REVIEW OF THE 30th ACTUARIAL REPORT ON THE CANADA PENSION PLAN

Document Properties

• Type of Publication: Actuarial Report

ACRONYMS USED IN THIS REPORT

AAE

Average Annual Earnings

AR27

27^th Actuarial Report on the CPP

AR28

28^th Actuarial Report on the CPP

AR29

29^th Actuarial Report on the CPP

AR30

30^th Actuarial Report on the CPP

AWE

Average Weekly Earnings

CIA

Canadian Institute of Actuaries

CPP

Canada Pension Plan

CPPIB

Canada Pension Plan Investment Board

CRA

Canada Revenue Agency

ESDC

Employment and Social Development Canada

FAMCR

First additional minimum contribution rate

GAD

Government Actuary’s Department

IAA

International Actuarial Association

ISAP

International Standard of Actuarial Practice

OCA

Office of the Chief Actuary

OSFI

Office of the Superintendent of Financial Institutions

OASDI

Old Age, Survivors and Disability Insurance (the U.S. Social Security program)

QPP

Québec Pension Plan

SAMCR

Second additional minimum contribution rate

YBE

Year’s Basic Exemption

YMPE

Year’s Maximum Pensionable Earnings

YAMPE

Year’s Additional Maximum Pensionable Earnings

EXECUTIVE SUMMARY

Authors

This report was prepared by a review panel of three independent actuaries: Sam Gutterman, Fellow of the Society of Actuaries, Fellow of the Casualty Actuarial Society and Member of the American Academy of Actuaries, Pierre Plamondon, Fellow of the Canadian Institute of Actuaries and Fellow of the Society of Actuaries, and Jill Wagman, Fellow of the Canadian Institute of Actuaries and Fellow of the Society of Actuaries.

Terms of Reference

The panel conducted its review of the 30^th Actuarial Report on the Canada Pension Plan in accordance with the following terms of reference:

“The peer reviewers will review the work of the Chief Actuary in completing the 30^th Actuarial Report on the Canada Pension Plan as at 31 December 2018. Following the review, they will provide a report to the Chief Actuary and the United Kingdom Government Actuary’s Department (GAD). GAD will then provide its opinion of the peer review to the Chief Actuary.

The review report should contain opinions on the following questions:

1. Is the professional experience of the Chief Actuary and the staff who worked on the report adequate for carrying out the work required?
2. Has the work been completed in compliance with the relevant professional standards of practice and statutory requirements?
3. Did the Chief Actuary have access to the information required to perform the valuation, and were relevant tests and analysis on the data completed as might be expected?
4. Were the actuarial methods and assumptions used in completing the report reasonable?
5. Does the 30^th Report fairly communicate the results of the work performed by the Chief Actuary and their staff?

In providing opinions on the questions listed above, the peer reviewers will also provide such recommendations as the peer reviewers deem appropriate with respect to future actuarial reports on the Canada Pension Plan prepared by the Office of the Chief Actuary.”

30^th Actuarial Report (AR30)

AR30 was prepared as at 31 December 2018.

For the base CPP, the report presents a best-estimate projection of pay-as-you-go contribution rates rising from 9.44% of contributory earnings in 2019 to 12.37% in 2070, remaining at a similar level thereafter. It also presents minimum contribution rates of 9.75% for the years 2022 to 2033 and 9.72% for the year 2034 and thereafter, down from 9.79% in AR27. Using these minimum contribution rates, AR30 projects ratios of assets-to-expenditures at relatively stable levels at 7.5 until 2034, and then increasing to 8.2 in 2050. Under a continuation of the current 9.9% contribution rate, AR30 projects ratios of assets-to-expenditures of 7.5 in 2019, increasing to 8.8 in 2050 and reaching 9.5 in 2095.

For the additional CPP, the report presents a first additional minimum contribution rate (FAMCR) applicable to earnings between the YBE and the YMPE at 1.49% in 2022 and 1.98% for 2023 and thereafter, and a second additional minimum contribution rate (SAMCR) applicable to earnings above the YMPE up to the YAMPE of 7.92% for 2024 and thereafter. Under the legislated first and second additional contribution rates of 2.0% for 2023 and thereafter and 8.0% for 2024 and thereafter, the ratios of assets-to-expenditures are projected to increase rapidly until 2025 and then decrease after that, reaching a level of 26 by 2080.

AR30 also presents the results of several sensitivity tests that show how different the results would be if specific assumptions, either individually or in combination, were varied. 

All of the results are estimates. All but the sensitivity tests present the Chief Actuary’s current “best estimates”, with no deliberate margins of conservatism or other intentional bias.

Clearly, the upcoming consequential experience of the COVID-19 pandemic will have an effect on the future of the CPP. However, at the time this report was prepared, it is premature for us to offer specific advice regarding its expected effects on factors including real investment return, economic growth and mortality. Once the implications are more apparent, we recommend that the OCA review its assumptions and, if appropriate, consider updating its projections.

It is essential to recognize that all of the results are estimates and not predictions. They simply present what the outcome will be if all of the actuarial assumptions are realized. These assumptions concern demographic and economic parameters over the next 75 years that are unknowable and, therefore, not amenable to precise prediction. Thus, it is important that readers look at the results of the sensitivity tests to enhance their understanding that the range of possible actual outcomes is wide and could even be wider than illustrated by the sensitivity tests.

Opinions

With respect to the five questions listed in the terms of reference it is our opinion that:

1. The professional experience of the Chief Actuary and the staff who worked on AR30 was adequate for carrying out the work required.
2. The work on AR30 complied with all relevant professional standards of practice and statutory requirements.
3. The Chief Actuary had access to the information required to perform the valuation and completed such relevant tests and analysis on the data as might be expected.
4. The actuarial methods used in completing AR30 were reasonable and the assumptions were also reasonable, both individually and in the aggregate.
5. AR30 fairly communicated the results of the work performed by the Chief Actuary and the staff who worked on the actuarial report.

The remainder of this report provides and discusses the basis for this opinion, including corresponding observations and recommendations as to future Actuarial Reports.

Recommendations

We compliment the Chief Actuary and the staff of her Office who prepared AR30 on their competence, commitment and professionalism. They were unfailingly helpful in clarifying issues raised by the review panel and in providing additional information. In the spirit of seeking to help the Chief Actuary and the staff of the Office to continue improving their work, our report includes the following recommendations:

Recommendation 1

We recommend that the OCA devote an increasing part of its continuing professional development activities to investment-related issues and to modeling.

Recommendation 2

We recommend that the management of the actuarial model be gradually enhanced through rigorous documentation of the model and preparation of appropriate training material. Further, in the longer term, we recommend the OCA consider modernizing the model to use a more current programming language which will not only enhance the OCA’s ability to attract talent from a larger pool of skilled programmers and actuaries, but also enable the model to be more easily updated and to process larger and more complex data sets as they become available in the future.

Recommendation 3

We recommend that the OCA continue to work with its data providers to address items on the OCA’s list of data enhancement priorities. More specifically, we recommend that:

• the $99,999 limit on employment earnings from a single employer in the Record of Earnings file be lifted; and
• where applicable, an audit confirmation be obtained regarding underlying data used as key inputs to the development of assumptions.

Recommendation 4

Given the uncertainty associated with future economic, investment and demographic experience, we recommend that two additional scenarios be analyzed: (1) relating to a ‘Japanese’ investment scenario, that is, one based upon a long-term continuation of low real long-term interest rates on federal bonds and (2) a scenario considering demographic, economic and investment factors, such as those experienced with the COVID-19 pandemic. Other scenarios could also be explored for study.

Recommendation 5

We recommend that the Chief Actuary continue to seek additional expert input to be considered in establishing the best-estimate actuarial assumptions and the range of variability examined in the sensitivity tests. Specifically, the Chief Actuary should seek input from demographic, economic and investment experts for their views on key assumptions, including a plausible range of variability that approximates an 80% confidence interval. The Chief Actuary should consider this input, but retain responsibility for the final assumptions that are chosen.

Recommendation 6

Regarding the setting of assumption on the real rate of return, we recommend that the Chief Actuary:

• continue to improve the degree of sophistication of its investment model to more closely match the CPPIB investment portfolio;
• review the speed and period of transition to the ultimate expected return in the context of opinions of investment experts;
• closely monitor the allocation of expenses between the basic CPP and additional CPP, with a view to refining this assumption in the future; and
• review the assumption on the use of leverage in the context of CPPIB’s strategic direction regarding the use of leverage and the resulting risk involved.

Recommendation 7

We recommend that the Chief Actuary continue to enhance the contents and accessibility of future Actuarial Reviews. For example, a separate file should be available on the website that provides an Executive Summary suitable for those who do not access the entire report, expand the use of charts to illustrate additional items to enhance the understanding by the reader of key factors involved with the AR, and include a certain number of years of historical data to assist the reader in understanding the key patterns involved.

Recommendation 8

We recommend that the Chief Actuary continue to analyze the incremental investment expenses incurred to apply the CPPIB’s active management strategy in order to assess the extent to which added value to the overall fund return over the long term may be expected to be consistently and reliably achieved.

Recommendation 9

We recommend that the Chief Actuary continue to maintain effective two-way communication with the CPPIB and Statistics Canada, with the goal of achieving continual improvements in the process of setting best-estimate assumptions and range of variability.

SECTION 1 – INTRODUCTION

This report presents the results of an in-depth review we conducted into the 30^th Actuarial Report on the Canada Pension Plan and the detailed actuarial examination on which it was based. We recommend that this practice should continue. This is the eighth such review that has been conducted.

This report follows closely the format of previous review reports. The observations, conclusions and recommendations, however, are entirely our own.

1.1 Terms of Reference

In accordance with our terms of reference, our review focused on the actuarial work done on the Plan. We were not asked to, and did not, review the merits of the current design, administration or investment arrangements of the Plan. Our review of those aspects was confined to how they interact with, and are reflected in, the actuarial work.

The terms of reference for our review were as follows:

“The peer reviewers will review the work of the Chief Actuary in completing the 30^th Actuarial Report on the Canada Pension Plan as at 31 December 2018. Following the review, they will provide a report to the Chief Actuary and the United Kingdom Government Actuary’s Department (GAD). GAD will then provide its opinion of the peer review to the Chief Actuary.

The review report should contain opinions on the following questions:

1. Is the professional experience of the Chief Actuary and the staff who worked on the report adequate for carrying out the work required?
2. Has the work been completed in compliance with the relevant professional standards of practice and statutory requirements?
3. Did the Chief Actuary have access to the information required to perform the valuation, and were relevant tests and analysis on the data completed as might be expected?
4. Were the actuarial methods and assumptions used in completing the report reasonable?
5. Does the 30^th Report fairly communicate the results of the work performed by the Chief Actuary and their staff?

In providing opinions on the questions listed above, the peer reviewers will also provide such recommendations as the peer reviewers deem appropriate with respect to future actuarial reports on the Canada Pension Plan prepared by the Office of the Chief Actuary.”

1.2 Procedures Followed

Our review was conducted as a close collaboration of the three panel members. The review work took place over the months from September 2019 through March 2020.

We received AR30 on 10 December 2019, the day it was tabled in Parliament. We received the working papers underlying the report on 14 November 2019.

We interviewed the Chief Actuary and members of the Office of the Chief Actuary (OCA), an independent unit of the Office of the Superintendent of Financial Institutions (OSFI), for one and one-half days. We met with officials from the Economic Analysis and Forecasting Division of the Department of Finance Canada, the Demography Division of Statistics Canada, and various functional areas within the Canada Pension Plan Investment Board (CPPIB). In all, we submitted over 110 requests for information or clarification to the various parties plus further oral questions at the time of our meetings. All of these parties responded promptly and fully to each request we made or provided reasons if such response was not possible.

We reviewed the papers presented at the CPP Seminar held in Ottawa on 28 September 2018, and at the QPP seminar held in Québec City on 1 November 2018.

Further, we reviewed the following documents:

• a paper prepared by Assia Billig and Tina Adjowa Magloe Francis of the Office of the Chief Actuary entitled "Technological transition and its potential impacts on social security programs”;
• the 2019 Annual report of the CPPIB;
• a PowerPoint presentation prepared by Assia Billig, Chief Actuary, entitled “Mortality projections for Canadian social security programs”;
• the technical report on methodology and assumptions of Statistics Canada used for the population projections for Canada (2018 to 2068);
• the report to the Social Security Advisory Board, USA, entitled “2019 Technical Panel on Assumptions and Methods”; and
• the Actuarial valuation of the Québec Pension Plan as at 31 December 2018.

We also studied with interest two recent publications of the OCA:

• “Measuring and Reporting Actuarial Obligations of the Canada Pension Plan: Actuarial Study No. 19” (April 2018); and
• “Technical Paper on the Additional Canada Pension Plan Regulations: Actuarial Study No. 20” (October 2018).

We reviewed the historical documents that are maintained on the website of the Office of the Chief Actuary, which we found to be useful.

We reviewed the Rules of Professional Conduct and the applicable Standards of Practice of the Canadian Institute of Actuaries, and the International Standards of Actuarial Practice of the International Actuarial Association.

We reviewed key provisions of the statute establishing the Canada Pension Plan.

We held several meetings in person and by teleconference and corresponded extensively by email.

After reviewing all of the information, and after much discussion among ourselves, we agreed on all of the opinions and recommendations presented in this report.

The Canada Pension Plan is a complex Plan that provides benefits composed of different components (base and additional CPP, which contain earnings-related and flat-rate benefits) payable on the occurrence of three different life-cycle events (retirement, disability and death) that have different eligibility criteria.

In addition, the actuarial computer models used to produce the results documented in AR30 are extremely complex. They project the many interrelationships of the Plan provisions and current population statistics with projections of future demographic, economic and financial experience.

In our work, we have tended to concentrate on what we consider to be the most important issues – in particular, the data used, the major methodology issues, the key actuarial assumptions, and the quality of the reporting by the Chief Actuary and her staff. As described in Section 4 (Data) of this report, we reviewed the sources of the data, and the processes used by the Chief Actuary to test and analyze the data; however, our mandate did not include a detailed audit of the data. Similarly, we reviewed the procedures used by the Chief Actuary to test the actuarial computer models; however, our mandate did not include a verification of the accuracy of the models.

