The EconomistFebruary 15th 2020 Finance & economics 632
1adjacent Brooklyn brownstones built in
the 1920s could be entirely different beasts.
Some markets might be too idiosyncratic
for i-buying, says Sean Black of Knock.
Prices jump in Palo Alto, a town south of
San Francisco that is popular with tech
workers, when a large company goes pub-
lic. Loft apartments in Tribeca, a neigh-
bourhood in Manhattan near the down-
town financial district, soar in years when
banker bonuses are fat.
Alex Rampell of Andreessen Horowitz,
a venture-capital firm that has invested in
Opendoor, says i-buyers create a pool of li-
quidity, allowing investors keen to buy
rental properties to do so at scale. “Institu-
tional investors buy to achieve a certain
rental yield, so they are less sensitive to
price uncertainty.”
The success of i-buyers also depends on
whether their algorithms get the price
right. The most important factor is loca-
tion, says Bridget Frey of Redfin. It inter-
acts with other factors, too. “You need loca-
tion to tell the algorithm what weight to
put on the thousands of other variables you
might look at.” Swimming pools add value
in San Diego but tend to decrease it in New
Jersey. In Atlanta proximity to a golf course
is highly prized. Before Zillow launched
there a worker traced every golf course on
Google Maps, so that it could be added as a
variable. For years Rich Barton, Zillow’s
founder, found it odd that the algorithm as-
signed a negative value to extra bedrooms.
“It seemed backwards. But once you’ve fac-
tored in square footage, extra rooms actu-
ally deduct from the value of a house.”
The process is not entirely devoid of hu-
man input. At Zillow’s offices in Seattle a
group of youthful workers spend their days
on Google Maps zooming in on pictures of
houses that sellers have requested prices
for, verifying that nothing looks too out of
the ordinary. Ms Frey would like to get to a
point where the algorithm beats the hu-
man. But at present Redfin also uses agents
to conduct home inspections, and defers to
them if their assessment differs from that
of the algorithm.
The bosses of the teams building the al-
gorithms all talk about their “buy-boxes”.
Rather than buying the most expensive or
the cheapest homes in any neighbour-
hood, they prefer the 60% or so in the mid-
dle. They find it easier to provide an offer
for average homes with confidence; over
more unusual homes there tends to be
greater uncertainty. And the more uncer-
tainty, the lower the offer they might have
to make—if they make one at all. “We
sometimes can’t quite figure out why that
particular home is so much cheaper or
more expensive than the rest,” says Stan
Humphries of Zillow.
That said, where i-buyers do operate,
they seem to get close to offering fair value.
Research by Zillow finds that, when sellersdecline the firm’s initial offer, their eventu-
al sale price is only 0.2% different. An inde-
pendent study by Mike DelPrete of the Uni-
versity of Colorado found that, on average,
the offers made by Zillow and Opendoor
were 98.6% of the price that standard in-
dustry models suggest, implying a 1.4%
discount compared with the market.
Getting value right is critical to how the
model works, says Glenn Kelman, the boss
of Redfin. “If we start buying homes cheap,
or trying to fix them up too much, our busi-
ness will start to be valued like a real-estate
investment firm. That is the opposite of
what we want.” Tech firms tend to trade at
higher valuations than property invest-
ment companies. I-buyers say they are in
the business of providing convenience and
liquidity, not flipping homes for profit.
A big question, though, is whether i-
buying can be profitable. None of them yet
make any money. Zillow’s home-buying
business spends $1.40 for each $1 of rev-
enue it receives. The firm makes most of its
revenue selling leads on buyers to agents it
is partnered with.Free agents
Other innovations are nibbling away at the
many other tasks that estate agents do.
Redfin and Opendoor use remote electron-
ic locks, which can let buyers into a home
by themselves. Your correspondent let her-
self into a lovely two-bedroom flat in Santa
Monica using Redfin’s app. Had she wanted
to buy it, she could have done so without
consulting an agent, by filling out an offer
form on the app.
But not all of the biggest prop-tech com-
panies in America are betting on estate
agents becoming redundant. Redfin’s fo-
cus is on lowering agents’ costs. Sellers
who list their home with Redfin pay com-
mission as low as 1%, instead of the usual
3% (though sellers must still pay tradition-
al commission rates to the buyers’ broker).
Compass, which was founded in 2012
and is now worth more than $6bn, is the
most focused on helping agents. Its tools
take the drudgery out of their work, in or-
der to make them more productive. Its plat-
form for agents analyses the best time to
list properties and automatically sends
them listings their buyers might like. Rob-
ert Reffkin, its founder, claims that agents
who use Compass make more deals. “If
Compass fails it is because my faith in the
role of the agent is misplaced.”
Lower fees, therefore, need not mean a
big hit to agents’ pockets. They might boost
productivity. They could encourage people
to move house more often, offsetting the
fall in fee rates. Clients, meanwhile, would
undoubtedly benefit. People rank buying a
home second only to divorce as the most
stressful time of their lives. If it becomes a
little less so, and cheaper into the bargain,
that would be a welcome change. 7T
he imf is familiar with unwelcome
edicts. Its job as the world’s lender of
last resort often involves demanding re-
form. But its staff may be discovering the
unpleasantness of being on the receiving
end. Kristalina Georgieva, the fund’s boss,
is reorganising the institution.
Ms Georgieva took over as the imf’s
managing director in October 2019 on a
wave of good publicity. As the first boss
from an emerging market, she regaled au-
diences with her own experiences of an
imfprogramme—in the 1990s she saw hy-
perinflation in Bulgaria wipe out her moth-
er’s savings in a week.
She arrived with a reputation for being
able to manage large bureaucracies, having
previously been chief executive officer of
the World Bank. A few months in, she has
managed two senior deputies out. On Feb-
ruary 7th she announced that David Lipton,
the first deputy managing director, and
Carla Grasso, another deputy managing di-
rector, would leave at the end of the month.
Mr Lipton’s reputation for diligence and
technical expertise meant that the news of
his early exit—his term had been due to
end in August 2021—was not well received
by staffers old and new. A rumour swirled
that his departure was a ploy by President
Donald Trump’s administration to choose
his successor. By convention, America ap-
points the first managing director.
In fact it seems that Mr Lipton’s depar-
ture was Ms Georgieva’s doing. Her prede-WASHINGTON, DC
The fund’s new boss makes her markThe International Monetary FundStructural
adjustment
Kristalina’s got a programme