Potential product reforms, such as raising the limit in the MIRon compensation for general damages.

Transcription

1 IBC -llo..*..til~ May 31,2013 B jc A I Insuran~e Bureau of Canada ~{~ ~ Bureau d assurance du Canada William A. Adams Vice-President, Western & Pacific Avenue, Suite 2603, Edmonton, AS T5J OH fax Burrard Street Suite 901, Vancouver, BC V6C 3A Ext fax: Mr. Alfred Savage Chair Automobile Insurance Rate Board 2440 Canadian Western Bank Place Jasper Avenue Edmonton, AB TSJ 3N6 Dear Mr. Savage, Insurance Bureau of Canada (IBC) is pleased to respond to the Alberta Automobile Insurance Rate Board's (AIRS) request for input regarding the 2013 review of the required premium for basic coverage beginning November 1, Attached is the actuarial analysis prepared by IBC's consulting actuary, Dr. Ron Miller of Insurance Services. IBC has and continues to express the view to governments and regulators throughout the country that, while the regulation of rates provides a degree of comfort to some stakeholders, in a competitive market the cost of auto insurance claims is the major factor in determining the price of the product. For example, premiums in Ontario are the highest in Canada, despite the province employing one of the most strict rate regulation regimes. The reason for Ontario's high premiums is high claims costs, which are the result of many years of excessive utilization of an overly rich no-fault benefit package that is susceptible to abuse and outright fraud. Since 2004, the price of auto insurance in Alberta has fluctuated only modestly from year-to-year because of the relative stability of the product's cost structure. On many occasions, the government and the industry have collaborated to help maintain this cost stability and overcome threats to it, such as the constitutional challenge to the Minor Injury Regulation (MIR). The result is one of the most affordable auto insurance products in the country. In 2011, Albertans, on average, spent 2.8% of their disposable income on auto insurance, compared to 3.2% in Atlantic Canada and 5.3% in Ontario. 1 While the Facility Association's market in Alberta remains the largest among the provinces it serves, it has been declining steadily and now stands at 6.1 %. 2 Cost Environment There are three related factors we want to highlight for their potential to have an impact on the cost of delivering automobile insurance in Alberta. They are: Changes in injury claims patterns; Changes in property damage claims patterns; and Potential product reforms, such as raising the limit in the MIRon compensation for general damages. 1 IBC with data from GISA and Statistics Canada. 2 Facility Association (FARM and RSP). Reprcscutiug the ro111pcmics that insure your ho111c, your cnr. your lms~ucss. Reprcswtcwt /cs socictt;s qui assrm ut votn hnbitatiou, votrc autolllobrlc, 1 otrc eutrcpnse 1

2 Changes in Injury Claims Patterns In preparation for last year's IWA, IBC carried out a survey of fifteen insurance companies operating in Alberta in an effort to determine whether any impact on claiming behaviour has been seen following the January 2012 court decision in Sparrowhawk v. Zapoltinsky. As you know, the Alberta Court of Queen's Bench ruled that the temporomandibular joint disorder (TMD) suffered by the claimant in this case was not appropriate for treatment under the Diagnostic and Treatment Protocols Regulation (DTPR) and was not a minor injury. Most of the companies surveyed by IBC reported that the number of TMD claims was increasing. This development concerned us because TMD is associated with relatively high settlement costs. Indeed, a closed claims study that IBC sponsored for the Alberta constitutional challenge found that, as far back as 2004, claims with TMD diagnoses were responsible for 18.4% of bodily injury settlement dollars paid by insurance companies, while representing only 7.7% of these claims. 3 We repeated our survey this year and were informed that this pattern is continuing. We also found that, along with TMD, companies are experiencing more Section 8 claims for psychological impairment and chronic pain, both of which some stakeholders consider to be outside the MIR's application. These claims, which typically take longer to close, are beginning to appear earlier in the claims management process. Insurers also reported that there appears to be a positive correlation between an injured person having legal representation and that person pursuing a claim for TMD, psychological impairment and/or chronic pain. In our view, these developments suggest that more stakeholders may be beginning to use the Section 8 process to assert those types of injuries (TMD, chronic pain, depression) that can take their claim out of the minor injury class in order to build the case for more lucrative bodily injury general damages awards. According to some of the companies we surveyed, there is already beginning to be evidence that more people are pursuing claims for general damages, outside the MIR, based on TMD, psychological impairment and/or chronic pain. This was reflected in a statement that appeared in the 2012 annual report of one large carrier in describing recent Alberta injury claims developments: "Due to the age of older minor injury claims and the changing environment, plaintiff lawyers are aggressively working files in order to build the case for their clients". 4 Changes in Property Damage Claims Patterns A legal decision from December 2012 threatens to have negative effects for property damage liability costs. In Taylor v. Hrytsak, the judge at the Provincial Court of Alberta awarded the plaintiff $17,000 after finding that a vehicle involved in a collision had experienced "diminished value". Already, companies report receiving more claims for compensation for the perceived diminished value of a vehicle after it has been repaired following a collision. This development is not surprising because the size of the award in the case cited established an attractive windfall target even for individuals with no plans to sell their repaired vehicles. Thus, this court decision has created an opportunity for claimants that, coupled with the developments since the decision in Sparrowhawk v. Zapoltinsky, leaves the auto insurance product vulnerable to a rising number of more expensive liability claims. Changes to the Minor lnjurv Regulation (MIR) IBC understands that, at the urging of certain stakeholders, one of the changes the government is considering, in its review of the regulations related to the auto insurance product, is a one-time 3 Alberta Closed Claims Study, May 2006, prepared by Barb Addie, F. C.I.A., Insurance Services. 4 Intact Financial Corporation. Annual Report

