Multi-Employer Pension Plans (interview with James Pierlot)

The transcript is edited for clarity and brevity.

[RK] Hello everyone! My name is Roman Kosarenko. I oversee pension investments at a large Canadian corporation. In my spare time, I write articles on LinkedIn. But today, I wanted to do something different. Today, I offer you a video interview with a person who is very knowledgeable about multi-employer pension plans (MEPPs).

Let me introduce James Pierlot, the founder and the Chief Executive Officer of Blue Pier. Blue Pier is multi-employer pension plan of a unique design. James is a member of the multi-employer pension plan advisory at the Financial Services Regulatory Authority of Ontario.

The topic of today's conversation is not Blue Pier as such, but MEPPs in general — their evolution, their place in the industry, and what their version of the retirement plan of the future could be.

First things first — I have no personal connection to James or his plan in any way, and neither of us is getting paid for this interview in any way. The views expressed here by me are mine and mine only, and don't necessarily reflect the views of my employer.

So, welcome James! And thank you for agreeing to take part in this experiment.

Since our topic today is multi-employer pension plans, I looked up some official data. According to StatCan, the number of active members of MEPPs was just over 1 million in 2019, which is almost the same as it was in 1992. We know that pre-2007 data was not measured in the same way, but we don't have any alternative facts to consider either. Are multi-employer pension plans a growing part of the retirement industry, or it is indeed a stagnant component for some structural reasons?

[JP] Thanks, Roman, for the introduction and for setting up this interview! It's always fun to talk about multi-employer pension plans and pension coverage in general, which obviously is an interest of mine.

Pension plan coverage has been in a secular decline for more than 30 years. The statistics show essentially what you say: membership in pension plans and multi-employer pension plans, particularly in the private sector, has been stagnant. What are the reasons for that? MEPPs have thrived where there is some large organization, such as a union or a government, which has the resources to establish them and a direct business interest and offers that to a constituency — for example, the union members, the employees etc.

In the public sector, we have seen growth in the membership of public sector pension plans in step with the growth of employment in the public sector. But in the private sector, union membership has been stagnant or declining, and pension plan membership has also been stagnant or declining fully in step with union membership.

There is really no entity historically, outside of unions or government, which has had the interest in establishing a multi-employer pension plan that is independent of the employers, independent of the members and which can welcome any employer. There's a lot of sunk costs in doing that. Other than Blue Pier, the only example that I am aware of where an attempt to do that was made back in 2016 was the Ontario Retirement Pension Plan, which never got established.

[RK] Multi-employer pension plans have evolved to be industry-specific. That brings demographic challenges into picture. We know, for example, that Ontario Teachers' Pension Plan has to fight the demographic wave of retiring plan members. Longevity can be industry-specific as well — there are known differences between, say, blue collar and white collar positions. It seems that the stability of a MEPP depends on the diversity of its membership. That, however, increases complexity and may lead to hidden risks to participating employers. How broad or diverse a pension plan of the future should be?

[JP] I see the challenge with MEPPs as not so much one of having diversity within the plan as having a proper demographic balance between the younger and the older members. The baby boom generation is a very large cohort. As it ages and moves through the system, there aren't nearly as many active members of pension plans as there are retired members. For Ontario Teachers', the ratio is about 2-3 active members for every retiree, and it used to be maybe 15.

Where a pension plan pools longevity risk, demographically diverse membership — different levels of education, collar type etc. — leads to different life expectancies at retirement among those groups. If they're all paying the same level of contributions then some are going to be subsidizing others. To take an extreme example, a coal miner may live seven years less than an investment banker, and if you put both in the same plan, one is going to be subsidizing the other due to pooling of longevity risk.

For MEPPs, the issue is figuring out how to put diverse groups into a plan and get the benefits of economy of scale and pooling, but without having these subsidies happening under the hood, that people don't know about and perhaps wouldn't like if they knew about them.

