their level of trust in institutions has fallen, their general level of trust in people in society has fallen, but they are sufficiently native in the digital world that they can look at a digital profile and it can cause them to trust someone who they don’t know sufficiently to get into their car and say, drive me to another city. I’ve got ongoing research– one of my Google faculty awards is supporting this– that are trying to unpack this further, go beyond surveys, look at platform data, run experiments.

But what makes me excited about this is that every time we’ve created a new system of trust in an economy, the world economy has grown dramatically. We moved from trading with people we knew in our village communities to trading with people we didn’t know once government food certification came along and you could buy milk not just from the person you knew– because you knew it was milk and not just powder mixed with water– you could buy the milk from the person in the neighboring village because now the government had set rules that outlawed adulteration of food. As many countries created courts and contracts and property rights, that expanded trading possibilities significantly. If you think about the everyday– just spend 30 seconds thinking about the things that you’ve bought in the last couple of days– a cup of coffee or whatever it is that you bought. Chances are that you’re relying on brand as a form of trust.

Today that’s the mainstream form of trust because you don’t want to start sign a contract every time you go to a coffee shop. I mean, it’s too expensive. And so, instead, you rely on the fact that the business’s desire for your continued business is going to cause them to invest the right amount in quality. And each of these has expanded commercial possibilities dramatically. And so we’re on the cusp of creating a new infrastructure. It’s a combination of existing things like government IDs, but new things like bringing the social network online through Facebook and LinkedIn, like increasingly rich peer feedback systems that allow you to learn from the experiences of others. And that’s core to what is making the sharing economy possible today. Because 20 years ago, we were trading stuff with people on eBay. The feedback system was enough to build trust in someone you didn’t know for them to be able to send you a package. What’s the worst that could happen? Maybe the product was broken? It wasn’t as advertised.

That baseball wasn’t real. It was fake. The jeans didn’t fit you. The Beanie Baby wasn’t as advertised. Or you didn’t get paid. Maybe you got paid late. But the risks were low. This was low stakes sort interaction. And the feedback system was enough to facilitate that kind of trust. But we’ve expanded this infrastructure to the point where we are engaging in this high stakes interaction with people now. If you think about it having a stranger sleep in your spare bedroom, which is what millions of people do on Airbnb, is a pretty high stakes kind of interaction in terms of the level of trust needed.

And yet it’s been created. So is getting into a stranger’s car and driving five hours if you’ve never met them before. So given that we’ve created this infrastructure of trust, that’s one of the reasons why I think that we are– it’s absolutely imperative that we start to think differently about how we govern things and, in particular, how we regulate business. There are plenty of examples of the sharing economy platforms butting heads with city and state governments around the world.

Many of you have heard about Uber and Lyft leaving Austin because they couldn’t come to an agreement about how drivers should be screened. They’re in the news every day, like there’s a different city that is either clamping down or being accommodating. Airbnb has been in a multi-year battle with New York state regulators about the legality of their B&B. Even a small start-up– and we were talking about this before the talk began– food start-up Josephine that allows people to prepare meals and sell them to other people. They were shut down in the East Bay recently over food regulation concerns. So part of the problem here is that we’ve got a regulatory system that is expecting full-time professionals and large companies as the suppliers.

And now we’ve blurred the lines between personal and professional. And so we’ve got people at a scale and a level of activity where the old regulatory boxes just don’t fit. And so if someone drives a Lyft for two hours a day, do we really need the same kind of oversight that we needed for someone who drives for 12 hours a day? Someone is preparing a meal twice a week. Do we really need the same kind of oversight that we needed for someone who’s running a restaurant? And you might say, yes. I mean, of course, we do need the same kind of oversight. Safety is safety. But we always make trade-offs. I mean, you don’t have a food inspector stationed in every restaurant testing every piece of food. We’ve already said, well, that’s prohibitively expensive for the government to do.

