RMIS data

Knowledge is power, or so the axiom goes. This is nosolve. Risk managers know that no matter how hard
truer than in the risk management industry, which hasthey work to improve their own data management
to handle an increasingly mind-numbing array ofsystems, Deragon says, they still will be faced with
information. Risk managers must collect whateverdifferent formats of data from other companies. The
data they can, analyze it and distribute it to a globalresult is an "it's not worth it" attitude in the industry.
base. For many, data management includes premiumGroups like the Pearl River, New York-based
numbers; data on locations, vehicles and humanAssociation for Cooperative Operations Research and
resources; exposure analyses; policy and coverageDevelopment (ACORD) are making major strides to
management; loss and claims data; and losscombat that, but the standard they are developing is
forecasting.very detailed. (See "Advancements in Claims Data
Thankfully, technology has improved the efficiency ofStandards," p.20.) Those that adhere to it gain a
these processes. The Internet provides an electronicgreat deal of functionality from their data, but
exchange among risk managers, brokers and insurers.adopting the standard requires a lot of effort. As a
Intranets offer an efficient, secure exchange of dataresult, the roll-out rate for data standards will be
between an enterprise's worldwide locations and itsslow. That does not help risk managers trying to
central risk management office.make sense of their company's loss history now.
But there is a problem. The success of any of these"The best thing for any major corporation to do is to
systems largely depends on the information providedstep back and collect their own data, regardless of
to risk managers by their commercial insurancewhat goes to the broker, and keep it in their own
brokers, third party administrators (TPAs) andin-house data management system," Deragon says.
insurance carriers. And therein lies the rub. For various"Some sophisticated risk managers are pushing for
reasons, insurers and other outside parties often arethat already. But getting support for it might be
unwilling or unable to give risk managers the loss anddifficult, especially since risk management is not
claims data they need to do their job morealways at the top of the executive profile."
efficiently. It is a problem that has dogged riskWho Owns Your Data?
managers for years, and for some in the riskPop quiz: You are a risk manager whose carrier has all
management community, the situation is reaching aof the loss data you need, and the technology to
head.deliver it to you. But when you ask for it, the carrier
Does Anyone Have the Information?refuses. What do you do?
One big issue risk managers must overcome whenMost risk managers naturally assume that their
chasing their loss and claims information is specificity,company owns their claims and loss information. Many
according to Vincent Oliva, vice president andinsurers, brokers and TPAs, however, insist that once
research group director of Stamford,such data comes to them, it belongs to them and
Connecticut-based Gartner Financial Services.they have no obligation to return it.
"Risk managers often buy a risk management"The issue of ownership of data has to be clarified,"
information system (RMIS) to slice and dice the dataHackenburg says. He believes, as do many risk
so they can run loss projections," Oliva says, "but themanagers, that any information pertinent to the client
historical data they get from insurers isn't verycompany's work product, insurance program, product
granular, and it makes the risk manager's job veryinformation, management processes, financial data
hard."and business methods is, in fact, client-owned data
By granularity, Oliva refers to the specificity of lossregardless of whose computer system it is in. "To
data, or in his words, "how deep the data goes." Thesuggest that a client doesn't own it once it has been
more granular the data, the more it has been brokenpartially, wholly or in pieces transferred to a carrier,
down, and the easier it is for a risk manager to pickbroker or TPA is a specious argument," he says.
and choose certain elements and run specific analysesThe argument becomes less clear, however, when
of it.the insurance company, broker or TPA takes the
"Let's say you've got fifteen claims in your packagingdata and adds value to it by applying proprietary
division that aggregated one million dollars in reservesprocesses, analytics or actuarial evaluations. Then,
and paid losses," Oliva says. "If your loss and claimsHackenburg says, ownership becomes an issue that
information is really granular, you can dig down onrequires a considerable amount of work from all
that information and get not just the total claims andparties to resolve who owns what. Acknowledging
loss figures for your packaging division, but withinthat, Hackenburg suggests that once a client pays
that, the type of each loss, the time it took to settleconsideration, fees, premiums and commissions, it has
them, the number of settlements versus judgmentsalso acquired the right of ownership to whatever
and so on. If you're an insurer, broker or TPA, youintellectual property has been applied.
have to be able to give this kind of data. Just givingFor example, a property insurer may draw and
summaries won't do risk managers much good."develop maps of the locations it is insuring. These
Sometimes, details are not the problem. An outrightso-called fire maps contain information about the
lack of useful information is, says attorney Lindaconstruction, occupancy, various exposures-both
Lamel, former CEO of New York-based Claims Onfrom internal operations and external forces-and
Line, Inc. What risk managers know about their lossavailable protection. The maps ostensibly help the
history is whatever data insurers, brokers and TPAsproperty insurer underwrite those risks, and many
provide. In some cases, the data made available iscarriers would consider them to be their property.
virtually useless to the risk manager.But given that they are maps of the client's facility,
"When I was an insurance regulator, it was usual forthey could also be used to improve the client's own
auto insurers to use gender and marital status whenloss prevention processes, loss control methods and
they rated comprehensive auto insurance," Lamelmaybe even its production techniques. In this regard,
says. "When the department asked what was thethe maps are as much an asset to the client as they
relevance of that data, the insurers admitted thereare to the insurer.
was no causal relationship between that data and"The client paid a premium to the insurance company
any sources of loss, and so they removed them asso the insurance company could make those maps,"
rating parameters. That always stood out to me. ItHackenburg says. Thus, the client could argue the
was something they were able to collect and thereright of ownership to the maps because it paid for
were actuaries trying to make rates off of it, but itthem.
