Instructions for Completion of Requested Input to OSFI/AMF Quantitative Impact Study of the New Standardized Approach to Calculating the Solvency Buffer for Market Risk

Document Properties

• Type of Publication: Instructions

Thank you for participating in this quantitative impact study. The purpose of this study is to assist the joint committee of OSFI, AMF and Assuris (Standard Approach Working Group) in assessing the impact of the proposed regulatory capital changes related to market risk.

In addition to supplying the requested information, your written comments on the results of this QIS will be useful and appreciated.

The basic information required for this study is to be entered in the enclosed Excel Workbook. A Worksheet is included in the Workbook for insurers to provide supplementary information and any comments they might have. The calculation of the solvency buffer for each component is intended to be in accordance with the Standardized Approach for Calculating the Solvency Buffer for Market Risk (Market Risk Paper). The instructions set out below contain extracts from the Market Risk Paper as well as additional explanations to assist insurers in completing the calculations and the Excel Worksheets.

All information is to be calculated as of December 31, 2009, using year end 2009 data. For insurers with a year end other than December 31, the insurer will use their year-end data. All amounts are in thousands of dollars.

It is recognized that under IFRS, certain liabilities presently included under the CALM methodology will be reclassified as Investment Contract liabilities since they do not meet the definition of insurance products under IFRS 4. Certain information is requested related to these liabilities. A brief description of products that have been determined to not meet the definition of insurance risk would be appreciated as well as the basis for the classification and the proportion of the deposit type/GIC products that are reclassified to be accounted for under IAS 39.

Certain information related to participating products is also requested in the Worksheets for purposes of understanding the impact of the capital standard on participating products.

Information is generally requested for Canada, United States, United Kingdom, Europe and Asia and other because these match up with the major geographic, currency and interest rate environments.

Summary Page

The Summary Page contains three separate sets of information.

The first set is the solvency buffers for each of the components of Market Risk. These are automatically sourced from each of the Worksheets. The “Reduction for Participating Products” is sourced from the calculation at the bottom of the Summary Page. Information on how to calculate these amounts is provided below.

The second set calculates the solvency buffer under the existing capital and accounting standards. The components should equal the amounts reported in your December 31, 2009 MCCSR^{Footnote 1} return or Report of the Appointed Actuary, as applicable. The amount titled “Reduction for Participating Products under Current MCCSR Rules” is to be calculated as the increase in required capital that would occur if the components of the solvency buffer (related to market risk only) under the existing capital and accounting standards were to be calculated as if all business was non-participating.

The third set of information is used to calculate the Adjustment for Participating Products under the new Market Risk capital standard as described below.

Companies are also reminded to enter their company name in the place provided at the top of the Summary Page.

Adjustment for Participating Products

The “Maximum Reduction in the Present Value of Future Dividends from Reducing Dividend Scales” is to be based on management’s judgment as to the amount of dividend reduction that could reasonably be made in a one year time horizon to offset significant adverse experience.

Insurers should calculate the dividend offset by major blocks of par business (each significant group of par business; generally defined as groups of policies with the same dividend scale). The total dividend offset calculated is carried forward to the Summary Page.

The methods used and assumptions made to calculate the dividend offsets should be justified and described in detail as requested in the Questions and Comments Worksheet.

The present value of the reduction in future dividend cash flows that would result should be calculated using the interest rates used to determine the interest rate risk solvency buffer for each jurisdiction. The second amount requested is the “Amount by which existing MCCSR required capital has been reduced (par vs. non-par requirements) for components other than those listed above”. This represents the amount by which the required capital under existing MCCSR rules (for all components except those designated on this Worksheet as related to market risk) would be increased if calculated as if all products were non-participating.

To assist the Standard Approach Working Group to determine the appropriate approach to the Adjustment for Participating Products, certain additional information is requested from the insurers, as follows.

Components of Dividends and Dividend Reduction – On the Asset and Liability Cash Flows Worksheet, the present value of the dividend cash flows are to be split between the investment, underwriting (mortality and morbidity) and expense components. Where this information is not readily available, insurers may use a simplified approach to make the necessary calculations. The method(s) of determining the components is to be explained in the Questions and Comments Worksheet of the QIS. Please also include an explanation of any special circumstances affecting the calculation of the dividend reduction including the treatment of terminal dividends and lapse supported products. Also, the Dividend Reduction calculated by the insurer is to be split into these three components and this information is also to be included in the Questions and Comments Worksheet. Where material, terminal dividends should be separately disclosed in the space provided.

