Loss Reserving 201 It's More than Numbers

Loss Reserving 201 It's More than Numbers Derek W. Freihaut September 17, 2015 Agenda Background/Loss Reserving 101 Key Considerations Claims Handling Reinsurance Underwriting Rates External Influences Other Key Information How Actuaries Handle 1 1

Background/Loss Reserving 101 INCURRED LOSS TRIANGLE Accident Months of Maturity Year 12 24 36 48 60 72 84 96 108 120 2001 10 20 30 40 50 60 70 80 90 95 2002 10 20 30 40 50 60 70 80 90 2003 10 20 30 40 50 60 70 80 2004 10 20 30 40 50 60 70 2005 10 20 30 40 50 60 2006 10 20 30 40 50 2007 10 20 30 40 2008 10 20 30 2009 10 20 2010 10 2 Background/Loss Reserving 101 AGE TO AGE FACTORS Accident 12 24 36 48 60 72 84 96 108 120 Year 24 36 48 60 72 84 96 108 120 Ultimate 2001 2.000 1.500 1.333 1.250 1.200 1.167 1.143 1.125 1.056 2002 2.000 1.500 1.333 1.250 1.200 1.167 1.143 1.125 2003 2.000 1.500 1.333 1.250 1.200 1.167 1.143 2004 2.000 1.500 1.333 1.250 1.200 1.167 2005 2.000 1.500 1.333 1.250 1.200 2006 2.000 1.500 1.333 1.250 2007 2.000 1.500 1.333 2008 2.000 1.500 2009 2.000 Average 2.000 1.500 1.333 1.250 1.200 1.167 1.143 1.125 1.056 Selected 2.000 1.500 1.333 1.250 1.200 1.167 1.143 1.125 1.056 1.053 Cumulative LDF 10.000 5.000 3.333 2.500 2.000 1.667 1.429 1.250 1.111 1.053 % of Ultimate 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% 90.0% 95.0% 3 2

Background/Loss Reserving 101 Incurred Accident Expected Incurred Percent of Expected Loss Dev B F Year Premium Loss Ratio Losses Ultimate Losses Method Method (1) (2) (3) (4) (5) (6) (7) (8) 2001 200 50% 95 95.0% 100 100 100 2002 200 50% 90 90.0% 100 100 100 2003 200 50% 80 80.0% 100 100 100 2004 200 50% 70 70.0% 100 100 100 2005 200 50% 60 60.0% 100 100 100 2006 200 50% 50 50.0% 100 100 100 2007 200 50% 40 40.0% 100 100 100 2008 200 50% 30 30.0% 100 100 100 2009 200 50% 20 20.0% 100 100 100 2010 200 50% 10 10.0% 100 100 100 Total 2,000 545 1,000 1,000 1,000 Col (6) = Col (2) x Col (3) Col (7) = Col (4) / Col (5) Col (8) = Col (4) + Col (6) x { 1 Col (5) } 4 Claims Handling Procedures Change in procedures How initial case reserves are set up Relative adequacy of case reserves Emphasis on speed of closing claims Desire to fight claims Staffing, number of claims per adjuster Different Third Party Administrator (TPA) 5 3

Claims Handling Procedures Change in reporting of claims Changes in how the accident date or report date is assigned Changes in the process of how incidents are labeled claims Underlying claims trends Changes in industry wide trends for claim frequency or severity Changes in trends specific to a coverage or region (FL PIP, CA WC) 6 Claims Handling Procedures Example Incurred Accident Expected Incurred Percent of Expected Loss Dev B F Year Premium Loss Ratio Losses Ultimate Losses Method Method (1) (2) (3) (4) (5) (6) (7) (8) 2001 200 50% 95 95.0% 100 100 100 2002 200 50% 90 90.0% 100 100 100 2003 200 50% 80 80.0% 100 100 100 2004 200 50% 70 70.0% 100 100 100 2005 200 50% 60 60.0% 100 100 100 2006 200 50% 50 50.0% 100 100 100 2007 200 50% 40 40.0% 100 100 100 2008 200 50% 30 30.0% 100 100 100 2009 200 50% 50 20.0% 100 250 130 2010 200 50% 40 10.0% 100 400 130 Total 2,000 605 1,000 1,450 1,060 7 4

Berquist Sherman Method Changes in case reserve adequacy distort methods based on reported incurred losses Methods based on paid losses ignore case reserves Berquist Sherman adjusts average case reserves to remove distortions Thus, valuable information in case reserves not lost 8 Reinsurance Changes in program structure Excess of loss versus quota share Change in retention level Commutations Changes in historical retentions Change in reinsurance premiums Impact on large claims or potential exposure above reinsurance layers Handling of expenses 9 5

Reinsurance Specific contract details Quota share percentage Retention Risk limiting features Loss ratio caps Loss corridors Swing rated premiums or commissions Risk transfer concerns Reinsurer ability to pay claims 10 1,200,000 Cumulative Gross Reported Loss for Claim #123 800,000 750,000 600,000 450,000 400,000 300,000 200,000 100,000 12 24 36 48 60 Months of Maturity 11 6

