Behavior – The New Frontier in Driver Classification presented by Don Bashline RightFind Technology If behavior is the answer, what’s the question? • Why do crashes happen? – Is it because the driver missed a mortgage payment? – Because the driver is a man? – Lives in a city? – Is young? Correlation vs. Causation • All the standard classification criteria correlate highly with claims • None of them represents causes of claims What Causes Crashes? • “The most important factor influencing traffic safety is individual human behavior” - Leonard Evans • More than 90% of crashes linked to road user - Indiana University study What Types of Behavior Matter? • Risk-taking behavior – speeding – running red lights – aggressive driving • Alcohol use • Errors of judgment or perception Can we predict these behaviors? • Doing so would give us a way to predict crash (and with it, claim) frequency for a driver • We could then classify and price this driver more accurately Yes we can! • Personality traits linked to crashcausing behaviors can be measured using standard techniques • More than 50 years of research demonstrates connection between personality and crash frequency How We’re Making it Work 1 • Affiliated with academic behavioral psychologist to develop test questions – survey of literature – administer questions – evaluate questionnaire results How We’re Making it Work 2 • Initial Results were Encouraging – certain personality traits correlated highly with insurance claim experience – these traits were not always those intuitively associated with aggressive driving How We’re Making it Work 3 • First Market Research Co. administered revised questionnaire across country and held focus groups – results of first survey largely confirmed – focus groups were particularly revealing How We’re Making it Work 4 • Retaining exam construction expert – will use proven items and new ones measuring same traits to develop tests – these tests should be robust, capable of validation What’s next for RightFind • We expect approval of our business method patent in 1st quarter of 2003 • end of 1st quarter 2003 will have array of exams available for licensing • We’ll also license technique to developers How will the tests be used? • Refining classes with large “between variance” - find good risks among young drivers, bad records • “Bad things happen to good people” improve predictability of driving record What will they look like? • Can be given on phone, as part of application, by agent, or on internet • Length from a few questions (1-2 minutes) to 10-15 minutes. • Questions directly related to personality attributes Why they’ll work • Classification refinement key to profitability • Equity - enable individual drivers to be rated on causative factors • Claims are infrequent - bad drivers will often have good driving record (and vice versa) but what about...? • • • • • Change is difficult Will regulators or consumers object? Can the tests be “gamed?” Dose/response problem How well can we measure this? Down the road for RightFind • Continue refining driver tests • Med Mal? Patients likelier to sue Docs with bad attitudes • Workers’ Comp? Management attitudes toward work and workers mean a lot • Commercial auto? Truck and taxi drivers Down the Road for Risk Classification • Eternal conflict: insurers want to use relevant information, but is it “socially acceptable?” • Genetic markers? • Brain mapping? Questions? Discussion? Thanks for your attention and for the invitation! Visit us at www.rightfindtech.com