So let’s make a trade-off. Let’s have spot checks instead. And so we’re willing to accept some imperfection in our regulatory system by weighing the costs and the benefits. And now, as the supply system changes, so do the costs and the benefits. And we have to rethink regulation. The fact that the platforms are providing pretty robust digital trust means that you can start to reduce the role of the government in providing those kinds of interventions that were needed to prevent market failure. So if you take the example of taxis, 25 years ago, I wouldn’t have gotten into a taxi in New York City unless it had a government body that was saying, well, this is a legal taxi. This is metered fare. This driver has been screened and is trustworthy. This car has been inspected. In the 1990s, the taxi market would have failed completely unless there was government intervention. It was absolutely necessary to keep the market functioning. But now you’ve got a new third party that is providing a lot of trust already. We’ve got GPS technology that maps our route.

Someone else can set a fare and stick to it. There’s a third party that’s interested in making sure that the drivers are high quality, that the vehicles are inspected. So I’m not saying we toss the government regulation out. But because of this new way of trusting that has been developed between peer-to-peers, we have to rethink the role of government and rethink other ways in which we start to regulate. Because it’s one thing to say, well, I’ve got this infrastructure, and it’s going to regulate the 5,000 taxi drivers in the San Francisco Bay area. It’s a completely different thing if you have half a million people who are part of the supply system.

Some of them are giving Waze carpool rides. Some of them are driving Lyfts two hours a day. Some of them are driving Ubers 12 hours a day. Once you have that kind of scale, the idea that the government is the sole regulator starts to make less sense. So what’s the alternative? Well, we’ve already started to shift a lot of the roles that were historically the government’s to platforms of different kinds.

And you guys work at Google, so you’re already familiar with this. But let’s put Google aside and look at some other examples. It used to be the case that we would trust the government. We would entrust the government with the responsibility of dividing the rights between the creators of content and the consumers of content. We had copyright that would say this is what you can do with a book when you buy it. And these are the rights that you retain if you publish the book. This is what you can do with a song that you buy. And these are the rights that you retain if you’re the creator of the song. That division was done through intellectual property law and copyright law by the government. We have de facto shifted a lot of that responsibility to platforms that provide us with streaming music, platforms through which we download music, platforms through which we buy electronic books. It’s a license now between you and the producer of that content that is mediated by the platform. There’s a host of other examples where this sort of responsibility that was the government’s has tacitly shifted to the platform.

And so we have to embrace the idea that many of the functions that used to be performed by the government can now be performed by a third party. So what are my answers to, well, how do we deal with regulation in a world where it’s quasi-professionals, the scale of supply is orders of magnitude higher than just a few hundred large corporations who can be regulated by the government? Well, I have three key ideas. One is to increase the level of pure regulation where people who are participating in the system start to become part of the regulatory solution as well. So take the food preparation example. Rather than having the restaurant inspector inspect every kitchen that is going to be used to produce a meal, whether it’s 10 meals a day or one meal every three weeks, instead have experts on the platform start to regulate the novices.

Start to have some internal peer-to-peer regulatory system as an alternative to it all flowing from the government. So that’s one thing that I think will start to expand. The second sort of government solution that I anticipate an expansion of is self-regulatory organizations. This doesn’t mean no regulation. It doesn’t mean deregulation. It just means some of the regulatory activity is conducted by an institution other than the government. So we’ve had self-regulation through history, from the merchant guilds in medieval times through the 20th-century economy. Nuclear power has successful self-regulation. The cotton industry has successful self-regulation. Financial services as a lot of self-regulation, and the people argue whether that’s successful or not. The chemical industry, there’s a lot of self-regulation. And so this is not a new idea. And the stakes in many of the industries which we accept self-regulation are comparable at least to what’s at stake with the short-term accommodation, point-to-point transportation. I think there’s also sort of some evidence that the American Medical Association is a self-regulatory organization.