made no difference."Insurers and brokers often counter that idea, saying
Another example Lamel offers involves the effortsthat sharing loss data with their clients only
of New York's governor to reduce the amount ofempowers those accounts to use their loss data as a
drinking and driving in the state. As a way ofprice guide and shop around for a better deal from
drumming up support for the initiative, he decided toanother carrier.
show how expensive this kind of behavior was. For"Are we living in a world of Karl Marx or a world of
that, he went to the state's insurance companies forAdams and The Wealth of Nations?" Hackenburg
data on DWI-related losses. "We all thought thisasks. "At the end of the day, knowledge is power,
would be the kind of data that auto insurers wouldand the insurance companies have got to figure out
have for underwriting and possibly denying coverageways of partnering with their clients and provide
to drivers with a history of DWI," Lamel says. "Whatsuperior services, including sharing their loss data with
we found is when auto insurers got claim forms, theythem." Otherwise, he says, clients will grow
did not have the space on them to indicate whetherfrustrated and will take their business to a carrier,
or not the driver had been drinking, so they had nobroker or TPA who is not quite so stingy with
data on it. None."information.
This taught Lamel that insurance may be thought ofLamel agrees, noting that insurance buyers do not
as an information industry, but it still has plenty oflook at insurance as a commodity anymore, nor do
room to improve, especially when it comes to havingthey shop purely on price. "That might be how
the kind of data risk managers need.agents and brokers look at things, and plenty of
The Standard Legacy Problemcarriers see agents and brokers as their ultimate
Another element of the data problem is theclient," she says. "But in this case, the risk manager is
industry's lack of a universal data standard. Majorthe ultimate client, and if the carriers are smart, they
contributors to the situation are the so-called legacywill acknowledge that if risk managers must pay
computer systems many insurance carriers continueextra for access to their own loss data, they will."
to use.Lamel points out that some brokers, especially those
Although the insurance industry was one of the firstkeen to gain a competitive edge, have become very
to embrace computer technology over forty yearswilling not only to share information with risk
ago, the machines they invested in were huge andmanagers, but to tailor it for their clients' specific
costly. As insurers accumulated and stored vastneeds. Many smaller brokers have made this their
amounts of data on these increasingly dated andtrademark, as have several larger players, including
rickety machines, they had to choose whether toWillis, Gallagher and Sedgwick.
invest in new, more compatible systems and transferGaining Leverage
all of their data, or simply continue to patch andFor those risk managers with a carrier, broker or
upgrade their legacy systems. Largely as a matter ofTPA that will still not share loss data, Lamel suggests
cost, many insurers have opted to stay with theirthere are other options. "If you have buying power,
existing systems. That means, however, a cripplingthat can really be the key to getting the information
data incompatibility-most legacy systems cannotyou want," she says. "If you are a key account, or if
share information with other computer systems.you have a twenty-year relationship with your carrier
Even if they could, insurers, brokers and TPAs haveor broker, you can probably negotiate for things that
different methods of organizing, analyzing and runningmight not otherwise be available to other accounts.
projections on data, making a bad incompatibilityIn this case, the leverage of being a key account is
situation even worse. This is especially true for riskcritical."
managers who have multiple accounts with multipleBut what is a risk manager without a major or
insurers and may have had accounts with manylong-term account to do? The answer, says
other third parties over the years. If each companyHackenburg, is to spend plenty of time up front with
maintains information in a different format, the riskthe carrier, broker or TPA before any formal
manager is left with the impossible task ofbusiness relationship begins. Define which entity owns
homogenizing all of that information in order to runwhat data and what rights of access the client will
historical loss analyses and make historically basedretain, even once the relationship ends. "Call it a
loss projections.prenuptial agreement," he says.
"The industry needs to standardize data formats andThe key is doing it sooner rather than later.
to develop free and clear access to that data,"Hackenburg says. "When somebody is after your
advises Richard Hackenburg, CEO of St. Louis-basedbusiness, that's when they are most willing to
Willis Missouri Inc. "That way, you can maintain anegotiate. Once you're hooked, things may not
principle of consistence, reducing costs over timealways be what they seemed."
because everybody's doing things in a consistentTo illustrate the point, Hackenburg recalls an old joke
fashion. The analytics applied to consistent data willin which a man dies and goes to Heaven. St. Peter
give better forecasts of what is likely to happen."gives the man a one-day tour beyond the Pearly
Jay Deragon, chairman of XS Voice, a Nashville,Gates, which, unsurprisingly, seems like a wonderful
Tennessee-based insurance technology consultingplace to spend eternity. It is beautiful, comfortable
firm, agrees, saying that getting usable informationand serene. But before letting the man in, St. Peter
from a carrier is akin to pulling teeth. From his pointgives him a day to tour Hell so he can make an
of view, the problem is that the insurance industry'sinformed decision as to where he would like to go.
early enthusiasm for computer technology hasThe man agrees and goes to Hell. To his surprise, as
waned, and now many carriers are content merely tonice as Heaven is, Hell is much better. There are
maintain the systems they have, investing the bulkraucous parties, all his friends are there, and it seems
of their energy into reserves and market conditions.like a really great time. So, when the man returns to
Working on data standards is simply a low priority.Heaven, he declines St. Peter's offer and goes back
"Even though data sharing is so popular amongto Hell. When he arrives, there are no parties, only
clients, many carriers are still not interested infire and brimstone. Bewildered, the man asks Lucifer
pursuing it," Deragon says. Because companies' losswhy things suddenly became so different. Lucifer
data often is spread out among various insurancesimply smiles and answers, "Yesterday, you were a
carriers and brokers, data incompatibility becomes aprospect. Today, you are a client."
systemic problem that no one insurer or broker can