At this moment, there is no formal proposal. There is a plan created through talks between a large number of Ontario universities. These talks reached an impasse. Recently, the University of Toronto with a smaller number of other institutions (Guelph and Queens) became interested in reviving the outcome of the talks. These narrower talks are supposed to start where the wider talks ended.
On this website, we refer to the last project of the talks as the new plan. More formally, the project is known as the University Pension Project. The most recent publicly available document describing the details of the New Plan can be found here: UPP plan. (OCUFA resources has some more, if outdated, materials.)
The new plan is multi-employer, jointly-sponsored, it has defined-benefit, and it will take over the liabilities for the old plan. How the latter issue is resolved, is an important consideration.
The current discussion about UPP plan involve three universities: the UofT, Queens, and Guelph. One thing that these universities have in common is that most of their unionized staff is represented by the same union (USW - Steelworkers). As you can see in the Table below, the current UofT plan is larger than the other two plans combined.
Table: Plan members and retirees at involved universities (2014-2015)
The next Table explains some potential arguments for and against a multi-university plan. We try to list all arguments that one can think of, for and against. Below, we discuss whether we think that our circumstances make some of these arguments more or less valid.
Table: Arguments for and against multi-employer plan.
In the new plan, multiple-employers would have to deal with multiple unions, with no employer and no union having a dominant power (UTFA has pretty much the same numbers as all members of Steelworkers across the three universities). The decisions would be made by a board of plan members and they would be taken in the interest of the plan and its members. The decision rules must be transparent to avoid conflicts and suspicions between members. Compare it with our current situation, where decisions about the plan are often reached through collective bargaining and on somewhat ad-hoc basis, where the interest of the plan and its members often is shared with negotiations about wages, or other terms and conditions, etc.
A similar argument applies when the plan is jointly sponsored. In such a case, the decisions must be jointly made by a board with equal numbers of the representatives of employees and employers. The goal of the board is to act in the interest of the plan and its members. If there is a disagreement, the plan rules describe the process in which the decision is going to be reached.
There is an argument that larger pension funds are able to invest in a different, perhaps more efficient way than smaller pension funds. For example, some infrastructure investments might be more possible only when there are enough money. Also, larger funds may cut on costs because of possible economies of scale. In fact, many of the successful Ontario plans are very large (for instance, OTPP has more than $100 bln assets.) However, we are not aware of research that supports such an argument.
This argument does not apply to our case. The University investment body, UTAM currently manages $8 bln in assets from the UofT pension plan plus the UofT endowment. The joint value of the 3-university pension plan would be slightly smaller than that (it would include $4 bln of the UofT pension assets, but not its endowment.) So, the combined pension fund would not have a larger funds that those that are managed under the current plan.
An argument for multi-employer plans is that they are more stable and more secure from the point of view of the employee. Even if one of the employee goes bankrupt, the other employers ensure that the plan continues to exist and operate.
Because the universities do not go bankrupt, we do not believe that the argument is very important in our case. Moreover, if any of the universities would go bankrupt first, most likely, it would not be the UofT. In such a case, it would be the other institutions benefiting from the extra insurance provided by the UofT, not the other way. We believe that the value of this extra insurance is not significant.
This is potentially a big problem. As we have already explained, because the defined benefit formula does not tie the pension to contributions (see Pension formula), the current plan has redistributional consequences. When the plans are combined, there is a potential for a similar redistribution that occurs between universities, not only within an university.
Because there is no actuarial analysis that focuses on this issue, we do not know how much redistribution can happen and in which direction. However, there are at least demographic and wage factors three factors that point in the same direction. Namely that there will be a systematic transfer from the UofT to other universities:
Age: Our member population is slightly younger than the population of other universities. Younger members are cheaper for the plan.
Retirement patterns: Our members tend to retire slightly later. Late retirements are cheaper for the plan because they collect pensions for fewer years and they contribute longer.
Wage profile: Our faculty earns more than faculty at the other universities (the UofT is much richer). That means that our salaries tend to hit the Cap much earlier, and as a result, our pensionable wage profile is much flatter. As we explained in redistributional consequences, flatter wage profiles tend to subsidize pensions of the steeper wage profiles.
The pension plan is chosen for a long time, so we need to think not only about the present conditions, but also about future situation. There is an argument to be made that the differences between the University of Toronto and other Ontario universities are likely to increase. The UofT is the leading Ontario university, in terms of teaching and research, we benefit from name recognition, higher research funding, rankings, popularity among foreign students, etc. It is unlikely that our position will change in the future. Indeed, it is likely that our position relative to other universities will only improve. (There is a discussion among the provincial government about future of Ontario universities and the tenor of the discussion is that the universities differ and that the differences should be further strengthened.)