1.3 The Canada Pension Plan

The CPP is a social insurance program that provides monthly income benefits and some lump sum benefits upon the retirement, death and disability of its participants. Virtually all working Canadians outside Québec contribute to the Plan.

Before 1997, contribution rates were set at a level that created relatively little advance funding of benefits and the funds not used for immediate benefit payments and expenses were loaned to the provinces at the federal long-term bond interest rate at the time the bonds were purchased. The Plan was amended in 1997 to:

• require an increased measure of advance funding;
• include a sunset clause regarding the investment of CPP assets in provincial revolving 20-year bonds;
• require that the funds not used for immediate benefit payments and expenses or for investment in those provincial bonds be invested in a diversified portfolio of investments; and
• establish an Investment Board to manage these investments.

Since the 27^th CPP Actuarial Report, the Canada Pension Plan was subject to a series of amendments:

• Part 1 of Bill C-26 – An Act to amend the Canada Pension Plan, the Canada Pension Plan Investment Board Act and the Income Tax Act (introduction of the additional CPP) which received Royal Assent on 15 December 2016;
• Division 19 of Part 6 of Bill C-74 – Budget Implementation Act, 2018, No. 1 (removal of age-related reductions in survivor benefits, introduction of child-rearing and disability drop-ins for the additional CPP, introduction of the post-retirement disability benefit for the base CPP, change in base CPP death benefit to a flat-rate of $2,500, authorization to make sustainability regulations for the additional CPP) which received Royal Assent on 21 June 2018;
• Division 2 of Part 4 of Bill C-86 – Budget Implementation Act, 2018, No. 2 (technical amendment regarding calculation of child-rearing drop-in amount under the additional CPP) which received Royal Assent on 13 December 2018;
• Division 6 of Part 4 of Bill C-97 – Budget Implementation Act, 2019, No. 1 (application for CPP retirement pension is waived upon reaching age 70, effective 1 January 2020, as well as administrative amendments regarding overpayments of salaries and wages) which received Royal Assent on 21 June 2019; and
• Amendments to the regulations regarding the calculation of the CPP contribution rates to clarify the determination of full funding rates and introduce the calculation of the additional CPP minimum contribution rates (awaiting formal consent by the provinces).

1.4 Statutory Actuarial Requirements

Section 115 of the Canada Pension Plan requires that an actuarial review be conducted once every three years and that a report is produced that indicates:

• Regarding the base CPP:
  • projected pay-as-you-go contribution rates, that is, each year’s contribution rate is just sufficient to cover that year’s benefit payments and expenses; and
  • a contribution rate, calculated by combining:
    1. a contribution rate, calculated in a prescribed manner, in respect of steady-state funding excluding changes that require full funding for post-1997 benefit improvements; and
    2. a contribution rate, calculated in a prescribed manner, in respect of full funding for post-1997 benefit improvements.
• Regarding the additional CPP:
• A first additional contribution rate, calculated in a prescribed manner, before and after a change in that rate caused by benefit improvements; and
• A second additional contribution rate, calculated in a prescribed manner, before and after a change in that rate caused by benefit improvements.

Section 113.1 of the Canada Pension Plan requires that a financial review of the Plan be prepared every three years by the federal and provincial Ministers of Finance. This review is to take into account the most recent report of the Chief Actuary prepared in accordance to Section 115. Section 115 states that projections must extend for at least 75 years into the future.

Separate funding objectives are established for the base and for the additional CPP.

• The financing objective of the base CPP is stated in terms of:
  • a steady-state contribution rate calculated as the lowest constant rate for which the projected ratio of Plan assets-to-expenditures 10 years after the end of the review period matches the corresponding projected ratio 60 years after the end of the review period; and
  • a full funding rate for increased or new benefits for which the cost equals or exceeds 0.02% of contributory earnings (the “de minimis” rule).

  The minimum contribution rate for the base CPP is the sum of the steady-state contribution rate and full funding rate.

• The financing objective of the additional CPP is stated in terms of the AMCRs (FAMCR and SAMCR) that must be determined before and after taking into account the full funding of any increased or new additional benefits. The AMCRs are defined as the lowest level contribution rates applicable after the end of the review period, such that the following conditions are met:
  • the present value of projected additional open group obligations is less than or equal to the projected additional assets and present value of projected additional contributions (open group assets);
  • the projected assets/expenditures (A/E) ratio of the additional Plan is the same in the 50^th and 60^th years following the end of the review period, but no earlier than in the years 2088 and 2098, respectively; and
  • the SAMCR is equal to 4 times the FAMCR.

The Calculation of Contribution Rates Regulations, 2007 describe the two funding objectives of the base CPP. These regulations were amended in 2018 to clarify the calculation and application of the “de minimis” rule (described above) and to set out the calculation of the additional minimum contribution rates for the additional CPP. These regulations are the Calculation of Contribution Rates Regulations, 2018 (awaiting formal consent by the provinces).

1.5 Main Results of the 27^th, 28^th and 29^th Actuarial Reports

The 27^th CPP Actuarial Report was prepared as at 31 December 2015. It presented a best-estimate projection of pay-as-you-go contribution rates for the Plan, rising from 9.13% of contributory earnings in 2016 to 11.80% in 2065, and remaining at a similar level thereafter. It also indicated a minimum contribution rate to be paid from 2019 and thereafter of 9.79% of contributory earnings. Using this minimum contribution rate of 9.79%, AR27 projected ratios of assets-to-expenditures falling slightly from 6.5 in 2016 to 6.4 by 2028, remaining at the same level fifty years later in 2078. Under a continuation of the current 9.9% contribution rate, AR27 projected ratios remained steady at 6.5 from 2016 to the early 2030s, then rising to 7.4 by 2090.

The 28^th CPP Actuarial Report was prepared to show the estimates for the Plan upon the introduction of the additional Plan (Bill C-26). AR28 concluded that the FAMCR of 2.0% and the SAMCR of 8.0% were sufficient to finance the additional CPP.

The 29^th CPP Actuarial Report was prepared to show the effect of Bill C-74 on the long-term financial states of the base and additional Plans. There was no supplemental actuarial report in respect of Bill C-86, because the cost impacts on the CPP were deemed to be small to negligible.

1.6 Improvements Following Recommendations of the Last Review Panel

The actuarial review panel for AR27 made eleven recommendations arising from its review, plus numerous other observations or suggestions for improvement, which the Chief Actuary has taken into account. In preparing AR30, the Chief Actuary has made numerous improvements in the work and reporting. Many of these improvements were made in response to the recommendations, observations and suggestions of the prior independent review panel. Where the recommendations of that panel have not been fully adopted, the Chief Actuary has provided a discussion of the partial progress made or has explained and supported where they were not fully achieved.

Our terms of reference do not call for a detailed evaluation of the appropriateness of the response of the Chief Actuary to the recommendations of the prior independent review panel. The Appendix, however, presents a summary of these responses.

1.7 Main Results of the 30^th Actuarial Report

AR30 was prepared as at 31 December 2018.

For the base CPP, the report presents a best-estimate projection of pay-as-you-go contribution rates rising from 9.44% of contributory earnings in 2019 to 12.37% in 2070 and remaining at a similar level thereafter. It also presents minimum contribution rates of 9.75% for the years 2022 to 2033 and 9.72% for the year 2034 and thereafter, down from 9.79% in AR27. Using these minimum contribution rates, AR30 projects ratios of assets-to-expenditures at relatively stable levels at 7.5 until 2034, and then increasing to 8.2 in 2050. Under a continuation of the current 9.9% contribution rate, AR30 projects ratios of assets-to-expenditures of 7.5 in 2019, increasing to 8.8 in 2050 and reaching 9.5 in 2095.

For the additional CPP, the report presents a first additional minimum contribution rate (FAMCR) applicable to earnings between the YBE and the YMPE at 1.49% in 2022 and 1.98% for 2023 and thereafter, and a second additional minimum contribution rate (SAMCR) applicable to earnings above the YMPE up to the YAMPE of 7.92% for 2024 and thereafter. Under the legislated first and second additional contribution rates of 2.0% for 2023 and thereafter and 8.0% for 2024 and thereafter, the ratios of assets-to-expenditures are projected to increase rapidly until 2025 and then decrease after that, reaching a level of 26 by 2080.

AR30 includes a reconciliation of the changes to the minimum contribution rate of the base CPP between AR27 and AR30, and changes to the first and second additional minimum contribution rates of the additional CPP between AR28 and AR30.

• For the base CPP, experience over the period 2016 to 2018 was better than anticipated, especially regarding investment income, which lowered the MCR. Changes made to the demographic assumptions also acted to lower the MCR. However, these reductions were partially offset by lower assumed real wage increases and changes to the assumed retirement take-up rates, assumed disability incidence rates, and investment assumptions.
• For the additional CPP, although demographic experience over 2016 to 2018 was better than anticipated and acted to lower the AMCRs, changes in assumptions, especially those related to asset allocation and the amendments under Bill C-74, have more than offset the decrease. As a result, the net result of all changes since the 28^th CPP Actuarial Report is an overall absolute increase in the FAMCR of 0.05% and corresponding increase in the SAMCR of 0.20%.

1.8 Interpreting the Results of CPP Actuarial Reports

In AR30, the financial performance of the CPP is measured through several key metrics, including:

• the projected pay-as-you-go contribution rates and asset-to-expenditure ratios by year to 2050 and then every fifth year through to 2095, under both the current legislated contribution rate and the minimum contribution rate;
• the minimum contribution rate of the base CPP and the first and second additional minimum contribution rates of the additional CPP calculated at the current valuation date and how these rates are projected to evolve over the next four triennial valuation reports, assuming that the Chief Actuary’s best-estimate assumptions are realized;
• a number of sensitivity tests, which illustrate the results that would be obtained under various changes in either future experience or actuarial assumptions; and
• a CPP balance sheet showing estimates of the assets as a percentage of the liabilities under an open group approach as at 31 December 2018 and 2030.

The current minimum contribution rates are the most significant of these results. The federal and provincial Ministers of Finance are to take it into account in their triennial financial review of the CPP. If the minimum contribution rate of the base CPP is higher than the legislated rate, and the federal and provincial governments do not agree on a course of action, the insufficient rates provisions in Section 113.1 of the Canada Pension Plan will apply to automatically increase the contribution rate and freeze benefits. Regarding the additional CPP, the course of action to be taken if the first and second additional minimum contribution rates are higher or lower than the legislated rates is described in regulations that are currently waiting for provincial approval.

The other results are also useful because they provide information regarding the long-term pattern of revenues and expenditures of the Plan and the unpredictability and variability of the different cash flows if the assumptions are changed or not realized. They also allow comparisons to be made with other countries’ social security programs.

It is essential to recognize that all of the results are estimates and not predictions. All but the sensitivity tests represent the Chief Actuary’s “best estimates”, with no deliberate margins for conservatism or other intentional bias. They present what the outcome will be if all of the actuarial assumptions are realized. The assumptions (for example, fertility rates, net migration rates, mortality rates, disability incidence rates, rates of labour force participation, retirement rates, rates of price increase, real rates of wage increase and real rates of return on investments) apply from 2019 for 75 years and are forecasts of unknowable future events and conditions and, therefore, are not amenable to precise prediction.

The estimates shown in AR30 are essential to provide guidance in financing the Plan and in performing other planning and management tasks. Yet, no matter how carefully they are prepared, they are still estimates. Thus, it is important that readers look at the sensitivity tests to enhance their understanding that the range of possible actual outcomes is wide and could even be wider than illustrated by the sensitivity tests.

1.9 Outline of this Report

Sections 2, 3 and 4 of this report address the first three questions in our terms of reference regarding Professional Experience, Professional and Statutory Requirements, and Data.

Section 5 (Methodology) and Section 6 (Assumptions) address question 4 in the terms of reference.

Section 7 (Communication of Results) addresses question 5 in the terms of reference.

Section 8 (Other Issues and Recommendations Thereon) provides further important commentary.

The Executive Summary provides an overview of our findings.

SECTION 2 – PROFESSIONAL EXPERIENCE

In this Section we address the following question:

“Is the professional experience of the Chief Actuary and the staff who worked on the report adequate for carrying out the work required?”

2.1 Background

AR30 as submitted by the Chief Actuary to the Minister of Finance, was tabled in Parliament on 10 December 2019. The Chief Actuary is Assia Billig, a Fellow of the Society of Actuaries (2004) and of the Canadian Institute of Actuaries (2005). Ms. Billig was appointed Chief Actuary within the Office of the Superintendent of Financial Institutions in April 2019.

She joined the Office of the Chief Actuary (OCA) in 2008, where she was involved in the preparation of statutory actuarial reports on the Canada Pension Plan and Old Age Security Program, as well as various national and international actuarial studies. Prior to joining the OCA, she worked in private pension consulting.

Ms. Billig is the vice-chair of the International Actuarial Association (IAA) Social Security Committee and the chair of the Technical Commission on Statistical, Actuarial and Financial Studies of the International Social Security Association. She has completed her undergraduate studies in Moscow State University and has PhD in Mathematics from University of Alberta.

The professional who worked most closely with Ms. Billig on AR30 is Michel Montambeault, Senior Actuary in the Office of the Chief Actuary, a Division of OSFI. Mr. Montambeault is a Fellow of the Society of Actuaries (1992) and of the Canadian Institute of Actuaries (1992). He spends his time on issues associated with the Canada Pension Plan and Old Age Security Program. He has worked on actuarial reviews of the Canada Pension Plan and the Old Age Security Program in the Office of the Chief Actuary for the last 30 years.

