Misunderstanding public pensions, vol. 2

Andrew Coyne’s column in the National Post today is an unusually clear example of a mistake that people on the right always make when talking about public pensions. The headline in fact says it all: Turning the CPP into 18 million RRSPs. Here’s the basic problem with Coyne’s analysis. A public pension, like CPP, is providing an insurance product to Canadian citizens, in the same way that the health care system provides an insurance product to Canadians. Having insurance is not equivalent to having a pile of money in a savings account. And yet people on the right – merely because they don’t like government – are constantly suggesting that we abolish these insurance schemes, and replace them with individual savings – effectively forcing individuals to self-insure. So people like David Gratzer want to get rid of public health insurance, and replace it with individual savings accounts. And now Coyne wants to phase out public retirement insurance and replace it with individual savings accounts. Both of these proposals are non-starters, because of the enormous loss of welfare that would accompany abolition of the insurance scheme. There is, in other words, no economic or public policy case to be made for these proposals, it’s just pure anti-government ideology. Their line of reasoning winds up being something like the following: “if the private sector can’t deliver this kind of insurance, then no one should be able to have it.”

There is one key difference between the health care case and that of public pensions. Most proponents of health savings accounts recognize that they are recommending more than just privatization of a public good, they understand that they are also abolishing an insurance arrangement (and so they usually make some provision to provide a “back end” insurance scheme to support the savings accounts). In the case of pensions, however, proponents of “privatization” schemes such as the one Coyne is recommending often don’t realize that they are abolishing an insurance scheme, because they don’t realize that this is what public pensions are. (Actually, he recommends that only the expansion be diverted into individual funds, not the core system, so the headline is misleading. For simplicity I’ll ignore that quibble.)

The core insurance product being offered by the public pensions system is a life annuity. A life annuity is a slightly obscure product that can be purchased from an insurance company, which offers protection against the risk of outliving one’s retirement savings. The basic idea is simple: in return for a large upfront payment, the annuity then pays out a fixed monthly sum, from the age of retirement until death. This means that you cannot outlive your savings, and so you don’t have to worry about being 95 years old and running out of money. Anyone interested can go fiddle around with an annuity calculator, such as the one offered by RBC Insurance (here). A word of advice though – in the field where it asks you to provide your age, don’t put in a number less than 55. Why? Because they won’t sell annuities to anyone younger than 55. This is what we in the public policy business call a market failure – and an unusually clear instance of one at that. An entire category of consumers (individuals under age 55) are simply not being served by private markets.

The underlying business model is that insurance companies sell a lot of these annuities, some of which they make money on (because the person dies young) and some of which they lose money on (because the person lives for a long time), and the two balance each other out. Furthermore, if you sell a lot of them, then the law of large numbers allows you to calculate with considerable confidence exactly how much you need to have on hand, in order to make the monthly payments.

Why do people want this type of insurance? Because it generates a massive efficiency gain (those interested in more details can see sections 1 and 2 of this paper). Each of us has a non-zero chance of living to the age of 100. In principle, this means that each one of us should be saving enough to finance 35 years of retirement income. This is a giant sum of money, and none of us is actually going to do that. In practice, maybe 1% of us will actually make it to the age of 100. Lots of us will die younger. So what we should do is pool our retirement savings, so that instead of each of us saving enough to support each of us until 100, we collective save just enough for one of us to make it to 100. (Same with health insurance – instead of each of us saving enough to pay for kidney dialysis, because each of us has a non-zero chance of getting diabetes, we collectively save enough for kidney dialysis for the 2% of us who are likely to get diabetes.)

Anyhow, to make a long story short, life annuities are subject to market failure, largely because of adverse selection problems. As a result, they are often a better deal when purchased collectively. This is essentially what a defined benefit pension scheme is. I participate in one of those as a University of Toronto employee. Basically, I make an upfront payment every month, and in return, I get a fixed monthly payment from the age of retirement until death. Whenever the stock market goes up, there are always some doofuses (doofi?) in the faculty association here who want to convert to a defined contribution scheme (i.e. an individual savings account), so they can use some of their alpha to get those sweet, sweet returns. This is a transparently bad deal, since the two are simply not equivalent – and if you do the math, on how much it would cost to purchase private annuities versus how much it costs to get the defined benefit pension, the pension easily comes out ahead.

I know this because my wife is a surgeon, and so has to save for her own retirement. She’s always dragging me along to meetings with her fancy Bay St. investment advisers. Whenever they find out I have a defined benefit pension, reactions vary from “Oh my god, you’re so lucky!” to “Wow, that thing is gold” to “Okay then, you don’t need me.” Unfortunately, not everyone is so lucky. In fact, defined benefit pensions are now pretty much confined to the public sector, because historically they have been unfunded, or partially funded, which in the public sector is fine but in the private sector creates perverse incentives (namely, to run up pension obligations and then declare bankruptcy). So how do all those unfortunate people in the private sector get their life annuities? From the government… the CPP (QPP) is basically a defined benefit pension scheme, which means that it is essentially a publicly-provided life annuity.

So why might we want to expand the CPP? A typical retirement portfolio should contain a mixture of savings and insurance. Basically, you should have enough annuities to cover your basic needs – so that if you wind up spending 10 years in hospice care, lying on a hospital bed with Alzheimer’s, those bills will be covered. Savings should be used to supplement these annuities, to cover things like trips to Florida while you’re still young enough to travel and enjoy it. Yet if you look at the typical Canadian’s retirement portfolio, it doesn’t have nearly enough annuity income. Furthermore, the current working generation is facing very significant uncertainty about how long they will live – who knows what medical technology will do to extend life in the next few decades? Again, this pushes in the direction of expanding annuity income, which for the average Canadian, can only be achieved through CPP expansion.

