MANITOBA PUBLIC INSURANCE

Transcription

1 MANITOBA PUBLIC INSURANCE AI.11 RATE STABILIZATION RESERVE AI.11 Discussion of the Rate Stabilization Reserve (RSR) AI.11.A Background The purpose of the Rate Stabilization Reserve (RSR) is to protect motorists from rate increases made necessary by unexpected events and losses arising from nonrecurring events or factors. Over the last several years, the Corporation and the Public Utilities Board have had differing perspectives on the appropriate target amount of the Basic RSR. Historical Timeline First Efforts The Percentage of Premium Method The Corporation s records indicate that the first public reference to the matter of retained earnings for the basic compulsory program is found in the 1989 Kopstein Report on Auto Insurance in Manitoba. At that time, private sector insurers were required to maintain a reserve equivalent to 30% of annual premium. Some interested parties took the position that public sector plans such as Autopac needed no reserves, because the government could be relied on to fund shortfalls. Essentially, Kopstein took a position between the two extremes and expressed support for retained earnings at approximately 15% of annual premiums. Since that time, insurance regulators have abandoned a premiums-based approach and have mandated successively more scientific and risk-based approaches, concluding with today s requirement that insurers utilize DCAT and MCT. In its Order 161/09 subsequent to the 2010 GRA, despite the Corporation s recommendations, the PUB ordered that the RSR target be set based on the Kopstein approach of 10% to 20% written premiums. It was the PUB s view that the RSR methodology should be one that is clearly understood by all parties. The indicated RSR range based on this method was $77 million to $154 million. The PUB Page 1

2 August 3, 2012 Rate Stabilization Reserve Part 1 AI.11 (Amended) also ordered that the DCAT, MCT, and RA/VaR be filed on no less than a triennial basis MPI Proposes Risk Analysis Framework In an effort to introduce more modern and mathematical approaches to assessing risk, the Corporation developed the first risk analysis as part of the 2000 GRA. Through successive rate hearings both MPI and the PUB made substantive modifications to what is now known as the Operational and Investment Risk Analysis (RA/VaR). However, given the evolution of insurance industry-specific approaches that have evolved since that time, and given the significant shortcomings and inadequacies of the Ra/VaR method, the Corporation believes that this method provides insufficient value and should no longer be used Minimum Capital Adequacy Test In 2005, the Corporation adopted a policy to base the RSR on the Minimum Capital Test (MCT), a test established by the Office of the Superintendent of Financial Institutions (OSFI) to measure the financial strength of private insurers. The test is a comparison of risk-adjusted available capital to risk-adjusted required capital. Private insurers are required to maintain an MCT score above 150% in order to avoid OSFI intervention. The Corporation proposed a target range based on 50% to 100% MCT to account for the differences between a monopoly and private insurer Dynamic Capital Adequacy Test The Corporation did further research as part of the 2010 GRA on the RA/VaR and MCT approach and concluded that neither approach truly addressed the stated purpose of the RSR. The Corporation argued that the RA/VaR was only based on events that have occurred since 1994/05, rather than plausible events that could occur now and in the future. Similarly, the MCT analyzed the risks inherent in the balance sheet, but did not study other operational risks inherent in the Corporation. It was also unclear if the MCT risk factors were reflective of the risks faced by a monopoly public insurer. Given the weaknesses of the existing RSR methods, the Corporation recommended an improved method for determining the RSR target Dynamic Capital Adequacy Page 2

3 August 3, 2012 Rate Stabilization Reserve Part 1 AI.11 (Amended) Testing (DCAT). The DCAT was recommended because it explicitly measures the potential financial impact from the Corporation s key risk factors and produces a RSR target that is directly related to the Corporation s risk level. The Corporation also informed the PUB at the 2010 GRA that the DCAT analysis would now be conducted internally and be integrated as part of the Corporation s risk management processes. The recommended RSR target based on the DCAT was $185 million for the 2010/11 year. AI.11.B Introduction From the 2010 GRA to present, the Kopstein or Percentage of Premium method was used to determine the PUB s RSR range, while the Corporation continued to use the DCAT for internal purposes. In its 2012 Order, the PUB stated that MPI file with the Board together with next year's GRA an updated DCAT, MCT and RA/VaR, together with the pros and cons of each approach and MPI's position on what methodology would best develop an appropriate RSR target range. The Corporation has filed the updated reports in AI.11: Part 2 Dynamic Capital Adequacy Test, Part 3 Minimum Capital Test, Part 4 Operational and Investment Risk Analysis, and Part 5 Percent of Premium approach. The sections below include the following information for each of the three RSR methodologies outlined in the Order along with the current Percentage of Premium methodology: - A high-level description of the methodology - The indicated RSR target produced by the methodology - The pros and cons of the methodology Page 3

