QUARTERLY VALUATION HIGHLIGHTS RISK SHARING POOLS. as at September 30, Ontario Alberta Grid and Alberta Non Grid New Brunswick and Nova Scotia

Transcription

1 QUARTERLY VALUATION HIGHLIGHTS RISK SHARING POOLS as at September 30, 2018 Ontario Alberta Grid and Alberta NonGrid New Brunswick and Nova Scotia FA Actuarial 1/16/2019 Should you require any further information, please call: Shawn Doherty, FCIA, FCAS SVP Actuarial & CFO (416)

2 Contents A. Executive Summary... 4 B. General Information... 8 B.1 Appointed Actuary and Hybrid Actuarial Services Model... 8 B.2 Intended Audience and Use... 9 B.3 Data... 9 B.3.1 RSP Valuation Data... 9 B.3.2 Industry AIX Data... 9 B.3.3 Other Data B.4 Actual vs (AvsP) B.5 Uncertainty C. ONTARIO RSP C.1 Valuation Highlights C.1.1 Ontario Product Reforms C.2 Booked results for the prior valuation implementation C.3 Booked results for the current valuation implementation C.4 a priori loss ratios C.5 Actual vs (AvsP) C.5.1 AvsP: Indemnity & Allowed Claims Expense C.5.2 AvsP: Paid Indemnity & Allowed Claims Expense C.6 Current valuation IBNR selections C.7 Premium Liabilities / Future Accident Years C.8 Actuarial Present Value Adjustments C.8.1 Selected Claims Payment Patterns C.8.2 Selected Discount Rate C.8.3 Selected Margins for Adverse Deviations C.9 Special IBNR Provisions / Adjustments D. ALBERTA GRID RSP D.1 Valuation Highlights D.2 Booked results for the prior valuation implementation D.3 Booked results for the current valuation implementation D.4 a priori loss ratios D.5 Actual vs (AvsP) D.5.1 AvsP: Indemnity & Allowed Claims Expense D.5.2 AvsP: Paid Indemnity & Allowed Claims Expense D.6 Current valuation IBNR selections D.7 Premium Liabilities / Future Accident Years page 1 of 74

3 D.8 Actuarial Present Value Adjustments D.8.1 Selected Claims Payment Patterns D.8.2 Selected Discount Rate D.8.3 Selected Margins for Adverse Deviations D.9 Special IBNR Provisions / Adjustments E. ALBERTA NON-GRID RSP E.1 Valuation Highlights E.2 Booked results for the prior valuation implementation E.3 Booked results for the current valuation implementation E.4 a priori loss ratios E.5 Actual vs (AvsP) E.5.1 AvsP: Indemnity & Allowed Claims Expense E.5.2 AvsP: Paid Indemnity & Allowed Claims Expense E.6 Current valuation IBNR selections E.7 Premium Liabilities / Future Accident Years E.8 Actuarial Present Value Adjustments E.8.1 Selected Claims Payment Patterns E.8.2 Selected Discount Rate E.8.3 Selected Margins for Adverse Deviations E.9 Special IBNR Provisions / Adjustments F. NEW BRUNSWICK RSP F.1 Valuation Highlights F.2 Booked results for the prior valuation implementation F.3 Booked results for the current valuation implementation F.4 a priori loss ratios F.5 Actual vs (AvsP) F.5.1 AvsP: Indemnity & Allowed Claims Expense F.5.2 AvsP: Paid Indemnity & Allowed Claims Expense F.6 Current valuation IBNR selections F.7 Premium Liabilities / Future Accident Years F.8 Actuarial Present Value Adjustments F.8.1 Selected Claims Payment Patterns F.8.2 Selected Discount Rate F.8.3 Selected Margins for Adverse Deviations F.9 Special IBNR Provisions / Adjustments G. NOVA SCOTIA RSP G.1 Valuation Highlights G.2 Booked results for the prior valuation implementation page 2 of 74

4 G.3 Booked results for the current valuation implementation G.4 a priori loss ratios G.5 Actual vs (AvsP) G.5.1 AvsP: Indemnity & Allowed Claims Expense G.5.2 AvsP: Paid Indemnity & Allowed Claims Expense G.6 Current valuation IBNR selections G.7 Premium Liabilities / Future Accident Years G.8 Actuarial Present Value Adjustments G.8.1 Selected Claims Payment Patterns G.8.2 Selected Discount Rate G.8.3 Selected Margins for Adverse Deviations G.9 Special IBNR Provisions / Adjustments H. Appendix 1: Changes in process introduced since the September 30, 2017 valuation I. Appendix 2: Recent Regulatory and/or Legislative Initiatives I.1 Ontario I.2 Alberta I.3 Harmonized Sales Tax I.4 Harmonized Sales Tax Class Action - Ontario J. Appendix 3: Court Decisions J.1 Supreme Court of Canada K. Appendix 4: General description of the RSP valuation process L. Appendix 5: Exhibits page 3 of 74

5 A. Executive Summary We have completed a valuation for all (Ontario, Alberta Grid, Alberta Non-Grid, New Brunswick, and Nova Scotia) Risk Sharing Pools ( RSPs ) as at September 30, 2018, with the results summarized in the table below 1. The previous valuation was completed at June 30, 2018 and included all RSPs. Valuation Summary (Nominal Basis) Risk Sharing Pool 2017 & Prior Beginning Indemnity Unpaid (000s) 2017 & Prior Accident Year Indemnity Change (000s) % of Beginning Unpaid 2018 Indemnity Loss Ratio Change from Prior Valuation Change against 2018 Earned Premium (000s) 2019 Indemnity Loss Ratio Change from Prior Valuation Change against 2019 Earned Premium (000s) unfavourable / (favourable) In total, the favourable prior accident year change of $10.7 million (column [2] in the table above) represents 0.9% (column [3]) of the $1,223.9 million beginning unpaid (column [1]) 2. As a provision for incorrectly reported member claims transactions was included with prior valuations (originally implemented with the December 31, 2017 valuation and further discussed below), the above summary (columns [1] and [2]) excludes the impact of claims correction transactions posted in the current quarter which did not impact the current valuation implementation. Throughout 2016 and 2017, a member has been conducting a comprehensive review of certain claims transactions as reported to the FA RSP system. As discussed in the December 2017 Quarterly Actuarial Highlights, a balance sheet adjustment was applied to selected nominal implementation loss ratios for the Ontario, Alberta non-grid, New Brunswick and Nova Scotia RSPs to offset an incorrectly reported member case reserve overstatement, with the resulting impact implemented with the March 2018 Operational Report (based on the December 31, 2017 valuation). Prior to the current valuation (as at September 30, 2018), the member posted case reserve corrections impacting accident year 2010 and onwards only. The submitted corrections were reviewed by FA management and determined to be reasonably aligned with the previously held provisions, which have been released resulting in a neutral implementation impact. As of the September 30, 2018 valuation, there remain case reserve corrections for the member related to Ontario RSP accident years , but this constitutes a relatively small level of overall case reserve (although FA management has opted to continue holding the current provision for those accident years until the correction transactions have been submitted 3 ). Selected Discount Rate at Sep/18 Change in Discount Rate from Prior Valuation Estimated $ Effect from sensitivity analysis (000s) [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] Ontario 771,742 (2,723) (0.4%) 127.0% 0.5% 1, % 0.1% % +39 bps (11,783) Alberta Grid 260,279 (4,336) (1.7%) 89.8% (2.0%) (3,249) 88.8% (1.7%) (2,699) 2.28% +41 bps (3,682) Alberta nongrid 144,367 (4,337) (3.0%) 109.3% (2.8%) (2,990) 107.5% (1.8%) (2,262) 2.29% +42 bps (2,183) New Brunswick 15,871 (324) (2.0%) 74.7% 0.5% % (0.1%) (17) 2.29% +42 bps (238) Nova Scotia 31,657 1, % 92.9% (0.7%) (194) 96.3% (0.3%) (91) 2.28% +42 bps (497) Total 1,223,916 (10,712) (0.9%) (4,586) (4,677) (18,383) "Unpaid", "Claims" and "Loss" include indemnity & allowed claims expense 1 The September 30, 2018 valuation result was implemented into the RSP Operational Results for the month of October The valuation implementation impact is discussed in the respective October 2018 Actuarial Highlights. 2 The beginning unpaid is the sum of the case reserves and selected nominal IBNR as per the valuation completed as at June 30, 2018, adjusted for a member s case reserve mis-statement. 3 The remaining case reserve correction transactions were submitted by the member in October 2018, reviewed by FA management and included alongside the current valuation implementation with the RSP Operational Results for the month of October The submitted corrections were aligned with the valuation provisions, which will result in a neutral implementation impact. page 4 of 74

6 The impact of roll-forward valuation 4 updates tend to be less material since the impact of actual emerged experience from the last full valuation has not been incorporated into revised assumptions. Changes in selected loss ratios for accident year 2018 (column [4] in the table on the prior page) were driven by low levels of reported comprehensive claims activity in the Alberta RSPs. The impact of these changes, relative to projected full year 2018 earned premium, is favourable by $4.6 million (column [6]). Changes in selected loss ratios for accident year 2019 (column [7]) were driven by updated a priori loss ratio selections as a result of updated written premium projection assumptions to reflect the Alberta grid rate mandatory coverages increase effective January 1, 2019 (AIRB Bulletin ). The impact of these changes has an anticipated full year 2019 favourable impact of $4.7 million (column [9]) in relation to the current projected complete AY2019 earned premium level. These AY2018 and AY2019 changes also collectively imply an immediate favourable impact in relation to policy liabilities with the valuation s implementation. As indicated in columns [10] and [11] in the table on the prior page, claims payment emergence patterns were updated and discount rates were increased, reflecting September 2018 Government of Canada yields, generating an initial estimated $18.4 million favourable impact due to the discount rate selection change (column [12]). The claims development margins for adverse deviations (MfADs) for all RSPs were reviewed with the prior valuation (as at June 30, 2018) as per our regular annual valuation process. In particular, selected claims development MfADs for older accident years (specifically related to accident years where claims development MfADs were previously increased due to periods of increased uncertainty) were reviewed and judgmentally reduced to reflect the decreasing uncertainty over time, resulting in an estimated $32.1 million favourable impact across all RSPS, implemented in the August 2018 Operational Reports. With the current valuation (as at September 30, 2018), the Ontario RSP (third party liability bodily injury and accident benefits) and Alberta Grid and Non-Grid RSP (third party liability bodily injury) claims development MfADs were further reviewed based on discussion and feedback with the FA s Actuarial Committee. The Ontario selected claims development MfADs were updated resulting in an estimated overall unfavourable impact of $13.9 million implemented in the October 2018 Ontario Operational Report, partially offsetting the favourable MfADs update impact of $30.4 million initially implemented in the August 2018 Operational Report (based on the June 30, 2018 valuation). No changes were made to the Alberta selected claims development MfADs from the prior valuation review. The selected investment income (25 basis points) margin for adverse deviation (MfAD) was not changed with the current valuation. Excluding the impact of the member corrections discussed earlier (which had a neutral valuation implementation impact), the Ontario RSP favourable prior accident year development 5 would have been $2.7 million, 0.4% of beginning unpaid (as determined with the June 30, 2018 valuation), and was driven by low levels of third party liability bodily injury recorded claims in recent years partially offset by older accident year large loss (AY2011 bodily injury and AY2014 accident benefits) unfavourable recorded claims activity in the quarter. Including the impact of the member corrections reported in the 4 Under the proposed schedule for fiscal year 2018, the off-half valuation quarters ending March 31, 2018 and September 30, 2018 would not reflect a full valuation update of assumptions, but would rather roll-forward key assumptions from the previous valuation. Loss development factors as brought forward through this process are interpolated assuming linear emergence. 5 The term development throughout this document refers to claims activity during the period, and favourable or unfavourable development is in relation to projections or underlying assumptions per the previous valuation. page 5 of 74

