Actuarial Highlights FARM Valuation as at December 31, Ontario Alberta. Facility Association Actuarial 11/9/2012

Transcription

1 FARM Valuation as at December 31, 2011 Ontario Alberta Facility Association Actuarial 11/9/2012

2 Contents A. Executive Summary... 3 B. General Information... 7 B.1 Transition to Hybrid Model for Actuarial Services...7 B.2 Intended Audience and Use...7 C. Ontario Private Passenger... 8 C.1 Claims Liabilities excluding Actuarial Present Value Adjustments...8 C.1.1 Prior Accident Years (2010 & prior)...9 C.1.2 Current Accident Year (2011) C.2 Premium Liabilities / Future Accident Years C.3 Special IBNR Provisions / Adjustments C.3.1 HST Adjustments C.3.2 Product Reform Adjustments D. Ontario Non-Private Passenger D.1 Claims Liabilities excluding Actuarial Present Value Adjustments D.1.1 Prior Accident Years (2010 & prior) D.1.2 Current Accident Year (2011) D.2 Premium Liabilities / Future Accident Years D.3 Special IBNR Provisions / Adjustments D.3.1 HST Adjustments D.3.2 Product Reform Adjustments E. Alberta Private Passenger E.1 Claims Liabilities excluding Actuarial Present Value Adjustments E.1.1 Prior Accident Years (2010 & prior) E.1.2 Current Accident Year (2011) E.2 Premium Liabilities / Future Accident Years F. Alberta Non-Private Passenger F.1 Claims Liabilities excluding Actuarial Present Value Adjustments F.1.1 Prior Accident Years (2010 & prior) F.1.2 Current Accident Year (2011) F.2 Premium Liabilities / Future Accident Years G. Actuarial Present Value Adjustments G.1 Selected Claims Payment Patterns G.1.1 Selected Discount Rate G.1.2 Selected Margins for Adverse Deviations page 1 of 42

3 H. Appendix 1: Key Performance Metrics for Valuations I. Appendix 2: General Valuation Issues I.1 Measurement Uncertainty I.2 Policy Liabilities I.3 Actuarial Present Value Adjustments I.3.1 Selected Discount Rate I.3.2 Selected Margins for Adverse Deviations page 2 of 42

4 A. Executive Summary We have completed a valuation for the Facility Association Residual Market ( FARM ) for private passenger ( PPV ) and non-private passenger ( non-ppv ) business segments in Ontario and Alberta as at December 31, 2011, with the results summarized in the table below with respect to claims indemnity amounts. PAY ultimate change % beginning unpaid 2011 loss ratio change: Ontario PPV (5.1 million) (3.0%) Ontario non-ppv 0.7 million 1.0% Alberta PPV 1.1 million 5.2% Alberta non-ppv 0.5 million 1.3% loss ratio change: As indicated above, only Ontario private passenger had an overall favourable prior accident years change from the last valuation. In addition to the ultimate selection changes for the prior accident years, loss ratios for both 2011 and 2012 were updated with the 2011 loss ratios increasing for all segments reviewed, while the 2012 increased only for Ontario private passenger. For this valuation, we did not perform a separate analysis on the selection of ultimate for the allowed claims adjustment expenses (i.e. legal costs in excess of applicable deductibles and professional fees). Rather, we brought forward the selected ultimate ratios for these expenses from the prior valuation (September 30, 2011). A separate analysis of the retroactive claims fee adjustment was not completed as part of this valuation although the provision for the retroactive claims fee adjustment is updated monthly as appropriate to reflect the most current estimate of ultimate loss ratio and taking account of retroactive claims fee payments made to date. For actuarial present value adjustments, the discount rate was updated from 1.22% to 1.53% for all jurisdictions and business segments, including those not directly included in the valuation. No changes were made to margins for adverse deviation. More information on the actuarial present value adjustments is available in section G. Ontario Selected Ultimate Loss Ratios (indemnity only) by Business Segment Please note: the chart axes are at different scales page 3 of 42

5 Ontario Selected Unpaid Claims Liability (indemnity only) by Business Segment, including development PfAD Please note: the chart axes are at different scales While the charts above only display accident years 1994 to 2011, there remained unpaid balances (case reserves) prior to accident year 1994 as at December 31, In particular for private passenger, there were case reserves for accident years inclusive at $1.6 million. For non-private passenger, case reserves for accident years inclusive were at $0.5 million. The above charts also include the PfADs for development (nominal basis) as well as the 2011 earned premium to provide context for the unpaid balances against current earned premium. In the upper right of the charts, the development PfAD and nominal unpaid claims balances are shown as a percentage of earned premium to provide potential future operational result context (for instance, due to release of PfAD and/or changes in ultimate). The charts also highlight the impact of the rapid change in exposures from 2002 to 2006 (as measured by earned premium - see the orange dashed line in the charts below). In particular, private passenger continues to have a significant amount of unpaid claims levels for accidents relative to a typical distribution of unpaid claims liability by accident year. Ontario Earned Premium and Selected Selected Claims Amounts by Business Segment Please note: the chart axes are at different scales More detailed information around the results for Ontario can be found in sections C and D. page 4 of 42