Okay, so there’s a clear and cogent argument in favour of CPP expansion. What is Coyne’s argument against it? Other than confusion about what pensions are, he seems to be troubled by some vague suspicions about the CPP Investment Board. (As some have pointed out, the idea that individuals get better financial management on the funds in their private RRSPs is wishful thinking.) And he seems to be worried that the government will start using it like a sovereign wealth fund, and that this will impact returns. And he seems to be suspicious of the idea that we need to be accumulating funds at all, to meet the demographic challenge posed by the retirement of the baby boomers.

Alright, well if it’s the float that he doesn’t like, then there’s an easy solution to that – bring CPP back to pure PAYGO (for explanation of what that means, see section 3 of this paper). That will put the CPP Investment Board out of business overnight. Of course he’d never go for that. Why? Because it’s inefficient not having a funded scheme. Here’s my point: forcing people to self-insure is also inefficient. In order to make the case for his proposal, Coyne needs to show that no insurance is better than public insurance. This is a tough case to make, because public insurance has to be pretty badly managed before it becomes worse than no insurance (the same way pizza has to be really bad before it becomes worse than nothing.)

P.S. This point about pensions is one that I’ve made before, if anyone is interested in seeing a different version of the argument. My hope is that if I explain it enough times, enough different ways, eventually it will begin to seep in.


Misunderstanding public pensions, vol. 2 — 10 Comments

  1. The problem is what’s actually being insured against and when payouts happen, is so mismatched that any efficiency gained is outweighed by the inefficiency in the system. If CCP kicked in at the point of extreme physical infirmity and/or 85 with much higher benefits, then you could make the case that it is successfully insuring against the the non-zero risk that we out-live both our savings and our fitness, and it could probably be provided for a fraction of the current cost.
    The thing is that is not the current system, the current system is a weird semi-redistributionary abomination made for political expediency. In as far as money falls out of the narrow circumstances listed above, it is money wasted/reduced-in-value the great value of money, its flexibility and liberty of use, has been taken from the owner. Thus for the purpose of “saving for retirement” as opposed extreme old age/physical infimity (ie. what we mostly talk about when we talk about ccp), a savings acount is far more efficient than the ccp

    • > If CCP kicked in at the point of extreme physical infirmity and/or 85 with much higher benefits,

      CPP does this already. Would-be retirees can increase their current CPP payments by up to 40% by delaying their enrollment until the age of 70.

      • THe issue is that if there were no assumption that everyone would get it merely as a a retirement scheme, but rather as an insurance against outliving their savings/physical infirmity/being in hospice for decades then we could insure against what it’s actually insuring against for a fraction of the cost, and create a massive amount of value by letting people control their own savings and retirement.

        The assumption that the general population will just stop working even if they are mentally/physically fit, regardless of savings, just because they hit a magic age decades out of life expectancy is a massive drain on the economy, and efficiency.

        Right wingers aren’t mad that ccp is public insurance as opposed to private, they’re mad that its a bad system which is structured so as to be politically bulletproof. it doesn’t continue to function the way it does because it works well, it does so because old people vote like crazy to protect any potential threat to their interests.

  2. What about this article in today’s ROB:


    He makes two arguments against defined benefit plans: the company might go bankrupt and never pay out, and that if you switch jobs, you may only get contributions + interest… which I guess just means that you lose the insurance against living too long, unless the job you’re taking on comes with a defined benefit plan. To be fair, he’s officially only pointing out that defined benefit plans aren’t as risk-free as they are portrayed as being. But he doesn’t mention the basic insurance case for defined benefit plans. If he did, then the worries about corporate balance sheets and people switching jobs would lead him to support public defined benefit plans, i.e. CPP expansion.

    • Note the editor’s remark at the bottom of that article. On his first point, yes, that’s why they’re so uncommon now in the private sector. On the second point, typically if you’re only in the plan for a couple years, you get a payout that’s a locked-in RRSP equal to the sum that is calculated to generate the payout that the pension would have given you. If you’re in for longer you just get the pension, so some people draw pensions from multiple employers upon retirement. In any case, the author undermines his case by saying that, in the case of a payout, the defined benefit plan becomes just as bad as a defined contribution plan… which kind of tacitly acknowledges that the defined benefit plan is better.

      Depending on the structure, defined benefit plans may encourage employees to remain in one job (i.e. from the employer’s perspective, they encourage retention), because contributions usually get you a percentage of your final salary, and so if you quit early your salary has not peaked. On the other hand, these are usually capped (ours is), so by the time you’re mid career there is no longer any penalty for leaving (e.g. I’m well over the ceiling, so if I quit now, my pension would pay out the same amount for the past contributions I’ve made as if I kept working).

  3. Would a guaranteed annual income (comprehensive social insurance?) reduce the need for public pension plan (retirement social insurance)?

    • > Would a guaranteed annual income (comprehensive social insurance?) reduce the need for public pension plan (retirement social insurance)?

      Canada already has a type of guaranteed income for seniors, in the form of OAS and GIS. Together, they do a reasonable job at preventing seniors from experiencing absolute (abject) poverty.

      Pension plans have a greater goal of preventing relative poverty, where the retirement standard of living falls unacceptably compared to that during one’s working career. This sort of poverty is not generally amenable to basic income approaches due to the tremendous transfers required.

      • Agree that OAS/GIS form a guaranteed income for seniors. But, if we were to expand that to all Canadians, would that lessen the need for CPP as well?

  4. I know we all feel like we are shouting into the void most of the time, so let me say that when I saw Coyne’s column my first thought was to your arguments on exactly this point (from ‘the efficient society’ perhaps?) which I then passed on to the few people I know who care about such things. Winning!