Pension Reform: It’s not rocket science
In a nutshell
With the latest re-vectoring of the federal government’s ideas on pension reform (PRPP), we compare it to some of the more meritorious proposals tabled for pension reform (or already operational in the Saskatchewan SPP). PRPP comes up wanting in comparison to the others based on the little that’s known about it and much which is not included/mentioned/defined in the proposal. Each proposal must be tested whether it meets the minimum requirements that real pension reform must address:
- Encourage or require adequate retirement savings rate of Canadians
- Provide low-cost vehicle for professional management of retirement assets (without constraint of CPP cap and accept transfers from existing wound-up DB, DC or RRSP plans)
- Provide low-cost longevity insurance options to deal with longevity risk (provide lifetime income, as well as low-cost decumulation strategies for un-annuitized assets
- Secure existing commitments to already earned private sector DB pension plan benefits: by strengthening regulatory framework to prevent plan underfunding and modify BIA/CCAA to increase priority of pension plan shortfalls in case of sponsor bankruptcy.
Requirements #1 and #2 can be met by individuals with a little saving discipline (‘just do it’) and a little financial education on how to create low-cost ETF based portfolios consistent with their risk tolerance implemented in a self-administered mode with an online discount broker. But more realistically, management of retirement assets could be sourced from a government created but independent large-scale non-profit CPPIB/OMERS/Teachers-like organizations or mutual rather than stockholder owned asset management companies (e.g. like Vanguard in the U.S.).
The table below looks at the various proposals: AFL (Alberta Federation of Labour), (broader than AFL) expanded CPP, Keith Ambachtsheer’s CSPP, Liberal Party White Paper recommendation of SCPP and the already operating Saskatchewan SPP.
Administration (especially payroll deduction, record-keeping, etc) for the CSPP, SCPP, expanded-CPP is assumed to be built upon existing CPP infrastructure so essentially minimal or no incremental cost would be incurred (in fact the additional assets would reduce the 0.59%/yr CPP admin cost).
There are no mechanisms to meet requirement #3 today as there are no low cost longevity insurance products (annuities or pure longevity insurance). Alternatively, it could be sourced from a government created but independent non-profit national insurer or mutual rather than stockholder owned company (i.e. a policyholder-owned mutual insurance company, there are still some in existence in the U.S. and Canada; a good example is New York Life which operates in both countries already). A pure longevity insurance program where individual longevity risk is protected however societal longevity risk is born by society as a whole (i.e. no risk to taxpayer and cross-generational subsidies.) All cost would be fully funded (no tax-payer risk) and there are no cross-generational subsidies.
Depending on one’s income requirements in retirement relative to available assets, a systematic withdrawal plan may meet the requirements of many who can live with the volatility of such a plan and/or wish to leave an estate and/or accumulate asset reserves for planned large expenditures or emergencies. Those who are prepared to accept a lower standard of living to secure lower income and volatility would pretty much have to accept a traditional immediate annuity (ideally an inflation indexed one; but there are few inflation indexed annuity sources available in Canada.)
Legend: Accumulation/Savings Mechanisms
Source (Who pays?): I=individual, E=Employer
Admin (Collection of savings): PD=Payroll Deduction., CA=Central Admin
Compulsion: V=Voluntary, A-E= Auto-Enroll2
Contribution Limit: N=No cap (over RRSP/TFSA limit to taxable acct), Y=Yes
Investment Management (No incremental admin cost if built on CPP infrastructure)
Cost: L= Low (<0.3%), M=Medium (<0.5%), H=High (<0.7%)
Alloc (Asset Allocation Flexibility): G=given, S=selectable, D=Default
LI (Longevity Insurance): Y=Yes (A=Annuity or LI=Pure LI), N=No
SWP (Systematic Withdrawal Mech.): Y=Yes, N=No, S=Selectable
There is clear consensus that we should build upon the CPP administrative infrastructure, so as to minimize incremental administration costs.
There is value to the AFL proposal in that (after a 40 year transition) it offers a potential doubling of the current maximum level of CPP benefits to about $23,400 in 2010 dollars of annual lifetime income for an additional contribution of 6% of YMPE split evenly between employer and employee. This could be a part of the overall solution, but is not the whole solution.