7 quarter, the Ontario RSP favourable prior accident year development would have been $9.2 million, 1.2% of the beginning unpaid. The table immediately below shows historical changes in valuation selected ultimates on an annual fiscal-accident year basis on the left with changes in the most recent quarterly valuations on a calendar-accident year basis 6 on the right. We have observed reductions in the overall magnitude of change despite having significant favourable experience in each of the previous 6 prior fiscal year-ends. Ontario RSP Changes in Prior Accident Year Selected Ultimates through time 7 Change in Selected Ultimates from prior Sept 30th Change in Selected Ultimates from Prior Quarter End Sep12 Sep13 Sep14 Sep15 Sep16 Sep17 Sep18 Dec17 Mar18 Jun18 Sep18 Dec18 AY2007 & Prior (18,685) 10,644 (4,111) 306 (4,620) 10,078 2,233 (43) (324) 3,099 (500) AY2008 (9,437) (92) (2,236) (1,632) (3,922) (2,670) (1,634) (473) 87 (1,059) (189) AY2009 (12,325) 8,326 5,510 (2,767) (6,439) (5,775) (2,626) (1,874) 422 (1,603) 429 AY2010 (25,024) 15,929 (7,623) (2,753) (7,409) (6,094) (7,359) (4,716) (106) (2,298) (240) AY2011 (24,649) (46,425) (37,295) (22,216) (7,733) (1,322) 1,332 (1,957) 395 1,128 1,765 AY2012 (73,806) (19,118) (43,289) (10,135) (5,257) 1,056 1,522 (97) (235) (134) AY2013 (24,834) (46,961) (2,493) (4,982) (8,739) (4,677) (2,055) (365) (1,642) AY2014 (20,591) (21,779) (17,319) 5,428 2, ,228 AY (12,028) (6,671) (5,123) 2, (3,971) AY2016 1,078 (2,602) 128 1,134 (1,670) (2,195) AY2017 (750) 2,763 (8,661) (3,789) Total (90,120) (85,424) (89,707) (139,903) (64,005) (44,291) (20,333) (14,357) 5,429 (11,106) (9,237) PAY change including AY2017 change at Dec17 not shown in change from Prior Quarter End section: (29,270) Similar summaries for the Alberta Grid and Alberta Non-Grid RSPs prior accident year development are shown immediately below and at the top of the next page with the favourable Alberta RSPs development driven by favourable third party liability bodily injury recorded claims activity (impacted by reported case reserve reductions across multiple member companies) in recent accident years (AY2014-AY2017) in contrast to the continued unfavourable prior accident year bodily injury development during prior valuations. Alberta Grid RSP Changes in Prior Accident Year Selected Ultimates through time Change in Selected Ultimates from prior Sept 30th Change in Selected Ultimates from Prior Quarter End Sep12 Sep13 Sep14 Sep15 Sep16 Sep17 Sep18 Dec17 Mar18 Jun18 Sep18 Dec18 AY2007 & Prior 4,775 5,786 1,791 (900) 458 1,151 1, AY2008 2,015 1,293 3,164 (1,893) (172) (312) (32) (148) AY2009 2,593 4,055 (270) (2,493) (440) (358) (585) (463) 180 (306) 4 AY2010 (657) 3,485 2,791 (4,147) 2, (47) 30 (62) (465) 450 AY2011 (800) 5,494 2,075 (2,387) 788 (908) 995 (504) AY2012 3,048 9,558 (3,542) 3,669 (1,464) (810) 93 (133) (332) (438) AY , ,339 (293) (1,950) (1,144) (472) 393 (727) AY ,602 9,649 (706) (1,100) (897) (629) AY ,132 5,832 1,434 2, (1,251) (784) AY ,994 1,537 3,474 (1,067) 226 (1,096) AY2017 3,291 (2,330) (3) (2,252) Total 7,925 23,161 30,121 (902) 42,559 22,617 4,326 4,199 (2,222) (1,191) (4,336) PAY change including AY2017 change at Dec17 not shown in change from Prior Quarter End section: (3,550) Excluding the impact of the member corrections discussed earlier, the Alberta Non-Grid RSP favourable prior accident year development would have been $4.3 million, 3.0% of beginning unpaid (as determined with the June 30, 2018 valuation). Including the impact of the member corrections 6 Due to FA s October 31 year-end, the runoff table is shown on a fiscal accident year basis. However, valuations are treated on a calendar accident year basis. As a result, the Change in Selected Ultimates from Prior Quarter End will not necessarily sum to the annual view for the most recent prior accident year. The valuation change discussions focus on the calendar accident basis. 7 The Changes in Prior Accident Year Selected Ultimates through time charts do not include the impact of balance sheet IBNR adjustments related to incorrect member case reserve reporting included with the current and prior valuations (introduced with the December 31, 2017). The impact of the incorrect member case reserve reporting flow through these exhibits as the claims transactions are correctly reported to the FA RSP system. page 6 of 74

8 reported in the quarter, the Alberta Non-Grid RSP favourable prior accident year development would have been $9.4 million, 6.3% of the beginning unpaid. Alberta Non-Grid RSP Changes in Prior Accident Year Selected Ultimates through time Change in Selected Ultimates from prior Sept 30th Change in Selected Ultimates from Prior Quarter End Sep12 Sep13 Sep14 Sep15 Sep16 Sep17 Sep18 Dec17 Mar18 Jun18 Sep18 Dec18 AY2007 & Prior 3,654 1,089 2, (6) AY2008 1,069 1,611 (1,170) (1,126) 229 (453) (77) (24) 34 (85) (2) AY2009 1,795 2,414 (361) (2,127) AY ,030 (2,659) (4,390) (376) (230) (452) (264) (146) (10) (32) AY2011 (7,537) (1,595) 2,299 (1,252) (1,491) 161 (52) (240) 112 (40) 117 AY2012 4,518 1,329 (1,991) 1,231 (1,255) (92) 391 (101) AY2013 4, (986) (517) (958) (5) 117 (1,182) 112 AY2014 5,967 3,532 (493) (2,451) (1,686) (132) (108) (525) AY2015 1,168 2,349 (5,638) (2,468) 47 (1,099) (2,119) AY2016 5,416 (3,873) (718) 273 (1,337) (2,091) AY2017 (388) 1,193 (2,105) (4,761) Total (582) 14,067 5,932 (4,578) 3,603 5,861 (12,348) (4,555) 1,597 (5,294) (9,381) PAY change including AY2017 change at Dec17 not shown in change from Prior Quarter End section: (17,632) Similar tables for the New Brunswick and Nova Scotia RSPs are included in their respective jurisdiction exhibits included with this report. Caution must be exercised in reviewing the variances in the New Brunswick and Nova Scotia special purpose RSPs as volumes are low and single claim transactions that are normal course for the business may look unusual and generate relatively significant variances that in nominal value terms are not that significant. The remainder of this report consists of 9 sections. Sections C through G are the detailed sections related to each of the RSPs, including valuation highlights and a discussion of actual vs. projected activity. General information about this report can be found in section B. The final 4 sections are appendices: the valuation process is described in detail in section K (Appendix 4); a summary of changes to the process during this fiscal year is provided in section H (Appendix 1); a summary of regulatory changes is provided in section I (Appendix 2) and recent applicable court decisions is provided in section J (Appendix 3); and supporting exhibits are provided in section L (Appendix 5). page 7 of 74

9 B. General Information This report summarizes the results of the valuation of the following Risk Sharing Pools ( RSPs ) as at September 30, 2018: Ontario; Alberta Grid; Alberta Non-Grid; New Brunswick; and Nova Scotia. The results of this valuation were reflected for the first time in the October 2018 Operational Reports for the above RSPs. The valuations have been prepared in accordance with Accepted Actuarial Practice and comply with the appropriate Standards of Practice of the Canadian Institute of Actuaries as well as applicable regulatory requirements. Accepted Actuarial Practice requires all policy liabilities recognize both the time value of money and provisions for adverse deviations. Unless specifically noted in this document, no explicit provision has been made for causes of loss which are not already reflected in the historical data, nor for otherwise unforeseen changes to the legal or economic environment in which claims are settled, including changes in the interpretation of existing legislation or regulation on matters currently before the courts. Automobile insurance product reforms occur from time to time and consideration is given to the associated impact, if any. Please see Section I for a discussion of recent product reforms and Section J for a discussion of recent court decisions considered for the purposes of this valuation. For ease of reference, we will use the term claims amount in reference to the more proper and descriptive term indemnity & allowed claims expense and the terms loss ratio, claims ratio, or claims amount ratio in reference to the ratio of claims amount to earned premium. General information regarding the Facility Association and on the Risk Sharing Pools in particular can be found on its website: B.1 Appointed Actuary and Hybrid Actuarial Services Model Liam McFarlane of Ernst & Young LLP is Facility Association s Appointed Actuary (effective as of June 1, 2013). Facility Association operates under a hybrid model in relation to the management and provision of actuarial services. Under this model, actuarial services are performed by both Facility Association s internal staff and its external actuarial consulting firm. The hybrid model approach maximizes the efficiency of resource allocation while providing access to additional expertise and capacity as needed. page 8 of 74

10 B.2 Intended Audience and Use This report is intended for the Member Companies of the Facility Association to provide additional information on the results of the most recent valuation of specific RSPs in relation to the results of prior such valuations. It is not intended, nor necessarily suitable, for any other purpose. B.3 Data Two primary data sets were used for the purposes of this valuation: RSP valuation data, which is aggregated premium and claim information primarily intended for valuation purposes; and industry AIX data, which is developed from detailed statistical records reported by insurers to the Insurance Bureau of Canada (IBC) 8 in accordance with the Automobile Statistical Plan. B.3.1 RSP Valuation Data Much of this analysis was based on RSP valuation data collected from Members and aggregated by IBC on behalf of Facility Association. The claims data excludes all loss adjustment expenses except certain specific reimbursed expenses ( allowed claims adjustment expenses ). The data is reconciled to information contained in Facility Association s Member Operational Reports, the results of which are reviewed by the Appointed Actuary for reasonableness. Procedures are in place to provide reasonable assurance that the data used is reliable and sufficient for the proper valuation of the liabilities. The valuation data, for the purposes of the valuation, is aggregated to the level of: RSP kind-of-loss / coverage accident year and half-year development half-year 9 Data elements captured include earned premium, claims 10 paid, case reserves, recorded claims (being the sum of claims paid and case reserves), and recorded claim counts. For the purposes of the valuation described in this report, the valuation data is as at September 30, B.3.2 Industry AIX Data Although the RSP valuation data is the primary source of data for valuation purposes, the following Industry AIX data file prepared by IBC (on behalf of GISA) is used to supplement the RSP valuation 8 IBC is the statistical agent of the General Insurance Statistical Agency (GISA), with responsibility of managing the Automobile Statistical Plan reporting. In addition, Facility Association outsources its IT to IBC. 9 Development quarter is also available for purposes of performing roll forward valuations in relation to valuation periods ending March 31 and September For purposes of this report, the terms claims or loss will refer to indemnity and allowed claims adjustment expenses unless otherwise indicated. page 9 of 74

11 data and is used in the determination of loss cost trend structures, being models describing changes in loss costs (average claim amount per exposure unit) over time, including the impacts of product reforms: industry experience (indemnity only) as per the AIX Development Exhibits for Private Passenger Vehicles in the applicable jurisdictions, compiled as at December 31, 2017 IBC (on behalf of GISA) assembles Industry AIX data from the submissions made under the Automobile Statistical Plan by each of the insurers writing automobile business in the applicable jurisdiction. As there are many insurers providing this information and due to remoteness from the individual data elements, it is not practical for IBC to directly put in place audit or audit-like procedures. However, IBC does perform various data edit checks which are designed to promote data integrity. Industry AIX data is relied upon without the benefit of any independent audit and has been used without modification. Nonetheless, the data is deemed to be reliable and appropriate for the purposes of this valuation and the trend analysis completed in relation to the data. B.3.3 Other Data Reliance has also been placed on other quantitative and qualitative information supplied by Facility Association without audit or independent verification. Wherever possible, such information was reviewed for reasonableness and internal consistency by the Appointed Actuary. B.4 Actual vs (AvsP) With each valuation, we project, by accident year, future claim activity (recorded and paid). Both projected recorded claim activity and projected paid claim activity is used as a means of providing feedback on our prior selections of ultimate. In addition, the paid projections are used directly as projected cash flows for claims in the determination of the discount rate selection for the policy liabilities. The challenge in interpreting actual versus projected (AvsP) variances as a feedback mechanism is how much of the variance is attributed to: process variance (i.e. randomness) inherent in the activities themselves (i.e. recorded and paid activity); model selection (i.e. that our emergence model is not a good representation or predictor of future emergence even if we ve correctly estimated ultimate); parameter selection within the model (i.e. that our emergence model can be a good representation of emergence, but we selected the wrong emergence factors); our selection of ultimate (i.e. that our emergence model and emergence factors selections are good, but we re applying the model and factors to the wrong ultimate); and changes to our model (i.e. changes made with the goal of improving its predictive capability). Nonetheless, the AvsP exercise is an important validation process for us. Our discussion in each RSP s AvsP section will focus on our interpretation of feedback the variances provide to our prior selections of ultimate, and how this provides information in relation to our current selections of ultimate. page 10 of 74