4 August 3, 2012 Rate Stabilization Reserve Part 1 AI.11 (Amended) AI.11.C Dynamic Capital Adequacy Testing Description: Dynamic Capital Adequacy Testing (DCAT) is a process of analyzing and projecting the trends of an insurer s capital position given its current circumstances, its recent past, and its intended business plan under a variety of future scenarios. It allows the Corporation to determine the implications that the business plan has on capital and identify the significant risks to which it is exposed. The following link (section 2500) provides the Standards of Practice on Dynamic Capital Adequacy Testing from the Canadian Institute of Actuaries effective January 1, 2012: Indicated RSR Target: A minimum RSR balance of $200 million as of February 29, 2012 based on the amended DCAT report (subject to approval by the Manitoba Public Insurance Board of Directors) Pros: 1. Assists Management and the Public Utilities Board in the identification, measurement and mitigation of key risks faced by the Corporation. 2. Creates a forward looking measure of risk (i.e. not a retrospective measure like other tests). 3. Uses company specific assumptions for adverse scenarios, ripple effects and management action, as opposed to prescribed rules that are the same for every company. 4. Produces an opinion that is based on the RSR targets set by the Manitoba Public Utilities Board. 5. Creates a clear linkage between the required RSR and the amount of risk faced by the Corporation. 6. Is a recognized method of the Canadian Institute of Actuaries and the Office of the Superintendent of Financial Institutions (OSFI). 7. The adverse scenarios and associated assumptions can be discussed and debated at the General Rate hearings and, if warranted, modified in the next DCAT report. Since the DCAT is done in-house, the Corporation can provide Page 4

5 August 3, 2012 Rate Stabilization Reserve Part 1 AI.11 (Amended) the impacts of alternate adverse scenarios at the request of the Public Utilities Board. Cons: 1. Certain risk factors may be difficult to quantify. 2. The results are not directly comparable to other jurisdictions. 3. Actuaries may have different opinions on likelihood and magnitude of given risk factors. AI.11.D Minimum Capital Test (MCT) Description 1 : The Minimum Capital Test (MCT) was developed by Canadian regulators to assess the key risks faced by the industry and to harmonize capital requirements across jurisdictions in Canada. It is a risk-based approach [that] better reflects the riskiness of individual P&C insurers and is consistent with approaches in other financial sectors. It assesses the riskiness of assets, policy liabilities, and offbalance sheet exposures, by applying various factors. Indicated RSR Target: A minimum capital (RSR) requirement of $283 million (i.e. 100% MCT score). Pros: 1. Is designed to assess the key risks faced by the insurance industry, the majority of which are relevant to Manitoba Public Insurance. 2. Is used by other insurers, including SGI and ICBC, and is recognized by OSFI. 3. Assesses the riskiness of assets, policy liabilities, and off-balance sheet exposures, by applying a consistent set of factors that were agreed upon by a task force of insurance experts. 4. Identifies risks based on the Corporation s current financial statements (or current risk profile). 1 All quotes in AI.11.C are from Notes on the Development of the Minimum Capital Test, Office of the Superintendent of Financial Institutions Canada, July 2003, or Guideline Minimum Capital Test (MCT) for Property and Casualty Insurers, Office of the Superintendent of Financial Institutions Canada, July Page 5

6 August 3, 2012 Rate Stabilization Reserve Part 1 AI.11 (Amended) 5. The calculation of the MCT score is completely objective (i.e. no judgment is required). 6. The MCT score is relatively easy to calculate. Cons: 1. It is a private sector test. Policyholder and regulatory concerns with respect to insurer solvency are of a different magnitude and kind compared to the situation with a Crown Corporation. 2. The method uses the same risk factors for all companies, which may not be reflective of the Corporation s risk level. 3. It requires the Public Utilities Board and the Corporation to make a judgment on the appropriate MCT score for a monopoly insurance provider (e.g. 50% to 100% MCT was recommended in the 2007 Rate Application). AI.11.E Operational and Investment Risk Analysis Description: The Operational and Investment Risk Analysis (Risk Analysis) is designed to provide a risk assessment relating to the financial outcomes for the Basic insurance program for Manitoba Public Insurance. As stated in Order 151/2000 number 3, the Risk Analysis incorporates the following criteria: At a 95.0% confidence level, both including and excluding operating costs; and At a 97.5% confidence level, both including and excluding operating costs Each of four scenarios should: Include only PIPP data; Include investment risk using a Value at Risk Analysis assuming a 25.0% equity component and a time horizon of between two and three years; Use variances between forecast and actual amounts for revenues, losses, operating expenses, and claim expenses; and Use actual correlations between all risk components recognizing the directional effect on net income. Indicated RSR Target: An RSR range of $193 million to $291 million. Page 6

