December 1, 2018

The 'Canadian Pension Model' Under Attack: Can It Survive And Prosper?

“Canada’s public sector DB plans frequently attribute their success to the ’Canadian Pension Model’. A recent World Bank study identifies superior governance, economies of scale, innovative investment   practices, responsible funding, and visionary leadership as important features of the Model….. Without disputing the virtues of the Model, we attribute the success of Canada’s public sector DB plans to large public subsidies made possible by practices which are neither admirable nor virtuous….”.

“Risk and Reward in Public Sector Pension Plans: A Taxpayer’s Perspective”Malcolm Hamilton and Phillip CrossFRASER INSTITUTE, December 2018


The Three Key Messages of this Letter

  1. The practices Hamilton and Cross refer to in their paper are indeed “neither admirable nor virtuous”, but are not caused by the ‘Canadian Pension Model’.
  2. Indeed, the Ontario Teachers’ Pension Plan story shows that complete implementation of the ‘Canada Pension Model’ is the best antidote to these undesirable practices.
  3. The ‘Canadian Pension Model’ would be a powerful instrument for producing fair, adequate pensions around the world at an affordable cost if it was globally adopted.

Setting the Valuation Cat among the Pension Pigeons

Authors Hamilton and Cross spare no words or numbers to justify their assertion that the success of  Canada’s public sector DB plans is due to “practices which are neither admirable nor virtuous”. Three logic steps lead them to their conclusion:

  1. Be clear about where pension funding comes from: It is meaningless to assert that to fund the typical public sector plan benefit, only 20% comes from the original contributions and 80% from the investment return on those contributions. A far more relevant calculation is that at a risk-free 1% real rate of return a typical public sector pension benefit costs about 40% of pay. However, if a typical amount of investment risk a typical plan takes on produces a typical 2.5% risk premium above the 1% real risk-free rate, the projected investment return of 3.5% would drop the required contribution rate from 40% to 20% of pay. This halving of the cost of the pension is not due to investment genius, but due to a willingness to bear investment risk and the power of return compounding.i
  2. Be clear about how plan funding is translated into employee compensation: if contributions were invested in risk-free, inflation-linked bonds yielding 1%, a 50-50 split in the required 40% of pay contribution rate results in 20% contribution rates each for the employer and for the employees. However, that is not what happens in practice. In practice, the 20% of pay contribution rate calculated based on the 3.5% real return assumption is deemed to be sufficient. The 50-50 split now reduces to 10% each for the employer and for the employees.  This practice skips over some obvious questions: what about the 20% of pay cost component (i.e., 40%-20%) to be earned through investment risk taking? Is that not the cost of underwriting the risk that the assumed 2.5% risk premium will in fact not be achieved? Who underwrites this risk? If it is the employer, its funding cost is not 10% of pay, but 10%+20%=30% of pay. Public sector employers underwriting the investment risk in their DB plans who do not recognize this 20% of pay risk underwriting cost in their bargaining processes, or in their reporting of the total cost of employee compensation, do their current and future tax-paying constituents a serious disservice. They are effectively giving that 20% of pay away for free.
  3. Be clear about who bears which risks in any specific pension plan design: to inject greater clarity into this often-fuzzy topic, the authors posit three plan designs: the Traditional DB Plan (TDBP) where the employers bear all risks, the Jointly-Sponsored Pension Plan (JSPP) where all risks are split 50-50, and the Target Benefit Plan (TBP) where the employer commits to a defined % of pay contribution rate, but where plan participants bear all the investment and other actuarial risks. Applying these definitions to the typical 40% of pay public sector pension plan assumed above, the employer-employee cost split in % of pay terms are 30%-10% for the TDBP, 20%-20% for the JSPP, and 10%-30% for the TBP. Hamilton and Cross note that without a plan design that is clear about how plan risks are to be borne, a never-ending bargaining  situation arises and full plan costs are likely to be under-reported. The objectives of the bargaining agents for plan members are clear: good pensions at the lowest possible cost to plan members, which ideally leads to the 30%-10% split TDBP. The objectives of the bargaining agents for tax payers should be to attract and retain competent employees at an affordable total compensations cost, which leads ideally to the 10%-30% split TBP.                  

Within this framing, Hamilton and Cross assert that the agents for plan members have been winning the bargaining game in Canada, with some 25% of public sector workers members of TDBPs, some 75% members of JSPPs, and 0% members of TBPs. The full costs of the TDBP and JSPP plans are indeed being under-reported. Why this imbalance? Hamilton and Cross suggest two reasons: 1. Public sector employers are not legally required to report the pension benefit guarantees embedded in TDBPs and JSPPs as pension costs, and hence part of total employment costs, and 2. The employer bargaining agents for tax payers are typically members of TDBPs or JSPPs themselves, and therefore cannot represent tax payers’ interests in a non-conflicted, unbiased manner.    

Relevance of the ‘Canadian Pension Model’

What do Hamilton and Cross’ analysis and conclusions have to do with the ‘Canadian Pension Model’? The answer requires a brief recount of the Model’s origins. The Government of Ontario launched a Task Force on Public Sector Pension Funds in September 1986. The Task Force produced its Report titled “In Whose Interest?” in November 1987. Its key recommendations included:

  • Pension plan designs should be changed so that taxpayers’ interests are better represented.
  • The cost of pension benefits should be assessed in a total compensation context.
  • Pension assets should be invested through separate arms-length pension organizations.
  • Government should not interfere in pension fund investment decision-making processes in any way.