6 Alberta Selected Ultimate Loss Ratios (indemnity only) by Business Segment Please note: the chart axes are at different scales Alberta FARM loss ratios have remained consistently low (below 50%) for both private passenger and non-private passenger over a number of years (2005 & 2006 being the most recent exceptions). For private passenger, the majority of exposures are subject to grid rates. The rather rapid private passenger deterioration in 2010 and 2011 bears watching. Alberta Selected Unpaid Claims Liability (indemnity only) by Business Segment, including development PfAD Please note: the chart axes are at different scales The charts above summarize the unpaid claims amounts and we include 2011 earned premium to provide context. Alberta private passenger has a small case reserve ($45 thousand) for accident year 1989 as at December 31, 2011, but otherwise, there were no unpaid balances for accident years 1996 and earlier. For nonprivate passenger, there were no unpaid balances for accident years 1994 and earlier. Rapid changes in earned premium volumes around 2002 leading up to reforms in 2004 have influenced the relative unpaid balances of the accident years relative to a typical book with moderate annual exposure changes (see earned premium through time in the charts below). For non-private passenger, earned premium levels peaked later and have decreased much more slowly than private passenger. page 5 of 42

7 Alberta Earned Premium and Selected Selected Claims Amounts by Business Segment Please note: the chart axes are at different scales More detailed information around the results for Alberta can be found in sections E and F. page 6 of 42

8 B. General Information This report summarizes the results of the valuation for indemnity amounts of the following FARM jurisdictions and business segments as at December 31, 2011: Ontario (private passenger and non-private passenger); and Alberta (private passenger and non-private passenger). The selected ultimate loss ratios from the prior valuation (September 30, 2011) were brought forward for the remaining jurisdictions and business segments. The selections of ultimate ratios for the allowed claims adjustment expense were similarly brought forward from the prior valuation for all jurisdictions and business segments. Retroactive claims adjustment provisions are derived based on formula as per the Plan of Operation. The results of this valuation were reflected for the first time in the March 31, 2012 Participation Reports for the FARM. As per historical practice, the updated selection of the discount rate is applicable to all jurisdictions and business segments. 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. Information that applies to FARM valuations in general is available in section H. This information will generally not change materially from one valuation to the next. For ease of reference, we will use the term claims amount in reference to the more proper and descriptive term indemnity and the terms loss ratio, claims ratio, or claims amount ratio in reference to the ratio of claims amount to earned premium. A valuation of ultimate selection for retroactive claims fees and for allowed claims expenses is not included in this discussion. B.1 Transition to Hybrid Model for Actuarial Services The Board of Directors of the Facility Association approved a plan in 2010 to move to a hybrid model in relation to actuarial services, whereby an in-house actuarial group would be formed to manage all actuarial services. Under this model, actuarial services would be completed both in-house and via use of the Facility Association s outside actuarial consulting resources. Cynthia Potts of Eckler Ltd. continues to be Facility Association's Appointed Actuary. The December 31, 2011 valuation work was completed by Eckler, using the same methodologies and the same approach to the selection of assumptions (including the level of data segmentation) as the September 30, 2011 valuation. The final valuation results are based on assumptions selected by the Appointed Actuary. 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 jurisdictions of the FARM in relation to the results of prior such valuations. It is not intended, nor necessarily suitable, for any other purpose. page 7 of 42

9 C. Ontario Private Passenger The left chart below presents life-to-date payments, case reserves, and nominal booked IBNR, together with this valuation s selection of ultimate claims amounts as at December 31, On the right, we show the associated dollar amounts, including a line indicating earned premium. This latter chart highlights the rapid change in earned premium from 2002 to During that period, claims amounts rose rapidly in 2003 but immediately began to decline so that the 2005 level was close to the 2002 level. This helps to explain why the loss ratio peaked in 2002, even though there was a rapid increase in both exposure levels and claims amounts. This also highlights the difficulty in ensuring relative rate adequacy across all rating variables during periods of rapid change. C.1 Claims Liabilities excluding Actuarial Present Value Adjustments The valuation did result in changes to the selections of prior accident years ultimate claims amount, as well as changes to the current accident year (2011) selected ultimate claims ratio (changes to the future accident year 2012 ultimate claims amount ratio are addressed in section C.2). The left chart below provides the components of the nominal unpaid claims liabilities, with the right chart showing the ratio of IBNR to case reserves. We note that this latter kpi for accident years 2001 to 2006 inclusive indicate a view that case reserves are significantly deficient and we are investigating (prior to 2001, the ratio to case is high due to low levels of case; where case is $0 and IBNR is not $0, we set the ratio to 100%). page 8 of 42

10 C.1.1 Prior Accident Years (2010 & prior) Throughout this section, results tables have accident years 2007 and prior grouped in a single row. Where annual earned premium changes are relatively stable, the most recent three prior accident years typically account for the majority (over 50%) of the unpaid balance. However, as clear from the chart in introductory remarks for this section, relatively stable changes in annual earned premium is not an accurate description for this portfolio. The most recent three prior accident years account for roughly 33% of the unpaid balance, and therefore do not account for most of the dollar amounts discussed. While our general focus will be on the most recent three accident years, the earlier accident years were reviewed separately, and are shown in the accompanying charts (back to 1994) and we will discuss them in detail as needed. The first table below summarizes the prior accident years selected ultimate claims amounts, the associated claims ratios, and the unpaid claims liability balances over the last five quarters of valuations. The valuation as at December 31, 2010 and March 31, 2011 included explicit special IBNR provisions to account for the additional impact of the following: July 2011 change in Ontario to a harmonized sales tax (HST) which expanded taxable services to include services consumed in the settlement of automobile insurance claims that were previously not subject to provincial sales tax; and September 2010 automobile insurance product reform. More information on these special provisions is available in section C.3, and the actual balances are provided in a later table. The next set of tables summarize change KPIs related to the above table. Commentary on the table contents and associated statistics is provided after the charts. page 9 of 42