The bar associations are self-regulatory organizations. The national Realtors Association is a self-regulatory organization. And they play a part in regulating health care provision, provision of legal services, and the selling of real estate. Part of the reason why you have self-regulation as part of the solution in these industries is that they tend to be more peer-to-peer than, say, the production of– I don’t know– cars. And so individual providers tend to be more prevalent in medicine, in law, in real estate. And so it seems like the presence of a self-regulatory organization is a natural complement to a supply model that is more peer-to-peer. And the final prescription that I’ve been giving governments around the country and, actually, around the world in terms of regulatory approach– and this is the last point I make before I stop for questions– is data-driven delegation. To me, this is my competing proposal to open data. Open data is very popular these days.

Every city government I talk talks about their open data initiative. And this has two sides to it. One side has to do with making available data that governments collect about us. But another side to it tends to be, well, if you’re operating in our city, then you need to give us some data. So Uber and Lyft hand over data to many city governments like New York City that provide details about every ride that they give.

And this is supposed to be part of a regulatory solution. I’m not a big fan of that part of open data. I think it’s much more efficient instead to delegate the regulatory responsibility to the party that holds the data. And so, let’s say that you want Airbnb hosts to pay hotel taxes. Whether or not that’s a good idea, let’s put that aside for a second. But let’s say that the government has decided that Airbnb hosts are like hoteliers and need to pay– whatever– 14% of their revenue as hotel tax.

They can either say, register with us and we’ll collect the tax from you. Or they can tell Airbnb, listen, you already have all of the data. Why don’t you just collect the tax for us and give us an audited response, and remit the tax to us? We don’t need a system for registering. And a lot of cities are trying this– delegate the responsibility to the party that has the data for actually enforcing whatever rules are necessary.

Or take the example of– there’s a widespread casual evidence in New York City that the yellow cabs in New York City discriminate against passengers of certain ethnicities. There are laws against this, but they have been impossible to enforce from the looks of it. I mean, what are you going to do? You can’t have people standing on the street corner flagging down taxis that don’t pick up people. Now we live in a world now where an increasing fraction of those rides are provided by Uber and Lyft. There’s a data trail about pick-ups. Or consider the situation that certain guests on Airbnb are facing, where they feel like people of certain ethnicities don’t get accommodation as easily as others. The data driven delegation approach here would not demonize the platform, but instead say, well, you are a partner.

You are part of the solution. You’re not creating the problem here. You’re just manifesting certain deficiencies that human society has in the interactions that are being captured on your platform. Why can’t you use your data trails now to detect and correct the kind of behavior that we consider undesirable? It’s been hard to capture this kind of behavior in the past. And so data-driven delegation is a way of saying, well, if there is a data trail through which behavior that might need to be corrected through regulation can be detected, delegate the responsibility of enforcing those kinds of rules, or even with coming up with rules on your own, to the platform of the entity that holds the data.

Don’t say, well, platform give me your data. That seems far less efficient. And YouTube practices data-driven delegation on a daily basis with their algorithms. And there are different ways in which decisions are made about whether video content should be streamed or not. I mean, it’s inconceivable that you have a government authority regulating content on YouTube at the same level of granularity and with the same level of accuracy that YouTube can do on its own. Now you might ask, well, what happens when the interests of the platform diverge from the interests of society? What happens when, say, the platform has a certain set of objectives that they want to fulfill, but society’s objectives are different? Maybe there are externalities, maybe we feel the platform is over-providing certain regulation, under-providing other regulation. So you have to give them rules. And you have to figure out a way in which they comply with these rules.

Well, the 20th-century solution would be to say, well, give us an audited report. But I’ve started to think of an audit as just one particular response to a specific query. We query publicly-traded companies every quarter and say give us the following summarized data. It’s of a specific kind. And every company, instead of turning over every single detail of their operations, every publicly-traded company puts out a 10-Q. Every year, they put out a 10-K. But that’s the response to one particular kind of query. And so I’m starting to see more potential in the idea that as you delegate regulatory responsibility to platforms, you also create some sort of limited API infrastructure where compliance with these delegated responsibilities can be verified by a third party.

