Chapter 20 A shaky third pillar: The vulnerability of retirement incomes

View the most recent version.

Introduction
Role of the third pillar
Importance of economic context
Financing problems of workplace pension plans
Declining coverage of workplace pensions
Shift from DB to DC
Overall assessment of the challenges in the third pillar and conclusions
Bibliography

Introduction

In the mid-1960s, when the Canada Pension Plan (CPP) was being introduced and the broad contours of Canada's current pension system took shape, the government of the day made it clear that the modest size of the CPP was designed to leave plenty of room for the voluntary private retirement savings schemes to flourish (LaMarsh 1968). Again, in the mid-1980s, at the conclusion of a prolonged debate on pension reform, the government rejected a more substantial role for the CPP in order to give room to private arrangements (Department of Finance 1984).

The private arrangements to which space was to be given were of two basic types: workplace pension plans, often referred to as 'Registered Pension Plans,' and tax-supported, individual savings accounts known as 'Registered Retirement Savings Plans' (RRSPs).¹ Both types of scheme are privately administered and are financed in significant measure by investment returns in addition to contributions based on employment earnings. Participation in workplace pensions is voluntary for employers but, in most cases, participation is not voluntary for employee plan members. Moreover, both types of scheme rely on supportive tax measures. Taken together, these two types of arrangement are referred to as the 'third pillar' in the three pillars approach to providing retirement income that has been employed by the World Bank (World Bank 1994). ²

Dividing the third pillar into workplace pension and RRSP components reflects distinctions in Canadian tax law. But, it is worth noting that workplace pensions can be categorized as either 'defined benefit' (DB) or 'defined contribution' (DC) plans. The former type of plan includes a formula according to which the amount of pension benefits payable at retirement age will be calculated. A typical formula might be:

Benefits = 1.75% x years of service x the average of a plan member's best five years earnings (1)

Contributions to a defined benefit pension plan are calculated intermittently based on past experience and future estimated experience.

A defined contribution (or money purchase) plan provides that contributions will be made to a plan at a specified rate. Contributions and investment returns accumulate over the course of working life. At retirement age, the accumulated assets are used to buy an annuity. The amount of the retirement benefit is unknown until retirement age and will depend on the amount of accumulated assets and the interest rates that underlie annuity prices at the date of retirement.³

Defined contribution pension plans function very much like RRSPs that include employer contributions and a prescribed rate of saving. Indeed, the distinction between DC pension plans and RRSPs has become blurred in recent years as many employers have chosen to put in place group RRSPs as opposed to DC pension plans. One distinct feature of group RRSPs as opposed to DC pension plans is that pension regulatory law does not apply to them.

To date, the decision to create a pension system that relies heavily on the third pillar as a source of income appears to have worked reasonably well. Over the period from the 1970s to the end of the twentieth century, poverty rates among the elderly declined substantially, although this was largely attributable to the first two pillars. The average purchasing power of all elderly household incomes increased by roughly 40%, from 1973 to 2000, when the average incomes of all elderly households reached approximately $35,000. The income gap between the elderly and the non-elderly narrowed from just over 20% in 1973 to just over 10% in the mid-1990s (after adjusting for differences in household size), and then increased to about 20% in 2000.⁴

The improvement in living standards was even stronger for the elderly households who are fully retired and have no earnings from employment. Their purchasing power came close to doubling over the period from 1973 to 2000 (from roughly $15,000 to $30,000) and the income gap in relation to younger households declined from nearly 50% to about 30% in the mid-1990s, and increased slightly by 2000. Thus, in 2002, a clear majority of Canadian retirees assessed their financial situation to be about the same in retirement as beforehand, with about one in three saying they are worse off. Being worse off was strongly associated with retiring for heath reasons and being involuntarily retired.⁵

In official policy discourse, the issue whether the elderly have and will continue to have adequate incomes gets minimal consideration. In all of the material that was prepared in connection with the reforms to the Canada Pension Plan in the late 1990s, nothing was said about this issue (Department of Finance 1996a). The same held true of the material prepared in connection with the proposed Seniors Benefit (Department of Finance 1996b). Through the latter part of the 1990s, the OECD had a major work program on ageing societies. The work done as part of this work program acknowledged that there are income shortfalls in many OECD countries for single, elderly women and for adult immigrants. But, for the most part, income adequacy was dealt with as a problem that had been solved (OECD 2001). Recently, the Régie des rentes du Québec put out a discussion document on the future of the QPP that addressed the issue of income adequacy by noting:

This paper looks at developments in the third pillar and asks whether there might be a broader issue of vulnerability to inadequate incomes than is suggested in the mainstream policy discourse. While the capacity of the third pillar to deliver retirement benefits in the future warrants a formal modelling exercise, what is offered here is an assessment of some of the key influences on that capacity.