12 B.5 Uncertainty The establishment of provisions for the unpaid, unrecorded, and/or unreported claims is based on numerical data and the interpretation of current and anticipated circumstances. It is a complex and dynamic process influenced by a large variety of factors. These factors include the experience of the respective RSPs and the experience of the voluntary market in the associated jurisdiction, claim frequency and severity, indemnity and allowed claims expense payment patterns, case reserving practices, and lags between when the event giving rise to the claim occurred, when the claim is reported to a member company, when the member company records claim information on their own system, and when that information is transmitted to Facility Association to be recorded. The process of determining the provisions necessarily involves uncertainty such that the actual results will deviate, perhaps substantially, from the best estimates made through the valuation process. page 11 of 74

13 C. ONTARIO RSP C.1 Valuation Highlights A summary of the valuation results through time is available in the A exhibit (see section L for all exhibits), with detail related to the current valuation provided in the B.1.1 and B.1.2 exhibits. Throughout 2016 and 2017, a member has been conducting a comprehensive review of certain claims transactions as reported to the FA RSP system. As discussed in the December 2017 Quarterly Actuarial Highlights, a balance sheet adjustment was applied to selected nominal implementation loss ratios to offset an incorrectly reported member case reserve overstatement, with the resulting impact implemented with the March 2018 Operational Report (based on the December 31, 2017 valuation). Prior to the current valuation (as at September 30, 2018), the member posted case reserve corrections impacting accident year 2010 and onwards only. The submitted corrections were reviewed by FA management and determined to be reasonably aligned with the -$6.4 million associated Ontario RSP nominal member case reserve adjustment previously held, which have been released resulting in a neutral implementation impact. As of the September 30, 2018 valuation, there remain case reserve corrections for the member related to Ontario RSP accident years , but this constitutes a relatively small level of overall case reserve (although FA management has opted to continue holding the -$1.8 million nominal current provision for those accident years until the correction transactions have been submitted 11 ). The nominal valuation results and exhibits presented in this section and in Appendix 5 do not include the impact of balance sheet IBNR adjustments related to incorrect member case reserve reporting included with the current and prior valuations (introduced with the December 31, 2017), i.e. the impact of the incorrect member case reserve reporting flow through these exhibits as the claims transactions are correctly reported to the FA RSP system. Including the member corrections reported in the quarter, the change in selected ultimate for prior accident years was $9.2 million favourable with this valuation (1.2% of the unpaid estimate as at last quarter), bringing the calendar year-to-date total favourable to $14.9 million (1.7% of the unpaid estimate as at the beginning of the 2018 calendar year). These changes are presented by accident year and government line in the tables below. Ontario RSP valuation changes in selected ultimate Ontario RSP valuation changes in selected ultimate (favourable) / unfavourable during Quarter (favourable) / unfavourable YTD Accident Year Third Party Accident Other Third Party Accident Other Total Accident Year Liability Benefits Coverages Liability Benefits Coverages Total 2013 & Prior (510) 339 (338) (509) 2013 & Prior 1,352 (4,608) (265) (3,521) 2014 (1,470) 2,851 (153) 1, (660) 3,551 (319) 2, (3,589) (550) (3,970) 2015 (3,264) 2,234 (517) (1,547) 2016 (1,104) (185) (905) (2,194) 2016 (3,328) 1,420 (822) (2,730) 2017 (1,122) (687) (1,980) (3,789) 2017 (6,120) (5,010) 1,443 (9,687) TOTAL (4,037) (1,271) (3,926) (9,234) TOTAL (12,020) (2,413) (480) (14,913) Excluding the impact of the member corrections discussed earlier (which had a neutral valuation implementation impact), the Ontario RSP favourable prior accident year development would have been 11 The remaining case reserve correction transactions were submitted by the member in October 2018, reviewed by FA management and included alongside the current valuation implementation with the RSP Operational Results for the month of October The submitted corrections were aligned with the valuation provisions, which will result in a neutral implementation impact. page 12 of 74

14 $2.7 million, 0.4% of beginning unpaid (as determined with the June 30, 2018 valuation), and was driven by low levels of third party liability bodily injury recorded claims in recent years partially offset by older accident year large loss (AY2011 bodily injury and AY2014 accident benefits) unfavourable recorded claims activity in the quarter. We have observed improvement over the past few valuation quarters in the accuracy of our recorded and paid claims emergence projections (as actual versus projected calendar quarter claims variance magnitudes have narrowed considerably), however, actual third party liability bodily injury experience has continued to emerge at a level below original expectations and, as more weight is given to this actual experience at each full valuation, the impact, while moderate as a proportion of the claim liabilities, is significant in relation to the current month and fiscal year to date earned premiums and can cause large swings in these loss ratios. In addition, as expected, the reform adjustment in relation to the retroactive application of changes to the prejudgement interest rate on general damages for non-pecuniary bodily injury loss in Ontario, introduced with the September 30, 2017 valuation, was fully aged out (unwound) with the current valuation (that is, no further adjustment is carried for accident years 2014 and prior). The selected loss ratio for accident year 2018 (current accident year) increased by 0.5 points to 127.0% driven by unfavourable physical damage (comprehensive) recorded claims activity in the quarter, and the selected loss ratio for accident year 2019 (future accident year) increased by 0.1 points to 127.1%. The selected loss ratios for AY2018 and AY2019 were also affected by reforms (prejudgment interest on nonpecuniary awards, interest rates for overdue payments on statutory accident benefits while in the dispute resolution process, and statutory accident benefits product changes effective June 1, 2016, and the bodily injury non-pecuniary deductible and threshold changes) to various extents. Further discussion on Ontario Product Reform adjustments follows in section C.1.1. Summary descriptions of recent regulatory and legislative initiatives are available in section I. The valuation process is described in more detail in section K, and a summary of changes to the process during this fiscal year is provided in section H. Policy liability projected cash flows and September 2018 government of Canada bond yields were used to determine the applicable discount rate. The selected investment income margin for adverse deviation ( MfAD ) was maintained at 25 basis points with the current valuation. Our usual process is to review and update the claims development MfADs annually with the June 30 valuation. However, with the current valuation (as at September 30, 2018), the Ontario (third party liability bodily injury and accident benefits) claims development MfADs were reviewed based on discussion and feedback with the FA s Actuarial Committee. The Ontario RSP selected claims development MfADs were updated resulting in an estimated overall unfavourable impact of $13.9 million implemented in the October 2018 Ontario Operational Report. See Exhibit D in section L for selected margins. C.1.1 Ontario Product Reforms With the September 30, 2015 valuation, various reform adjustments related to changes introduced in Ontario 2014 Bill 15 (Fighting Fraud and Reducing Automobile Insurance Rates) and Ontario 2015 Bill 91 (Building Ontario Up Act (Budget Measures)) were first introduced. The various reform adjustments were reviewed during the Ontario Industry PPV trend analysis and incorporated with page 13 of 74

15 the prior valuation (as at June 30, 2018) and maintained with the current valuation (as at September 30, 2018). A summary of recent Ontario regulatory and/or legislative initiatives is included in Section I.1. In the table immediately below, we have included a summary of the reform adjustments with impact by accident year - further discussion related to selected reform adjustments follows (the following discussion was updated to reflect that the retroactive non-pecuniary PJI adjustment to Ontario RSP bodily injury unpaid amounts, impacting AY2014 and prior, which was introduced during the valuation was fully aged out and removed with the current valuation (as expected) and the review and update of the reform adjustments using Ontario Industry PPV data). Reform Ontario 2014 Bill 15 Ontario 2015 Bill 91 Product reform changes Non-Pecuniary Bodily Injury PJI SABS Interest Rate for claims in dispute SABS product change Bodily Injury Tort Threshold/Deductible Adjustment Decrease in AY2015 & subsequent bodily injury a priori loss ratios Decrease in AY2015 & subsequent accident benefits a priori loss ratios Increase in AHY2016/2 & subsequent bodily injury a priori loss ratios Decrease in AHY2016/2 & subsequent accident benefits a priori loss ratios Decrease in AY2015 & subsequent bodily injury a priori loss ratios Ontario 2014 Bill 15 included changes to the Insurance Company Act, via introduction of Section 258.3(8.1) to move the non-pecuniary 12 bodily injury award prejudgment interest calculation to follow that of pecuniary awards as per Ontario s Court of Justice Act Section (where the rate varies depending on the timing of the action 13 ) as opposed to Section (where the interest rate is set at 5% via Ontario Rules of Court). Thus, the prejudgment interest rate for non-pecuniary awards is now aligned with that of pecuniary. This change came into force on January 1, 2015, but it was not clear if this applied only to actions brought on or after January 1, 2015, or if it applied to all settlements on or after January 1, In the Ontario Court of Appeal decisions in El-Khodr v. Lackie (September 19, 2017; 2017 ONCA 716) and Cobb v. Long Estate (September 19, 2017; 2017 ONCA 717), the court of appeal ruled that the change to prejudgment interest for non-pecuniary losses from a set level of 5% to the level that applies to pecuniary losses were implemented to achieve particular policy objectives and therefore should have retrospective application (i.e. to be applied to all settlements on or after January 1, 2015). The Ontario Court of Appeal ruling in El-Khodr v. Lackie was appealed to the Supreme Court of Canada; on June 7, 2018, the Supreme Court of Canada dismissed the application for leave to appeal from the judgment of the Court of Appeal for Ontario. 12 Pecuniary awards are defined on the Ontario Attorney General s website as Damages that can be measured in money (i.e., special damages) with special damages further defined as Damages intended to compensate a plaintiff for a quantifiable monetary loss. Examples of such losses include: lost earnings, medical bills, and repair costs. In contrast, non-pecuniary awards are defined as Damages that cannot be measured in money, but nevertheless are compensated for with money (i.e., general damages) with general damages further defined as Damages for non-monetary losses suffered by a plaintiff. These damages are not capable of exact quantification. Examples of such losses suffered include pain, suffering, and disfigurement. 13 Under the Court of Justice Act Section 128.1, the applicable interest rate is based on the date the cause of action arose ; however, the Insurance Act Section 258.3(8) states that no prejudgment interest is payable for.. any period of time before the plaintiff served the notice.., thus effectively making prejudgment interest calculated from the date of notice of claim, as opposed to being based on the date of the accident. page 14 of 74