So I’m still early in my thinking there, but I wanted to toss that idea out there because I’m sure that some of you might have some thoughts about it. But let me stop there. But this is where you can reach me in case in the next 15 minutes of questions, you don’t get to ask your question, or you decide not to. I’m easy to find on a wide variety of platforms. So thank you for coming. And I look to your questions. [APPLAUSE] Yes? AUDIENCE: I really liked what you said about trust. And one of the points that I believe you made was that one of the reasons why that trust in the sharing economy is so high is because people are able to learn from other people’s past experiences and read reviews and these sorts of things.

But one question that I have is that institutions and resources determined by public opinion really isn’t that new of a thing. If you think about even voting in government, most officials are elected by the public. And people are well aware that those decisions are made based on public opinion. And yet, trust in politicians is extremely low. So, really, the kind of trust that exists within the sharing economy is somewhat new and revolutionary, and possibly even quite an exception. I was wondering what you feel has happened within the sharing economy that has made that trust succeed in a way that it hasn’t in the past.

ARUN SUNDARARAJAN: That’s a great question. I think the ways in which we could come up with crowd-based decisions in the past have been at a very coarse level of granularity. It hasn’t been possible to delegate responsibility for collectively making a decision on small decisions to a group of interested stakeholders because the process of running an election is expensive. And so the big decisions, like who should run the government or who should be my elected representative have been determined by voting.

But a vast majority of other decisions have not. And so now with the ability to capture information from people who have an interest either because they have participated in some process, they have been customers of a particular business, the costs of collecting that information have dropped dramatically to the point where it is now cost-feasible to have detailed, granular feedback about every restaurant, about every hotel, about every nail salon, about every small business. And so given that we have that kind of information, the way in which we collect this feedback is naturally aligned with interested parties contributing information.

And so I think part of the reason why there’s a higher level of peer-to-peer trust in the sharing economy than from, say, population-scale voting systems is that it’s more specific to something that you are interested in. And the participation is also from people whose incentives are aligned, who actually are informed about the thing that they are providing feedback on rather than randomly being asked to make a choice of pretty coarse granularity. AUDIENCE: [INAUDIBLE] ARUN SUNDARARAJAN: All right. So the question was you focused on the successes, what have we learned from the failures? I’ll answer that in two ways. First is to say that there are certain industries that I don’t think will make the transition to crowd-based capitalism any time soon, like high-end health care, meaning complex, health care– open heart surgery, unlikely to be a crowd-based platform for that. I wouldn’t use it if there was one, for sure. On the other hand, I cut my finger cooking. And I need someone to stitch it up. I can certainly see a crowd-based platform where registered nurses who live in the neighborhood come and provide that for a couple hundred dollars, once the trust gets to the level where we can delegate health care responsibility to a registered provider.

So there are certain industry segments, like heavy manufacturing, those things are likely to maintain their industrial capitalism structure. There’ve also been a number of platforms in the sharing economy that have started and then shut down. Some of them are just because of VC enthusiasm. There’s a lot of money sloshing in. Bad ideas will get funded. It’s the part of the process of creative destruction. I think some of the platforms, you have to get to– there’s a distinction, for example, between offering something on demand, instantly– you place the order and it comes to you– versus offering something as a scheduled service. I think we’re starting to realize that some services– on-demand laundry, for example– lend themselves better to the scheduled approach rather than I ask you to pick it up.

And you come and pick it up in an hour, where the logistics start to become more complicated. I think one of the things we’ve learned from some of the failures is that you need a certain immense– just because Uber and Lyft are successful with the on-demand, instantaneous model doesn’t mean that every other industry needs that. If I want a ride, I want it right now. And I want it in three minutes. But if I want my laundry done, or I want dinner this evening, it’s quite feasible that the successful model is going to be one where instead of saying I want it now, bring it to me, you might say in the morning, well, this is what I want, and I want it at o’clock, or come and pick it up at in the afternoon.