Impact of Adverse Lapses, Anti-selective Mortality and Other Costs caused by a Dividend Reduction – It is expected that any non-trivial reduction in dividends will have negative impacts on the insurer due to adverse lapses, anti-selection, unit expense increases or even class actions by policyholders. For purposes of this QIS, these impacts are not to be reflected in the cash flows. Insurers are, however, asked to provide supplemental information to the Standard Approach Working Group by disclosing in the Questions and Comments Worksheet the estimated amount of these negative impacts of the dividend reduction taken. Also provide an explanation as to how the estimated costs were determined.

We also request the insurer disclose the present value of all the costs (excess mortality, excess cash values, excess expenses, etc.) resulting from additional lapses of 10% (over best estimate and PfADs) of the policies per year to be included in the Questions and Comments Worksheet. For example, if the best estimate assumption with PfAD is 5% then use a 15% assumption for this calculation. This should be calculated based on the worst scenario (scenario used to calculate the interest rate risk buffer for each major geographic segment) and, if necessary, approximations may be used.

Adjustable Products

The Standard Approach Working Group recognizes that there may need to be a suitable adjustment for products with adjustable features that mitigate the impact of market risk; however, this has yet to be determined. For purposes of this QIS, no adjustment is included for adjustable products.

To assist the Standard Approach Working Group in understanding the extent of products that might be subject to an adjustment, insurers are requested to provide additional information regarding the description of blocks of adjustable products (e.g. non par adjustable T100 individual life) and the present value of the related liability cash flows using the base scenario interest rates.

Insurers are also asked to indicate which risks (investment risk, underwriting risk, expense risk) the product may be adjusted for and a brief description of the adjustable feature (e.g. premiums adjustable every 5 years). A Worksheet titled “Adjustable Products” is provided for this purpose. Additional lines should be added to the extent necessary.

Interest Rate Risk

Interest rate risk is the risk of economic loss resulting from market changes in interest rates. The most significant part of this risk is the risk of economic loss due to the effect of the volatility and uncertainty of future interest rates on the mismatch of cash flows from interest sensitive assets with liability cash flows.

Projection of Cash Flows

The asset and liability cash flows are to be projected separately by month for the first year and yearly thereafter. Add all cash flows over 30 years and present the total asset and liability cash flows in “year 31 and beyond”. Calculate and disclose the durations of the over 30 year asset cash flows and the over 30 year liability cash flows. For each month and year presented, deduct the liability cash flows from the asset cash flows to arrive at the net cash flows.

The asset, liability and net cash flows are to be separately disclosed for each of the following liability categories:

• non par products

• pass through products

• investment contracts. That is deposit type contracts that will be accounted for as investment contracts under IFRS (excluding any cash flows from segregated funds and service contracts).

• participating products

• capital and surplus

• the total

For participating products, the par dividend cash flows are to be provided by month for the first year and yearly thereafter.

The asset, liability and net cash flows are to be projected separately for Canada, United States, United Kingdom, Europe, Asia and other for each of the categories above.

For each month and year presented, calculate and disclose the present value of the asset cash flows, liability cash flows and the net cash flows for each of the categories using the discount rates determined under the instructions titled “QIS Discount Rates”. For cash flows over 30 years, calculate the present value to year zero and show the total present value in “year 31 and beyond”. Note that the total present value of the assets less the total present value of the liabilities should equal the present value of the net cash flows.

The cash flows are to be determined using the instructions titled “Cash Flows”.

Cash Flows

When IFRS Phase II for Insurance Contracts is introduced the liability cash flows for capital purposes will be projected using assumptions that are consistent with IFRS. For purposes of this QIS, cash flows should be calculated using the assumptions that are consistent with those used in preparing the CALM actuarial liabilities except with no reinvestment assumptions for assets or liabilities. The cash flows are to be included in the Worksheet titled “Asset and Liability Cash Flows”.

Provisions for Adverse Deviation

The cash flows will be determined as at December 31, 2009 using best estimate cash flows including Provisions for Adverse Deviations (PfADs) except for PfADs for interest rate risk (C3 provisions). Provisions for asset default risk (C1 provisions including PfADs) should be included in the cash flows (whether deducted from asset cash flows or included in the liability cash flows which ever is the insurer’s practice). Liability cash flows should include PFADS for insurance (mortality, morbidity and lapse) and expense risk PfADs.

PfADs are required for CALM purposes and are often used by insurers for Asset Liability Management. When risk margins are determined under IFRS Phase II, we will review the appropriateness of including the risk margins in cash flows.

Insurers are asked to indicate in the Questions and Comments Worksheet whether PfADs should be included in the cash flows. If possible, please quantify the impact of excluding the PfADs on the solvency buffer calculated.

