WIXLIB

The Challenges Facing P&C Insurers inImplementing Price Optimization and the BasicFramework to Confront them2014 CAS RPMWashington, DCJun Yan, Ph.DPeter Tomopoulos, ACAS, MAAAKelli Broin, FCAS, MAAA

AgendaWhy Price Optimization?Price Optimization is Far Beyond a Mathematically Optimized PricePrice Optimization FrameworkChallenges Insurers Face Implementing Price OptimizationConcluding Remarks1Deloitte Consulting LLPConfidential and Proprietary – Copyright 2014 – Deloitte Development LLP

Why Price Optimization?2Deloitte Consulting LLPConfidential and Proprietary – Copyright 2014 – Deloitte Development LLP

Balancing premium volume with target loss ratio Traditionally, insurance companies and actuaries have relied on the insured’s expectedclaim cost along with claims history (experience rating) and underwriter judgment(schedule rating) to determine the appropriate premium. Price optimization integrates the risk’s insurance price and the market demand forinsurance. Market demand reflects:CustomerBehaviorElasticity ofDemandCompetitorPricing Price optimization enables control of balancing premium volume and loss ratio– Can increase premium volume without impacting the target loss ratio, OR– Can decrease target loss ratio without impacting premium volumePremiumLossRatioPremiumLoss RatioOR3Deloitte Consulting LLPConfidential and Proprietary – Copyright 2014 – Deloitte Development LLP

Three levels of pricing sophisticationInsurance pricing can be classified in three levels of sophistication:Basic Rating Plans, Underwriting Models, and Market Demand ModelsPrice OptimizationIncreasing SophisticationPricingLevels1. Basic Rating Loss cost (pure premium)model based on exposureand traditional ratingvariablesComponentsof PricingLevels4 Rating variables driven byindustry and regulatoryrestrictionsDeloitte Consulting LLP2. Underwriting3. Market Demand Refinement of Level 1 (basicrating) loss cost estimate Reflects strategic, regulatory,and market constraints Generally used to rationalizethe process of assigning tiers,credits, debits and companyplacement Builds upon results from Level 2(underwriting) and Level 1 (basicrating) Predictive variables could berating and non rating frominternal and external datasources Applies customer price elasticityto optimize priceConfidential and Proprietary – Copyright 2014 – Deloitte Development LLP

Current market capabilitiesInsurers have instituted basic forms of price optimization techniques in recent years, butfew have integrated them within a strategic and operations pricing platform needed toovercome challenges to successfully implement optimization techniquesPrice OptimizationCurrent Market Capabilities Very few companies have fully developed such integrated platforms Most companies have some level of competitive pricing insights but lack integration with customer andmarket demand behaviorPricingLevelsCurrentMarketCapabilities51. Basic Rating2. Underwriting3. Market Demand Vast majority of companieshave sufficient or leadingpractice capabilities in rating There is a large degree ofvariability in approachescompanies take in this area Some companies may havedeveloped customer demandmodels Ability to rerate historicalbook (in-force and lapsedpolicies) varies by company Most companies have somecapability but generally notintegrated with customerbehavior Very few companies havedeveloped a well integratedplatform for operationDeloitte Consulting LLPConfidential and Proprietary – Copyright 2014 – Deloitte Development LLP

Key concerns impacting implementationAs insurers attempt or have attempted to optimize their rates, key concerns have beenidentified that are becoming critical for companies to address to incorporate these priceoptimization techniques in their pricing structures.Key Concerns6Ignoring underwriter adjustmentsBy not reflecting the individual riskcharacteristics at a policy level, manyof the characteristics are not reflectedin base rating.Large-scale effort to implementLarge amount of data to collect, storeand analyze; and then, how do youreact to all that information?Difficulty to monitor and measureAs more characteristics areconsidered in an optimization model,the result will become a “black box”and hard to understandRestrictions for commercial linesInformation on final price is limitedDeloitte Consulting LLPConfidential and Proprietary – Copyright 2014 – Deloitte Development LLP

A different view to deal with the core challengesWhen a strategic and operational pricing platform effectively integrates profitability,market demand and customer price elasticity, it will allow insurers to successfully managefuture market cycles and dynamicsPrice OptimizationEvolving Practices Platform enables dynamic analysis of risk, profitability and growth simultaneously Platform allows for scenario testing to assist in operation and strategic decision making of price changesfor highly segmented marketPricingLevelsEvolvingCapabilities71. Basic Rating Well developed loss costbased rating that generatetens of thousands ofindividual marketsegments/price points Companies can easily rerate entire historical book forany rating changeDeloitte Consulting LLP2. Underwriting Reflect risks at the individualpolicy level in pricing a policy Companies have ability toeffectively generate andcapture competitive rates foreach policy and new businessquote3. Market Demand Data from rating andunderwriting components alongwith new business and renewaldata, and data from competitivemarkets can be used to developcustomer price elasticity andmarket demand curves. Curve is combined with segment/tiering analyses to refineknowledge of customer behaviorand to take appropriate actionsConfidential and Proprietary – Copyright 2014 – Deloitte Development LLP

Price Optimization is Far Beyonda Mathematically Optimized Price8Deloitte Consulting LLPConfidential and Proprietary – Copyright 2014 – Deloitte Development LLP