And those models may end up dominating. I guess what else have we learned from some of the failures? Well, there’ve been a number of attempts to create peer-to-peer rental markets for assets other than cars and space in your home. If you think about it, space in your home and your car are the two typically most valuable assets an individual has. The home is utilized at a high level, but it’s so valuable relative to a person’s typical income that even a small amount of slack can overcome the transaction costs, and you’ve got a market like Airbnb. A car is underutilized vastly, and so peer-to-peer rental markets work. But they haven’t worked for things like vacuum cleaners and power drills and all of these celebrated 2012 sharing economy examples.

I think what’s happening there is that the transaction costs are still too high. The value that is created from lending clothing– every city I’ve gone to has a peer-to-peer clothing rental market that is struggling with logistics. So I think once we get to the point where the objects that we’re sharing are self-aware– the digitization of the physical– an object knows where it is. It knows how much it’s being used. And it can call on transport, if needed. And so if there’s a drone delivery network that can pick it up and take it, then you’re going to lower the transaction costs to the point where a lot more peer-to-peer rental markets will become feasible for lower value assets. I think about this, and I think about the only thing more frightening than a network of drones flying above me is if they’re carrying power drills as well. So it may not be the power drills that start it, maybe the $500 dress that you want to rent from someone.

I think those are some of the things that we’ve learned from some of the failures, that the transaction costs are too high for certain peer-to-peer, and that on-demand may not be instantaneous. On-demand may not be– unit economics may not work out, but a scheduled model may work well. Well, I don’t have a good answer. So the question was, we had a high level of trust in the 1950s. In the United States, people would hitchhike. That trust level seems to have dropped. What happened? And will we see a similar drop-off in trust in the sharing economy as it gets more commonplace? So I don’t know why trust has dropped off so much in especially the US economy over the last 40 years. I’ve been tracking these numbers from the General Social Survey. They ask people as part of that [INAUDIBLE]. Every year or two years, they have the giant survey about everything from like gun control to choice. And one of the questions they ask is, in general, you feel people can be trusted or you can’t be too careful. And the response is a reliable metric that we’ve got 40 years of data about the level of trust in society.

So I’ve seen that number fall from about 50% in the mid ’70s to about 30% today. So our level of trust in other people, in general, has gone down. The fall has been steepest among people under the age of 35 where less than one in five people under the age of 35 today believe that other people are trustworthy. The thing is that in parallel with this fall in general trust has been an increase in digitally mediated trust. We’re getting more accustomed to reading digital cues because we’re growing up reading Yelp reviews, reading Amazon reviews, understanding what’s real and what’s fake about a Facebook or LinkedIn network. And so our ability to make assessments about someone else based on digital information about them is getting more sophisticated. And that’s part of the reason why when although the average person will definitely not hitchhike in the United States, it’s likely that as city to city car-sharing comes along, people will participate in it. Or the average person wouldn’t stand on the side of the road in France and hitch a ride, but they would trust the person on BlaBlaCar enough.

Will this trust dissolve over time? I hope not. I think that an alternative path that it could take is that it could raise the general level of trust in society because more and more of our everyday economic activities are going to be taking that leap of trust with someone we don’t know. The vast majority of these are positive experiences. And so, as we get accustomed, as we stop buying from brands and start to get our rides, our food, our accommodation from other people, and as these end up being positive experiences, this may start to inject a greater level of general trust in society because we’ll discover through our own experience that most people are, in general, trustworthy.

And now that we also have these digital feedback systems, it raises the bar for behavior, as well. Because if you were running a– if you were, say, selling food to your community in the past and someone had a bad experience, maybe they’ll tell their friends. But it’s not going to permeate across your entire demand base. Today if you’re selling food through Josephine and someone gets food poisoning, the entire community is going to know instantly through your feedback. And that will shutter your business. So it raises the bar on you to behave, take actions that are more trustworthy. AUDIENCE: So how do you see the long-term [INAUDIBLE]? ARUN SUNDARARAJAN: Well, to the second point, it seems likely that there will be some things– there are some people who stay in an Airbnb because they say, oh, it’s cool. It’s different. It’s not what my parents did. But Airbnb is becoming, for example, increasingly the preferred solution for more and more people.