Future Income Taxes

Future income tax cash flows resulting from timing differences in an IFRS environment will be excluded from the cash flows as deferred income tax balances are not expected to vary with interest rates in an IFRS environment. For purposes of this QIS, future income tax cash flows will be included in the cash flows as they are generally included by insurers in their Asset Liability Management.

Asset Cash Flows

For assets with fixed cash flows that are matched with policy liabilities, insurers should use the same cash flows as used in preparing the CALM valuation. These cash flows are net of investment expenses and expected default provisions (including C1 PfADs). No reinvestment of asset cash flows should be assumed. Only contractual asset cash flows should be projected

For assets with fixed cash flows matched with surplus, insurers would use the cash flows that would be the gross cash flows adjusted for investment expenses and any expected default provisions (including C1 PfADs). A fixed cash flow is a cash flow with contract guaranteed amount and whose value is not contingent on future market prices or future interest rates.

For assets with non-fixed cash flows including equities and real estate, insurers would use their balance sheet value at time zero less the present value of those assets’ fixed cash flows. Common share dividends are to be excluded from the cash flows.

Real Estate

For real estate, insurers are permitted to include the fixed cash flows in the applicable period they are contractually expected to be received (i.e. leases in force – no renewals to be assumed). Insurers would then include, at time zero, the balance sheet value less the present value of fixed cash flows. Where the fixed cash flows are not included in the applicable period, insurers would use the real estate’s full balance sheet carrying value at time zero.

Preferred Shares

For preferred shares, the fixed cash flows would be allowed to be recognized to the first call or redemption date and the redemption proceeds would be recognized on the redemption date. Insurers would then include, at time zero, the balance sheet value less the present value of fixed cash flows. For preferred shares that are currently redeemable, insurers would use the balance sheet value at time zero.

Innovative Tier 1

For innovative Tier 1 instruments, the fixed cash flows would be allowed to be recognized to the first call or redemption date and the redemption proceeds would be recognized on the redemption date. Insurers would then include, at time zero, the balance sheet value less the present value of fixed cash flows.

Pooled Funds - Pass Through Products

Where the account value of a policy varies directly with a pooled fund and the pooled fund is held by the insurer, cash flows should be projected in the Pass Through Product Risk Worksheet. The pooled funds including mutual and segregated funds may contain bonds, equities or other assets.

Pooled Funds - Non Pass Through Products

Where the account value of a policy does not directly vary with a bond fund, cash flows should be included so that the value of the bond fund changes appropriately with the change in interest rates in each scenario. Disclose the balance sheet value of these bond funds included in the base scenario in the Questions and Comments Worksheet.

For mutual and pooled funds that hold assets with non-fixed cash flows (e.g. equities, real estate), insurers would use their balance sheet value at time zero.

Securitized Assets

For securitized fixed income assets, such as bonds and mortgages that are used to match policyholder benefits, insurers may include the underlying fixed cash flows in the projected asset cash flows.

Disclose the balance sheet value of these securitized bonds included in the base scenario in the Questions and Comments Worksheet.

For securitized assets that are held in non-fixed cash flows, insurers would use their balance sheet value at time zero.

Interest Rate Swaps

For an interest rate swap, an amount of nominal cash flow should be added to the asset or liability cash flows at the term of the fixed rate portion of the swap. An offsetting amount should be added to the time zero cash flows. The reported offset may be positive or negative, and an asset or a liability. However, any reported nominal cash flows should be such that when the interest rate scenarios are applied, the change in the net present value amount of the nominal cash flows equals the change in market value of the swaps. If a swap is tied to a specific investment, the swap may be netted with the investment’s reported interest cash flows.

Reverse Mortgages and Collateral Loans

For reverse mortgages and collateral loans with fixed interest rates, cash flows should be projected using reasonable assumptions including mortality. If the assets have variable interest rates, the assets should be shown as time zero cash flows.

Policy Loans

For policy loans with interest rates that are fixed or subject to guaranteed minimums, the cash flows should be projected using insurance risk PfADs (e.g. mortality and lapse rates).

For variable rate policy loans that are not subject to guaranteed minimums, policy loan amounts should be shown as time zero cash flows.

Participating Dividends and Cash Flows

For participating business, the dividend cash flows (par best estimate cash flows) included in the December 31, 2009 actuarial liabilities are to be included in the Worksheet in the applicable column. Separate columns are provided for regular and terminal dividends (where material). Where terminal dividends are not material they may be included in with regular dividends. Also the present value of these dividend cash flows are required using the QIS discount rates. As discussed in the Adjustment for Participating Products above, the present value of the dividend cash flows are to be split into components related to investment experience, underwriting experience and expense experience.