Some kinds of subsidies are inevitable because there will always be variances in at-retirement life expectancy, even within demographically similar groups. And that's fine, because that's kind of random. But where you have material demographic differences within the same pool — just because the CPP pools everyone together doesn't mean that every other plan should do it.

[RK] I want to contrast some of the aspects of the MEPP model with the single-employer pension plans, which is a subject I know very well. From my perspective, the line of responsibility is always very clear with single-employer pension plans. It is always clear who is the first defendant in a potential lawsuit.

Multi-employer pension plans are very diverse, and their governance can be quite complicated, even to regulators. Some MEPPs own their service providers. The interlocking relationship among unions, their plans and their service providers is difficult to tackle. How important is good governance for a pension plan in general? Is it a significant challenge for existing multi-employer pension plans?

[JP] I would challenge the notion that the lines of responsibility in a single-employer pension plan are always quite so clear. There is never a single person who is responsible for everything. In a large company, the administrator of the plan is usually a board of directors or people on the human resources and compensation committee. Those people are a revolving door: directors come — directors go, committee members come — committee members go.

If the board of directors is administering a pension plan, it serves two masters and has conflicting responsibilities. On one hand, they have to maximize shareholder value, being agents of the shareholders. On the other hand, they are supposed to be a fiduciary, taking care of the interests of the pension plan members. That creates a structural conflict with the interests of the shareholders.

In a 1995 case involving Imperial Oil, the Supreme Court of Canada articulated that as a "two hats" situation. When the board of directors of a company is serving the pension plan members' interests, they have to put on their fiduciary hat in service of the members, and when they're serving the shareholders, they have to put on their fiduciary hat in service of the shareholders.

For the most part, employers manage that conflict fairly well. However, when things go wrong with that model, they go very wrong, and we've seen some spectacular examples of failure.

In a multi-employer pension plan, of course, you have the separation of the sponsor and the administrator roles. The sponsor is the person who establishes the plan to serve some business interest: attract and retain employees etc. The administrator is the fiduciary, responsible for statutory compliance, benefit delivery, proper administration, investment management. The administrator is a board of trustees, the sponsor is the employer — everybody knows what their job is.

That's of course in an ideal world. In some multi-employer pension plans, trustees may have other allegiances because they were appointed, say, by a union, or they have an implicit representation of an employer group. That presents an issue, and it has been an issue in certain MEPPs, as I'm sure you are aware.

Another good example would be the jointly sponsored pension plans (JSPPs) in the public sector. They have more of an independent expert board of trustees, and the sponsor function is explicitly separated from that. The board of trustees function is really to ensure sustainability of the plan over time.

I think the two key drivers of good pension outcomes in pension management are good governance and scale. Good governance involves having an expert board of trustees with specific and relevant subject matter expertise. It could be actuarial science, investments, governance, law, accounting… The board of trustees cannot be experts in everything, but they have to have knowledge of something that is relevant to the operation of a pension plan. And they have to be independent of both the members and the employers. Their fundamental accountability needs to be toward the constituency that matters, which is the plan members.

[RK] In the industry press, the MEPP model is often presented as the best of both worlds: to a plan member it can be a defined benefit plan, and to an employer it looks like a defined contribution plan. It is also a trusteed plan, where the people who oversee the plan must act in the best interest of the members. Yet, somehow, MEPPs are hardly a breakaway success.

An RSP plan is arguably an inferior product since it's legally an individual contract with a for-profit financial institution, and there is no one acting in the best interest of the subscriber. RSPs have tax disadvantages to employers and management expense disadvantages to subscribers. And yet, RSP asset growth outpaces the asset growth of trusteed plans, including public plans. Why do you think the model, which is fundamentally more sound, is not more successful?

[JP] I'm going to take issue with the idea that pension plans, whether a MEPP or any other plan, can provide certainty of contributions and certainty of benefits. Something must be volatile in the basic pension equation, where contributions plus investment returns equals benefit payments plus expenses. You can never really say that there's no risk to the employer or no risk to members, because you're talking about entities with very long-term time horizons. Over time, pension risks are really on a continuum.