3 increase in the award limit for general damages in the MIR. The current limit has been subject to annual inflation indexation since Governments in Atlantic Canada recently reviewed their limits on general damages that may be awarded for minor injuries. In 2010, the Nova Scotia government raised the limit from $2,500 to $7,500, and indexed it to changes in the Consumer Price Index (CPI). The New Brunswick government has just reformed its regulation, raising the limit to $7,500 effective July 1, While industry data is not mature enough to show the full impact of raising the limit in Nova Scotia, preliminary data indicates that, as between collisions that occurred in 2009 and those happening in 2011, the cost per car for bodily injury claims rose 9%, from $ to $ We believe that this increase is, in part, simply the result of the higher non-pecuniary awards now available to those incurring minor injuries, but also that, to some extent, it may reflect an incentive effect that these higher awards have created. As we see in the actuarial analysis submitted with this letter, it appears that the statistical plan may already be reflecting cost pressures arising from the Sparrowhawk v. Zapoltinsky decision, as well as perhaps some measure of the typical erosion of the effectiveness of savings measures over time after a reform. In the case of the more recent Taylor v. Hrytsak decision, there is only anecdotal evidence to date from insurers that it may be changing the dynamics of property damage claims. Nonetheless, the fact that these new sources of cost pressure are clearly present suggests to us that, during this period in which the auto insurance market remains fairly stable, it would be highly prudent for the government to initiate an in-depth review of the claims cost environment in Alberta, and to work with stakeholders to develop solutions to the problems identified and ensure that the cost stability Albertans have come to expect continues into the future. IBC would welcome the opportunity to participate in these discussions. Commentary on the Industry-Wide Adjustment (IWA) On many occasions, we have expressed the view that a process reliant on competition is more appropriate than uniform adjustments for regulating the price of auto insurance. This view is based on several reasons: Uniform adjustments are more likely to disrupt the stability of the market. If a single company sets its premiums incorrectly in a competitive market, only that company feels the repercussions. But if an industry-wide indication is wrong, all companies and their consumers feel the effects. Experience in jurisdictions with rate regulation processes reliant on competition shows that premiums follow market conditions, and that competition tends to smooth out the impact that sudden changes in costs may have on prices. Strict regulation is costly. Last year, IBC conducted a Canada-wide study on the costs to the industry and its consumers that arise from funding the operations of the nearly 30 federal and provincial bodies having responsibilities for regulating insurance carriers and intermediaries. The results of the study showed that, in 2010, the direct cost that regulators charged to the industry was $80 million. It should be noted that this figure is, by no means, exclusive to auto insurance rate regulation, but rather covers the many governing bodies and types of regulation that the industry is subject to. Further, this estimate does not include the very substantial costs incurred by companies which need to use numerous expert personnel to carry out the activities required by regulators. In the case of auto insurance rate regulation, for example, depending upon the jurisdiction, 5 IBC with data from GISA. 3

4 companies can incur application fees, additional actuarial costs, and extensive staff time to comply with the different rate filing instructions used by different rate boards. As well, they incur the systems and human resources costs of reporting the additional data that rate regulators request, and the cost associated with the delay in getting indicated rates through the approval process and to the market. A more competitive rate-making environment encourages companies to differentiate themselves. This provides consumers with more choice through innovative product offerings and service models as companies compete more aggressively for the business of consumers. Last fall, IBC welcomed the opportunity to engage with the government and the AIRB in discussions on the Premiums Regulation. During that review, the industry proposed transitioning from an IWA process for basic coverage and a file-and-use process for optional coverage to a file-and-approve process for the entire product. The industry believes this model is more appropriate for a competitive market and will continue to provide a level of regulatory oversight that benefits consumers. Indication IBC does not endorse the assignment of a single rate adjustment indication for the entire industry. But because the IWA process calls for the provision of an indication, we retained Dr. Ron Miller of Insurance Services to conduct an actuarial analysis for this submission. On page 22 of Dr. Miller's enclosed report, the estimate given for the all-industry average of indicated required premium for basic coverage for November 1, 2013 is $670.82, while the projection of average street premium per car year, before annual adjustment, is stated as $ These numbers imply that the indicated all-industry rate change is 13.5%. This indicated rate increase results largely from the confluence of two factors: the first factor is the influence of rising costs for bodily injury claims. Dr. Miller found that bodily injury claims from previous years are closing at a higher cost than was initially predicted. In reviewing past exhibits from GISA, we have found, that because of this development, the ultimate cost per car for Third Party Liability (TPL) claims for accident-years 2008 to 2011 has been adjusted upwards. The annual adjustments are set out in the table below. Changes to Ultimate Cost per Car for TPL Claims for Accident-Years (AY) 2008 to Exhibit Year AY 2008 AY 2009 AY 2010 AY 2011 AY ~34:o2' /;.... '.(1~:9z:~;'!Ifl!lff 1 ijli~'{4~:1 :~,~ ;.9/'i~;;,,.',;l:f?~;Ji'1.~$i Tt'!EX:cl :G~ 442f51 '. < The table also shows that the largest increases happened between the 2011 and 2012 GISA publications. In fact, between the publication of the 2011 and 2012 exhibits, the estimated ultimate cost per car for TPL claims for accident-years 2008, 2009, 2010 and 2011 increased by 6 IBC with data from GISA. 4

