Chapter Two — What Pensions Do

2.1 Introduction

In this chapter, I describe the effects of Ontario's defined benefit (DB) occupational pension system on a number of policy domains: the provision of retirement income security, the promotion of savings, the functioning of labour markets and the province's economic well-being in general.

Describing these effects ought to be a straightforward task. Unfortunately, it is not. The data on which such a description must be based was neither readily available nor particularly reliable — a complaint first made by a Canadian pension commission in 1912, and repeated by virtually every commission since. Commission staff and contract researchers — notably Richard Shillington of Informetrica Ltd. — made extensive efforts to obtain, integrate and verify the relevant data. However, though now likely as good as it is going to get, the data available to the Commission remains more uncertain than it ought to be in such an important field of public policy. There are several reasons for this.

First, many of the standard sources of pension data for Ontario were simply not reliable. The three principal agencies that collect Ontario pension data — the Financial Services Commission of Ontario (FSCO), Statistics Canada (StatsCan) and Canada Revenue Agency (CRA) — each does so for its own purposes and from a somewhat different perspective. FSCO, the primary agency, regulates not just pension plans but other kinds of financial institutions. Consequently, though its executive head, the Superintendent, is authorized under the Pension Benefits Act to “conduct surveys and research programs and to compile statistical information related to pensions and pension plans,” FSCO does not see itself as having the mandate or resources to collect or analyse data not directly related to its regulatory functions, or to support policy making within a broader social policy framework. It is further hampered by its reliance on cost-recovery financing, which inevitably limits its ability to develop a substantive policy research capability. As a result, FSCO simply does not collect a great deal of data that would be useful for policy making and, unfortunately, when data sets do exist, they are sometimes erroneous, miscoded or intermittent.

StatsCan and CRA — both federal agencies — produce an array of pension data. However, StatsCan relies in part on FSCO data, whose limitations are mentioned above. Moreover, because of differences in definitions and methodology as between FSCO and StatsCan, it is often difficult for third parties such as academic researchers or policy analysts to integrate studies produced by these two bodies. For similar reasons, data generated by CRA is not always consistent with data from StatsCan. For example, the StatsCan definition of a multi-employer pension plan (MEPP) differs both from that of FSCO and of CRA. This has resulted in coding errors by the other agencies, which in turn makes analysis of the long-term growth of these important plans very difficult. Likewise, the definition of pension plan members in Ontario fails to take account of the fact that a worker may be contributing to one plan while being a deferred member of several others.

And finally, neither federal agency produces data sets that can be directly and easily used by Ontario policy analysts. For example, readily available federal data often does not distinguish between workers resident in Ontario whose employment and pensions are regulated by Ontario legislation, and those whose employment and pensions are regulated by federal legislation. The latter group is considerable: it includes Ontarians employed in federally regulated businesses such as banks, railways, telecommunications and airlines (whose pension plans are regulated by the Office of the Superintendent of Financial Institutions), as well as members of the federal public service and the armed forces who happen to reside in Ontario. Nor does the available data readily permit measurement of the sectoral or occupational distribution of workers enrolled in pension plans.

How do these data difficulties affect the Commission's work? I cite three examples among many.

First, even basic information concerning the number of workers contributing to, or retirees collecting benefits from, DB pension plans regulated by the province's Pension Benefits Act could not — and still cannot — be provided at a proper level of reliability.

Second, Chapter Three explores the extent and cause of a decline in DB pension coverage, one of the key issues confronting the Commission. However, it is almost impossible to determine coverage by economic sector for plans under Ontario jurisdiction. FSCO does not track sectoral coverage on a consistent basis. StatsCan does do so, but its confidentiality and suppression policies preclude publication of much of its data. Consequently, only with the greatest difficulty has the Commission been able to assess the straightforward claim that the decline in employment in the manufacturing sector accounts for a significant proportion of the decline in overall pension coverage. And if it is difficult to reliably link occupational pension plan coverage to the social and economic characteristics of plan members on a current basis, it is almost impossible to do so retrospectively.

