RAA 2019: INSIGHTS GAINED FROM HURRICANE IRMA CLAIMS
AGENDA IDENTIFYING CLAIMS DATA VALUE FOR BUSINESS PURPOSES Overview of 2017 Catastrophes and Hurricane Irma Contribution Context of major US-landfalling hurricane (Re)insurance industry impacts JLT Re Clients HU Irma Claims Analyses High-level summary and insightful tools used Defining investigation paths: claims frequency, severity, and loss development Claims CAT Insight Dashboard within JLT Re s CATography Summarizing High-Level Findings / Trends Lessons learned about claims frequency and severity LAE development over time and proportion of total loss What are we not understanding yet about HU Irma and major hurricanes? Recommendations for additional analyses for 2018 HU Michael and future storms RAA 2019: JLT Re Insights Gained from Hurricane Irma Claims 2
GLOBAL INSURED CATASTROPHE LOSSES 1970 TO 2018 Top 5 Insured Loss Years from Catastrophes 1. 2017 2. 2011 3. 2005 4. 2012/8 #Events Includes NFIP insured losses Source: JLT Re, Swiss Re RAA 2019: JLT Re Insights Gained from Hurricane Irma Claims 3
HURRICANE IRMA EVENT OVERVIEW Hurricane Irma first made US landfall as a Category 4 Hurricane at Cudjoe Key carrying 130 mph wind speeds, and its central pressure was 929 mb after striking multiple Caribbean islands at Category 5 strength JLT Re CATography Claims Irma s track was fortunate for mainland US landfall Graphic at right shows proximity of hurricane force winds (orange shape) to highly-populated areas of Miami-Dade and Broward counties Hurricane force winds degraded to tropical storm force just before Orange County Final industry losses include these factors: Loss amplification due to demand surge, claims inflation, LAE inflation, etc. Business Interruption or Time Element losses Storm Surge and Flood Losses RAA 2019: JLT Re Insights Gained from Hurricane Irma Claims 4
CATEGORY 4 HURRICANE IRMA HISTORICAL CARIBBEAN HURRICANE Hurricane Irma first made US landfall as a Category 4 hurricane at Cudjoe Key packing 130 mph winds and central pressure of 929 mb after striking multiple Caribbean islands at Category 5 strength Date Company Loss Estimate Notes Sept. 11 th Sept. 11 th Sept. 18 th Sept. 18 th Sept. 20 th Oct 12 th FBR Capital Enki Research AIR Worldwide CoreLogic RMS Georgia Ins. Dept. $10bn to $30bn $19bn $32bn- $50bn $13.5- $19bn wind $9bn-$16bn flood $32.5bn - $49.5bn $550mn Nov. 2018 Florida OIR $11.08bn Insured losses in Florida and Caribbean Insured loss of $19bn with economic loss greater than $49bn $25bn-35bn from US; $7bn-15bn from Caribbean, does not include NFIP, economic loss not provided Flood: insured residential flood $5-8bn, insured commercial flood $4-8bn, $20-30bn non insured residential flood Economic loss: $42.5bn to $65bn; $22.5bn - $30bn US, $10bn - $20bn Caribbean; excludes NFIP estimate; economic loss range $60bn - $95bn 103,000 claims Encompasses 1M claims of which 76k are still open (7.6% total count) RAA 2019: JLT Re Insights Gained from Hurricane Irma Claims 5
STABILIZING LOSS ESTIMATES 2017 HIM LONGER TO DEVELOP, MAGNITUDE OF CHANGE LOWER 2017 HIM Storm Industry Losses vs. Previous years Losses are taking longer to develop Magnitude of change from initial storm estimate to subsequent losses is not as volatile as the other major storms in 2012, 2011, and 2008 2005 Katrina s loss developed over a similar timeframe as is being observed for 2017 storms 1-year post-landfall Source: JLT Re, PCS RAA 2019: JLT Re Insights Gained from Hurricane Irma Claims 6
CONTINUED DEVELOPMENT OF IRMA CLAIMS FLORIDA OIR LOSS UPDATES SINCE LANDFALL JLT Re has been tracking the claims information reported to FL OIR More than 400 days from landfall and Irma losses are still increasing Despite HU Michael s different path through the panhandle, it is also showing similar claims development as Irma the year before $6.55Bn, Dec 15, 2017 $5.02Bn, Jan 11, 2019 $11.08Bn, Nov 14, 2018 Source: JLT Re, Florida OIR RAA 2019: JLT Re Insights Gained from Hurricane Irma Claims 7
