April 07, 2018

Very misleading THE article comparing DB with DC pensions

An article about the USS pensions dispute in the Times Higher Education by David Voas (Let's defend pensions not defined benefits) gives a somewhat inaccurate account of the issue. It also makes a comparison between Defined Benefit and Defined Contribution pensions that is very misleading.

At the centre of the issue between the union and the employers is what kind of scheme it is. The employers want to end the guaranteed pension that is based on years of service and earnings, and replace it with a defined contribution arrangement where instead of a pension on retirement one gets a pot of money that depends only on what has been paid in and investment returns.

The article argues compares a young person aged 30 near the start of their career with someone nearing retirement, and concludes that the youngster would do a lot better under Defined Contribution. This is very misleading. For most people (almost all in fact) they would get a lot better penson in retirement from a Defined Benefit scheme.

The article actually makes a number of statements that are wide of the mark. And it is not sufficiently analytical.

For example it says, "defined benefit schemes are defunct unless underwritten by the taxpayer." That is not really true. There are schemes in the private sector that are not defunct. And there is no reason why the USS should be defunct without taxpayer support. The problem is with the way the regulations are being applied and the weakness of the employers. While it is true that most of the DB pensions schemes are in the public sector, such as the Teachers Pension Scheme that is normal in the post-92 universities, what Voas is getting at, I think, is a repetition of City group think that applies simplistic generalities.

His description of the problem, though, goes from vagueness to missing the whole point. He writes,"The problem is twofold. Employers struggle to afford the cost of guaranteeing that the benefits promised in 50 years' time will be paid ... Employees are also hit, though, because everyone ends up paying above the odds." (emphasis added) What does this mean? It sounds very bad. It is not factually true. There is no conclusive evidence that employers struggle to pay the pensions in 50 years. All that is required is that the scheme stays open until then and invests in high return assets such as company shares (equities).

The next paragraph is an explaination but is in fact a gross oversimplification to the point of being misleading, revealing that Voas has not understood the arguments: "The basic equation is simple enough: contributions plus investment income equals benefits. Contributions - and even benefits - are fairly predictable, albeit affected by changes in life expectancy. The challenge is to estimate real returns on investment over several decades."

That is not where the problem lies. There would actually be no problem if it were as Voas says. The USS investment portfolio has a high return, an average of 12 percent per year over the past five years, for example, and is expected to continue to do well into the future, since it is skewed towards equities.

The problem is that the liabilities (present value of future benefits) are being valued counterfactually. That is, NOT using the real returns on investments as described. If things were as Voas says, there would not likely be a problem. The problem is that the scheme is being told to assume very poor returns - index linked gilts have a negative return. So the problem is a result of the excessive prudence being forced on the scheme by the regulations and the weakness of the support from the employers.

In order to compare DB with DC, he contrasts two academics with the same salary, a woman aged 30 with an older man nearing retirement. They both receive a defined benefit of 1/75 of salary for a year's work. He claims that for the older man it is a windfall: "he receives substantilly more than his current contribution would support. For the 30 year old it is very poor value. She could conservatively expect her contributions to grow by 2 per cent per annum, thereby doubling her money by age 65. This fund would be worth more than the defined benefits." This is very difficult to make sense of. Why compare these two individuals in this way, by just taking a snapshot of one year. What pension would the young woman get when she retires? How would that compare with the man? Surely what is needed is to compare the pensions obtainable from DB and DC over a whole career, not just one year taken at different stages in life. It is also very superficial. We are not told if the returns are after inflation. All defined benefits are uprated for inflation by the CPI.

And so it goes on. It reads as if the article were simply intended to make the case for DC over DB. But that goes against all the evidence from academic studies and practical knowledge from actuaries. It also goes against the studies that have been done for the UCU by its actuarial advisers, which show that for the great majority of members, it is very considerably more expensive to provide a pension of a given annual amount by DC. Estimates have put this at between 40 percent and 100 percent more.

