While more than five million working Canadians rely on workplace pension plans for retirement security, the predominant form — a defined benefit (DB) plan — is in danger of extinction, at least in the private sector. Traditional DB plans are a non-starter for any organization that does not already sponsor such a plan.
And even though we have seen a massive migration to defined contribution (DC) pension plans, these are far from ideal since the transfer of risk to employees — who have minimal investment knowledge and an aversion to risk — will always be a problem. So is there a third way?
In the middle of the risk-sharing spectrum are hybrid plans, which include any plan with both DB and DC features. If employers no longer want to absorb all the risk, then surely plans that share the risk with workers is the way to go.
Harry Arthurs, chair of the Ontario Expert Commission on Pensions, stressed the importance of innovation in plan design in his 2008 report:
“How can our system of voluntary pension plans be sustained and enlarged while protecting the security of members’ entitlements and affordability for sponsors?... The achievement of these seemingly contradictory objectives is indeed possible but only if we can move beyond conventional understanding of DB and DC plans.”
However, the concepts of DB and DC are so deeply ingrained that plan designs falling outside these strict boundaries make people vaguely uncomfortable — especially the regulators who will have to rule on whether a given hybrid design is acceptable.
Some examples of hybrid plans illustrate the challenge.
The most common type of hybrid plan involves a core DB plan and supplementary DC plan. This is the arrangement Air Canada and its flight attendants union agreed to adopt in September 2011.
It can be argued this is not a true hybrid plan but rather two separate, conventional plans mashed together. A combination DB-DC plan is fraught with problems. It’s harder to administer, since it usually requires different service providers for the DB and DC portions. The cost of administration is nearly double since both a DB and a DC record need to be kept for each member.
The main drawback, however, is explaining it to members since the DB and DC pieces cannot easily be added together. One involves a lifetime annuity and the other an account balance, and any attempt to homogenize them is artificial.
Of course, such plans aren’t all bad. They share risk between the employer and employees, just not in the most elegant fashion. The other feature in their favour is existing pension and tax legislation already accommodates them.
Target benefit plans
Target benefit plans are the most talked about hybrid plans these days, even though they aren’t yet permitted in single-employer situations. A target benefit plan is like a DB plan as it pays a pension over an employee’s lifetime — the amount of which is expressed in a formula. It is also like a DC plan as employer contributions are fixed; hence, in the case of a funding deficit, benefits would be reduced rather than contributions being increased.
The other similarity to DC plans is the ultimate benefit depends on investment performance and, therefore, employees take on the investment risk. Unlike typical DC plans, members do not have investment options so the investment risk is pooled among all members. The longevity risk (the risk of outliving your retirement assets) is also pooled.
Another key difference is younger members essentially subsidize older members the same as they would in a DB plan. That’s because a dollar of benefit under the plan’s target formula is worth more for an older member than for a younger one. This is one of the biggest strengths — and biggest failings — of DB plans.
Target benefit plans exist only in multi-employer situations and generally involve industry-wide bargaining. As the provincial pension reform panels concluded, there is no reason they can’t exist in single-employer situations as well — but expect special conditions to be imposed on them, such as joint governance.
Employers should prefer target benefit plans to DB plans because contributions are fixed. They may also prefer them to DC plans because there are no investment options, so employers avoid the onerous task of educating employees on making their own investment decisions. The biggest challenge for employers is in managing expectations, since pensions can be reduced and this is not always clear to plan members.
Employees should prefer target benefit plans to DC plans since the investment risk and longevity risk are both shared among all members rather than borne individually.
Cash balance plans
The concept of cash balance plans surfaced in the United States in the mid-1980s. As with any DC plan, each member has an individual account and receives periodic statements showing an account balance. Each year, the account is credited with a fixed percentage of earnings, such as five per cent of base pay. Interest is then added to the account balance but, unlike regular DC plans, the interest credited is based on some external index — such as the consumer price index (CPI) or T-bills — rather than actual investment performance.
Hence, the employer still bears some investment risk. This is the DB element in cash balance plans. At retirement, the member’s account balance is payable in a lump sum or converted into a lifetime pension or annuity.
A cash balance plan differs from a target benefit plan in that an account balance is communicated rather than a DB pension amount. It differs from a DC plan in that members do not have investment options and less investment risk.
Employers will tend to prefer cash balance plans to DB plans because risk is reduced. They will also find these plans easier to administer than a DB plan or typical DC plan. Cash balance plans are also a non-confrontational way to eliminate early retirement subsidies that might have existed under a prior DB plan.
But they are not acceptable under Canadian pension legislation or tax law, so it will be interesting to see if eventual legislation will accommodate them.
Modified DB plan
The final example of a hybrid plan is something we will call a “modified DB plan.” This type of plan behaves exactly like a conventional DB plan except it would pay a lump sum rather than an annuity in the event of plan windup. Furthermore, that lump sum would be determined using a discount rate derived from the yields on prevailing corporate bonds (AA or possibly single A) rather than from government bonds, as is currently the case.
While this lump sum can be worth less than the benefit promise under a traditional DB plan, the difference is not that great in reality. Remember, DB plans can also have a deficit on windup that, in the case of insolvency, would usually cause pensions to be reduced.
Even in the midst of the recent financial crisis, the funded ratio under a modified DB plan would have remained close to 100 per cent. This is possible because of the choice of discount rate. Rather than trying to minimize volatility by matching assets to an immutable DB promise, this plan design changes the pension promise to match the assets.
From an employer’s perspective, the appeal is clear enough. Providing a DB pension eliminates the problem of educating plan members to make investment choices and the modified DB plan significantly reduces the problem of funding shocks, especially those that occur when workers are least able to make additional contributions, as in the midst of the last financial crisis. Finally, it dampens the volatility in pension expense, which has been the primary driver behind the decline in DB plans.
Certainly, employees would prefer a traditional DB plan but a modified DB plan is better than a target benefit plan which, in turn, is better than a DC plan for most employees — so this may be the ideal compromise. There is no way of knowing whether a modified DB plan will be acceptable under new legislation to accommodate hybrid plans.
Should a modified DB plan be an acceptable hybrid design? Provided that the benefit promise is clearly communicated, why not?
Fred Vettese is chief actuary at MorneauShepell in Toronto. He can be reached at (416) 383-6454 or firstname.lastname@example.org.