11 Accident Year 2007 & prior Change during quarter in Selected Ultimate Claims Amount including "special" provisions ($000s) Change during quarter in Selected Ultimate LRs including "special" provisions Change duiring quarter in Select Ultimate Claims Amount as % beginning Unpaid Dec 2010 Mar 2011 Jun 2011 Sep 2011 Dec 2011 Dec 2010 Mar 2011 Jun 2011 Sep 2011 Dec 2011 Dec 2010 Mar 2011 Jun 2011 Sep 2011 Dec (4,236) (10,399) 1, (0.2%) (0.5%) 0.1% - (3.0%) (8.1%) 1.5% 0.1% ,526 (4,110) 2,559 (1,535) 2.3% (6.2%) 3.9% (2.4%) 7.2% (19.3%) 15.8% (8.7%) 2009 (614) (1,916) (226) (1,402) (1.2%) (4.0%) (0.4%) (2.9%) (2.2%) (7.6%) (1.0%) (6.8%) 2010 (1,892) (2,117) (1,117) (2,219) (4.6%) (5.3%) (2.8%) (5.6%) (6.1%) (8.1%) (5.0%) (11.0%) 2010 & prior (5,216) (18,542) 2,919 (5,051) (0.2%) (0.8%) 0.1% (0.2%) (2.3%) (9.2%) 1.7% (3.0%) change type total changes in amounts ($000s) by change change type year counts by change type change type % year counts by change type favourable (6,988) (19,245) (2,263) (5,949) favourable favourable 75.0% 66.7% 62.5% 60.9% neutral neutral neutral 4.2% - 8.3% 8.7% unfav'able 1, , unfav'able unfav'able 20.8% 33.3% 29.2% 30.4% change type largest changes for single accident year change type largest changes for single accident year change type largest changes for single accident year favourable (1,892) (4,110) (1,117) (2,219) favourable (4.6%) (6.2%) (2.8%) (5.6%) favourable <(999.9%) <(999.9%) <(999.9%) <(999.9%) AY: AY: AY: unfav'able 1, , unfav'able 2.3% 0.1% 3.9% 0.3% unfav'able >999.9% 121.6% 100.0% 35.7% AY: AY: AY: This quarter, 30% of the active accident years showed unfavourable variances. The amount of unfavourable change ($0.9 million) was overwhelmed by the amount of favourable change ($5.9 million) for the remaining accident years, resulting in an overall change that is $5.1 million favourable (3.0% of beginning unpaid). This change is within our standard +/-5% not unusual range relative to the beginning unpaid estimate. The larger changes in dollar terms tend to occur in recent accident years, as one would expect. However, we do note a bias toward favourable changes in the most recent accident years. With changes in selected ultimate come changes in the associated claims ratios. The claims ratio point changes can add some perspective to the amount changes. In this particular case, accident years 2007 and prior saw changes under 0.5 points, but accident years 2008 to 2010 inclusive saw changes greater than 2 points (with 2010 dropping over 5 points). Accident year 2010 showed the largest dollar change in two of the valuation quarters considered and had the largest favourable change in three of the four valuations. This is not particularly surprising, given that the most recent year is the least mature and generally has the largest unpaid liability. Accident year 2008 also shows up three times on the largest change table, showing the largest unfavourable change twice and largest favourable once. The chart above showing ultimate changes relative to the beginning unpaid balance in general shows the older accident years are within the +/-5% band this quarter. For the most recent three accident years, 2008 and 2009 show a change between -5% and -10%, but 2010 is slightly outside of the +/-10% band. page 10 of 42

12 Having discussed the changes in selected ultimate, our attention now moves to the associated selected IBNR levels for the current and the prior valuations with regard to the prior accident years, including special provisions previously mentioned. Further commentary follows on the changes in selected ultimate claims amounts in relation to claims activity and changes in our IBNR selections. As discussed in section H, it is possible to retrospectively determine the level of redundancy or deficiency in the booked IBNR level in comparison to the subsequent valuation as at the same date. This is summarized below: Generally, selected IBNR levels for a particular prior accident year are expected to decrease over time, as claims transactions are recorded and ultimately settled and closed, and this is generally the case as indicated in the first section above. We also expect that incremental recorded claims amounts will usually be positive amounts, otherwise either indicating there have been recoveries on payments made (for salvage and subrogation) or that prior case reserves have proven redundant overall (i.e. case reserve take downs are higher than claim payments). As indicated below, negative amounts for incremental recorded claims indemnity have shown up frequently during the recent quarters. Partial triangles for the three most recent accident years are displayed and discussed below, focused on accident year 2010 in relation to 2008 and page 11 of 42

13 The cumulative recorded claims ratio for 2010 initially behaved liked 2009 but more recently has looked like 2008, due to changes in the most recent incremental activity as shown below. The first focuses on dollar amounts, while the second normalizes for differences in earned premium by accident year. We have highlighted in both the chart above and below where the change in incremental recorded claims activity for accident year 2010 appears to have taken place i.e. during the first accident quarter with product reform. Not surprisingly, accident benefits coverage saw the most variance in the incremental changes against earned premium. We have summarized some additional KPIs in the charts below. page 12 of 42

