Loss Experiences – April, 2011, Tornadoes (PCS Cat # 46)

[LOGO] ROBERT BENTLEY GOVERNOR JIM L. RIDLING COMMISSIONER STATE OF ALABAMA DEPARTMENT OF INSURANCE 201 MONROE STREET, SUITE 502 POST OFFICE BOX 303351 MONTGOMERY, ALABAMA 36130-3351 TELEPHONE: (334) 269-3550 FACSIMILE: (334) 241-4192 INTERNET: www.aldoi.gov DEPUTY COMMISSIONER CHARLES M. ANGELL (acting) CHIEF OF STAFF RAGAN INGRAM CHIEF EXAMINER RICHARD L. FORD STATE FIRE MARSHAL EDWARD S. PAULK GENERAL COUNSEL REYN NORMAN # BULLETIN 2012-02 TO: All insurers authorized to write property and casualty insurance FROM: Jim L. Ridling Commissioner of Insurance JLR DATE: January 12, 2012 RE: Loss experiences – April, 2011, tornados (PCS Cat # 46) EFFECTIVE: Immediate This Bulletin addresses the manner in which losses experienced as a result of the tornado outbreak in Alabama in April, 2011 (PCS Cat #46), must be handled in rate filings submitted to this office. 1. In relevant part, a rating system shall "[g]ive consideration to past experience within the state . . . over such period of years as appears to be fairly representative of the frequency of the particular hazard or peril, including, where pertinent, the . . . catastrophe hazards, if any . . ." Ala. Code § 27-13-27(3). 2. Analysis by this office of the April, 2011, tornado events in Alabama indicates these were a 1-in-250 year event, that is, there is a 0.4% probability that this series of tornados and their accompanying damage would occur in any given year. 3. All Property filings: (a) Common ratemaking practice permits insurers to use a 20-30 year experience period for establishing a rate load for non-modeled wind losses. Thus, any single year's tornado losses are assigned a probability of recurring of 3.3% to 5%, well in excess of the appropriate probability for the April, 2011, tornados. (b) All future rate filings that incorporate the April, 2011, tornado losses in their supporting data must therefore make appropriate adjustments so as to ensure that the April, 2011, tornado losses do not have assigned to them a greater weighting or probability than 0.4%. For example, if a 20-year experience period is used to estimate a non-modeled wind loss load, and if 2011 is included in that experience period, then only 8% of the April, 2011, tornado losses should be included in that 20-year loss history. Page 1 of 2 Bulletin No. 2012-02 January 12, 2012 (c) April, 2011, tornado losses should not be included in data utilized to determine frequency, severity, or pure premium trend factors. (d) Alabama's non-hurricane wind rate loads should be tiered high to low going from the northern part of the state to the southern part to reflect the expectation of higher non-hurricane wind losses per amount of insurance in the northern part of the state versus the southern part. This is supported by non-hurricane wind data by territory groupings for major insurers in Alabama, as well as weather data with regard to tornado and other wind events. (e) All Homeowners filings proposing territorial base rate changes must provide: (1) Indicated and proposed rate level changes by territory (2) Hurricane, Non-Hurricane Wind, and All-Other-Perils rate loads, allocated to territory, used to develop the territorial indications. # 4. All Auto filings: (a) Common ratemaking practice utilizes 3 years of experience for establishing comprehensive rates. Thus, any single year's tornado losses are assigned a probability of recurring of 20% to 50%, depending on whether 2011 is the newest or oldest year in the experience period, well in excess of the appropriate probability for the April, 2011 tornados. (b) All future rate filings that incorporate the April, 2011, tornado losses in their supporting data must therefore make appropriate adjustments so as to ensure that the April, 2011, tornado losses do not have assigned to them a greater weighting or probability than 0.4%. For example, if a 3-year experience period is used to project a comprehensive loss ratio or pure premium, and if 2011 is included in that experience period, then only 0.8% to 2.0% of the April, 2011, tornado losses should be included in that 3-year loss history, depending on the weight given to 2011. (c) April, 2011, tornado losses should not be included in data utilized to determine frequency, severity, or pure premium trend factors. 5. Any filings not adhering to these guidelines will be disapproved by this office. JLR/JFM/bc Page 2 of 2