7 August 3, 2012 Rate Stabilization Reserve Part 1 AI.11 (Amended) Pros: 1. The analysis is easy to update. 2. The operational risk calculations are transparent. 3. The mathematical calculations of historic operational risk are objective. Cons: 1. The operational risk calculation is based entirely on an analysis of financial outcomes over the last eighteen years without any adjustment to reflect the Corporation s current cost levels. 2. The methodology does not reflect the Corporation s current risk profile as many changes to the Corporation s policies, procedures, contracts, etc. have been made over the past eighteen years (e.g. reinsurance contracts, reserving guidelines, investment policies, etc). 3. The operational risk analysis is not based on plausible adverse events that could occur. Rather, it is based on events that have actually occurred in the past eighteen years. 4. If the Corporation improves its forecasting process, it will have a lower operational risk margin, since the margin depends on the deviations between actual and forecasted results. Alternatively, the risk margins can be increased by poor forecasting or other changes that are unrelated to operational risk. 5. The historical results (currently only 18 data points) do not provide a credible sample for statistical modeling. For example, a general rule of thumb for obtaining a statistically reliable correlation coefficient would be a minimum of 30 data points for normally distributed variables and as much as 100 or more for non-normally distributed variables. 6. Using Value-at-Risk (VaR) may not be appropriate for a time horizon of 2.5 years. In all of the Corporation s research regarding VaR the Corporation has never seen an author advise calculating VaR for a period longer than one year. 7. The risk margins produced do not provide the Corporation or the Public Utilities Board with a clear understanding of the risks faced by the Corporation, how these risks should be mitigated, or what management or Public Utilities Board action may be required. Page 7

8 August 3, 2012 Rate Stabilization Reserve Part 1 AI.11 (Amended) 8. The method is not used or recognized by any other insurer, regulator, or professional body. AI.11.F Percentage of Premium (Kopstein) Method Description: In the 1989 Kopstein Report on Auto Insurance in Manitoba recommendation 7.1 stated: That the government of Manitoba issue a public directive to the corporation setting an Autopac retained earnings surplus target of about 15% of premiums. (This would amount to $40 to $50 million at prevailing premium levels.) The government directive should indicate that, if Autopac surplus falls below 10 percent or exceeds 20 percent of premiums, the corporation should and would be expected to take remedial action. The Percentage of Premium method for determining the RSR, which is the method currently used by the PUB, is based on the recommendation from the Kopstein report. The RSR range for a given year is determined based on 10% to 20% of net written premiums. Indicated RSR Target (2012/13): An RSR range of $77 million to $153 million. Pros: 1. The indicated RSR range is easy to understand and calculate. 2. The mathematical calculations are objective. Cons: 1. The method assumes that the Corporation s risk level is a function of its annual premium level; however, the Corporation s main risks are from changes to assets and liabilities which are significantly larger than annual premiums. 2. It does not assist Management or the Public Utilities Board in the identification, measurement and mitigation of key risks. 3. The method does not create a clear linkage between the required RSR and the amount of risk faced by the Corporation. Page 8

9 August 3, 2012 Rate Stabilization Reserve Part 1 AI.11 (Amended) 4. The indicated RSR range does not change when the Corporation s risk profile changes. 5. To the Corporation s knowledge, the method is not recognized or used by any other Regulator or professional body. AI.11.G Why the DCAT should be used to set the RSR Target The Corporation recommends that the DCAT be used to set the RSR target. The Corporation s position on this matter has not changed since it last filed a Discussion of the Rate Stabilization Reserve in AI.18.1 in the 2010 Rate Application. The key benefits to using the DCAT to set the RSR target, compared to using the MCT, RA/VaR, or Percentage of Premium (PoP) methodologies, are as follows: 1. Assists Management and the Public Utilities Board in the identification, measurement, and mitigation of key risks faced by the Corporation. DCAT: MCT: RA/VaR: Risks are clearly identified and quantified with supporting rational. Adverse scenarios are run through the Corporation s financial model showing the expected impact on RSR. Realistic Management and Regulatory actions are incorporated into the scenarios. It is not clear how the risk factors were developed, if they are appropriate for a public insurer, or what specific risk events they represent. The current level of operational risk is assumed to be reflective of the Corporation s historical forecasting experience. Specific operational risks are not identified. The Value-at-Risk method of measuring investment risk is not appropriate over a 2.5 year time period. The method is independent of the risk level of the Corporation. 2. Creates a forward looking measure of risk (i.e. not a retrospective measure like other capital tests). Page 9