14 The left chart shows the change in recorded claims amount as a percentage in the change in ultimate. The ultimate selection at any given valuation anticipates future changes in recorded activity. As such, a change in recorded activity itself may not lead to a change in the ultimate selection however, if the change in recorded is significantly different than anticipated in the prior valuation, this may lead to a change in ultimate. To the extent that this kpi is not 0%, it may indicate a variance in change in recorded from anticipated that may have contributed to a change in ultimate selection. (We are working on developing a kpi of actual vs. projected change in recorded as a % of change in ultimate.) The two right most charts show the booked redundancy / (deficiency) level both in dollar terms and as a percentage of the unpaid balance that was booked as at December 31, These charts clearly show that the older accident years (2001 and prior) showed a higher redundancy / (deficiency) ratio volatility, whereas the more recent accident years showed a higher dollar level of redundancy / (deficiency). These patterns are not unusual. In addition to the quarter-on-quarter view, we also consider the change over the last year (that is, comparing the differences at December 31, 2011 with the values as at December 31, 2010): There were fewer unfavourable years on the year-over-year basis, and the associated total unfavourable amount was well below the favourable year amounts, resulting in a $26.4 million favourable change. Accident year 2010 had both the largest amount of change and the biggest change in loss ratio (the largest changes relative to beginning balances have been older years, with small dollar changes applied to much smaller beginning balances). Accident year 2010 s selected ultimate loss ratio has dropped from 106% to 87%. page 13 of 42

15 The charts associated with the above discussion are presented below: The far right chart provides a view of change over a twelve-month period relative to the beginning unpaid balance. Our rule of thumb is that claims development would exceed the development margin for roughly 1-in-4 accident years (based on the view that the margin moves the confidence level of the unpaid balance to the 75 th percentile). Given there are 23 active accident years, we would expect to see roughly 6 years with development greater than the beginning margin and we found 2, which is not necessarily outside a range of what to expect. That said, we believe it is unusual to see so many accident years showing favourable changes in ultimate at or in excess of the beginning margin for the more recent accident years and we are investigating ways to improve the selection process. C.1.2 Current Accident Year (2011) The table 1 below compares the current accident year determination last valuation with this valuation. The selected ultimate claims ratio for accident year 2011 increased 3.8 percentage points, with both the expected estimate loss ratio and incurred development estimate ratio deteriorating. There remains a significant gap between the two estimates, where the incurred development estimate is affected by the significant reduction in claims activity reported to us by the Servicing Carriers. We are seeing dramatic reductions in both paid and recorded cumulative activity particularly at 12 months as indicated in the table below. Contributing to the difficulty of determining a selection for accident year 2011 was the unusually mild fall/winter in southern Ontario we anticipate that this will also have a favourable impact on the 2012 results relative to the underlying a priori assumption. However, the mild weather in the fourth quarter does not explain why the levels appear low at 3, 6 and 9 months of age. 1 The weights in the table are approximate, as the actual selection of loss ratio is done at a government line level, that is, third party liability vs accident benefits vs all other. page 14 of 42

16 The tables below provide the associated government line detail. Third party liability claims recorded activity is lower than anticipated at each age; particularly as it was (and still is) anticipated that the reforms would lead to some accident benefits amounts shifting to third party liability. We don t believe what we are seeing can be attributed to the mild 2011 fall/winter season, as we were already seeing improvements at 6 months (June 30). Accident benefits recorded activity is lower than the reform impacts suggested, driving the large gap between the ultimate estimate based on development and on the a priori (i.e. expected) claims ratio. Further, it is difficult to anticipate the changes in development patterns that results from a change in the coverage claims mix between catastrophic claims and non-catastrophic claims. Our belief is that catastrophic claims 2 are not directly affected by the product reform and tend to develop at later ages, whereas the non-catastrophic claims are affected by the reforms (reducing their loss cost levels) and tend to develop earlier. All other things being equal, we would expect the later ages now to develop more, relative to earlier ages, than in the past (as the catastrophic claim amounts are developing relative to a lower recorded non-catastrophic claim base). Unfortunately, we do not have sufficient detail to attempt to quantify this potential change. As such, we continue to put higher weight on the expected claims estimate than we might have otherwise, given the age. 2 We may find that to the extent catastrophic claims apply to not-at-fault individuals, some of the catastrophic claim amounts move over to third party liability. page 15 of 42

17 The paid and recorded claims ratios for all other coverages seem low as well. Were it only at 12 months, it might support the view of the mild fall (Q4) but it is also low at 9 months (incorporating the summer quarter Q3). We may be seeing an impact of overall rate level change, with premium per exposure increasing faster than claims per exposure and leading to a reduced ratio. This is an area we will investigate in the future. C.2 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 establish a priori claims ratios for the accident year underlying the unearned premium levels. The selected accident year 2012 claims ratio increased from 63.2% to 69.7%. We use the 2012 ratio for accident year 2013 as well (applies to determining future costs associated with the unearned premium reserve). C.3 Special IBNR Provisions / Adjustments C.3.1 HST Adjustments The 2011 HST provision is estimated to be $0 at December 31, 2011, consistent with an assumption that applicable case reserves have been appropriately adjusted to include HST. C.3.2 Product Reform Adjustments The 2010 Product Reform provision is estimated to be $0 at December 31, 2011, consistent with an assumption that case reserves pertaining to accident years 2010 and prior will increasingly reflect the effect of product reforms during the first half of calendar year Accident year 2011 case reserves are assumed to fully reflect the product reforms. page 16 of 42