16 A reform adjustment related to changes in non-pecuniary prejudgment interest provisions, introduced in Ontario 2014 Bill 15, was included in the Ontario Industry PPV trend analyses (and thereby in the calculation of a priori estimates used in the selection of current valuation estimates of ultimates). The reform adjustment, applied to the bodily injury coverage, resulted in an estimated 5% reduction on accident years 2015 and subsequent selected loss costs. With the current valuation (as at September 30, 2018), no additional reform adjustment was included as we have assumed the retroactive impact of this product reform change has been fully reflected in outstanding case reserves (i.e. the retroactive non-pecuniary PJI adjustment to Ontario RSP bodily injury unpaid amounts, impacting AY2014 and prior, which was introduced during the valuation was fully aged out and removed with the current valuation (as expected)). Ontario 2014 Bill 15 included changes to the Insurance Company Act, via Ontario Regulation 236/14 changes to Ontario Regulation changing Section 51 on Overdue Payments. Ontario Regulation Section 51(4) aligns the interest rate payable for Statutory Accident Benefits Schedule (SABS) claims in dispute under Section 280 of the Insurance Act with the prejudgment interest rate described in subsection of the Ontario Courts of Justice Act. Thus, the SABS interest rate for overdue payments while in the dispute resolution process was reduced from a rate of 1% per month to a rate aligned with current market rates. This change came into force on January 1, A reform adjustment related to changes in the SABS Interest Rate for overdue payments while in the dispute resolution process, introduced in Ontario 2014 Bill 15, was included in the Ontario Industry PPV trend analyses (and thereby in the calculation of a priori estimates used in the selection of current valuation estimates of ultimates). The reform adjustment, applied to the Medical Expense and Disability Income sub-coverages, resulted in an estimated 1% decrease on accident years 2015 and subsequent selected loss costs. Ontario 2015 Bill 91 included changes to the Insurance Company Act, via Ontario Regulation 251/15 changes to Ontario Regulation changing the available benefits payable under the SABS. Changes to the SABS include changes to monetary limits and benefit durations payable. These changes came into force on June 1, A reform adjustment related to SABS product changes impacting policies written after June 1, 2016, introduced in Ontario 2015 Bill 91, was included in the Ontario Industry PPV trend analyses (and thereby in the calculation of a priori estimates used in the selection of projected future loss costs). The reform adjustment, applied to the medical expense and disability income sub-coverages, resulted in an estimated 14% decrease on selected loss costs for policies written after June 1, A corresponding adjustment, applied to the bodily injury coverage, resulted in an estimated 6.4% increase on selected loss costs for PPV policies written after June 1, Ontario 2015 Bill 91 included changes to the Insurance Company Act, via Ontario Regulation 221/15 changes to Ontario Regulation 461/96 amending the current tort deductible and monetary threshold beyond which the tort deductible does not apply. This change came into force on January 1, 2015, but it was not clear if this applied only to actions brought on or after January 1, 2015, or if it applied to all settlements on or after January 1, In the Ontario Court of Appeal decisions in El-Khodr v. Lackie (September 19, 2017; 2017 ONCA 716) and Cobb v. Long Estate (September 19, 2017; 2017 ONCA 717), the court of appeal ruled that the changes to the tort deductible and monetary threshold were implemented to achieve page 15 of 74

17 particular policy objectives and therefore should have retrospective application (i.e. to be applied to all settlements on or after January 1, 2015). A reform adjustment related to changes in bodily injury tort threshold and deductibles, introduced in Ontario 2015 Bill 91, was included in the Ontario Industry PPV trend analyses (and thereby in the calculation of a priori estimates used in the selection of current valuation estimates of ultimates). The Ontario Industry PPV trend analyses reform adjustments, applied to the bodily injury coverage, resulted in an estimated 6% decrease in accident years 2015 and subsequent selected loss costs. The Facility Association view, consistent with the recent Ontario Court of Appeal decisions in El-Khodr v. Lackie and Cobb v. Long Estate, is that recent changes to the bodily injury tort threshold and deductibles are on a settlement date basis, with the current valuation (as at September 30, 2018), no additional reform adjustment was included as we have assumed the retroactive impact of this product reform change has been fully reflected in outstanding case reserves. C.2 Booked results for the prior valuation implementation It is helpful to consider how the portfolio looked after the prior valuation was implemented. In this case, the August 2018 booked results were based on assumptions derived from the prior (June 30, 2018) valuation and were discussed in the associated monthly Actuarial Highlights. The charts immediately below show the associated levels of claim liabilities 14 booked by accident year 15. The left chart displays life-to-date payments, case reserves, IBNR, and the total including actuarial present value adjustments against accident year earned premium. The right chart shows the associated dollar amounts for the components of the claim liabilities and the then-current projected amount of 2018 full year earned premium (the red hash-mark line) to provide some perspective % ONTARIO Accident Year Loss Aug 31, 2018 $ millions 400 ONTARIO Accident Year Unpaid Claim Aug 31, 2018 % proj EP apv adj.: 22% nominal unpaid: 266% 160.0% % % 100.0% % % % % 50 (20.0%) (50) PRIOR Paid Indemnity & Allowed Claims Expense Ratio Case Ratio IBNR (nominal) Ratio Ultimate Indemnity & Allowed Claims Expense (M/S) as % EP Case Reserves IBNR (nominal) APV Adjustment (M/S) proj EP M/S refers to Member Statement values that is, actuarial present value adjustments at the selected discount rate. The tables at the top of the next page show the associated Member Statement (M/S) policy liabilities. 14 Claim liabilities refer to provision for unpaid indemnity and allowed claims expenses. Allowed claims expenses are first party legal and other expenses as listed in the RSP Claims Guide. Claims expenses paid through the member company expense allowance are NOT included in this discussion. 15 The loss ratio chart has been limited to show the most recent 20 accident years; the unpaid provision chart has been limited to show the most recent 20 accident years, and show all accident years older than 20 years collectively as PRIOR. page 16 of 74

18 claim liabilities ($000s) amt % case 620, % ibnr 324, % M/S apv adjust. 78, % M/S total 1,024, % premium liabilities ($000s) amt % unearned prem 180, % prem def/(dpac) 48, % M/S apv adjust. 18, % M/S total 248, % policy liabilities ($000s) amt % claim 945, % premium 229, % M/S apv adjust. 97, % M/S total 1,272, % C.3 Booked results for the current valuation implementation The October 2018 booked results were based on assumptions derived from the current (September 30, 2018) valuation and are discussed in the associated monthly Actuarial Highlights. The charts immediately below show the levels of claim liabilities booked by accident year 16 on that basis. The left chart displays life-to-date payments, case reserves, IBNR, and the total including actuarial present value adjustments against accident year earned premium. The right chart shows the associated dollar amounts for the components of the claim liabilities and the current projected amount of 2018 full year earned premium (the red hash-mark line) to provide some perspective % ONTARIO Accident Year Loss Oct 31, 2018 $ millions 400 ONTARIO Accident Year Unpaid Claim Oct 31, 2018 % proj EP apv adj.: 23% nominal unpaid: 270% 160.0% % % % % % % % (50) PRIOR Paid Indemnity & Allowed Claims Expense Ratio Case Ratio IBNR (nominal) Ratio Ultimate Indemnity & Allowed Claims Expense (M/S) as % EP Case Reserves IBNR (nominal) APV Adjustment (M/S) proj EP M/S refers to Member Statement values that is, actuarial present value adjustments at the selected discount rate. The tables at the top of the next page show the associated Member Statement (M/S) policy liabilities. 16 The loss ratio chart has been limited to show the most recent 20 accident years; the unpaid provision chart has been limited to show the most recent 20 accident years, and show all accident years older than 20 years collectively as PRIOR. page 17 of 74

19 claim liabilities ($000s) amt % case 620, % ibnr 336, % M/S apv adjust. 80, % M/S total 1,037, % premium liabilities ($000s) amt % unearned prem 179, % prem def/(dpac) 49, % M/S apv adjust. 19, % M/S total 248, % C.4 a priori loss ratios The Ontario RSP a priori loss ratios were carried forward from the June 30, 2018 valuation, and are presented in the B.1.4, B.2.3, B.3.3, and B.4.3 exhibits in section L. C.5 Actual vs (AvsP) policy liabilities ($000s) amt % claim 956, % premium 228, % M/S apv adjust. 99, % M/S total 1,285, % Variances in projected recorded and paid emergence and the associated actual emergence is presented in the two following tables. Ontario RSP Third Party Liability Accident Benefits Other Coverages Total 2018 Actual Actual Actual Actual 2018 Accident [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] Year =[2][1] =[5][4] =[8][7] =[1]+[4]+[7] =[2]+[5]+[8] =[11][10] 2013 & Prior (29) , (16) (385) (369) 684 1, (75) (1,560) (1,485) 93 2,693 2,600 (4) (155) (151) ,386 1,199 3, (3,697) (54) (554) (500) 3, (2,998) ,783 2,211 (1,572) 4,403 4,063 (340) 11 (788) (799) 8,197 5,486 (2,711) ,285 2,412 (5,873) 11,835 4,126 (7,709) 164 (1,825) (1,989) 20,284 4,713 (15,571) ,504 27,625 (1,879) 24,487 23,428 (1,059) 25,464 25, ,455 76,544 (2,911) Total 41,655 32,864 (8,791) 45,380 35,618 (9,762) 25,565 21,784 (3,781) 112,600 90,266 (22,334) 2017 & prior 12,151 5,239 (6,912) 20,893 12,190 (8,703) 101 (3,707) (3,808) 33,145 13,722 (19,423) *projected recorded claims based on to Ultimate emergence model as at 2018Q2 As indicated above, total recorded emergence at $90.3 million was $22.3 million (19.8%) less than the $112.6 million projected. Prior accident year favourable experience was impacted by member correction transactions reported during the quarter, resulting in a $5.8 million decrease in recorded activity (reductions in case reserves) in the period impacting AY2010-AY2017. These correction transactions were consistent with valuation IBNR adjustments included with the prior valuation (RSP 2018-Q2). Favourable development in the current accident year (AY2018) was also impacted by favourable member correction transactions reported during the quarter. Notwithstanding the impact of the member corrections reported in the current quarter, we continue to observe high levels of older accident year (AY2014 and prior) unfavourable accident benefits recorded claims activity, offsetting lower levels of recent accident year (AY2015-AY2018) accident benefits recorded claims activity. page 18 of 74

20 Ontario RSP Third Party Liability Accident Benefits Other Coverages Total Actual Paid Actual Paid Actual Paid Actual Paid Paid Claims Paid Claims Paid Claims Paid Claims in in in in Accident [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] Year =[14][13] =[17][16] =[20][19] =[13]+[16]+[19] =[14]+[17]+[20] =[23][22] 2013 & Prior 5,985 7,231 1,246 7,638 4,921 (2,717) (199) 13,829 12,159 (1,670) ,245 3, ,040 1,724 (3,316) 24 3 (21) 8,309 5,352 (2,957) ,666 3,575 (91) 4,810 2,904 (1,906) (35) 8,557 6,525 (2,032) ,601 2,166 (1,435) 4,589 4,583 (6) (16) 8,288 6,831 (1,457) , (968) 8,698 8,049 (649) 1, (1,713) 12,432 9,102 (3,330) ,507 16,500 (1,007) 4,794 4,651 (143) 26,500 23,427 (3,073) 48,801 44,578 (4,223) Total 35,905 34,030 (1,875) 35,569 26,832 (8,737) 28,742 23,685 (5,057) 100,216 84,547 (15,669) 2017 & prior 18,398 17,530 (868) 30,775 22,181 (8,594) 2, (1,984) 51,415 39,969 (11,446) *projected paid claims based on Paid to Ultimate emergence model as at 2018Q2 As indicated above, total paid emergence at $84.5 million was $15.7 million (15.6%) less than the $100.2 million projected. While we continue to observe high levels of older accident year (AY2014 and prior) unfavourable accident benefits recorded claims activity, we are not observing corresponding paid claims activity, perhaps suggesting a change in emergence patterns and case reserve levels in relation to increased large loss claims reporting. Additional detail and summary charts akin to those found in the monthly Actuarial Highlights are presented in the sections that follow. There are constraints in this analysis in that full 25-quarter levels are not available for all metrics considered, as the detail has not all been fully captured and entered into the Facility Association databases (for instance, projected recorded and paid activity prior to 2013 ). C.5.1 AvsP: Indemnity & Allowed Claims Expense Actual recorded activity (paid and case reserve changes) over the last 25-calendar quarters is shown in the charts immediately below, including the prior 24 quarter average level. Ontario RSP Actual by Calendar Quarter 120,000 Ontario PAYs (latest prior 24 qtr avg = 28,951) Ontario CAY (latest prior 24 qtr avg = 50,053) PAYs Actual prior 24 qtr avg CAY Actual projected amount 120,000 prior 24 qtr avg projected amount 100, ,000 80,000 80,000 60,000 60,000 40,000 40,000 20,000 20,000 activity variances from the previous quarter s projections are shown in the charts at the top of the next page, including the prior 24-quarter standard deviation levels. page 19 of 74