Legally, for example, in a specified Ontario multi-employer pension plan (SOMEPP), you might have the employer contributions limited to a fixed amount set in a collective agreement or on a per hour basis etc. But if there is a funding difficulty in that plan, then there is an implicit call on one or more of the participating employers. In the collective bargaining process, the union might say: "Well, you know, we've got a problem with the pension plan, and we need you to help out with that." Even though the employer has no legal obligation to do so, it might be a very bad business decision not to negotiate on that point.

So there is never certainty of contribution and certainty of benefit. No one can make promises. A pension promise is going to be paid or not paid to the extent that there's some kind of asset backing it. That asset could be an actual financial asset sitting in a pension fund, or it could be a call on the assets of an employer.

There's also the fact that you don't know what the liability actually was for any given individual until that person passes away.

I'm kind of curious where you get the RSP and MEPP numbers. The Statistics Canada data that I look at show that in the trusteed plan sector, which is largely MEPPs in the public sector, or multi-employer-like plans, you've got assets of close to 2 trillion dollars. From 2014 to 2018, those have grown at about 17%, which is double the rate of what StatCan reports for the RSP data, about 8%. That seems to suggest that the model is successful, in the public sector at least, and is gathering more assets.

If you compare to 1992, there was a much higher rate of pension coverage in the private sector at the time, which potentially absorbs some of the savings that would otherwise have gone to RSPs.

[RK] One of the perceived advantages of the MEPPs is that the onerous solvency requirements don't apply to them — either explicitly, as in specified Ontario MEPPs, or implicitly they should be balanced through some combination of negotiated contribution increases or possibly benefit reductions. Yet, benefit reductions rarely, if ever, happen in practice, and a sufficient increase in contributions is difficult to negotiate when multiple employers bargain using different schedules. What is your take on the contingent funding liabilities that are implied in those negotiated contribution DB plans? Can you have your cake and eat it too?

[JP] These days, we are seeing a transition across the country and among pension regulators from a solvency funding regime to a going concern or a going concern plus funding regime with provisions for adverse deviation. It leads to a funding security level that's on some midpoint between solvency and going concern.

If you go back far enough, everything that's old is new again. It used to be the case that plans were only funded on a going concern basis. Then there were big plan failures, and the regulators and policymakers came along and said: "This doesn't make sense. We need to have proper funding of pension plans." And they implemented the solvency rules, where a plan is supposed to have enough money so that, if the sponsor goes bankrupt at any given point, there will be enough to settle out the benefits by way of commuted value transfers, annuity purchases etc.

That definitely improved benefit security and outcomes for members. But then interest rates declined and investment returns in the portfolios declined. The declining interest rates and the increasing life expectancy drove up the liabilities, so people said: "Solvency is too expensive!" Now we want to go back to going concern and going concern plus because over the long term everything will be fine. If we get the rate of return that is suggested by the discount rate we've chosen and our historical rates of return…

Everything does work out in the long term in that scenario. The problem is with the average lifespan of the entities sponsoring pension plans, which is under 10 years for a company or a business in Canada, whereas pension plans are very long time horizon entities. What if a sponsoring employer disappears and there's a solvency deficit? Is some other employer going to be asked to step in and fill in that gap?

It's not a popular view these days, but solvency makes sense as a funding paradigm, because it really gets you closer to full settlement of the obligation on a risk-free basis.

Some people will say: "Well, that will throw up the surplus." There will be arguments about surplus at some points. That is true, for sure there will. But I would argue that, just because you have to deal with the upside and sharing the upside, is not a good reason not to have upside, nor to structure a funding regime, which over time creates a probability that pension benefits will be underfunded, and people will not get what they were promised.

[RK] One of the apparent trends in Canada is the conversion of defined benefit plans to a jointly sponsored plan model, where the funding responsibility is shared and no longer subject to collective bargaining. In one characteristic example, the unionized members of the Ontario's Workplace Safety Insurance Board pension plan gave up their right to bargain over pension contributions in order to permanently block either a defined contribution plan or a target benefit plan. Is the JSPP model just so much better, more robust? Are there any weaknesses in that?