18 D. Ontario Non-Private Passenger The chart on the left below presents life-to-date payments, case reserves, and nominal booked IBNR, together with this valuation s selection of ultimate claims amounts as at December 31, The chart to the right shows the associated dollar amounts, including a line indicating earned premium. This shows a rapid increase in earned premium level from 2001 to a peak in 2004, with a subsequent reduction settling on a level that has been sustained over the last three years. D.1 Claims Liabilities excluding Actuarial Present Value Adjustments The valuation did result in changes to the selections of prior accident years ultimate claims amount, as well as changes to the current accident year (2011) selected ultimate claims ratio (changes to the future accident year 2012 ultimate claims amount ratio are addressed in section D.2). The components of the nominal claims liability are presented below. The chart on the right indicates a higher implied case deficiency for accident years D.1.1 Prior Accident Years (2010 & prior) Axis for right chart was limited to max of 200% Throughout this section, results tables have accident years 2007 and prior grouped in a single row. The most recent three prior accident years account for the majority (just over 50%) of the unpaid balance. As such, our general focus will be on the most recent three accident years. The earlier accident years were reviewed separately, and are shown in the accompanying charts (back to 1994) and we will discuss them in detail as needed. page 17 of 42

19 The first table below summarizes the prior accident years selected ultimate claims amounts, the associated claims ratios, and the unpaid claims liability balances over the last five quarters of valuations. The valuation as at December 31, 2010 and March 31, 2011 included explicit special IBNR provisions to account for the additional impact of the following: July 2011 change in Ontario to a harmonized sales tax (HST) which expanded taxable services to include services consumed in the settlement of automobile insurance claims that were previously not subject to provincial sales tax; and September 2010 automobile insurance product reform. More information on these special provisions is available in section D.3, and the actual balances are provided in a later table. The next set of tables summarize change KPIs related to the above table. Commentary on the table contents and associated statistics is provided after the charts. page 18 of 42

20 Axis for right-most chart limited to 120% This quarter, 50% of the active accident years showed unfavourable changes, and the amount of unfavourable change ($2.5 million) was greater than the amount of favourable change ($1.8 million) for the remaining accident years, resulting in an overall change that is $0.7 million unfavourable (1.0% of beginning unpaid). This change is within our standard +/-5% range relative to the beginning unpaid estimate. The most recent 3 prior accident years accounted for half of the 6 years with favourable changes, particularly accident year The majority of the unfavourable changes came from accident years 2002 to 2007 inclusive. With changes in selected ultimate come changes in the associated claims ratios. The claims ratio point changes can add some perspective to the amount changes. In general, the loss ratio changes are less than 1 point, although 2002, 2007, and 2010 fall outside of this band. For the left most chart, we have forced the axis to a maximum of 120% - the 2001 kpi was 681% as indicated in the table above the chart. Having discussed the changes in selected ultimate, our attention now moves to the associated selected IBNR levels for the current and the prior valuations with regard to the prior accident years, including special provisions previously mentioned. Further commentary follows below on the changes in selected ultimate claims amounts in relation to claims activity and changes in our IBNR selections. page 19 of 42

21 As discussed in section H, it is possible to retrospectively determine the level of redundancy or deficiency in the booked IBNR level in comparison to the subsequent valuation as at the same date. This is summarized below: Generally, selected IBNR levels for a particular prior accident year are expected to decrease over time, as claims transactions are recorded and ultimately settled and closed, and this is applies to the most recent 3 prior accident years. However, for 2007 & older, selected IBNR ended at December 31, 2011 at a level higher than at December 31, For these years, only 1995 and 2007 ended with less IBNR than was selected December 31, For the most part, this reflects changes in ultimates. We also expect that incremental recorded claims amounts will usually be positive amounts, otherwise either indicating there have been recoveries on payments made (for salvage and subrogation) or that prior case reserves have proven redundant overall (i.e. case reserve take downs are higher than claim payments). As indicated below, positive amounts for incremental recorded claims indemnity have generally occurred. Accident Year 2007 & prior Change in Recorded Indemnity ($000s) Chg in Recorded as % Change in Ultimate Booked Redundancy / (Deficiency) as % Booked Unpaid Dec 2010 Mar 2011 Jun 2011 Sep 2011 Dec 2011 Dec 2010 Mar 2011 Jun 2011 Sep 2011 Dec 2011 Dec 2010 Mar 2011 Jun 2011 Sep 2011 Dec , (978) % 21.9% 55.5% 17.8% 2.4% 0.2% (3.7%) 1.0% (9.0%) 2008 (147) (551) (171) % 65.4% 53.4% (308.8%) (3.0%) 1.0% (0.2%) 0.6% (0.6%) (134.7%) (134.7%) (124.4%) (50.3%) 4.8% - 1.6% 2.2% (1.0%) , ,829 (459) 895.0% (90.5%) 160.1% 30.9% 4.8% 7.4% 1.4% (0.4%) 0.5% 2010 & prior 4, , >999.9% (51.9%) (83.5%) 25.8% 2.7% 2.0% (1.2%) 0.8% (4.4%) The cumulative recorded activity by age as a percentage of earned premium is presented below, indicating a variance between 2008 and the later two years. This is obviously not related to product reform. We note that accident year 2008 had a relatively high loss ratio, and the variance we note below may relate to 2008 simply being a bad year. We have summarized some additional KPIs in the charts below. In order to maintain focus on the most significant changes, we have capped the axes for the left chart at +/-150%, and the right at +/-25%. page 20 of 42

