These days, there’s almost no aspect of commerce that hasn’t been radically transformed by technology. Uber lets you summon a privately driven car to chauffeur you, Seamless let you order takeout without ever speaking to a soul, and thanks to Airbnb, you can sleep in another person’s bed.

But the shift to having everything at your fingertips at the tap of an app hasn’t been without its bumps. Ride-sharing apps have been blamed for worsening traffic congestion, Amazon has been accused of subjecting its warehouse workers to brutal working conditions, and Airbnb has been a source of great debate in places like New York City.

To learn more about online marketplaces today, we sat down with Apostolos Filippas, Ph.D., an assistant professor of information systems at the Gabelli School of Business.

Listen here:

Full transcription below:

Patrick Verel: These days there’s almost no aspect of commerce that hasn’t been radically transformed by technology. Uber lets you summon a privately driven car to chauffeur you. Seamless lets you order take out without ever speaking to a soul. And thanks to Airbnb, you can sleep in another person’s bed. But the shift to having everything at your fingertips at the tap of an app, hasn’t been without its bumps. Ride sharing apps have been blamed for worsening traffic congestion, Amazon has been accused of subjecting its warehouse workers to brutal working conditions, and Airbnb has been a source of great debate in places like New York city. To learn more about online marketplaces today we sat down with Apostolos Filippas, an assistant professor of information systems at the Gabelli School of Business.

I’m Patrick Verel, and this is Fordham News.

Talk to me about the study you published in May. Now you’ve found what would seem to be a happy medium for Airbnb, correct?

Apostolos Filippas: Yes. So platforms like Airbnb, HomeAway, and so on, they often face a critique that pertains to the externalities that guests impose on people living in the buildings. So what do I mean by that is that when you’re an Airbnb host, people come into your building and maybe they’re not familiar with the building space. They ask questions on your neighbors, they have parties later on at night than longterm residents do, and so on. So there’s some externalities that your neighbors have to face. This is a common critique of home sharing platforms. In my opinion, it’s a real critique. So this is a problem that home sharing platforms do create.

So in this paper with John Horton, we’re proposing a solution to this problem. And our solution is a very simple one. Our solution is that whether a building allows home sharing or not, should not be up to the tenant, to every individual tenant. It should not be up to the city, but it should instead be up to the building owner. And that might mean of course an owner that owns the entire building or all the tenants of a building deciding by vote on the home sharing policy. So what this allows for is it allows people, it allows tenants, to sort into buildings that either allow or do not allow for Airbnb according to their preferences.

And we believe this is a good thing and we show in the paper that this has many advantages. Because for example, if you’re a student and you absolutely want to monetize, to make some extra money off of your apartment once you’re away in the summer, for example, you’ve got something available for three months, but you still have to pay rent. This is a very good way to make some extra money on the side. On the other hand, if you have a family, if you, for example, have two kids and you absolutely do not want to be next to guests from all over the world, that might be noisy, you don’t know anything about, and so on, then you absolutely want to pick a building that does not allow for home sharing. So this allows for, as you said, a happy medium, allows people to go and live in buildings that either allow for home sharing or do not allow for home sharing.

PV: Do you know of any buildings that are like this? Where they’ve already, either the owner of the building or the tenants in it, have all come together and agreed upon this is what we’re going to do or this is what we’re not going to do?

AF: Absolutely. So in New York City, in most co-ops, this was the modus operandi for a long time. So the buildings came together, the building tenants came together, and they voted on a home sharing policy. Now most of these co-ops ended up not allowing for Airbnb. Just because, to be a long term co-op owner in New York City, well you don’t really care about that, the few extra money that you can make through Airbnb, right? So they preferred probably their privacy over that few extra money over Airbnb. I believe the city of Sydney is now moving towards this direction and they are putting forth the legislation that will allow for buildings to decide whether they will allow for home sharing or not.

PV: So this would be something where the city itself is becoming hands off on it and it just becomes every building for themselves?

AF: Fantastic. So this is a very good comment. The city becoming hands off would be allowing every tenant to decide what to do with their building. So it’s not hands off, it’s one level away from hands off. It tells tenants to decide on the building level. And this is what creates all of the positive effects of this policy.

PV: Now you also study the kinds of rating systems that are used on sites like Airbnb, right?

AF: That’s right. Pretty much every online marketplace is using rating systems. Uber has a rating system; you have to rate your driver after every ride and your driver rates you after every ride. Airbnb has a rating system; you have to rate your host after every stay and the host rates you as well, and so on. And this rating systems are extremely crucial to a well functioning marketplace. The reason is that in this marketplaces, we’re called to trust strangers every day with these transactions. You sleeping to strangers homes, you get into strangers cars and ask them to drive you around, and ratings allow us to be more certain about the quality of that person.

They also have a secondary effect. They also make sure that everybody’s on their best behavior. If you know you’re being rated, then you’ll treat an apartment, an Airbnb apartment better. If you know that you’re being rated, and you’re an Airbnb host, you’ll put more effort into providing a good experience. In this research, we found that these rating systems suffer from a fundamental problem, and the problem is that … Have you ever taken an Uber ride? You probably have, right?

PV: Oh yeah.

AF: So what was the rating of your driver? Did you remember?

PV: No, not really. I mean, I am sure it was fairly high. I wouldn’t get in one with one star, let’s put it that way.