10 August 3, 2012 Rate Stabilization Reserve Part 1 AI.11 (Amended) DCAT: MCT: RA/VaR: The adverse scenarios are tested against the Corporation s pro forma financial statements and are assumed to occur in the rating period of the current application. Required capital is based on the current balance sheet, which is an improvement over the RA/VaR and Percentage of Premium methods. However, it is unclear what specific future risks the required capital is providing protection from and whether such an amount is appropriate for MPI. The indicated RSR is based completely on historical forecasting performance. Historical results are not at the same cost level as future experience and include events that are not reflective of future risk (e.g. $250 million reduction in claim liabilities). The indicated RSR range does not change based on the Corporation s risk profile. 3. Uses company specific assumptions for adverse scenarios, ripple effects, and management action, as opposed to prescribed rules that are the same for every company. DCAT: MCT: RA/VaR: The adverse scenarios are developed specifically for MPI. It is not clear how the risk factors were developed, if they are appropriate for a public insurer, or what specific risk events they represent. Uses historical MPI forecasting performance for operational risk; however, the sample size of 18 data points (i.e. years of history) is not a statistically reliable sample. Also, past forecasting performance is not necessarily relevant to future forecasting performance. The method is independent of the risk level of the Corporation. 4. Produces an opinion as to the adequacy of the basic insurance plan reserve that is based on the RSR targets set by the Manitoba Public Utilities Board. DCAT: A satisfactory opinion is based on the Regulatory rules in this province. Page 10

11 August 3, 2012 Rate Stabilization Reserve Part 1 AI.11 (Amended) MCT: RA/VaR: It is unclear how the private sector minimum regulatory target of 150% MCT is applicable to a public insurer. In effect, the PUB would set the target, which would, by implication, be adequate, with no inherent process to identify and quantify plausible risks. In effect, the PUB would set the target, which would, by implication, be adequate, with no inherent process to identify and quantify plausible risks 5. Creates a clear linkage between the required RSR and the nature as well as amount of risk faced by the Corporation. DCAT: MCT: RA/VaR: The indicated RSR target is the amount of RSR the Corporation requires at the evaluation date to maintain RSR above zero under all plausible adverse scenarios. It is not clear how the risk factors were developed, if they are appropriate for a public insurer, or what specific risk events they represent. It is not clear why the indicated level of RSR is required. The method is independent of the risk level of the Corporation. 6. It is a recognized method of the Canadian Institute of Actuaries and OSFI. DCAT: MCT: RA/VaR: Is used by all Canadian Property and Casualty insurers that report to OSFI. Is used by all Canadian Property and Casualty insurers that report to OSFI. Is not used or recognized by any other Regulator or professional body. To the Corporation s knowledge, is not used or recognized by any other Regulator or professional body. 7. The adverse scenarios and associated assumptions can be discussed and debated at the General Rate hearings and, if warranted, modified in the next DCAT report. Page 11

12 August 3, 2012 Rate Stabilization Reserve Part 1 AI.11 (Amended) DCAT: MCT: RA/VaR: The adverse scenarios are tailored to MPI s operations and can be continuously improved over time. The RSR target is based on risks specific to MPI. The risk factors are supplied and cannot be changed. There is no judgment required to calculate the required capital. The risk factors are not specific to MPI. The method uses prescribed calculations and is not responsive to risks specific to MPI. The method uses prescribed calculations and is not responsive to risks specific to MPI. AI.11.H Capital Targets used by other Public Insurers Saskatchewan Government Insurance uses an MCT target range of 75% to 150% for the Auto Fund. If these targets were used for MPI s basic compulsory program the RSR range would be $212 million to $425 million. The Insurance Corporation of British Columbia has a statutory minimum capital requirement of 100% MCT for its basic line of business. They also have a Capital Management Plan (approved by their regulator) with a target MCT of 130%. If these targets were used for MPI s basic compulsory program the RSR range would be $283 million to $368 million. Based on the Corporation s amended DCAT (subject to MPI Board of Directors approval), MPI s RSR should be $200 million. The RSR suggested by the Saskatchewan and British Columbia method would result in MPI requiring an RSR much higher than we believe is reasonable or warranted, based on plausible adverse scenarios appropriate to MPI s Basic program. To achieve this level of RSR would mean increasing rates to generate higher retained earnings for the reserve. Page 12

13 August 3, 2012 Rate Stabilization Reserve Part 1 AI.11 (Amended) AI.11.I Recommendation 1. That the PUB adopt the Dynamic Capital Adequacy Test as the method for determining the MPI required RSR and agree to an RSR target for 2012/13 of $200 million (subject to the approval of the amended DCAT report by the MPI Board of Directors). The above analysis and supporting documentation provides a sound rationale for the PUB to adopt the Dynamic Capital Adequacy Test as the method for determining the required RSR. The DCAT explicitly measures the potential financial impact from the Corporation s key risk factors and produces a RSR target that is directly related to the Corporation s risk level and directly responsive to the purpose of the RSR. 2. That the PUB continue to monitor the Basic insurance MCT score and the RSR targets established by other Crown auto insurers on the basis that it is an industry standard approach used by other similar entities. 3. That the PUB discontinue monitoring the results obtained through the Risk Analysis/VaR calculations, based on the many shortcomings identified herein. 4. That the PUB discontinue monitoring the results obtained through the Percentage of Premium ( Kopstein ) approach for the primary reason that the Basic insurance plan s greatest risks are to its assets and liabilities and bear no relationship to its annual revenues. Page 13