5 an average of 8.1%. By comparison, the increase between the 2010 and 2011 exhibits for accident-years 2008, 2009 and 2010 was an average of 2.8% and the change between the 2009 and 2010 exhibits for accident-years 2008 and 2009 was an average of 0.4%. The substantial adjustments between the 2011 and 2012 publications coincide with the court decision in Sparrowhawk v. Zapoltinsky. A second factor contributing to the magnitude of the indicated rate increase for 2013 is related to previous decisions of the AIRB regarding price adjustments. As noted above, Dr. Miller estimates that the average street premium for the basic auto insurance product as of November 1, 2013, prior to implementation of the IWA adjustment, will be $ That the average street premium will be this low and the corresponding difference between street and indicated premium so wide, is in part a reflection of the Board's decision at last year's IWA. Thus, a year ago, the kinds of bodily injury cost pressures that are described in this letter and documented in the accompanying actuarial analysis were already very much in evidence. It was for this reason we understand that the Board's actuary, Oliver Wyman, identified an indicated adjustment for 2012 of +11.5%. However, the AIRB decided to permit an IWA of only +5% last year, leaving a sizable gap between the cost escalation trend, which was well under way, and the premium adjustment. Unfortunately, on an all-industry basis, the gap between average premium and costs has continued to grow over the past year. IBC is, of course, quite aware that not all companies have applied the +5% permitted adjustment. While companies have up to three years to apply the adjustment, we see the fact that some companies have not taken it as providing strong and positive evidence to support the industry's long-held position that a single, one-size-fits-all adjustment factor is not at all appropriate in a competitive market. It also shows that, even in a restrictive rate regulation context, companies' pricing decisions are based as much as they can be on each company's own book of business and market strategy. At the same time, however, we want to caution that individual company responses to past IWA directives should not take away from the requirements in the Premiums Regulation to decide on an all-industry adjustment that is reflective of industry-wide experience. Conclusion The principal messages we want to convey to the AIRB are summarized as follows: According to IBC's consulting actuary, the indicated rate adjustment for basic coverage, effective November 1, 2013 is 13.5%. We believe that the evidence, including Oliver Wyman's analysis and information from individual companies, will be sufficient to allow the AIRB to arrive at an adjustment decision that truly reflects the cost trends and projections that are put before it. In this regard, we believe that it will be particularly important that, on an all-industry basis, basic product rates are at or close to actuarial adequacy when the rate regulation system transitions to a more competitive process. There is little doubt that the pressures on bodily injury claims costs, as well as - albeit more recently - property damage tort claims, have become more pronounced. Consequently, we are urging the government to undertake an in-depth review of the factors affecting claims costs with a view to developing solutions that can protect the future stability of Alberta's automobile insurance market. Finally, we want to reiterate our support of the government moving forward with changes to the Premiums Regulation that permit competition to play a greater role in determining auto insurance premium levels. In our view, this will bring benefits to Alberta's driving public through lower premiums for good drivers, stronger incentives for safe driving, increasing product innovation, and more consumer choice. 5

6 We look forward to reviewing Oliver Wyman's analysis when it becomes available, and subsequently meeting with the AIRB on June 11, ~ William A. Adams Vice-President, Western & Pacific Enclosure (1) 6

7 Actuarial Analysis re Alberta Private Passenger Automobile Insurance Industry-Wide Rate Level Adjustment for Basic Coverage Effective November 1, 2013 submitted to Alberta Automobile Insurance Rate Board for its June 2013 Review prepared for Insurance Bureau of Canada by Insurance Services Inc. May 31, 2013

8

9 Table of Contents 1. Overview Introduction Scope My Approach Data Restatement Major Caveats Assumptions and Model Parameters in the Analysis Incurred Count and Incurred Amount Loss Development Factors Unallocated Loss Adjustment Expense Factors Trending Model Independent Variables in Trending Model Variable = Change in Level re the 2004 Reforms for TPL-BI Variable = Change in Level re the 2004 Reforms for Other Sub-Coverages Variable = Change in Trend re the 2004 Reforms for TPL-BI Variable = Change in Trend re the 2004 Reforms for Other Sub-Coverages Variable = Change in Level at 01/03/ Acceptance or Rejection of Independent Variables Summary of Findings re 2004 Reforms for Major Sub-Coverages Variable = Change in Level re 2004 Reforms Variable = Change in Trend re 2004 Reforms Summary of Findings re Change in Level at 01/03/2007 for Major Sub-Coverages Variable = Change in Level at 01/03/ Forward Projections Payment Pattern Fixed and Variable Expenses Alberta Health Levy Discounting for the Time Value of Money Forward New Money Pre-Tax Investment Rate re Cash Flow on Underwriting Discounting Process Profit and Contingency Margin Average Street Premium High Level Summary of Findings Required Rate Change Exhibit 1 - Required Premiums and Overall Rate Change for Basic Coverage Exhibit 2 - Trending Model and Forward Projections Exhibit 3 - Derived Accident Half-Year Payment Patterns... 37