Ms. Billig and Mr. Montambeault co-signed the 30^th Actuarial Report of the CPP. They were assisted by a relatively large staff of actuaries. The professionals who worked on AR30 were:

2.2 Observations

There are very few actuaries working in Canada with experience in valuing social security programs like the CPP and the QPP. The data sources, macroeconomic modelling and range of assumptions involved in actuarial valuations of these social programs are more complex than for employer-sponsored pension plans. Therefore, occupational pension plan experience is useful but not as useful as previous experience with social programs similar to the CPP and the QPP. Ms. Billig and Mr. Montambeault have considerable experience and understanding of the issues involved in valuing the Canada Pension Plan. In addition, Mr. Patrick Dontigny has a solid knowledge of the CPP actuarial model (Actucan) and Mr. Maxime Delisle represents an important addition to the actuarial team regarding investment matters.

The staff of the Office of the Chief Actuary is of sufficient size to spend adequate amounts of time on CPP and related matters, such as improving methodologies and data sources, performing inter-valuation studies, enhancing documentation, evaluating the cost of new benefits and liaising with other government departments and other social security actuaries, all of which contribute to the quality of the work and of the report.

We are pleased to observe that staff levels are being maintained. We understand from our discussions with Ms. Billig that there is an appropriate program of staff recruiting and succession planning in place. There is a mix of more experienced and relatively new personnel on the staff of the OCA. Staff continuity has been excellent. We also note that staff are moved from one technical area to another in their primary work and/or peer review work. Thus, staff are gaining in-depth knowledge of the entire workings of the Plan and applicable actuarial models.

We are satisfied that Ms. Billig and the staff who assisted her in preparing AR30 have the relevant experience and are qualified to carry out the actuarial valuation.

Notwithstanding this, we note that:

• The investment program being followed by the CPPIB continues to become more complex and more sophisticated. Also, the assets under management by the CPPIB continue to increase and will accelerate with the adoption of the additional CPP. Compared to the base CPP, the additional CPP will be financed much more heavily by investment income than by current contributions (which is the reverse of the base CPP). To respond to the increasing size, complexity, sophistication and importance of investment-related issues within the base and additional CPP, we continue to believe that the investment expertise of the Chief Actuary and their staff should continue to be progressively enhanced over time.
• Only few members of the OCA actuarial team have a thorough knowledge of all parts of the actuarial model used for CPP valuations. It is important that this knowledge be shared among a larger number of OCA staff and, more generally, that the modeling expertise of the Office be further enhanced.

2.3 Opinion on Professional Experience

In our opinion, the professional experience of the Chief Actuary and the staff who worked on AR30 was adequate for carrying out the work required.

2.4 Recommendations

Recommendation 1

We recommend that the OCA devote an increasing part of its continuing professional development activities to investment-related issues and to modeling.

SECTION 3 – PROFESSIONAL AND STATUTORY REQUIREMENTS

In this Section, we address the following question:

“Has the work been completed in compliance with the relevant professional standards of practice and statutory requirements?”

3.1 Background

To address this question, we have considered each of the following:

• Canadian Institute of Actuaries Rules of Professional Conduct: The Chief Actuary and their co-signatory are Fellows of the Canadian Institute of Actuaries (CIA), the professional body governing the education, qualification, conduct and work of actuaries in Canada. The CIA promulgates the professional rules and ethical standards with which a member must comply and thereby serve the public interest. The Rules of Professional Conduct are the Institute’s highest level of guidance to its members. Failure to adhere to the rules results in disciplinary proceedings.

• CIA Standards of Practice: These standards govern the work performed by actuaries in Canada. These include general standards governing all areas of practice and practice-specific standards governing work in specific areas. Part 1000 – General Standards apply to all areas of actuarial practice. Part 7000 – Social Security Programs came into effect February 1, 2018 and applies to an actuary when performing or reviewing, advising on, or opining on work related to social security programs.

• Canada Pension Plan: This statute provides the terms of reference of the Chief Actuary when preparing an actuarial report in relation to the CPP. Section 113.1 identifies the actuarial information required by the federal and provincial Ministers of Finance when recommending changes to CPP benefits or contribution rates, or both. Section 115 stipulates the timing, contents and certain other aspects of the Chief Actuary’s triennial report.

In the Subsections below, we consider each of these in turn.

3.2 Canadian Institute of Actuaries (CIA) Rules of Professional Conduct

The following Rules of Professional Conduct of the CIA are particularly relevant to this review:

• Rule 1: A member shall act honestly, with integrity and competence, and in a manner to fulfil the profession’s responsibility to the public and to uphold the reputation of the actuarial profession.
• Rule 2: A member shall perform professional services only when the member is qualified to do so and meets applicable professional continuing qualification standards.
• Rule 3: A member shall ensure that professional services performed by or under the direction of the member meet applicable standards of practice.

We are satisfied that the Chief Actuary and her staff have met the requirements of the CIA Rules of Professional Conduct.

Further to Rule 2, Section 2 of this report expands on our assessment of the professional experience of the staff of the Office of the Chief Actuary. Under the auspices of Rule 2, the CIA has also promulgated Continuing Professional Development (CPD) requirements that are applicable to practicing actuaries. Until 2019, these requirements obliged an actuary to obtain at least 100 hours of CPD over a two-year period, and the CPD activities should be relevant to the actuary’s area of practice. The actuary had to devote required minimum amounts of CPD time to technical skills and professionalism. At least 24 hours of CPD time must be obtained biennially by participating in “structured” activities such as participating in professional meetings or seminars. We have reviewed the CPD records of the Chief Actuary and her co-signatory to AR30, as well as the professional staff of the OCA who are Fellows of the Canadian Institute of Actuaries, and confirmed that they all met the CIA’s CPD qualification requirements.

Notwithstanding the foregoing finding, the reader should note that Recommendation 1 in the preceding Section suggests an increased focus on investment and modeling issues in OCA’s future CPD activities.

Further to Rule 3, the next two Subsections expand on our assessment of the Chief Actuary’s compliance with the CIA General Standards of Practice and the CIA Social Security Programs Standards of Practice.

3.2.1 Canadian Institute of Actuaries (CIA) General Standards of Practice – Part 1000

The General Standards of Practice of the CIA are extensive and detailed. The topics covered include numerous matters relevant to AR30 such as:

• materiality;
• knowledge of the circumstances of the case;
• approximations;
• subsequent events;
• data sufficiency and reliability;
• the model;
• quality assurance;
• control procedures;
• reasonableness of results;
• documentation;
• actuary’s use of another person’s work;
• selection of assumptions and methods;
• provision for adverse deviations;
• comparison of current and prior assumptions; and
• reporting.

The CIA standard requires that the actuary should take a subsequent event into account if the subsequent event:

• Provides information about the entity as it was at the calculation date;
• Retroactively makes the entity different at the calculation date; or
• Makes the entity different after the calculation date and a purpose of the work is to report on the entity as it will be as a result of the event.

Under Bill C-97 – Budget Implementation Act, 2019, No. 1, the application for a CPP retirement pension is waived upon reaching age 70, effective 1 January 2020. Bill C-97 was introduced in 2019 and received Royal Assent on 21 June 2019. The Chief Actuary considered this amendment to be a subsequent event for the purpose of this report, since it became known to the Chief Actuary after the valuation date but before the report date and was determined to have an effect on the financial state of the CPP. In addition, amendments to the regulations regarding the calculation of the CPP contribution rates were proposed in 2018 to clarify the determination of full funding rates and introduce the calculation of the additional CPP minimum contribution rates. These regulations, as well as proposed regulations regarding the sustainability of the additional CPP, namely the Calculation of Contribution Rates Regulations, 2018 and the Additional Canada Pension Plan Sustainability Regulations, are currently awaiting formal consent by the provinces.

In AR30 the Chief Actuary took into account all the above listed amendments and proposed regulations, as well as the subsequent event of the amendment under Bill C-97. Further, AR30 took into account updated population estimates from Statistics Canada that became available in January 2019, which was also considered to be a subsequent event for the purpose of this report because it had an effect on the determination of certain demographic assumptions.

The CIA standard on data sufficiency and reliability requires the actuary to apply such procedures as are necessary for the actuary to arrive at a conclusion as to the sufficiency and reliability of the data relied upon. The Chief Actuary relies on data from multiple sources and has considered the qualifications, competence, integrity, and objectivity of the parties providing the data. In addition, the Office of the Chief Actuary issued an in-depth report dated February 2019 entitled CPP Actuarial Report External Data Validation Process (for the 30^th CPP Actuarial Report as at 31 Dec. 2018) which describes the source of each data files, the information provided in each data file, the validation processes and the OCA’s tolerance for errors.

The CIA standard states that when the work involves the use of a model, the actuary should choose a model appropriate to the purpose and requirements of the work; and understand any limitations in the model that might make the results of the model inappropriate for the intended purpose or might produce a misleading result. While the Chief Actuary is confident that the model used is appropriate and reliable, we did discuss the fact that within the OCA, there are only a few team members who fully understand the details of the model and all of its complexities. Further, the model is complex and there is a lack of robust documentation. Although there are efforts being made to cross-train other staff, being that the model is written in Fortran (not a “modern” programming language), it may make it more difficult to attract the necessary talent with the required expertise to maintain the model. As a result, the model should be modernized over a period of years, and rigorously documented.

The CIA standard on assumptions requires that the assumptions, individually and in the aggregate, should be appropriate. We have concluded that the assumptions adopted for AR30 are reasonable, both individually and in the aggregate, and are therefore appropriate. 

The CIA standard on provision for adverse deviations (such a provision is sometimes referred to as a “margin for conservatism”) states that the actuary “should not include a provision [for adverse deviations] if the related work requires an unbiased calculation.” Section 113.1 of the Canada Pension Plan requires that the Chief Actuary determine the lowest constant contribution rate that, if maintained over the foreseeable future, results in specified projected asset-to-expenditure ratios that are constant. The Chief Actuary interprets this requirement as necessitating an unbiased calculation, and we agree.

Accordingly, the Chief Actuary uses assumptions that represent her “best estimate” for each relevant assumption. The consequence is that the overall valuation results, other than the sensitivity tests, are likewise the Chief Actuary’s “best estimates” and do not include a provision for adverse deviations.

In our view, the work on AR30 complies with the relevant portions of the CIA General Standards of Practice.

3.2.2 Canadian Institute of Actuaries (CIA) Social Security Programs Standards of Practice – Part 7000

The Social Security Programs Standards of Practice of the CIA apply to an actuary when performing or reviewing, advising on, or opining on work related to social security programs. The standards came into effect October 15, 2017 and were revised effective February 1, 2018. The standards are extensive and, as an example, cover the following matters relevant to AR30:

Part 7220 - Data

• “7220.01 Where sufficient, reliable, and relevant data are not available for the valuation of a specific benefit, the actuary should make appropriate assumptions and/or introduce appropriate methods to compensate for any perceived deficiencies in the data.”

Part 7310 – Methods

• “7310.01 The actuary should value the social security program assuming that it continues indefinitely as a going concern.”
• “7310.02 The actuary should select an actuarial cost method that is consistent with the circumstances affecting the work.”
• “7310.03 The actuary’s work should take into account the benefits, relevant policies, and administration practices of the social security program, as of the calculation date, and should take into account any virtually definitive amendment to these items that is expected to have a material effect on benefits, unless the circumstances affecting the work require otherwise.”

Part 7320 - Assumptions

• “7320.01 The actuary should select assumptions that reflect the projection period and the expectation that the social security program will continue indefinitely as a going concern, but may adjust such assumptions to reflect short-term considerations, where appropriate.”
• “7320.02 The actuary should select either best estimate assumptions or best estimate assumptions modified to incorporate margins for adverse deviations to the extent, if any, mandated by law or by the circumstances affecting the work, and should provide the rationale for the decision made with respect to the inclusion or exclusion of margins.”

Part 7350 - Margins

• “7350.01 The actuary should not include any margins for adverse deviations when the circumstances affecting the work require a best estimate calculation.”

Part 7360 - Sensitivity Testing

• “7360.01 The actuary should perform sensitivity testing of adverse scenarios to illustrate plausible material risks to which the social security program may be exposed and to aid in the understanding of the effect of adverse changes to assumptions.”

Part 7400 - Experience Analysis

• “7400.01 The actuary should conduct an experience analysis, including a comparison of actual and expected experience for the period between the prior calculation date and the current calculation date.”

Part 7500 – Reporting on the Valuation of a Social Security Program

This part outlines possible report content and is extensive. Approximately thirty different disclosures are suggested, ranging over areas such as methodology, data, assumptions, results and analysis. AR30 provides all the relevant suggested disclosures.

In our view, the work on AR30 complies with the relevant portions of the CIA Social Security Programs Standards of Practice.

3.3 Canada Pension Plan Statute

The Canada Pension Plan stipulates the frequency, approximate timing and certain contents of the Chief Actuary’s triennial reports to the federal and provincial Ministers of Finance. In AR30, the Chief Actuary and their staff have complied with all of these statutory requirements.

3.4 Opinion on Professional and Statutory Requirements

In our opinion, the work on AR30 complied with all relevant professional standards of practice and statutory requirements.

3.5 Recommendations

Recommendation 2

We recommend that the management of the actuarial model be gradually enhanced through rigorous documentation of the model and preparation of appropriate training material. Further, in the longer term, we recommend the OCA consider modernizing the model to use a more current programming language which will not only enhance the OCA’s ability to attract talent from a larger pool of skilled programmers and actuaries, but also enable the model to more easily updated and to process larger and more complex data sets as they become available in the future.

SECTION 4 – DATA

In this Section, we address the following question:

“Did the Chief Actuary have access to the information required to perform the valuation, and were relevant tests and analysis on the data completed as might be expected?”

4.1 Background

Appropriate data are required for “current status” data inputs into the computer model, for “validation” (back-testing) of the model, and to develop appropriate actuarial assumptions for future years. Data from ESDC and CRA normally flow to the OCA through Service Canada. Examples of such data are:

The status and validation data, and the historical data used to develop assumptions, appear to be objectively derived, factual and up to date.

The valuation data on benefits, earnings and contributions received from ESDC are tested in detail by OCA for internal consistency and reasonableness. The data from other sources are reviewed by OCA for internal consistency and consistency with past data. Any irregularities are checked out with the data source and any data errors are corrected.