EXAMINING CLIENT CLAIMS DATA JLT RE HU IRMA CLAIMS ANALYSES Client Claims Analyses for Portfolios Impacted by Hurricane Irma Multiple loss bordereaux / clients (use RMS, AIR, or mix of both) Claims information as-of 31-Dec, 31- Mar, or 30-Jun Exposure at time of landfall or within 1 month provided Mix of personal and commercial LOBs Approximately $400M in incurred loss examined across all claims (another 10-25% in loss reserves) Irma Hazard Courtesy of RMS Event Response Three metrics we examined: Claims Hit Rate = # Claims / # Exposed Risks Damage Ratio = Incurred Loss / Total Exposed TIV Losses Paid How often is there an indemnity payout? Contribution of indemnity to total incurred loss? relative to geography, hazard (wind speed), and exposure characteristics RAA 2019: JLT Re Insights Gained from Hurricane Irma Claims 8
LEVERAGING VARIOUS DATA TOOLS UNCOVERING INSIGHT FOR VARIOUS BUSINESS PURPOSES Catastrophe Modeling Use IRMA experience to investigate cat modeling (+blending) fit-for-purpose on individual portfolios Refine cat modeling assumptions and data inputs Review LAE loading of EP curves JLT Re CATography Claims Portfolio Management Review impact of data quality on future loss assessments Identify exposure accumulations or other pain points for review Future Event Response Use claims experience to develop damage factors for future events Review regional claiming behaviors High-level review of claims adjusters UW or Point-of-Sale Metrics Inform UW guidelines on certain exposures Review individual brokers policies vs. claiming behaviors observed RAA 2019: JLT Re Insights Gained from Hurricane Irma Claims 9
DEMO: CLAIMS CAT INSIGHT DASHBOARD Top Three Insights Varying claiming frequency behavior in Tri-County region Impact of Hurricane Deductibles Proportion of LAE of total incurred loss higher than in past storms
INSIGHT #1 HIGH HAZARD COUNTIES VS. TRI-COUNTY REGION All Personal Lines Claims Collier Lee Palm Beach Broward Miami-Dade Lee Miami-Dade Broward Collier Palm Beach RAA 2019: JLT Re Insights Gained from Hurricane Irma Claims 11
INSIGHT #2 HU DEDUCTIBLES ARE WORKING! 22% Expense Ratio 19% Expense Ratio RAA 2019: JLT Re Insights Gained from Hurricane Irma Claims 12
INSIGHT #3 PROPORTION OF LAE IN TOP 5 LOSS COUNTIES RAA 2019: JLT Re Insights Gained from Hurricane Irma Claims 13
LACK OF SURPRISES DEVIATIONS OF DAMAGEABILITY EXPECTATIONS NOT PREVALENT Florida has the longest history of claims information for the insurance industry vs. any other U.S. state; luckily also the most detailed data BUT no two cat events are the same, so we continue to analyze and track LAE Comparisons to cat modeling assumptions were somewhat inline with expectations; deviation categories grouped into these themes Sample size prevented deep-dives into certain groups of claims, like MFD or renters policies Examples of 1995-2001 construction showed more damageability than pre-1995 (limited counties) County outliers per bordereaux for claims hit ratios were unique (one client: Orange County, another: Pinellas) Coastal Location Small Floor Area Pre-1995 Construction Manufactured Homes Old Roof Flat Roof RAA 2019: JLT Re Insights Gained from Hurricane Irma Claims 14
TRI-COUNTY AREA TRENDS CHALLENGES TO MONITORING LAE AND TOTAL INCURRED LOSS Miami-Dade and Broward Counties consistently at top of list for highest number of claims despite relatively low wind speeds Loss reserves were high for Tri-County area 6+ months following landfall Closed claims without indemnity payment for Miami-Dade were 20+% for some clients Proportion of County Claims Count by WS Band JLT Re CATography Cat Insights RAA 2019: JLT Re Insights Gained from Hurricane Irma Claims 15