10

11 1. Overview 1.1 Introduction Insurance Bureau of Canada (IBC) commissioned Insurance Services Inc. (EISI) to provide an actuarial analysis to the Alberta Automobile Insurance Rate Board (AIRB, or the Board) for its June 2013 Annual Review of Automobile Insurance Premiums for Basic Coverage (i.e.: Third Party Liability and Accident Benefits), which will culminate in an industry-wide rate adjustment to be effective November 1, 2013 for Private Passenger Automobile Insurance. The author of this report is as follows: Ronald R. Miller Insurance Services Inc. 30 Killdeer Crescent Toronto, ON M4G 2W8 Telephone: (416) Fax: (416) The author is a Fellow of the Casualty Actuarial Society and a Fellow of the Canadian Institute of Actuaries. 1.2 Scope My report is based on analysis of the Basic Coverage data at the sub-coverage level as included in the all-industry Private Passenger (excluding Farmers) Loss Development Exhibit recently published by GISA. It includes the following: projection of reported-to-date claim experience to ultimate values including Unallocated Loss Adjustment Expense (ULAE), but excluding the Alberta Health Levy costs trending of ultimate claim frequency, claim severity, and claim cost per car year of experience to levels appropriate for the policy (rating) year starting November 1, 2013, based on a regression model which tests, amongst other things, for changes in level and changes in trend at the time of implementation of the 2004 product reforms projection of claim payment patterns based on paid-to-date claim amount development projection of fixed and variable expenses and Alberta Health Levy costs to levels appropriate for the policy (rating) year starting November 1, 2013 projection of average street premium per car to a level judged to be applicable to the policy (rating) year starting November 1, 2013, prior to any industry wide adjustment decision the AIRB may make this cycle 1

12 1.3 My Approach I have reviewed and analyzed the currently available data and have performed the following steps: projection of ultimate claim counts and amounts from the raw reported-to-date Private Passenger (excluding Farmers) Alberta all-industry Green Book 31/12/2012 Loss Development Exhibit experience incurred-to-date claim count and claim amount data by accident half year by application of Loss Development Factors selected by me, and the loading of the resulting claim amounts for ULAE by application of the ULAE factors to be published in the Green Book (with one exception - see Section 2.2 below) forward projection of car years of exposure and ultimate claim frequencies, claim severities, and claim costs per car year of exposure for the next few accident half years, from the resulting ultimate claim frequencies, claim severities, and claim costs per car year of exposure above, using a log linear regression model which, amongst other things, tests for a change in level and a change in trend at the time of implementation of the 2004 reforms (mostly at 01/10/2004, except for some earlier reforms to the tort bodily injury coverage which were effective 24/01/2004) derivation from these forward projections of corresponding values appropriate for the policy year starting 01/11/2013 by the application of appropriate overlapping exposure weights projection of cumulative paid amount Loss Development Factors from the raw paid-to-date Private Passenger (excluding Farmers) Alberta all-industry Green Book 31/12/2012 Loss Development Exhibit experience paid claim amount data by accident half year, and conversion of these into projected payout patterns appropriate to the policy year starting 01/11/2013 projection of fixed expense amount per car year and the variable expense margin appropriate to the policy year starting 01/11/2013 based on the IBC 2011 Calendar Year Automobile Expense Survey report data projection of Alberta Health Levy costs appropriate to the policy year starting 01/11/2013, based on the Health Cost Recovery Amounts up to 2012 as shown in the AIRB s 2012 Annual Report together with judgment projection of the forward interest rate applicable to cash flow on underwriting and appropriate to the policy year starting 01/11/2013, based on current and expected near term yields on Government of Canada bonds matched approximately to the average duration of the claim liabilities 2

