Buildings burn. They shake. They blow up, and they blow down. The events and situations causing damage to buildings are unchanging.
Executive SummaryWith massive proprietary databases and expertise in data science, large property insurers used to have an information advantage over smaller competitors. But the advantage is disappearing in the presence of a new "data-driven underwriting paradigm," giving smaller companies access to detailed data and predictive modeling technology provided by third-party vendors. Here, executives of one such vendor detail the promise of the new paradigm in allowing underwriters from firms of all sizes to avoid skyrocketing loss ratios of recent years, and also allowing property owners to better understand and mitigate risks while they experience the added benefit of dealing with a simpler insurance-buying process.
For more than 300 years, insurance companies have indemnified property owners for the consequences of these perils. The pillars of property underwriting—construction, protection, occupancy, exposure—are the same now as they were a century ago. It seems that the science of underwriting in this well-established and ostensibly well-understood line of insurance should be perfected by now.
Property insurance continues to be a challenging line of business for many insurers, however. Despite centuries of experience with property perils and exposures, property loss ratios skyrocketed in the recent past, forcing the market to make abrupt and painful pricing corrections. Even with sharply higher rates, insurance executives are concerned about both the long-term and short-term profitability of the line.
One perennial problem is inadequate underwriting information. Throughout the history of property insurance, underwriters have contended with insufficient, incomplete and unreliable data. Various third-party information sources have been available since at least the 1870s when Sanborn maps—which presented block-by-block information on buildings in urban areas—were introduced. In the past, however, third-party data was limited in its scope and application. Underwriters were forced to rely significantly on information provided by property owners and their agents in insurance applications, which is notoriously inaccurate.