9.1.1 Why innovation is important and so difficult
How can our system of voluntary pension plans be sustained and enlarged while protecting the security of members’ entitlements and affordability for sponsors? In this chapter I argue that the achievement of these seemingly contradictory objectives is indeed possible — but only if we can move beyond conventional understanding of defined benefit (DB) and defined contribution (DC) plans, and of why they are able to provide the attractive outcomes for which they are rightly prized. Doing so will enable us to think about innovative pension designs, which constitute a blend of DC and DB models, refinements or adaptations of the DB model, or the introduction of new types of plans that are its functional equivalent. In my view, the introduction of innovative plan designs based on new ways of thinking represents the best hope for reinvigorating our DB system.
If innovation is so critical, why is it so difficult to achieve? Broadly speaking:
- When the pension system is working well, no one sees the need for innovation; and when it is not, people are too pessimistic to take chances.
- In a complex, highly regulated field such as pensions, where arrangements involve the fortunes and lives of many people over many decades, a burden of persuasion rightly rests on proponents of change to demonstrate that the “new” will constitute an improvement over the “old.”
- The Pension Benefits Act (PBA) and the Income Tax Act (ITA) lack specific mechanisms to facilitate the introduction of new pension designs; moreover, changing either statute requires political will and a priority position on the legislative agenda, both of which are hard to come by — particularly at the same time.
- Technical concepts — such as “defined benefit” or “defined contribution” — have a powerful hold on pension professionals and policy makers. To describe or discuss innovative plan designs, it is necessary either to re-define these concepts or invent a new (and perhaps clumsy) conceptual vocabulary.
- The intellectual investment of regulators, sponsors, unions, pension professionals and service providers in the status quo sometimes causes them to over-value existing arrangements and to mistrust innovation.
These last two barriers to innovation warrant special attention.
9.1.2 Re-examining the defined benefit/defined contribution distinction
Debates on pension issues often juxtapose “classical” or “pure” DC plan designs with “classical” or “pure” DB designs. Differences between the two are canvassed extensively in Chapters One and Two, and can be briefly summarized as follows:
- In DC plans, the investment risk is borne by individual plan members; in DB plans, it is borne by the employer and/or spread across the plan membership (the point is hotly debated in the context of surplus ownership and elsewhere) — and, it varies as among single-employer pension plans (SEPPs), multi-employer pension plans (MEPPs) and jointly sponsored pension plans (JSPPs).
- In DC plans, the longevity risk is generally borne by individual plan members (although in principle, individual members might purchase a life annuity upon retirement); in DB plans, it is spread across the entire present and future membership of the plan.
- In DC plans, where contributions are made according to a set formula, regulatory oversight is reasonably light; in DB plans, where funding is on the basis of complex valuations based on multiple, volatile factors, oversight is intense and ongoing.
- In DC plans, the clear, segregated individual account format leaves little room for insolvency risk; in DB plans, members — particularly of SEPPs — do bear the risk that a sponsor may fail with an under-funded plan.
- In DC plans, if members leave, their entitlements can be easily ascertained, monetized and transferred to another plan or a locked-in account; in DB plans, individual entitlements are difficult to disentangle and not easily transferable.
- In DC plans, contribution formulae may be adjusted from time to time, or contributions may be increased automatically as wages rise, but the member’s entitlement is the fixed capital accumulation on retirement; in DB plans, pensions adjust automatically prior to retirement as salaries and length of service increase and are sometimes inflation-adjusted after retirement as well and supplemented by ancillary benefits.
- In DC plans, members end up with a range of options around investment decisions, including making their own; in DB plans these are made by professional advisors. (This point is not trivial. A recent U.S. study showed that from 1995 to 2006, the investment performance of DB plans, on average, exceeded that of DC plans by 1% a year; over the 11 years under study, the cumulative effect was a 14% advantage in favour of DB plans. Further investment advantages in favour of DB plans are described below.)