13 forward projection of the average street premium per car year to the level judged to be applicable to the policy (rating) year starting 01/11/2013 (before consideration of any adjustment that the AIRB may mandate as the result of the current annual review, or of any interim individual company Section 6 rate changes that the AIRB may otherwise approve subsequent to those which are fully reflected in the currently available data, to 1 st Quarter 2013), but this cycle making a judgmental adjustment to impute an additional amount in recognition that the written premium data up to early 2013 reflects significantly less than a 100% uptake so far of the +5% 01/11/2012 IWA decision last year The final indicated rate change may be derived by the following additional steps: deriving discount factors (discounting to average date of writing a policy) appropriate to the policy year starting 01/11/2013, based on the forward interest rate together with the payment patterns discussed above converting the ultimate claim costs, Health Levy costs, and fixed expense amount per car year appropriate to the policy year starting 01/11/2013 discussed above to appropriate discounted amounts using these discount factors, loading for forgone investment income due to the estimated average lag interval between writing a policy and receiving the premium, and loading by the variable expense margin plus selected profit and contingency margin, to produce the indicated required premiums per car year appropriate to the policy year starting 01/11/2013 loading these indicated required premiums per car year were for fixed expenses by distributing the indicated required premium for fixed expenses proportionally to the remaining required premium items to arrive at final indicated required premiums per car year appropriate to the policy year starting 01/11/2013 deriving the indicated required overall rate change appropriate to the policy year starting 01/11/2013 as the ratio of the total Basic Coverage indicated required premium per car year to the projected average street premium per car year, less one 1.4 Data Restatement The analysis underlying this report uses the data from the recently released GISA Alberta Private Passenger (ex. Farmers) 31/12/2012 Loss Development Exhibit. This is the second cycle where this exhibit has been produced from the new exhibit system implemented by GISA. As part of my analysis, I compared the TPL and AB data in this exhibit against its analogue from the 31/12/2011 exhibit which was also produced from the new exhibit system. For exposure and premiums, this comparison revealed only minor changes in overlapping data points, generally less than 0.5% in magnitude and in the most recent periods. For incurred claim counts, and paid and incurred claim amounts, this comparison also revealed only minor changes in overlapping data points, all less than 0.5% and generally in the most recent periods. Such changes are expected as some insurers refile data. This situation contrasts with the situation last cycle, when the exhibit system had changed, which showed somewhat larger and unusual data swing patterns. 3

14 1.5 Major Caveats There are many caveats which are applicable to my report. Principal amongst these are the following: all projections of average future claim costs are uncertain, the more so for the longer tail sub-coverages where claims do not settle for a substantial period of time because of significant product reform under Bill 53 and related initiatives, the subsequent Charter Challenge, and the inevitable erosion of the effectiveness of the Minor Injury Cap over time, this level of uncertainty in average future claim costs is higher than would otherwise be the case especially for the most recent years even though the Charter Challenge was ultimately defeated and the MIR was determined to be constitutional, the evolution of this challenge and recent court decisions likely lead to ultimate claim costs and development patterns which are different from what they would have been in their absence. Chronological highlights here are: 02/2008 decision of the lower court striking down the MIR ab initio 06/2009 Court of Appeal decision reversing the lower court s decision 12/2009 denial of leave to appeal by the Supreme Court of Canada 01/2012 Decision of the Alberta Court of Queen s Bench in Sparrowhawk v. Zapoltinsky regarding whether TMD is a minor injury This makes the projections of future development patterns and future costs here more uncertain because of possible disturbances to development patterns for recent periods, stemming from a number of sources, including the following: insurers generally increased their unpaid claim liability by year end 2008 after the lower court decision, often only through a change in bulk reserves (not reported under ASP), but in some cases, by increasing TPL-BI claim case reserves, and such increases in total reserves generally reflected an expectation of ultimate claim costs as a probability-weighting of values appropriate if the MIR remained struck down and if it were ultimately restored on appeal insurers continued to settle claims, including MIR claims, after the lower court decision, but closure rates declined insurers generally decreased their unpaid claim liability in this connection by year end 2009 after the Court of Appeal decision reversing the lower court decision and just after the denial of leave to appeal by the Supreme Court of Canada, often only through a change in bulk reserves (not reported under ASP), but in some cases, by decreasing TPL-BI claim case reserves 4

15 because of the timing of the denial of leave to appeal by the Supreme Court of Canada, a few insurers may not have had time to fully reduce their case reserves in this connection by 31/12/2009 as of 31/12/2009, there may be more incurred but unreported TPL-BI claims than usual, if some claimants in recent accidents may have been waiting to hear what the Supreme Court of Canada would say and delayed filing proof of loss until after the 17/12/2009 refusal of the Supreme Court of Canada to grant leave to appeal during the 2012 calendar year, there may be increases in case incurred claim severities for recent prior period claims asserting a TMD injury, and a corresponding increase in severity for new claims asserting such an injury claim count and claim amount development patterns for the and subsequent accident periods along future calendar period diagonals may in the fullness of time prove to differ from each other for the Third Party Liability - Bodily Injury sub-coverage the last several calendar half year diagonals in the incurred claim amount data development triangles for TPL-BI show development factors since the 2004 reforms at each development level often to be generally increasing for the latest several accident half years, with especially significant increases on the latest two such diagonals - this is likely a reflection of the inevitable erosion over time of the proportion of injuries falling under the Minor Injury Threshold, as claimants and their counsel seek ways to bypass the threshold, and the 01/2012 Sparrowhawk decision may be one of the factors underlying this changes along the latest two diagonals Alberta gasoline prices at the pump have been quite volatile in recent years, except for a fairly stable stretch in 2009 and 2010, and including a very volatile stretch in periods of higher gasoline prices may lead to a reduction in road exposure and therefore fewer claims, while periods of lower gasoline prices may have the opposite effect - no account of these fluctuations in gasoline prices has been taken into account in the analysis here all projections of future average street premium amounts are uncertain, but some more recent data, not part of the regular GISA 31/12/2012 Green Book exhibits, sheds some useful light on this issue (see Section 3.9 below) - for the purposes of this analysis I have assumed no net effect on average rates based on interim individual company Section 6 rate changes to compulsory coverage which the Board may approve subsequent to those fully reflected in this additional data, except that I have added an adjustment (that might or might not come to pass) to impute an additional amount recognition that the written premium data up to early 2013 reflects significantly less than a 100% uptake so far of the +5% 01/11/2012 IWA decision last year 5