These future developments may force the universities to adapt. For an example, imagine that a lower ranking university faced with decreasing enrollment and research funding, may want to reduce the benefit package offered to its employees. One way to adapt is to change pension plan.
Under a single-employer plan, an university is relatively flexible to adjust to a new situation. Under multi-employer plan, any adjustment must be negotiated with multiple other partners. The decision taking will be shared among multiple universities. However, because what is good for one university, does not have to be optimal for another, the decision-making may lead to inefficiencies. In particular, the UofT may be forced to reduce its benefit package only because it is required by the financial condition of the other members of the plan.
Although the new plan describes conditions under which the university could leave it, such exit is always costly and it is an adequate solution to the flexibility problem.
Given that steeper profiles receive higher pensions (see above and redistributional consequences), each member would like to have the wage profile as steep as possible. More precisely, choosing between two wage profiles that lead to the same lifetime income, a member would prefer a profile that is steeper, because the latter leads to higher pensions and lower contributions. The employer of this member may go along and design wage schemes that are steeper than it would be otherwise.
To see how such a manipulation can happen, imagine that during the collective bargaining at the University X, the local union or faculty association considers two types of wage increases: (a) merit-based or (b) seniority-based. (That is a real-life choice. UTFA asks this question every single bargaining round.) The merit-based raises tend to go to younger, more active individuals. The seniority-based go to older employees. The latter are more attractive from the point of view of pensions, even if the former are more attractive from the point of view of providing incentives for research.
Suppose that the University X chooses merit-based raises, and University Y goes for seniority-based ones. The relatively steeper wage profiles at Y would mean a systematic transfer from X to Y.
The issue is pertinent to multi-employer plans and it does not arise with a single employer. The reason is that a single-employer does not benefit from gaming the wages against pensions as both come from the same pocket. In multi-employer plans, the wages come out from the university pocket directly, but the cost of pensions is shared with all the other universities. Hence, the university has incentive to manipulate the wages of its employees to increase their pensions, because at least part of the cost of the latter will be born by other universities.
Is there an evidence that this problem is real? Yes, there is and the problem is substantial according to a recent study that just came out in the J. of Public Economics. (In short, unionized teacher's wages in U.S. spike in few years before the retirement. This can be attributed to the design of the pension formula, the fact that wages are negotiated and paid by districts, but pensions are paid out of state-wide pension plans.)
In a jointly-sponsored plan, employees and employers share equally
cost,
risk, and
governance.
Some of the most successful pension funds in the world are Ontario jointly sponsored pension plans. Examples include OTPP, Hoopp, OMERS.
The next Table explains some potential arguments for and against a jointly sponsored plan. We try to list all arguments that we can find; below, we discuss the validity of each of the arguments.
Table: Arguments for and against jointly-sponsored plan.
Cost sharing means that the employees and employers contribute half of the service cost of the plan. (Recall that the service cost of the plan is the amount of money that the plan needs in any given year to fund liabilities that are generated in any given year. In other words, each year that we work, we earn some pension benefit (see Pension formula). The service cost of the plan is the amount of money that we need to pay to the plan to finance this benefit.)
Risk-sharing means that, additionally, any deficit of the plan has to be corrected by an increase in the contributions, reduction of benefits, or removing of the indexation from the current pensions. The cost of correcting the deficit is split equally between employers and employees. Similarly, any surplus may lead to a reduction in contributions or improvement in the benefits.
To see how it works in practice, imagine that instead of our current plan, we had a jointly sponsored plan.
- The service cost of our current plan is 20% of the salary. The cost sharing of JSPP means that the employees would contribute 10% of their salaries to pensions.
- The current special payments that go to cover the going concern deficit are equal to 46% of the service cost (the special payments are equal to $79 mln and the total value of the service cost is $173 mln). If our plan was jointly sponsored, the risk-sharing would raise employees contributions from 10% to 14.6% of their salaries. That would be 4.6% more of salary for exactly the same benefit. The extra contribution would be required to cover the deficit.
See risk comparisons to get a better sense about risk imposed on employees according to their age..
The plan is managed by the sponsor board half of the members of which represent employers and the other half represents the employees. No decision about the plan can occur without the agreement of the both sides of the board.