[JP] In a single-employer pension plan, which is defined benefit, the employer is ultimately the guarantor of the benefits and if there is some kind of a deficit, the employer has to pay it off. In a JSPP, the word "sponsor" means taking funding risk, having some kind of upside and downside. If the plan has a surplus, the employer and the employee contributions will both go up and both go down, commensurately with the funded ratio of the plan. The allocation is typically 50/50, so what a JSPP does is it takes half the funding risk away from the employer and transfers it to the members as a group.

There are a number of advantages to that model. The employers have less funding risk, which they obviously like. The transfer of funding risk to members makes them take a greater interest in the operations of the plan and perhaps to be more realistic in terms of the benefits that they expect to get out of it. If they want more pension, they are going to have a direct impact on their paycheck. That leads to a tension and perhaps more realistic expectations.

In a JSPP, unlike other types of multi-employer pension plans, it's not possible to reduce the nominal amount of the pension benefit paid to a retiree, which makes JSPPs similar to single-employer pension plans. That puts the retirees into a special, privileged category. If there is some kind of problem with the funding of the plan, the administrator can't call on them to share in the pain, to keep the structure sustainable in the interests of everyone.

Some JSPPs have addressed that by introducing conditional inflation protection. If there's a problem with the funding of the plan, then they cannot provide an inflationary increase for pensioners, and that makes pensioners participants in the funding of the plan.

For a demographically homogeneous group then, the JSPP can make a lot of sense, if the employers are willing to accept the funding risks associated with the JSPP. Potential problems may arise from demographic differences within the longevity risk pool.

[RK] What you are talking about is generational fairness. In a JSPP, current employees take the risk, but those who are retired don't. Should generational fairness be built into a MEPP by design?

[JP] Sustainability of a MEPP, or any pension plan, should not trade on information asymmetries. If people knew that there were persistent subsidies within a pension structure, then they probably wouldn't be supportive of it. If they don't know about it, and that's the only reason why it's working, then that is a problem from a systemic, structural point of view.

Retirees can have risk and still have adequate pensions and good and happy retirements. A couple of years ago, the federal government introduced the idea of a Variable Payment Life Annuity (VPLA). A traditional annuity with fixed and equal periodic payments necessitates a matching risk-free asset, which of course has a very low rate of return. Instead, in a VPLA, the members can participate in the upside of the capital markets in retirement as they did during the accumulation phase, and have a larger pension, even though the pension amount might fluctuate from year to year. Risk is not necessarily a bad thing, even for retirees, because it presents an opportunity of getting greater value over time.

I think that a lot of this comes down to a communication exercise. You need to make sure that if you promise someone a pension you should be delivering it. But maybe the promise needs to be contextualized a bit, to say: "This is the target pension that we're going for, but we can't guarantee that it will always be that amount. It might be more, it might be less, but it should be within this range."

It's about probabilities, not certainties, because when you promise people certain things about the future, you're doing something that you cannot do, because no one can predict the future.

[RK] Regardless of the funding model, Canadian MEPPs, with the sole exception of Quebec, are subject to the same commuted value standards. These standards specify the value of accrued benefit that the terminated member can opt to take out of the plan. The discount rate applied is substantially lower than the going-concern discount rate, so that creates asymmetries, again. What is your position on commuted value? As an embedded put option, is it unfair to the remaining plan members?

[JP] I think it's clear that if you value a pension for one purpose for the people who are in the plan by discounting it at a relatively high rate (5-6%), and then the people who leave the plan get a bigger amount for their promised pension benefit, because you're discounting it at a lower rate, you have advantaged that person to the cost of everyone else. As they take money out, the funded ratio of the plan for the remaining members drops below what it was for them.

From a purely fairness point of view, I'm not sure why we would have a system, where we give a commuted value to a person at a different discount rate, or using a different valuation method, from the value for the people who remain in the plan.