The dividend cash flows are also to be included in the liability cash flow in the Worksheet in the appropriate column. Thus the liability cash flows are the best estimate liability cash flows (including dividends) and PfADs.

Deposit Liabilities

For deposit taking type business which will be reclassified out of actuarial liabilities under IFRS 4 and classified as Investment Contracts, the related liability cash flows, PfADs, asset cash flows and the Account Value are to be placed in their separate column in the Worksheet. The Account Value is the amount the insurer expects to carry the liability at under IFRS accounting if they carry these liabilities using the effective interest method^{Footnote 2}.

Assets Backing Surplus and Other Liability Cash Flows

Cash flows from assets backing capital and surplus and any other liability cash flows not related to products should be included in the respective “Capital and Surplus” columns by region in the Cash Flow Worksheet, and are to be included in all calculations of market risk.

Universal Life

One of the basic assumptions used in this QIS is that only contractual cash flows should be projected. No reinvestments should be assumed. However with universal life, the contract continues after the end of any interest guaranteed period in the inside investment account. Therefore, a reinvestment assumption is required in order to project realistic cash flows for premiums, policy charges and benefits and expenses.

After IFRS Phase II is introduced, the reinvestment assumption used should be consistent with the reinvestment assumption used in the IFRS valuation. It is not yet clear what that will be. Insurers should use assumptions that are consistent with their CALM valuation. Ensure these reinvestment assumptions vary appropriately with the scenario that is being tested. Where the universal life contract has minimum interest guarantees, the effect of these guarantees should be reflected in the scenario that is being tested. Disclose the method used to reflect the effect of these guarantees in the Questions and Comments Worksheet.

Where the performance of a universal life contract inside account benefit is tied to the performance of specific assets and these assets are held by the insurer, then the cash flows on these assets and liabilities should be included in the Pass Through Product Worksheet.

If the matching assets are not held, then the cash flows should be projected using assumptions that are consistent with their CALM valuation.

No adjustments will be made to the universal life cash flows due to anticipated changes in lapse rates and expense charges with the change in interest rates in each scenario. A solvency buffer for these changes will be considered under insurance risk.

Base Scenario Discount Rates

The Standard Approach Working Group considers that it would be preferable if the discount rates to be used for the interest rate shocks be consistent with IFRS Phase II for Insurance Contracts. The determination of an appropriate discount rate is still under discussion. However, for the purposes of this QIS we are testing three discount rates, risk fee rates and two risk free rates plus a spread. For the base scenario, assume the discount rates are the risk free interest rates plus a spread based on corporate A rated bonds as set out in the attached worksheet.

The risk free interest rates for Canada are the spot rates for Government of Canada bonds. The rates for the United States are the spot rates for applicable United States Treasuries. The rates for United Kingdom are spot rates for UK sovereign benchmark bonds. For Europe the spot rates are the Government of Germany bonds. For Japan the spot rates are Government of Japan bonds. The rates for other jurisdictions are to be determined by insurers on an equivalent basis for sovereign risk bonds of the applicable jurisdiction. Interest rates need only be applied to jurisdictions which are material in the circumstances.

The discount rates are to be calculated based on the methodology set out below.

The discount rates to be used in the base scenario (Base Scenario Discount Rates) are:

1. For cash flows from day 0 to year 20 year, the risk free interest rates plus the spread.

2. For cash flows from years 20 to 30, linearly interpolated rates between 20-year discount rate and the terminal reinvestment rate determined in c) Text for screen readers: c = 3 below.

3. For cash flows from beyond 30 years, the terminal reinvestment rate, calculated as the average of the last 10 year-end risk free spot rates for 20-year obligations plus the average of the 20-year spreads for last 10 year-ends.

All interest rates are as at December 31, 2009. For insurers with a year end other than December 31, the insurer will use the discount rates as at December 31, 2009.

The rates for Canada, United States, United Kingdom, Europe and Japan are set out on the attached worksheet. For other jurisdictions please use an appropriate risk free rate and disclose the methods and rates used.

Interest Rate Shock Method

The standardized approach will use a projected cash flow methodology that measures the economic impact of a sudden change in interest rates at time zero. The procedure to calculate the solvency buffer is to project the contractual cash flows from all assets and liabilities (best estimate liability cash flows, PfADs (excluding C3) and participating dividends) held on the reporting date using the previous instructions titled “Cash Flows” unless otherwise instructed below. We have provided the discount rates to be used for the Base Scenario, the Stressed Discount Rates and the additional discount rate scenarios for each significant jurisdiction in the Worksheet titled “Discount Rates”.