22 Axis for left-most chart limited to +/-150%) Axis for right-most chart limited to +/-25%) In addition to the quarter-on-quarter view, we also consider the change over the last year (that is, comparing the differences at December 31, 2011 with the values as at December 31, 2010): 50% of active accident years had unfavourable changes on the year-over-year basis, but the associated total unfavourable amount was well below the favourable year amounts, resulting in a $1.7 million favourable change. Accident year 2008 had the most favourable change across all three metrics above. The charts associated with the above discussion are presented below. Accident year 2008 s result stands out in the two left charts below. Axis for right-most chart limited to +50%) page 21 of 42

23 The chart on the far right above indicates the changes relative to the beginning unpaid. We expect roughly ¼ of active accident years to show unfavourable development in excess of the development margin. While the chart does not show all accident years, 10 of 23 accident years showed unfavourable development in excess of the beginning margin. While high, this is not necessarily unusual (and tends to apply to older accident years where beginning balances were low). D.1.2 Current Accident Year (2011) The table 3 below compares the current accident year determination last valuation with this valuation. The selected ultimate claims ratio for accident year 2011 increased 1.2 percentage points. The main drivers of the change is the change in the development estimate and the relative weights applied to the estimates from the two methods (the Dec. 31 weights reflect weights applied at the government line level that are not uniform this forces the weight applied to (a) to be greater than 100% when the government line results are summed up). We are seeing cumulative recorded activity lower than the recent accident years at a similar age, but it is not significant enough to suggest a difference other than regular process variance. D.2 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 establish a priori claims ratios for the accident year underlying the unearned premium levels. The selected accident year 2012 claims ratio was decreased from 70.7% to 68.1%. We use the 2012 ratio for accident year 2013 as well (applies to determining future costs associated with the unearned premium reserve). 3 The weights in the table are approximate, as the actual selection of loss ratio is done at a government line level, that is, third party liability vs accident benefits vs all other. page 22 of 42

24 D.3 Special IBNR Provisions / Adjustments D.3.1 HST Adjustments The 2011 HST provision is estimated to be $0 at December 31, 2011, consistent with an assumption that applicable case reserves have been appropriately adjusted to include HST. D.3.2 Product Reform Adjustments The 2010 Product Reform provision is estimated to be $0 at December 31, 2011, consistent with an assumption that case reserves pertaining to accident years 2010 and prior will increasingly reflect the effect of product reforms during the first half of calendar year Accident year 2011 case reserves are assumed to fully reflect the product reforms. page 23 of 42

25 E. Alberta Private Passenger The chart on the left below presents life-to-date payments, case reserves, and nominal booked IBNR, together with this valuation s selection of ultimate as at December 31, The chart on the right shows the associated dollar amounts, including a line indicating the level of earned premium. The period leading into and immediately after the 2004 reforms saw a rapid change in the earned premium volumes. Criteria for FARM risks were tightened as well, helping to drop exposures (and earned premium) levels. Since the reforms in 2004, the majority of the FARM Alberta Private Passenger exposures are subject to Grid rates. E.1 Claims Liabilities excluding Actuarial Present Value Adjustments The valuation did result in changes to the selections of prior accident years ultimate claims amount, as well as changes to the current accident year (2011) selected ultimate claims ratio (changes to the future accident year 2012 ultimate claims amount ratio is addressed in section E.2). The charts below present kpis around the components of the claims liability. E.1.1 Prior Accident Years (2010 & prior) Throughout this section, results tables have accident years 2007 and prior grouped in a single row. Where annual earned premium changes are relatively stable, the most recent three prior accident years typically account for the majority (over 50%) of the unpaid balance. However, as clear from the chart in introductory remarks for this section, relatively stable changes in annual earned premium is not an accurate description for this portfolio. The most recent three prior accident years account for roughly 30% of the unpaid balance, and therefore do not account for most of the dollar amounts discussed. While our general focus will be on the most recent three page 24 of 42