These recommendations led to the creation of the Ontario Teachers’ Pension Plan (OTPP) in 1991. In line with the Task Force’s recommendations, the Plan design was eventually set up as a 50-50 JSPP ‘partnership’ between the Ontario Teachers’ Federation and the Ontario Government (the ‘Partners’). The Plan was to be administered by an independent, arms-length organization led by a professional Board selected and approved by the Partners. The OTPP Partners would negotiate the ‘pension deal’, the OTPP organization would implement it.  

Fast forwarding 27 years to today, OTPP is generally regarded as the leading example of the ‘Canadian Pension Model’ in action. Three key facts from OTPP’s 2017 Annual Report support this perception:

  • OTPP has achieved top global ‘value-for-money’ rankings in both investments and benefit administration as measured by an independent benchmarking organization.
  • OTPP has an estimated $10B asset surplus assuming a 2.75% net real liability discount rate. This surplus will not be spent, but kept as a ‘rainy day’ reserve.
  • OTPP has moved away from the standard 50-50 JSPP pension plan design model towards the TBP model by adding conditional inflation protection (CIP) on future service to its arsenal of long-term plan sustainability levers. This CIP risk lever, which potentially reduces real benefits paid to retired as well as active teachers if the Plan is underfunded, falls directly on plan members. 

These reported facts from the 2017 Report do not tell the whole OTPP success story. Its professional Board has played, and continues to play a critical role in achieving funding fairness:

  • The Board rather than the two OTPP Partners is responsible for setting the liability discount rate. Given its independence from the Partners, it is no accident that OTPP’s 2.75% real rate is the most conservative among Canadian TDBP and JSPP pension plans.
  • The Board facilitated the adoption of CIP by the Partners as an additional funding policy lever after the Global Financial Crisis (GFC). Its implementation helped restore OTPP’s surplus funded position after the GFC. 

Within this context, the question “what do Hamilton and Cross’ analysis and conclusions have to do with the ‘Canadian Pension Model’?” can be addressed. In theory, an independent, professional Board of the pension management organization should be a key success component of the Model. The Board has a fiduciary duty to steer the pension deal bargaining agents in the dual directions of prudence and fairness. The OTPP case shows that good theory can make for good practice. Hamilton and Cross are right to point out that this arms-length discipline disappears when the bargaining agents and the pension organization Board are the same people. That is still the case in too many public sector pension plan situations in   Canada, and for that matter, in other countries such as the UK and the USA as well.

How the ‘Canadian Model’ Can Power Pensions around the World

To repeat, three important success drivers of the ‘Canadian Pension Model’ are 1. A clear mission, 2. Arms-length professional governance, and 3. No ‘agency’ constraints to achieving the mission. Our October 2018 Letter titled “Saving Retail Retirement Savers: What It Will Take” showed that relative to leaving people to their own retirement saving and investment devices, a well-managed professional collective approach can easily reduce the cost of a target pension by 50%. A just-published study titled “The Value of a Good Pension” estimates even greater cost savings.ii Why these dramatic improvements in turning retirement savings into lifetime pensions? They include automatic savings, significant fee reductions, higher risk-adjusted investment returns, and the pooling of longevity risk.   

In a different framing, our August 2017 Letter titled “The ‘Canada Model’ for Pension Fund Management: Past, Present, and Future” used the CEM Benchmarking Inc. databases to contrast the 10yr investment performance of eight ‘Canada Model’ pension funds relative to that of 132 other pension funds in the database. The eight Canada funds outperformed the other 132 funds by an average 0.5%/yr. on a risk-adjusted basis. Two important ‘Canada Model’ performance drivers were 1. Insourcing of the investment management function by hiring teams of highly-qualified investment professionals, thus materially reducing the cost of investing, and 2. Materially greater exposure to private real estate, infrastructure, and equity investment markets. The result is direct, intense engagement by ‘Canada Model’ funds in long-term sustainable wealth-creation processes not available to funds without these capabilities. Logically, these competitive advantages should translate into higher long-term risk-adjusted returns, and thus far that has been the case.

In Conclusion

The opening lines of this Letter promised that it would convey three messages: 

  1. The practices Hamilton and Cross refer to in their paper are indeed “neither admirable nor virtuous”, but not caused by the ‘Canadian Pension Model’.
  2. Indeed, the OTPP story suggests the complete implementation of the ‘Canadian Pension Model’ is the best antidote to these undesirable practices.
  3. The ‘Canadian Pension Model’ would be a powerful instrument for producing fair, adequate pensions around the world at an affordable cost if it was globally adopted.


Keith Ambachtsheer    



  1. There continues to be considerable debate and controversy about this approach to pricing investment risk insurance. My most recent contribution to the debate was the September Letter titled “Short-Term vs. Long-Term Investment Risk: Really Understanding the Difference”. I show that post-WWII, the S&P500 stock index was indeed risky relative to a 10yr Treasury Bond portfolio for 3yr holding periods, but not for 30yr holding periods. In the latter  context, stocks outperformed bonds in all 30yr periods tested. While this does not guarantee that will be the case for the next 30yr holding period, it does suggest the price of risk may be set by short-term investors. In that case, plausibly, long-term investors are being paid for risks not relevant to them. 
  2. Published in November 2018 by Healthcare of Ontario Pension Plan (HOOPP), Ryerson University’s National Institute of Aging (NIA), and the CommonWealth pension management organization. The study can be accessed HERE.


KPA Advisory Services is pleased to share this edition of The Ambachtsheer Letter with all readers; if you wish to become a KPA Advisory Client/gain access to ALL Letters, please see the Services page on our website.

The information herein has been obtained from sources which we believe to be reliable, but do not guarantee its accuracy or completeness

Advisory Service clients have access to full issues of the Ambachtsheer Letter.

Become an Advisory Service Client
or Login
Back to Top