This piece is extremely poor scholarship.

- 2 comments by 0 or more people Not publicly viewable

  1. David Voas

    Perhaps I may be allowed to comment on the comment (which I’ve only just seen).

    My key claim is that defined benefit (DB) schemes are not inherently superior to defined contribution (DC) schemes: it all depends on accruals, contributions, investment returns, and so on. DB schemes may (if the employer is absolutely reliable) be superior in reducing risk, but reducing risk should not necessarily be the highest priority. If I give you the choice between £20 and an amount drawn from a uniform distribution between £18 and £27, the latter would be the sensible choice for most people.

    As accrual rates deteriorate and contributions increase, the statement that “For most people (almost all in fact) they would get a lot better pension in retirement from a Defined Benefit scheme” becomes increasingly untrue. DB is reasonably attractive in the current USS scheme. Under the proposals made by the UCU in January, never mind the agreement reached at ACAS in March and then rejected by union members, it wouldn’t be.

    I agree that recent pension legislation has been unhelpful and indeed counter-productive, hence the line in my article just before the one quoted: “while the operation was a success, the patient died.” I’m not signing up to “City group think”. I’m pointing out that IF we are indeed governed by this legislation, then the choice is between paying over the odds (by paying too much for too little, all the while making needless deficit recovery contributions) and switching to DC.

    There is no magic that makes DB better than DC – the same money is being invested by the same people (the USS) in potentially the same way. The difference is that the investment and longevity risks are borne by the employer with DB and by the employee in DC. That’s why any change needs to be accompanied by the payment of a risk premium, as I argued.

    Of course if we can change the regulatory system, so much the better. We don’t disagree about that. But given the recent history of scheme closures and the dearth of sympathy in government for universities, my guess is that the chances are close to nil.

    14 Apr 2018, 19:05

  2. David Voas

    Let me try to elaborate my argument about the two people, one aged 30 and the other nearing retirement. Let’s assume that each is paid £45,000. Under the current career average scheme, membership of USS for one year with an income of £45,000 entitles you to receive a pension of 45,000/75 = £600 during each year of retirement. (The scheme adjusts for CPI.) If they are retired for 20 years, then they each stand to receive 20×600 = £12,000, ignoring discounting, survivor’s pension, etc., plus a lump sum of 3×600.

    The employer plus employee contributions will be the same for each: 21% of salary, or £9,450. (Employer contributions are 18%, but a proportion is earmarked to fund the deficit, expenses, etc.; the contribution to the DB scheme is 13% from employers and 8% from employees.) The young academic will sacrifice income now for a benefit that begins in three and a half decades; the present value of that future benefit of £13,800 is less than half the nominal sum. The older person enjoys the same benefit almost immediately and hence receives full value.

    The old person receives a windfall from that year of service: the benefits accrued are worth more than the contributions can support. The young person loses out: the future benefits (once discounted) are worth much less than her contributions.

    You write that “Surely what is needed is to compare the pensions obtainable from DB and DC over a whole career.” Why should we assume that people spend their entire lives as academics? I was well into middle age when I entered academic life. Many young adults will work at universities for a while before moving on to better things. The point is that for this year (and for each year that the current scheme continues), academics like me who are approaching retirement will be adding to the scheme’s liabilities without making a corresponding contribution, while young adults are having their contributions locked away and invested for decades in return for a defined benefit with a present value that is inferior to the current payments. It might not matter too much if we could absolutely guarantee that the DB scheme will remain intact for the next three and a half decades, but that’s far from being the case. There’s every chance that they will overpay now (to support their elders), only for the DB scheme to end before they receive its redistributive advantages.

    14 Apr 2018, 19:06

Add a comment

You are not allowed to comment on this entry as it has restricted commenting permissions.

Blog archive


Search this blog

Not signed in
Sign in

Powered by BlogBuilder