I used to pride myself on being the world’s dumbest buyer of real estate. I like to overpay for a house, overpay for renovations, contract at fictitious prices for non-existent products, fail to account for the risk that a state could revoke residents’ freedoms and necessitate a move, etc. It seems, however, that I’ve been unseated. “Zillow Quits Home-Flipping Business, Cites Inability to Forecast Prices” (WSJ, November 2):
Real-estate firm Zillow Group Inc. is exiting from the home-flipping business, saying Tuesday that its algorithmic+ model to buy and sell homes rapidly doesn’t work as planned.
In a statement Tuesday, Chief Executive Rich Barton said Zillow had failed to predict the pace of home-price appreciation accurately, marking an end to a venture the company once said could generate $20 billion a year. Instead, the company said it now plans to cut 25% of its workforce.
The move represents a big hit to Zillow’s top line. Home-flipping was the company’s largest source of revenue, but it has never turned a profit.
Zillow, which released earnings Tuesday, said its home-flipping business, Zillow Offers, lost $381 million last quarter, as measured by adjusted earnings before interest, taxes, depreciation and amortization. That resulted in a combined adjusted Ebitda loss of $169 million across all of Zillow.
They were using the fraudulent EBITDA measure, which excludes the interest they had to pay to hold onto these houses. So $381 million/quarter would have been the minimum loss.
How is it possible for people who do nothing but real estate all day every day to lose their own money in this spectacular fashion? It isn’t surprising when some big developers go bust. They are in an arrangement where they keep the upside if a high-risk project goes well and stick a bank with the downside if the economy tanks (1990 calling!). But Zillow didn’t have this incentive to take crazy risks. And the real estate market did not tank. To the contrary, we’re in a period of inflation not seen since around 1980. A monkey should have been able to make money buying houses in this market since the cost of borrowing is lower than the rate of inflation in house prices.
Related:
- If Zillow is right about inflation, is everyone who rents stupid?
- Did the Zillow icon become a rainbow flag on your phone?
- Zillow welcomes me to the gayborhood
- Coronavirus will make the suburbs cool again? (April 2020; I wasn’t completely wrong! Maybe the American suburbs aren’t cool, but they’re certainly now far too expensive for most Americans to live in.)
> A monkey should have been able to make money buying houses in this market since the cost of borrowing is lower than the rate of inflation in house prices.
I don’t have any better answer than the article does about why they’ve done so badly with all the brainpower they could surely have brought to bear on the problem. However, I have noticed during a couple recent trips to healthcare facilities that feature big TV screens on the waiting room walls showing reality TV that there seems to be more competition than ever in the “house flipping” genre. I wonder how much the producers of those shows have relied on Zillow for their market insights?
As far as the rest of us monkeys go, at least one of the monkeys that hosts NPR Marketplace, lamenting Zillow’s decision tonight, said with a [sigh] that he was still going to be keeping an eye on their valuation of his property…so apparently they still have someone who believes in them! I’ll bet they make a comeback!
I remember when you posted the inflation article your basic assumption was due to inflation wall street firms could just buy real estate and sell it for a year later and earn a 10% return. This failed to account for the fact that the government has made it legal for people just to skip paying rent.
TS: inflation in real estate has been so high (about 20 percent/year) and interest rates are so low that, absent spectacular stupidity, it should have been possible to make money even without trying to rent out a house.
The algorithmic valuation of real estate is just stupid – the houses are all unique items, each with its own set of peculiarities and failings. An experienced human appraiser takes all of it into account, even when the forms don’t have boxes for things like “it smells funny”. An experienced realtor talks to people to get a feeling where the market goes, and what kinds of properties will be in demand and what kinds will not.
Zillow apparently thought that the comps and interpolation can approximate all of that… well, they got played by people who knew they were selling above actual value of their properties or buying at a discount.
Yeah, this is my take of it too. It’s a real John Henry, man vs. machine type situation. It looks like humans are better at valuing property than AI (unsurprisingly), and since this is a competitive game, being worse means losing everything: the human will accept the AIs bid when it errs in the high side, but will reject it when it errrs on the low side. And so you lose for every error. I would love to know if the AI experts at Zillow correctly accounted for this game aspect of the problem when doing their back testing, which I’m sure showed great performance.
House flipping is just going long. If you do it with borrowed money, then you’re going long with leverage. As such, you’re completely dependent on the market continuing to go up.
The other thing that struck me about the article was no mention of asymmetrical information. The seller of the house has much more information than Zillow. That’s less true in a normal real estate transaction.
Tom: But the market did continue to go up during the period in which Zillow lost money on the houses that it was holding.
Philip, real estate market went up in the past year but with multiple local maxim and minima. So I guess not surprisingly Zillow algos could not forecast and execute right buy / sell points as in most cases houses are not really liquid investment.
Houses are complicated, and if their formula didn’t account for even a single factor–let’s say, cat pee smell–then a house with that factor will seem much cheaper than it should based upon what they do measure (square footage, number of bathrooms, neighboring house prices, etc.) So they probably wound up with a bunch of smelly or otherwise flawed properties…
Maybe this is why we need the Metaverse! Internet VR Real Estate Smell-o-Vision! In 3D! You can live in the house you want without actually living there, turn off the cigar and cat pee smell, and tune out the neighbor’s late-night Young Thug + Drowning Pool mix!
123: Zillow had home inspectors that would look at each house before they wrote a check. See https://www.bloomberg.com/news/articles/2021-10-17/zillow-pauses-home-purchases-as-snags-hit-tech-powered-flipping and https://www.inputmag.com/culture/zillows-home-flipping-business-is-done-its-algorithms-are-to-blame and https://www.millionacres.com/real-estate-investing/articles/why-is-zillow-pumping-the-brakes-on-house-flipping/
Oops. Forgot the link to that timeless classic: “Bodies.” I’m being a little glib, but I think you might actually be on to one of the reasons why Zillow had trouble.
Perhaps this is the product of computer, financial and business nerds who’ve never swung a hammer or fixed their own toilet, watching HGTV and deciding that home flipping is a trivial undertaking. If the uneducated, blue collar folks these nerds looked down on in high school and college can do it, anyone can. Oooops! Home flipping is much more art than science, more “boots on the ground” than theoretical computer model.
<>
I predict Florida real estate prices will continue to rise due to the inexorable demand to leave Brandonvilles and move to DeSantisvilles.
Therefore, philg’s poor timing in real estate will continue. Although dumping Maskholechusetts may be a decent decision.