16 all results in this report apply to all-industry averages and do not apply to and may not be appropriate for any particular individual company, especially given that each individual company had its compulsory insurance rates frozen as of 30/10/2003 at levels then in force (for new business) and at levels in force over the preceding year (for renewal business) - as with any competitive market, at that particular point in time, different insurers had different relative rate adequacy levels, but each was frozen at its own particular level just the same, and this issue continues to some degree going forward, although there has been some amelioration of this situation in recent years through Section 6 rate filings the underlying experience data analyzed in this report is for the all-industry Private Passenger Automobile (excluding Farmers, but including Fleets rated on a per car basis, and Trailers but with their exposure suppressed) class of business, but the indicated industry-wide rate level adjustment is meant to and does in fact apply to the Private Passenger Automobile (including Farmers, but excluding Fleets) class of business - the Farmers Private Passenger book of business accounts for a relatively small approximately 3% adjustment which should be added to the number of cars, with average premium levels for Basic Coverage at roughly half that of that for Private Passenger (excluding Farmers), but is expected to be subject to percentage changes in premiums and presumably in claim costs on account of the reforms similar to those of Private Passenger (excluding Farmers), and the Private Passenger Fleets on a per car basis book accounts for a relatively small approximately 0.7% adjustment which should be subtracted from the number of cars, with average premium levels (including premiums but not exposure for trailers) for compulsory coverage at roughly 60% to 70% above that of Private Passenger (excluding Farmers) - as such, the indicated compulsory insurance required premiums and projected street premiums per car in my report, which do reflect the inclusion of Fleets on a per car basis and Trailers with their exposure suppressed, but do not reflect the inclusion of Private Passenger Farmers business, may both be slightly overstated for the Private Passenger (including Farmers, but excluding Fleets and Trailers) class of business, but the indicated percentage rate change based on these is not expected to be materially affected all results in this report apply to all-industry all classification cell averages and do not apply to and may not be appropriate for any particular territory or other classification cell, even at the all-industry level, let alone at the individual company level - this may be a particular issue for any individual company whose distribution of business is significantly different from that of all-industry at the territorial level, or along other classification dimensions Effective 01/07/2006, the GST rate was reduced from 7% to 6%, and effective 01/01/2008, it was further reduced from 6% to 5% - to the extent that some components of claim and other expense costs are subject to the GST, these changes are expected to reduce such costs slightly, and this issue has not been addressed explicitly in my analysis, but for expenses, it is implicitly accounted for by starting from expense data from 2008, and for claims, it is implicitly accounted for to some degree by the regression methodology used for trending 6

17 the historical payment patterns estimated in this report are based on the analysis of the underlying raw paid-to-date Green Book development data which in part relates to the product as it was prior to the Bill 53 and related reforms, in part relates to the product as was after the reforms but before the lower court decision, in part relates to the product as it is after the 02/2008 strikedown of the MIR but before the reversal of this decision by the Court of Appeal, in part relates to the product as it is after the 06/2009 reversal of the lower court decision by the Court of Appeal but before the denial of leave to appeal by the Supreme Court of Canada, in part relates to the product as it is after the denial of leave to appeal by the Supreme Court of Canada, and in part relates to the environment as it is after the Sparrowhawk decision - some shift in the payment patterns on account of the reforms and on account of this situation, especially for the Third Party Liability - Bodily Injury (TPL- BI) sub-coverage should be expected, and I have made an adjustment to the pattern I might otherwise have selected for this latter sub-coverage, based on judgment Orders-in-Council 520/2006 to 522/2006 revised certain regulations under the Insurance Act - in particular, by: increasing, effective 01/03/2007, certain caps on accident benefits under the Accident Benefits Regulation, including the cap on the funeral benefits ($2,000 up to $5,000) and weekly disability income indemnity cap ($300 to $400 for employed, $100 to $135 for others - this reform is expected to increase claim costs slightly and is explicitly addressed in my analysis by the introduction of an additional level variable in my trend analysis indexing the Minor Injury Amount (cap on non-pecuniary damages in cases of minor injury) based on a calculated change in the Alberta CPI, effective 01/01/2007 and each subsequent January 1 - for 2007, the $4,000 amount increased by 3.6% to $4,144, for 2008 by 4.7% to $4,339, for 2009 by 3.8% to $4,504, for 2010 by 0.3% to $4,518, and for 2011 by 0.9% to $4,559 - this reform is expected to increase claim costs slightly and is not explicitly addressed in my analysis Effective 01/06/2013, rates in the fee schedules for Chiropractic and Physical Therapy services are increasing modestly, which may lead to a modest increase in AB-MR claim costs, but no account of this change has been taken here In recent previous cycles, the ULAE factor, Private Passenger Commission loading, and general expense loading for the latest year (now 2012 this cycle) were taken from the values published by IBC in its Automobile Expense Survey (AES) report for the latest year, but unfortunately, this survey was discontinued after the last cycle, and no new data is available The new exhibit system has produced overlapping loss development data which differs somewhat from the analogous data from the old system for older accident periods 7