The Chief Actuary has advised us that she and her staff had access to sufficient and reliable data to complete their work, and in AR30 has provided her opinion that “the data on which this report is based are sufficient and reliable”.

4.2 Observations

We have the following observations:

• The Chief Actuary appears to have had access to the data required.
• The data are extensive and appear to be reasonably complete and available on a timely basis.
• The data are tested for reasonableness and internal consistency by the OCA and any deficiencies are resolved before the data are used.
• The Record of Earnings (ROE) file of all workers who ever made a contribution to the CPP appears to be sufficiently complete (except for very recent transactions) and accurate, including dates of birth, although we did not validate this data.
• The Service Canada's Master Benefit File appears to be sufficiently complete, although there is some concern about the inability to verify survivorship with respect to those residing outside of Canada.
• The Canada Pension Plan Investment Board (CPPIB) has a Reference Portfolio that represents its baseline policy for determining asset allocations. Actual asset allocations can and do deviate systematically from the Reference Portfolio as the CPPIB seeks to achieve incremental returns at reasonable risk. When the CPPIB decides to deviate from the Reference Portfolio it conducts a risk-budgeting analysis to determine that the deviation is justified on an expected risk-adjusted basis. The Chief Actuary made assumptions as to future asset allocations based on the Reference Portfolio, the risk-budgeting policy, the current actual asset allocation, the asset allocations of other large Canadian pension funds, and her judgement with respect to possible future CPPIB asset allocations.
• The CPP and QPP seminars have provided a great deal of useful information, with increased relevance over time (for example, a shift to a longer-term focus). These seminars should continue to engage presenters who are known to hold divergent views and encourage presenters to summarize the range of plausible viewpoints.
• The OCA maintains contacts with other Departments and Agencies, such as the CPPIB, ESDC, CRA, Statistics Canada, Finance Canada and the Bank of Canada, as well as with external agencies such as Retraite Québec, the Conference Board, the CD Howe Institute, and the University of Toronto's Policy and Economic Analysis Program. All of this provides useful input.
• The OCA has identified its priorities for data enhancement that could lead to improved analysis.
• Where applicable, the OCA should seek audit confirmation from data providers regarding underlying data used as key inputs to the development of assumptions.

One of the OCA's data enhancement priorities pertains to the current limit of $99,999 that applies to employment earnings from a single employer in the ROE file. In 2019, the cap on employment earnings under the base CPP (the YMPE) was $57,400, well below this $99,999 limit. Under the newly adopted additional CPP, the YAMPE will represent 114% of the YMPE. As nominal wage growth continues, the $99,999 limit will eventually be reached. Even today, for a current worker with multiple employers, it can be surpassed. Clearly, this limit will need to be lifted.

4.3 Opinion on Data

In our opinion, the Chief Actuary had access to the information required to perform the valuation and completed such relevant tests and analysis on the data as might be expected.

4.4 Recommendations

Recommendation 3

We recommend that the OCA continue to work with its data providers to address items on the OCA's list of data enhancement priorities. More specifically, we recommend that:

• the $99,999 limit on employment earnings from a single employer in the Record of Earnings file be lifted; and
• where applicable, an audit confirmation be obtained regarding underlying data used as key inputs to the development of assumptions.

SECTION 5 – METHODOLOGY

In this Section, we address the following question:

"Were the actuarial methods used in completing the report reasonable?"

5.1 Background

The results presented in AR30 are based on a macro-simulation model of the Plan's operations, which projects the elements of income and outgo and the accumulation of both the base CPP and the additional CPP assets year by year up to the year 2095. Those projections are used to determine projected pay-as-you-go contribution rates and the minimum contribution rate (MCR) of the base CPP, based on the financing objective set out in the CPP statute, and are used to determine the first additional minimum contribution rate (FAMCR) and the second additional minimum contribution rate (SAMCR) of the additional CPP as set out in the CPP statute.

5.2 Macro-simulation Model

The OCA's macro-simulation model starts with current and past statistics on the population (numbers of people distributed by age and sex) and earnings (distributed by age, sex and broad earnings levels) of residents of Canada outside of Québec. The model projects each of the following, in turn, for each calendar year during the projection period:

• the number and characteristics (for example, age, sex, earnings) of the population of Canada less Québec;
• the number and characteristics of eligible CPP contributors and beneficiaries;
• the amount of CPP contributions made and benefits received by eligible CPP contributors and beneficiaries;
• the investment income;
• the expenses; and
• the assets accumulating in the base and additional CPP funds.

Thus, the model combines the projections of the contribution income and benefit outgo with the projections of investment income and expenses to arrive at the total amounts of projected asset.

The model projects anticipated experience in future years based on demographic and economic assumptions related to the CPP as a whole. These assumptions include demographic assumptions such as fertility, migration and mortality, and economic assumptions such as regarding labour force participation rates, price inflation and wage escalation. Due to the use of different reference portfolios of the base and additional CPP, different investment return assumptions are used.

The Record of Earnings (ROE), the data file for each individual who has ever made a contribution to the CPP, is not directly used for the valuation itself. Although certain assumptions and adjustments are set based on a review of the ROE file, and certain back-testing is done against the ROE file, the benefit and contribution projections are built on population forecasts of grouped data. Thus, the fundamental valuation concept differs from the person-by-person seriatim used for actuarial valuations of most occupational pension plans. Further, the actuarial valuation of occupational pension plans effectively assume a closed population. The CPP valuation, in contrast, projects a constantly changing population and the Plan is implicitly assumed to continue in perpetuity.

The model is calibrated using a back-testing procedure. Model output for years prior to the valuation date is compared against historical values. Discrepancies are investigated and resolved. Resolution may include the development of adjustment factors to better calibrate the model to historical experience. These experience adjustment factors are generally modest, but they serve the important function of "truing up" the projected results to past observed values. This ensures that the use of grouped data, or minor inadequate refinement in the assumptions, do not unduly distort the overall results. We are of the opinion that the adjustments made are reasonable.

The model relies principally on a deterministic, rather than a stochastic, approach. That is, for each year in the projection period, each run of the model produces:

• a (deterministic) single set of projected results

  rather than

• a (stochastic) probability distribution of possible results derived from projections of the expected results and of the underlying volatility of one or more of the assumptions of the model (this allows estimates of probability to be assigned to ranges of outcomes, thereby increasing the information available).

The results of stochastic analysis appear in the individual sensitivity tests, which are described in Subsection 5.8 hereunder.

Moreover, once an assumption reaches its ultimate value, each subsequent year's projected results are based on a continuation of that assumption. There is no provision in the model for subsequent assumptions to deviate from the ultimate value. In part this is reasonable as the projections cover a long time-horizon, which will inevitably see cycles and stochastic deviations in experience, which are expected to offset over time. As a consequence, the model results give the impression of smooth changes in most of its outputs, without reversals. It is likely that future experience will be more volatile than is reflected by the projections.

The model is written in the Fortran program language. Fortran was developed in the 1950s and is especially suited to numeric and scientific computing. Fortran has continued to be updated over the years with the most recent release in 2018. Fortran and C++ have been benchmarked as the fastest languages in terms of performance; however, Fortran is rarely used today, except in scientific computations. The biggest challenge with managing models written in Fortran is finding and attracting talent who are skilled in its use. Languages like Python, R or C++ are vastly more popular among students and programmers and are supported through community-based resources, which make training and trouble-shooting much more adaptable and accessible. Further, as more demands for updates arise and larger data sets become more readily available, Fortran's abilities are limited compared to some of the more modern languages.

5.3 Form of Output

The model produces the following outputs which are discussed in Section 5.4:

• projected demographic and financial results, including the pay-as-you-go contribution rates, the asset-to-expenditure ratios based on the current statutory contribution rate, and other income and expenditure details for each year up to 2050 and thereafter every fifth year up to 2095;
• the current minimum contribution rate for the base CPP, which in AR30 consists of the steady-state rate and full funding rate, as well as projected minimum contribution rates for each of the next four triennial actuarial reviews;
• the first and second additional contribution rates for the additional CPP;
• a CPP balance sheet showing estimates of the assets as a percentage of the liabilities under an open group approach as at 31 December 2018 and 31 December 2030;
• reconciliations of AR30 results with the results in AR27 (base CPP) and AR28 (additional CPP); and
• sensitivity tests showing the results of applying alternative assumptions.

5.4 Actuarial Cost Analyses – Base CPP

5.4.1 Pay-As-You-Go – Base CPP

When the base CPP was initially established, it was financed on a "pay-as-you-go" basis with a small reserve fund. Although that financing approach was replaced in 1997, the projected pay-as-you-go costs provide useful information about the future financial status of the base CPP. Paragraphs 115(1.1)(a) and (b) of the Canada Pension Plan require the Chief Actuary to present "pay-as-you-go" projections year-by-year for the first 30 years and thereafter every five years up to at least 75 years after the valuation date. In AR30, the projection extends to the year 2095.

5.4.2 Minimum Contribution Rate - Base CPP

The methods used to compute the minimum contribution rate involve a combination of "steady-state funding" and "full funding". Thus, the "minimum contribution rate" is computed as the sum of:

1. the contribution rate determined by the steady-state method for all benefits other than benefit improvements resulting from changes to the Canada Pension Plan that occurred after 1997; and
2. the contribution rate determined by the full funding method for benefit improvements due to post-1997 changes to the Canada Pension Plan.

The steady-state method produces a contribution rate that is the lowest constant rate that, if maintained over the foreseeable future, results in projected asset-to-expenditure ratios that are generally constant. The asset-to-expenditure ratio for a year is the ratio of the projected assets at the end of the year to the projected expenditures in the following year. In practice, the steady-state rate is computed as the lowest level contribution rate, starting three years after the review date (called the "review period"), that produces the same projected asset-to-expenditure ratios in the 10th and the 60th years following the review period. The use of these years is stipulated in the Calculation of Contribution Rates Regulations, 2018. In AR30, the asset-to-expenditure ratios for 2031 and 2081 are used for this purpose.

Paragraph 113.1(4)(e) of the Canada Pension Plan requires that post-1997 benefit improvements be separately identified and funded on a "full funding" basis. That is, the steady-state contribution rate must be augmented to reflect benefit improvements that are deemed to be earned in the future. There must also be a temporary increase in the contribution rate to liquidate any unfunded liability resulting from the benefit improvement. The temporary increase is to apply for several years that is consistent with common actuarial practice. The Chief Actuary has chosen 15 years for this purpose; we concur that this accords with common actuarial practice in Canada.

5.5 Actuarial Cost Analyses – Additional CPP

The financing methods used to compute the AMCRs (first additional minimum contribution rate or FAMCR and second additional minimum contribution rate or SAMCR) are defined in the Calculation of Contribution Rates Regulations, 2018 as the lowest level of contribution rates, applicable after the end of the review period, such that:

1. the present value of projected additional open group obligations is less than or equal to the projected additional assets and present value of projected additional contributions (open group assets);
2. the projected assets/expenditures (A/E) ratio of the additional Plan is the same in the 50th and 60th years following the end of the review period, but no earlier than in the years 2088 and 2098, respectively; and
3. the SAMCR equals the FAMCR multiplied by the ratio of the earnings replacement rate of the second tier of the additional Plan to the replacement rate of the first tier (33.33%/8.33%, or 4).

The first requirement satisfies Paragraph 113.1(4)(d)(ii) of the Canada Pension Plan which states that contribution rates must be no lower than the rates that result in projected contributions and investment income that are sufficient to fully pay the projected expenditures of the additional Canada Pension Plan over the foreseeable future.

The second requirement satisfies Paragraph 113.1(4)(d)(i) of the Canada Pension Plan which states that, beginning with the year 2024, contribution rates are the lowest constant rates that can be maintained over the foreseeable future.

5.6 Actuarial Balance Sheet

An actuarial balance sheet compares Plan assets to actuarial liabilities for contributors and beneficiaries under the present Plan provisions as at 31 December 2018 and 31 December 2030 on an open group basis.

In an open group actuarial balance sheet, it is assumed that the Plan is ongoing into the future, and the balance sheet takes into account future contributions of current and future members, as well as benefits of current and future members. Benefits and contributions are discounted at the assumed rate of return. Under the open group approach, using the legislated contribution rate of 9.9%, the base CPP assets are 100.6% of the base CPP liability at 31 December 2018 and 100.5% of the liability at 31 December 2030.

Under the open group approach, using the legislated first and second contribution rates of 2.0% and 8.0%, respectively, the additional CPP assets are 107.8% of the additional CPP obligations at 1 January 2019 and 106.4% of the obligation at 31 December 2030.

We concur with this paragraph of Section C.4 of Appendix C of AR30: "The CPP is intended to be long-term and enduring in nature, a fact that is reinforced by the federal, provincial, and territorial governments' joint stewardship through the established strong governance and accountability framework of the Plan. It therefore follows that if the base Plan's financial sustainability is to be measured based on its asset excess or shortfall, it should be done on an open group basis that reflects the partially funded nature of the base Plan, that is, its reliance on both future contributions and invested assets as means of financing its future expenditures. The inclusion of future contributions and benefits with respect to both current and future participants in the assessment of the base Plan's financial state confirms that the base Plan is able to meet its financial obligations over the long term."

5.7 Reconciliations

Detailed reconciliations are conducted and presented in Section 6 of AR30. The current results for the base CPP are reconciled in Section 6.1 against the results presented in AR27, while the current results of the additional CPP are reconciled in Section 6.2 against the results presented in AR28. These identify the principal causes of the changes in results and measure the impact of each on the results. The detailed reconciliations also serve as a check on the results of AR30.

5.8 Uncertainty and Sensitivity/Scenario Tests

In addition to the results based on best-estimate assumptions selected by the Chief Actuary, several sensitivity and scenario tests are presented and discussed in Appendix E – Uncertainty of Results. Showing the results of these tests not only illustrates the extent of probable variation in projected results, given the impossibility of forecasting accuracy over a long-term horizon. These show the results using alternative assumptions and thereby provide useful information regarding the uncertainty of the valuation results. In the following sections, our discussion is presented in the same order as appearing in AR30.