Third, the Commission has tried to ascertain the extent to which DB coverage is being replaced by defined contribution (DC) coverage rather than by no coverage at all. However, statistics on retirement income are not sufficiently disaggregated by source to enable the Commission to measure precisely the proportion of income attributable to different types of pension plans. It is therefore unable to do more than approximate what might be the future consequences for Ontario's income security policies and its economy of the current downward trend in DB coverage.

In short, the data developed by the Commission to support the analysis in this and succeeding chapters is the best that could be managed (and more accurate, I believe, than data used in previous pension studies in Ontario or Canada). However, it is still too approximate, still inferior to data used to support studies and regulatory activity in some other jurisdictions, and still inadequate as the basis for future pension policy development.

This last point bears repeating. Ontario's pension system can be strengthened only if we proceed from an informed and careful diagnosis of the difficulties of the present system. But no such diagnosis is possible if professionals, researchers and advocates lack the tools to undertake statistical research. In Chapter Ten I make several recommendations that should improve their situation. For now, I merely note that Ontario's Ministry of Finance has already begun to enhance its capacity to collect and analyse pension data.

Back to top

2.2 The Contribution of Occupational Pensions to Ontario's Retirement Income System

Pension systems should not only alleviate poverty among the elderly but also prevent a significant fall in the living standards of workers upon their retirement. Success in achieving the latter objective is usually measured by reference to a suitable, but much-debated, “replacement rate” — the substitution of retirement income for wages from employment.

In Canada, as in many other countries, the pension system comprises both private and public elements. Private elements include various forms of savings, some of them tax-sheltered, such as Registered Retirement Savings Plans (RRSPs), the recently initiated Tax Free Savings Accounts (TFSA) and especially occupational pensions provided voluntarily by employers or unions, or under their joint sponsorship. Public elements include the federal Old Age Security (OAS) program, the Guaranteed Income Supplement (GIS), the Canada/Quebec Pension Plans (C/QPP) and other schemes established by the provinces.

Relative to many OECD countries, Canada's public retirement income programs are quite modest. As a result, occupational pension plans and other forms of private savings play a more important role in providing retirement income security and in achieving a suitable replacement rate. Moreover, the public and private systems have become tightly integrated so that clawbacks at a rate of 50% are built into the design of both the federal GIS and the Ontario Guaranteed Annual Income System (GAINS). Thus, integration reinforces the significant role accorded occupational pensions in retirement income policy.

The difficulty is, however, that in jurisdictions such as ours, with voluntary private pension systems, coverage of occupational pension plans has never exceeded 50% of the workforce. In fact, in Ontario, while the absolute number of workers with pension coverage has been increasing, the percentage of those with coverage has been declining slowly since the 1970s and now stands at 34.7% of the paid labour force under provincial jurisdiction — somewhat below the Canadian average (including federal workers) of 38.5%. Nonetheless, the absolute numbers are still impressive: approximately 2.18 million Ontarians are covered (1.76 million by plans under provincial jurisdiction; 420,000 by plans under the jurisdiction of the federal government or some other province) and a further 1.0 million Ontarian retirees are collecting pension benefits (850,000 under Ontario-regulated plans, the balance under plans regulated by other Canadian jurisdictions). These workers and retirees, together with their families and dependants, constitute a significant presence in virtually every community in Ontario.

In general, average retirement incomes and replacement rates have improved since the early 1970s. As a result, the financial security of Canadian seniors has improved considerably relative to what it once was, and relative to that of seniors in many OECD countries. However, this significant achievement for Canadian social policy owes at least as much to the maturation of the C/QPP system, and to the introduction of other age-related public support systems, as it does to occupational pension plans.

Moreover, the experience of different groups of retirees varies, depending on the extent to which their retirement incomes are derived from public as opposed to private plans. Private retirement programs — occupational pension plans and RRSPs — constitute a much larger proportion of the incomes of middle- and high-income retirees than of low-income retirees. Longitudinal studies show that Ontario retirees in the most affluent quintile rely on occupational plans and private savings for 41% of their income, and on public plans for only 16%. By contrast, retirees in the poorest quintile receive 57% of their income from the public system and only 21% from private pensions and RRSPs. Indeed, most lower-income retirees have nothing more than public plans to rely on. Only about 25% of families in the poorest quintile had an occupational pension, and only 2% of families in this group had two.