Before turning to that, three points are worth adding:

1) It has become common practice in Canada and in many countries to compare income data of all people aged 65 and over with younger populations, and to draw inferences directly about the adequacy of retirement and pension provisions from that comparison. But, incomes of people 65 and over include a surprisingly large amount of income from employment. If the comparison is made only with older people who have left the labour force, the income situation of the elderly does not look as good.

Chart 20.1 shows that the incomes of all people 65 and over reached 90% of the level of the non-elderly by the mid-1990s before receding slightly. However, for fully retired households,⁶ average income of the over-65s was only 70% of that of younger households, even after taking account of differences in household size.

2) The elderly in Canada tend to have a much more compressed distribution of income than do the non-elderly. There is a heavy concentration of incomes just above and just below the Statistics Canada low income cut-off. This means that there tends to be less deep poverty among the elderly (i.e. incomes well below the low cut-off line), but it also means that there are more older people living close to the low income cut-off than would be true for younger populations.

3) For the low income elderly, the flat rate Old Age Security (OAS) program and the income-tested Guaranteed Income Supplement (GIS) are extremely important sources of income. (The OAS and GIS account for the following percentages of income in the first three deciles: 76.2, 63.3 and 64.0.) In addition, the OAS is part of the foundation on which the CPP and workplace pensions are built. But, because the program is only price indexed, its value relative to average earnings in society declines in the face of real wage growth. The relative living standard of the elderly declines in this situation.

Role of the third pillar

The rapid growth of income of the elderly over the period from the mid-1970s to the end of the century reflected real income growth from a number of major sources. By far the fastest growing source of income over this period was income from the C/QPP which increased from a mere $668 per elderly household in 1973 to $6,763 in 2000. (All dollar amounts in this section of the paper are in year 2000 dollars.) This huge increase reflects a growing portion of the elderly becoming eligible for C/QPP benefits - from 28 to 93% of elderly households receiving C/QPP benefits - as well as the rapid phasing-in of full retirement benefits - a process that took place over the ten years from 1967 through 1976. The portion of elderly household income coming from C/QPP increased from 2.8% to 19.1% of total income received by elderly households.

But, workplace pension income was the second fastest growing source of income. Average amounts received per elderly household increased from $2,527 to $10,436 over the period from 1973 to 2000. The increased amounts received reflected both an increase in the portion of elderly households receiving income from the source (66.3% in 2000, versus 27.9% in 1973), and an increase in the amount received per household that received income from this source (from $9,049 in 1973, to $15,881 in 2000). Workplace pension income accounted for 10.4% of income received by elderly households in 1973, and 29.4% in 2000. The increase in workplace pension income was particularly rapid between 1981 and 1989.

By way of contrast, income received by older households from investments was somewhat erratic, and in 2000, averaged $6,782 compared to $4,066 in 1973. Average amounts received from OAS/GIS remained relatively stable over the period, increasing from $7,434 to $7,978. Finally, average earnings from employment fell from $8,314 to $4,049. This was accounted for primarily by a drop in the number of households with earnings from employment from 32.6% in 1973, to 22.9% in 2000. Most of this decline was accomplished by 1989. Where they were present, earnings from employment tended to be substantial and very unequally distributed (see Table 20.1, below).

Workplace pensions clearly made an important contribution to income growth among older Canadians over the latter part of the twentieth century. But, the contribution of workplace pensions to income improvement has not been distributed evenly across the elderly population. Table 20.1, which is based on data from the year 2000, provides an insight into this reality.

Table 20.1 shows that in the year 2000, workplace pension income was providing a larger share of total income to Canadians 65 and over than was either OAS/GIS or C/QPP (but less than the two public programs combined). This outcome reflects the fact that workplace pensions provide larger amounts of income to smaller portions of the elderly than do OAS/GIS or C/QPP. Two-thirds of elderly households received income from workplace pensions, and 29% of their income came from that source. But, 97% and 93% of the elderly received incomes from OAS and C/QPP, but received only 23% and 19% of their incomes respectively from these sources.