Stressed Discount Rates

For each test scenario, the Stressed Discount Rates will be calculated as follows:

1. For cash flows from day 0 to year 20, adjust the Base Scenarios Discount Rates by calculating:

   • the adjustment to the 90-day discount rate (T)

   • the adjustment to the 20-year discount rate (B)

   • the adjustments for all intervening years by applying linear interpolation to the coefficients used to calculate the two adjustments (T) and (B) above.

2. For cash flows from years 20 to 30, interpolate the rates between the adjusted 20-year discount rate and the adjusted terminal reinvestment rate determined in c) below.

3. For cash flows beyond 30 years, add the adjustment (L) to the terminal reinvestment rate.

1. Base Scenario

   1. Discount the asset and liability cash flows to year zero using the Base Scenario Discount Rates to produce the asset present value and liability present value.

   2. Deduct the present value of the liability cash flows from the present value of asset cash flows. This gives a net present value of the cash flows of this base scenario.

2. Test Scenarios

   1. For each test scenario, discount the asset and liability cash flows using the QIS stressed discount rate to produce the asset present value and liability present value.

   2. Deduct the present value of the liability cash flows from the present value of asset cash flows. This gives a net present value of the cash flows of each test scenario.

The test scenarios are as follows:

1. Lower short term interest rate (after shock T2), lower long term interest rate (after shock B2) and lower terminal reinvestment rate (after shock L2)

2. Higher short term interest rate (after shock T1), lower long term interest rate (after shock B2) and lower terminal reinvestment rate (after shock L2)

3. Higher short term interest rate (after shock T1), higher long term interest rate (after shock B1) and higher terminal reinvestment rate (after shock L1)

4. Lower short term interest rate (after shock T2), higher long term interest rate (after shock B1) and higher terminal reinvestment rate (after shock L1)

Definition of Interest Rate Shocks

• T1 is the estimated 99.5% percentile of the potential upward change in 90 day discount rates over one year

• T2 is the estimated 99.5% percentile of the potential downward change in 90 day discount rates over one year

• B1 is the estimated 99.5% percentile of the potential upward change in 20 year discount rates over one year

• B2 is the estimated 99.5% percentile of the potential downward change in 20 year discount rates over one year

• L1 is the potential upward change in the terminal reinvestment rate (10 year average of the 20 year spot rate) over one year

• L2 is the potential downward change in the terminal reinvestment rate (10 year average of the 20 year government spot rate) over one year

The interest rate shocks (T1, T2, B1 and B2) to be used are the following linear functions of the square roots of the current risk free rates r, where r is expressed as a decimal (for example five percent will be expressed as 0.05). This function is based on a simplified Cox-Ingersoll-Ross model fitted to the historic data. See Appendix III for a more detailed description of the method used for the shocks.

If the current 90-day risk free rate is r_0.25 , then:

 

If the current 20-year risk free rate is r₂₀ , then:

 

If the risk free rate for term t is equal to r_t , where t is between 90 days and 20 years, then the changes interpolated for the time t rate under the four test scenarios, which in all cases are added to the base rates, are:

 

When adjusting the terminal reinvestment rate:

L1 is + 50 basis points

L2 is - 50 basis points

When interest rates are stressed the liability cash flows should be adjusted to account for any interest rate guarantees.

Participating Products

Cash flows from participating products will be those used in the calculation of best estimate policyholder obligations adjusted for PfADs, including dividends. The dividend cash flows used in the base scenario should be the best estimate cash flows that reflect any anticipated future change in the dividend rate based on the base scenario discount rates. No adjustment to cash flows will be made for anticipated reductions or increases in dividends that may result from increases or decreases in interest rates in each scenario. A reduction in the solvency buffer for potential reductions in dividends will be separately calculated for the participating products.

III. Resulting Solvency Buffer

For insurers operating only in Canada, the solvency buffer is the difference between the total net present value of the base scenario and the lowest total net present value of the test scenarios.

For insurers operating in Canada and the United States, the solvency buffer for purposes of this QIS (i.e. final approach has not yet been determined) will be based on the scenario which produces the highest buffer on a combined basis for the two currency/interest environments. This recognizes that interest rates in the two countries tend to move in similar directions. The test scenario to be used is the scenario which produces the greatest difference between the total net present value of the base scenarios and the lowest total net present value of the test scenarios (in total for Canada and the United States), but if a scenario results in a negative buffer for a jurisdiction, that jurisdiction’s buffer will be set to zero for purposes of calculating the combined buffer for that scenario. The resulting solvency buffer for each country, therefore, cannot be lower than zero. See example in Appendix I. The SAWG will also use the QIS data to test the impact of using the worst scenario buffer in each country.

