Improving Risk Assessment Accuracy: The Motivation for Usage Based Insurance

With virtually every country requiring some form of insurance for automobile owners, drivers are constantly looking for ways to save money by lowering their monthly or annual premiums. When that search intersects with insurance companies' desire to accurately assess individual customer's risk to limit their liability exposure, the result is usage based insurance (UBI). With several different varieties of usage based insurance available, UBI appears to offer potential benefits to both consumers and firms. This paper examines the various types of usage based insurance available in the United States and Europe before considering the potential advantages and drawbacks of more large-scale adoption of this new insurance offering.

In contrast to the typical fixed rate insurance which determines a driver's premium using fixed formulas based on categorization, usage based insurance uses a driver's individual behavior to determine the cost of their insurance (NAIC 2014). Instead of using past indicators of risk, usage based insurance attempts to calculate a driver's future risk based on their current driving habits (DeGraff 2013, 9). Because risk categories such as age, gender, and credit history are often imprecise and even inaccurate predictors of future risk, insurers are turning to usage based insurance in an effort to more accurately identify which drivers are safe and which are not (NAIC 2014). Doing so would allow insurers to charge customers premiums more closely aligned to their actual risk. Safer drivers would be rewarded with a lower rate while risky drivers would bear more of the cost for their liability (Insurance Journal December 8, 2011). Different variations of usage based insurance have emerged over the past several decades. Among the most common are distance based, pay as you drive (PAYD), and pay how you drive (PHYD). One of the primary difference between the different methods is that distance based insurance uses existing technology built into cars while both pay as you drive and pay how you drive rely on telematic devices to collect data (Rose 2013, 1).

The simplest form of usage based insurance is known as distance based insurance. While many insurers ask applicants to estimate their yearly mileage, evidence shows that such estimates typically under-report mileage totals – especially if doing so will lead to financial savings for consumers (Rose 2013, 1). In response, some insurers have begun using odometer readings to verify these mileage estimates and establish a more accurate record of yearly automobile usage (IINC April 25, 2011). This method of insurance has gained higher popularity in Australia and other countries as well as the United States (Litman 2012, 11; IINC April 25, 2011). An initial concern with this method was that consumers would simply lie about their odometer readings in order to obtain lower raters. In response, insurers implemented a process of odometer auditing (Litman 2012, 11). In these yearly audits, a certified mechanic or other entity inspects the odometer and makes an official record of the total mileage driven (Litman 2012, 10). Some critics have pointed out that drivers may tamper with their odometers to reduce the mileage recorded, however manufacturers are increasingly installing tamper resistant and tamper evident devices (Litman 2012, 12). If an audit discovers evidence of tampering, the insurer is then able to take adverse action against the customer (Litman 2012, 12).

Distance Based Insurance

Distance based insurance is touted by insurers and a cheap and simple alternative to both the prevalent insurance regimes and the more invasive telematic devices. Because it utilizes a car's built-in odometer, distance based insurance does not require drivers to purchase or install any additional tracking devices (NAIC 2014). While drivers are required to undergo a yearly odometer audit, the cost of such audits is estimated at less than $10 per vehicle (VTPI 2013). Many drivers are more than willing to pay that amount particularly if they believe they will save money due to a low distance traveled (Business Week August 21, 2013). The same widespread support does not exist for the mandatory implementation of such a scheme (Litman 2012, 18). Consumers prefer the option to switch their coverage type without being forced to do so. This may be due to consumers who frequently travel long distances worrying that their premiums will increase under such an arrangement (Litman 2012, 18).

Both of the other forms of usage based insurance – pay as you drive and pay how you drive – supplement a car's built in equipment with some type of telematic device which records data and transmits it to the insurer (Rose 2013, 2). The most straightforward telematic approach is pay as you drive. This version is highly similar to the distance based option discussed above. However, instead of relying on odometer measurements, pay as you drive requires the installation of a device which measures the distance traveled by the car (Rose 2013, 2). Typically, this device uses GPS tracking to measure the distance traveled although it may also measure the total amount of time the car is in motion (Litman 2012, 13). In addition to GPS tracking, the telematic device may record the time of day during which a trip occurred (Litman 2012, 13). This data can be used to penalize drivers who are frequently on the road between midnight and 4 AM (Henry September 30, 2012). The most frequent technique used to measure mileage using a device connected to an automobile's On-Board Diagnostic (ODB) port (WSJ December 9, 2013). These independent records are touted as having more accuracy than the self reported data upon which distance based insurance plans rely. One basic drawback to the implementation of both pay as you drive and pay how you drive is the added cost of the telematic device (NAIC 2013). Some insurers have discussed eliminating the need for an additional tracking device by having customers install an application on their smart-phone in order to avoid this cost (Golia September 14, 2011).