At the low end of the elderly income distribution, workplace pension income is notable primarily by its absence. In the middle of the distribution, the percentages of people receiving workplace pension income are increasing quite significantly (from 28.8 in decile 4, to 77.6 in decile 6), but the amounts received are still quite modest. But, in the top three deciles, both the percentage of the elderly population receiving workplace pension income and the amounts received are quite considerable.

Bearing in mind what Table 20.1 reveals about the role of workplace pension income by decile, it will surprise virtually no reader that there are differences by gender. Among Canadian men 65 and over, 68.1% receive workplace pension income compared to 47.9% of Canadian women. Older men in the fourth decile of the male distribution are slightly more likely than women in the seventh decile to receive workplace pension income (69.9% compared to 69.3%). That said, it is striking that in the top three deciles of the female distribution, workplace pension income is very widespread with the percentage of older women receiving income from this source ranging from 84.2% up to 88.6%.

It is also striking that workplace pension income grew very rapidly as an income source for older women between 1996 and 2000. In 1996, only 34.2% of women 65 and over received workplace pension income compared to 47.9% in 2000. Moreover, the average amounts received increased from $2,696 to $4,233 in 2000. As will be discussed more fully below, there is reason to hope that gender differentials might diminish further in the future. Gender differentials may however, be diminishing in the context of an overall situation that is deteriorating.

Another subset of the elderly population for whom workplace pensions have played a somewhat smaller role than would be inferred from the average case is adult immigrants to Canada. Among adult immigrants who receive workplace pension income, the amount received is very close to the average for the entire elderly population; it was 94% of the average for the entire elderly population in 2000. There was however, a somewhat larger gap in the percentage of elderly households in receipt of workplace pension income (49.8%) versus the elderly population as a whole (56.8%). Thus overall, workplace pension income for adult immigrants 65 and over was 21% lower than for the elderly population in total ($5,683 versus $6,900).

Unfortunately, it is harder to be optimistic about the future prospects for the receipt of workplace pension income by adult immigrants to Canada. The well-documented increase in the difficulties they have faced getting good employment in recent years would normally have negative effects on the receipt of workplace pension income in the future (Jackson 2005).

Importance of economic context

The third pillar has made an important though somewhat uneven contribution to the improved well-being of elderly Canadians. As has been noted, this is attributable in significant measure to the maturation of workplace pension plans as well as the maturation of the C/QPP. As a result, a good deal of commentary on the improved income situation of Canadians 65 and over attributes the improvement to the strength of the Canadian retirement income system. What is often overlooked is the way in which the third pillar interacted with a very specific set of economic circumstances to produce both the positive absolute and relative income situation of the elderly noted above.

One feature of workplace DB pensions is the absence of formal arrangements to protect pensions against inflation. As of the year 2000, only 9% of all DB pension plans and 16% of all pension plan members had full protection against inflation. Another 29% of DB plan members had partial protection against inflation. As a result, the purchasing power of workplace pension incomes is subject to a significant degree of risk from inflation (Statistics Canada 2000).

Table 20.2 shows the loss in purchasing power over each of the last five decades of the twentieth century that would have been experienced by recipients of pensions that were not price indexed. It gives the amount of purchasing power that remained at the end of each decade based on an income that began to be provided at the start of the decade, but received no protection against inflation.

The dramatic effect of the high levels of inflation in the 1970s is evident. A nonindexed pension that began to be paid in 1970 would only buy 44% of the goods and services at the end of the decade that it could purchase at the beginning of the decade. The majority of the purchasing power - 56% of it to be exact - was lost to inflation.

The loss of purchasing power in the 1970s will not come as a surprise to older readers. But, bearing in mind the low inflation nature of the 1990s, the fact that 16% of purchasing power was lost in that decade may come as something of a surprise. Bear in mind, too, that a decade now represents little more than half of the average life expectancy of a 65-year-old Canadian, and the importance of inflation risk to older persons' purchasing power is clear.

In considering the well-being of older Canadians, the real incomes (incomes net of inflation) they have and how those incomes move through time is very important. Equally important is the question of how those incomes compare with the incomes of active workers. Are the incomes of retirees allowing them to share pretty much the same standard of living as younger people?

As was noted above, it is very common to form judgements about the well-being of people 65 and over by comparing their income to that of younger people. One thing about this comparison in the late 1980s and 1990s in Canada deserves far more attention than it gets. The average real rate of wage growth of the working-age population was virtually nonexistent. Table 20.3 highlights this point.