These contrasting positions are made more extreme because wealthier Ontario families tend to have resources in all areas required for the provision of retirement income security: home equity, retirement assets and personal investments. They are also more likely to have more than one private pension plan: for top-quintile earners, roughly 80% of prime-aged families had at least one occupational pension plan, and at least 40% of them had two. They are also more likely to have some post-retirement employment earnings, being generally well-educated and often self-employed. And finally, wealthier retirees also receive public pensions, although as noted, these are subject to being taxed or clawed back.

Not surprisingly, then, since the early 1980s, the capacity of wealthier and lower-income families to meet their retirement income needs has been diverging. While average retirement savings of couples grew during the period 1986-2004, most of this growth was attributable to higher-income earners. For those in the bottom quintile, savings were stagnant — perhaps because they were obliged to spend their relative meagre earnings for current needs. In sum, there has been an increase in inequality of retirement income that mirrors the long-term trend toward greater inequality in earned income.

This inequality, finally, is not distributed randomly among different demographic groups. Our research shows that, as is the case with wages, males have higher retirement incomes than females, although the gap is closing. And, as noted in the next chapter, recent immigrants are more likely than others to find themselves with little retirement income beyond what the public system provides — and even the public system does not provide them with full benefits. It seems likely, too, that members of visible minorities are similarly disadvantaged.

What is the position of seniors who have no occupational pensions and must rely entirely on public schemes for their retirement income? Today, a single 65-year-old would receive no more than $18,000 a year from public sources. This person would be living just on the cusp of poverty, according to several accepted benchmarks. In terms of income replacement ratios, public plans do, on average, provide the lowest quintile of retirees with income benefits greater than 100% of their pre-retirement earnings throughout their retirement. However, about 20% of this group — mostly single women — had replacement ratios below 80% by the time they were 70 years old, which would mean they too are living in poverty. Longitudinal studies show little change in these replacement rates over time.

Furthermore, in more recent years, with the maturation of the C/QPP, the most rapid source of income growth for older Canadians has been not from public plans but from occupational pensions and annuities, often purchased by private savings, including RRSPs or DC plan lump sum payouts. Several studies attest to the growing relative and absolute importance of occupational pension plans and RRSPs in the provision of retirement income security. Thus, in 1984, Canadian seniors received 76.8% of their income from public plans and 23.2% from occupational pension plans and RRSPs; in 2004, they received 59.8% from public plans and 40.2% from these private sources. Whether this trend will continue in Ontario in light of the difficulties in its pension system, described in Chapter Three, is a question that requires careful study.

To conclude, the public and private elements of our retirement income system are related, even integrated, in certain ways. But while the public pension system does mitigate the worst effects of income polarization, patterns of private retirement saving (which include occupational pensions and RRSPs) are determined by, and therefore tend to reproduce and possibly exacerbate, the disparities in income that Ontarians experience during their working lives.

Back to top

2.3 Occupational Pensions as a Strategy to Promote Savings

Defined benefit plans — which account for about 80% of all occupational pension coverage — represent only one of several possible strategies of saving for retirement. How do they compare with the alternatives?

Since they operate through collective pooling of contributions and assets over a long period, DB plans in effect provide some degree of security or “insurance” against many of the risks confronting retirees, such as outliving their savings (the “longevity risk”) or experiencing poor investment returns (the “investment risk”). Because DB pensions spread these risks among all plan members, young and old — even across generations of plan members — they help to ensure that retirees will have a source of income until their death. Some DB plans provide disability or other ancillary benefits and some provide ad hoc or formulaic protection against inflation. Such advantages are not normally available to individuals enrolled in DC plans or those who by choice or necessity save for their own retirement. In addition, DB plans — especially large plans — are highly cost-efficient in terms of their administration and fund management, and provide members with a level of investment expertise that few individual savers themselves possess. Indeed, studies suggest that large DB plans produce significantly higher net investment returns than other savings vehicles, when service costs and fees are taken into account.

