Underwriting Technology – Focus on Automation and Analytical Models January, 2014 BLOG POST
Identifying, understanding and underwriting risks and exposures are critical for insurers Insurance business differs from other industries in two important ways. First, unlike other businesses it attracts risk rather than shedding it. Second, the true and final cost of the product is not known at the time of sale. Hence, any loss or profit is uncertain. Today, as insurers are struggling to achieve financial targets amid weak economic environment, it becomes essential to efficiently choose risk and have accurate pricing so as to minimize loss and improve top-line. That brings underwriting into the picture. Underwriting tries to predict future losses and helps insurers effectively price the products that provide protection against these losses. Huge catastrophe losses in recent past have also put pressure on insurers to redefine and reassess the basics of underwriting so as to make profit in both hard and soft market cycle. Today, insurers are investing in the underwriting technology for better understanding of future losses so as to make profit. Predictive modeling and data mining are two of the tools widely used in the underwriting process.
Insurers’ technology priorities for underwriting Insurance underwriting is witnessing rapid changes on the back of technology development. Four key drivers of increased acceptance of underwriting technology by insurers are technology advancement, internal and external data availability, tough business environment and quest for competitive edge. Most insurers are adopting efficient underwriting processes, and few aggressive players are exploring new advanced technologies like data mining and predictive analytics. These advanced systems represent the next generation of underwriting tools to achieve insurers’ goals of predicting future losses and pricing the products accordingly.
State-of-the-art underwriting ecosystem
Underwriting Technology – Focus on Automation and Analytical Models
Page 2
Insurers worldwide are investing in various tools and technologies across the underwriting ecosystem, such as fully electronic workflow, automation, data verification and predictive modeling. However, the two key focus areas have been predictive modeling and automation. Predictive modeling has been used by insurers in claims fraud detection and consumer insights before gradually migrating to underwriting. Technically, it takes into account the historical data and tries to identify patterns that were previously unknown, to formulate an algorithm that can be used to solve a specific business problem. Unlike traditional underwriting approach, which relies solely on past loss data, it is a combination of internal historical data of carriers and various external variables related to the policy underwritten. These solutions are also helpful for insurers to take corrective actions based on portfolio segmentation, which is important to understand the different levels of profitability with a different portfolio segment. P&C insurance business has seen a significant impact from the automation of underwriting systems, so much so that there has been a long debate in the industry if models can truly replace skilled and experienced underwriters. The industry has well accepted automation in underwriting process, but it still relies on the experienced underwriting professionals in reviewing large and complex risks. Risks such as worker’s compensation and other simpler personal risks have indeed gained speedy delivery and removal of manual data inputs from automation. Companies use automation to save time and resources in simpler and less complex risks and direct this saving towards more complex and larger risks.
Conclusion Traditional underwriting has been using historical loss data alone for pricing. On the other hand, predictive modeling uses past data and also takes into account external data to identify trends and predict losses. While historical loss experience is important, it is little related to future losses as it is not representative of the underlying future exposures. Focus on future risk exposure is important for understanding and estimating the loss potential. For instance, in an auto policy, the projected or estimated kilometers driven can be requested during the beginning of the policy period, which can again be confirmed at the end of it. Insurance industry has improved efficiency of underwriting system through analytical models but there is still a long way to go where future exposures are also taken into consideration for developing tools and models.
Source: SOA, Insurance Networking, Insurance Journal, Boire Filler Group, Gen Re
Underwriting Technology – Focus on Automation and Analytical Models
Page 3