By Katie Kuehner-Hebert
Most insurers have adopted some form of risk analytics for underwriting and fraud prevention, but the industry has a ways to go to fully leverage such tools to make more real-time decisions across the entire organization, according to an Accenture Risk Management 2012 Risk Analytics Study: Insights for the Insurance Industry.
The study, based on interviews of 465 risk managers and executives -- 22 percent of whom were property and casualty insurers (19 percent life insurers) -- stated that carriers could gain a leg up on their competitors if they utilized sophisticated analytics for risk aggregation, exposure and investment concentration, reinsurance management and capital calculations.
The industry, at least, recognized the need to improve risk management capabilities in light of both regulatory and competitive pressures, according to the study: The more quickly insurers responded to events -- and perhaps more importantly, predicted events before they occurred -- the better they managed their business above and beyond their competitors.
Currently, many insurers perform classical data warehousing, but the future trend is for such companies to have high-performance analytical architectures that would provide multiple benefits, particularly in risk management, said Markus Salchegger, senior director, responsible for Risk Management Insurance in Austria, Germany and Switzerland, and one of the authors of the report.
The goal would be to integrate risk analytics processes across the organization, to aid in decisions regarding operations, capital management and management processes, Salchegger said.
This approach -- called a "risk-adjusted operating model" -- could help insurers manage their enterprise-wide risks and align their risk management programs with business and regulatory concerns.
"But insurance companies have to be prepared to make this happen, because the underlying data has to be in the right shape," he said. "Of course, you cannot change one organ of an entire body without orchestration of all the other organs."
In Accenture's study, 88 percent of the insurance respondents agreed that developing an integrated approach to risk and analytics would give their organization a competitive advantage; however, few have actually achieved such a level of integration. Fewer than one-third (31 percent) of P&C firms claim to have achieved a fully integrated view of risk aggregated across models. One in five of all insurers said that separate risk models are used for each type of risk.
"Companies could use risk analytics for better online risk exposure monitoring and pricing, but this requires alignment of tools to interconnect with such architectures," such as modeling, stress tests, and underlying source systems that flow data throughout the organization, Salchegger said.
"One of the crucial things is the increased granularity of the data and required segmentation capabilities, to better identify where the data comes from," he said.
"It can come from various business units or geographies, which might have different perspectives."
According to the study, North American insurers said that the use of risk analytics is most important for risk selection and pricing (63 percent), fraud (61 percent) and investment portfolio optimization (50 percent). The primary objectives of risk analytics for these firms are more accurate underwriting (84 percent), better claims outcomes and fraud prevention (75 percent), calculation of economic capital (68 percent) and better prospecting decisions (56 percent).
Nearly 70 percent of North American insurers are currently investing in risk analytics for underwriting and investment functional areas, according to the study. Organizations spend more than one-half of their risk analytics investments on underwriting, while distribution sees the least capital. Only one-fourth (26 percent) are currently investing in analytics for claims and distribution.
The study's conclusions rang true for David Zona, chief underwriting officer at Fireman's Fund in Novato, Calif. Zona agreed that there are some carriers -- like Fireman's Fund -- that are leaders in risk analytics, particularly in the integration of predictive analytics capabilities within decision-making processes.
"Insight is important, but the application of insight is critical," he said.
The ultimate goal of risk analytics is to make decisions in real time, and Fireman's Fund is able to do that in two key areas: modeling expected costs and managing the impact of tail events, Zona said.
For example, the insurer is building underwriting models to predict expected loss costs for private passenger auto and commercial property. While this type of modeling is fairly prevalent in the industry, he said Fireman's Fund's capabilities are more effective because risk analytics is integrated into the decision-making process at the individual risk level.
To manage tail risks, which have ultra-low frequency but ultra-high severity, the insurer has invested in risk analytics capabilities to build tools and to deploy them into real-time decision-making processes for reinsurance purchasing, capital allocation and asset allocation.
"For example, our approach to total cost of CAT management is fully operationalized down to the individual risk level," Zona said.
"This is a good example of what the Accenture study refers to as the 'real-time optimization' of analytics."
Fireman's Fund next plans to invest in distribution analytics relative to the insurer's spending on underwriting, investments and claims applications, he said.
"We see this as an area of opportunity to build more insight for the distribution function, so that we can better serve our agent customers and better steer our field resources towards higher-yield opportunities," Zona said.
Mark Verheyen, senior vice president and chief risk officer at CNA Financial Corp. in Chicago, said his company is engaged in risk analytics to help determine the correlation of claims experience to other elements of CNA's balance sheet, such as looking at how claim inflation and interest rates are correlated to do a better job of assessing risk.
Verheyen particularly is involved in capital modeling, stress testing and scenario testing of portfolios, to predict what could severely impact the company.
"Some of the outcomes with this type of risk analytics and volatility modeling directly impact how much capital we feel we need to allocate to a particular line of business, given its risk propensity and how well it aligns with our risk position -- whether it diversifies our exposure or concentrates our exposure," he said.
According to Accenture's study, insurers were also investing in data management and risk analytics capabilities to meet future regulatory requirements, particularly those with European operations in anticipation of the adoption of Solvency II, which spells out requirements for data quality, specifically, the appropriateness, completeness and accuracy of data.
The key building blocks of Solvency II -- capital calculation models, governance, Own Risk and Solvency Assessment (ORSA), internal controls and supervisory reporting/public disclosure norms -- are being adopted by an increasing number of regulators across the globe.
In the study, 65 percent of P&C insurers reported performing data quality controls when collecting historical data; 69 percent of all insurers had a data policy in place, and 41 percent had a data quality department.
"But beyond that, insurance companies need to ask the right questions to get additional benefits out of the data," Salchegger said.
"They should be viewing data in a more holistic way, so they need to have data management throughout the company. If a company really wants to benefit from new analytics technologies, data management will be key to their future. This starts with data governance -- providing the right roles and responsibilities within the firm, throughout business units and geographies, designating who is responsible for what."
Insurers will also need to improve their modeling and stress-testing capabilities to meet regulatory requirements, according to the study. However, only 55 percent rated themselves either above average or excellent in that area.
"In view of the increasing importance of scenario modeling and stress testing for insurers, it was not encouraging that the insurance firms participating in our study -- particularly P&C firms -- scored below the global average for using stress testing that is integrated into strategic decision making for large projects," the study's authors concluded.
CNA has European operations that are engaged in the type of capital modeling that is required by Solvency II, and the company extends many of those capabilities to its U.S. operations, albeit with different risk measurements, Verheyen said.
The insurer is also investing in capabilities ahead of possible capital modeling changes within the United States, spearheaded by the National Association of Insurance Commissioners. States are also set to implement new ORSA requirements over the next several years, which will likely include scenario and stress-testing impacts to business plans, and the demonstration of adequate capital to mitigate for those impacts, he said.
"If companies haven't been investing in this type of risk analytics, there will be pressure for them to improve their game," he said. "We see significant value in building out investment in models to help us quantify that risk. This will continue to happen regardless of where regulators will ultimately go."
KATIE KUEHNER-HEBERT is a freelance writer based in California, with a specialty in finance. She can
be reached at firstname.lastname@example.org
April 12, 2013
Copyright 2013© LRP Publications