To be sure, DB pension plans have been criticized as “paternalistic” because workers cannot “take control” of their own money or choose their own investments. However, studies have frequently shown that workers seldom know what benefits they have a right to receive, let alone how their pension fund actually works. Even when given individualized investment choices, as are sometimes available under DC plans, they often turn out to lack the financial knowledge and skills necessary to optimize their investment returns. And, of course, by contrast with members of DB plans, DC plan members must not only manage the investment risk, but confront the longevity risk as well.

Finally, DB plans are not only collective but, quite often, mandatory for everyone employed in a workplace that has such a plan. This is a positive feature of occupational pension plans. Study after study confirms that, given the choice, people tend not to save enough for retirement, if indeed they save at all. Researchers of the “behavioural finance” school label this tendency as “un-knowledgeable,” “apathetic” or “myopic” — terms that apply not simply to those with low incomes (who are under strong pressures to spend what they earn) but also to those with significant capacity for savings. No doubt that is why some jurisdictions — notably Australia — have moved to universal, mandatory occupational pension systems with opt-out provisions. Under their systems, all employers must contribute but, typically, contributions remain voluntary for employees, with incentives provided by the government in the form of matching payments. In the United States, whose Pension Protection Act of 2006 encourages but does not require employers to initiate plans with opt-out provisions, preliminary research shows that most workers are still not saving nearly enough to support themselves in retirement.

These features of DB plans — security, cost-efficiency, investment expertise, mandatory membership — strongly recommend them as a savings strategy. Of course, DB plans are not perfect. They depend crucially on the willingness of the employer to sponsor a plan in the first place and to continue to support it in the face of a challenging business environment and rising pension costs. Plans may fail due to poor administration, the insolvency of the sponsor or some external cause. They may suffer temporary financial reverses. And they may not — and often do not — provide adequate protection against inflation.

But for all their shortcomings, DB plans appear to be, like democracy, the worst system of retirement savings — except for all the others. Those shortcomings are addressed in subsequent chapters. So, too, is the need to enhance the understanding by plan members of the risks and benefits associated with different savings strategies and, if they are covered by a DB plan, of what that plan does and does not offer them.

Back to top

2.4 The Effect of Occupational Pensions on Labour Markets

Occupational pension plans are not simply part of the larger system of retirement income security and an attractive vehicle to encourage savings — they are also a significant influence on labour markets.

Employers provide pensions because they believe that doing so helps them to attract and retain the workers they need and want. And indeed, pensions do seem to produce the effects imagined and intended by employers: workers with pensions stay in their jobs longer than those without. However, the precise cause-and-effect relationship between pensions and job tenure is unclear. On the one hand, because the design of many DB plans is back-loaded, workers know that they must stay to receive the full value of their pension. On the other, workers with pensionable jobs may be reluctant to leave them because of the other non-pension benefits typically associated with such jobs. In fact, studies have shown that workers with pensionable jobs enjoy an overall remuneration premium of up to 29% over other workers. Nor is this conclusion counter-intuitive given that pensions are most commonly provided in larger workplaces and in unionized workplaces, both of which also tend to provide higher levels of wages and benefits.

Paradoxically, pensions can be used not only to reinforce job tenure but also to facilitate its termination. This may happen in several ways. First, the availability of a pension requires workers and employers to address the issue of a “normal” retirement age, even though mandatory retirement policies are now unlawful in Ontario and most Canadian provinces. Second, pensions may be used strategically by employers to reduce their workforce by providing early retirement incentives to redundant workers. And third, as employers begin to confront anticipated shortages of skilled labour, they may decide to use pension plans as one element of a flexible retirement strategy that would allow workers approaching retirement to reduce their hours of work gradually rather than all at once, and to maintain the level of their earnings through some combination of wages and pension benefits.