The SCPP (the Liberal Party White Paper) and CSPP (Ambachtsheer) proposals are both very thoughtful and comprehensive.
The attractiveness of the SCPP proposal is that it also addresses the problems with the existing private sector DB pensions with a discussion of required regulatory changes and protection of already earned DB pensions in case of bankruptcy. However assets are locked in until retirement, annuitization is compulsory upon retirement, assets are explicitly targeted to be managed by the private sector (with no explicit cap on investment management fees). While I think this proposal has great potential, I would feel better with a little less compulsion, for example adding the AFL proposal as a base, and allowing opt-out clauses on contributions above the AFL requirements (e.g. annuitization, lock-in)
In Keith Ambachtsheer’s CSPP proposal (which he calls “paternalistic libertarian”) everything is ultimately voluntary, though defaults are set up to encourage ‘good’ behavior and you must explicitly opt-out from these defaults (e.g. auto-enroll, start 50% target auto-annuitization at age 45). Some will argue that a little more compulsion would be desirable (this is where one might want to blend in the AFL proposal as a base.)
I recommend some additional features that I believe would further improve both the SCPP and CSPP proposals: (1) pure longevity insurance option would be a far cheaper and more effective means than annuities for implementing individual longevity insurance needs of Canadians (a single premium of about 7% of assets at age 65 is used to purchase a life contingent lifetime annual payout of the same order of magnitude as the single-premium) and starting upon retirement a systematic withdrawal plan from a balanced portfolio defaulting to 4%/year draw based on the previous year-end asset value (this guarantees that one never runs out of assets and given that the longevity insurance lifetime income kicks in at 85, there would be a high probability that an estate and/or emergency funds will be available as desired or needed; the 4% default can be over-ridden by the individual) and (3) possibly even a self-managed option for the assets analogous to the self-managed RRSP where one could implement a very low cost ETF based diversified portfolio).
There may even be value to a dual approach, i.e. a combination of an expanded-CPP and DC plan riding on the same administrative platform and having some of the best features of the SCPP and CSPP.
P.S. You might also be interested in looking at my Pension Reform proposals that I presented in November 2007 to the Ontario Expert Commission on Pensions
1 In addition to $2,500/yr max, there an option to transfer $10,000/yr from an RRSP
2Auto-Enroll (A_E) means that both individual/employer are automatically enrolled at some specified level, when either or both may opt-out is indicated by A-E/OO
3 There is reason to expect lower SPP investment management cost once it’s scaled up from $260M with funds from other individuals from across Canada and the new higher limit effective January 1, 2011
4For-profit private sector managed investments in Canada historically came with high fees >>1%
5Keith Ambachtsheer’s CSPP (Canada Supplementary Pension Plan) is a comprehensive proposal for a supplementary pension plan which includes: (1) mechanisms for (pre-retirement) accumulation and (post-retirement) decumulation resulting in “affordable post-work lifetime payment streams”, (2) (potentially) universal coverage and portability, (3) “pension delivery institutions that are transparent and cost-effective and operate solely in the best interest of the people they are meant to serve”. He suggests a 60% income replacement rate including CPP/OAS/GIS, 10% contribution rate from $30K to $110K (with contributions going to TFSA for those with $30-45K incomes and RRSPs for those with >$45K incomes. CPPIB-like investment management costs of <0.3%, auto-enrol at the default level with opt-out option or option to opt-in for additional contributions, investment option could be default or more conservative, default set to 50% annuitization staring at age 45 but again with opt-out option. The proposals also include an RRSP transfer option, but there is no mention on how to deal with current private sector pension crisis.) There appears to be no requirement to lock-in accumulated funds until retirement as with the other proposals. CSPP proposes to use the existing CPP administrative infrastructure, so it can be assumed that minimal additional administrative costs would be incurred, just as in an expanded-CPP.
