Instructions – Mortgage Insurer Appointed Actuary’s Report (AAR) Supplementary Tables

Each mortgage insurer should submit a workbook containing supplemental data (Supplementary Tables) in addition to the Appointed Actuary’s Report (AAR). The Supplementary Tables must contain specific data elements presented in specific formats. This workbook is a template for the Supplementary Tables.

This workbook includes following tabs:

• cover
• list of supplementary tables
• data dictionary
• supplementary tables (Table 1-19)
• drop down menu
• data validation

General instructions

Prepare the Supplementary Tables in accordance with the general instructions provided in this section. Instructions and guidance specific to the individual workbook tabs are provided in the following sections:

Filing

The Supplementary Tables (this workbook) should be submitted as a structured return in the Regulatory Reporting System (RRS) no later than 60 days after the end of the insurer’s fiscal year.

Basis

Information entered in this workbook should be

• determined as of the insurer’s fiscal year-end unless instructed otherwise.
• prepared on a consolidated basis.
• expressed amounts in thousands of Canadian dollars unless instructed otherwise.
• determined for insurance contract liabilities measured under IFRS 17.

Workbook modifications

The workbook should not be modified in any way including

• adding or deleting columns
• renaming, adding, or deleting tabs
• changing the format of cells

Granularity

Insurers should allocate values where they are requested at a more granular level than the level of aggregation at which the insurer has determined the amount. Where the insurer has allocated an amount, describe the allocation approach in the comment box of the cover page.

Data entry

• Blue cells are for manual entry.
• Green cells indicate the use of a drop down menu.
• Gray cells are automatically calculated.
• White cells have fixed text.

Please note that all colored cells are made accessible in the Excel spreadsheet through tooltips and locked cells.

Refer to the “data dictionary” tab for further description of what is expected in each column including drop down menu, column type and data type.

Cells should not be left “blank” unless there is no input for the cell. “0” should be entered if the amount is zero.

Insurers may only enter the values provided in drop down menus where they have been provided. Values other than those provided in drop down menus may not be entered.

Machine readable data entry

This workbook has been updated to support ingestion of the data into OSFI’s databases. Data entry in this workbook follows certain general principles.

Each tab in the workbook collects data on a certain topic (for example, table 2 collects reference curve information). Refer to the tab “list of supplementary tables” for a complete overview of the tables.

Within each tab, each distinct data value (referenced as value) is entered in a separate row (for example, each peer review information required is entered in a separate row in table 19).

Each value may be described by a value_category and one or more descriptors depending on the tab and the nature of the data collected.

In some tabs, values are described by descriptors in addition to the value_category. For example, insurers provide ultimate undiscounted estimate at prior year end in tables 10 and 12 (value) broken down by

• aggregation basis (gross, net or ceded)
• aggregation type (accident year, underwriting or policy year)
• actuarial lines of business

The values for each of these additional descriptors should be entered or chosen from a set of allowable entries (drop down menu).

Please see the data dictionary, drop down menu, and data validation for additional information on permitted values and limitations on the values that may be entered.

Data quality

Insurers should conduct a quality check of the Supplementary Tables prior to submission. In particular, insurers should ensure:

• values are selected exclusively from the options provided in drop down menus in all green cells.
• values are expressed in thousands for dollar amounts or percentages as instructed.
• information is consistent throughout the workbook.
• values are consistent with the validation rules presented in the “Data Validation” tab.

We may ask insurers to refile Supplementary Tables to correct errors including the failure to follow these instructions.

Table 1 - Portfolio Reporting

List each unique link between a portfolio and an actuarial line of business on its own row. Specifically:

• a portfolio associated with multiple actuarial lines of business will appear on multiple rows (once per actuarial line of business).
• an actuarial line of business associated with multiple portfolios will also appear on multiple rows (once per portfolio).

Table 2 - Reference Discount Curve

The total discount rate should equal the risk-free rate plus the applicable illiquidity premium. Risk-free rates and total discount rates may be expressed as forward rates or spot rates provided the two are consistent. Describe the basis for the discount rates in the comment box of the cover page.

Table 14 - Coverage Units

Provide projected coverage units separately for transactional, portfolio or multi-unit insurance. Leave blank where no coverage is provided.

Table 15 - Annual Estimate of Future Cash Flows (Undiscounted) for LRC

Provide future cash flows separately for transactional, portfolio or multi-units insurance. Do not discount the future cash flows. Leave blank if there is no cash flow in the year.

Table 16, 17 and 18 - Sensitivity Tests

These tables require insurers to determine the impact of shocks to key macroeconomic variables (that is, discount rates, unemployment rate (UR), and house price) on insurance contract liabilities.

If other variables are related to these macroeconomic variables, consistent adjustments should be made to those variables. For example, an insurer may assume that renewal mortgage rates and discount rates are related, then it would adjust mortgage rates in response to the discount rate shocks.

For rows where the “value_category” is “reported liabilities”, set the “value” to the baseline insurance contract liability component amounts reported in the AAR.

For Table 16

Recalculate insurance contract liabilities using end-of-period discount rates, adjusted by -50 and +50 basis points at all durations. Separate the liabilities into present value of future cash flows and RA for LIC and LRC, and CSM for LRC only.

For Table 17

Provide LRC calculated using: 1) the reported liabilities’ unemployment rates used, 2) those rates minus 1.0%, and 3) those rates plus 1.0%. Separate the liabilities into the present value of future cash flows, RA and CSM.

For Table 18

Provide insurance contract liabilities as determined using: 1) the house prices at the reporting date (that is, the house prices used in determining reported liabilities), 2) 90% of those house prices, and 3) 110% of those house prices. Separate the liabilities into present value of future cash flows, RA and CSM (for LRC) for LIC and LRC respectively. Future house prices should be projected based on the shocked house prices as at the reporting date.

Determining the impact of sensitivity tests for scenario-based or stochastic models

IFRS 17 specifies that the cash flows used to measure insurance contract liabilities should be estimate of future cash flows. Insurers may, where appropriate, use deterministic assumptions to project future cash flows. In principle, IFRS 17 expects insurers to determine future cash flows as the probability-weighted mean of the cash flows produced from a set of scenarios. Insurers may, but are not required to, use stochastic methods to estimate future cash flows.

Mortgage insurers may use scenarios and/or stochastic techniques to estimate future cash flows. These approaches may apply to the parameters that are shocked. The interpretation of the shocks described for these tables may be dependent on the techniques selected by the insurers for their models.

We expect insurers to use judgement in adapting the required shocks to their particular models. For example, an insurer using a stochastic model might produce future cash flows for the unemployment rate shock by

• adjusting the unemployment rates under each scenario by 1.0%
• selecting a representative scenario that reproduces the reported liability and adjusting the unemployment rates in that scenario by 1.0%
• selecting a deterministic set of assumptions that reproduces the reported liability and adjusting the unemployment assumptions or some other reasonable approaches.

We expect the insurer to select a reasonable approach to applying the shock to their own models in a manner that they expect to be analogous to a simple shock applied to a deterministic scenario and to describe the methodology selected in the comment box of the cover page.