[RK] James you are a lawyer by trade, and you studied law in Montreal, in Toronto and also in France. There are certain parallels between Canadian and European pension challenges, as both sides need to deal with regulatory fragmentation. It's interesting that the recently announced Pan-European Pension Product (PEPP) is an individual savings vehicle, somewhat similar to Canadian RSP or an IRA in the US. If you were to speculate, could the Canadian MEPP model evolve to permit individual membership?

[JP] Absolutely. I've long advocated for individual membership in a pension plan.

It's a curious aspect of the Canadian tax rules for retirement saving going back more than 100 years that you cannot be in a pension plan unless you have an employer who is willing to contribute to it on your behalf. You can't even use your own money to fund a pension in many situations because of 50/50 funding rules.

We are in a situation where employers don't want to set up pension plans, because of the costs and risks of operating them. In the private sector, 70% of all workers are employed by companies that have fewer than 100 employees. They just can't set up pension plans — it's too complex for them. So we end up with close to 70% of the workforce, more than 12 million people in Canada, without pensions.

If individual membership were permitted, I could join an established pension plan, and it could have participating employers in it and participating members. I could then go to a new job with a new employer, and part of my compensation deal might be: "I want you to put in 10% of my pay into this pension plan." And that plan would follow you wherever you go. A structure like that makes a lot of sense. It's not to say we don't need employers — we do need employers. Employers can stay part of the mix. It's not necessarily employers, but not necessarily no employers either.

Individual membership is certainly conceivable. If you are in a multi-employer pension plan, and you leave your employer, you no longer have a connection to that employer, but you leave your money in the MEPP — well, you are an individual member of the plan!

[RK] Portability is one of the great problems of our industry. In fact, the commuted value was originally meant to solve the portability problem. What is the role of multi-employer pension plans here? Can the MEPPs lessen the issue of portability?

[JP] Absolutely — the MEPP is the perfect vehicle for that, particularly if it was possible to have individual membership. The more employers are in it, the higher is the probability that when you move to a different employer that employer will be a participating employer of the MEPP, and so you won't have to change accounts at all. Obviously, it would have to be very big before all of the employers would be in there.

Individual membership takes care of the problem of portability to a large extent. People change jobs every 5 or 6 years, and they end up having these little pieces of pension all over the place. They need to figure out a way to amalgamate them into an RSP. Of course, if they're transferring from a DB plan then they will get a tax hit because of the prescribed transfer limits. So many Canadians end their career having worked for different employers with pension plans, or maybe a group RSP, or maybe a deferred profit sharing plan — and they've lost an opportunity to participate in a big pooled vehicle that operates in their interest.

The Financial Services Regulatory Authority of Ontario recently established a technical advisory committee on missing members. One of the outputs of that committee was a public consultation document which proposed that assets for missing pension plan members could potentially be amalgamated in larger structures. That is obviously an opportunity to take care of people who leave employers and aggregate the assets into a single vehicle that operates in their interest.

[RK] James, your plan can provide both defined benefit and defined contribution benefits, but, for the moment, DC benefits don't include any form of lifetime income. You mentioned the recently permitted Variable Payment Life Annuities (VPLA). Do you foresee changes in the way DC MEPPs distribute income to retired members?

[JP] A multi-employer pension plan is, compared to a single-employer plan, an ideal vehicle, if it has the scale to provide retirement income effectively to people. Ontario DC plans can already pay retirement income in the form of variable benefits, which are similar to benefits payable from a Registered Retirement Income Fund (RRIF). But those are not benefits guaranteed for life.

Before the VPLA, only certain grandfathered DC plans were entitled to pay annuity income to members. Once the VPLA becomes available and is fully implemented through federal and provincial statutory and regulatory changes, it offers a new option. The Variable Payment Life Annuity is like any other life pension except that it provides continued participation in the capital markets throughout retirement. Retirees can potentially get higher lifetime pension payments because there are still equities in the portfolio etc.

