Risk Management Information Systems design

Many risk managers have attempted to takeBut given that the fundamental concepts of ERM are
enterprise risk management (ERM) from a slicknot yet standardized, how could an information
consulting pitch to a practical management system.system be designed from the ground up to support
But while ERM has helped many of theseit? There are systems that will, with the help of an
professionals improve the strategic structure of theiranalyst or actuary, allow risk managers to develop
risk financing programs, few have fully achieved theirand run simulations of limited sets of risks. Few,
ambitions. One obstacle is the risk managementhowever, are designed to collect the requisite data in
information system (RMIS) built without anthe first place.
enterprisewide orientation toward risk data.Because the insurer can predefine its risk through
For ERM programs to fulfill their potential, the RMIScoverage definitions, exclusions, retentions,
must focus on the risk financing needs and processesdeductibles and limits, these risk-limiting tools
of the entire company-i.e., reporting based on itsultimately shape the structure of today's RMIS. The
specific financial and operational dynamics. It cannotrisk manager, however, cannot predefine risks and
just tally the insurance companies' claims and losses,cannot describe every loss incident in terms of the
as it does now. The system should incorporatecoverage definitions intended to serve the needs of
occurrence descriptions and retained loss costs. Itthe insurer. Risk managers need an information
should support a range of risk financing methods andstructure that extends beyond the insurers'
the financial analysis and reporting needs of the riskboundaries.
manager.Without standardized methods of management and
The recommendations that follow do not describe aanalysis-and the technology to link the information
total ERM system. (Indeed, building a separate ERMtogether-it is difficult to implement ERM programs
system would be like constructing an independent sixand information systems. And the lack of information
sigma program. Both must be built into othersystems to collect the loss experience data on
enterprise processes to be effective.) Rather, thenontraditional risks prevents the development of ERM
recommendations that follow offer suggestions forprocedures and methodologies. The absence of each
the next steps in the evolution of RMIS design, whichelement hinders the evolution of the other.
will, if adopted, make RMIS an integral part of ERMMaking ERM tractable will require a pioneer effort to
practices.develop the intellectual tools, the prerequisite data
ERM: Great Concept, Intractable Implementation?standards and information systems that will let us
Current professional and academic schools of thoughtachieve a real breakthrough. Unfortunately, today's
dictate that ERM should achieve proper allocation ofRMIS provides no support for this kind of analytics.
risk capital across three major risk categories-financial,And a lack of compelling market demand for
credit and operational risk.enterprise risk assessment tools has failed to induce
To this end, financial risk management is highlyIT entrepreneurs to invest in the development of
standardized. (This is possible because of thesystems that support ERM.
extensive statistical data available from large, openA Cost/Benefit Analysis
markets-equity, bond, currency, derivative andRisk managers already use elements of
commodity trading systems-and the traders' interestenterprisewide risk management to improve the
in any analytical systems that provide a competitiveefficiency of risk spending. They make estimates of
advantage.) Credit risk management methods arethe scope and size of risks facing the firm and thus
less developed than those for financial riskallocate risk financing resources to bring the firm
management, but they are rapidly evolving.closer to an optimal allocation of risk capital. The
Operational risk is the least developed.estimates start with risk mapping-plotting the
Operational risk includes traditional property/casualtyexpected frequency and severity of each risk (often
risks, but it is also a catch-all term for any risk that isdisplayed on an x-y coordinate chart).
not financial- or credit-related. This includes risks thatThis is followed by scenario analysis, which
are typically beyond the scope of the traditional riskstress-tests the potential loss amounts. A low
manager: business control risks, corporateprobability (95 percentile) sequence of adverse
governance risks and capital-intensive project risks.outcomes is developed from the chain of events
For these, we lack statistical data and validatedfollowing a major loss event. The total cost of the
statistical methods to gauge the risks, and thereforepath associated with these adverse outcomes is then
few transfer markets have developed for them.calculated.