26 accident years, the earlier accident years were reviewed separately, and are shown in the accompanying charts (back to 1994) and we will discuss them in detail as needed. The first table below summarizes the prior accident year selected ultimate claims amounts 4, the associated claims ratios, and unpaid claims liability balances over the last five quarters of valuations. During the periods included in the presentation, there were no special IBNR provisions required. Commentary on the table contents and associated statistics is provided below, after the charts. Accident Year 2007 & prior Change during quarter in Selected Ultimate Claims Amount including "special" provisions ($000s) Change during quarter in Selected Ultimate LRs including "special" provisions Change duiring quarter in Select Ultimate Claims Amount as % beginning Unpaid Dec 2010 Mar 2011 Jun 2011 Sep 2011 Dec 2011 Dec 2010 Mar 2011 Jun 2011 Sep 2011 Dec 2011 Dec 2010 Mar 2011 Jun 2011 Sep 2011 Dec (432) (0.1%) 0.1% - - (2.3%) 1.3% 1.1% 3.1% 2008 (172) (86) (0.7%) (0.4%) - 1.1% (12.8%) (7.8%) 1.0% 24.3% 2009 (45) 138 (160) (43) (0.2%) 0.8% (0.9%) (0.2%) (3.0%) 9.9% (11.0%) (3.4%) 2010 (73) (0.5%) 0.1% 0.8% 2.4% (1.6%) 0.4% 3.8% 12.1% 2010 & prior (723) ,060 (0.1%) 0.1% - 0.1% (2.8%) 1.2% 0.7% 5.2% change type total changes in amounts ($000s) by change change type year counts by change type change type % year counts by change type favourable (977) (845) (348) (409) favourable favourable 57.1% 57.1% 38.5% 23.1% neutral neutral neutral 21.4% 14.3% 23.1% 15.4% unfav'able 254 1, ,468 unfav'able unfav'able 21.4% 28.6% 38.5% 61.5% change type largest changes for single accident year change type largest changes for single accident year change type largest changes for single accident year favourable (430) (224) (160) (365) favourable (1.2%) (0.7%) (0.9%) (0.2%) favourable (18.5%) <(999.9%) <(999.9%) <(999.9%) AY: AY: AY: unfav'able unfav'able 0.2% 1.0% 0.8% 2.4% unfav'able 63.2% 17.4% 4.8% 24.3% AY: AY: AY: This quarter, 62% of the active accident years showed unfavourable variances, and their associated change amount ($1.5 million) overwhelmed the favourable years ($0.4 million), resulting in an overall change that was $1.1 million unfavourable (5.2% of beginning unpaid). This change in comparison with the unpaid claims liability balance selected last valuation is just outside our +/-5% not unusual range. The centre chart below shows how each individual accident year s selected ultimate claims ratio has changed with this valuation and the majority of the changes are within +/-1 point (2010 had the largest unfavourable movement at an additional 2.4 points). 4 The tables here will refer to amounts including special provisions. For Alberta private passenger, special provisions have been $0 for all accident years during this review period. page 25 of 42

27 Axis for right most chart limited to -100%) Having discussed changes to the selected ultimates, our attention now moves to the associated selected IBNR levels for the current and the prior valuations with regard to the prior accident years (as previously indicated, there were no special provisions required during the periods reviewed). Further commentary on the changes in selected ultimate claims amounts in relation to claims activity and changes in our IBNR selections follow. As discussed in section H, it is possible to retrospectively determine the level of redundancy or deficiency in the booked IBNR level in comparison to the subsequent valuation as at the same date. This is summarized below. Generally, selected IBNR levels for a particular prior accident year are expected to decrease over time, as claims transactions are recorded and ultimately settled and closed. This appears to be the pattern here. We also expect that incremental recorded claims amounts will usually be positive amounts, otherwise either indicating there have been recoveries on payments made (for salvage and subrogation) or that prior case reserves have proven redundant overall (i.e. case reserve take downs are higher than claim payments). As indicated below, positive amounts for incremental recorded claims indemnity have generally occurred. page 26 of 42

28 In addition to the quarter-on-quarter view, we also consider the change over the last year (that is, comparing the differences at December 31, 2011 with the values as at December 31, 2010): Accident Year 2007 & prior Change in Selected Ultimate Claims Amount including "special" provisions ($000s) year-to-year Change in Selected Ultimate LRs including "special" provisions year-to-year Change in Select Ultimate Claims Amount as % beginning Unpaid (year-over-year) Dec 2010 Mar 2011 Jun 2011 Sep 2011 Dec 2011 Dec 2010 Mar 2011 Jun 2011 Sep 2011 Dec 2011 Dec 2010 Mar 2011 Jun 2011 Sep 2011 Dec % 2008 (2) - (0.1%) 2009 (110) (0.5%) (7.3%) % 10.5% 2010 & prior % 2.9% change type total changes in amounts ($000s) by change year counts by change type type change type change type % year counts by change type favourable (1,352) favourable 10 favourable 76.9% neutral - neutral 2 neutral 15.4% unfav'able 2,120 unfav'able 2 unfav'able 15.4% change type largest changes for single accident year change type largest changes for single accident year change type largest changes for single accident year favourable (347) favourable (1.1%) favourable <(999.9%) AY: 2003 AY: 2007 AY: 2001 unfav'able 1,647 unfav'able 2.8% unfav'able 35.5% AY: 2005 AY: 2010 AY: 2005 As opposed to the quarter, the year-on-year view indicates only 15% (2) accident years with unfavourable changes; however, those changes in dollar amounts ($2.1 million), overwhelmed the favourable years ($1.4 million) resulting in an overall change that was unfavourable by $0.8 million or 2.9% of the beginning balance (this latter metric is within our +/-5% range). The centre chart below shows that the selected ultimate claims ratio changes have generally been within +/- 2 points. As previously discussed, we are looking at our processes in consideration of potential that selections are reduced early then later increased again. page 27 of 42