At the same time, the payment will fluctuate, and the retirees will be a part of a risk pool. They will get a guaranteed pension payment until they die, even though they don't know what the amount will be.

Based on many years of providing educational programs to pension plan members and professionals, I believe people will not voluntarily choose to purchase annuities, for the most part. But that doesn't mean that the VPLA can't be helpful, because pension plan members tend to accept defaults, such as default contribution rates. If you give them the opportunity to participate in the pension plan, they'll be unlikely to opt in, but if you put them in and say: "You can opt out" — they will be unlikely to opt out. Similarly, most pension plan members use the default investment in their DC plan.

If you default them into a choice of a Variable Payment Life Annuity and — similarly to the current rules when they leave a DB plan they can take a commuted value if they're more than 10 years away from retirement date — you could actually make the default a life annuity, and that would probably result in a significant uptake of that vehicle.

I also want to mention the Advanced Life Deferred Annuity (ALDA), because that is another new option alongside VLPA that the federal government has introduced. The ALDA would provide longevity insurance from, for example, age 85. That, combined with variable benefits, could give you something similar to a Variable Payment Life Annuity.

[RK] James, you mentioned that people would not voluntarily buy annuities. How would you explain the fact that in the Saskatchewan Pension Plan, which is one of the grandfathered plans permitted to issue annuities, about a half of the new retirees voluntarily choose in-plan annuities?

[JP] That's difficult for me to comment on, because I don't know the circumstances of that particular plan. A lot of it, I think, comes down to education. If the administrator of the pension plan has communicated to the members the advantages of annuities, and if that plan has a Variable Payment Life Annuity, and they've talked to the retirees, and the retirees are happy… Pension plan members talk to each other, so perhaps that's why.

In all of the discussions that I've had with pension plan members — and I'll admit that it's anecdotal evidence — they don't tend to want to buy annuities. They think: "If I purchase an annuity and if I will only live a couple of years after it starts, then the insurance company will take all the money." They don't put a high value on the insurance they get from an annuity. And then they have bequest motives etc.

Perhaps there is an example of a pension plan where members take out annuities, but it should be a 100%, not 50%, unless they have some special knowledge about their life expectancy. That brings into play the issue of adverse selection, or anti-selection risk. People are more likely to buy an annuity if they know that they're likely to live long — and less likely to buy one if they're not going to live long, or they think they won't. Insurance companies know that, they have to price annuities accordingly. That is one of the reasons why group annuity purchases are so much less costly than individual annuity purchases. With group annuity purchases you don't have to deal with anti-selection risk so opting into an annuity for that reason alone is rather problematic.

[RK] My last question for today is about pension regulation, which is rather notorious for its complexity. What would be the most important ways to reduce the regulatory burden on employers participating in in a multi-employer pension plan?

[JP] The first point I would note is that, compared to participating or establishing and operating a single-employer pension plan, the regulatory burden on employers in MEPPs is really light, and the risks are quite light. That doesn't mean that there are no regulatory obligations, but it's certainly a whole lot less work and risk for an employer to participate in a MEPP than it is to operate their own plan.

But that being said, some of the ways that regulators could help employers participating in MEPPs would be authorizing or allowing employers to have electronic-only communication, sign up, enrollment etc., providing safe harbor liability protection to employers who comply with enrollment, reporting and remitting requirements to make sure that they understand that they have no liability for the administrative side at all.

My favorite wish is to simplify the pension adjustment reporting requirements by having a reform of the tax rules so the accrual regime that we have now and the factor of 9 are changed to a lifetime accumulation allowance structure. The compliance would be dealt with by the plan custodians, not by the participating employers, and all of the pension adjustment reporting regime would go away.

I think there are opportunities to improve the employer experience in multi-employer pension plans. But that being said, it's a lot better than running your own plan.

[RK] Thank you very much, James, for your answers, they are definitely very deep. I will endeavor to publish a transcript of this video. Thank you for your participation today!

[JP] Thank you for the opportunity. It was a very enjoyable discussion!