When their characteristics are starkly contrasted in this fashion, DC and DB plans appear to be so different from each other that a binary distinction between them is inevitable: a plan is either DC or DB, but it cannot be both. However, this is patently not the case. Many plan designs have emerged that are, strictly speaking, neither “pure” DC nor “pure” DB. They may be hybrids containing layered elements of both DC and DB plans, or sufficiently distinct that they might be considered a species in themselves.
In my view, setting aside the notion of “pure” or “classic” DC and DB plans is the first step toward re-imagining the pension system as one that includes a wide range of innovative designs with characteristics of both or neither. The second step is to revisit DB plans to see what makes them so attractive.
The key advantage of DB over DC plans is the greater capacity of the former to insulate individual workers from risk. In DB plans, as noted, the investment risk is, to some extent, transferred from the members to the sponsor, while the longevity risk is spread across the entire present and future membership of the plan — the larger the membership, the more efficiently managed the risk. However, DC plans do not offer the opportunity for either form of risk management. Individual members manage their own accounts on retirement. If they manage badly, encounter adverse market conditions or live longer than they anticipate, they will find themselves without sufficient pension income and will have to depend on their Canada/Quebec Pension Plan (C/QPP) and Old Age Security (OAS) public pensions. The result will likely be a reduction in their standard of living. DB plan members do not have to deal with these contingencies on their own. Barring the catastrophic failure of their pension plan — a relatively rare event against which they may be partly buffered by the Pension Benefits Guarantee Fund (PBGF) — they will receive the pension they expect, and it will last them as long as they live. Risk-spreading makes the difference, and size allows risk-spreading and other efficiencies to work better.
This analysis seems to lead to the following conclusion: the way ahead for occupational pensions in Ontario lies with large innovative plans rather than with relatively small, conventionally structured DB and DC plans. However, this is not to say that large innovative plans represent a universal “cure” for the DB system:
- Although innovative plans will approximate DB plans in some important respects, they will not replicate them in every particular.
- Innovative plans will therefore not forcibly displace “classic” DB plans; a shift from the latter to the former will occur only if sponsors and the unions with which they negotiate their pension plans are persuaded that an innovative alternative provides them with similar or superior outcomes at lower cost.
- Innovative plans are not inevitably large plans; some small- and medium-sized sponsors will wish to operate their own plans rather than merging them with a larger plan because they believe that doing so will better enable them to use pensions as a tool to recruit and retain key workers.
- Innovative plans should not be entitled to favourable regulatory treatment just because they are innovative; they must be shown to manage risks in an acceptable way.
- Innovative plans will not appeal to employers who have no interest in providing pension coverage for their workers; however, they should appeal to employers who are considering or are being pressed to initiate a pension plan but are hesitating because of the well-known criticisms of DB plans.
Nonetheless, if innovative designs can arrest the decline in pension coverage — as I think they may — they will make a significant contribution to Ontario’s social and economic policy. And if they can actually reverse the decline so that a higher proportion of workers gains access to pensions over the next 10 or 20 years, that contribution will be ranked as very valuable indeed.
9.2 Promoting Innovative Plan Design
Many innovative models already exist in Ontario and in other jurisdictions that either combine elements of DC and DB designs, or constitute a significant departure from them. A few examples make the point.
One hybrid plan model (popular, for example, among universities) combines DB and DC models by offering a DC pension, but with a minimum DB floor of protection. Thus, the advantages of both designs may be achieved to some extent. Other combinations are possible, subject always to existing legislative constraints.
Cash balance plans
Cash balance plans are, in essence, DC plans. However, the plan sponsor guarantees a defined rate of investment return on members’ accounts, often fixed by reference to an outside index (AAA Corporate Bond rates, for example) and in excess of what members could expect to receive by themselves investing in a conservative portfolio. The effect of this guarantee is to relieve individual members of the expense of professional investment management and to shift part of the investment risk to the plan sponsor. To this extent, cash balance plans become the functional equivalent of DB plans.
However, these plans do not qualify for registration under Canada’s tax laws, effectively preventing their operation in Ontario. Moreover, though common in the United States, cash balance plans have been controversial because the sponsor retains all investment gains in excess of the promised minimum. For present purposes, I merely cite them — without endorsing them — as another example of innovation that is perhaps worth investigating.