29 Axis for right-most chart limited to -100% The chart above on the far right shows changes year-over-year relative to the beginning unpaid balance, where the green dashed line indicates the beginning margin for adverse claims development. Of the 14 active accident years, we would expect roughly ¼ or 4 to show unfavourable development in excess of the beginning margin, and 2 years have done so which is within a reasonable range of our expectation. E.1.2 Current Accident Year (2011) The table 5 below compares the current accident year determination last valuation with this valuation. As indicated above, the incurred development estimate is much higher than the expected estimate and is also being given more weight, resulting in the deterioration of the selection of the ultimate claims ratio. The incurred estimate is influenced by the recorded claims ratio. The cumulative recorded claims amount as a percentage of earned premium by age is shown in the next table. Both 2010 and 2011 have shown higher levels of case reserves, leading to higher recorded levels than we ve seen in recent years, leading to the higher loss ratios indicated in the loss ratio chart at the beginning of this section, and in particular the higher incurred development estimate in the table above. 5 The weights in the table are approximate, as the actual selection of loss ratio is done at a government line level, that is, third party liability vs accident benefits vs all other. page 28 of 42

30 E.2 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 establish a priori claims ratios for the accident year underlying the unearned premium levels. The selected accident year 2012 claims ratio was decreased from 46.9% to 44.6%. We use the 2012 ratio for accident year 2013 as well (applies to determining future costs associated with the unearned premium reserve). page 29 of 42

31 F. Alberta Non-Private Passenger The chart on the left below presents life-to-date payments, case reserves, and nominal booked IBNR, together with this valuation s selection of ultimate as at December 31, The chart on the right shows the associated dollar amounts, including a line indicating the level of earned premium. The period leading-into and immediately-after the 2004 reforms saw a rapid change in the earned premium volumes which have only recently begun to decline in a meaningful way. F.1 Claims Liabilities excluding Actuarial Present Value Adjustments The valuation did result in changes to the selections of prior accident years ultimate claims amount, as well as changes to the current accident year (2011) selected ultimate claims ratio (changes to the future accident year 2012 ultimate claims amount ratio is addressed in section F.2). Additional kpis related to the claims liability are presented in the charts below. F.1.1 Prior Accident Years (2010 & prior) Throughout this section, results tables have accident years 2007 and prior grouped in a single row. Where annual earned premium changes are relatively stable, the most recent three prior accident years typically account for the majority (over 50%) of the unpaid balance. This is the case here, as the most recent three prior accident years account for roughly 55% of the unpaid balance. While our general focus will be on the most recent three accident years, the earlier accident years were reviewed separately, and are shown in the accompanying charts (back to 1994) and we will discuss them in detail as needed. page 30 of 42

32 The first table below summarizes the prior accident year selected ultimate claims amounts 6, the associated claims ratios, and unpaid claims liability balances over the last five quarters of valuations. During the periods included in the presentation, there were no special IBNR provisions required. Commentary on the table contents and associated statistics is provided below, after the charts. This quarter, 20% of the active accident years showed unfavourable variances, but their associated change amount ($1.1 million) overwhelmed the favourable years ($0.6 million), resulting in an overall change that was $0.5 million unfavourable (1.3% of beginning unpaid). This change in comparison with the unpaid claims liability balance selected last valuation is within our +/-5% range. Both accident year 2008 and 2010 appear 3 times in the largest change table for amounts and loss ratio points. The centre chart below shows how each individual accident year s selected ultimate claims ratio has changed with this valuation and all but one of the changes are within +/-1 point. 6 The tables here will refer to amounts including special provisions. For Alberta non-private passenger, special provisions have been $0 for all accident years during this review period. page 31 of 42

33 Axis for right most chart limited to -20% Having discussed changes to the selected ultimates, our attention now moves to the associated selected IBNR levels for the current and the prior valuations with regard to the prior accident years, (as previously indicated, there were no special provisions required during the periods reviewed). Further commentary on the changes in selected ultimate claims amounts in relation to claims activity and changes in our IBNR selections follow. As discussed in section H, it is possible to retrospectively determine the level of redundancy or deficiency in the booked IBNR level in comparison to the subsequent valuation as at the same date. This is summarized in the table below. Generally, selected IBNR levels for a particular prior accident year are expected to decrease over time, as claims transactions are recorded and ultimately settled and closed. The odd up-and-down pattern evident for the most recent three accident years as indicated above reflects changes in ultimate selections over time. We also expect that incremental recorded claims amounts will usually be positive amounts, otherwise either indicating there have been recoveries on payments made (for salvage and subrogation) or that prior case reserves have proven redundant overall (i.e. case reserve take downs are higher than claim payments). As indicated below, negative amounts for incremental recorded claims indemnity have occurred with some frequency during the recent quarters. page 32 of 42

34 In addition to the quarter-on-quarter view, we also consider the change over the last year (that is, comparing the differences at December 31, 2011 with the values as at December 31, 2010): The year-on-year view indicates 27% of accident years with unfavourable changes and those changes in dollar amounts ($0.8 million), were overwhelmed by the favourable years ($2.7 million) resulting in an overall change that was favourable by $2.0 million or 4.0% of the beginning balance (this latter metric is within our +/-5% range). The associated charts are below. The far right chart shows the change in ultimate year-over-year as a percentage of the beginning unpaid balance, and included as the green dashed line is the beginning margin for adverse claims development. In general, we expect roughly ¼ of active accident years will have unfavourable development in excess of the associated beginning margin over the course of a year. In this case, we expected roughly 4 and there were 2, which is within the range of reasonableness. page 33 of 42