Like Table 20.2, Table 20.3 draws on data from the last five decades of the twentieth century. It shows how the value of a price-indexed pension would change over the course of each decade compared to average wages and salaries. In other words, a price-indexed pension that began to be paid in say 1960, would have the same purchasing power at the end of the decade, but it would only be worth 76% of what it was worth at the start of the decade compared to average wages and salaries. There was a 24% loss in relative purchasing power.

The measures in Table 20.3 are a direct reflection of real wage growth in each decade. The low numbers in the 1950s and 1960s reflect strong, real wage growth. The numbers for the 1980s and 1990s are a sad reflection of the lack of real wage growth over that period. If we think of expressing incomes of the elderly as a percentage of incomes of the non-elderly, the former were growing rapidly thanks to the maturation of the C/QPP and workplace pensions, and the latter were stagnating thanks to nonexistent, real wage growth.

One other condition made the 1980s and 1990s special with respect to the third pillar. That is the high rate of return on financial assets. This was largely, but not exclusively, a storey about the long bull run in the stock market from 1982 to the year 2000, and this commentary will focus on stock market returns. It is important to note however, that people who bought long-term bonds in the early part of this period also made unusually strong returns on them.

Stock markets across much of the globe participated in the long bull run. At the end of 2000, the Toronto Stock Exchange index stood at 7.53 times its level as of the end of 1982. In Canadian dollars, the US-based Standard and Poor's index stood at 19.49 times its 1982 level. To put the giddiness (and unreality) of these numbers into perspective, Canada's nominal GDP grew by a factor of 2.83, and corporate profits grew by a factor of 5.08. The fact that stock market returns outpaced corporate earnings over such a long period is indicative of the price/earnings ratios of stocks climbing to unprecedented levels and setting the stage for an inevitable collapse (Shiller 2000).

All parts of the third pillar benefited enormously from this development. The investment returns of defined benefit pension plans easily exceeded the rate assumed by plan actuaries when calculating plan liabilities. Pension surpluses became common place. The appropriate way to apply surpluses was often in dispute between employers and unionized plan members, and occasionally the disputes were settled in court. Nonetheless, the surpluses also financed a mix of contribution holidays by employers and benefit improvements for plan members. The surpluses were often used to facilitate special early retirement packages in downsizing situations. In some cases, they were also used to make ad hoc adjustments to the pension incomes in pay.

Higher returns on investments also made a direct contribution to individual savings arrangements. Exactly how these higher returns might manifest themselves (higher incomes with no change in retirement age, earlier retirement, or lower saving rate) is not clear. What is clear is that the circumstances of the 1980s and 1990s made it easier to achieve retirement income objectives through the third pillar.

Clearly, rates of return on financial assets declined precipitously in the period from mid-year 2000 to the early part of 2003. The effects of this will be noted below. For the moment, two points bear emphasis. First, to state the obvious, the strong growth in the absolute and relative incomes of older Canadians in the latter part of the twentieth century got a strong boost from particular features of the economic environment: low inflation, low real wage growth and high returns on financial assets. The same institutional arrangements cannot be expected to produce the same results in different circumstances. Second, the combination of population ageing and decelerating population growth that have well-known adverse effects on contribution rates to pension plans financed on a pay-go basis, will have similar, though less direct effects on contribution rates to workplace DB plans.

Financing problems of workplace pension plans

All Canadian jurisdictions impose financial requirements on workplace DB plans that are quite strict by international standards. Pension assets are to be separated from the assets of sponsoring employers, and are held in trust for the plan members. Moreover, once in not more than every three years, an actuarial valuation must be performed on the pension plan. A basic feature of these valuations is the preparation of two actuarial balance sheets in which the amount of assets held in the pension fund is compared to the liabilities of the plan; the liabilities are the value of the benefit promises made to the plan members.

One balance sheet is prepared on a 'going concern' basis meaning that it is assumed that the plan will remain in place indefinitely. The other is prepared on a 'solvency' basis meaning that it is assumed that the plan is going to be wound up on the effective date of the valuation. While there are several differences between the two types of balance sheets, it will suffice to note here that the solvency valuations are required to rely more heavily on current market values of assets. Solvency valuations also have to determine the value of liabilities using current interest rates in the market as opposed to interest rates that might reasonably be expected over longer periods of time.

If going concern liabilities exceed assets, an actuarial deficit (unfunded liability) results that has to be amortized (paid off) over no more than fifteen years. If solvency liabilities exceed solvency assets, a solvency deficiency is said to exist and it must be amortized over no more than five years. The overriding object of these financing exercises is to create a situation in which the employer sponsoring the pension plan can go bankrupt, and there will be sufficient assets in the pension fund to pay the benefits promised to the plan members.