AF: Okay. Absolutely you wouldn’t, right? In fact, I would claim that you wouldn’t get into anything with less than 4.5 out of five stars, right? If you notice the next time you take an Uber ride, you’ll see that your driver probably has a 4.7, 4.8. I mean in most cases, even a 4.99 out of 5 star rating. So in our paper that’s called Reputation Inflation, we find that reputation systems have this fundamental flaw of they tend to exhibit more inflated ratings over time. Ratings start off from relatively average of, let’s say, three stars out of five and boom, five years later everybody has a five out of five. And that’s pretty similar to what happens to some extent in the U.S. universities. If you go back to the 50s, the average GPA was probably three out of four. And nowadays the average GPA is something around 3.7, 3.8 out of four. It manifests itself whenever we have a system where two peers, two people are rating one another.

The reason that we give good ratings to people, inflated ratings, even more than we believe they deserve, is that we do not want to harm their future prospects. For example, if you’re a professor, you do not want to harm the labor market prospects of a student by giving them a bad grade. If you are an Uber rider, you do not want a driver to not find jobs in the future because you gave them a bad rating. You don’t want them to get into trouble because of your rating. So you kind of hold back from giving bad ratings unless something terrible happens, unless someone’s speeding a hundred miles per hour or so.

PV:  Or in the instance of my wife, who took an Uber to LaGuardia recently, and the driver informed her that he needed to stop and go to the bathroom and did so on the side of the highway because he had to pee. And to which she was thinking, why did you pick me up if you had to go to the bathroom? I don’t know what rating she gave him, but I’m sure it wasn’t … I’m sure she docked him at least a point.

AF: Maybe you’ll be surprised. Maybe she just didn’t give a bad rating, she gave a five out of five. Maybe she didn’t even give a rating.

PV: I’ll have to ask her now. I never did find out what she gave him.

So you’re saying … It’s interesting about the professors too. So it’s not just because you might get a bad rating on or something. It’s because you legitimately think, “If I give them this C minus, I’m going to hurt them down the line.”

AF: Yep. So this was a competing hypothesis in our study. We’re thinking maybe everybody’s giving good ratings because what’s happening here is that you don’t want to receive a bad rating yourself. Students rate professors as well, drivers rate riders as well. Airbnb hosts rate Airbnb guests as well. But what we found is that even if you do not give any information about who gave you this rating, as long as their rating has consequences for the person receiving the rating, people will start inflating their scores.

PV: Oh. So it has nothing to do with the fact that you might get a bad rating yourself?

AF: It’s not about this tit for tat behavior as much as it is about you not wanting to hurt the other person, the person that you’re rating.

PV: I guess that’s kind of nice, we all are not looking to hurt each other, right? Even if you’re willing to overlook this, something that went wrong in the house or in the car that you were in. Ultimately you’re like, “I don’t want to hurt this person.” But it would seem to also defeat the whole purpose as well of a rating system.

AF: Yeah, absolutely. It gives us a little bit of an optimistic outlook for society. We still do not want to hurt other people through our ratings. But you put it very nicely, kind of defeats the purpose. There is a fundamental trade off there between how consequential the rating is, and how willing people will be to give this rating.

So I believe that in the future we’ll start moving to rating systems that ask very objective questions, not subjective ones, not rate your ride from one to five. They will ask, for example, “Did the driver take a direct route? Yes? No? “Was the house clean? Yes? No?” We’ll ask more protective questions and then we’ll try to use these answers to help people improve rather than punish them without giving them the opportunity to improve.

PV: We’re talking about the sharing economy, right? So sharing has become obviously super successful when it comes to cars and homes. So the natural question then is what’s stopping us from sharing everything?

AF: That’s a really good question, Patrick. So when the sharing economy first came out, you had pundits trying to look into their crystal ball and they were predicting a world where we do not own anything anymore. We just rent and share whatever we need to consume, right? However, 10 years later almost, we do not really see this world. We see maybe even more consumption taking place.

So the sharing economy, in another paper with John Horton and Richard Zeckhauser, we’re trying to answer this question. What is going to happen? Because due to this sharing economy. Will consumption expand or contract? And consumption we find will always expand because of the sharing economy. There is something that just remains there and use and we can now use. On the other hand, the question of whether ownership will expand or contract. That’s a little bit harder to answer.

Will fewer people own cars because of the sharing economy or will more people own cars? The answer there is much more ambiguous. So you can think about these two countervailing forces. On the one hand, maybe you do not need to own a car because you can get all of these services on Uber, or you can rent a car when you need it on all of these platforms on one hand. On the other hand, maybe now you can own a car because now you can have the money to own a car because you can rent it out and make some money for yourself. Maybe now you can own an apartment that wouldn’t make sense before because you can rent it out on Airbnb short term and make some money. So whether ownership expands or contracts, well we’ll find out in a few years.

PV: In light of all this, what’s your take on the gig economy?

AF: It has almost been 10 years since we started having this discussion about online marketplaces, peer to peer market, and so on. And in the beginning of all this phenomenon we did have two sides. One side said, “Everything is going to be gig economy and sharing economy base.” And the other side that never believed that it would amount to anything. Now, 10 years later, we see that it’s part of our lives, right? We all know about Airbnb, we all know about Uber. Employers sometimes find freelancers online instead of hiring somebody full time. It just makes sense for some activities. And it has benefited a lot of different demographics. For example, students that want to make some back on the side, they always used to be freelancers. Now that can get higher quality jobs easier as freelancer. Moms that have little stretches of time, two, three hours when they want to work, they can find jobs that last just two, three hours and so on. So I believe the sharing economy and the gig economy has unlocked a lot of value, and it’s going to be a part of our lives in the years to come, yes.


Patrick Verel can be reached at or (212) 636-7790.