By Bradley Hope and Leslie Scism
Two Sigma Investments spent the last 15 years honing its
quantitative approach to investing. Now, it's turning its
number-crunching prowess on new prey: insurance.
The firm is joining with American International Group Inc. and
Hamilton Insurance Group Ltd. to provide an automated, online
analytical system to issue policies in minutes to smaller
businesses like clothing shops, beauty parlors and medical
offices.
Under the arrangement, agents will go online and type in the
barest of information about their client -- business name and
address and a few other details -- and algorithms will begin
filling in information about the prospective policyholder. That's
in contrast to longstanding practices in which insurers often ask
dozens of questions for the process to get started.
For Two Sigma, it's the first significant attempt to branch out
to a new business from investing. The firm is known as one of the
most successful "quants," which are quantitative hedge funds that
use computer systems to trawl data, make predictions and trade
automatically in the markets.
Founded in 2001 by a computer scientist, David Siegel, and a
mathematician, John Overdeck, who had both worked in hedge funds in
the past, it increasingly is casting itself as the Google Inc. of
Wall Street, a technology company that can take its data-crunching
skills to task on a variety of finance and economics problems.
Last October, it hired Alfred Spector, Google's former vice
president of research and special initiatives, to become its chief
technology officer.
Two Sigma thinks of itself as a company that focuses on
"problems involving risk and economic activity," rather than a
hedge fund, Mr. Siegel said. The insurance deal is its first major
effort outside of the markets, but he said over time it plans to
target other opportunities where it can leverage those skills.
Small and medium-size business insurance is a "gigantic market"
and a "data science" problem, he said.
"It's not just about taking manual processes and automating
them," Mr. Siegel said. "By using our data science, we think we can
do a better job of underwriting."
The algorithms being used by the Two Sigma group's venture,
called Attune, will draw on detail that Two Sigma has obtained from
vendors who collect public records and other sources that can tell
AIG and Hamilton what they believe they need to know to understand
the risk to be insured.
That could include detail about square footage and construction
features of the business premises, changes to mechanical systems
over the years as outlined in government building permits,
proximity to fire hydrants and sources of flooding, as well as code
violations, other regulatory run-ins and legal entanglements, the
firms said.
In most instances, the algorithms would be expected to assess
the risk and determine a policy's price almost instantaneously.
The effort to speed up the issuance of policies isn't the first.
In February, Warren Buffett's Berkshire Hathaway Inc. launched the
online coveryourbusiness.com, where business owners can go directly
to obtain policies.
"Every person is buying more and more online, even paper towels,
so why not insurance?" said Rakesh Gupta, the Berkshire
initiative's chief operating officer. Believing simplicity for the
applicant is key, the Berkshire team is continuing to whittle down
the number of questions traditionally asked by "questioning all the
questions."
AIG Chief Executive Peter Hancock said he was attracted to the
partnership by Two Sigma's data-analysis record. They have been
"learning the hard lessons of what analytics can and can't do," he
said.
Global insurance conglomerate AIG is well-known for covering
complex risks at big companies. While it hasn't focused on basic
property and liability policies for small enterprises, it does
provide them workers-compensation insurance and specialized
coverage to protect against such things as cyberrisk.
Bermuda-based Hamilton was established in late 2013 by a group
of investors, led by insurance-industry veteran Brian Duperreault
and including principals of Two Sigma. Their aim: applying powerful
computing to underwriting and pricing insurance. Mr. Duperreault
previously ran insurance brokerage and benefits firm Marsh &
McLennan Cos. and the property-casualty insurer now known as Chubb
Ltd., and earlier was a senior executive at AIG.
As Hamilton's chairman and CEO, Mr. Duperreault reached out to
Mr. Hancock early this year because of Mr. Hancock's interest in
data science and AIG's decades of underwriting and claims
experience. "I thought it could be a very interesting combination,"
Mr. Duperreault said. "We would take the technology we have and use
their data and come up with something even better."
Terms of the deal aren't being publicly disclosed but Mr.
Duperreault said: "We're all equal partners."
Write to Bradley Hope at bradley.hope@wsj.com and Leslie Scism
at leslie.scism@wsj.com
(END) Dow Jones Newswires
September 27, 2016 16:23 ET (20:23 GMT)
Copyright (c) 2016 Dow Jones & Company, Inc.
American (NYSE:AIG)
Historical Stock Chart
From Apr 2024 to May 2024
American (NYSE:AIG)
Historical Stock Chart
From May 2023 to May 2024