The glow was barely off the celebration of the new millennium when workplace DB balance sheets were engulfed in a perfect financial storm. The stock market bubble burst in the spring of 2000 and continued to fall through the spring of 2003. The asset side of workplace DB balance sheets took a beating. Trusted pension plan assets of just over $600,000 billion fell by more than $60 billion between mid-year 2000 and the end of 2003 despite an increase in workplace pension contributions from $12 billion in the year 2000 to $19 billion in 2002 (Statistics Canada 2003).

Although it got far less attention, the liability side of the balance sheets was also taking a beating. The yields on long-term Government of Canada bonds fell over the 10 years from the end of 1995 through the end of 2004, from 9.4% to 4.9%. To put this decline in context, a one percentage point drop in a DB pension discount rate will typically increase current service costs by 20%.⁷

The stock market decline and falling long-term bond yields caused pension balance sheets that had been running up surpluses in a relatively easy manner through the 1980s and 1990s to swing hard in the deficit direction. DB plans in which contribution holidays had been taken shortly before, suddenly required full contributions to cover newly accruing benefits and special payments to get rid of actuarial deficits. Employer contributions to workplace DB pension plans shot up from $6.4 billion in 2000 to $11.6 billion in 2002 (Statistics Canada 2003).

Not surprisingly, many employers and pension consultants expressed a great deal of pain in the face of the DB financing problems. These problems provided an additional spur to the shift from DB to DC that had been underway for some time, and that will be discussed below.

At the same time, the damage done to pension balance sheets produced another type of problem. Companies came to the brink of bankruptcy with substantial unfunded liabilities - Air Canada and Stelco being the most prominent cases in point. In these cases, it would have been the plan members rather than the employers who would have felt the pain as benefits would have had to be cut to match the benefit payments with the available assets.

The fact that both types of problems arose (increased contributions and threatened benefit cuts) as a result of the damage done to balance sheets, highlights the dilemma in resolving DB financing issues. The measures that hold out the greatest promise for reducing the volatility of pension contributions are likely to accentuate the risks to plan members that assets will not be sufficient to cover liabilities in the event of bankruptcy.

In the case of DB pension plans, important institutional players (employers, unions, regulators, and to some degree, politicians) found themselves in the centre of the storm and spoke on it, as they should. Far less attention focussed on the reality that the change in financial markets that produced the DB crisis was affecting DC plans as well. However, in the DC world, there was an important difference: it was individual plan members who were taking it on the chin. Asset accumulations were contracting and annuity prices were going up.

Given the mistreatment of this issue in so much popular discussion, including that of pension professionals, it is worth noting that the negative impact of changes in financial markets does have second order effects in both DB and DC plans.

In the case of DB plans, the first order effect is that employers have to make special payments to pension plans (there is a small number of plans in which the special payments are made by both employers and active plan members). But, there may be second order effects that shift some or all of the burden to plan members through measures such as: pension benefits improvements that were planned may not be made; wage and salary increases might be smaller than otherwise planned; or, non-pension benefits may be curtailed in some way. Strangely, there is no systematic evidence on who actually bears risk in DB plans despite all of the strong statements one hears on the subject.

In the case of DC plans, the impact of poor performance is conceived of in terms of lower benefit amounts per month for the plan members. Clearly though, the impact could also be manifest in delayed retirement.⁸ In an extreme case of poor performance, an employer might even feel pressure to top up low pensions which would entail the secondary shifting of risk from the plan members to the employer.

Declining coverage of workplace pensions⁹

Following a long period of steady increase in the portion of employed Canadians covered by workplace pension plans, a reversal in this process began in the mid-1980s and carried on into the 1990s. Thus in 1991, 45.4% of employed Canadians were members of workplace pension plans, and in 2001, only 40.1% were members of workplace pension plans. Between 1992 and 2002, the absolute number of people who belong to workplace pension plans grew by 2.9%, but the employed labour force grew by 15.2%.

For purposes of estimating the impact on future retirement incomes, it is important to know what portion of employed people belong to workplace pension plans. However, it is also worth noting that the actual number of workplace pensions declined by 26% between 1992 and 2002. The fact that the decline in the number of plans significantly exceeds the decline in plan membership suggests that the part of the workplace pension population that shrunk the most was smaller plans. To some extent, the decline in small plans may have been offset by the growth of group RRSPs that do not show up in data on pension coverage. Some evidence on this point is provided by Frenken and Maser (1992). However, it appears that the use of RRSPs levelled off in the mid-1990s and is not offsetting a decline in workplace pension coverage (Palameta 2003).

