Trading Places: Transforming Personal and Commercial Lines through Loss Control

By Jim Weiss

Personal lines (PL) insurance has become rather impersonal over the years. Once a homeowners or auto policy has been booked, insurers might not hear from a policyholder until there’s a billing question or, much worse, a claim. In contrast, commercial lines insurance has traditionally been less predicated on simply waiting for something bad to happen.

It’s no secret why loss control has historically been the province of commercial lines (CL) insurers. Complex, multimillion-dollar CL risks arguably justify greater preventive handling — because of their catastrophic risk potential — than those of private individuals. An engineer or building inspector is likely to have more impactful suggestions about a manufacturing plant or construction site than about the average ranch or bi-level home. Similarly, the average household vehicle doesn’t generate enough claims in a lifetime to warrant the experience-rating approach used to coinsure large or midsize commercial fleets. Given those diseconomies of scale, it is not surprising that PL insurers’ efforts would more likely be centered on predicting losses (for purposes of pricing and underwriting) than on controlling them.

Unlike CL, the homogeneity and increased availability of information about PL exposures allow insurers to consider detailed information, such as the building characteristics of a home or the actual manner in which a vehicle is operated, when processing applications and renewals. Such approaches push the bounds of feasibility for CL because of the unique nature and complexity of information surrounding each risk.

But a funny thing has happened as PL insurers have accrued ever more predictive capabilities: They’ve begun to derive insights that — if effectively communicated back to policyholders — can help prevent the very losses being predicted, thereby reducing premiums. In other words, PL is beginning to look more like CL.

Usage-Based Auto Insurance Switches Gears

A prime example of the “commercialization” of PL is usage-based auto insurance, also known as UBI. Credibility concerns for personal auto have traditionally limited experience rating to broad strokes such as “accident-free” discounts. UBI now allows policyholders voluntarily to install small telematics devices in their vehicles that collect more detailed experiential information, such as when, where, how often, and the manner in which those vehicles are operated. This information may then be analyzed and used to determine insurance discounts. In this way, UBI is analogous to experience rating, because policyholders’ premiums are more related to their safety record as opposed simply to their demographics.

Conventional wisdom suggests that safer drivers will naturally “self-select” UBI programs with the confidence that they will likely earn discounts. Such thinking is contradicted by recent estimates that a maximum of only 5 percent of policyholders will enroll in UBI programs by the end of 2014.1 For perspective, consider that Verisk’s Safety Scoring® models estimate the riskiest 20 percent of drivers to be up to ten times more likely to experience an accident than the least risky 20 percent. Therefore, it appears a solid majority of very-low-risk operators are not opting into UBI despite multiple carriers in most states offering such programs. Further, given the strong ability to identify which driving patterns could result in accidents, insurers whose efforts focus solely on selection and pricing may miss a significant opportunity to educate riskier operators.

Figure 1
Screenshot from UBI portal

The weekly journal feature of Verisk’s online portal allows UBI policyholders to view their trips over a multiweek period. Each trip is classified as low-, medium-, or high-risk according to Verisk’s Safety Scoring models. Policyholders can click on trips to see which maneuvers were classified as most risky and view animations of their trips superimposed on a map. Those insights can allow policyholders to improve their score and potentially earn greater discounts under the Safety Scoring rating rule, which ISO has filed in more than 35 jurisdictions.

Some PL insurers have begun to respond by offering a more compelling customer experience. For example, recent UBI offerings have included supplementing or replacing discounts with value-added services (VAS), including online portals, emergency or roadside assistance, and concierge services. It’s estimated that 57 percent of U.S. drivers would be receptive to UBI if it offered VAS — and potentially willing to pay additional premium for it.2 By relying on data previously used only for predictions, insurers can now use the resulting insights to provide policyholders with certain real-time information (for example, notifications of traffic jams) to help reduce accident risk. Studies from the fleet sector point to promising results. For example, one natural gas company reduced accident rates by 50 percent for its fleet by sharing weekly “scorecards” with drivers that highlighted areas for improvement.3 Online portals such as the one shown in Figure 1 provide analogous information to UBI policyholders. If VAS were half as effective as fleet telematics in reducing accidents, then insurers offering UBI would likely more than cover their costs of discounts and underlying telematics technology.