21 Ontario RSP Actual vs Summary: Variances 17 by Calendar Quarter 30,000 20,000 10,000 (10,000) (20,000) (30,000) (40,000) (50,000) (60,000) (70,000) (80,000) Ontario PAYs (latest prior 24 qtr std dev = +/10,539) Ontario CAY (latest prior 24 qtr std dev = +/14,245) PAYs Actual less prior 24 qtr std dev CAY Actual less prior 24 qtr std dev On Latest $ thousands PAYs CAY Actual less 28,951 50,053 std dev 10,539 14,245 AP <> std dev 11 2 % <> std dev 47.8% 8.7% norm <> std dev 31.7% 31.7% 30,000 20,000 10,000 (10,000) (20,000) (30,000) With respect to recorded indemnity & allowed claims expense, the prior accident years (PAYs) variances (left chart above) indicate bias 18 in the projection process (such that projections tend to be too high in retrospect, where projections are available, with only 3 times in the past 23 quarters where actuals were higher than our projections for the PAY recorded amount), and while the magnitude of the variances remain high, improvements in the variance magnitude are clear. That being said, 48% of variances related to the available projections were outside of one standard deviation, suggesting the projection process has performed worse than simply projecting the prior 24-quarter average amount. Given the significant reductions in ultimate estimates in valuations since the 2010 reforms, it is difficult to determine at this point how much of the poor projection result is due to the various causes as discussed in Section B.4. However, our current view is that the current AvsP variances support the view that the historical valuation ultimate selections, in hindsight, were redundant to some degree (hence our reduced ultimate selections). The PAY recorded variance fell outside of the one standard deviation band during the latest quarter. While the variance was impacted by member corrections described in the previous section, the PAY recorded claims activity in the quarter was reviewed and confirmed, with the remaining variance attributed to process variance. The current accident year ( CAY ) recorded variances (right chart above) fell outside of one standard deviation 9% of the time (where projections are available), suggesting that the projection process has performed better than simply projecting the prior 24-quarter average amount, but this is, admittedly, not a difficult hurdle to overcome (as the 24-quarter average does not take into account the obvious and expected CAY pattern of recorded activity increases as the CAY moves from to Q4). Unlike the PAY discussion, bias in the projections has not been indicated at the 95% confidence level on a lagging 23-quarter basis, with 7 of the actuals higher than the associated projection. 17 Projections are available only for valuation quarters 2013 and beyond. For purposes of presentation, we have set the actual vs projection variances to 0 for all prior periods. 18 For the binomial distribution with 23 trials and an assumed 50% success probability, the 95% confidence range is 7 to 16. That is, for the 23 quarters presented, if the recorded projection was unbiased, with a 95% confidence, we would expect between 7 to 16 variances above 0. Less than 7 variances above 0 would indicate that our projections are biased high, and greater than 16 variances would indicate that our projections are biased low. page 20 of 74

22 We have included, for reference, additional charts immediately below related to levels influencing recorded activity. Ontario RSP Levels that influence 19 activity by Calendar Quarter Ontario PAYs (latest prior 24 qtr avg = 314,152,762) Ontario CAY (latest prior yr = 281,635) PAYs Beginning IBNR prior 24 qtr avg CAY Ending YTD Earned Premium 600, , , , , , , , , , , , ,000 50, % 20.0% 15.0% 10.0% 5.0% Ontario PAYs (latest prior 24 qtr avg = 9.7%) Ontario CAY (latest prior 24 qtr avg = 31.7%) PAY as % of Beginning IBNR prior 24 qtr avg CAY as % CAY YTD Earned Premium projected amount 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% prior 24 qtr avg projected amount 0.0% 0.0% We track beginning prior accident years IBNR as recorded activity comes out of IBNR. Changes in the prior accident years beginning IBNR (see upper left chart above) occur for several possible reasons: to offset actual recorded activity (through loss ratio matching); the annual switchover as a current accident year becomes a prior accident year (occurs in January); and IBNR levels potentially change with each new valuation. The chart on the lower left above shows that from to 2015-Q4 inclusive, prior accident year recorded activity had been consistently below 10% of beginning IBNR, although the emergence model projections prior to our 2014 model refinement was projecting recorded at more than 15%. With the benefit of hindsight, it may be that what now appears to be redundancy in our previous IBNR selections may have played a part in these inaccurate projections, as may have the previous practice (up to valuation 2014 Q2) of projecting emergence at a government line level. It is difficult to assess whether recent increases (since 2016-) in the prior accident year recorded activity relative to IBNR is being driven by increases in recorded activity/changes in case reserving practices, the significant reductions in IBNR over the prior valuations, or changes in payment emergence patterns. 19 Our recorded activity projections are based on selected ratios of life-to-date recorded activity to ultimate, converted to a recorded to beginning IBNR ratio, where ratio selection is based on our review of historical results. We find it helpful to show CAY emergence relative to earned premium for the purposes of the AvsP discussion. page 21 of 74

23 CAY recorded activity relative to year-to-date earned premium (bottom right chart above) may be showing a potential (deteriorating) trend in relation to and Q2 recorded activity (similar trends for and Q4 are not quite as clear at this point). These deteriorations may reflect reductions in earned rate levels or, alternatively, increases in loss costs that are not being offset by earned rate changes. At this point, we are not able to draw definitive conclusions, other than this seems to support the selected accident year a priori loss ratios showing similar deterioration (both are deteriorating 4-5% per year). C.5.2 AvsP: Paid Indemnity & Allowed Claims Expense The charts immediately below show actual paid activity in each of the most recent 25 calendar quarters, along with a prior 24-quarter average to show how each quarter s actual compares with the average amount of the preceding 24 calendar quarters. Ontario RSP Actual Paid activity by Calendar Quarter 90,000 80,000 70,000 60,000 50,000 40,000 30,000 20,000 10,000 Ontario PAYs (latest prior 24 qtr avg = 47,732) Ontario CAY (latest prior 24 qtr avg = 25,631) PAYs Actual Paid prior 24 qtr avg CAY Actual Paid projected amount 60,000 50,000 40,000 30,000 20,000 10,000 prior 24 qtr avg projected amount The charts immediately below show the actual less projected paid variances for the last 25 calendar quarters, along with bands for the prior 24-quarter standard deviations to show how the variances from projection compare with historical standard deviations. Ontario RSP Actual vs Summary: Paid Variances 20 by Calendar Quarter Ontario PAYs (latest prior 24 qtr std dev = +/6,651) Ontario CAY (latest prior 24 qtr std dev = +/9,350) PAYs Actual less Paid prior 24 qtr std dev CAY Actual less Paid 20,000 15,000 prior 24 qtr std dev 10,000 10,000 5,000 (10,000) (20,000) (5,000) (30,000) (10,000) (40,000) (15,000) On Latest $ thousands Paid PAYs CAY Qtrly Avg Paid (prior 24 qtrs) 47,732 25,631 std dev 6,651 9,350 AP <> std dev 17 % <> std dev 73.9% 0.0% norm <> std dev 31.7% 31.7% With respect to paid indemnity & allowed claims expense prior accident years variances (left chart above), 74% of the variances (where projections are available) have fallen outside of one standard deviation, suggesting the projection process has performed worse than projecting simply based on the 20 Projections are available only for valuation quarters 2013 and beyond. For purposes of presentation, we have set the actual vs projection variances to 0 for all prior periods. page 22 of 74

24 preceding 24-quarter average. The variances suggest that the projection process is biased (with only 4 times in the past 23 quarters where actuals were higher than our projections for the PAY paid amount), although there was evidence of improvement up until A similar change is not as evident with the PAY recorded amount at 2017-, perhaps suggesting a change in claims payment emergence patterns, we continue to review and monitor, particularly with respect to potential changes in claims payment and case reserve levels in relation to increased accident benefits large loss claims reporting. The PAY paid variance fell outside of the one standard deviation band during the latest quarter. The activity was reviewed and confirmed, with the variance attributed to process variance. In contrast, the current accident year paid variances (right chart middle of the prior page) tend to show actuals higher than projected (although not by much). While the CAY paid variances fell outside of one standard deviation 0% of the time (where projections are available), suggesting the projection process has performed better than projecting simply based on the preceding 24-quarter average, however, bias has been indicated at a 95% confidence level on a lagging 23-quarter basis, with 17 times in the past 23 quarters where actuals were higher than our projections for the CAY paid amount. We have included, for reference, additional charts immediately below related to levels influencing paid activity. Ontario RSP Levels that influence 21 Paid activity by Calendar Quarter 1,200,000 1,000, ,000 Ontario PAYs (latest prior 24 qtr avg = 817,349) Ontario CAY (latest prior yr = 281,635) prior 24 qtr avg PAY Beginning Unpaid IBNR CAY Ending YTD Earned Premium case 400, , , , , , , , , ,000 50,000 Ontario PAYs (latest prior 24 qtr avg = 5.9%) Ontario CAY (latest prior 24 qtr avg = 15.9%) 10.0% PAY Paid as % Beginning Unpaid prior 24 qtr avg projected amount 35.0% CAY Paid as % CAY YTD Earned Premium prior 24 qtr avg projected amount 9.0% 8.0% 7.0% 30.0% 25.0% 6.0% 5.0% 4.0% 20.0% 15.0% 3.0% 2.0% 1.0% 10.0% 5.0% 0.0% 0.0% We track beginning prior accident years unpaid balance (case and IBNR) as paid activity comes out of the unpaid balance. Changes in the prior accident years beginning unpaid balance (see upper left chart above) occur for several possible reasons: 21 Our paid projections are based on selected ratios of life-to-date paid activity to ultimate, converted to a paid to beginning unpaid ratio, where ratio selection is based on our review of historical results. We find it helpful to show CAY emergence relative to earned premium for the purposes of the AvsP discussion. page 23 of 74

25 to offset actual paid activity (may reduce case or IBNR or both); the annual switchover as a current accident year becomes a prior accident year (occurs in January); and IBNR levels potentially change with each new valuation. Similar to our comments related to current accident year recorded activity as a percentage of year-todate earned premium, there appears to be a deterioration in the current accident year paid ratios to earned premium that supports our selections of a priori loss ratios (deteriorating at about 4-5% per accident year). C.6 Current valuation IBNR selections Exhibit B.1.1 (see section L for all exhibits) summarizes the overall change in ultimate with this valuation and B.1.2 shows selected loss ratios over the most recent 4 valuations for comparison purposes on an all coverages basis. The B.2 exhibits provide information for third party liability, B.3 exhibits for accident benefits, and B.4 exhibits for the other government line. C.7 Premium Liabilities / Future Accident Years In order to provide a basis for estimating the full premium liability level for monthly statements (i.e. the level of premium deficiency liability / deferred policy acquisition cost asset to carry) we leverage the a priori loss ratios for the accident year underlying the unearned premium levels. The test of recoverability leverages assumptions set by the Appointed Actuary. These include the member expense allowances (taking into account the Board approved allowances) and policy administration / maintenance expense assumptions. C.8 Actuarial Present Value Adjustments C.8.1 Selected Claims Payment Patterns Payment patterns are selected through the emergence models (the same used for projecting future claims paid and recorded activity for the AvsP process), leveraging a paid to ultimate metric. C.8.2 Selected Discount Rate The projected future claims paid cash flows are matched to a simulated portfolio of Government of Canada benchmark monthly bonds (yields anchored to the valuation date), and 15 basis point investment expense is assumed. A discount rate of 2.22% per annum was selected for the valuation of the claim liabilities and premium liabilities at September 30, 2018, up from 1.83% selected with the June 30, 2018 valuation. The chart to the right shows the Government of Canada benchmark bond yield curves at June 2018 and September Annual Effective Yield 3.00% 2.50% 2.00% 1.50% 1.00% 0.50% Government of Canada Benchmark Bond Yields Jun18 Sep % Duration (mths) page 24 of 74