Given what was noted above about the different role of workplace pension income in the incomes of older men versus older women, it is worth noting that the longstanding gap in workplace pension coverage between employed men and employed women has been largely eliminated in recent years. In 1991, 49% of employed men and 41% of employed women belonged to workplace pension plans. By 2001, the comparable number were 41% and 39%. Unfortunately, this equality was created not by a levelling up of women, but by a levelling down of men. As Morissette and Drolet have noted, there has been a sharp drop in pension coverage among young men aged 25 to 34 (Morissette and Drolet 1999).

A word of caution does have to be offered on this gender equalizing process. The benefits to which people become entitled in retirement reflect not only pension coverage (and earnings) at a particular moment in time, but on pension coverage through time. Since workplace pensions typically fall somewhat short of providing full portability, and women change jobs more frequently than men, it may be harder for women to maintain continuity in pension coverage through time. Because of their longer life expectancy, women retirees are also somewhat more vulnerable to the effects of inflation on nonindexed pensions.

Differences in pension coverage between men and women is just one of many social and economic variables with respect to which pension coverage varies. This variability is important in thinking about the overall decline in pension coverage just noted. Two extreme possibilities present themselves: there was a uniform decline across all social and economic variables, or there was no decline within social and economic groups, but the prominence of different groups in the employed labour force changed so as to bring about the decline (i.e. groups with traditionally low coverage increased in importance).

In their discussion of declining pension coverage of younger men, Morissette and Drolet draw attention to the importance of sectoral shifts in employment and declining unionization in explaining lower pension coverage.¹⁰

Shift from DB to DC

Running parallel to the general decline in pension coverage just discussed has been a shift in coverage from DB to DC plans. The significance of this shift rests in the fact that a larger share of the future elderly will have their workplace pension incomes more directly exposed to investment risks at or around the date of retirement. Moreover, as was noted above, the risks may be reflected in either or both of income uncertainty, or uncertainty about the age of retirement.

The shift in the balance of coverage from DB to DC has not been dramatic, but it has been steady for some years. Thus over the period from 1992 to 2002, the percentage of plan members who belong to DC plans increased from 9% to 15%. This trend was more pronounced in the corporate sector than elsewhere; DC membership as a percentage of plan membership in the corporate sector increased from 12% to 21%. It should be noted however, that these data might understate the shift to DC as group RRSPs that don't show up in the data seem to have grown in prominence in recent years (Frenken and Maser 1992).

DC plans have always held some attraction for small employers thanks to the certainty of the employer's financial commitment under DC plans and their administrative simplicity. Some longer term trends have added to these traditional considerations. Among the longer term influences are: the increasing complexity of DB plans that has resulted from tax and regulatory change; shift in employment to non-standard work including part-time work; young worker resistence to DB plans based on their weak record in dealing with termination of employment before retirement; generally slack labour markets; the influence of employer preferences in the US and the UK where the shift to DC has been more pronounced; and, general policy level support for DC coming from institutions like the World Bank and the OECD.

Two further influences of more recent origin are also worth noting. Conflicts over DB surplus that have gone to court have generally been resolved in favour of plan members. This has led to a sense of grievance in some quarters that employers who sponsor DB plans face downside risk with no chance of capturing upside gains; in the rhetoric of the day, they face asymmetric risk. This view leaves no room for the secondary shifting of risk noted above. Also, the financial problems faced by DB plans that were noted above have added to the attractiveness of DC to some employers while simultaneously accentuating their disadvantage for plan members.

While the trend to DC is clear and has many motivations, it is also a slow continuous trend, not an avalanche or a tidal wave. Moreover, there are some clear mitigating influences, not the least of which is union and plan member pressure. Also, to the extent that a pension plan is viewed as a means of attracting and retaining workers, a DC plan is less likely to be attractive to the growing number of older workers than is a DB plan. In addition, pension benefit levels and retirement ages might both become more unpredictable. The first order effects of this will be felt by plan members, but there may be second order effects that will be difficult for employers such as people working later or retiring earlier than suits employers' purposes, or facing strong moral pressure to make up for poor pensions.