HU DEDUCTIBLE IMPACTS GEOGRAPHICAL VARIATIONS FROM WHAT PREVIOUSLY OBSERVED Claiming Behavior & Deductible Influence? Standard 2% wind deductibles for homeowners policies seems to assisted in closing approximately 20% of claims without indemnity payments. Trend seemed to be independent of hazard intensity, thus damage severity. One hypothesis: relationship to mandatory evacuation orders? The Over-Evacuation Problem Almost 7 million Florida residents evacuated before HU Irma landfall, largest US HU evacuation in history Most coastal counties ordered evacuations for areas susceptible to storm surge, but many left their homes because of wind fears (i.e., hearing Category 4 or 5 on the news) Palm Beach Co s director of emergency management reported ~50% of those who evacuated were from areas where evacuations had not been ordered Source: NPR JLT Re CAT Insights - Personal Lines for Palm Beach County RAA 2019: JLT Re Insights Gained from Hurricane Irma Claims 16
GROWING LAE, AOB IMPACTS FUTURE INDICATORS OF NEW CAT MODELING ADJUSTMENTS? Proportion of LAE in Over First 6 months Assessed 3 to 6 months post-landfall LAE was ~18 to 25% for personal lines of loss bordereaux examined Some insurers in 2018 were expecting ultimately to exceed 25% LAE in the Tri-Counties JLT Re CAT Insights Personal Lines for Tri-County Tracking LAE over time for the industry and individual clients Discussions around loading cat modeling results for LAE with an amount, comparing to Irma experience Whether to use one storm s experience to change any shifts in the OEP curves for use in analyzing reinsurance programs? RAA 2019: JLT Re Insights Gained from Hurricane Irma Claims 17
MODEL-TO-INCURRED RATIOS (M2I) MODELING GROSS LOSS COMPARISONS FOR TOP COUNTIES M2I Ratios used to compare gross model loss (M) to claims indemnity loss (I) A result of 1.0 shows model prediction inline with claims experience Traditional cat models do not include expenses explicitly within Sample M2I plot below is composed of modified results to protect the innocent Model Apples Model Oranges Model Overestimation or M2I > 1.0 Pinellas Lee Collier Monroe Event M2I (all counties) Miami-Dade Broward Brevard Polk Orange Volusia Palm Beach M2I = 1.0 or model matches incurred loss Model Underestimation or M2I < 1.0 RAA 2019: JLT Re Insights Gained from Hurricane Irma Claims 18
BRINGING IT ALL TOGETHER MAJOR INSURANCE PERSPECTIVES & LESSONS LEARNED Primary Insurance Companies with or without Florida exposures Lessons learned transferable to other catastrophes and geographic regions Cat model validation and blending can be supported New understanding of LAE is evolving in Florida; are there other states? Reinsurers Covering HU Irma Claims Different claiming behaviors between personal and commercial lines Scaling of EP curves per portfolio could be adjusted considering 2017 HIM Reinsurance Brokers Supporting Both Sides Opportunity to leverage new technologies for deeper analysis Supporting reinsurance submissions Assisting insurance companies in understanding their experience against the market; where are the deviations from average? RAA 2019: JLT Re Insights Gained from Hurricane Irma Claims 19
? S THANK YOU & DISCUSSION
CONTACTS Taronne Tabucchi VP & Senior Risk Engineer E: taronne.tabucchi@jltre.com Chris Zumbrum Catastrophe Risk Management Services Officer E: chris.zumbrum@jltre.com This presentation by JLT Re (North America) Inc. is intended to provide only general information based on sources we believe are reliable. JLT Re (North America) Inc. makes no representations or warranties, express or implied, as to the accuracy of any of the information herein, which is not intended to be taken as advice with respect to any individual situation and cannot be relied upon as such. Any statements concerning tax, accounting, legal or regulatory matters should be understood to be general observations based solely on our experience as reinsurance brokers and risk consultants. We are not tax, accounting, legal or regulatory professionals and any such information provided is not professional advice. These matters should be reviewed with your own qualified advisors in these areas. This document may not be copied or reproduced in any form without the express permission of JLT Re (North America) Inc., except that clients of JLT Re (North America) Inc. need not obtain such permission when using this report for internal business purposes. RAA 2019: JLT Re Insights Gained from Hurricane Irma Claims 21