Contingent benefit plans
Contingent benefit plans — in contrast to cash balance plans — are structured to return investment gains, if there are any, to the beneficiaries. They are, in effect, DB plans that specify that certain benefits will be paid only if the plan funding is healthy.
For example, several Ontario pension plans now pay full indexation of retirement benefits only if the plan can afford these payment increases; otherwise — pending replenishment of the fund — partial or no indexation is provided. Or to take another example, pension plans in the Netherlands are career average plans in which the wage base is adjusted upward if the plan funding is adequate. The effect is that over time, and so long as funding will support this gradual shift, they may come to resemble final average plans. These examples suggest other possible experiments with the contingent benefit formula. For example, one might design a pension plan under which 90% of the promised benefits are absolutely guaranteed, but the other 10% are paid only if the plan is healthy.
Target benefit plans
In a target benefit plan, contributions are fixed, and on the basis of a best estimate of what the funding will provide, benefits are promised. However, if it is later determined that the “target” benefit cannot be achieved with the available resources — contributions and the return on investment — both accrued and future benefits can be adjusted downward and, for that matter, upward, if and when the plan’s fortunes recover. Thus, from the viewpoint of the member, the plan may be perceived as a DB plan. Indeed, the benefit is defined and has all the characteristics of a DB plan except one: it is contingent on the plan’s success. However, from the viewpoint of the plan sponsor or sponsors, the plan functions like a DC plan. Contributions are fixed or defined: once those contributions have been made, the sponsor has no obligation to make good any notional deficiency, because the benefits will be adjusted rather than paid in full.
Target benefits may presently be provided only by multi-employer plans, subject to two limited exceptions. First, DB single-employer JSPPs are precluded under existing legislation from offering target or variable benefits while ongoing, but may reduce benefits on wind-up if the plan is under-funded. Second, collectively bargained DB SEPPs, while still ongoing, may provide for fixed or defined contributions and for the reduction of benefits in the event that a funding deficiency results. However, unlike JSPPs, this particular DB model does not provide for unions to participate in plan governance, is therefore apparently unattractive to them, and is in fact rarely used.
In my view, it ought to be possible to offer target benefits in a single employer plan under certain circumstances. The jointly governed target benefit pension plan (JGTBPP), proposed in Chapter Eight, may well appeal to some prospective sponsors, unions and workers — especially when the alternative is no pension coverage at all.
Given the risks associated with self-administered SEPPs (discussed in Chapters Four and Eight), and the limited appeal of the present provisions for reducing benefits in collectively bargained plans (discussed above), I have proposed that the new JGTBPP should be available where workers both belong to a trade union or equivalent organization, and are represented on the plan’s governing body. These interlocking requirements should ensure that beneficiary interests will be considered in fixing the levels of employer contributions and target benefits, lend economic force to workers’ demands that the sponsor/employer fund the plan at the proper level, and ensure that member representatives will have an effective voice in the plan’s governance process.
Member-funded pension plans
Member-funded pension plans (MFPPs) — a new DB plan design developed in Quebec — contemplates that, as in a DC plan or an Ontario-style MEPP, the sponsor’s obligation will be limited to the contribution of a fixed amount. Whatever additional funds are required to pay the promised benefits will be contributed by the members who thus collectively assume the financial risk. Given the limited capacity of workers to do so, however, these plans are to be subject to very strict funding rules that effectively require them to be fully funded at all times. MFPP governance structures are carefully prescribed and they must be embedded in a collective bargaining relationship.
There are significant differences in detail and degree among these existing and proposed innovative, union-negotiated, member-governed target benefit and/or multi-employer plans, and within each plan type. However, they share certain common characteristics: they depart in some significant way from the model of “pure” DB plans; they deliver somewhat similar but not identical benefits; and they are generally big enough to spread risks widely, achieve efficiencies of scale and to exercise leverage in investment markets. Although each may have some shortcomings, such plans appear collectively to represent some part of the future of the Ontario pension system: they already account for 60% of all plan members and 70% of all DB plan members — percentages that should grow if my recommendation to initiate JGTBPPs is accepted.