There is also a question whether DC plans will remain as simple administratively in the future as they have in the past. Canada's Joint Forum of Financial Market Regulators (a body that brings together pension, insurance and securities regulators) has recently issued guidelines on the operation of capital accumulation plans, by which the regulators mean any form of savings plan that employers operate for employees. The guidelines suggest that the providers of DC plans have significant responsibilities to plan members with respect to the range of investment choices that are available, and with respect to plan members' access to appropriate counselling (Joint Forum of Financial Market Regulators 2004).

As the term guideline suggests, these are voluntary. But, they are not totally lacking in legal significance, as compliance (or otherwise) with the guidelines will likely be an issue in any civil litigation. It is noteworthy, too, that the guidelines, unlike regulatory law, are relevant to both group RRSPs and DC pension plans.

Overall assessment of the challenges in the third pillar and conclusions

Workplace pensions play an important role in providing income to older Canadians. For Canadians 65 and over, whose situation has been the focal point of this paper, workplace pensions lift people from adequacy to comfort. Although it has not been documented above, workplace pensions also play an important role in facilitating retirement prior to age 65. Relative to C/QPP and OAS, workplace pensions tend to pay out more of their income prior to age 65.

It is equally clear that workplace pensions currently face severe pressures. This is especially true of the DB variety of workplace pension plan. Given the role that workplace pensions play in the retirement income system, their overall contraction is less likely to manifest itself in more older Canadians with incomes below low income measures, than in fewer people stating that their standard of living is as good as it was during their employment. It will also be reflected in fewer people being able to retire comfortably before age 65.

There has been a tendency in some quarters to declare that DB pension plans are a thing of the past, at least in the private sector. This is clearly a huge exaggeration, although it does accurately capture a trend. That said, it is also worth noting that hybrids, that have long existed in the form of multi-employer pension plans and DC plans with DB guarantees, are now emerging in new forms. There are a number of quite large pension plans in the near public sector that have formal arrangements for sharing special payments, within limits, between employers and plan members. In addition, the Quebec Federation of Labour has called for the creation of employee-run DB plans under which plan members would make all special payments.

Clearly, the third pillar has been weakened, and there is a more broadly based vulnerability of retirement income in Canada than one would infer from policy discourse. Cries of calamity would overstate the case. But, the complacency about the adequacy of retirement incomes that dominates official policy discourse understates the case. In context, it is worth noting that two recent studies that have taken totally different approaches from each other and from this paper have reached the conclusion that as many as one-third of the future elderly may have inadequate incomes. Thus Schellenberg (2004) found that one-third of people aged 45 to 59 who were surveyed in the 2002 General Social Survey foresaw having inadequate income in retirement, and the same conclusion was reached in an analysis of the data on 45- to 64-year-olds derived from Statistics Canada's Survey of Financial Security in 2001 (Maser and Dufour 2001). Then again, as was noted at the outset, this is an issue that deserves some formal modelling.

Bibliography

Baldwin, Bob and Pierre Laliberté. 1999. Incomes of Older Canadians: Amounts and Sources, 1973-1996. Research Paper No.15. Ottawa. Canadian Labour Congress.

Canadian Institute of Actuaries. 2005. Canadian Economic Statistics 1924-2004.

Coile, Courtney and Phillip Levine. 2005. Bulls, Bears, and Retirement Behaviour. Working Paper No. 10779. Washington. National Bureau of Economic Research.

Department of Finance. 1996a. An Information Paper for Consultations on the Canada Pension Plan. Ottawa.

Department of Finance. 1996b. The Senior's Benefit: Securing the Future. Ottawa.

Department of Finance. 1984. Action Plan on Pension Reform: Building Better Pensions for Canadians. Ottawa.

Eschtruth, Andrew D. and Jonathan Gemus. 2002. Are Older Workers Responding to Bear Markets? Boston. Center for Research on Retirement, Boston College.

Frenken, Hubert and Karen Maser. 1992. "Employer sponsored pension plans - who is covered?" Perspectives on Labour and Income. 4, 4, Winter. Statistics Canada Catalogue no. 75-001-XIE.

Gardner, Jonathan and Mike Orszag. 2003. How Have Older Workers Responded to Scary Markets? Watson Wyatt Technical Paper 2003-LS05.

Jackson, Andrew. 2005. Work and Labour in Canada: Critical Issues. Toronto. Canadian Scholars' Press Inc.

Joint Forum of Financial Market Regulators. 2004. Guidelines for Capital Accumulation Plans (CAP Plans).

LaMarsh, Judy. 1968. Memoirs of a Bird in a Gilded Cage. Toronto. McClelland and Stewart.