14 MANITOBA PUBLIC INSURANCE AI.11 RATE STABILIZATION RESERVE AI.11 Discussion of the Rate Stabilization Reserve (RSR) AI.11.A Background The purpose of the Rate Stabilization Reserve (RSR) is to protect motorists from rate increases made necessary by unexpected events and losses s arising from nonrecurring events or factors. Over the last several years, the Corporation and the Public Utilities Board have had differing perspectives s on the appropriate target amount of the Basic RSR. Historical Timeline First Efforts The Percentage of Premium Method The Corporation s records indicate that the first public reference to the matter of retained earnings are for the basic compulsory program is found in the 1989 Kopstein Report on Auto Insurance in Manitoba. At that time, private sector insurers were required to maintain n a reserve equivalent to 30% of annual premium. Some interested parties took the position that public sector plans such as Autopac needed no reserves, because the government could be relied on to fund shortfalls. Essentially, Kopstein took a position between the two extremes and expressed arr REVISED ED support for retained earnings at approximately 15% of annual premiums. Since that time, insurance regulators have abandoned a premiums-based approach and have mandated successively more scientific and risk-based approaches, concluding with today s requirement that insurers utilize DCAT and MCT. In its Order 161/09 subsequent to the 2010 GRA, despite the Corporation s recommendations, the PUB ordered that the RSR target be set based on the Kopstein approach of 10% to 20% written premiums. It was the PUB s view that the RSR methodology should be one that is clearly understood by all parties. The indicated RSR range based on this method was $77 million to $154 million. The PUB Page 1

15 June 15, 2012 Rate Stabilization Reserve Part 1 AI.11 also ordered that the DCAT, MCT, and RA/VaR be filed on no less than a triennial basis MPI Proposes Risk Analysis Framework In an effort to introduce more modern and mathematical approaches to assessing risk, the Corporation developed the first risk analysis as part of the 2000 GRA. Through successive rate hearings both MPI and the PUB made substantive modifications to what is now known as the Operational and Investment Risk Analysis (RA/VaR). However, given the evolution of insurance industry-specific approaches that have evolved since that time, and given the significant shortcomings and inadequacies of the Ra/VaR method, the Corporation believes that this method provides insufficient value and should no longer be used Minimum Capital Adequacy Test In 2005, the Corporation adopted a policy to base the RSR on the Minimum Capital Test (MCT), a test established by the Office of the Superintendent of Financial Institutions (OSFI) to measure the financial strength of private insurers. The test is a comparison of risk-adjusted available capital to risk-adjusted required capital. Private insurers are required to maintain an MCT score above 150% in order to avoid OSFI intervention. The Corporation proposed a target range based on 50% to 100% nclr MCT to account for the differences between a monopoly and private insurer Dynamic Capital Adequacy Test REVISED ED The Corporation did d further research as part of the 2010 GRA on the RA/VaR and MCT approach and concluded that neither approach truly addressed the stated purpose of the RSR. The Corporation argued that the RA/VaR was only based on events that have occurred since 1994/05, rather than plausible events that could occur now and in the future. Similarly, the MCT analyzed the risks inherent in the balance sheet, but did not study other operational risks inherent in the Corporation. It was also unclear if the MCT risk factors were reflective of the risks faced by a monopoly public insurer. Given the weaknesses of the existing RSR methods, the Corporation recommended an improved method for determining the RSR target Dynamic Capital Adequacy Page 2

16 June 15, 2012 Rate Stabilization Reserve Part 1 AI.11 Testing (DCAT). The DCAT was recommended because it explicitly measures the potential financial impact from the Corporation s key risk factors and produces a RSR target that is directly related to the Corporation s risk level. The Corporation also informed the PUB at the 2010 GRA that the DCAT analysis would now be conducted internally and be integrated as part of the Corporation s risk management processes. The recommended RSR target based on the DCAT was $185 million for the 2010/11 year. AI.11.B Introduction From the 2010 GRA to present, the Kopstein or Percentage of Premium method was used to determine the PUB s RSR range, while the Corporation continued to use the DCAT for internal purposes. In its 2012 Order, the PUB stated that MPI file with the Board together with next year's GRA an updated DCAT, MCT and RA/VaR, together with the pros and cons of each approach and MPI's position on what methodology would best develop an appropriate RSR target range. The Corporation has filed the updated reports in AI.11: Part 2 Dynamic Capital Adequacy Test, Part 3 ir Minimum m Capital Test, Part 4 Operational and Investment Risk Analysis, and Part 5 Percent of Premium approach. The sections below include the following information for each of the three RSR methodologies outlined in the Order along with the current Percentage of Premium methodology: REVISED - A high-level description of the methodology - The indicated RSR target produced by the methodology - The pros and cons of the methodology Page 3