Recommendation 9-1 — Innovation in plan designs should be promoted and facilitated by the proposed Pension Champion (see Recommendation 10-5).
Legislation and regulations that inhibit innovation should be reviewed and revised with a view to enabling the introduction of plan designs not now permitted, if such designs meet criteria to be set out in the statute. If the federal Income Tax Act represents an obstacle to otherwise desirable innovation, the Ontario government should approach the federal Minister of Finance with a view to removing such obstacles.
9.3 Promoting Larger Plans
As noted above, spreading certain risks across large populations results in more predictable outcomes and less volatility. Examples of these variables include life expectancy and age at retirement. This size advantage is compounded along almost every vector of plan success. For example, large plans pay far lower fees on their investments than small ones, as Table 1 illustrates.
|Size Of Pension Fund||Investment Fees For Large-Cap Equities|
|Individual account||250–300 basis points|
|$10 million||60 basis points|
|$1 billion||42 basis points|
|$10 billion||28–35 basis points|
Table 2 tracks the impact of investment expense ratios and shows how profoundly they can affect the aggregate pension benefits and working income replacement rates of retired plan members. The example used assumes an annual contribution to a plan of $10,000 over 40 years by or on behalf of a worker making $50,000 per year.
|Effective Expense Ratio||0%||0.4%||1.5%||3%||5%|
|Annual contributions (over 40 years)||$10,000||$10,000||$10,000||$10,000||$10,000|
|Accumulated value (after 40 years)||$777,000||$707,000||$551,000||$400,000||$272,000|
|Annual pension payment||$45,000||$41,000||$32,000||$23,000||$16,000|
|Working income replacement rate||90%||82%||64%||46%||32%|
However, lower investment fees are but one of the many advantages enjoyed by large plans over smaller ones and over individual savers. In terms of income generation, large plans are in a position to hire expert staff to initiate and execute their investment strategies, to make attractive private placements of their investment funds, and to spread the investment risk by acquiring a wider range of investment vehicles. In terms of administrative expense, large plans are able to reduce their unit costs of administration by spreading them across a large plan membership, and they are typically able to offer members enhanced levels of information, education and service. Finally, large plans are more likely to survive than smaller ones, if only because the enterprises (or groups of enterprises) that sponsor them are likely to be more stable or resilient than those that sponsor small plans.
The cumulative effect of all of these advantages is extremely significant. It is so significant, in fact, that plan size may be a greater determinant of a member’s pension than plan design. Or, to make a more modest claim — holding plan design constant — large plans will generally perform better than small ones.
Keith Ambachtsheer, in his recent C. D. Howe paper, suggests that assets of $10 billion are necessary to claim the economies of scale and leverage in financial markets. In fact, most JSPPs and some very large DB SEPPs have already achieved that critical size, and a number of MEPPs have come close. On the other hand, most DC plans and smaller SEPPs (especially non-union SEPPs, which account for about 17% of all plan enrolments) have not. Any strategy for preserving and enhancing our pension system ought therefore to place special emphasis on assisting small- and medium-sized plans to either become large, or to combine forces with other small- and medium-sized plans in order to claim the advantages available to large plans. It must also ensure that the very favourable investment expense ratios thus achieved are passed on to plan members.
Recommendation 9-2 — Pension policy and legislation ought to facilitate the growth and operation of large-scale pension plans or to enable and encourage cooperation among small- and medium-sized plans.
The logic of this position leads me to explore the very largest plan of all in Canadian terms — the Canada Pension Plan (CPP) — to see whether it exhibits design features that could be transferred to private or public sector occupational pension plans.
The CPP is huge. In 2006, 12.1 million Canadian workers made contributions to the CPP, while 3.4 million Canadians received benefits as retired workers, and another 1.1 million received retirement benefits as survivors of deceased retired workers. Because of amendments in 1996, the CPP is now partially funded and the Canada Pension Plan Investment Board now has $120 billion of investable funds managed by a professional staff of 400 analysts.