In the single-employer single-sponsor pension plans, the employer is responsible for the deficit (section 75, Ontario Pension Benefit Act). If our case, the University of Toronto would have to finance it from its own assets: endowment, perhaps sale of buildings, etc. Thus, the only threat to the benefits comes from the bankruptcy of the employer, i.e., if the assets of the University are not enough to cover the liabilities of the plan. (In reality, the protection is even stronger than that. It is unlikely that the government would stand idly by if the University were to sell buildings. More likely, the government would step in and bail us out.)
However, pension protection under JSPP is significantly weaker - the benefits under JSPP are secure only as long as the plan exists (see exceptions in Section 75 of PBA). In the case of wind-up of the JSPP (see below), the benefits accrued under the new plan can be limited even if the employers continue operating and are in a good financial standing.
That information is considered sufficiently important that it is the regulator requires members to be specifically told about such a possibility before they give their consent to the conversion to the jointly sponsored plan - see Ratification.
It is important to point out that the reduced protection applies only to the benefits acquired under the new plan. Subsequent regulation says the employer is on hook for the solvency deficit of the old plan liabilities in case of the new plan wind-up (see Special contributions on wind-up of JSPP). Thus, all the benefits acquired under the current plan are secure as long as the UofT is not bankrupt.
Wind-up
If the JSPP winds up with a deficit (and the only reason for winding up is that there is a substantial deficit), the earned benefits are reduced proportionally to the fraction of the plan that remains funded. For instance, suppose that the plan has 20% deficit. Then, the pension benefits would be reduced to 80% of what they were supposed to be.
There are two ways that the JSPP can end up winding up:
- by the decision of the administrator (i.e., sponsor board). The decision to wind up the plan belongs to the administrator of the plan. The founding documents can specify another entity to do so. Importantly, if the decision is made by the sponsor board, the employee groups have to consulted. See below for the discussion of the incentives, or who would like to wind up the plan.
- by the Superintendent of Ontario pension plans. The Ontario Pension Benefit Act as a list of possibly situations in which the Superintendent may force the wind-up decision on the plan. The first of them is that a significant fraction of the employers stop paying contributions. That effectively may mean that the employers can unilaterally force the Superintendent hand.
To see why would a JSPP wind-up, we need to consider incentives. First of all, nobody wants to wind up a plan with a surplus -as everybody would like to participate in this surplus. On the other hand, if the plan has deficit, this deficit has to be paid. One way to avoid paying the deficit is to wind-up the plan.
- If you are an employer, and the deficit is significant, you may want to wind-up.
- If you are an employee who just become a member , you have no pensions earned, and you expect that your future contributions will be used to pay off the deficit. You will want to wind up.
- If you are a current retiree, or an employee who is close to retirement, you don't care about the deficit (you will not pay any more contributions), however you wil be greatly affected by the pension reduction in case of wind-up. You don't want to wind up.
- If you are an employee in the middle of your career, the decision depends on the amount of the deficit (that will be covered from your future contributions) and the amount of pensions that you earned (which would be reduced under wind up). The decision will depend on where you loose more.
Volatile contributions and benefits
Because the plan is required to amortize the deficits, and 1/2 of the cost of the amortization will fall on the employees, the contribution rates wil vary from year to year. The degree of volatility will depend on the mechanism of deficit corrections. Additionally, the sponsor board (the managing body of the plan) may decide to change future benefits (the benefits that have been earned cannot be changed as long as the plan exists).
To get some idea of how volatile the contributions can be, think about this. Our current (going concern) deficit is $573 mln. Imagine that instead of our current plan, we are under the jointly sponsored plan, and we are responsible for amortizing this deficit. Currently, we pay around $79 mln a year extra. Because our current service cost is $173 mln (see Actuarial evaluation of the University of Toronto pension plan), we would need to raise our contributions by 46%! (Just to be clear what it means: if our current contribution rates are 10%, then the amortization would raise them to 14.6%.)
To see how he contribution rates changed in other jointly sponsored plans, see historical contribution rates at OTPP.
One way to reduce the volatility of contribution rates is to make sure that the plan has a sufficient surplus that serves as a financial cushion (see below). surplus instead of raising contributions. The surplus will be replenished in
The financing formula of many typically includes room for a special reserve fund. In bad times, the plan can draw on this extra surplus instead of raising contributions. The surplus will be replenished in better times.
The current design of the UPP assumes that a small fraction of our contributions will go to this reserve fund. The exact value is going to be determined in the negotiations, but it is likely that it will be equal to 0.5% of our salaries, which makes it equal to 2.5% of the service cost. Over years, this extra contribution will create an extra funding surplus. The amount of the surplus can be seen on the graph below.
