iBuyers like Zillow have an algorithmic blindspot : Planet Money : The Indicator from Planet Money With troves of data at their disposal, iBuyers like Opendoor, Redfin and Zillow have been trying to make money buying homes online and selling them — quickly. Armed with high-tech algorithms, what could go wrong? Lemons, for one thing.

iBuyers, Zillow, and 'the lemons problem'

  • Download
  • <iframe src="https://www.npr.org/player/embed/1053689886/1053718924" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript




Welcome to THE INDICATOR FROM PLANET MONEY. I'm your host, Adrian Ma.


And I'm Stacey Vanek Smith.

MA: You know, Stacey, selling your home can be a real pain.

VANEK SMITH: Oh, yeah.

MA: Just listen to some of these frustrated home sellers.

UNIDENTIFIED PERSON #1: Tell me about it. I just spent hours cleaning and making repairs to get my house market-ready.

UNIDENTIFIED PERSON #2: I just had a buyer pull out because of financing, and don't get me started on agent fees.

MA: Oh, no. There's got to be a better way.

VANEK SMITH: And now there is a better way - sell your home to an iBuyer.

MA: The I is for instant. We use algorithms to calculate your home's worth and make you an offer so you can sell your home for cash in a matter of days.

UNIDENTIFIED PERSON #1: Wow, that's amazing. Thanks iBuyer.

UNIDENTIFIED PERSON #2: Wow, that's amazing. Thanks iBuyer.


VANEK SMITH: Does it come with ginsu knives? That is the question.

MA: (Laughter).

VANEK SMITH: So iBuyers - these are companies like Opendoor, Redfin and Zillow. And they started doing this a few years ago - buying homes. For the home seller, the idea was that this would be fast and easy. And for iBuyers, I mean, this seemed like a slam dunk. After all, they had all this data on, like, what homes people were searching for, what neighborhoods people were interested in, what homes people were clicking on and liking. And, you know, they also knew, like, regional data and all these trends. They had so much information.

MA: And so they could make these really precise calculations on how much they should pay for a house and still turn a profit.


MA: So today on the show, the mighty algorithm takes on the housing market. What could possibly go wrong?

iBuyers - basically, they buy homes online, throw on a coat of paint and turn around and sell them. This industry really got started around 2014 and took off with companies like Opendoor, Zillow and Offerpad. But Gregor Matvos of Northwestern University - he says they only concentrated in certain places, like Phoenix, Atlanta and Las Vegas.

GREGOR MATVOS: They are very present in markets in which you have, let's say, standardized housing in a very, fairly narrow price range. It's really houses that you can price pretty darn well using a computer algorithm.

MA: Think homes that are relatively nice, relatively new and would sell relatively quick.

VANEK SMITH: These guys were looking for, like, the Toyota Corolla of homes and not, like, the Ferrari. But, like Gregor says, in areas with a lot of newer kind of cookie-cutter-style houses, iBuyers were using all this data that they were collecting on home values and customer interest to just scoop up dozens or, in some cases, hundreds of properties. And to some people, you know, this wasn't landing very well. People started to worry that, you know, these iBuyers had an unfair market advantage.


UNIDENTIFIED PERSON #3: And let's talk about some what ifs. What if there was a company that everybody used, everybody used, everybody knew of to look for houses?

MA: Yes, like this real estate agent from Las Vegas. You may have seen his viral TikTok video.


UNIDENTIFIED PERSON #3: And so that company - they just sit back and they just collect all the data. They just know what zip code is looking at what zip code and how much those people can afford. And let's say that billion-dollar company uses that information to go into that zip code and start purchasing houses.

VANEK SMITH: So our TikToker doesn't name names, but, you know, he's talking about iBuyers like Zillow. And he goes on to suggest that if these companies buy enough homes, it could distort the market. They could actually artificially inflate home prices. Zillow and Redfin, by the way, have issued statements denying this.

MA: And anyway, Gregor does not think this is happening. First off, these companies have been losing hundreds of millions of dollars a year on iBuying. And despite their growth, iBuyers' share of recent home purchases is only about 1%. And even in Phoenix, where remember iBuyers reportedly own 10% of houses for sale, Gregor says that figure would have to be way bigger for any company to control the market.