Most Canadians understand the CPP to be a DB plan since its funding is driven by a triennial actuarial report and its benefits are clearly defined by legislation. However, if one looks more closely, the CPP bears much more resemblance to a MEPP target benefit plan. Thus, the benefit structure of the CPP has changed many times since its inception in 1966, and in response to financial difficulties, at one point it decreased benefits for future retirees by close to 10%. If current contribution rates cannot sustain promised benefits, and no alternative is agreed to, benefits will automatically be reduced again. In this respect as well, the CPP more resembles a “target benefit” MEPP than it does a “pure” SEPP DB plan.
Thus, the example of the CPP reinforces my belief not only that size matters to the success of a plan, but also that the “target benefit” strategy is worth emulating.
9.4 A New Strategy for Ontario’s Occupational Pensions System
The two themes of this chapter — the need to complement “pure” DB or DC pension designs with innovative alternatives, and the need to provide all pension plans with the advantages of scale now enjoyed by a few — lead me to propose a new overall strategy for Ontario’s pension system.
Under this strategy, small DB plans, DC plans and individual savers with no access to occupational plans would be permitted and encouraged — but not required — to associate themselves with super-plans (those with assets in the range of $10 billion). These super-plans might operate in several ways. They might:
- extend full membership to workers previously enrolled in small plans and provide them with DB coverage;
- provide the nucleus around which new MEPPs might be formed by enlisting numbers of small SEPPs;
- offer standard MEPP-like target benefits to members of associated plans while requiring sponsors to make only fixed contributions; or
- provide investment or administrative services to associated plans, which would otherwise continue to operate on their original terms.
In order for any of these things to happen, however, a new approach to pension regulation would have to be adopted.
First, the government would have to allow Ontario’s existing large plans to amend their membership criteria and mandates. Some of those plans are created by trust documents or other private instruments; some are the product of special statutes. In either case, a fairly simple procedure should be established to make it possible for them to broaden the scope of their activities and the qualifications for plan membership, if they wish to do so. For example, large plans should be allowed to offer investment services to smaller plans and sell investment vehicles to individuals. Many of these plans, I believe, would welcome such an opportunity, as they are reaching a point of maturity past which their future net cash flows will shrink, and their ability to undertake new investments will be severely reduced. Without new members and new cash flow, their existing investment expertise will be under-utilized and, ultimately, difficult to sustain.
Second, the government or the proposed Pension Champion should encourage existing MEPPs to extend their mandates — again, by fairly simple means such as a resolution of the trustees or a referendum of the members. For example, construction MEPPs might be willing to combine with each other and to include workers in adjacent industries such as trucking or building products. This would obviously require inter-union cooperation, in which regard the present Ontario Municipal Employees Retirement System (OMERS) plan may represent a useful precedent.
Third, both SEPPs and MEPPs should be allowed to commingle their investments, as public sector unions have done in British Columbia under the British Columbia Investment Management Corporation (BCIMC), and federally under the Public Service Pension Investment Board. This might best be accomplished on a sectoral basis using TIAA-CREF in the United States as a model. TIAA-CREF manages the retirement plans of 3.4 million individuals from 15,000 institutions (mainly academic) and has $420 billion of assets under management. Obviously, the government would have to enact enabling legislation to permit commingling.
Fourth, it may be possible to encourage the emergence within the financial services industry of a large-scale, low-cost private agency to provide pensions to individual workers and members of small plans who must now buy pension coverage at very high cost or not at all. Aggregating their pensions in a commingled fund would enable them to secure retirement income on more reasonable terms than are presently available. One way to accomplish this might be for the government to call for proposals to establish such an agency and to run it for a fixed term — say 10 or 15 years — subject to renewal if the winner’s performance has been satisfactory or after new bids are sought.