Questions for overall designBefore designing a price optimization model, there are leading strategic questionsinsurers need to considerLeading QuestionDoes price optimization have to be asoftware driven black box? Wheredoes the Operations Researchsoftware kick in?Expected Outcomes1)2)3)1)What if a price optimization is atransparent process, but not a blackbox?How can we turn a software drivenprice optimization process to be acontrollable open box?9Deloitte Consulting LLP2)3)1)2)Demand modelingLoss cost estimation or loss ratioestimationIntegration of loss cost and demandingfor optimized pricingPrice optimization process should betotally manageable and controllableThe performance of price optimizationwill be easily monitoredPrice optimization will not be purely amathematical result; it will turn to be adecision making tool for strategic pricingSimplify the focus of price optimizationCreate a simple segmentationspecifically for price optimizationConfidential and Proprietary – Copyright 2014 – Deloitte Development LLP

Questions for overall design - ContinuedBefore designing a price optimization model, there are leading strategic questionsinsurers need to considerLeading QuestionWhat is the ultimate goal of priceoptimization?What are the major premiumcomponents?What are the major components forprice optimization?10 Deloitte Consulting LLPExpected Outcomes1)2)Increase Profit (decrease loss ratio)Increase premium volume1)2)3)BenchmarksUnderwriting adjustmentsMarket adjustment1)Risk premium with expenses beforemarket adjustment ( 85%)Premium adjustment based oncustomer demanding and marketcompetition ( 15%)2)Confidential and Proprietary – Copyright 2014 – Deloitte Development LLP

Benefits of a risk-based approachAn approach that directly incorporates underwriting adjustments into determining a pricefor insurance offers a number of key advantagesQuotePrice11 Deloitte Consulting LLPPrice ComponentsSourceOutputBasic RatingRating ents /Underwriting TiersTechnicalPriceMarketDemandMarket DemandAdjustmentMarket PriceConfidential and Proprietary – Copyright 2014 – Deloitte Development LLP

Advantages of including underwriting componentsThe platform is built on a policy level where projected loss ratio, estimated conversion /retention rate, and estimated price elasticity are scored for each policy, which can then besegmented for further optimizationBasic RatingUnderwriting Pricing By coverage, on exposure level Rated on the policy level To be used in rating To be used to identify rating plan deficiency Target: Loss Cost Target: Loss Ratio based on manual premium Factor based point estimation Loss ratio ranking and lift curve basedestimation Not separated by new business and renewalbusiness Separated by new business and renewalbusiness Only using regulated rating variables Much wider range of variables are applied Rating lookup tables are fixed in ITimplementation Lookup tables are dynamically adjusted in ITimplementationA key advantage of including underwriting components is that the insured’s priceelasticity and demand behavior is on the final price at the policy level and not thecoverage and sub-component level. Optimizing price on the sub-coverage levelcreates a gap between the results and the insured’s price behavior.12 Deloitte Consulting LLPConfidential and Proprietary – Copyright 2014 – Deloitte Development LLP

Price Optimization Framework13 Deloitte Consulting LLPConfidential and Proprietary – Copyright 2014 – Deloitte Development LLP

Rely on comparative raters to provide insight intoinsurers’ rates and relative position to your own ratesUsing competitor rates relative to your own and tracking win/loss rates, a price elasticity /demand curve can be createdIllustrative Price Points forIndividual Risks25New Business Conversion Rate & Price Elasticity1.81.620Price Elasticity1.4Price PointConversion y123Individual RiskComp. 1Comp. 2Comp. 345Comp. 40Price Difference with Competitors Examine company’s price versus competitor prices on individual risks and win/loss Plot win rate versus price position to determine basic price elasticity curve Can be segmented by customer type, sales channel (e.g. on-line, agent), income, region, etc.14 Deloitte Consulting LLPConfidential and Proprietary – Copyright 2014 – Deloitte Development LLP

Optimize pricing decisions based on UW tier groupsIntegrating output from an UW predictive model with the demand curve yields a platformfor insurance companies to make and optimize their pricing decisions based onunderwriting tier groups.Integrate risk-based approach using output from Underwriting scoring engine Incorporates variables that are predictive of risk and cost but that cannot be used in traditional ratemaking Analyzes risks at the policy level, not coverage levelTier Tier Factor-0.40-0.38-0.34-0.30-0.28-0.26-0.24-0.20 15 Deloitte Consulting LLPConfidential and Proprietary – Copyright 2014 – Deloitte Development LLP

Pricing lab for scenario testingThe pricing lab provides a simple framework for companies to perform scenario analysisand to drive business decisions that align with the company‘s strategic goals, whilebalancing profit and growthInputsApplicationsBy dimension (profitability & elasticity): Underwriting Tier & CorrespondingTier Factor Manual Premium Target Loss Ratio Estimated Loss Cost Testing and comparing pricingstrategies Optimizing pricing relatedunderwriting rules Competitor Pricing Information Price Elasticity and Demand ModelsA pricing lab will produce a matrix for each scenario showing the premium and lossratio impacts for each combination of dimensions.16 Deloitte Consulting LLPConfidential and Proprietary – Copyright 2014 – Deloitte Development LLP