17 June 15, 2012 Rate Stabilization Reserve Part 1 AI.11 AI.11.C Dynamic Capital Adequacy Testing Description: Dynamic Capital Adequacy Testing (DCAT) is a process of analyzing and projecting the trends of an insurer s capital position given its current circumstances, its recent past, and its intended business plan under a variety of future scenarios. It allows the Corporation to determine the implications that the business plan has on capital and identify the significant risks to which it is exposed. The following link (section 2500) provides the Standards of Practice on Dynamic Capital Adequacy Testing from the Canadian Institute of Actuaries effective January 1, 2012: mplete_janu Indicated RSR Target: A minimum m RSR balance of $190 million as of February 29, Pros: 1. Assists Management and the Public Utilities Board in the identification, measurement RE and mitigation of key risks faced by the Corporation. 2. Creates a forward rd looking measure of risk (i.e. not a retrospective measure like other tests). 3. Uses company specific assumptions for adverse scenarios, ripple effects and management ment action, as opposed to prescribed rules that are the same for every company. 4. Produces an opinion REVris inrin that is based on the RSR targets set by the Manitoba Public Utilities Board. REVISED ED 5. Creates a clear linkage between the required RSR and the amount of risk faced by the Corporation. 6. Is a recognized method of the Canadian Institute of Actuaries and the Office of the Superintendent of Financial Institutions (OSFI). 7. The adverse scenarios and associated assumptions can be discussed and debated at the General Rate hearings and, if warranted, modified in the next DCAT report. Since the DCAT is done in-house, the Corporation can provide Page 4

18 June 15, 2012 Rate Stabilization Reserve Part 1 AI.11 the impacts of alternate adverse scenarios at the request of the Public Utilities Board. Cons: 1. Certain risk factors may be difficult to quantify. 2. The results are not directly comparable to other jurisdictions. 3. Actuaries may have different opinions on likelihood and magnitude of given risk factors. AI.11.D Minimum Capital Test (MCT) Description 1 : The Minimum Capital Test (MCT) was developed by Canadian regulators to assess the key risks faced by the industry and to harmonize capital requirements across jurisdictions in Canada. It is a risk-based approach [that] better reflects the riskiness of individual P&C insurers and is consistent with approaches in other financial sectors. It assesses the riskiness iness of assets, policy liabilities, and offbalance sheet exposures, by applying various factors. Pros: Indicated RSR Target: A minimum capital (RSR) requirement of $283 million (i.e. 100% 0% MCT score). REVISED 1. Is designed to assess the key risks faced by the insurance industry, the majority of which are relevant to Manitoba Public Insurance. 2. Is used by other Rr insurers, including SGI and ICBC, and is recognized by OSFI. 3. Assesses the riskiness of assets, policy liabilities, and off-balance sheet exposures, by applying a consistent set of factors that were agreed upon by a task force of insurance experts. 4. Identifies risks based on the Corporation s current financial statements (or current risk profile). 1 All quotes in AI.11.C are from Notes on the Development of the Minimum Capital Test, Office of the Superintendent of Financial Institutions Canada, July 2003, or Guideline Minimum Capital Test (MCT) for Property and Casualty Insurers, Office of the Superintendent of Financial Institutions Canada, July Page 5

19 June 15, 2012 Rate Stabilization Reserve Part 1 AI The calculation of the MCT score is completely objective (i.e. no judgment is required). 6. The MCT score is relatively easy to calculate. Cons: 1. It is a private sector test. Policyholder and regulatory concerns with respect to insurer solvency are of a different magnitude and kind compared to the situation with a Crown Corporation. 2. The method uses the same risk factors for all companies, which may not be reflective of the Corporation s risk level. 3. It requires the Public Utilities Board and the Corporation to make a judgment AI.11.E on the appropriate MCT score for a monopoly insurance provider (e.g. 50% to 100% MCT was recommended in the Rate Application). Operational and Investment Risk Analysis Description: The Operational and Investment Risk Analysis (Risk Analysis) is designed to provide a risk assessment sessment relating to the financial outcomes for the Basic insurance program for Manitoba Public Insurance. As stated in Order 151/2000 number 3, the Risk Analysis incorporates the following criteria: At a 95.0% confidence level, both including and excluding operating costs; and At a 97.5% confidence level, both including and excluding operating costs Each of four scenarios should: REVISED ED REVng t Include only PIPP data; Include investment risk using a Value at Risk Analysis assuming a 25.0% equity component and a time horizon of between two and three years; Use variances between forecast and actual amounts for revenues, losses, operating expenses, and claim expenses; and Use actual correlations between all risk components recognizing the directional effect on net income. Indicated RSR Target: An RSR range of $193 million to $291 million. Page 6