It should be noted that the possibilities of UBI have not been lost on CL insurers, many of whose policyholders may already use telematics for fleet management. Although adoption lags PL, where at least eight of the top ten PL auto carriers have launched UBI, commercial auto programs are also available that range from offering discounts for fleets that use telematics to safety-based pricing (similar to PL) that accelerates the estimation process for experience rating.

Building Characteristics Come into Focus

Further evidence of the increased importance of loss control for some PL insurers is reflected by use of aerial imagery. Advanced pricing tools (such as ISO Risk Analyzer®) use property characteristics including age of roof, number of rooms, square footage, lot size, and amenities (for example, pools) to estimate homeowners loss costs. However, like most advanced analytics, this approach is only as accurate as the information that empowers it.

The most readily available data about building characteristics typically resides with local tax assessors. But that information runs the risk of being incomplete or outdated. Not every homeowner’s first inclination is to report the latest repair or alteration to the local taxman or insurance agent, and even if it were, there are likely to be wide variations in the quality of data collected between different jurisdictions. As a result, during pricing and underwriting, many PL insurers send property inspectors to homes to the tune of $100 million per year, by Verisk’s most recent estimates. Those costs are to some degree wasted when the assessor information is deemed fully accurate or, conversely, the policy application is not accepted.

Figure 2
Aerial image related to virtual property inspection

house.jpg

This high-resolution image, snapped from 14,500 feet in the air, can be analyzed to estimate the dimensions of the property and structure — tasks once necessitating a property inspection — for use in underwriting or advanced pricing approaches. Mere visual inspection reveals hazards such as a water slide. An agent with access to these images might offer loss control suggestions, such as fencing off the basketball court or capping the chimney.

In contrast, consider how property inspections are an essential part of doing business for CL. Some inspectors’ findings are geared as much toward improving risks as estimating them (for example, through best practices to safeguard risky machinery or improve fire protection). For risks as predictable as those in the average home, a rigorous on-site inspection may not be as necessary. For example, some PL insurers may decide to use high-resolution aerial imagery to expedite claims handling. For a fraction of the cost of having someone scale and survey a damaged roof, an analysis of a high-resolution aerial image can predetect the roof’s dimensions and finer details, such as satellite dishes and individual shingles. This approach may also help with claims settlement and accelerating the adjustment process in cases, for example, when a damaged property is cordoned off.

There are number of reasons for insurers to take advantage of aerial imagery data before a claim occurs. PL insurers can use aerial images to verify critical information about insured properties during the underwriting process and get more mileage out of their advanced pricing. Figure 2 displays an image from a virtual property inspection, from which one can detect dimensions, number of stories or units, roof type and material, and pools or trampolines. This information can serve as the basis for more informed (pricing or underwriting) decisions, but it can also be the beginning of a conversation. Because PL insurers have already identified which property characteristics may present the greatest risk, a logical next step is for agents to suggest property modifications to the policyholder — which may help lower premiums to the point of paying for themselves over time. Agents can also offer potential coverage solutions to excluded hazards on the property.

Trading Places

So, it would appear the tables are turning. While CL carriers have traditionally dedicated significant effort to working with policyholders to control losses, PL insurers have been motivated by economics to devote greater resources toward loss prediction. However, technologies such as telematics and aerial imagery are changing the game by providing cost-effective methods for PL insurers to convert predictive insights into preventive ones. In turn, pricing and underwriting decisions can evolve into loss control discussions that help reduce risk and improve the policyholder experience. In this way, technology is actually helping put the human touch back in personal lines insurance.

Jim Weiss, FCAS, MAAA, CPCU, is manager of Personal Automobile Actuarial at ISO Insurance Programs and Analytic Services.

  1. Source: “Report: Three in four insurers moving forward with pay as you go
  2. Source: “Drivers are overwhelmingly receptive to usage-based auto insurance, according to Towers Watson survey,” 72% of the 79% of drivers receptive to UBI, or 57% of total drivers, would be willing to pay for VAS.
  3. Source: “Encana reduces accident rate by 49% in 15 months with driver safety initiative”