Lipsett, Brenda and Mark Reesor. 1997. Employer Sponsored Pension Plans - Who Benefits? Working paper W-97-2E. Ottawa. Human Resources Development Canada.

Maser, Karen and Thomas Dufour. 2001. The Assets and Debts of Canadians: Focus on Private Pension Savings. Catalogue no. 13-596-XIE. Ottawa. Statistics Canada.

Morissette, René and Marie Drolet. 1999. The Evolution of Pension Coverage of Young and Prime-Aged Workers in Canada. Analytical Studies Branch Research Paper Series. Catalogue no. 11F0019MIE1999138. Ottawa. Statistics Canada.

Myles, John. 2000. The Maturation of Canada's Retirement Income System: Income Levels, Income Inequality and Low Income Among the Elderly. Ottawa and Tallahassee. Statistics Canada and Florida State University.

Organisation for Economic Co-operation and Development (OECD). 2001. Ageing and Income: Financial Resources and Retirement in Nine OECD Countries. Paris.

Organisation for Economic Co-operation and Development (OECD). 2000. Reforms for an Ageing Society. Paris.

Organisation for Economic Co-operation and Development (OECD). 1998. Maintaining Prosperity in an Ageing Society. Paris.

Palameta, Boris. 2003. "Profiling RRSP Contributors." Perspectives on Labour and Income. 15, 1, Spring. Statistics Canada Catalogue no. 75-001-XIE.

Régie des rentes du Québec. 2003. Adapting the Pension Plan to Quebec's New Realities. Québec.

Schellenberg, Grant. 2004. The Retirement Plans and Expectations of Non-Retired Canadians Aged 45 to 59. Analytical Studies Branch Research Paper Series. Catalogue no. 11F0019MIE2004223. Ottawa. Statistics Canada.

---

Notes

1. The term workplace pension plan is ussed throughout the text to refer to pension plans established by individual employers in both the public and private sectors for their employees, and it also refers to plans established for employees of particular groups of employers (e.g., municipla employee plans or plans for carpenters in a particular region). The term does not include group RRSPs, nor does it include Supplementary Executive Retirement Plans. It is, in effect, a synonym for the term Registered Pension Plan.
2. In the World Bank typology, the first pillar is made of pensions financed from general government revenues that may be offered on a univeral flat rate basis or on an income or means-tested basis. Canada's Old Age Security and Guaranteed Income Supplement are first-pillar programs in teh World Bank typology. The second pillar is mandatory programs designed to replace pre-retirement earnings, and may take the form of mandatory defined benefit plans like the Canada and Quebec pension plans, or mandatory individual savings accounts. The Bank has become well known for promoting mandatory individual savings accoutns in recent years. The third pillar is made up of group or individual arrangements that are not required by law, and may be inititiated by employers, unions or individuals.
3. In recent years, regulatory law has permitted some alternatives to monthly annuity payments for both DB and DC plans as have been available for payouts from RRSPs for some years.
4. Income data in this section and in the section below titled "Role of the third pillar" come from special tabulations of data from the Statistics Canada's Survey of Labour and Income Dynamics and the Survey of Consumer Finances. On historical tends, see also: Baldwin and Laliberté (1999) and Myles (2000).
5. Claims about retirees' perceptions of their well-being are based on special tabulations of data from Statistics Canada's General Social Survey (2002)
6. Fully retired households are households with no earnings from employment. The equivalence scales used in calculating household incomes is the OECD new equivalence scale tht assigns a value of 1.0 to the first member of the household, 0.5 to the second member, and 0.3 to additional members.
7. Although performance of the stock market has improved since the early part of 2003, the balance sheets of DB pension plans are still suffering thanks to the need to overcome the losses of the early years of the decade and, more importantly, due to the persistence of low interest rates. A silver lining in all of this is that the financing difficulties of DB pension plans have provoked a thorough debate on the proper means of financing them.
8. Although there is some debate on the matter, it has been argued that increased labour force participation by the 55- to 64-year-old age group in the US and UK since the year 2000 is attributable to wealth lost in the stock market during that period. See: Eschtruth and Gemus (2003), Gardner and Orszag (2003), and Coile and Levine (2005).
9. Data on pension coverage are from Statistics Canada 2003 as are data referred to in the discussion below of the "Shift from DB to DC."
10. A thorough assessment of the social and economic characteristics of workplace pension plan members is provided in Lipsett and Reesor (1997).