26 Sensitivity to the discount rate assumption is presented in Exhibit C (see section L). C.8.3 Selected Margins for Adverse Deviations The margin for adverse deviation ( MfADs ) for investment income was maintained at 25 basis points with the current valuation. The claims development MfADs for all RSPs were reviewed with the prior valuation (as at June 30, 2018) as per our regular annual valuation process. In particular, Ontario RSP selected claims development MfADs for older accident years were reviewed and judgmentally reduced to reflect the decreasing uncertainty over time, resulting in an estimated $30.4 million favourable impact, implemented in the August 2018 Operational Reports. Our usual process is to review and update the claims development MfADs annually with the June 30 valuation. However, with the current valuation (as at September 30, 2018), the Ontario (third party liability bodily injury and accident benefits) claims development MfADs were further reviewed and updated based on discussion and feedback with the FA s Actuarial Committee. This is summarized in Exhibit D (see section L). The estimated unfavourable implementation impact of updating the claims development margins was an increase in the nominal claims PfAD of $13.9 million. C.9 Special IBNR Provisions / Adjustments As discussed in the Highlights section C.1 and in the Ontario Product Reform adjustments section C.1.1. During the current valuation (September 30, 2018), a balance sheet adjustment was applied to selected nominal implementation loss ratios for the Ontario RSP to offset an incorrectly reported member case reserve overstatement impacting AY2006-AY2009. This balance sheet adjustment resulted in an overall nominal reduction in claims liabilities of $1.8 million The remaining case reserve correction transactions were submitted by the member in October 2018, reviewed by FA management and included alongside the current valuation implementation with the RSP Operational Results for the month of October The submitted corrections were aligned with the valuation provisions, which will result in a neutral implementation impact. page 25 of 74

27 D. ALBERTA GRID RSP D.1 Valuation Highlights A summary of the valuation results through time is available in the A exhibit (see section L for all exhibits), with detail related to the current valuation provided in the B.1.1 and B.1.2 exhibits. The change in selected ultimate for prior accident years was $4.3 million favourable with this valuation (1.7% of the unpaid estimate as at last quarter), bringing the calendar year-to-date total unfavourable to $7.8 million (2.5% of the unpaid estimate as at the beginning of the 2018 calendar year). These changes are presented by accident year and government line in the tables below. Alberta Grid RSP valuation changes in selected ultimate Alberta Grid RSP valuation changes in selected ultimate (favourable) / unfavourable during Quarter (favourable) / unfavourable YTD Accident Year Third Party Accident Other Third Party Accident Other Total Accident Year Liability Benefits Coverages Liability Benefits Coverages Total 2013 & Prior 761 (17) (320) & Prior 1,544 (5) (393) 1, (627) (57) 55 (629) 2014 (1,192) (47) 134 (1,105) 2015 (784) 6 (6) (784) 2015 (1,257) 46 (57) (1,268) 2016 (979) (8) (109) (1,096) 2016 (1,675) 10 (272) (1,937) 2017 (1,674) (239) (340) (2,253) 2017 (2,178) (727) (1,681) (4,586) TOTAL (3,303) (315) (720) (4,338) TOTAL (4,758) (723) (2,269) (7,750) Favourable third party liability development is driven by bodily injury reported case reserve reductions, across multiple member companies, partially offset by unfavourable older accident year bodily injury (AY2009-AY2011) claims settlements. The selected loss ratios for accident year 2018 (current accident year; AY2018 down 2.0 points to 89.8%) and accident year 2019 (future accident year; AY2019 down 1.7 points to 88.8%) both decreased. Change in selected loss ratio for AY2019 were driven by updated a priori loss ratio written premium projection assumptions to reflect the Alberta grid rate mandatory coverages increase effective January 1, 2019 (AIRB Bulletin ). Summary descriptions of recent regulatory and legislative initiatives are available in section I. The valuation process is described in more detail in section K, and a summary of changes to the process during this fiscal year is provided in section H. Policy liability projected cash flows and September 2018 government of Canada bond yields were used to determine the applicable discount rate. The selected investment income margin for adverse deviation ( MfAD ) was maintained at 25 basis points with the current valuation. Our usual process is to review and update the claims development MfADs annually with the June 30 valuation. However, with the current valuation (as at September 30, 2018), the Alberta Grid (third party liability bodily injury) claims development MfADs were reviewed based on discussion and feedback with the FA s Actuarial Committee, but no changes were made to the selections from the prior review (as at June 30, 2018). See Exhibit D in section L for selected margins. page 26 of 74

28 D.2 Booked results for the prior valuation implementation It is helpful to consider how the portfolio looked after the prior valuation was implemented. In this case, the August 2018 booked results were based on assumptions derived from the prior (June 30, 2018) valuation and were discussed in the associated monthly Actuarial Highlights. The charts immediately below show the associated levels of claim liabilities 23 booked by accident year 24. The left chart displays life-to-date payments, case reserves, IBNR, and the total including actuarial present value adjustments against accident year earned premium. The right chart shows the associated dollar amounts for the components of the claim liabilities and the then-current projected amount of 2018 full year earned premium (the red hash-mark line) to provide some perspective % Alberta Grid Accident Year Loss Aug 31, 2018 $ millions 180 Alberta Grid Accident Year Unpaid Claim Aug 31, 2018 % proj EP apv adj.: 13% nominal unpaid: 199% % % % % % (20) Paid Indemnity & Allowed Claims Expense Ratio Case Ratio IBNR (nominal) Ratio Ultimate Indemnity & Allowed Claims Expense (M/S) as % EP M/S refers to Member Statement values that is, actuarial present value adjustments at the selected discount rate. The tables immediately below show the associated Member Statement (M/S) policy liabilities. claim liabilities ($000s) amt % case 194, % ibnr 126, % M/S apv adjust. 21, % M/S total 342, % D.3 Booked results for the current valuation implementation Case Reserves IBNR (nominal) APV Adjustment (M/S) proj EP premium liabilities ($000s) amt % unearned prem 82, % prem def/(dpac) (6,886) (8.7%) M/S apv adjust. 4, % M/S total 79, % policy liabilities ($000s) amt % claim 321, % premium 75, % M/S apv adjust. 25, % M/S total 421, % The October 2018 booked results were based on assumptions derived from the current (September 30, 2018) valuation and are discussed in the associated monthly Actuarial Highlights. 23 Claim liabilities refer to provision for unpaid indemnity and allowed claims expenses. Allowed claims expenses are first party legal and other expenses as listed in the RSP Claims Guide. Claims expenses paid through the member company expense allowance are NOT included in this discussion. 24 Accident year 2004 was an incomplete year and therefore has been excluded from the loss ratio chart. page 27 of 74

29 The charts immediately below show the levels of claim liabilities booked by accident year on that basis. The left chart displays life-to-date payments, case reserves, IBNR, and the total including actuarial present value adjustments against accident year earned premium. The right chart shows the associated dollar amounts for the components of the claim liabilities and the current projected amount of 2018 full year earned premium (the red hash-mark line) to provide some perspective % Alberta Grid Accident Year Loss Oct 31, 2018 $ millions 180 Alberta Grid Accident Year Unpaid Claim Oct 31, 2018 % proj EP apv adj.: 11% nominal unpaid: 196% % % % % % (20) Paid Indemnity & Allowed Claims Expense Ratio Case Ratio IBNR (nominal) Ratio Ultimate Indemnity & Allowed Claims Expense (M/S) as % EP M/S refers to Member Statement values that is, actuarial present value adjustments at the selected discount rate. The tables immediately below show the associated Member Statement (M/S) policy liabilities. claim liabilities ($000s) amt % case 192, % ibnr 124, % M/S apv adjust. 17, % M/S total 333, % D.4 a priori loss ratios Case Reserves IBNR (nominal) APV Adjustment (M/S) proj EP premium liabilities ($000s) amt % unearned prem 88, % prem def/(dpac) (9,138) (11.1%) M/S apv adjust. 3, % M/S total 82, % policy liabilities ($000s) amt % claim 316, % premium 78, % M/S apv adjust. 20, % M/S total 416, % This valuation s a priori loss ratios are summarized in the charts and table at the top of the next page (by government line and accident half-year), with a comparison to the last valuation (as at June 30, 2018) a priori loss ratios. We do not typically provide this detail for roll-forward valuations, as a priori loss ratios are generally not updated with those valuations. However, as a grid rate increase (effective January 1, 2019) was approved since our last valuation, we opted to update the a priori loss ratios to take this into account. page 28 of 74

30 140.0% RSP: Alberta Grid Total 160.0% RSP: Alberta Grid Third Party Liability 120.0% 140.0% 100.0% 80.0% 60.0% 120.0% 100.0% 80.0% 60.0% 40.0% 40.0% 20.0% 20.0% '07 '08 '09 '10 '11 '19 Paid as % EP Case as % EP IBNR as % EP Favourable Unpaid as % EP Unfavourable Unpaid as % EP Prior Selected a priori LRs Current Selected a priori LRs '20 '07 '08 '09 '10 '11 '19 Paid as % EP Case as % EP IBNR as % EP Favourable Unpaid as % EP Unfavourable Unpaid as % EP Prior Selected a priori LRs Current Selected a priori LRs ' % 180.0% 160.0% 140.0% 120.0% 100.0% 80.0% 60.0% 40.0% 20.0% '07 RSP: Alberta Grid Accident Benefits '08 '09 '10 '11 '19 Paid as % EP Case as % EP IBNR as % EP Favourable Unpaid as % EP Unfavourable Unpaid as % EP Prior Selected a priori LRs Current Selected a priori LRs ' % 100.0% 80.0% 60.0% 40.0% 20.0% (20.0%) (40.0%) '07 RSP: Alberta Grid Other Coverages '08 '09 '10 '11 '19 Paid as % EP Case as % EP IBNR as % EP Favourable Unpaid as % EP Unfavourable Unpaid as % EP Prior Selected a priori LRs Current Selected a priori LRs '20 CURRENT PRIOR CHANGE Accident Third Party Accident Other Third Party Accident Other Third Party Accident Other Total Total Period Liability Benefits Coverages Liability Benefits Coverages Liability Benefits Coverages Total 2013 / % 95.0% 49.0% 78.0% 94.0% 95.0% 49.0% 78.0% 2013 / % 114.0% 72.0% 95.0% 107.0% 114.0% 72.0% 95.0% 2014 / % 97.0% 50.0% 80.0% 97.0% 97.0% 50.0% 80.0% 2014 / % 117.0% 74.0% 97.0% 110.0% 117.0% 74.0% 97.0% 2015 / % 100.0% 51.0% 82.0% 99.0% 100.0% 51.0% 82.0% 2015 / % 122.0% 77.0% 101.0% 112.0% 122.0% 77.0% 101.0% 2016 / % 105.0% 55.0% 86.0% 100.0% 105.0% 55.0% 86.0% 2016 / % 128.0% 80.0% 103.0% 112.0% 128.0% 80.0% 103.0% 2017 / % 111.0% 57.0% 88.0% 101.0% 111.0% 57.0% 88.0% 2017 / % 133.0% 84.0% 106.0% 114.0% 133.0% 84.0% 106.0% 2018 / % 108.0% 58.0% 86.0% 98.0% 108.0% 58.0% 86.0% 2018 / % 126.0% 85.0% 96.0% 100.0% 126.0% 85.0% 96.0% 2019 / % 106.0% 59.0% 81.0% 91.0% 108.0% 59.0% 82.0% (1.0%) (2.0%) (1.0%) 2019 / % 127.0% 86.0% 97.0% 103.0% 130.0% 86.0% 99.0% (3.0%) (3.0%) (2.0%) 2020 / % 109.0% 60.0% 82.0% 93.0% 111.0% 60.0% 84.0% (2.0%) (2.0%) (2.0%) The valuation results used to form the basis of the a priori estimates for the September 30, 2018 valuation continued to use selected ultimates from the March 31, 2018 valuation (consistent with the prior valuation a priori loss ratio assumptions). Trend structure models based on industry indemnity results as at December 31, 2017 were used (again, consistent with the prior valuation a priori loss ratio assumptions). Written premium projection assumptions were updated to reflect the Alberta grid rate mandatory coverages increase effective January 1, 2019 (AIRB Bulletin ). The current valuation Alberta Grid RSP a priori loss ratios are presented in the B.1.4, B.2.3, B.3.3, and B.4.3 exhibits in section L. page 29 of 74