Despite all the positive contributions of DB plans to labour markets, however, the fact that they tend to bind workers to their jobs has also been the subject of adverse comment. Portability is the crucial issue. Workers are understandably reluctant to leave their DB pensions stranded behind them in order to move on to new jobs where their skills might be in higher demand, their productivity might be enhanced and their overall rewards might be greater. This reluctance might diminish if DB pensions were fully portable. However, they seldom are — a situation that has given rise to strikingly different critiques. On the one hand, as early as 1961, the Ontario Committee on Portable Pensions recommended the establishment of an agency that would facilitate portability. Subsequent studies have made recommendations to the same effect. On the other, proponents of flexible labour markets point to DB pensions as a factor contributing to reduced labour mobility and therefore inefficiency, as compared to DC plans, which present no obstacle to movement. (Ironically, as Morley Gunderson points out in his study for the Commission, DB and DC plans are both associated with reduced quits and lower turnover rates.) In Chapter Five, borrowing from the 1961 Commission's recommendations, I review possible arrangements that might both preserve labour mobility and enhance the portability of DB pensions.

As to more general labour market effects, some stakeholders, particularly in manufacturing, argued in their submissions that occupational pensions represent a significant cost that Ontario employers can ill afford in an era when global competition, technological innovation, the high Canadian dollar and other developments are all generating pressures to lower compensation costs. Gunderson's study suggests possible limits to this argument. As he notes, “pensions can have important incentive effects that can be a part of strategic human resource planning on the part of firms...” Presumably the sacrifice of these effects must be reckoned as at least partial offsets to any savings gained by reducing pension costs. Moreover, as he also points out, pensions form part of a “total compensation package;” if employers reduce or eliminate pensions, market forces may return some or all of the savings employers anticipate to workers in the form of wages or non-pension benefits.

This is not a debate that can be resolved in the present context, or indeed elsewhere. Conditions change; theories proliferate; the evidence is likely to remain inconclusive. There is no doubt that Ontario's manufacturers and other employers are experiencing significant competitive pressures, and that they will try to contain labour and other costs — including those associated with pensions. This they have every right to attempt to do, especially in the context of a voluntary pension system. But whether they will succeed, and whether changes in pension preferences, provision and policy can, will or should assist them in restoring their competitive position over the long term is far from clear.

Back to top

2.5 Occupational Pensions and the Economy

Pension plans have enhanced the dynamism of Ontario's economy in at least three ways: through the investment of their funds; by increasing the purchasing power of pension beneficiaries; and by reducing the reliance of older Ontarians on tax-supported income supplements and needs-related, in-kind benefits. However, as funds mature, as the proportion of Ontario workers enrolled in occupational pension plans declines, as the percentage of retirees with such pensions dwindles, and as older people come to represent a growing segment of the general population, each of these economic effects is likely to change significantly — and not for the better.

2.5.1 Pension funds as investors

Canadian pension funds, with over $950 billion of Canadian assets as of 2005, are the country's second largest source of capital after the chartered banks. Funds based in Ontario-regulated plans account for some $380 billion, 40% of the total. At the end of the second quarter of 2006, according to Statistics Canada, stocks and equity funds accounted for 39% of pension fund assets; bonds and bond funds, 32%; real estate, 7%; short-term investments, 4%; mortgages, 1%; and other assets, 17%. Pension funds are thus a prominent feature of Ontario's financial landscape.

The investment profile of pension funds tends to vary according to plan size. The larger plans, in particular, tend to be a major presence in the financial industry, promoting new financial strategies, institutions and products, as well as improved corporate governance practices. These funds often hire their own in-house investment staff, and possess the resources and long-term investment horizons that enable them to invest in alternative asset classes such as private equity, real estate and public infrastructure. While less liquid, these assets tend to provide higher rates of return. Indeed, a few large plans have established their own investment companies or departments to handle private equity projects; some pool funds through partnerships set up in cooperation with other plans; and yet others invest through more conventional commercial partnerships.

According to some estimates, only plans with more than $10 billion in assets are able to operate across the full investment spectrum. While there are only five such plans in Ontario, they collectively account for 35% of all active plan members under Ontario jurisdiction. By contrast, small- and medium-sized pension funds — the vast majority of plans — generally lack the infrastructure, resources, expertise or inclination to undertake similar portfolio diversification on their own account. As a result, they tend to purchase standard financial products; pay significant fees to financial intermediaries, advisors and service providers; and, in consequence, generate less favourable returns. This discrepancy in market behaviour and financial results between large and small plans prompt some suggestions later in this report.