The vertical axis measures the size of the funding surplus relative to the value of liabilities. The horizontal axis counts the number of years since inception. The two lines compare the surplus for two different designs: UPP (red line) and clean JSPP (blue line). The latter is the same plan as JSPP, but with no transfer of old plan liabilities (see risk comparisons). The two lines diverge at 15 years of inception, which is when the transfer of the old plan liabilities is supposed to happen (see below).
There are two things to keep in mind when choosing the fraction of the contributions that goes to special surplus:
- The current amount of 5% seems small. As you can see on the graph, after 30 years, even without transfer of the liabilities, the size of the extra surplus is not going to be larger than 7% of the total value of liabilities, and considerably smaller if the old plan liabilities are transferred. Many JSPP (like OTPP, or CATT) are much more aggressive in building up rainy day fund, with a (more or less explicit) goal of 25%. If we wanted to reach the same level in 30 years, the fraction of extra contributions would have to be of order of 15%.
- This extra contribution is a direct transfer from the current generations of employees to the future ones. The larger the fraction, the more the current generations pay for the surplus that is going to benefit future generations.
- In the same time, the majority of risk induced by the change of the plan will fall on the future generations see risk comparisons). The extra payment can be treated as a way of compensating future generations for this risk.
In order to encourage the UofT participation, the new plan benefits were designed to be as close as possible to the current plan. In particular, the structure of the key pension benefit is the same: 1.6% times CPP and 2% above that but below the Cap times the serviceable years (see Pension Formula).
The other benefits may differ from the current plan. Indexation becomes conditional (except for the old plan benefits????) - see below. It is possible that the new plan will have to choose between early retirement and survivor benefits because keeping both of them will be too expensive. Early retirement option: TBA
The current plan offers indexation at 75% of the CPI inflation. In some of the past years, indexation was increased to 100% as a result of the collective bargaining.
The UPP plan does not guarantee indexation. The indexation at 75% is going to be funded, in the sense that the contributions are chosen so to cover indexation at 75% of the inflation. However, if the plan is under-funded (i.e., it has a deficit), the sponsor board of the plan may decide to remove indexation. The rationale for making indexation conditional is that it is the only way, in which the retirees can participate in the risk of future investments (the majority of the risk falls on the employers and employees)
We do not know what happens to the indexation of the benefits under the old plan. If they are not transferred to the new plan, then clearly the old plan rules apply and the indexation at 75% is guaranteed. However, if the old plan rules will apply to the new plan benefits, it will lead to a situation, in which the new plan will be responsible for two types of benefits: some with guaranteed indexation, and some with conditional one.
Our current early retirement benefit allows for retirement with unreduced pensions at age 60 (see more details about current benefits). It is likely that early retirement is much more expensive. At this moment, we do not know the relative costs of these two benefits.
Few points to keep in mind when we think which benefit to keep in our plan.
- Early retirement without reduced pension benefit is quite generous. To see it, notice that if you decide to retire one year before 65, you are going to lose 2% of your pension (see Pension formula). In the same time, you are going to collect your pensions for 5% more time (the life expectancy at 65 is around 85, hence 1 year of 20 makes it 5%) more time. Thus, one year of early retirement leads to approx. 5%-2%=3% more in pensions as compared to somebody who retire on time for each year of early retirement. You will also save on the contributions to the plan that you were supposed to provide in the last year.
- The benefit will become more expensive as the life expectancy is increasing.
- Canada Pension Plan (CPP) does not have early retirement without reduced pension. In fact, pension is reduced 7.2% for each early retirement year, which effectively means that somebody choosing to retire early receives roughly the same amount of pensions as without early retirement. In fact, CPP increases pensions by 8.4% for each year of retiring late (up to age 71) .
- There is an argument that early retirement benefit is good for workplace satisfaction, as it allows co-workers who are not longer happy (hence, less productive) leave early. However, there is no reason why not to keep the early retirement out of the multi-university plan, and, instead, negotiate it locally with the employer. Because workplace satisfaction is in the university interest, they will be more receptive to add early retirement incentives to their local benefit package. In the same time, keeping a costly (with a cost that is only going to grow) benefit out of the plan design, can make it more flexible to adjust to future challenges, and also differences between university.
- Survivor benefit is more valuable to men, because they live, on average, shorter than women. In the same time, men live shorter, on average, hence they collect less in pensions.
- We do have precise information about the use patterns of early retirement and survivor benefits.