Pay how you drive plans also utilize telematic devices to transmit data to the insurance company. However, the data that is recorded by these devices is far more detailed than that captured by the simple GPS pay as you drive modules. The premise is that a driver's level of risk is determined by their typical driving habits in addition to the total number of miles driven (Rose 2013, 4). That is, a safe driver driving 60,000 miles a year should not be charged the same premiums as a risky driver who drives that same distance. As a result, in addition to the total mileage driven, pay how you drive plans generally track the exact date and time of day at which the trip occurred (NAIC 2014). More sophisticated devices have introduced accelerometers which measure rapid acceleration and braking (Rose 2013, 2). The idea behind this is that drivers who accelerate quickly and brake abruptly are more likely to be involved in an accident and therefor represent a larger liability risk to the insurer (NAIC 2014). Some critics have argued that telematic devices are unable to distinguish between repetitive braking during rush hour traffic and harsh braking due to unsafe driving practices (Jergler October 2, 2013). As a result, some safe drivers may pay higher premiums than their actual driving habits would indicate (Jergler October 2, 2013). Other telematic devices also measure the frequency of lane changes as insurers often view frequent lane changes as a warning sign of risky behavior (Rose 2013, 2).

In the United States, usage based insurance was first introduced by Progressive Insurance (Rose 2013, 2). In 1998, Progressive began offering a program called Autograph (Litman 2012, 13). For a $15 equipment fee, a GPS device was installed on the customer's car to track total mileage (Litman 2012, 13). Once installed, however, the device also gave drivers access to other services like roadside assistance and theft recovery at a significant discount while saving drivers an average of 25% compared to traditional insurance plans (Litman 2012, 13). In the decade and a half since then, the market for usage based insurance has exploded. From 2012-2013, Progressive reported $1 billion in premium revenue directly from usage based insurance policies (Rose 2013, 2). According to the company, 1.6 million drivers have signed up for its pay how you drive program – now called the Snapshot Discount – since 2008, and Progressive has used telematic devices to collect data on approximately 8 billion miles driven (Jergler October 2, 2013). Progressive's offering has also been joined in the marketplace by numerous other companies. Allstate's Drivewise, GM's On-star, Safeco's Rewind, and Travelers' Intellidrive are just a few of the pay as you drive and pay how you drive programs available to consumers (Dewri, Annadata, Eltarjaman, & Thurimella 2013, 1). A recent study predicted that 25% of the total automobile insurance market in the United States would rely on telematic devices to provide usage based insurance by the year 2020(Rose 2013, 2). The total value of that market share is expected to reach $30 billion by that same year (Rose 2013, 2). Consumer surveys have shown that up to 79% of drivers would consider purchasing usage based insurance with that number growing to 89% if the driver believes that the change would result in monetary savings (Insurance Journal September 5, 2013).

Usage based insurance utilizing telematic devices is also gaining popularity in Europe – the world's largest car insurance market. According to data from 2011, over one million customers were insured under a usage based policy (Rose 2013, 2). While such policies are available in every country, they are especially prevalent in the crowded insurance markets of Italy, Spain, France, and the United Kingdom (Rose 2013, 2). Such a distribution is unsurprising given the widespread car ownership in those countries compared to others in the Eurozone. Much like the market in the United States, the forecast for usage based insurance in Europe is quite optimistic. According to a white paper by the Statistical Analysis System (SAS) Institute, the value of insurance premiums for policies utilizing telematic devices is expected to top $66.7 Billion by 2020 (Rose 2013, 2). With such exponential growth, it is unsurprising that insurers and even policy think-tanks in other countries such as Australia are looking to expand the market share of usage based insurance (VTPI 2013). At this point, such policies are just one option in an open marketplace, but some have suggested that mandating their use could provide societal benefits (Litman 2012, 19).

Usage based insurance plans have been widely praised for their potential benefits both to consumers and insurers – not to mention positive environmental externalities. Perhaps the most basic claim of proponents is that such policies save drivers money (Litman 2012, 18). According to Progressive, their Snapshot program saves drivers roughly $30 million every year equivalent to $150 per driver (Insurance Journal December 8, 2011). The savings are realized by identifying safe drivers which pose a lower risk of liability which in turn allows the insurance company to reduce those drivers' monthly or annual premiums (Perr & Knight 2012). With insurance premiums rising by 35% in 2013, the cost savings alone may be enough to convince some drivers to give pay how you drive policies a try (Newcomb August 30, 2013). If usage based insurance policies allows car owners to reduce the cost of car insurance – and by extension the fixed cost of owning a car – a study showed that car ownership could become more affordable increasing the total number of vehicle owners by 1% (Litman 2012, 15). On the insurers' side, being able to identify safe and risky drivers allows companies to avoid large payouts by either increasing premiums or denying coverage to drivers with the highest risk (Perr & Knight 2012).