Though we have accurate data on the actuarialFor example, an earthquake damages a key facility.
dimensions of the frequency and severity of manyThis damage prevents delivery of products, leading
risks, operational risks often are multidimensional.to disruption of contracts and revenue loss. The lost
Across an enterprise, risks have widely varying timerevenue subsequently prevents wage increases,
horizons, degrees of certainty and predictability. Theleading to a labor union action, which further disrupts
nature of an occurrence or event can vary widelyproduction. Unreliable production drives away potential
(e.g., discrete versus continuous occurrences,new customers, further reducing future sales.
speculative versus fortuitous outcomes). And theAn initial event often has ripple effects. The full cost
correlations between risks typically are not wellof the loss extends far beyond the original damage
understood.to the facility. Stress-testing or scenario analysis
Operational risks frequently derive from specializedallows the firm to paint a more complete picture of
functions where evaluating the risks requiresrisks, and to gauge the extent of the firm's exposure
experience and expertise (e.g., information systemsto catastrophic events.
security, environmental health and safety, contractualTo improve these analyses, the risk manager needs
risks). Within those business functions, specialists areto use RMIS to capture more data on the
often unwilling or unprepared to conform their riskdownstream effects of the initial loss event. Invisible
assessment methods to a broader system. So whilecosts could be calculated and incorporated into the
we may be able to get their participation in creatingoverall risk picture. This might include the cost of
assessments, the assessments cannot be easilyovertime hours for recall and remediation of a
aggregated with other loss probability distributionsdefective product, lost sales due to bad publicity, or
across the organization. Even if we are somehowthe added cost of debt service due to a downgrade
able to aggregate risk assessments, the credibility ofof the firm's financial rating.
the results may be questioned by the decision makerUnlike financial risks or even most traditional property
to whom it is presented because its method ofcasualty risks, there is virtually no statistical history on
calculation is not clear, or required assumptions arethese kinds of costs. And yet, these are the costs
disputed.that most often threaten the viability of a company
All of this reflects a lack of commonly understoodin the wake of a catastrophe.
and accepted ERM principles, concepts and standardsWithout more advanced RMIS technology, risk
around which to build business processes andmanagers are limited to recording the company's loss
systems.experience or collecting other firms' case histories and
Where Current Generation RMIS Falls Shortusing techniques like modeling and Monte Carlo
Current generation RMIS technology was designedsimulations.
primarily to support insurance claims processing, and itSo, would the cost of developing a robust,
does this quite well. It organizes data in a way thatERM-supportive RMIS exceed its benefits? The costs
most closely resembles the claims processingare immediate and tangible; the benefit is difficult to
systems used by insurance companies. The basicestimate or demonstrate. Risk managers already
data record is for an insurance claim, meaning thatstruggle with how to explain the value of a loss that
incidents must at least be potential insurance claimsis prevented or financed, particularly as measured by
to be supported. The data to fill these claims recordsthe net present value of the improved capital
are normally provided by the insurer or third partyallocation. Even if the risk reduction is significant, it is
administrator and loaded into the database by thea potential future benefit, not an assured, immediate
RMIS provider. In other words, the system isexpense reduction.
primarily intended for electronic storage and retrievalWhether the risk assessments from RMIS are likely
of traditional insurer loss runs. This is great if you areto lead to enough marginal benefits to offset the
running a claims department, but ERM requires muchcost of data tracking and analysis depends on the
more.company's risk profile. Large firms stand to gain the
If the goal of ERM is to maximize the firm's netmost from refining the efficiency of risk capital
income, then the fundamental premise of ERM is thatallocation. But as the cost of the computing tools
risk decisions are capital allocation decisions. Riskneeded to collect data and perform the sophisticated
managers strive to assign the right amount of capitalmodeling and analyses continue to decrease, the
to a mix of risk financing or mitigation methods tobenefits grow for all organizations. Ultimately, RMIS
optimize results. To accomplish this, they need tomay pay for itself by empowering an organization to
understand their company's risk tolerance in light ofavoid or effectively finance that one catastrophic
their organization's cash flows, debt position, creditloss that would otherwise slash the company's
rating and price-earnings ratio (if publicly traded).financial results.