The solvency buffer for other jurisdictions (United Kingdom, Europe and Asia) is the difference between the total net present value of the base scenario and the lowest total net present value of the test scenarios for each jurisdiction, respectively.

Additional Discount Rate Scenarios

In addition to the information requested above, two additional set of calculations are requested. These additional calculations will use the same information and the same methodology as in the previous calculations described above, except that the different Base Scenario Discount Rates will be used and the resulting interest rate shocks will need to be adjusted accordingly. For specific instructions for the additional scenarios please see Appendix II.

Except for the use of these different interest rate shocks all other aspects of the calculations should be the same as previously described. Results of the calculations should be reported on the worksheet titled “Interest Rate Risk Addendum”, which is same as the Worksheet titled “Interest Rate Risk”. For purposes of this QIS, the cash flows to be used for the additional discount rate scenarios will be the same cash flows used for the Base Discount Rate Scenario.

Interest Rate Spread Risk

This component has been excluded for purposes of this study. This may be tested at a later time.

Equity Risk

Equity market risk is the risk of economic loss due to changes in the prices of common shares this includes both the systematic and specific components common share price fluctuation.

The solvency buffer for all investments classified as equities (including equity index securities and managed equity portfolios) excluding preferred shares will be calculated using a simple immediate (time zero) deterministic downward shock of 30% to the market value of the equities.

Where equities are being used to back equity investment options in non participating products, such as universal life insurance, the equity risk is being passed through to policyholders. The solvency margin will be replaced with the correlations requirement in the current MCCSR guidelines as defined in the Pass Through Products Risk section below.

Separate solvency buffers should be calculated for equities held to back participating and non- participating business.

Equity and Real Estate Mutual and Pooled Funds

The solvency buffer for investments in mutual and pooled fund will be calculated using the shock for equities or real estate. The factor to be used for each fund is the highest factor (i.e. equity shock or real estate shock) applicable to any security that the fund holds or is permitted to invest in. Balanced funds that hold equities will be treated as if they are equity funds.

Real Estate Risk

Real estate market risk is the risk of economic loss due to changes in the amount and timing of cash flows from investments in real estate.

For income producing real estate, the solvency buffer will be calculated in two parts, a credit risk part and a market risk part.

The first part will be a credit risk based calculation on the present value of the fixed lease cash flows that are contractually expected to be received (i.e. leases in force – no renewals to be assumed). These are the same cash flows as was used in the calculation of Interest Rate Risk Buffer. The categories^{Footnote 3} and the related factors used in this QIS are consistent with those used in the Credit Risk QIS completed in the fall of 2009. The present value of the lease cash flows should be allocated to the categories listed in the QIS forms based on rating agency issuer ratings of the specific lessees. The present value related to lessees without external ratings will receive the highest factor.

The second part will be a market risk based calculation using an immediate (time zero) deterministic downward shock to the residual market value of the real estate. The residual value is the market value at year-end 2009 net of the present value of the fixed cash flows that are contractually expected to be received as determined above. The calculation of this part will be a 30% immediate decline in the residual market value of real estate. The total real estate carrying value, total market value and total solvency buffer are to be included by jurisdiction in the Real Estate Risk Worksheet.

For non income producing real estate; own use real estate; oil and gas properties; and timberland and agriculture properties, the solvency buffer will be calculated using a 30% downward shock to the market value of the real estate. The solvency buffer is the difference, if positive between the carrying value on the balance sheet at year-end 2009 (moving average market), and 70% of the market value at year-end 2009. The solvency buffer must be determined on a property by property basis. The total real estate carrying value, total market value and total solvency buffer are to be included by jurisdiction in the Real Estate Risk Worksheet.

We recognize that under Canadian GAAP carrying value may differ from the market value. The total carrying value and the total market value of real estate is required to be disclosed in order that this differential may be appropriately incorporated into our analysis to reflect the fact that available capital will likely be different under IFRS than at year end 2009.

Separate solvency buffers should be calculated for real estate held to back participating and non- participating business.

Pass Through Products Risk

The existing capital requirements include a calculation of capital required for index linked pass through products. The amounts included on the existing MCCSR page 35.010 should be included on this worksheet up to and including Line 19 of 35.010 (accept as modified to include a new threshold of CF<70% and the change in factor from 15% to 30%), which responds to the market risk related to such products. For products with a 30% factor, insurers could alternatively use for those products (and their matched assets) the requirements for credit risk and other categories of market risk, including the interest rate risk and equity risk.

Currency Risk

Currency market risk is the risk of economic loss due to changes in the amount and timing of cash flows arising from changes in currency rates of exchange.