18 2. Assumptions and Model Parameters in the Analysis My analysis of the Loss Development Exhibit and other data incorporates a number of assumptions and selection of model parameters, the most important of these are discussed in the following sub-sections. 2.1 Incurred Count and Incurred Amount Loss Development Factors I selected incurred count and incurred amount loss development factors by sub-coverage and accident half year based on analysis of the recently released GISA Alberta Private Passenger (ex. Farmers) 31/12/2012 Loss Development exhibit. As a default, the selected factors were taken as the average of the latest four such factors, except for the first, which was taken as the average of the latest two such factors for the same half year (because of seasonality). I then made a number of adjustments to these default factors, by judgment at the sub-coverage level, as follows: TPL-BI: because of the generally high incurred age-to-age claim amount development factors observed on the latest two calendar half year diagonals (see table below) and uncertainty as to whether such higher values are likely to continue in the future (or perhaps rather decrease or perhaps even increase further), this cycle, instead of adopting the default factors, I selected factors for both incurred counts and amounts as the average of the latest six such factors, except for the first, which was taken as the average of the latest three such factors for the same half year (because of seasonality), except that the 17 th factor for amounts was increased by about 0.4% to account for an expected lag in development post reform - had I adopted the default factors (with adjustment) instead (which are similar to the factors adopted by GISA s actuary for Green Book purposes), the resulting projection of the ultimate claim cost for the policy year starting 01/11/2013 would have been about 6.5% higher Alberta PPAxF TPL-BI Historical Incurred Amount Age-to-Age Development Factors Based on Analysis of GISA PPAxF Loss Development Data as of 31/12/2012 Accident Age / Age (Months) Half Year 12/06 18/12 24/18 30/24 36/30 42/36 48/42 54/48 60/54 66/60 72/66 78/72 84/78 90/84 96/

19 The paid amount develop factors selected as discussed in Section 2.4 below may be used to produce another and different set of estimates of the ultimate claims amounts here, and the following table provides a comparison of these two sets of estimates for accident half years and subsequent: Alberta PPAxF TPL-BI Ultimate Claim Cost per Car (excluding ULAE) Comparison of Incurred vs Paid Projections Based on Analysis of GISA PPAxF Loss Development Data as of 31/12/2012 Accident Half Year Incurred (I) Projection Paid (P) Projection Difference P Less I % Difference P / I $ $ ($5.85) -2.8% $ $ ($7.57) -3.3% $ $ ($6.37) -3.2% $ $ ($10.86) -4.5% $ $ ($14.56) -7.4% $ $ ($6.26) -2.6% $ $ ($15.62) -7.4% $ $ ($19.17) -7.9% $ $ ($9.58) -5.0% $ $ ($13.85) -5.9% $ $ $ % $ $ $ % $ $ $ % $ $ $ % $ $ $ % $ $ $ % Average $ $ ($3.84) -1.8% Because of the significantly larger age-to-age development factors for Paid versus Incurred, the Paid projection meteorology is much less stable than the Incurred projection methodology. Nevertheless, the above table shows that for the and subsequent accident periods, the two sets of projections are on average quite close, and that for the last six periods (i.e. those in 2010, 2011, and 2012), the Paid projections are higher than the Incurred projections, which suggests that it is unlikely that the Incurred projections here are being overstated because of possible current overly conservative case reserves. 9

20 AB-FB: for both incurred counts and amounts, the factors were taken as the weighted average factors, except for the 1 st, which was taken as the weighted average of such factors for the same half year AB-DB: for both incurred counts and amounts, the factors were taken as the weighted average factors, except for the 1 st, which was taken as the weighted average of such factors for the same half year AB-DI: for incurred amounts, the 6 th factor was taken as the weighted average factor AB-SB: for both incurred counts and amounts, the factors were replaced by factors selected by judgment AB-UM: for both incurred counts and amounts, the factors were taken as weighted average factors, except for the first, taken as the weighted average such factors for the same half year In many cases, apparently spurious factors in the tail were ignored. Detailed supporting exhibits are not included in this report. In my view, especially because of observed generally higher incurred claim amount age-to-age development along the latest two calendar half year diagonals for TPL-BI, which may reflect erosion of the threshold from the 01/2012 Sparrowhawk decision amongst other causes, significant uncertainty remains at this time as to the values of the eventual ultimate claim amounts for TPL-BI for recent accident periods, and development patterns for these may in the fullness of time prove to differ from their analogues from the recent past and from each other, the more so for the more recent periods. These factors were applied by sub-coverage and accident half year to the reported-to-date incurred counts and amounts in the PPA LDE data to project ultimate counts and amounts, the amounts being before inclusion of Unallocated Loss Adjustment Expenses. 2.2 Unallocated Loss Adjustment Expense Factors Claim data reported under the Automobile Statistical Plan (ASP) includes Pure Loss (indemnity) and Allocated Loss Adjustment Expenses (ALAE), but not Unallocated Loss Adjustment Expenses (ULAE). It is therefore necessary to adjust such claim amounts to include ULAE in order to get a proper picture of total claim costs. I selected ULAE factors as those published by GISA in the introductory pages of the 2012 PPA ALR for the most recent years, except that the factors for 2008 and 2009 were replaced by their average (see footnote below table following). The 2012 factor was taken as the same as the 2011 factor, since the old IBC Automobile Expense Survey (AES) was discontinued after the last cycle, and no new data is available - it is my understanding that a similar approach for 2012 is being taken for production of the GISA Green Book exhibits. For earlier years, I selected the factors as those previously published by GISA and/or IBC. These factors were applied uniformly by sub-coverage and accident half year to the ultimate claim amounts projected in section 3.1 above. 10