It’s not quite there yet for business travelers. But for people who are taking a vacation, it’s not so much about it being cool and fringe anymore. I mean, they are the biggest provider of short-term accommodation in the world now. So it’s no longer a fringe activity. I think that kind of model will persist because it will just become normal. It enters the cultural dialogue. It becomes the way in which you find accommodation. You don’t think twice about staying in a hotel now.

It’s just the thing that you do when you travel to another city. And so I don’t think the acceptance or rejection of Airbnb in a decade is going to be colored by, is this cool? Or, is this fringe? Or, is this not? It’ll just be the way in which you find accommodation. To the economic factors, well, that’s an industry by industry question. The unit economics associated with Airbnb are fundamentally superior to those of the hotel industry because you are tapping into excess capacity. And so the unit economics of peer-to-peer car rental are fundamentally superior to that of Hertz because Getaround now has more members in the San Francisco Bay area than Zipcar. They’ve done it in an extremely short period of time. And this is not Uber-level funding.

They’re a modest-sized company, but they’re already that big. How are they able to scale so rapidly? Well, because scaling there involves tapping into existing cars, not leasing a fleet of cars yourself. So I’m optimistic about many slices of crowd-based capitalism here because the underlying economics are just fundamentally superior. I think most of the subsidies that we’ve seen from venture money have been in the point-to-point transportation space, the ride-sharing space.

Over there, we’ve got another transition that’s coming that all of you are familiar with, the autonomous vehicle transition. And that will take longer than, I think, Uber and Lyft think it will take. I don’t expect that we will have autonomous vehicles that are mass market and widespread in five years. I’m sure the technology is ready, but there’s a cultural acceptance. There’s a regulatory acceptance. There’s an interaction between the autonomous and the non-autonomous. All of these need to be ironed out. And it’s mostly the acceptance factor. But those start to come in. And so as these subsidies start to phase, the cost structure will start to improve to the point where, for an increasing fraction of the population, on-demand transportation will be as economical or more economical than buying a car. I mean, right now it’s the case for me. I live in Manhattan. 75% of us don’t own cars anyway because owning a car is just useless. You can’t find parking.

It’s really expensive. And so you use subway. You use taxi. You use on-demand transportation. And you have similar, like San Francisco, LA, the big cities. But at least half the population of the US lives in a densely enough populated area where if the costs fall enough, the on-demand alternative will be cost comparable to owning your own car. And so if you think about that, you think about the fact that Americans spend over a trillion dollars every year buying new and used cars– just buying, forget about maintenance and all of that. So what the investors in Lyft and Uber and there are for others– Go-Jek in Indonesia, Grab in Singapore, Didi Chuxing in China, Ola in India– all of them are unicorns.

But what the VCs are betting on isn’t that these guys will be the new taxi. What they’re betting on is that they will capture a slice of the trillions of dollars that people spend buying cars every year. So if you capture 5% of that trillion, you’ve justified your valuations. So over there, I think the model will work, but it’ll expand dramatically as the unit economics improve because of automation. Well, there’s one more point, as well. Because of automation, but before that, because also that you have gotten enough of a density where genuine carpooling through a market starts to become possible. And so you’ve got people driving from San Francisco to Mountain View and then driving back from Mountain View to San Francisco.

Even today, you look at the highway. Most people have one car. Most cars have one person. If you get enough of a density in a market where you can start to get people to share that space, then that model starts to become viable. And so that’s another thing that will drive down unit economics. But I’m not making a living driving the car. I happen to be driving it from one place to another anyway. And I’m just making some extra money by having someone else subsidizing my costs. AUDIENCE: [INAUDIBLE] ARUN SUNDARARAJAN: Well, I can’t quantify how much of a reason– so the question was, I argued that part of the reason why someone would become a driver on BlaBlaCar is social reasons. To what extent is this a convincing reason and how much is a similar reason present in other industries in the sharing economy? My conversations with BlaBlaCar have suggested that this is a big part of why people supply on BlaBlaCar. You get lonely traveling a long distance by yourself.