Highlights of the effects of selected alternative scenarios/assumptions (covering investment rates of return, mortality and economic growth) are shown in the Executive Summary of AR30 as Highlights of Uncertainty of Results, with a more extensive discussion and additional sensitivities/scenarios included in Appendix E. We concur with the prominence that this discussion is given. We concur that this provides the uncertainty of the best-estimates appropriate attention. It would be useful in Section 1.5 to indicate where in the subsequent sections of AR30 the effects of uncertainty and the sensitivity to other key assumptions are provided.

5.8.1 Sensitivity to Investment Policy - Base CPP

The first area examined for sensitivity is that of investment policy. Appendix E, Subsection E.2 of AR30 shows the impact of four alternative portfolios, compared to the best-estimate portfolio, on the base CPP minimum contribution rate. This shows that the minimum contribution rate is expected to decrease as more investment risk is taken. However, this expected decrease comes at the price of escalating risk (measured in AR30 by the standard deviation of one-year portfolio returns).

The next sensitivity test presented in Subsection E.2 shows the impact of investment strategy on the probability of a significant loss over a short period of time and, in turn, on the base CPP minimum contribution rate. This is accomplished by showing the probability of a 10% and 20% nominal cumulative loss in the portfolio during the 2031-2033 period for the best-estimate portfolio and each of the four alternative portfolios. The test shows that, assuming best-estimate assumptions are used before 2031, under all portfolios a 10% nominal cumulative loss during the 2031-2033 period would increase the minimum contribution rate by 0.51%, and a 20% loss would increase the minimum contribution rate by 0.69%. It also shows that the probability of such losses increases significantly as the equity allocation increases.

Unfortunately, this material fails to highlight for the reader the range of potential outcomes over the projection period associated with the alternative portfolios. It only shows the likelihood and impact of the losses on the 2033 minimum contribution rates, assuming the best-estimate portfolio assumptions are realized in the period leading up to 2031. This section of AR30 might have been improved by stochastically modelling the impact of the alternative portfolios on the minimum contribution rate over a 75-year projection period. Alternatively, a "Japanese" investment scenario, consisting of a low real rate of return over an extended period (or at least a longer transition to the time that the ultimate real rate of return is reached), could be utilized to help demonstrate the impact of a possible low investment return scenario. These would highlight the genuinely higher risk of more aggressive asset allocations or an extended period of low returns.

5.8.2 Sensitivity to Investment Policy - Additional CPP

Subsection E.2 also shows the impact of four alternative portfolios for the additional CPP, which differ from the alternative portfolios tested on the base CPP, given the CPPIB's two-pool investment structure. The impact of these alternative portfolios, compared to the best-estimate portfolio, is shown for the additional CPP's first additional minimum contribution rate (FAMCR). Similar to the base CPP, this shows that the first additional minimum contribution rate is expected to decrease as more investment risk is taken, but that this expected decrease comes at the price of escalating risk (measured in AR30 by the standard deviation of one-year portfolio returns).

Another sensitivity test presented in AR30 with respect to the investment policy of the additional CPP is determining the impact of alternative investment policies on the probability that the FAMCR falls within specified ranges. Range A (less than or equal to 1.69%) and range E (greater than or equal to 2.21%) identify, as specified under the Additional Canada Pension Plan Sustainability Regulations, when immediate action would be required. If the FAMCR falls within range B (1.70% to 1.79%) or range D (2.11% to 2.20%), no action is initially required. If, however, the rate remains in the same range for a second consecutive valuation, immediate action would be required. Therefore, ranges B and D are considered "warning" ranges. No action is required if the FAMCR falls in range C (1.80% to 2.10%) which, under the Best Estimate Portfolio, has a probability of 80%.

5.8.3 Sensitivity to Investment Environment and Financial Shocks

The next series of tests discussed in Subsection E.3 examine the sensitivity of the base Plan MCR and additional Plan AMCRs to investment environment and financial market shocks that would result in the MCR reaching the legislated contribution rate of 9.9% and the AMCRs falling into a range that triggers default actions defined by the Additional Canada Pension Plan Sustainability Regulations.

The first scenario examined a range of short-term returns in the future and the second scenario examined a permanent change in the expected rates of return. The outcomes of these alternative possible scenarios over different time frames are provided, along with associated probabilities. While useful, these results would be easier to understand if presented in a tabular format.

5.8.4 Individual Sensitivity Tests and Scenario Tests

Other sensitivity tests in AR30 examine one assumption at a time. Other tests examine separate economic growth and population aging scenarios. The tests illustrate the effect of changes, both lower-cost and higher-cost, in each of seven key assumptions and two scenarios (18 tests in total). These tests are discussed in detail in Appendix E, Subsection E.4.

The individual sensitivity tests examine the effect of changes, both lower-cost and higher-cost, in each of seven key assumptions: fertility, mortality, migration, price increases, real wage increases, real rates of return, and disability incidence rates. Except for the sensitivity tests that examine the impact of changes in fertility rates, mortality improvement rates and real wage increases, stochastic considerations are used to provide estimates of lower-cost and higher-cost scenarios within an 80% probability range for each assumption. That is, alternative scenarios are presented using the respective assumption at its projected 10th and 90th percentile. It is also true that any stochastic analysis has to be augmented by considerable professional judgement. In many cases one needs to consider the likelihood of a "regime" switch or structural changes.

Under the base CPP, the resulting higher-cost and lower-cost values for each assumption are used as inputs into the model to project revised (a) minimum contribution rates (b) pay-as-you-go rates, and (c) asset-to-expenditure ratios if there were no change in the current 9.9% contribution rate.

Fourteen individual sensitivity tests are run. In five of those cases (involving real rate of return on investments, real wage increases, mortality rates, fertility rates and disability incidence rates), the resulting minimum contribution rate exceeds 9.9%, a rate that would entail a review of benefits and contributions by the Ministers of Finance. This points out that considerable attention needs to be given to these assumptions.

Under the additional CPP, the resulting higher-cost and lower-cost values for each assumption are used as inputs into the model to project revised (a) first additional minimum contribution rate, (b) second additional minimum contribution rate, and (c) asset-to-expenditure ratios if there were no change in the current 2.0%/8.0% contribution rate.

Fourteen individual sensitivity tests were run. However, it is noted that fertility, net migration and disability incidence rates do not significantly impact the AMCRs. Of the remaining eight tests, three (real rate of return on investments, real wage rates and mortality rates) result in a FAMCR greater than 2.20% which would result in immediate action required by the Additional Canada Pension Plan Sustainability Regulations. This points out that particular attention needs to be given to these assumptions.

The approach to the individual sensitivity tests used in AR30 is the same one that was used in AR27. Each report shows how much variation should be expected, with roughly equal plausibility, in each direction and for each assumption or set thereof.

Regarding the two types of scenarios that were tested, AR30 presents one set of projections in the context of high and low economic growth. This combines optimistic and pessimistic economic assumptions regarding participation rates, unemployment rates, age of retirement benefit take-up, and real wage increases.
The report also presents two "combined" population aging sensitivity tests: the "Younger Population Scenario" and the "Older Population Scenario". Both scenarios are presented to indicate the variation in results from two possible combinations of key demographic assumptions (considering the interrelationship of the various assumptions) that provide a reasonable range of possible future outcomes. The assumptions that are combined are: fertility, mortality, net migration and labour force participation rates.

We believe that the illustration of alternative scenarios adds considerable value to the actuarial review. In light of this, we recommend that the OCA consider the use of additional alternative scenarios, possibly incorporating key elements of both economic and demographic assumptions on an internally consistent multi-variable basis.

Recent developments relating to the COVID-19 pandemic underscore the importance of multi-factor sensitivity analysis of the CPP. These developments to date demonstrate the interrelations and high correlation of some of the demographic, economic and investment factors under tail conditions. Although such conditions are impossible to accurately predict, their use in scenario development illustrates the value of such analysis in the assessment of the sustainability of the CPP. They can also be useful in setting the values used in sensitivity testing of the program.
We believe that sensitivity and scenario tests represent valuable tools, if used appropriately. They provide the reader information that may be used to estimate the financial impact of a change in a particular best-estimate assumption. They also point out that, because of the uncertainty involved in any projection of the future, the best-estimate results will not turn into reality, as well as the importance of the key assumptions used.

The reader should be cautious in interpreting the information provided about the likely variations in the assumptions. In particular, the reader should be aware that actual outcomes can range more widely than are shown by the sensitivity and scenario tests and that the likelihood of experiencing results outside the range shown may be even higher than is projected. This is due to the inherent unpredictability and interrelationship between some of the key assumptions, rather than a defect in the Chief Actuary's analysis.

We note that:

• Determining a plausible range for future variability is difficult and subjective;
• Past variability is not an unfailing guide to future variability; and
• The future is inherently unknowable, so it is prudent to take into account a range of expert opinion when making and interpreting projections. This will help both in setting best estimates and also in setting ranges for any sensitivity analysis.

The OCA may need to reintroduce more subjectivity in the choice of probability distributions and alternative scenarios used for certain sensitivity tests.

5.9 Opinion on Methodology

In our opinion, the actuarial methods used in completing AR30 were reasonable.

5.10 Recommendations

Recommendation 4

Given the uncertainty associated with future economic and demographic experience, we recommend that two additional scenarios be analyzed: (1) relating to a 'Japanese' investment scenario, that is, one based upon a long-term continuation of low real long-term interest rates on federal bonds and (2) a scenario considering demographic, economic and investment factors, such as those experienced with the COVID-19 pandemic. Other scenarios could also be explored for study.

Recommendation 5

We recommend that the Chief Actuary continue to seek additional expert input to be considered in establishing the best-estimate actuarial assumptions and the range of variability examined in the sensitivity tests. Specifically, the Chief Actuary should seek the input from demographic, economic and investment experts for their views on key assumptions, including a plausible range of variability that approximates an 80% confidence interval. The Chief Actuary should consider this input, but retain responsibility for the final assumptions that are chosen.

SECTION 6 – ASSUMPTIONS

In this Section, we address the following question:

"Were the assumptions used in completing the report reasonable?"

6.1 Background

The actuarial review that is required to be made every three years under Section 115 of the Canada Pension Plan requires that the Chief Actuary look back in time, to review the operations of the program, and also look forward to make an estimate of its future operations. For the forward-looking part of the process, the Chief Actuary builds a model that incorporates the details of the benefit, contribution and investment elements of the CPP and reflects the expected behaviour of the factors that determine the year-by-year development of these elements. The model for a plan as complex as the CPP is by necessity complex itself. The assumptions incorporated into the model for a particular actuarial review reflect the Chief Actuary's judgement, based on her interpretation of past experience and the best available evidence about the likely course of future experience.

The nature of the actuarial process is to make projections (not predictions) about the future based on the evidence available and then to review them periodically. Where appropriate, the actuary makes "mid-course corrections" in the assumptions as the emerging experience of the plan deviates from the previous assumptions and the expectations for likely future experience change. In assessing whether to change an assumption and if so, by how much, the actuary must weigh:

• long-term historical data;
• short-term historical data;
• evidence that a "regime switch" or structural change has taken place;
• recent amendments to the Canada Pension Plan;
• policy (for example, CPPIB investment policy, ESDC administration policies and government policies on inflation control and immigration levels);
• expert academic and professional research; and
• other external sources of relevant information.

The assumptions are intended to apply over the long-term future, so the actuary will normally give substantial weight to long-term historical data. However, where the actuary judges that more recent data for a particular assumption indicate a regime switch or a trend that is likely to continue for the long-term future, the actuary will recognize that switch or trend in the assumption. That is, if the actuary expects that future conditions will differ from those of the historical period, that expected difference will be considered.

For many of the assumptions used in the model, the Chief Actuary has adopted a method that actuaries describe as "select and ultimate". Under this approach, the particular assumption gradually changes over a period of years (the "select period") from one that initially is close to actual recent experience to one that reflects the actuary's best-estimate of the long-term future (the "ultimate" assumption). The length of the select period can differ for different assumptions. The choice is based on the actuary's judgement and depends on factors including the nature of the assumption involved and on how significantly the ultimate assumption differs from recent experience.

The results of the actuarial process at any given time do not yield the "right" answer, but should lie somewhere within a range that can be regarded as "reasonable". Previous actuarial reports on the CPP have focused on certain key assumptions. All assumptions used in those reports can be described as "best-estimate", that is, the assumptions were set, in the judgement of the Chief Actuary, such that adverse or favourable deviations of actual future experience from each of those assumptions are about equally likely. AR30 follows this same approach.

The Chief Actuary and her staff gather expert opinions on the key assumptions from a variety of sources. The review panel did the same. This is a difficult and sometimes frustrating exercise. If one visits, say, five experts, one can expect to come away with multiple answers, sometimes covering a wide spectrum of possibilities. Indeed, how a decision is made as to what the best estimate should be is not based a unique theory or scientific source. There is no mathematical formula that guarantees that one arrives at the "best" estimate. The most one can do is to arrive at an estimate that is "reasonable". The panel members collectively recognize this limitation when it assessed the work of the Chief Actuary.

The major actuarial assumptions in AR30 can be conveniently divided into two groups:

• demographic assumptions that deal with changes in the covered population (for example, fertility, migration and mortality rates) and events (for example, death, disability, recovery and retirement) that trigger the starting or stopping of CPP benefit payments or contributions; and
• economic assumptions that deal with such issues as employment, wages, prices and returns on investment.

6.2 Demographic Assumptions and Our Opinion

6.2.1 Fertility

Fertility rates varying by age group and year are applied to the female population to project the number of births each year.

The total fertility rate, the most commonly used fertility metric, is a convenient way of summarizing a set of age-specific fertility rates. It indicates the average number of children that would be born to a woman in her lifetime based on the age-specific rates in a given calendar year. The total fertility rate for 2017 for Canada, the most recent year available prior to the development of AR30, was 1.545 (for Québec was 1.602), while that for 2018 (available after the development of AR30) was 1.50. The ultimate total fertility rate assumption for AR30 for Canada was 1.62 (1.65 for Québec), a slight decrease from 1.65 for AR27 (1.68 for Québec).