20 June 15, 2012 Rate Stabilization Reserve Part 1 AI.11 Pros: 1. The analysis is easy to update. 2. The operational risk calculations are transparent. 3. The mathematical calculations of historic operational risk are objective. Cons: 1. The operational risk calculation is based entirely on an analysis of financial outcomes over the last eighteen years without any adjustment to reflect the Corporation s current cost levels. 2. The methodology does not reflect the Corporation s current risk profile as many changes to the Corporation s policies, procedures, contracts, etc. have been made over the past eighteen years (e.g.. reinsurance contracts, reserving guidelines, investment policies, etc). 3. The operational risk analysis is not based on plausible adverse events that could occur. Rather, it is based on events that have actually occurred in the past eighteen years. 4. If the Corporation improves its forecasting process, it will have a lower operational risk margin, since the margin depends on the deviations between actual and forecasted results. s. Alternatively, the risk margins can be increased by poor forecasting or other changes that are unrelated to operational risk. 5. The historical results R (currently only 18 data points) do not provide a credible sample for statistical modeling. For example, a general rule of thumb for obtaining a statistically reliable correlation coefficient would be a minimum of 30 data points for normally distributed variables and as much as 100 or more for non-normally ly Rnts distributed variables. REVISED ED 6. Using Value-at-Risk (VaR) may not be appropriate for a time horizon of 2.5 years. In all of the Corporation s research regarding VaR the Corporation has never seen an author advise calculating VaR for a period longer than one year. 7. The risk margins produced do not provide the Corporation or the Public Utilities Board with a clear understanding of the risks faced by the Corporation, how these risks should be mitigated, or what management or Public Utilities Board action may be required. Page 7

21 June 15, 2012 Rate Stabilization Reserve Part 1 AI The method is not used or recognized by any other insurer, regulator, or professional body. AI.11.F Percentage of Premium (Kopstein) Method Description: In the 1989 Kopstein Report on Auto Insurance in Manitoba recommendation 7.1 stated: That the government of Manitoba issue a public directive to the corporation setting an Autopac retained earnings surplus target of about 15% of premiums. (This would amount to $40 to $50 million at prevailing premium levels.) The government directive should indicate that, if Autopac surplus falls below 10 percent or exceeds 20 percent of premiums, ms, the corporation should and would be expected to take remedial action. The Percentage of Premium method for determining the RSR, which is the method currently used by the PUB, is based on the recommendation from the Kopstein report. The RSR range for a given year is determined based on 10% to 20% of net written premiums. Pros: Cons: Indicated RSR Target (2012/13): An RSR range of $77 million to $153 million. REVISED ED 1. The indicated RSR range is easy to understand and calculate. 2. The mathematical matical calculations are objective. 1. The method assumes that the Corporation s risk level is a function of its annual premium level; however, the Corporation s main risks are from changes to assets and liabilities which are significantly larger than annual premiums. 2. It does not assist Management or the Public Utilities Board in the identification, measurement and mitigation of key risks. 3. The method does not create a clear linkage between the required RSR and the amount of risk faced by the Corporation. Page 8

22 June 15, 2012 Rate Stabilization Reserve Part 1 AI The indicated RSR range does not change when the Corporation s risk profile changes. 5. To the Corporation s knowledge, the method is not recognized or used by any other Regulator or professional body. AI.11.G Why the DCAT should be used to set the RSR Target The Corporation recommends that the DCAT be used to set the RSR target. The Corporation s position on this matter has not changed since it last filed a Discussion of the Rate Stabilization Reserve in AI.18.1 in the 2010 Rate Application. The key benefits to using the DCAT to set the RSR target, compared to using the MCT, RA/VaR, or Percentage of Premium (PoP) methodologies, are as follows: 1. Assists Management and the Public Utilities Board in the identification, measurement, and mitigation of key risks faced by the Corporation. DCAT: MCT: RA/VaR: Risks are clearly identified and quantified with supporting rational. Adverse scenarios are run through the Corporation s financial model showing the expected impact on RSR. Realistic Management and Regulatory actions are incorporated into the scenarios. It is not clear how the risk factors were developed, if they are esent.re appropriate for a public insurer, or what specific risk events they represent. esent. REVISED ED The current enren level of operational risk is assumed to be reflective of the Corporation s historical forecasting experience. Specific operational risks are not identified. The Value-at-Risk method of measuring investment risk is not appropriate over a 2.5 year time period. The method is independent of the risk level of the Corporation. 2. Creates a forward looking measure of risk (i.e. not a retrospective measure like other capital tests). Page 9