It’s nice to have other people in the car. It seems that, on average, we live in a society where people have insufficient connection with other people. There’s a lot of research that has suggested that, physiologically, we need social connection. There’s something analogous to physical pain that lights up in your neural network when you don’t have enough social connection. Neurologists call it social pain. So that’s certainly part of it that would cause you to make a decision that just purely on the economic aspects isn’t something that you’d be drawn to.

So it’s a combination of reasons. I think a lot of recent Airbnb hosts– the fastest rate of growth in Airbnb hosting tends to be in not the millennial demographic, but in the over 50 demographic. And I think over there, empty nests, and I have spare space, and it’ll be nice to have people staying in my home also plays a role that is operating in conjunction with, OK, I can trust this person enough and I make some money as well.

So I don’t think that it’s the only reason. Well, the one place where it’s the only reason is with couch surfing. Couch surfing is a platform that allows you to sleep on someone’s couch in exchange for letting them sleep on your– or people sleep on your couch. But there’s no money there because couch surfing is not really a marketplace. It’s a social network. The typical couch-surfing person goes to a new city and says, where’s the party? It’s not like, I need a place to stay. It’s a pure gift economy in that the exchange of accommodation is purely incidental, and the real objective is to build the connection. So over there, the social reason is dominant. But in other situations, it’s a combination of factors. Well, I mean, you certainly have a point with– so the question was, there are a couple of conjectures that are given about why there’s been a drop in societal trust over the last 50 years.

One is the sensationalization by the media of negative events with increasing intensity. And the other is, I guess, flawed regulatory intervention and reaction to bad events that have suppressed certain things that would have otherwise proceeded well. I mean, both are possibilities. I certainly think that if you look at the level of trust or the level of activity and, as a consequence, the implied level of trust that BlaBlaCar users in the United Kingdom have, it’s a little different from the rest of Europe. And that could be attributable, in part, to the fact that you don’t have the sensationalized “hitchhiking is dangerous” kinds of stories in France and Germany as much as you do in the United Kingdom. And so that’s certainly a possibility, that certain activities have been labeled as negative because of a relatively small number of extremely bad things that have happened. But I think that there is something more basic that has happened when it comes to explaining that drop of trust– I mean, I don’t know exactly what it is– because the drop in the level of trust in society hasn’t just been on specific kinds of activities.

It’s been fairly broad-based, which is why I think the sharing economy holds so much promise of re-injecting this trust. To your second point about is it flawed regulatory intervention? If that is the reason, then the sharing economy is probably in better shape because a lot of people think of these platforms that we’ve talked about as somehow making an attempt to reduce regulation.

But that’s actually not the case. The changes that they’re trying to induce, or the changes that they will induce, are different. They are a change in certain government regulations that don’t fit anymore and actually an increase in total regulation, except that the regulation is done by someone other than the government. And so if you think about the amount of regulation that will be done in a crowd-based capitalism economy, it will actually be more. It just won’t all come from the government. And where I’m getting to is that because in some ways the equivalent to government regulation that emerges from a platform is targeted at solving a real problem and is created by a crowd of people. The chances that it will be flawed regulation that might suppress the activity are a lot lower. I mean, it’s naturally aligned with encouraging the activity rather than discouraging it. So if your conjecture about regulatory intervention leading to lower trust is correct, then the sharing economy is a solution.

The regulatory systems of the sharing economy, they’re more participatory. They’re more ethical in some ways. They’re more targeted at solving actual market failures. There’s some promise there. All right. Thank you very much enjoy. I enjoyed the discussion. [APPLAUSE].

Digital Economy