MATVOS: So if they are trying to do this manipulation, they're not doing it very well.

VANEK SMITH: Not only does market manipulation not seem to be happening, but also profits don't seem to be happening. Zillow, for instance, announced last week it was quitting the iBuying business and laying off a quarter of its staff. Zillow says its iBuying operation lost more than $420 million in the third quarter of this year alone.

MA: So how did this happen? Well, you know when we mentioned earlier how iBuyers fix up homes a bit before selling them? Zillow says they had trouble finding enough contractors to do the work, which made selling really difficult. But also they kind of admitted their algorithm just didn't work like they hoped. Gregor and his colleagues have been studying iBuyers for a few years now. And he says even though these companies have troves of data they can slice and dice, they just can't get around this particular metaphorical fruit.

MATVOS: iBuyers fundamentally are exposed to what in economics we call the lemons problem.

VANEK SMITH: The lemons problem.

MA: Why do lemons get such a bad rap?

VANEK SMITH: Exactly. In this case, we are talking about lemons as in, like, a car. Like, you buy a used car, and it turns out to have some unexpected defect. It's a lemon.

MA: Right. And the lemons problem is a problem of asymmetric information. That's when a seller has way more knowledge of a thing than the buyer. And this asymmetric information problem has reared its head with iBuyers.

MATVOS: If they want to transact quickly, they can't take the time to very, very precisely value the house. They already need a pretty darn good computer algorithm to price things. But they're still not as good as someone who can take the time and walk back and forth and do a very thorough inspection and then take two months to close down the house sale.

VANEK SMITH: So algorithms - for all of the power and insight they have, they do have some weaknesses and some blind spots. Like, you know, an algorithm is not going to know if, like, the neighbors are really noisy or if the basement smells weird. And of course, the companies doing the iBuying - they know this. So they try to, you know, stick to middle-of-the-road houses and pay just a very reasonable price for the middle-of-the-road house to protect themselves from, you know, unexpected basement smells and make sure that, no matter what happens, they can turn a little profit.

MA: But on the flip side, for algorithms, they also miss, you know, what you might call the lemonade problem, right? What if a house is above average in some not so quantifiable ways? Like, it's drenched in beautiful sunlight all day or it's downwind from a pastry shop so the air smells like pie in the mornings.

VANEK SMITH: I would pay extra for pie smell.

MA: (Laughter).

MATVOS: The person who's selling it says, wait a second. The average price for my awesome house - why in the world would I sell you my awesome house for the average price? I'd rather not sell you my house at all. The person with a terrible house says, wait, the average price for a terrible house? Yes, please. So as the buyer who doesn't know what house you're getting, when you offer the average price, you only get terrible houses.

VANEK SMITH: Yeah, pricing a house is complicated. There are just things the data can't capture - lemons and lemonade - and ultimately, Zillow could not figure out a solution for these algorithmic blind spots. The housing market eluded them. The housing market beat the algorithm.

MA: Yeah, maybe for now. But, you know, Gregor doesn't think that iBuying is doomed.

MATVOS: I think there is room for iBuyers. I think they provide a service that we need. I just don't know that - the fundamental economics aren't that great.

VANEK SMITH: The fundamental economics - you know, the pricing problem, the lemons problem, the costs. It all adds up to iBuyers losing money.

MA: Which, to be fair, is a time-honored way of doing things in the tech industry, right?

VANEK SMITH: (Laughter) No, that's true. That's true.


VANEK SMITH: But wait, there's more.


MA: This episode of THE INDICATOR was produced by Julia Ritchey with help from Isaac Rodrigues. It was fact-checked by Taylor Washington. Our senior producer is Viet Le. Our editor is Kate Concannon. And THE INDICATOR is a production of NPR.

UNIDENTIFIED PERSON #4: THE INDICATOR is not responsible for any injuries sustained while iBuying.

[POST-PUBLICATION CORRECTION: An earlier version of this episode incorrectly said that iBuyers reportedly own 10% of homes in the Phoenix area. The story has been updated to clarify that this number refers to 10% of homes for sale.]

Copyright © 2021 NPR. All rights reserved. Visit our website terms of use and permissions pages at www.npr.org for further information.

NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.