The existing capital requirements include a calculation of capital required for currency risk. For purposes of this QIS, include on the Summary Page the amounts reported on Page 90.025 Line 489 Column 36 of the insurer’s MCCSR 2009 year end filing.

Liability Market Options Risk

Liability market options risk is the risk of economic loss due to changes in the amount and timing of cash flows related to all market related options and guarantees in the liabilities. These include minimum interest rate guarantees in and accumulation annuities and guarantees of segregated fund performance. In calculating cash flows under each interest rate scenario, interest rate guarantees should be taken into account with future cash flows changed as appropriate.

There are no QIS requirements with respect to Segregated Fund Guarantees (Liability Market Options Risk) at this time as the requirements are currently under development. For purposes of this QIS, include on the Summary Page the amounts reported on Page 90.010 Line 230 Column 8 of the insurer’s MCCSR 2009 year end filing.

Asset Market Options Risk

There are no requirements with respect to Asset Market Options Risk at this time. Under each interest rate scenario, assets with callable/extendible features should be taken into account with future cash flows changed as appropriate. For other options, amounts included in the current MCCSR are to be reported.

For purposes of this QIS, include in the Asset Market Options Risk Worksheet the amounts reported on pages 31.021 and 32.021 line 90, column 5 of the insurer’s MCCSR 2009 year end filing. The amount on page 32.021 related to qualifying participating shares must be doubled to reflect non-par factors.

Questions and Comments

Space is provided for comments covering a number of topics. Insurers are required to respond to the specific questions and are encouraged to provide additional comments.

Appendix I - Resulting Solvency Buffer Example

Where an insurer operates in Canada and the United States, the solvency buffer for interest rate risk is determined using the same test scenario for both countries as interest rates are highly correlated. The following illustrates how to calculate the buffers for interest rate risk for these two jurisdictions.

Interest Rate Risk - Solvency Buffer

(amounts in thousands)

*Determining Worst Scenario:

Appendix II - Additional Interest Rate Scenarios

1. Base scenario discount rate without a spread

   In the interest rate shock test provided in the QIS, the interest rates are based on risk free rates plus a spread. Insurers are required to complete the same calculations for the base scenarios and interest rate shocks but using risk free interest rates without a spread and for years beyond 30 a terminal reinvestment rate. The discount rates are:

   1. For cash flows from day 0 to year 20, the risk free interest rates.

   2. For cash flows from years 20 to 30, the linearly interpolated rates between 20-year discount rate and the terminal reinvestment rate determined in c) Text for screen readers: c = 3 below.

   3. For cash flows beyond 30, the terminal reinvestment rate, calculated as the average of the last 10 year-end risk free spot rates for 20-year obligations.

   Since the shocks in the base scenario are calculated using the risk free rates (i.e. the spread is not included in r), the shocks will be the same for this scenario as they are for the scenario that includes spreads.

2. Base scenario discount rate with spread with a terminal reinvestment rate beyond year 30

   A second additional series of interest rate shock tests are required to be completed for informational purposes only. For these shock tests, insurers will use the risk free rates plus a spread up to year 30, and a terminal reinvestment rate beyond 30. The discount rates to be used are calculated as follows:

   1. For cash flows from day 0 to year 30, the risk free interest rates plus the spread.

   2. For cash flows beyond 30 years, the terminal reinvestment rate, calculated as the average of the last 10 year-end risk free spot rates for 30-year obligations plus the average of the 30-year spreads for the last 10 year-ends.

   If the risk free rate for term t under this scenario is equal to r^t , where t is 90 days through 30 years, then the four test shocks to be added to the scenario rates are given by:

    

   The shock to the terminal reinvestment rate under test scenarios i) and ii) is a subtraction of 50 basis points, and the shock under test scenarios iii) and iv) is an addition of 50 basis points.

   We request comments on the transition of discount rates (from year 20) to the terminal reinvestment rate on the space provided in the Comments and Questions Worksheet. Specifically we are interested in the insurers’ comments with respect to potential linear interpolation (from years 20) to the terminal reinvestment rate and suggestions for alternatives.

Appendix III - Interest Rate Shocks in the Standardized Approach

This note explores a few types of interest rate models that can be used to determine the magnitude of the interest rate shocks in the standardized approach market risk component. The recommendation, based on a simplified Cox-Ingersoll-Ross model, is to use interest rate shocks for the 90-day and 30-year rates that are linear functions of the square roots of the current rates.