1,200,000 Cumulative Gross Reported Loss for Claim #123 Claim Entered Litigation 800,000 750,000 600,000 450,000 400,000 300,000 200,000 100,000 12 24 36 48 60 Months of Maturity 12 1,200,000 Cumulative Gross Reported Loss for Claim #123 Claim Entered Litigation Claim Settled 800,000 750,000 600,000 400,000 300,000 450,000 200,000 100,000 12 24 36 48 60 Months of Maturity 13 7

1,200,000 Cumulative Gross Reported Loss for Claim #123 Claim Entered Litigation Claim Settled 800,000 750,000 600,000 400,000 500,000 300,000 450,000 200,000 100,000 12 24 36 48 60 Months of Maturity 14 1,200,000 Cumulative Gross Reported Loss for Claim #123 Claim Entered Litigation Claim Settled 800,000 750,000 600,000 400,000 200,000 500,000 250,000 100,000 300,000 450,000 12 24 36 48 60 Months of Maturity 15 8

AGE TO AGE FACTORS 12 24 36 48 Losses 24 36 48 60 Unlimited 3.00 1.50 2.22 0.75 Ltd to $500K 3.00 1.50 1.11 1.00 Ltd to $250K 2.50 1.00 1.00 1.00 As retentions increase, development patterns can project more future development and be more volatile Expected loss costs increase as retentions increase 16 Incurred Accident Expected Incurred Percent of Expected Loss Dev B F Year Exposure Loss Cost Losses Ultimate Losses Method Method (1) (2) (3) (4) (5) (6) (7) (8) 2001 200 0.50 95 95.0% 100 100 100 2002 200 0.50 90 90.0% 100 100 100 2003 200 0.50 80 80.0% 100 100 100 2004 200 0.50 70 70.0% 100 100 100 2005 200 0.50 60 60.0% 100 100 100 2006 200 0.50 50 50.0% 100 100 100 2007 200 0.50 40 40.0% 100 100 100 2008 200 0.50 30 30.0% 100 100 100 2009 200 0.70 20 20.0% 140 100 132 2010 200 0.90 10 10.0% 180 100 172 Total 2,000 545 1,120 1,000 1,104 17 9

Common example Net analysis Retention changes over program history How to limit losses in development triangles? 18 Most accurate separate triangles for each retention Most conservative use highest historical retention Blended use historical retentions by period Most responsive use current retention Recent periods expected to experience most development Availability may be limited historical data Augment with benchmarks Approach should be documented Note: Currently valued loss runs should be available Limit losses to actual historical retentions by period Apply selected patterns 19 10

Rates and Underwriting Rates Rate level increases/decreases Changes in schedule rating Concern with adverse selection Underwriting Changes in what s acceptable to write Change in focus on profit or growth New markets/rates Changes in mix of business Geographical mix Limits or deductibles written 20 Rates Example Incurred Accident Expected Incurred Percent of Expected Loss Dev B F Year Premium Loss Ratio Losses Ultimate Losses Method Method (1) (2) (3) (4) (5) (6) (7) (8) 2001 200 50% 95 95.0% 100 100 100 2002 200 50% 90 90.0% 100 100 100 2003 200 50% 80 80.0% 100 100 100 2004 200 50% 70 70.0% 100 100 100 2005 200 50% 60 60.0% 100 100 100 2006 200 50% 50 50.0% 100 100 100 2007 200 50% 40 40.0% 100 100 100 2008 200 50% 30 30.0% 100 100 100 2009 300 50% 20 20.0% 150 100 140 2010 450 50% 10 10.0% 225 100 213 Total 2,350 545 1,175 1,000 1,153 Col (6) = Col (2) x Col (3) Col (7) = Col (4) / Col (5) Col (8) = Col (4) + Col (6) x { 1 Col (5) } 21 11

Acquisition Example New block of business acquired Segmentation of current analysis (e.g. new state added to a bystate analysis) Difference in historical patterns Differences systematic or not Relative size Differences in claims handling and underwriting Persistence of differences after acquisition Distinct from organic growth (one time vs gradual shift) 22 Growth Example Growing book of business Compare long term & short term averages of age to age factors Be careful about overreacting Volume weighted average Shrinking book of business Recent age to age factors may show different trend, more volatility Volume weighted average Stability vs responsiveness Growth/shrinkage in a certain subdivision 23 12

External Influences External Influences Law changes FL PIP FL sinkhole SCHIP ACA Tort reforms Inflation or other economic factors New competitors New claims sources FL public adjusters 24 Other Key Information Items DDR Death, Disability, and Retirement Unearned premium and unearned premium reserve required Extra Contractual Obligations (ECO) claims Need to understand how they are included in the data and reserved for Large claims/specific claim detail Subrogation and salvage Need to understand how they are included in the data Treatment of Expenses ALAE/ULAE vs DCC/A&O 25 13

How Actuaries Handle Trust But Verify Verification Loss data Claim count and average severity triangles Diagnostics Comparison exhibits Company expectations Time Tools Alternative methods Berquist Sherman Judgment Non traditional methods Wait Documentation 26 Questions 27 14

Thank you for your time and attention Derek Freihaut dfreihaut@pinnacleactuaries.com 309 807 2313 Commitment Beyond Numbers 28 15