The treatment of old plan benefits in UPP plan is unfair to the new plan members as it leaves them with a part of the risk for the old plan liabilities. The exact degree of the risk is unknown, but the ballpark estimate suggests it to be around a 30% of the current going concern value of the old plan liabilities.
Before we describe the issue and how the old plan benefits are treated under the UPP plan , we describe the problem that the treatment is attempting to solve.
At the commencement of the new plan, each of the universities has old plans with earned benefits (including retirees and current members who are still working but have already earned some benefits under the old plan). These benefits are liabilities for the old plan.
The earned benefits have to be fulfilled.
The universities do not want to run old pension plans separately from the new plan.
The universities could transfer all their assets to the new plan and make the new plan responsible for the old plan benefits, i.e., transfer the old plan liabilities.
But that would be wrong and unfair for at least two reasons:
Problem 1: Each of the multiple employers had different past plans and faces different scales of deficits. For the sake of example, suppose that Guelph does not have any deficit and the University of Toronto does. If the University does not transfer enough assets to cover their deficit, then the new plan members from Guelph will have to pay with their contributions for the deficits that were incurred by the UofT. It would not be fair to Guelph.
Problem 2: However, there is another important reason. Suppose that that at the commencement, the all employers transfer assets that are equal exactly to the present value of liabilities computed according to the going concern method. That would solve Problem 1. However, that would also leave the new plan with all the risk of providing old plan benefits when going forward. To see why, imagine that the next year the investment return would not perform as well as expected. In such a case, the new plan members would have to cover the deficit with their own contributions.
That is not fair to the future members of the new plan. The reason is that the old plan liabilities where incurred under the assumption that the University is solely responsible for them. There is no reason why the new plan members should take over the risk that the University is responsible.
Notice that the transfer of assets equal to the present value of liabilities computed under hypothetical wind up scenario would be OK. In such a case, the new plan Administrator could simply use the assets to buy a risk-less investments (Treasury bonds) and use them to pay out all the future liabilities coming out from the old plan. There would be no old plan liabilities risk for the new plan members. Of course, if the University could make a transfer equal to the solvency deficit, then they would be also able to amortize the current solvency, and we would not have any of these discussions.
Here is how UPP plan describes the treatment of the old plan benefits:
At the commencement, the Universities will transfer all its assets to the new plan. Additionally, it will promise to cover the Going Concern deficit in 15 years. You can think about it that it will divide the Going Concern deficit installments and pay back equal amount in each year. (More precisely, it will do something like we do when we pay back 15-year mortgage in equal installments.) This will take care of Problem 1.
In the first year after the commencement, the actuary will re-calculate the going concern deficit deficit for the old plan liabilities. If there is a surplus, it will be used to reduce the number of payments for the original deficit from the previous point. If there is a deficit, the University will pay it back in equal installments in the next 15 years (so this payment will end in the 16th year after commencement).
The procedure described in (2) will repeat year after year for the first 15 years. After that, there will be no more re-calculations and adjustment and (except for the payments that are still flowing that were generated in the previous point) no more new payments. From now on, all the old plan liabilities are on the shoulders of the new plan members.
Although this procedure seems complicated, it has a very simple result. At the end of the 15th year of running the new plan, there will be effectively no going concern deficit deficit with respect to the old plan liabilities. That means that Problem 2 is not going to matter for the first 15 years. At the same time, at the beginning of the 16th year, Problem 2 comes back: all the old plan liabilities become responsibility of the new plan and its members.
The UPP plan essentially shits Problem 2 15 years into the future. Is it fair? No. But how much should we care? That depends on whether you are young (it means, you have more than 15 years to the retirement) or older. If the latter, this issue will affect you to a lesser degree. (One reason why a senior member could be affected is because of conditional indexation. Another is that the new plan may want to create some cushion to prepare for the extra risk at 16 year and raise contribution rates earlier.)
It also depends on how much of the old plan liabilities will still be around in 15 years. The truth is we don’t know exactly because we do not have access to precise actuarial modelling. However, the model that we used to evaluate risk comparisons of different pension plans suggests that there will be around 56% of liabilities left. See the graph.
The employees will be responsible for half of the associated risk.
There are solutions
One would be to extend 15 years to something longer and make a careful actuarial analysis of how many of the liabilities will still remain.
Another would be to make the University amortize the solvency deficit instead of the going concern deficit. Of course, the University is not able to do it over short period of time. (That is why we are where we are.) But it should be possible and easier for them to amortize the solvency deficit over 15 years (in the same way that it will be impossible for us to pay our mortgage in 5 years, but it is often possible, if difficult to do it over 15 years).