Fertility trends are based on historical fertility rates by age of mother. The Chief Actuary assumes that recently observed fertility rates for females under age 30 will decrease slightly in the near future, while fertility rates observed from ages 30 to 50 will increase. This is consistent with an increase in average age at birth.

Similar to other assumptions, the approach used in AR30 (and in past actuarial reports on the CPP) is to develop one fertility assumption for Canada and a separate one for Québec. These assumptions are then used to develop separate population projections for Canada and for Québec. From these, the projected population of Canada less Québec is derived. The ultimate fertility rate assumed for Canada is 1.62, while for Québec it is 1.65. This differential is reasonable, although some converge might occur.

The long-term fertility assumption depends on several factors that are difficult to predict. Fertility rates at all child-bearing ages declined sharply in Canada in the 1960s and early 1970s as the result of social, economic and medical factors, including improved contraception methods that became more accessible. Since the mid-1970s, fertility rates at ages under 30 have continued to trend downward, while the rates at higher ages increased until 2008 and have fluctuated since then, partly due to the financial crisis of 2008 (fertility rates tend to decrease when unemployment rises). As a result, the average age of motherhood has increased and is expected to increase somewhat further.

In the future, fertility rates could remain at current levels or even decline to the lower levels experienced in several other developed countries (for example, Japan and Italy at 1.4), or increase in the direction of the higher rates recently experienced in the U.S. (about 1.7) or France (1.9). The assumed ultimate rate of 1.62 in AR30 is close to the Statistics Canada medium assumption of 1.59. The United Nations' assumed rates are about 1.54 in 2040 and about 1.70 in 2095. The total fertility rate assumption is within a reasonable range of possible ultimate rates.

We consider that a useful enhancement to the methodology used to develop estimates of future fertility rates would be to incorporate the use of individual age fertility rate input, which is available for use by the OCA from Statistics Canada. In addition, we suggest that the effects of the gradual increase in the average age of mothers at birth be further considered, especially in light of the 2018 fertility experience.

The sensitivity tests for the fertility assumption are a lower-cost ultimate total fertility rate for Canada of 1.92 and a higher-cost ultimate rate of 1.32. These rates define the range of values recently observed in the G7 countries, in particular in France and Italy, respectively. The test results may be summarized as follows:

The above table illustrates that changes in fertility can have a relatively large effect on the cost of the basic CPP. However, sensitivity test results should be interpreted with caution. Readers should form their own opinion about the plausibility of the lower-cost and higher-cost assumptions. Moreover, they should realize that changes in assumptions are not likely to occur in isolation. For example, a radical change in fertility rates may be accompanied by other changes that would mitigate their impact (e.g., changes in average ages at retirement, levels of immigration or labour force participation rates).

Opinion on Fertility

In our opinion, the AR30 fertility assumption is reasonable.

6.2.2 Mortality

The mortality assumed is developed in two steps: (1) general Canadian population mortality and (2) mortality for the several types of beneficiaries (e.g., retirees, the disabled, dependents and survivors). The purpose of the second step is to adjust the general population mortality rates to account for the specific mortality experience of CPP retirement and survivor beneficiaries. Mortality rates for disability beneficiaries are based on actual experience for that segment of population.

The basis for the general Canadian population included the Canadian Human Mortality Database, the Statistics Canada Life Tables for 2011-2013 and for 2014-2016, while the basis for beneficiary-specific mortality included Statistics Canada Life Tables for 2014-2016 and beneficiary-specific mortality for the most recent data available for Canada. Mortality for each of these sources was smoothed or extrapolated in areas such as above age 99 for males and 102 for females, where there was an insufficient volume to be fully credible.

For this report, the mortality rate projections are anchored from the 2015 mortality rates of the Statistics Canada Life Tables (CLT 2014-2016 Tables).

Assumed mortality rates for years 2016 to 2034 are derived by applying the cumulative estimated annual mortality improvement rates to the mortality rates of 2015. Annual mortality improvement rates (MIRs) are analyzed by age, sex and period. The historical MIRs are graduated using a best-fit log-linear regression. For ages 65 and over, the annual MIRs for 2016 to 2017 are further projected using the trends derived from the administrative data on Old Age Security (OAS) beneficiaries, representing 98% of the general Canadian population.

Mortality improvement rates for males at most ages are currently higher than those for females but are assumed to decrease to the same level as female rates from 2035 onward. The mortality assumptions of the 30^th CPP Report produce slightly lower life expectancies for males and females than those of the 27^th CPP Report, primarily due to the recent slowdown in the rate mortality improvement. The gap between male and female life expectancy continues to shrink somewhat, but does not disappear.

Several key risks exist to the mortality improvement rates over the short-term period. For example, deaths due to opioid overuse was a large contributor to the lack of improvement in life expectancy at birth between 2016 and 2017. It is too early to tell the course of these premature deaths, but this relatively high level may continue for several years, which contribute to the uncertainty in future mortality improvement rates. We believe it would be worthwhile in future actuarial reviews to determine how to treat this cause of death in the future.

In addition, it is currently difficult to discern the relative slowdown over the last several years in overall mortality improvement due to other factors. Possibly experience over the next three years will provide some light on this puzzle.

In any event, partly because of these changes in underlying mortality, it will be important for the next triannual review of mortality to avoid relying on the experience of a single year to anchor future mortality rate projections.

The AR30 ultimate mortality improvement rates are the same as those assumed in AR27 at all ages and both sexes. These are a 0.8% annual improvement through age 89, decreasing to 0.0% at age 110. Compared to the improvement rates used in the U.S. Social Security (OASDI) 2015 Trustees Report, the AR30 ultimate improvement rates for males and females under 65 are lower than the U.S. rates but larger for males and females aged 65 to 99 (note that U.S. mortality experience is significantly higher than those of Canada). As to CPP assumptions versus U.K. assumptions, the CPP assumed mortality improvement rates are lower at all ages for both males and females than those assumed in the United Kingdom (where initial mortality rates are also higher than in Canada).

Separate studies of mortality for retirees, those who are disabled and survivors were performed. A study was conducted of mortality for four pension level categories for retiree categories, that indicated that mortality improvement rates, at least until the age group 90-99, have been greater for those at higher pension levels. Consideration should be given to enhance this analysis, possibly with the use of clustering of a small number of pension level categories in the determination of future mortality improvement projections for retirees. Although similar relationships may arise for the disabled and survivors, it is unlikely that the results of this further analysis will result in material differences in cost projections. Nevertheless, these relationships should be reviewed to confirm their materiality.

The sensitivity tests for the mortality assumption were implemented by adjusting the assumed rates of mortality improvement through deterministic methods. Based on the best-estimate assumptions used in AR30, the life expectancy at age 65 for retirees in 2050 would be 23.3 years for males and 25.6 years for females, with corresponding life expectancies for the lower-cost and higher-cost assumption shown in the first column of the following table. The lower-cost assumption postulates no future mortality improvement after the year 2035. Thus, AR30 gives cost estimates for constant, non-improving life expectancy after 2035. The test results may be summarized as follows:

In addition, AR30's Executive Summary indicates that MCR for a scenario in which there are no mortality improvements at all of 9.03 for the base CPP and 1.72 and 6.88 for the additional CPP.

Opinion on Mortality

In our opinion, the AR30 mortality assumption is reasonable.

6.2.3 Migration

The rate of net annual migration to Canada since 2001 has varied from a low of 0.48% of the population to a high of 1.11% (if we assume the net flow of non-permanent residents equals zero, then the range is 0.46% to 0.74%). The volatility in the earlier years of the 20th century was much more extreme. Migration varies from year to year in response to demographic, economic, social, political and administrative changes. Historically, this has been the most volatile component of change in the Canadian population. Over 2009 to 2018, the average net migration rate for Canada was 0.62% of the population (assuming the net flow of non-permanent residents equals zero).

The migration rate for Québec is calculated and projected on its own. For Québec, the ultimate net migration rate is 0.43% (average over 10 years) starting in 2021.

Historically, the average annual net flow of non-permanent residents was assumed to be equal to zero because of the large variations (both positive and negative) in this variable and the small weight they carry in terms of contributions to the CPP. This is the assumption in AR30.

The AR30 assumption for net annual migration to Canada decreases from the 1.11% level experienced in 2018 to 0.86% in 2019, 0.73% in 2020, and 0.62% in 2021 and beyond. This ultimate rate is the same as the ultimate net migration rate of 0.62% assumed in AR27.

Migration is a major driver of population growth. The assumed ultimate net migration rate in AR30 (0.62% in 2021) is lower than the medium projection rate used by Statistics Canada in its projections (0.68% in 2029-30 and beyond). We are comfortable with this difference.

Statistics Canada indicated to the peer reviewers that postcensal estimates indicate that the number of emigrants appearing in general information sources is underestimated in Canada. Statistics Canada uses indirect techniques for estimating more precisely the number of emigrants and this results in a number of projected emigrants that is higher than the number projected by the OCA, with a resulting lower net migration rate projected by Statistics Canada. It is suggested that the OCA consider, in future actuarial reports, the adjustments made by Statistics Canada to better estimate the number of emigrants.

There is regular debate among demographers as to whether future external immigration will take the form of a percent of total population or a constant number of individuals. Both AR30 and Statistics Canada have taken the view that immigration will take the form of a percentage of total population, while the Trustees of the U.S. Social Security Trust Fund have utilized the approach of a constant rate of population. In addition, AR30 has utilized the constant amount approach in migration from Québec to the rest of Canada. At this time, this apparent inconsistency is reasonable. However, it may provide a useful perspective to understand the effects of the constant approach.

There has been a significant upward spike among temporary foreign-born student. It is unclear how this will affect future consequential more-permanent immigration. It is now assumed that that if this occurs, it will reduce other permanent immigrations. This bears watching to ascertain whether this assertion is appropriate.

The sensitivity tests for the net migration assumption are based on an 80% confidence interval. The assumptions are an ultimate lower-cost annual net migration rate for Canada of 0.68% and a higher-cost rate of 0.57%. The test results may be summarized as follows:

Opinion on Migration

In our opinion, the AR30 migration assumption is reasonable.

6.2.4 Disability Incidence

The assumption regarding the incidence of disability takes the form of rates that vary by age, sex and calendar year. These can be summarized as an aggregate rate based on the current population distribution. The experience indicates aggregate rates for the period between 2015 and 2017 of 2.94 new disabilities per thousand eligible male workers and 3.62 new disabilities per thousand eligible female workers. The AR30 assumption is that disability incidence will experience aggregate rates for years 2020 and later of 2.95 for males and 3.65 for females for years 2020 and beyond. The corresponding adjusted ultimate rates in AR27 were 3.10 for males and 3.65 for females for years 2017 and beyond. 

The use of historical data as the basis for assumptions about the future must always be performed with care. In this case, very little weight can be given to experience data for the years before 1995, when major changes in the administration of the disability provisions occurred that led to a significant decline in disability incidence rates. The Chief Actuary must also take into account the effect of relevant changes in the law, such as those introduced by the 2007 Plan amendments, which relaxed the minimum qualifying period, effective 3 March 2008, for those with 25 or more years of contributions. 

We note that, based upon a non-representative survey of the private sector disability insurance industry, disability incidence appears to be increasing, especially due to mental and stress-related illnesses. We recommend that this area be monitored closely to determine whether a corresponding change in disability incidence assumptions is warranted.

A review of whether or extent that disability incidence differs by earnings level may contribute to the understanding of future disability incidence patterns. We recommend assessing these historical patterns.

Another related assumption of disability incidence is the termination of those who were collecting disability benefits. The two components of these terminations are deaths and recoveries. The AR30 assumption regarding both of these components are reasonable.

The sensitivity tests for this assumption are an ultimate (2020 and beyond) lower-cost rate per thousand eligible workers of 2.10 for males and 2.80 for females, and an ultimate higher-cost rate of 3.75 for males and 4.50 for females. This is again based on an 80% probability range. The test results may be summarized as follows:

Opinion on Disability Incidence

In our opinion, the AR30 disability incidence rate assumption is reasonable.

6.3 Labour Force and Economic Assumptions and Our Opinion

6.3.1 Labour Force

Retirement Take-up Rates

The contributions to the Plan and benefits paid from the Plan are affected by the ages at which individuals start their CPP pension. Prior to the implementation of the 2009 Plan amendments, contributions to the Plan by individuals, and by employers on their behalf, stopped when the individual started to collect a retirement pension, and neither contributions nor benefit accruals were resumed, even if the individual returned to employment. The Plan’s normal retirement age is 65, the earliest commencement age is 60, and the latest commencement age is 70. Prior to the implementation of the 2009 Plan amendments, pensions were reduced by 0.5% for each month by which the pension start age was below 65 or increased by 0.5% for each month by which the pension start age was after age 65.

The 2009 Plan amendments to the CPP, effective 1 January 2012, removed the need for those under age 65 to cease working for the month before and the month of benefit commencement (the Work Cessation Test). Those who choose to receive the retirement benefit while continuing to work must still participate in the CPP by making continuing contributions (matched by their employer) until age 65, and they receive commensurate benefit increases. Between ages 65 and 70, pensioners who are working can opt to contribute to the CPP and, if so, their employers must also contribute, with commensurate benefit increases to the pensioner.

Further, the early and late retirement pension adjustment factors were changed to factors that do not involve a subsidy. From 2016, the Pre-65 Downward Monthly Adjustment Factor is 0.6% and the Post-65 Upward Monthly Adjustment Factor is 0.7%. In accordance with subsection 115(1.11) of the Canada Pension Plan, the Chief Actuary must recalculate the pension adjustment factors and specify them in every third triennial report. This was last documented in Actuarial Study No. 18: Canada Pension Plan Actuarial Adjustment Factors, published in April 2017. This study indicated that the 0.6% early retirement reduction factor and the 0.7% late retirement addition factor were deemed to be suitable at that time.

Although these adjustment factors were expected to encourage participants to ask for benefits at a later age, the removal of the Work Cessation Test has in part offset this effect.