23 June 15, 2012 Rate Stabilization Reserve Part 1 AI.11 DCAT: The adverse scenarios are tested against the Corporation s pro forma financial statements and are assumed to occur in the rating period of the current application. MCT: Required capital is based on the current balance sheet, which is an improvement over the RA/VaR and Percentage of Premium methods. However, it is unclear what specific future risks the required capital is providing protection from and whether such an amount is appropriate for MPI. RA/VaR: The indicated RSR is based completely on historical forecasting performance. Historical results are not at the same cost level as future experience and include events that are not reflective of future risk (e.g. $250 million reduction in claim liabilities). The indicated RSR range does not change based on the Corporation s risk profile. 3. Uses company specific assumptions s for adverse scenarios, ripple effects, and management action, as opposed to prescribed rules that are the same for every company. DCAT: The adverse scenarios are developed specifically for MPI. MCT: It is not clear how the risk factors were developed, if they are RA/VaR: appropriate for a public insurer, or what specific risk events they represent. REVISED ED ibed r,rev Uses historical MPI forecasting performance for operational risk; however, the sample size of 18 data points (i.e. years of history) is not a statistically reliable sample. Also, past forecasting performance is not necessarily relevant to future forecasting performance. The method is independent of the risk level of the Corporation. 4. Produces an opinion as to the adequacy of the basic insurance plan reserve that is based on the RSR targets set by the Manitoba Public Utilities Board. DCAT: A satisfactory opinion is based on the Regulatory rules in this province. Page 10

24 June 15, 2012 Rate Stabilization Reserve Part 1 AI.11 MCT: RA/VaR: It is unclear how the private sector minimum regulatory target of 150% MCT is applicable to a public insurer. In effect, the PUB would set the target, which would, by implication, be adequate, with no inherent process to identify and quantify plausible risks. In effect, the PUB would set the target, which would, by implication, be adequate, with no inherent process to identify and quantify plausible risks 5. Creates a clear linkage between the required RSR and the nature as well as amount of risk faced by the Corporation. DCAT: MCT: RA/VaR: The indicated RSR target is the amount of RSR the Corporation requires at the evaluation date to maintain RSR above zero under all plausible adverse scenarios. It is not clear how the risk factors were developed, if they are appropriate for a public insurer, or what specific risk events they represent. It is not clear why the indicated level of RSR is required. The method is independent of the risk level of the Corporation. 6. It is a recognized method of the Canadian Institute of Actuaries and OSFI. DCAT: MCT: RA/VaR: Is used by all Canadian Property and Casualty insurers that report to OSFI. REVIS EVISED ED Is used by all Canadian Property and Casualty insurers that report to OSFI. Is not used or recognized by any other Regulator or professional body. To the Corporation s knowledge, is not used or recognized by any other Regulator or professional body. 7. The adverse scenarios and associated assumptions can be discussed and debated at the General Rate hearings and, if warranted, modified in the next DCAT report. Page 11

25 June 15, 2012 Rate Stabilization Reserve Part 1 AI.11 DCAT: The adverse scenarios are tailored to MPI s operations and can be continuously improved over time. The RSR target is based on risks specific to MPI. MCT: The risk factors are supplied and cannot be changed. There is no judgment required to calculate the required capital. The risk factors are not specific to MPI. RA/VaR: The method uses prescribed calculations and is not responsive to risks specific to MPI. AI.11.H The method uses prescribed calculations and is not responsive to risks specific to MPI. Capital Targets used by other Public Insurers Saskatchewan Government Insurance uses an MCT target range of 75% to 150% for the Auto Fund. If these targets were used for MPI s basic compulsory program the RSR range would be $212 million to $425 million. The Insurance Corporation of British Columbia has a statutory minimum capital requirement of RE 100% MCT for its basic line of business. They also have a Capital Management Plan (approved by their regulator) with a target MCT of 130%. If these targets were used for MPI s basic compulsory program the RSR range would be $283 million to $368 million. Based on the Corporation s oration s calculated DCAT, MPI s RSR should be $190 million. The REVISED RSR suggested by the Saskatchewan SRS and British Columbia method would result in MPI requiring an RSR much higher than we believe is reasonable or warranted, based on plausible adverse scenarios appropriate to MPI s Basic program. To achieve this level of RSR would mean increasing rates to generate higher retained earnings for the reserve. Page 12

26 June 15, 2012 Rate Stabilization Reserve Part 1 AI.11 AI.11.I Recommendation 1. That the PUB adopt the Dynamic Capital Adequacy Test as the method for determining the MPI required RSR and agree to an RSR target for 2012/13 of $190 million. The above analysis and supporting documentation provides a sound rationale for the PUB to adopt the Dynamic Capital Adequacy Test as the method for determining the required RSR. The DCAT explicitly measures the potential financial impact from the Corporation s key risk factors and produces a RSR target that is directly related to the Corporation s risk level and directly responsive to the purpose of the RSR. 2. That the PUB continue to monitor the Basic insurance MCT score and the RSR targets established by other Crown auto insurers on the basis that it is an industry standard approach used by other similar entities. 3. That the PUB discontinue monitoring ng the results obtained through the Risk Analysis/VaR calculations, based on the many shortcomings identified herein. 4. That the PUB discontinue monitoring the results obtained through the Percentage of Premium ( Kopstein ) approach for the primary reason that the Basic insurance plan s greatest risks are to its assets and liabilities and bear no relationship to its annual revenues. REVISED Page 13