Pension funds are able, and sometimes willing, to provide “patient capital.” Both the federal and provincial governments have therefore expressed interest in persuading them to invest in infrastructure projects such as water and waste disposal facilities, hospitals, schools, highways and transit systems, courthouses and correctional institutions. Public-private partnerships sometimes serve as the vehicle for pension fund investment in these infrastructure projects, although this approach is controversial in some circles. The Bank of Canada recently suggested improvements in the framework governing the financing of public infrastructure projects, with a view to encouraging even more extensive investment by Canadian pension funds in such projects.

Finally, the firms that manage Canadian pension funds, many of which are located in Ontario, produce significant earnings for their clients through trading gains, dividends and interest payments on longer-term investments. These pension managers often become successful wealth-generating organizations in their own right, a significant presence in Ontario's financial services sector and important contributors to its economy.

2.5.2 The impact of pension funds on the purchasing power of retirees

While the data is far from reliable, and reaches back only to the beginning of this decade, it suggests that occupational pensions account for about 20% of the overall purchasing power of retired individuals and couples — their largest single source of income. To be sure, this percentage is not constant among age cohorts; it differs among retirees with different overall levels of income and is lower for those over 85. However, income from occupational pensions amounting to some $36.1 billion in the aggregate clearly represents an important component of the income available to older Ontarians. Moreover, expenditures by retirees of this large sum on goods and services generate multiplier effects that, in turn, benefit businesses, workers and communities across the province.

2.5.3 The relationship between occupational pensions and government revenues and expenditures

As Chapter Three explains in greater detail, occupational pension coverage has been declining for several decades. If this decline continues, what will be the effects upon the province's financial health?

On the one hand, the existence of occupational pension plans reduces the potential cost of publicly funded income support systems, which would otherwise come under pressure to pay higher benefits and, likely, to cover a larger proportion of retirees. Studies undertaken for the Commission by Richard Shillington and Keith Horner show how the clawback provisions of the federal GIS and provincial GAINS programs presently operate to accomplish this result. As a corollary, if fewer clawbacks occur due to declining pension coverage, these programs are likely to cost taxpayers more. On the other hand, benefits paid by occupational pension plans are taxable income in the hands of retirees and thus make a positive contribution to the fiscal health of the province. The shrinkage of pension coverage represents a potential threat to this revenue source.

It is sometimes suggested that personal RRSPs can and will fill the void in retiree income left by the diminishing coverage of occupational pension plans. However, the Horner study suggests that this is unlikely to be the case, as RRSP contribution rates seem to be declining in tandem with the decline in occupational pension coverage. The consequences of a decline in pension coverage must therefore be confronted head on.

Horner has modelled the impact of a 10% decline in occupational pension coverage from the current level of just under 35% to about 31.5%. He analyses two scenarios, one where the 10% decline is not replaced by other savings, and one where increased RRSP contributions offset the decline in pension income by one-half. He suggests that by 2030, a combination of reduced tax revenues and increased benefit expenditures in public income-support programs would cost the provincial and federal governments together between $819 million and $1401 million, depending on which assumption is used concerning RRSP contributions. Horner also notes that the strongest effects are likely to be those on tax revenues — nearly 70% of the total. By 2030, the impact of a 10% pension decline in current pension coverage would yield a decline of about 0.7% in the net cash flows of the federal and provincial governments combined. These are non-trivial consequences.

Finally, mention must be made of a development that may perhaps mitigate the predicted negative effects of declining pension coverage. Due to the increased labour force participation of women, the overall proportion of adults in “paid work” in Canada has increased in recent decades. Conceivably, the resulting increment in household earnings and savings could offset the decline in pension income, allow retirees to achieve desired income replacement rates even without pensions and therefore insulate public income support programs from the pressures forecast above. Time — and more and better research — will tell.

Back to top

2.6 Conclusion

In light of the generally positive contribution of occupational pensions to various domains of Ontario's social and economic policy, to its workplaces and capital markets, and to its financial well-being, it seems both odd and unfortunate that this chapter must end by noting that the system has for many years experienced a long, slow decline. The next chapter examines the extent, and possible causes and consequences, of that decline.

Back to top