The retirement take-up rate represents the ratio of the number of individuals who elect to start receiving their retirement pension at a specific age to the total number of individuals of a cohort who are eligible for a retirement pension. Separate retirement take-up rates are assumed for each year, each sex, and each age from 60 to 70 inclusive.

For cohorts reaching age 60 in 2021 and thereafter, the retirement take-up rates at age 60 are assumed to be 27.0% (males) and 29.5% (females). The retirement take-up rates at age 65 are assumed to be 46.4% in 2026 and thereafter for both sexes. The rates result in a projected average age at take-up of 63.2 years in 2040, which is slightly higher than was observed (62.4) in 2012.

Some observers suggest that, because of improvements in health and life expectancy, together with the prospect of tight labour markets associated with the retirement of the baby boomers, there could be a tendency for individuals to retire at older ages in the future. Others suggest that these considerations must be balanced against entrenched social expectations of early retirement. As a result of this difference of views, take-up experience should be followed closely for use in future actuarial reviews.

There is no separate sensitivity test for retirement take-up rates. However, they are part of a combined variable called “Changes to Labour Market” as part of the “Economic Growth” sensitivity test.

Unemployment and Labour Force Participation Rates (higher and lower economic growth scenarios)

The assumption for net job creation in Canada is established so that the assumed rate of unemployment, 5.9% in 2019, increases to a constant 6.2% from 2030 onwards (the same ultimate assumption as AR27). This is in line with various economic forecasts and reflects moderate economic growth.

The development of projected numbers and profiles of contributors begins with the development of calendar year labour force participation rates by age-sex categories and the application of these rates to the projections of the total population in each of those categories. The projections are done separately for Canada and Québec. The participation rates are “cohort-based”. The participation rates for all age groups are expected to increase due to:

• the attractive employment opportunities resulting from labour shortages;
• the aging of cohorts with stronger labour attachments, especially for women and individuals with higher educational attainment; and
• a shortage of retirement savings encouraging longer labour force attachment.

Under the best-estimate scenario, the participation rate of those aged 18 to 69 is expected to increase from 75.9% in 2018 to 79.2% in 2035. These rates are consistent with similar rates produced by other Canadian economic forecasters.

The resulting labour force projections are then used, in combination with projections of assumed net jobs created, to give projections of employed and unemployed workers.

A deterministic model is used to generate the lower-cost and higher-cost scenarios for these assumptions. A probabilistic range is not used for these assumptions because the pressures from an expected tight labour market as the baby boom retires seem unlike any labour situation experienced in the past (i.e., a regime switch). Instead, the lower-cost ultimate assumed unemployment rate is 4.2% starting in 2030, together with an ultimate aggregate labour force participation rate for ages 18 to 69 of 84.6% starting in 2035. By 2035, retirement rates at age 60 are assumed to gradually decrease to levels that are 20 percentage points lower than the best estimates, i.e., 7.0% and 9.5% for males and females respectively. This results in an increase in the projected average age at take-up from 63.2 to 64.2 in 2040. The higher-cost ultimate assumed unemployment rate is 8.2% starting in 2030, together with an ultimate aggregate labour force participation rate of 76.2% starting in 2035. By 2035, retirement rates at age 60 are assumed to gradually increase to levels that are 20 percentage points higher than the best estimate, i.e., 47.0% and 49.5% for males and females, respectively. This results in a decrease in the projected average age at take-up from 63.2 to 62.3 in 2040.

The sensitivity test results may be summarized as follows:

Opinion on Unemployment, Labour Force Participation and Retirement Rates

In our opinion, the AR30 assumptions as to the rates of unemployment, labour force participation and retirement benefit take-up are reasonable.

6.3.2 Price Increases

The rate of price inflation is a necessary assumption for an actuarial review of the CPP. CPP contributions, benefit payments and investment returns are all affected by inflation. However, the extent and timing of these effects are not offsetting. The net result is that an increase in the inflation assumption results in a decrease in minimum contribution rates for both the base and additional CPP, and vice versa.

Inflation in Canada was extremely volatile during the 20th century, with long runs of both very high and very low inflation. However, since 1991 the Bank of Canada and the Minister of Finance have jointly established inflation-control targets. These targets have been agreed upon for five years at a time. In October 2016, it was announced that the Bank of Canada and the Minister of Finance agreed to extend the targets for another five years to 2021. This framework has been remarkably successful at keeping the inflation rate in Canada generally within a range of +/-1% around the policy target.

The Bank is currently assessing the possible use of a dual mandate of targeting inflation as well as GDP growth or employment. However, the Chief Actuary considers it very likely that the Bank will renew its inflation target commitment (2%, with a range of 1% to 3%) or at least that this target will constitute an important part of the future Bank’s mandate. The Chief Actuary thus used a price increase assumption in AR30 of 2.0% for the year 2019 and thereafter.

In the long term, there may be upward challenges, such as pressure from U.S. inflation (note that the U.S. Federal Reserve Board has a target inflation rate of 2.3% expressed in the same terms as Canadian inflation rate) and potential tight labour markets. Other research shows that an aging population, such as expected to occur in Canada, will put downward pressure on inflation.

The sensitivity tests for this assumption are a higher-cost scenario with an ultimate price increase rate of 1.5% in 2019 and beyond and a lower-cost scenario with an ultimate rate of 2.6%, in 2019 and beyond. This is based on a stochastic approach with an 80% confidence interval. The results of these tests may be summarized as follows:

Our inquiries lead us to believe that the current framework with respect to prices will continue for a long time and that there is no reason to expect a change to the current target of 2%.

Our review of the opinions of some economists and financial forecasters also found a concentration of views of long-term inflation rates around 2%. Thus, we consider that maintaining same the long-term assumption of 2.0% in AR30 as appropriate.

Opinion on Price Increases

In our opinion, the AR30 price increase assumption is reasonable.

6.3.3 Real Wage Increases

Both contributions and initial benefits under the CPP are affected by wage increases. Subsequent benefit increases are affected by inflation. The wage increase assumption is separated into two parts: the inflation assumption (discussed in Subsection 6.3.2) and the real wage increase assumption (the portion of wage increases above inflation), which is discussed here. 

In AR30, the real wage increase assumption is applied to both average annual earnings (AAE, used to project contributory earnings) and to average weekly earnings (AWE, an index used to adjust the Year’s Maximum Pensionable Earnings and the Year’s Additional Maximum Pensionable Earnings). 

There are five main factors that influence increases in the real wage: (1) changes in general productivity, (2) the extent to which changes in productivity are shared between labour and capital, (3) changes in the compensation structure offered to employees, (4) changes in the average number of hours worked, and (5) changes in labour’s terms of trade (i.e., the difference in the way the prices of goods produced by workers (measured by the GDP deflator) in comparison to shifts in the prices of goods consumed by workers (measured by the CPI)).

The most important of these factors in setting the real wage increase assumption is growth in productivity. An aging population is expected to have a dampening effect on productivity growth. We recommend a review of the assumption methodology underlying the real wage assumption, particularly as to whether simplification is appropriate, focusing on the number of factors considered and the economic/demographic components included.

The real increase in AAE is assumed to be the same as the real increase in AWE. An ultimate real wage increase of 1.0% has been assumed to be the best-estimate for years 2025 and thereafter. This was a decrease from that assumed in AR30 of 1.1%.

The University of Toronto’s updated forecasts (February 2019), provided after the real wage assumption was set for AR30, projected ultimate real wage increases of 1.4% starting in 2026. In 2019, the Conference Board of Canada forecasted a lower real wage growth rate of 0.5% for 2020, gradually increasing to 0.9% in 2040 and thereafter.

A deterministic approach was used to establish the sensitivity tests for the ultimate real wage increase assumption for 2025 and beyond. This gives a first scenario at 1.7% (from 2025) versus a second scenario at 0.3% (from 2019). These sensitivity tests have opposite effects on the base and additional CPP. The results of these tests are shown below:

Opinion on Real Wage Increases

In our opinion, the AR30 real wage increase assumption is reasonable.

6.3.4 Real Rate of Return on Investments

If the CPP was totally unfunded (that is, if the contributions each year were just enough to cover that year’s benefit payments and operating expenses), then the CPP contribution rate would be equal to the pay-as-you-go rate and no assumption for the rate of investment return would be required. However, sizeable funds are projected to accumulate for both the base and additional CPP and the rate of investment return becomes a material factor in determining the contribution rates for the Plan. The base CPP assets totaled $372 billion at the end of 2018 (12.4% higher than expected in AR27) and are projected to grow significantly over the coming decades.

As with the assumed increases in employment earnings and benefit payments, part of the assumed nominal rate of investment return is attributable to general price inflation. Here we focus on the real rate of investment return (that is, net of the rate of inflation). 

The foundation of the CPPIB’s investment strategy is a two-asset portfolio, referred to as the “reference portfolio”. The reference portfolio is used to establish how much risk the CPPIB is willing to take and consists of a global equity benchmark and a Canadian government nominal bond benchmark.

With the introduction of the additional CPP, the CPPIB Board approved two different reference portfolios for each component of the Plan. The reference portfolio for the base CPP is 85% global equity and 15% Canadian government nominal bonds (the reference portfolio was 72% equity and 28% bonds in fiscal 2015-2016). The reference portfolio for the additional CPP is 50% global equity and 50% Canadian government nominal bonds.

In order to invest the assets of the base and additional CPP according to their reference portfolios, the CPPIB developed a two-pool investment structure made up of a Core pool and a Supplementary pool. The Core pool is a diversified portfolio of equities, fixed income and real assets, whereas the Supplementary pool consists of fixed income securities. The base CPP is invested 100% in the Core pool, while the additional CPP is invested in both the Core and Supplementary pools such that the overall risk level is consistent with its reference portfolio.

As at December 31, 2018 the asset mix of the base CPP is 56% equities (32% public, 24% private), 18% fixed income (17% marketable bonds, 6% non-marketable bonds, 9% credit, -14% cash) and 26% real assets. The -14% cash is debt used for leverage purposes. For the base CPP, the best-estimate real rate of return assumption in AR30, net of investment expenses, is 2.9% in 2019 and 2020, then increases by 1.0% to 3.8% in 2021. It then fluctuates slightly before rising to an ultimate assumption of 4.0% in 2029 and later. This produces an average real rate of return over the 75-year projection period of 3.95%.

For the additional CPP, the best-estimate real rate of return assumption in AR30, net of investment expenses, starts out at -0.7% in 2019, due to the assumed implementation expenses of $9 million and the assumed increase in bond yields. The assumed real rate of return then increases to 0.4% in 2020 and 2.3% in 2021, thereafter increasing steadily to an ultimate assumption of 3.6% in 2029 and later. This produces an average real rate of return over the 75-year projection period of 3.38%.

The ultimate rates reflect a “building block approach” whereby, in real terms:

• Long-term Government of Canada bonds are assumed to yield 2.6% per year from 2029 onward.
• Marketable bonds (Core pool) are assumed to return 2.2% per year by 2029 and thereafter.
• The Supplementary pool is expected to have a higher proportion of long-term bonds relative to the Core pool, so the ultimate expected rate of return is slightly higher than the Core pool marketable bonds at 2.3% from 2029 onward.
• Credit, which includes investments in corporate bonds, private debt and private mortgages is assumed to return an ultimate real rate of 2.9% from 2029 onward. 
• Canadian and foreign developed market equities are assumed to return 3.1% per year higher than the yield on long-term Government of Canada bonds. Emerging market equities are assumed to be 0.90% higher than developed market equities. This results in a lower-than-historical total equity risk premium.
• The ultimate total real rate of return for public equities is 4.3% per year beginning in 2024.
• The ultimate total real rate of return for private equities decreases from 5.7% in 2019 to 4.9% per year beginning in 2025.
• Real estate, infrastructure and natural resource investments are assumed to provide a real return that increases from 1.6% in 2019 to an ultimate rate of 4.1% in 2029 and beyond.
• The actual CPP portfolio also holds non-marketable provincial bonds of steadily diminishing duration and importance since this component of the portfolio (6% of the total) is being gradually wound down by 2042. These bonds are assumed to earn differing returns over the intervening period based on the expected make-up of the portfolio and the Chief Actuary’s expectations for changes in bond yields.

While the expected ultimate real rates of return are reasonable, the sudden 1% increase in 2021 to the expected real return of the base CPP and 1.9% increase in 2021 to the expected real return of the additional CPP seems highly unlikely and faster than we would have expected, given that if interest rates were to rise so quickly, it would cause other negative effects on the economy. We recommend that the speed and period of transition to the ultimate expected return be reviewed in the context of opinions of investment experts.

The OCA characterizes the asset mix of the base CPP as at 31 December 2018 to be 56% equities, 18% fixed income securities and 26% in real assets. Although the CPPIB has not adopted long-term asset mix targets, the Chief Actuary assumes an ultimate allocation of assets equal to 42% equities, 32% fixed income securities and 26% real assets, reflecting the CPPIB’s planned near-term increase in commitment to real estate and infrastructure investments.

It is assumed that the use of leverage decreases over time. We recommend that this assumption be reviewed in the context of CPPIB’s strategic direction regarding the use of leverage and the resulting risk involved.

The Chief Actuary also assumes that annual investment management expenses will equal 1.0% of Core pool assets. This is 0.8% higher than the estimated 0.2% cost of passively implementing the assumed asset mix. However, this 0.8% is assumed to be exactly offset by the added value produced by the CPPIB’s investment management, resulting in an assumed net cost of 0.2%. Thus, it is assumed that no net value is added or lost by the CPPIB’s investment management.

The sensitivity tests for this assumption are to increase or decrease the rate of return on all of the CPP assets to reflect an 80% confidence interval. For the base CPP, this gives a range from 2.35% to 5.55% and for the additional CPP, this gives a range of 2.48% to 4.28%, both with 80% probability. The results of these tests are summarized below and show the sensitivity of the MCR, FAMCR and SAMCR to this assumption:

We believe that the modelling approach used by the Chief Actuary has increased in sophistication relative to the approach used in development of AR27, particularly as the CPPIB has adopted its two-pool investment structure and continues to increase the sophistication of its approach to investing the CPP assets. We also expect that the Chief Actuary will continue her research and analysis concerning the size and sustainability of the equity risk premium. Such analysis should not be confined to the Canadian marketplace, since the CPP fund is heavily invested in non-Canadian assets. Also, the analysis should not be confined to a review of the past, since the future may differ substantially.

