A Delicate Dance CONTINUED
beyond the parameters of traditional deterministic models.
“And you need the ability to consume new data points that
support your rating segmentation or reserve ranges very rapidly,” Roberts said. Because today’s environment of heavy
regulation and greater competition makes it necessary to consider larger amounts of data from numerous additional sources,
Roberts explained, “you can no longer consider just the commonly used historical data dimensions.
“There are so many niche markets you can exploit if you have
the data to support your rates,” said Roberts, whose work re-
quires him to integrate his actuarial and IT skill sets. “If you can
identify specific categories of similar risks, you can apply a much
more appropriate rate to that group using additional criteria or
tiered rating plans.”
These pursuits are necessary for success, said Glenn Watson,
assistant vice president and applications development manager
for Grange Insurance, based in Columbus, Ohio. “That is not triv-
ial to be able to take multiple different data points and bring them
together and correlate them as something to pursue to provide
competitive advantage,” he said.
Regulatory and competitive pressures demand stochastic
modeling for the predictive results of thousands of possible
scenarios, Gold said. “Sometimes you need more than one set
of stochastic simulations—you need a second- or third-order
stochastic model,” he explained.
Such data-hungry models require exceptional processing
power and storage capacity. As an example, Gold said, he has a
large client that needs at least 5,000 cores at peak periods. This
is the horsepower of about 200 computers strung together in
a closely coupled high-performance computing environment.
Since stochastic calculations could take hundreds of years on
a personal computer, Gold said, actuaries no longer will be able
to do under-the-radar work-arounds from their desktops. They
will need help from their IT team, and both sets of professionals
will have to collaborate closely.
Making this transition from working separately to working
in tandem means confronting long-standing organizational and
Actuaries and IT professionals often seem to have opposing
goals. This can lead to misperceptions, miscommunication, and
avoidance strategies that hamper an organization’s overall goals.
“Actuaries find IT stifling while IT thinks actuaries are cowboys,” Gold said. And either side can be hesitant about working
with the other. “I have seen cases both ways, where they avoid
each other like the plague,” Gold said.
Actuaries’ specialized needs can be threatening to IT, Smith
said, which can harm the relationship. This is often the case
when actuaries want to bring in third-party vendors. Since IT
professionals are hired to be the guardians of uniformity and security, Smith said, they naturally worry about security breaches,
making them resistant to change the status quo.
Large insurance companies are overwhelmingly driven by
security issues, Gold said. They must protect policyholder information and develop role-based security. “And when they have
something that works, they may be reluctant to change it again,”
Gold said, adding that he regularly notices this when his company is selling its software.
Data accessibility is another cause for relational friction. IT
professionals can create seemingly unnecessary hurdles, making
it more difficult for actuaries to do their job. And this can frustrate actuaries who work in an increasingly information-intense
environment, said Watson. “There is more data to consume, and,
by God, they are going to consume it,” Watson said of actuaries.
Besides being concerned about data security and integrity,
said Watson, if actuaries don’t follow a disciplined process, “it
can mess up their data.” Sometimes actuaries grow frustrated
by the rules, but an audit trail is necessary to prevent future
problems, Watson said, and “to protect them from themselves.”
Working quickly to fulfill various needs often results in actuaries forgetting to follow protocols that ensure they are using the
most current version of information in the future, Watson said.
“What actuaries need to appreciate is that software solutions
have to abide by basic development principles,” Roberts said. That
includes the development process, quality assurance, and testing.
Models and scenario testing are growing in complexity, and
greater visibility is required, Roberts said. “The whole development process has to be auditable, rigorous, traceable, updatable,
and it should comply with generally accepted development
principles,” he explained.
Getting new software, programming, or data also can be
frustrating because IT departments have limited resources.
Roosevelt Mosley, principal and consulting actuary for Pinnacle
Actuarial Resources in Bloomington, Ill., gave the example of a
client who finally got the nod to upgrade a SAS tool after a year-and-a-half wait. “IT has all these priorities, and to get into the
IT list of priorities is like begging, borrowing, and stealing and
calling everyone you know,” Mosley said.
This also manifests itself when developing software
solutions with actuaries, said Mosley, who works on actuarial software with a major developer. “Since everything
in the IT world today gets translated into a time and
cost estimate, there is a pressure on IT to reduce
the laundry list, not expand it.” As a result, Mosley
said, features are delayed to the next version or
Perhaps the most limited resource for actuaries is time. “Actuaries seem to always be in
a rush to deliver the next piece of information—and the pressure is gradually getting
worse,” Watson said.
But in the heat of getting a task done
as quickly as possible, actuaries easily
grow frustrated with the limitations
of resources, protocols, and data