21 The selected ULAE factors are as follows for accident years 1994 and subsequent: Alberta Private Passenger (Excluding Farmers) Automobile Insurance Green Book Unallocated Loss Adjustment Expense Factors (except 2008 and 2009) Accident Year ULAE Factor * * ** * because of a distortion caused by the way these factors are derived from calendar year data and the fact that insurers increased total reserves materially during 2008 on account of the lower court decision striking down the MIR and reduced them materially during 2009 on account of the appeal court decision and Supreme Court denial of leave to appeal, both in the Green Book and here, the factors for both 2008 and 2009 are taken as the average of those indicated by the Automobile Expense Survey data (1.084 for 2008 and for 2009) ** because the Automobile Expense Survey was discontinued, no new data is available here, and the 2011 factor has been adopted for 2012 These factors were applied to the ultimate claim amounts (before ULAE) projected in section 3.1 above. For 2004 and prior, the factors are derived from aggregate data submitted under the old IBC Expense Allocation Program, and for 2005 and subsequent, they are derived from aggregate data submitted under the new IBC Automobile Expense Survey. 11

22 2.3 Trending Model I used a log linear regression trending model applied at the accident half year and sub-coverage level (see Exhibit 2). The input to this model is the historical earned exposure and premium data and projected ultimate claim counts and amounts (from sections 3.1 and 3.2 above), all derived from the PPA LDE data, for the and subsequent accident half years - this provides a balance between the number of years included which were before and which were after the year of the reforms (2004). Because the trending model is a linear regression on the logarithms of the dependent variables of claim frequency, claim severity, and claim cost per car year, it provides a consistent treatment of these three dependent variables (i.e. the tautology claim frequency x claim severity = claim cost per car year is preserved in all fitted and projected values). This is not necessarily the case with other trending models Independent Variables in Trending Model This model tests for the significance of, and potentially fits, parameters related to up to 7 independent variables, as follows: Overall Level Change in Level re the 2004 Reforms Change in Level at 01/03/2007 Change in Level for 2 nd half years compared to 1 st half years (i.e. Seasonality re 2 nd half years) Change in Level for half years with Weather related Catastrophe Occurrences Overall Trend Change in Trend re the 2004 Reforms For the Third Party Liability - Bodily Injury (TPL-BI) sub-coverage, in the last cycle, I made adjustments to the values (which would otherwise have been appropriate) of the independent variables relating to Change in Level re the 2004 Reforms and to Change in Trend re the 2004 Reforms. 12

23 In the last cycle, for the Change in Level variable, a further adjustment was made in recognition of a pattern in the data which then appeared to indicate that the effect on TPL-BI loss costs after the 2004 Reforms did not occur suddenly, but rather phased itself in over an extended period of time, but with some disruption to this pattern on account of the 02/2008 lower court decision striking down the MIR, the subsequent 06/2009 decision of the Court of Appeal, and the 12/2009 denial of leave to appeal by the Supreme Court of Canada. However, since the last cycle, the projections of post reform ultimate claims cost per car no longer show this anomalous pattern, many of them having undergone significant development in the last year, and so I have dropped the phase in approach. For both the Change in Level and Change in Trend variables, these adjustments were made in recognition of the fact that part of the reforms here (i.e. gross to net income loss and offset of collateral source amounts) became effective at 24/01/2004, before the 01/10/2004 date when most of the reforms were implemented, and I have continued this approach this cycle. My adjustments for the earliest periods are based on KPMG s a priori estimate that 30% of the total savings for TPL-BI would be due to the part of the reforms which ended up being implemented earlier (see page 20 of KPMG s 13/12/2004 report to Alberta Finance Report I Costing Analysis of 2004 Auto Reform : 80 / ( ) = 30%), and on judgment, and are as follows: Variable = Change in Level re the 2004 Reforms for TPL-BI: Time 0 = 01/01/2004 Value for = 30% of ( / 181) =.262 Value for = 30% of % of.50 =.650 Value for and subsequent = Variable = Change in Level re the 2004 Reforms for Other Sub-Coverages: Time 0 = 01/01/2004 Value for =.500 Value for and subsequent = Variable = Change in Trend re the 2004 Reforms for TPL-BI: Time 0 = 24/01/2004 for parts implemented early, and 01/10/2004 for rest Value for = 30% of ( x 23 / 365) =.066 Value for = 30% of ( / 365) + 70% of.125 =.294 Value for = 30% of ( / 365) + 70% of.500 =.706 Value for and subsequent = prior Variable = Change in Trend re the 2004 Reforms for Other Sub-Coverages: Time 0 = 01/10/2004 Value for =.125 Value for =.500 Value for and subsequent = prior