31 D.5 Actual vs (AvsP) Variances in projected recorded and paid emergence and the associated actual emergence are presented in the two following tables. Alberta Grid RSP Third Party Liability Accident Benefits Other Coverages Total 2018 Actual Actual Actual Actual 2018 Accident [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] Year =[2][1] =[5][4] =[8][7] =[1]+[4]+[7] =[2]+[5]+[8] =[11][10] 2013 & Prior 1,131 2,201 1,070 (17) (17) 2 (318) (320) 1,133 1, , (624) 2 (54) (56) (6) , (633) , (923) (12) (15) (3) 1, (914) ,560 1,044 (516) 2 (23) (25) (56) (124) (68) 1, (609) ,149 1,810 (3,339) 94 (33) (127) (807) (1,001) (194) 4, (3,660) ,994 11,953 (4,041) 1,415 1,311 (104) 11,641 6,572 (5,069) 29,050 19,836 (9,214) Total 26,465 18,092 (8,373) 1,518 1,201 (317) 10,762 5,155 (5,607) 38,745 24,448 (14,297) 2017 & prior 10,471 6,139 (4,332) 103 (110) (213) (879) (1,417) (538) 9,695 4,612 (5,083) *projected recorded claims based on to Ultimate emergence model as at 2018Q2 As indicated above, total recorded emergence at $24.4 million was $14.3 million (36.9%) less than the $38.7 million projected. The current accident year (AY2018) accounted for 64.4% of the favourable variance, $9.2 million, in particular being driven by low levels of recorded comprehensive claims activity reported in the current quarter. Favourable recent accident year third party liability is driven by reported bodily injury case reserve reductions across multiple member companies, partially offset by unfavourable older accident year bodily injury (AY2009-AY2011) claims settlements. Alberta Grid RSP Third Party Liability Accident Benefits Other Coverages Total Actual Paid Actual Paid Actual Paid Actual Paid Paid Claims Paid Claims Paid Claims Paid Claims in in in in Accident [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] Year =[14][13] =[17][16] =[20][19] =[13]+[16]+[19] =[14]+[17]+[20] =[23][22] 2013 & Prior 5,078 5, (8) (131) 5,220 5, ,621 1,167 (454) ,631 1,297 (334) ,302 2, (64) 10 2 (8) 2,442 2, ,806 1,917 (889) (20) 8 (12) (20) 2,967 2,038 (929) ,564 3, (96) (155) (285) (130) 3,966 3,787 (179) ,412 5,309 (103) ,362 6,273 (1,089) 13,519 12,332 (1,187) Total 20,783 20,057 (726) 1,608 1,444 (164) 7,354 6,092 (1,262) 29,745 27,593 (2,152) 2017 & prior 15,371 14,748 (623) (169) (8) (181) (173) 16,226 15,261 (965) *projected paid claims based on Paid to Ultimate emergence model as at 2018Q2 As indicated above, total paid emergence at $27.6 million was $2.2 million (7.2%) less than the $29.7 million projected. Additional detail and summary charts akin to those found in the monthly Actuarial Highlights are presented in the sections that follow. There are constraints in this analysis in that full 25-quarter levels are not available for all metrics considered, as the detail has not all been fully captured and entered into the Facility Association databases (for instance, projected recorded and paid activity prior to 2013 ). D.5.1 AvsP: Indemnity & Allowed Claims Expense Actual recorded activity (paid and case reserve changes) over the last 25-calendar quarters is shown in the charts at the top of the next page, including the prior 24 quarter average level. page 30 of 74

32 Alberta Grid RSP Actual by Calendar Quarter 18,000 16,000 14,000 12,000 10,000 8,000 6,000 4,000 2,000 Alberta Grid PAYs (latest prior 24 qtr avg = 8,554) Alberta Grid CAY (latest prior 24 qtr avg = 23,461) PAYs Actual prior 24 qtr avg CAY Actual projected amount 40,000 35,000 30,000 25,000 20,000 15,000 10,000 5,000 prior 24 qtr avg projected amount activity variances from the previous quarter s projections are shown in the charts immediately below, including the prior 24-quarter standard deviation levels. Alberta Grid RSP Actual vs Summary: Variances 25 by Calendar Quarter 8,000 6,000 4,000 2,000 (2,000) (4,000) (6,000) (8,000) Alberta Grid PAYs (latest prior 24 qtr std dev = +/3,660) Alberta Grid CAY (latest prior 24 qtr std dev = +/4,731) PAYs Actual less prior 24 qtr std dev CAY Actual less prior 24 qtr std dev On Latest $ thousands PAYs CAY Actual less 8,554 23,461 std dev 3,660 4,731 AP <> std dev 7 3 % <> std dev 30.4% 13.0% norm <> std dev 31.7% 31.7% 15,000 10,000 (5,000) (10,000) (15,000) With respect to recorded indemnity & allowed claims expense, the prior accident years (PAYs) variances (left chart above) fell outside of one standard deviation 30% of the time (where projections are available), suggesting the projection process has performed no better than projecting the prior 24-quarter average amount. The variances show that actuals have been generally higher than projected for the available projections but the magnitude of the variances have not necessarily been extremely high and bias 26 has not been indicated at the 95% confidence level on a lagging 23-quarter basis, with 16 times in the past 23 quarters where actuals were higher than our projections for the PAY recorded amount. While there may be various causes for this as outlined in Section B.4, we believe the main driver of these variances is that the prior valuation selections of ultimate have proven, in hindsight, to be deficient to some degree. 5, Projections are available only for valuation quarters 2013 and beyond. For purposes of presentation, we have set the actual vs projection variances to 0 for all prior periods. 26 For the binomial distribution with 23 trials and an assumed 50% success probability, the 95% confidence range is 7 to 16. That is, for the 23 quarters presented, if the recorded projection was unbiased, with a 95% confidence, we would expect between 7 to 16 variances above 0. Less than 7 variances above 0 would indicate that our projections are biased high, and greater than 16 variances would indicate that our projections are biased low. page 31 of 74

33 The PAY recorded variance fell outside of the one standard deviation band during the latest quarter. The variance was reviewed and confirmed, with the variance attributed to process variance. The current accident year ( CAY ) recorded variances (right chart middle of the prior page) fell outside of one standard deviation 13% of the time (where projections are available), suggesting that the projection process has performed better than simply projecting the prior 24-quarter average amount. However, there does appear to be evidence of bias in the projection process (as actuals were higher than projections 17 times in the past 23 quarters). This result would support the view that our historical selections of ultimate, in hindsight, were deficient to some degree. The CAY recorded variance fell outside of the one standard deviation band during the latest quarter. The variance was reviewed and confirmed, with the variance attributed to process variance. We have included, for reference, additional charts immediately below related to levels influencing recorded activity. Alberta Grid RSP Levels that influence 27 activity by Calendar Quarter Alberta Grid PAYs (latest prior 24 qtr avg = 74,944,238) Alberta Grid CAY (latest prior yr = 111,395) PAYs Beginning IBNR prior 24 qtr avg CAY Ending YTD Earned Premium 140, , ,000 80,000 60,000 40,000 20, , , , , , ,000 80,000 60,000 40,000 20,000 Alberta Grid PAYs (latest prior 24 qtr avg = 11.4%) Alberta Grid CAY (latest prior 24 qtr avg = 30.8%) 20.0% 18.0% 16.0% 14.0% 12.0% 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% PAY as % of Beginning IBNR prior 24 qtr avg projected amount 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0% CAY as % CAY YTD Earned Premium prior 24 qtr avg projected amount We track beginning prior accident years IBNR as recorded activity comes out of IBNR. Changes in the prior accident years beginning IBNR (see upper left chart above) occur for several possible reasons: to offset actual recorded activity (through loss ratio matching); the annual switchover as a current accident year becomes a prior accident year (occurs in January); and IBNR levels potentially change with each new valuation. 27 Our recorded activity projections are based on selected ratios of life-to-date recorded activity to ultimate, converted to a recorded to beginning IBNR ratio, where ratio selection is based on our review of historical results. We find it helpful to show CAY emergence relative to earned premium for the purposes of the AvsP discussion. page 32 of 74

34 D.5.2 AvsP: Paid Indemnity & Allowed Claims Expense The charts immediately below show actual paid activity in each of the most recent 25 calendar quarters, along with a prior 24-quarter average to show how each quarter s actual compares with the average amount of the preceding 24 calendar quarters. Alberta Grid RSP Actual Paid activity by Calendar Quarter 35,000 30,000 25,000 20,000 15,000 10,000 5,000 Alberta Grid PAYs (latest prior 24 qtr avg = 21,147) Alberta Grid CAY (latest prior 24 qtr avg = 11,307) PAYs Actual Paid prior 24 qtr avg CAY Actual Paid projected amount 20,000 18,000 16,000 14,000 12,000 10,000 8,000 6,000 4,000 2,000 prior 24 qtr avg projected amount The charts immediately below show the actual less projected paid variances for the last 25 calendar quarters, along with bands for the prior 24-quarter standard deviations to show how the variances from projection compare with historical standard deviations. Alberta Grid RSP Actual vs Summary: Paid Variances 28 by Calendar Quarter 20,000 15,000 10,000 5,000 (5,000) (10,000) Alberta Grid PAYs (latest prior 24 qtr std dev = +/4,649) Alberta Grid CAY (latest prior 24 qtr std dev = +/3,820) PAYs Actual less Paid prior 24 qtr std dev CAY Actual less Paid prior 24 qtr std dev On Latest $ thousands Paid PAYs CAY Qtrly Avg Paid (prior 24 qtrs) 21,147 11,307 std dev 4,649 3,820 AP <> std dev 5 % <> std dev 21.7% 0.0% norm <> std dev 31.7% 31.7% With respect to paid indemnity & allowed claims expense prior accident years variances (left chart above), 22% of the variances (where projections are available) have fallen outside of one standard deviation, suggesting the projection process has performed better than projecting simply based on the preceding 24-month average. With 13 times of the past 23 quarters where actuals were higher than projected, there does not appear to be evidence of bias in the projection process, in particular no bias over the more recent periods. The current accident year paid variances (right chart above) have not fallen outside of one standard deviation (where projections are available), suggesting that the projection process has performed better than simply projecting the prior 24-quarter average amount. There does not appear to be evidence of bias 28 Projections are available only for valuation quarters 2013 and beyond. For purposes of presentation, we have set the actual vs projection variances to 0 for all prior periods. 8,000 6,000 4,000 2,000 (2,000) (4,000) (6,000) (8,000) page 33 of 74

35 in the projection process over the more recent periods, with 13 times of the past 23 quarters actuals being higher than projected. We have included, for reference, additional charts immediately below related to levels influencing paid activity. Alberta Grid RSP Levels that influence 29 Paid activity by Calendar Quarter Alberta Grid PAYs (latest prior 24 qtr avg = 252,932) Alberta Grid CAY (latest prior yr = 111,395) 350, ,000 PAY Beginning Unpaid prior 24 qtr avg IBNR case 200, , ,000 CAY Ending YTD Earned Premium 250, , , , , ,000 80, ,000 50,000 60,000 40,000 20,000 Alberta Grid PAYs (latest prior 24 qtr avg = 8.3%) Alberta Grid CAY (latest prior 24 qtr avg = 13.2%) 12.0% PAY Paid as % Beginning Unpaid prior 24 qtr avg projected amount 20.0% CAY Paid as % CAY YTD Earned Premium prior 24 qtr avg projected amount 10.0% 18.0% 16.0% 8.0% 14.0% 12.0% 6.0% 10.0% 4.0% 8.0% 6.0% 2.0% 4.0% 2.0% 0.0% 0.0% We track beginning prior accident years unpaid balance (case and IBNR) as paid activity comes out of the unpaid balance. Changes in the prior accident years beginning unpaid balance (see upper left chart above) occur for several possible reasons: to offset actual paid activity (may reduce case or IBNR or both); the annual switchover as a current accident year becomes a prior accident year (occurs in January); and IBNR levels potentially change with each new valuation. D.6 Current valuation IBNR selections Exhibit B.1.1 (see section L for all exhibits) summarizes the overall change in ultimate with this valuation and B.1.2 shows selected loss ratios over the most recent 4 valuations for comparison purposes on an all coverages basis. The B.2 exhibits provide information for third party liability, B.3 exhibits for accident benefits, and B.4 exhibits for the other government line. 29 Our paid projections are based on selected ratios of life-to-date paid activity to ultimate, converted to a paid to beginning unpaid ratio, where ratio selection is based on our review of historical results. We find it helpful to show CAY emergence relative to earned premium for the purposes of the AvsP discussion. page 34 of 74