Finally, the voluntary pension sector could be expanded beyond the employment relationship to encompass self-employed persons, holders of retail franchises, individuals who work for successive employers on short-term contracts, and employees in enterprises without a plan or in SMEs too small to sustain one. Many of these individuals belong to voluntary organizations, professional bodies or other associations that might be prepared to function as an “affinity group” for the delivery of pensions, as they do presently for the purchase of life insurance and other benefits. “Affinity” plans might operate in a fashion somewhat akin to the Quebec MFPPs, described above. However, such an approach would require amendment of the Income Tax Act, which presently requires that the sponsor–member nexus be based on an employment relationship.
What all of these proposals have in common is that they seek to expand access to affordable pensions to individuals who do not now enjoy it. This can be done economically only if sponsors or other pension providers can aggregate their funds into large pools of investment capital, reduce the cost of investment advice, lower administrative expenses, spread the investment risk and the longevity risk across a significant plan population, and provide members with the ability to project their retirement income with reasonable certainty (albeit subject to some degree of risk). This last point is of special importance in plans that offer target benefits, as almost all of the possibilities outlined above will do. If target benefits are provided, the risks associated with this type of plan must be transparent to all participants.
Recommendation 9-3 — Legislation and regulations should be enacted to enable and promote large commingled target benefit plans that might provide affordable pension coverage to Ontarians who do not presently have pensions or for whom the costs of obtaining a pension are unnecessarily high.
The effect of these innovations in plan design could be further enhanced by adopting legislation that, when a pension plan is available in a given workplace, participation should be mandatory and automatic, subject to the right of a worker to opt out. Recent experience in the United States suggests that this “nudge” strategy — previously mentioned in Chapter Eight — may significantly increase participation rates. The U.S. experience should be monitored to see whether it is, in fact, as positive as it initially seems to be. If it is, Ontario should consider enacting similar legislation.
Finally, I visualize that the role of the proposed Pension Champion will be to put flesh on these admittedly barebones proposals; to test the waters with pension providers, potential plan members and other stakeholders; to devise a legislative framework for this initiative; and to ensure that all of this is accomplished in a manner that supports those ongoing pension plans that would prefer to continue to operate as they had previously.
9.5 Ontario’s Long-term Pension Strategy: Stakeholder Views and Additional Possibilities
In concluding this chapter on innovation in plan design, I feel obliged to report that a significant number of submissions raised the possibility that an expanded or two-tier CPP, or a new provincial counterpart plan, might offer many of the advantages that I seek to capture in Recommendation 9-3 for a new strategy to promote large-scale target benefit plans. I was particularly struck by the fact that this idea was raised in different ways in briefs from stakeholders as disparate as the Canadian Federation of Independent Business and the Canadian Labour Congress.
Details varied. Some urged an expansion of the existing CPP structure so as to either increase the benefit (now 25% of the retiree’s best 40 years of accruals), or to increase the maximum earnings on which the benefits accrue (now $44,900 — approximately the average wage), or both. Others suggested allowing employers and employees to voluntarily make extra contributions to the CPP (to be invested by the Canada Pension Plan Investment Board) to buy commensurate extra benefits (administered at a low marginal cost within the existing CPP framework). No matter the details, those making these submissions maintained that a public scheme of some sort would enhance pension coverage, improve benefit portability, and contain costs.
I had neither the mandate nor the time nor the resources to investigate the possibility that an extension of the public pension system might complement, supplement or even replace the existing voluntary occupational pension system, which was the focus of my Commission’s work. However, it is an idea that clearly deserves careful study over the longer term while efforts to reinvigorate the present system are under way.
Recommendation 9-4 — The government of Ontario should investigate the advantages and disadvantages of expanding the mandate of the Canada Pension Plan, or creating a comparable provincial plan, so as to enhance pension coverage, control costs and improve benefit portability.
The idea of expanding the CPP, among others, has attracted considerable interest outside Ontario as well, and proposals are afoot to discuss it at a national “pension summit.” I believe that if such a summit does occur, it ought also to explore a broad spectrum of proposals for large-scale, long-term changes in the architecture of the pension system.
Recommendation 9-5 — The government of Ontario should support the call for a national pension summit whose agenda should extend to all ideas for significantly expanding pension coverage, including the innovative proposals contained in this report.