Because the data collected by the telematic devices is often made available to consumers, drivers are able to monitor their own safe driving scores which often leads to a decline in risky behaviors (Rose 2013, 2). According to Progressive, drivers enrolled in their Snapshot program are 10% less likely to be involved in an accident and 20% less likely to be given a speeding ticket (Insurance Journal December 8, 2011). A similar reduction in accidents can also be achieved with simple distance based or pay as you drive policies which tend to reduce the total number of miles driven each year as drivers look to cut down on costs (Litman 2012, 5). A 2003 study found that just a 10% reduction in total vehicle mileage could reduce accidents and traffic casualties by 14-18% (Edlin 2003). As distance based insurance policies reduce the total number of miles driven, it is expected that automobile pollution will also decline. According to the Victoria Transport Policy Institute pay as you go pricing has the potential to reduce fuel consumption and CO2 emissions by up to 13.5% (Litman 2012, 13). Pay how you drive insurance policies have gained popularity in Europe as a means of eliminating gender discrimination – previous category based pricing typically charged men 2-3 times as much for insurance as women (Rose 2013, 3). By measuring drivers' actual performance, insurers are able to determine rates based on individual risk rather than the less accurate categorical risk (NAIC 2014).

The positives of usage based insurance do not come without some negative drawbacks. From the perspective of insurance companies, telematic devices represent a potential overload of data (Perr & Knight 2012). With the devices recording data every second, the amount of data that must be sorted through in order to create a useful measurement of driving habits measures on the order of a terabyte – equivalent to 1,000 gigabytes – per customer; a daunting task for even a streamlined organization (Rose 2013, 3). Naturally, analyzing that much data is an expensive endeavor and that is without taking into account the telematic devices themselves (NAIC 2014; Newcomb August 21, 2013; ). Pay how you drive insurance is also fraught with risks for consumers. As with any GPS tracking device, there are grave privacy concerns (Jergler October 2, 2013). A recent study by the University of Denver found that even if exact GPS coordinates are not recorded, a simple measurement of trip length can be used to track a customer's movements (Dewri et. al. 2013). Privacy rights activists are naturally in arms about such possibilities although insurers have countered that such concerns are moot due to the widespread use of smart-phone applications which already track individuals' every move (Jergler October 2, 2013). Beyond the collection of data, the objection has been raised that insurance companies could potentially sell data they collect to independent companies (Jergler October 2, 2013). According to privacy experts, such a move would raise serious concerns and would likely violate various state consumer protection laws (Jergler October 2, 2013). No such arrangements have been formed, but their specter could dissuade drivers from embracing more comprehensive telematic monitoring.

While no insurance algorithm is likely to identify individual drivers' risk with complete accuracy, usage based insurance offers a promising alternative to the current fixed rate insurance regime. By taking into account driver's current risk factors rather than categorizations based on past risk, pay how you drive insurance plans offer safe drivers a chance to save while simultaneously allowing insurers a method of identifying the drivers who truly represent a potential liability. Such benefits are likely to continue driving the rapid expansion of such programs into a multi-billion dollar market offering in less than a decade. The growth of usage based insurance may be held back somewhat by privacy concerns and insurance companies' ability to process data, but the wide-spread use of smart-phones with tracking applications provides support for the idea that acceptance is just a matter of time.

Works Cited:

Business Wire. "Considering Telematics?" August 21, 2013.

DeGraff, Katie. "Usage Based Insurance: A Revolution is Underway." Ohio Insurance Institute.April 2013.

Dewri, Rinku, Prasad Annadata, Wisam Eltarjaman, & Ramakrishna Thurimella. "Inferring Trip Destinations from Driving Habits Data." University of Denver. 2013.

Edlin, Aaron. "Per-Mile Premiums for Auto Insurance." Economics for an Imperfect World: Essays In Honor of Joseph Stiglitz, MIT Press: 2003

Golia, Nathan. "Mobile Advancements Lead to Rise in Usage-Based Insurance." Insurance & Technology. September 14, 2011.

Henry, Jim. "Drivers Accept Monitoring Devices, to Earn Discounts on Auto Insurance." September 30, 2012.

Insurance Information Network of California (IINC). "Mileage Based Auto Insurance." April 25, 2011.

Insurance Journal. "70 Million Drivers Missing Out on Car Insurance Savings: Insurer." December 8, 2011.

Insurance Journal. "How U.S. Consumers Feel About Usage-Based Auto Insurance." September 5, 2013.

Jergler, Don. "Researchers Question Privacy of Usage-Based Auto Insurance." Insurance Journal. October 2, 2013.

Litman, Todd. "Distance-Based Vehicle Insurance as a TDM Strategy." Victoria Transport Policy Institute. December 10, 2012.

National Association of Insurance Commisioners (NAIC). "Usage Based Insurance and Telematics." Last updated January 2, 2014.

Newcomb, Douglas. "Are You Ready to Sign Up for 'Usage-Based' Auto Insurance?" August 30, 2013.

Perr & Knight. "Usage Based Insurance." 2012. Accessed February 13, 2014.

Rose, Stuart. "Telematics: How Big Data is Transforming the Auto Insurance Industry." SAS Whitepaper Series. 2013.

Victoria Transport Policy Institute (VTPI). "Pay As You Drive Vehicle Insurance." TDM Encyclopedia. March 12, 2013.

Wall Street Journal (WSJ). "Better Drivers Pay Less for Car Insurance with Snapshot® Usage-Based Insurance Program from Progressive." December 9, 2013.