6Liberal Party White Paper recommendations, the SCPP is a portable target-benefit plan, which is extremely well thought out with a lot of details specified. It addresses all important elements including enhanced protection for existing private sector DB plans by strengthening the regulatory framework as well as proposing CCAA/BIA changes to increase priority of pension plan shortfalls to at least preferred unsecured level, a mechanism to deal with stranded pension assets when a plan is wound up, re-evaluation of appropriateness of how CPP CPI adjustments are calculated, as well as even a look at increasing the CPP YMPE. They also address the need for better saving mechanisms for non-working and self-employed Canadians. The plan has a default contribution rate set at 9% for employee and 9% for employer (opt-out is permitted for either) with accumulated assets locked in until retirement. The White Paper recommendations extend to CPP improvements, such as allowing some retroactive payments for missed benefits. The key concern in my mind is that the funds would be managed by tapping “private sector expertise” without and explicit stated cap on the cost associated with the fund management (say 0.3% max); other options could be fund management under CPPIB or similar management organizations like OMERS/Teachers or creating Vanguard like investment management corporation which is mutual rather than stockholder owned. Additionally some flexibility should be considered at the decumulation phase by allowing partial annuitization or pure Longevity Insurance option with the balance of the assets continue to be invested with provision of a Systematic Withdrawal option (say annual 4- 5% of remaining balance). SCPP proposes to use the existing CPP administrative infrastructure, so it can be assumed that minimal additional administrative costs would be incurs as in an expanded-CPP.
7 SPP (Saskatchewan Pension Plan (2009 Annual Report, Investments and Statement of Investment Policies and Goals.) is open to all Canadians and it has recently increased the allowable maximum annual contribution to $2,500; it also permits a $10,000 annual transfer from an RRSP (so I suspect you could in fact effectively contribute an annual maximum of $12,500 which would allow a person earning <$70,000 to contribute the full tax deferrable 18%/yr). Administration costs appear to be about 0.75%/yr for the $268M fund (not bad compared to $118B CPP’s 0.59%/yr). Investment management costs are targeted to be <1%, this is an additional cost). The SPP’s four year rolling return is 2.3% or 0.2% higher than their benchmark beforeexpenses, so it appears to underperform a passive implementation.
9 AFL (Alberta Federation of Labour) proposal is essentially an ‘expanded CPP’ system. The difference is that the AFL has a very specific objective of doubling from 25% to 50% of the CPP’s YMPE (currently about $47,000), whereas the ‘expanded CPP’ proposals tend to explore more broadly an increase in both the current 25% level and the YMPE. The paper suggests that 6% additional CPP contribution over 40 years would be sufficient to achieve this (this is much lower than the current 9.9% rate for the first 25%, which actually allows the dropout of the lowest 7-8 years and includes some make-up payment for underfunding that accumulated by the 1990s), premiums would be phased in over 7 years, while full benefits would take 40 years to achieve. They also discuss a supplemental plan based on an expanded CPP where there is also a doubling of the YMPE. (Kesselman looks at various flavours of pension reform built upon the CPP framework) According to Greg Hurstthe CPP costs are not that low: operating expenses 0.59% ($41/member), investment management fees 0.32% and investment related transaction costs 0.08%; but building on top CPP infrastructure you could assume that operating costs would not increase from current (absolute levels) while investment management costs for incremental assets would continue to be of the order of 0.32% as well. Some of the issues with the CPP or exp-CPP or AFL/CPP options are that: (1) there are no mechanisms mentioned to transfer existing but wound-up DB plans, existing DC plans or RRSPs, (2) CPP plan is effectively an annuity and does not provide mechanism to accumulate assets for emergencies, major expenditures, saving for an estate or savings in excess of that permitted in the CPP-based plan. What this has going for it, in addition to raising the CPP payout level to a more realistic level, is that the current administrative costs could likely support an expanded CPP, so only the incremental investment costs would have to be borne.
8In order to deal with the needs of those near or already in retirement, it is necessary to permit (the gradual) transfer of existing retirement savings to the new low cost investment and decumulation vehicle. The taxpayer bears none of the costs (these are charged against the investments) and assumes no risk (market risk, cohort longevity risk and inflation risk are borne by investors except to the extent that they are managed by the investment managers) under the ‘required’ plan scenario.