Many insurers might either decline to underwrite such a risk or quote a very high rate, as well as exacting terms and conditions surrounding the contract. Insurers would likely also demand additional mitigation measures before initiating coverage.

Many individual attributes with varying degrees of positive or negative effects contribute to an overall risk equation. This multivariate effect manifests itself in a myriad of ingredients, including geographic, physical, social, and political attributes. Improved risk management strategy begins with data being collected in an intelligent and precise manner by qualified and certified professionals. The data-collection baseline is then enhanced through an analytic approach to deploy the data effectively. The analytic paradigm combines specific subject-matter expertise, data aggregation, interpretation, verification, validation, and modeling steps. It can also involve the distillation of many companies' loss experience over a long period of observation. This type of quantification and analysis is fast becoming the standard of modern insurance operations. Subsequent steps to validate the risk equation often build from this starting point.

Communities vary significantly regarding factors such as geography, weather, roadways and traffic patterns, physical infrastructure, and investments in firefighting, police, flood mitigation, and building-code enforcement — attributes that strongly affect insured loss frequency and severity.

To administer a program in support of loss cost estimation, insurance providers can develop and maintain insured-loss data by working closely at the grassroots level with local officials across the nation in the assessment of community loss-mitigation capabilities. Such continual collaboration would keep the data fresh while providing educational avenues to show which community attributes can be changed to help lower expected losses.

As the relevant data becomes more readily available and as analytic sophistication grows, insurers and their vendor-partners can develop refined statistical models to help both communities and insurance operations staff better understand complex risk-quality relationships, permitting more accurate underwriting and rating of individual risks.

Understanding how individual community attributes affect loss is difficult, but analytics experts are now using time-tested, objective metrics to evaluate communities and to determine how one community measures up against another. A suitable mea-surement program can include the following:

• Close working relationship with local fire and building officials
• Clear metrics that can be used to rank/score departments nationally
• Identification of attributes that contribute to measurement of losses and mitigation
• Qualified staff with the necessary certifications and training to conduct objective surveys

Municipal fire and building-code authorities and other community officials would use such measurement processes to help administer important programs that evaluate components of community infrastructure, such as fire-hazard risk, compliance with up-to-date building codes and construction guidelines, and the geospatial threats of flood damage from surrounding bodies of water.

Many communities strive to show they are safe places to live and conduct business. When communities actually are safer, citizens, policyholders, local businesses, and public authorities benefit from stronger, well-protected communities with reduced risks and losses. The insurers benefit from more accurate risk assessment.

Predictive Modeling Insight
Experts in the field of risk assessment make better judgments with better information on both the subject at risk and the surrounding “risk ecosystem” (e.g., adjacent community risks, local mitigation capabilities, topography of land/water, prevailing weather patterns, and even catastrophe disposition to events caused by society or nature). Such risk-assessment data can deliver the highest value when it surpasses the level of isolated, silo-based information and reaches the level of actionable insight and knowledge for decisive action.

Today's risk-based modeling methods enhance ratemaking engines with more and better data. Property underwriters have long respected this reality and are expanding their awareness by combining all exposures into their geographic risk considerations.

Some carriers have adjusted where they concentrate policies (or are in the process of doing so), while others have already lost their competitive edge.

In the workers' compensation meltdown of just ten years ago, companies that resolved to forgo a losing market, maintain underwriting discipline, and get their pricing right, survived to compete another day.

In the Florida homeowners' marketplace, insurers are competing on risk selection as opposed to pricing. Large insurers are leaving, while the residual market facility has too many policies. In turn, start-up companies are able to cherry pick the best risks. These decisions are largely guided by catastrophe models as opposed to predictive models, but the concept is the same.

The knowledge of future likelihoods can permit more refined risk assessment for individual and aggregated pools of insurance contracts. This predictive knowledge can allow even more accurate risk-based pricing and targeted marketing approaches for insurers. Meanwhile, communities can leverage the same information and insight to improve the quality of life and the safety of their citizenry.

Advanced statistical methods, cross-industry experts and data sets, text mining, and geospatial information have all contributed to a new breed of data components for modeling loss costs at each risk address. In addition to smarter underwriting, new and emerging analytic methods optimize customer-lifetime-value portfolio management, corroborating that customer-centric approaches can be the best form of innovation and value generation.

A key driver of decision-making action for insurers and their policyholders is predictive modeling, the process of incorporating multiple building blocks of contextual information into a framework of statistical analysis. With predictive analytics, compound elements of risk data can be better observed in relation to loss frequency and loss severity. Also, more intelligent insights about what may occur can be generated. Increased accuracy in forecasting aggregate losses on portfolios of risks is what makes predictive analytics so useful today.

The improvements under way in obtaining better data and increasing the quantity of data — in conjunction with more sophisticated modeling technologies — are the factors that will keep predictive analytics relevant tomorrow.

As the insurance industry continues to make progress in the deployment of analytics throughout the enterprise, history continues to favor the most informed underwriters. For those companies using analytics effectively, their shareholders and customers are rewarded with value, and their markets and regulatory environments are endowed with greater stability.

Ranked Risks
A keen understanding of accurate pricing is an obvious competitive advantage, but there is a continuum of possible outcomes to every distribution of ranked risks. Rating plans using predictive analytics can assign a unique price to any risk, turning pricing into an even more potent underwriting tool. Depending on the strength of a company's underwriting strategies, traditional insurance plans with few pricing points can result in long-term profitability at one end of the continuum or invite adverse results and even outright financial ruin at the other end. This is generically the same for all lines of business and for every industry — proper segmentation of customers and their needs and costs is essential to a healthy company, and the healthiest companies continually analyze data to improve themselves beyond their internal benchmarks as well as external competitive indices.

Carriers caught in the transition from traditional to modern methods of pricing and ratemaking and in the general adoption of analytics as a competitive reality find themselves either gleeful or anxious, based on the successes of their investments in analytics. Those who waited to see what can happen are now buying their way into the game via experienced consultants. But the latecomers now need immediate execution of their information technology and business process departmental functions because they no longer have time to determine those elements by trial and error. Errors in execution are now costlier than ever. The pace of improvement and innovation in predictive modeling and the availability of new data sources continue to accelerate.

Advanced analytics play a critical role in the underwriting practices of the modern insurance enterprise. New risk-based modeling methods enhance ratemaking engines and improve risk management decision-making across all lines of business. Carriers must take advantage of emerging analytic methods to optimize customer service and the lifetime value of policyholders in their books of business. Those carriers who sit and wait will soon experience the negative effects of the cost of doing nothing. G

Kevin B. Thompson is senior vice president of ISO's Insurance Services Department. Marty Ellingsworth is president of ISO Innovative Analytics (IIA), a unit of ISO focused on delivering advanced predictive analytic tools to the property/casualty insurance industry. More information is available at www.iso.com.