36 D.7 Premium Liabilities / Future Accident Years In order to provide a basis for estimating the full premium liability level for monthly statements (i.e. the level of premium deficiency liability / deferred policy acquisition cost asset to carry) we leverage the a priori loss ratios for the accident year underlying the unearned premium levels. The test of recoverability leverages assumptions set by the Appointed Actuary. These include the member expense allowances (taking into account the Board approved allowances) and policy administration / maintenance expense assumptions. D.8 Actuarial Present Value Adjustments D.8.1 Selected Claims Payment Patterns Payment patterns are selected through the emergence models (the same used for projecting future claims paid and recorded activity for the AvsP process), leveraging a paid to ultimate metric. D.8.2 Selected Discount Rate The projected future claims paid cash flow are matched to a simulated portfolio of Government of Canada benchmark monthly bonds (yields anchored to the valuation date), and 15 basis point investment expense is assumed. A discount rate of 2.28% per annum was selected for the valuation of the claim liabilities and premium liabilities at September 30, 2018, up from 1.87% selected with the June 30, 2018 valuation. The chart to the right shows the Government of Canada benchmark bond yield curves at June 2018 and September Sensitivity to the discount rate assumption is presented in Exhibit C (see section L). Annual Effective Yield 3.00% 2.50% 2.00% 1.50% 1.00% 0.50% 0.00% Government of Canada Benchmark Bond Yields Jun18 Sep Duration (mths) D.8.3 Selected Margins for Adverse Deviations The margin for adverse deviation ( MfADs ) for investment income was maintained at 25 basis points with the current valuation. Our usual process is to review and update the claims development MfADs annually with the June 30 valuation. However, with the current valuation (as at September 30, 2018), the Alberta Grid (third party liability bodily injury) claims development MfADs were reviewed based on discussion and feedback with the FA s Actuarial Committee, but no changes were made to the selections from the prior valuation (as at June 30, 2018) and these are summarized in Exhibit D (see section L). D.9 Special IBNR Provisions / Adjustments There are currently no special IBNR provisions or adjustments. page 35 of 74

37 E. ALBERTA NONGRID RSP E.1 Valuation Highlights A summary of the valuation results through time is available in the A exhibit (see section L for all exhibits), with detail related to the current valuation provided in the B.1.1 and B.1.2 exhibits. Throughout 2016 and 2017, a member has been conducting a comprehensive review of certain claims transactions as reported to the FA RSP system. As discussed in the December 2017 Quarterly Actuarial Highlights, a balance sheet adjustment was applied to selected nominal implementation loss ratios to offset an incorrectly reported member case reserve overstatement, with the resulting impact implemented with the March 2018 Operational Report (based on the December 31, 2017 valuation). Prior to the current valuation (as at September 30, 2018), the member posted case reserve corrections. The submitted corrections were reviewed by FA management and determined to be reasonably aligned with the -$5.0 million associated Alberta Non-Grid RSP nominal member case reserve adjustment previously held, which have been released resulting in a neutral implementation impact. The nominal valuation results and exhibits presented in this section and in Appendix 5 do not include the impact of balance sheet IBNR adjustments related to incorrect member case reserve reporting included with the current and prior valuations (introduced with the December 31, 2017), i.e. the impact of the incorrect member case reserve reporting flow through these exhibits as the claims transactions are correctly reported to the FA RSP system. Including the member corrections reported in the quarter, the change in selected ultimate for prior accident years was $9.4 million favourable with this valuation (6.3% of the unpaid estimate as at last quarter), bringing the calendar year-to-date total favourable to $13.1 million (7.0% of the unpaid estimate as at the beginning of the 2018 calendar year). These changes are presented by accident year and government line in the tables below. Alberta NonGrid RSP valuation changes in selected ultimate Alberta NonGrid RSP valuation changes in selected ultimate (favourable) / unfavourable during Quarter (favourable) / unfavourable YTD Accident Year Third Party Accident Other Third Party Accident Other Total Accident Year Liability Benefits Coverages Liability Benefits Coverages Total 2013 & Prior 132 (1) (13) & Prior (207) (4) (104) (315) 2014 (509) (9) (6) (524) 2014 (860) (763) 2015 (1,730) (45) (345) (2,120) 2015 (2,801) (59) (312) (3,172) 2016 (632) (212) (1,247) (2,091) 2016 (2,369) (40) (746) (3,155) 2017 (2,462) (148) (2,152) (4,762) 2017 (2,781) 264 (3,156) (5,673) TOTAL (5,201) (415) (3,763) (9,379) TOTAL (9,018) 222 (4,282) (13,078) Excluding the impact of the member corrections discussed earlier (which had a neutral valuation implementation impact), the Alberta Non-Grid RSP favourable prior accident year development would have been $4.3 million, 3.0% of beginning unpaid (as determined with the June 30, 2018 valuation), and was driven by low levels of third party liability bodily injury recorded claims activity impacted by reported case reserve reductions across multiple member companies. The selected loss ratios for accident year 2018 (current accident year; AY2018 down 2.8 points to 109.3%) and accident year 2019 (future accident year; AY2019 down 1.8 points to 107.5%) both decreased. Change in selected loss ratio for AY2019 were driven by updated a priori loss ratio written premium projection assumptions to reflect the Alberta grid rate mandatory coverages increase effective page 36 of 74

38 January 1, 2019 (AIRB Bulletin ), assuming a partial increase in non-grid mandatory coverage rates and some risks previously transferred to the Alberta grid pool being transferred to the non-grid pool. Summary descriptions of recent regulatory and legislative initiatives are available in section I. The valuation process is described in more detail in section K, and a summary of changes to the process during this fiscal year is provided in section H. Policy liability projected cash flows and September 2018 government of Canada bond yields were used to determine the applicable discount rate. The selected investment income margin for adverse deviation ( MfAD ) was maintained at 25 basis points with the current valuation. With the current valuation (as at September 30, 2018), the Alberta Non-Grid (third party liability bodily injury) claims development MfADs were reviewed based on discussion and feedback with the FA s Actuarial Committee, but no changes were made to the selections from the prior review (as at June 30, 2018). See Exhibit D in section L for selected margins. E.2 Booked results for the prior valuation implementation It is helpful to consider how the portfolio looked after the prior valuation was implemented. In this case, the August 2018 booked results were based on assumptions derived from the prior (June 30, 2018) valuation and were discussed in the associated monthly Actuarial Highlights. The charts immediately below show the associated levels of claim liabilities 30 booked by accident year 31. The left chart displays life-to-date payments, case reserves, IBNR, and the total including actuarial present value adjustments against accident year earned premium. The right chart shows the associated dollar amounts for the components of the claim liabilities and the then-current projected amount of 2018 full year earned premium (the red hash-mark line) to provide some perspective % Alberta NonGrid Accident Year Loss Aug 31, 2018 $ millions 120 Alberta NonGrid Accident Year Unpaid Claim Aug 31, 2018 % proj EP apv adj.: 11% nominal unpaid: 178% 120.0% % % % % % (20) Paid Indemnity & Allowed Claims Expense Ratio Case Ratio IBNR (nominal) Ratio Ultimate Indemnity & Allowed Claims Expense (M/S) as % EP Case Reserves IBNR (nominal) APV Adjustment (M/S) proj EP M/S refers to Member Statement values that is, actuarial present value adjustments at the selected discount rate. The tables at the top of the next page show the associated Member Statement (M/S) policy liabilities. 30 Claim liabilities refer to provision for unpaid indemnity and allowed claims expenses. Allowed claims expenses are first party legal and other expenses as listed in the RSP Claims Guide. Claims expenses paid through the member company expense allowance are NOT included in this discussion. 31 Accident year 2004 was an incomplete year and therefore has been excluded from the loss ratio chart. page 37 of 74

39 claim liabilities ($000s) amt % case 133, % ibnr 54, % M/S apv adjust. 12, % M/S total 200, % premium liabilities ($000s) amt % unearned prem 59, % prem def/(dpac) 6, % M/S apv adjust. 3, % M/S total 69, % policy liabilities ($000s) amt % claim 188, % premium 65, % M/S apv adjust. 15, % M/S total 270, % E.3 Booked results for the current valuation implementation The October 2018 booked results were based on assumptions derived from the current (September 30, 2018) valuation and are discussed in the associated monthly Actuarial Highlights. The charts immediately below show the levels of claim liabilities booked by accident year on that basis. The left chart displays life-to-date payments, case reserves, IBNR, and the total including actuarial present value adjustments against accident year earned premium. The right chart shows the associated dollar amounts for the components of the claim liabilities and the current projected amount of 2018 full year earned premium (the red hash-mark line) to provide some perspective % Alberta NonGrid Accident Year Loss Oct 31, 2018 $ millions 120 Alberta NonGrid Accident Year Unpaid Claim Oct 31, 2018 % proj EP apv adj.: 10% nominal unpaid: 174% 120.0% % % % % % (20) Paid Indemnity & Allowed Claims Expense Ratio Case Ratio IBNR (nominal) Ratio Ultimate Indemnity & Allowed Claims Expense (M/S) as % EP Case Reserves IBNR (nominal) APV Adjustment (M/S) proj EP M/S refers to Member Statement values that is, actuarial present value adjustments at the selected discount rate. The tables at the top of the next page show the associated Member Statement (M/S) policy liabilities. page 38 of 74

40 claim liabilities ($000s) amt % case 127, % ibnr 58, % M/S apv adjust. 10, % M/S total 195, % premium liabilities ($000s) amt % unearned prem 60, % prem def/(dpac) 5, % M/S apv adjust. 3, % M/S total 68, % E.4 a priori loss ratios policy liabilities ($000s) amt % claim 185, % premium 65, % M/S apv adjust. 13, % M/S total 264, % This valuation s a priori loss ratios are summarized in the charts and table at the top of the next page (by government line and accident half-year), with a comparison to the last valuation (as at June 30, 2018) a priori loss ratios. We do not typically provide this detail for roll-forward valuations, as a priori loss ratios are generally not updated with those valuations. However, as a grid rate increase (effective January 1, 2019) was approved since our last valuation, we opted to update the a priori loss ratios to take this into account, assuming a partial increase in non-grid mandatory coverage rates and some risks previously transferred to the Alberta grid pool being transferred to the non-grid pool. page 39 of 74

41 The valuation results used to form the basis of the a priori estimates for the September 30, 2018 valuation continued to use selected ultimates from the March 31, 2018 valuation (consistent with the prior valuation a priori loss ratio assumptions). Trend structure models based on industry indemnity results as at December 31, 2017 were used (again, consistent with the prior valuation a priori loss ratio assumptions). Written premium projection assumptions were updated to reflect the Alberta grid rate mandatory coverages increase effective January 1, 2019 (AIRB Bulletin ), assuming a partial increase in non-grid mandatory coverage rates and some risks previously transferred to the Alberta grid pool being transferred to the non-grid pool. The current valuation Alberta Non-Grid RSP a priori loss ratios are presented in the B.1.4, B.2.3, B.3.3, and B.4.3 exhibits in section L. page 40 of 74