Constant Volatility Models

1. These models have the general form:

   dr = f(r,t)dt + σdW

   A simple model for observed rates is:

   dr = σdW

2. Models in this class are prone to generating negative rates, especially at high confidence levels, because the shocks will be constant regardless of the current rate.

Cox-Ingersoll-Ross Model

3. This model is widely accepted and used in the industry. It models the risk-neutral short rate as:

    

   Under this model, the rate becomes more volatile as it gets higher, but it also reverts to a long-run value b.

4. The differences between the real-world short rate estimated from the real-world 90-day rate and the risk-neutral short rate are too subtle to have any material impact. We therefore take the above equation as a model of the real-world short rate.

5. The probability density function for the short rate under this model contains a Bessel function, but we cannot have special functions in a standardized approach.

6. Attempted fittings of the model to assorted data sets using multiple techniques produce values of a that were very low, or even negative.

7. Points 5 and 6 above lead us to try the assumption of a = 0. Now the model is very tractable, as it can be written in constant form:

    

8. The reduced model is also more conservative because, without a mean reversion term, it will produce larger potential swings in the interest rate for a given confidence level.

9. Fitting this model to the Bank of Canada’s data series of 3-month treasury bill auction yields from 1955 to 2007, with time measured in days, produces a value of σ = 0.000251. This implies that, if the current 90-day rate is r, then over a period of one year at a 99.5% confidence level, the rate will move up by 0.163√r + 0.0066 or less, or move down by 0.163√r − 0.0066 or less.

10. The Cox-Ingersoll-Ross model actually models the entire yield curve at any given time. However, there is no way under this model to reconcile observed real-world long-term yields with the short-term yields that were in effect at the same time.

11. Therefore, we again fit the reduced model, only this time using the Bank of Canada’s long- term benchmark bond yields from 1976 to 2007 instead of 90-day treasury yields. Since there is a large difference between real-world and risk-neutral rates over 30 years, we infer the risk- neutral short rate from each observed long-term yield under the model’s risk-neutral assumptions.

12. We now obtain a new value of σ = 0.000169. This implies that, if the current 30-year yield is r, then over a period of one year at a 99.5% confidence level, the rate will move up by 0.098√r + 0.0024 or less, or move down by 0.098√r − 0.0024 or less.

13. The minimum and maximum rates over the next year produced by this model, for both the 90-day and 30-year yields, are intuitively plausible (see tables next page).

14. There have been interest rate movements over one year that have fallen outside of the predicted intervals. Between January 1994 and January 1995, the 90-day rate went from 3.63% to 7.98% (the given model predicts a 99.5% maximum value of 7.40%), and between January 2001 and January 2002 the rate went from 5.14% to 1.97% (the predicted 99.5% minimum is 2.10%). Further, between January 1982 and January 1983 the 30-year rate went down from 16.33% to 12.48%, while the model predicts a 99.5% minimum value of 12.60%.

15. Nonetheless, the shocks may still be too high, because both the short- and long-term rates are shocked simultaneously. Unless these rates are perfectly correlated, the probability that both will move by their 99.5-th percentile amounts simultaneously is in fact lower than 0.5%.

16. So, in order to account for the correlation between the short and long rates, we assume that the W’s in the two processes are correlated. Direct analysis of correlations between monthly changes in the long-term benchmark bond rate and various shorter-term rates showed that the correlation with the 90-day rate was 39%, the correlation with the 6-month rate was 58%, and the correlation with the 1-3 year rate was 83%. However, the lower correlations with the 90-day rate are explained by the fact that this rate is managed by the central bank, whereas long-term rates are entirely determined by the market. Thus, simply using the correlation with the 90-day rate is likely to underestimate the shocks, especially since we will be interpolating to get the shocks for intermediate rates that are not centrally managed.

17. We thus used the correlation based on the 1-3 year rate, on the basis that this rate is free from direct central bank influence while still being of sufficiently short term. When the correlation between the changes in the square roots of the series is calculated, the correlation is 77.2%. Based on the simultaneous 99.5-th percentile of a bivariate normal distribution with this correlation, the shocks for when both the long- and short-term rates move in the same direction simultaneously are:

     

18. The Bank of Canada long term benchmark bond series (V122544) is in fact calculated using bonds having maturities close to 30 years. For intermediate durations, the shocks should be calculated by linearly interpolating the square root and constant coefficients of the shocks. In particular, the shocks for the 20-year rate derived from those for the 90-day and 30-year rates are:

     

19. The square-root model for the shocks breaks down if current rates are very low. Thus, for the simultaneous downward shocks, the 90-day rate should be set to zero if the current 90-day rate is lower than 48 bps, and the 20-year rate should be set to zero if the current 20-year rate is lower than 26 bps.