We also note that the Chief Actuary’s assumptions do not include an additional allowance for the CPPIB outperforming the normally expected returns for its asset classes. According to the CPPIB’s 2019 annual report, the CPPIB’s Investment Planning Committee (the primary management committee with investment responsibilities) has three primary measures of success, one of which is to deliver net dollar value-added relative to each Reference Portfolio. In support of this measure, the compensation structure for all levels of employees ties incentive compensation to total fund investment results, with equal weighting given to total return and dollar value-added relative to the Reference Portfolios. In fiscal 2019, the CPPIB delivered a net dollar value-add of $6.4 billion, and indicates that $29.2 billion of compounded dollar value-add have been generated since inception of active management at April 1, 2006.

According to Section 3230.03 of the Canadian Institute of Actuaries Standards of Practice, actuarial practice is to anticipate such higher value-added incremental returns only if “the actuary has reason to believe, based on relevant supporting data, that such additional returns will be consistently and reliably earned over the long term.”  While we agree with the Chief Actuary’s approach at the current time, we believe it is important to closely monitor the relationship both between incremental returns and incremental expenses, and the allocation of expenses between the basic CPP and additional CPP, with a view to refining this assumption in the future.

Finally, it is important to understand that the 4.0% ultimate real rate of return assumed in AR30 for the base CPP and the 3.6% ultimate real rate of return assumed for the additional CPP are not performance targets for the CPPIB to strive to attain. Rather, these assumptions are the Chief Actuary’s estimates of what the returns will ultimately be over a long-term period, given the supporting assumptions as to future asset mix and future real rates of return for the various classes of assets. The mandate of the CPPIB is to earn a maximum rate of return without undue risk of loss.

Opinion on Real Rate of Return on Investments

In our opinion, the Chief Actuary’s real rate of return assumption is reasonable.

6.4 CPP Operating Expenses and Our Opinion

CPP operating expenses include those incurred by ESDC, CRA, Service Canada, OSFI, and the Department of Finance. AR30 includes Table 7 and 19, which show the projection of CPP operating expenses for the base and additional CPP, respectively, as well as Tables 99 and 100, which relate those expenses as percentages of total earnings.

In calendar year 2018, operating expenses of the base CPP, excluding CPPIB (see section 8.1 of this report), amounted to $636 million. When expressed as a percentage of total annual earnings, they have averaged 0.092% over the last 10 years. Base CPP operating expenses have been mostly stable in the past (with the exception of a one-time expense that arose in 2012) and are projected to be stable in the future. Accordingly, a ratio of 0.090% is assumed for 2019 and thereafter.

For the additional CPP, the short-term estimates regarding operating expenses are provided by the ESDC and the CRA for fiscal years 2019-2020 to 2021-2022. For CPP30, it is assumed that the operating expenses for 2017 and 2018 (the start-up costs) will be charged to the Additional CPP Account in calendar year 2019 in addition to the expenses incurred in that year. Over the long term, it is projected that the additional CPP operating expenses will represent 0.010% of total earnings for 2024 and thereafter. Because the difference in the operating ratio between the two funds is currently expected to be so different, we recommend that the OCA monitor the allocation of expenses between the base CPP and additional CPP to ensure reasonableness.

Opinion on CPP Operating Expenses

In our opinion, the Chief Actuary’s assumption regarding CPP operating expenses is reasonable.

6.5 Assumptions in the Aggregate and Opinion Thereon

The Chief Actuary’s actuarial assumptions are her best estimates, based on her review of past experience and judgement about the likely course of future experience. For most assumptions, there is considerable room for actuarial judgement. As a result, the range of values that would be considered reasonable can be quite wide. In our review of the major actuarial assumptions, we concluded that each of them is within the reasonable range.

Notwithstanding all this, we caution the reader that, since the range of reasonable assumptions is wide, so too is the range of reasonable valuation results. The sensitivity tests in AR30 present plausible ranges of results. However, actual results could be outside of the lower and higher cost projected results.

Opinion on Assumptions in the Aggregate

In our opinion, the assumptions made in completing AR30 are, in the aggregate, reasonable.

6.6 Recommendations

Recommendation 6

Regarding the setting of assumption on the real rate of return, we recommend that the Chief Actuary:

• continue to improve the degree of sophistication of its investment model to more closely match the CPPIB investment portfolio;
• review the speed and period of transition to the ultimate expected return in the context of opinions of investment experts;
• closely monitor the allocation of expenses between the basic CPP and additional CPP, with a view to refining this assumption in the future;
• review the assumption on the use of leverage in the context of CPPIB’s strategic direction regarding the use of leverage and the resulting risk involved.

SECTION 7 – COMMUNICATION OF RESULTS

In this Section, we address the following question:

“Does the 30^th Report fairly communicate the results of the work performed by the Chief Actuary and his staff?”

7.1 Background

AR30, as tabled in the House of Commons on 10 December 2019, is a bound soft-cover book, separately published in English (215 pages) and French (228 pages).

The English version of AR30 is available from the OSFI website at:
http://www.osfi-bsif.gc.ca/Eng/Docs/cpp30.pdf

and the French version at:
http://www.osfi-bsif.gc.ca/Fra/Docs/cpp30.pdf

Supporting information for AR30 is available in great detail. For example, the OSFI website provides links to:

• all current and prior “Actuarial Reports” produced since the inception of the CPP;
• “Actuarial Studies”;
• documents relating to the “CPP Actuarial Peer Review”;
• “Speeches and Presentations” by the OCA staff;
• material presented at CPP-related and QPP-related “Inter-Disciplinary Seminars”; and
• various other CPP-related documents.

In addition, the supporting working papers for AR30 provide a comprehensive overview of the methodology and assumptions used in deriving the projections shown in AR30. Detailed projection tables are available, in CD-ROM format, to the public upon request and are also provided to all provincial governments. We believe that the information in AR30, together with the supporting CD-ROM, is sufficient for any reader to obtain a full understanding of the analysis.

7.2 Observations

AR30 is a very informative document. It includes a great deal of detail, a comprehensive Executive Summary, and many useful tables and charts. The overall conclusions are clearly set out.

AR30 has both a broad audience and a technical audience. The broad audience is mainly interested in the high-level results of the actuarial review. The technical audience of actuaries, economists, demographers, policy analysts, and others is interested in more extensive detail regarding the Plan provisions, data, methodology, assumptions, demographic projections and financial projections.

The Chief Actuary recognizes the dual nature of the audience and has changed the layout of the report to include the high-level results in the main body of the report, while placing the more technical information in appendices. Thus, the form of the report continues to be a compromise, containing more detail than is needed by the broad audience, and less than may be desired by technical readers.

To provide what may be a more accessible version, a separate file should be available of the website that only provides an Executive Summary, suitable for those who do not access the entire report. This might include more information than is currently given in the Executive Summary, recognizing that many of those who read the shorter form will not read the rest of the report.

We anticipate that the Chief Actuary will continue to enhance the format and information provided in the actuarial valuation report in future such reports. For example, even more charts could be provided to illustrate several additional areas of this complex topic. The 2019 Technical Panel on Assumptions and Methods (Report to the Social Security Advisory Board, United States) provides good examples. In addition, certain figures and graphs included in the Actuarial Report of the Quebec Pension Plan as at 31 December 2018 that should be considered for inclusion that would make the document more accessible to unsophisticated readers.

We also believe that in several of the tables included in the report, a certain number of years of historical data should be added to provide perspective to the reader to assist in understanding the patterns involved. At the same time, certain tables, for example Table 7, could be shortened by excluding the data for some of the individual years.

In the presentation of the actuarial balance sheet, a non-specialized reader could think that this confirmation of long-term financial sustainability results directly from the fact that the ratio of assets to liabilities is above 100% in 2018 and in 2030. Considering that this ratio was below 100% in previous actuarial reports such as AR26 and that the conclusion was the same, we think that the report should mention that the assets-to-liabilities ratio is not an absolute measure of the Plan's financial sustainability, that the CPP can tolerate fluctuations of this ratio below and above 100% and still be on solid financial ground, and that that the actuarial balance sheet is complementary to the minimum contribution rate as a measurement of the long-term financial sustainability of the Plan.

In addition, the actuarial report could mention that, since 2011, the CPP actuarial balance sheets appear in the notes to the Public Accounts of Canada, which reinforces the necessity to present this information in CPP actuarial reports.

7.3 Opinion on Communication of Results

In our opinion, AR30 fairly communicated the results of the work performed by the Chief Actuary and the staff who worked on the actuarial report.

7.4 Recommendations

Recommendation 7

We recommend that the Chief Actuary continue to enhance the contents and accessibility of future Actuarial Reviews. For example, a separate file should be available on the website that provides an Executive Summary suitable for those who do not access the entire report, expand the use of charts to illustrate additional items to enhance the understanding by the reader of key factors involved with the AR, and include a certain number of years of historical data to assist the reader in understanding the key patterns involved.

SECTION 8 – OTHER MATTERS RELATED TO AR30 AND RECOMMENDATIONS THEREON

In this Section, we address two other issues that we considered in our review, namely:

• CPPIB expenses; and
• external guidance in selecting assumptions.

8.1 CPPIB EXPENSES

CPPIB expenses are not included in CPP operating expenses in AR30 but are reported as allocated to the CPPIB assets and accounted for separately in the investment expenses assumption. We support this method of reporting.

From the AR30 working papers and the CPPIB Fiscal 2019 Annual Report, we have the following information regarding total CPPIB expenses, measured as a fraction (in basis points, or bps) of average assets under CPPIB management:

The upward trend in total CPPIB expenses reflects a larger portfolio and a progressively more active investment management strategy, as well as a performance-based fee structure that has resulted in increasing cost when performance has been beyond target (as has been the case in recent years). Over the last three calendar years, total expenses, as a percentage of assets, averaged 96 bps. Intuitively, one would expect that, over time, as operations become more established and as assets under management grow, expenses as a percentage of assets under management will stabilize, if not decrease (unless the growth in performance-based fees outpaces the growth in assets). We have been advised by the CPPIB that they are still building their active management capabilities and that recent performance has been very strong.

Fiscal year 2007 was the first year of implementation of the CPPIB active management strategy. At that time, the CPPIB also created the CPP Reference Portfolio, which is a hypothetical portfolio against which the value-added returns from active management are measured. In general, the objective of active management is to generate returns in excess of those from this Reference Portfolio, after reduction for the additional expenses incurred from active management. Thus, the additional returns from a successful active management program should at least equal the cost incurred to pursue active management.

The following table compares the CPPIB added value (from which external management fees and transaction costs have already been deducted) with the internal operating expenses of the CPPIB. The net added value, after all CPPIB expenses, is then shown.

Over the period 2007-2018, the net value added was extremely variable, with both large positive and large negative amounts.

Consistent with Canadian actuarial practice, AR30 reflects an assumption that the added value from active management will exactly equal the incremental expenses incurred to pursue that active management strategy. That is, no net added value is assumed to be either gained or lost. We agree with this approach at the current time and believe it is important to closely monitor the relationship both between incremental returns and incremental expenses, and the allocation of expenses between the basic CPP and additional CPP, with a view to refining this assumption in the future.

8.2 External Guidance in Selecting Assumptions

AR30 is a vitally important document. Its audience is not only the federal and provincial governments, who are responsible for the governance and administration of the CPP, but also includes the millions of present and former contributors and beneficiaries who rely on the CPP for their financial security. The assumptions used in the report should be, and be seen as, the best available unbiased forecasts of future events.

Because of the wide range and complexity of the assumptions and methodologies involved in actuarial reviews of the CPP, it is desirable for the Chief Actuary to seek out the advice and guidance of experts, including actuaries, demographers and economists, to help ensure that a wide range of analysis and opinion is considered and to enhance the credibility of the actuarial reviews.

To this end, the Office of the Chief Actuary hosted the CPP Seminar on Demographic, Economic and Investment Perspectives for Canada, Years 2018 to 2050 on 28 September 2018. Representatives of the OCA also attended a seminar hosted by Retraite Québec on 1 November 2018. Participation at these events helped the OCA to formulate best-estimate assumptions and methodologies for AR30.

After the tabling of each of the last seven triennial actuarial reports on the CPP (AR17, AR18, AR21, AR23, AR25, AR26 and AR27), OSFI engaged a panel of three independent actuaries to conduct a post-release review of the actuarial reports, similar to the review described in this report. All of these actuarial review panel reports have included a number of recommendations for improvements in, or revised approaches to, the processes, sources of data, methodologies and assumptions utilized in preparing actuarial reports on the CPP. This process provides a level of assurance to the public and also helps the Chief Actuary in gathering a range of views regarding the complex methodologies and assumptions involved.

The Chief Actuary has developed rigorous processes for the selection of assumptions. All decisions on assumptions are made in consultation with her internal staff. She draws upon the expertise of officials from other government departments and agencies who participate in interdisciplinary seminars, maintains effective two-way communication with the CPPIB, and devotes a considerable amount of time to keeping abreast of experts’ views on demographic and economic matters. She also considers the comments and advice contained in the reports of the actuarial review panels that reviewed the previous actuarial reports.

8.3 Recommendations

Recommendation 8

We recommend that the Chief Actuary continue to analyze the incremental investment expenses incurred over time to implement the CPPIB’s active management strategy in order to assess the extent to which added value to the overall fund return may be expected to be consistently and reliably achieved over the long term.

Recommendation 9

We recommend that the Chief Actuary continue to maintain effective two-way communication with the CPPIB and Statistics Canada, with the goal of achieving continual improvements in the process of setting best-estimate assumptions and range of variability.

Appendix

Responses of the Office of the Chief Actuary to the recommendations of the AR27 independent review panel

This appendix outlines the responses provided by the OCA to the recommendations formulated in the 27^th CPP Actuarial Report.