Michael Otsuka is Professor of Philosophy at the London School of Economics. He has also served as the pensions officer for the LSE’s UCU branch. This article originally appeared on Wonkhe.com, the home of higher education policy, people and politics; http://wonkhe.com/blogs/would-a-shift-from-bonds-to-growth-assets-keep-the-uss-afloat/
This article is published in full below
Last week, two key documents entered the public domain that cast light on the current triennial valuation of the USS. On Tuesday, the University of Sheffield publicly posted the 58-page proposed Actuarial Valuation that USS had released to employers at the beginning of the month for consultation. On Friday, the Universities and Colleges Union (UCU), which formally represents the interests of scheme members, posted a 16-page response by their actuarial advisors at First Actuarial.
USS maintains that, to continue to provide the same level of pension benefits, contributions would need to increase from the current 26% (18% employer plus 8% employee) of salaries to a much higher rate of 32.6%. As 18% and 8% are widely regarded as the upper limits of what employers and employees are willing to pay, the implication of USS’s proposed valuation is deep cuts in defined benefit pensions.
First Actuarial presents a very different picture. They argue that USS can prudently continue to provide the current level of benefits without increasing contributions at all. In other words: keep calm and carry on.
What accounts for this difference in perspective? To answer this question, I need to say a few words about an actuarial valuation of a pension scheme.
The valuation serves two purposes: (1) to establish whether the £60 bn of assets in the pension fund as of the 31 March 2017 valuation date are sufficient to pay for defined benefit pensions that have been promised up to that date (past promises), and (2) to set the level of employer and employee contributions that will be required going forward to make and later fulfil new pensions promises (future promises).
Under the current scheme rules, annual employer and employee contributions on all salaries below £55,550 give rise to the following promise: lifetime pension payments in retirement worth 1/75th of each year’s salary, revalued for inflation.
The difference between USS and First Actuarial is rooted in the fact that contributions made during the working lives of scheme members need to be invested and to achieve positive real returns, to cover these promises in retirement. If they are simply deposited into an interest-free savings account, they will fall short.
First Actuarial calculates that contributions for past promises need to be invested, so they grow by 1.36% per annum in real terms (i.e., beyond the CPI rate of inflation) to reach the ‘break even’ point of covering all past promises. This is similar to the required rate of return that Con Keating reports as having been told that USS itself has calculated.
First Actuarial also calculates that contributions going forward for future promises will need to grow at a somewhat greater rate—by 1.85% per annum in real terms—to reach the break-even point of covering those new promises. Since this break-even rate is 36% higher than the break-even rate for past promises, this figure is corroborated by USS’s statement that the cost of providing future promises is now about 35% greater than the cost of providing past promises.
So far, therefore, USS and First Actuarial are on the same page: they are in agreement regarding the investment returns that are required to meet both past and future pension promises.
They are, however, in deep disagreement over investment strategy and choice.
USS describes its current portfolio as “broadly half in equities, one-third in bonds and the balance in infrastructure, property and other assets.” First Actuarial maintains that the pension fund should remain at least as heavily invested as it now is in growth assets such as equity and property. They also note that the “best estimate expected return on UK and overseas equities and property is well above [the higher break-even rate of] CPI + 1.85%. The USS in-house team, the scheme actuary, the actuaries advising UUK and ourselves all agree on this.”
The best estimate is one that the scheme will meet or exceed 50% of the time. The regulations, however, require that pensions promises be funded on a prudent basis, involving a greater than 50% chance of success. In line with others, USS construes prudence as requiring a two-in-three, or 67%, chance of success.
USS maintains that its current portfolio has a 67% chance of achieving no more than 0.53% below CPI per annum (i.e., negative real returns) over the next ten years. Such a pessimistic forecast is largely down to the currently extremely low expected return on bonds, in which USS is now one-third invested.
USS believes, however, that economic conditions will have returned to a more normal equilibrium by the end of the coming decade, at which point bond yields will have reverted to their higher 2014 levels. From that point and for two further decades, the current portfolio will have a 67% chance of achieving a return of 2.8% above CPI per annum. If therefore, we average over these three decades, the current portfolio has a 67% chance of achieving returns of 1.69% above CPI per annum.
This is slightly below the higher break-even rate of CPI + 1.85% for future pension promises.
First Actuarial would maintain that there is a simple solution to this problem: rebalance the portfolio so that it is somewhat more heavily weighted towards equity and property and other growth assets. There are various shifts in that direction that would, even by the lights of USS’s investment team, have a prudently adjusted 67% rather than 50% of meeting all future promises.
USS would probably respond that even a two-in-three chance of success isn’t prudent enough for a portfolio so weighted towards equities and property. This is because the returns on such growth assets are more volatile and variable than the returns on bonds. Hence a one-third chance of failure might involve catastrophic downsides that would be unacceptable in the case of equities and property but not bonds.
First Actuarial has a rejoinder to this challenge, which appeals to the cash flow analysis of the scheme that they develop in their paper. Their view is that for an open, ongoing, scheme such as USS—uniquely grounded in an enduring multi-employer sector of well-established and often venerable universities—the risk of investment in equities and property is minimal. This is because incoming contributions plus a modest level of investment income from growth assets will be sufficient to meet all pension promises in any given year, for at least the next 50 years. Hence the scheme is protected against the disinvestment risk of having to sell equities or property at an inopportune time. Since one can hold these assets over the long run, one can be confident of being able, over decades, to reap returns that comfortably exceed the break-even rates.
USS offers a very different approach to the investment of contributions. In accordance with their much-discussed ‘Test 1’, they call for a rebalancing of the portfolio in the opposite direction—towards bonds and other ‘liability-driven investments’ (LDI). These investments have a much lower expected return than equities and property. But this cost is thought to be outweighed by the security provided by the fact that their cash flows more precisely match the cash flows of the liabilities (in this case, the promised pensions). A long-dated index-linked government bond, for example, offers guaranteed regular payments revalued by inflation over a long number of years, in a manner that relatively closely approximates the shape of pension payments.
Therefore, even though USS maintains that their current growth-weighted portfolio will have a 67% chance of achieving CPI + 2.8% ten years from now (which is well above both breakeven rates), their Test 1 mandates a shift from years 11-20 out of that growth portfolio and into one weighted towards bonds and other LDI. The shift to LDI will be sufficiently great that, by year 20 and thereafter, USS estimate that the portfolio will have a two-thirds chance of achieving a return of no greater than CPI + 1.7%, which is below the break-even rate for future pension promises. When, moreover, one combines this with the dismal returns that USS forecasts during the first ten years, we arrive at the low figure of CPI + 1.14% as the average return per annum that USS expects its portfolio to have a two-thirds chance of achieving over the next thirty years. This figure is below the break-even rate for past as well as future pension promises.
In comparing USS’s approach to investment with First Actuarial’s, the following question arises: According to USS’s own best estimates, the expected 30 year returns on growth assets such as equity and property are 3.64% and 3.23% above CPI respectively, whereas the expected return on LDI such as index-linked bonds is CPI minus 0.76%. Therefore, even if cash flows from growth assets track pensions liabilities less well than cash flows from LDI, the fact that the expected total volume of cash flows from growth assets is much higher than the expected total volume of cash flows from equivalently priced LDI ensures that the liabilities will be covered, even on a prudent rather than a best estimate basis. Why match pensions liabilities with a small stream of income that has the same shape as the liabilities rather than fund them from a less well matched but large stream of income that should more than cover the liabilities, whatever the mismatch?
The soundness of any call in future weeks for cuts to defined benefit pensions will turn on the presence or absence of a compelling answer to this question.
The Ongoing debates
There are two debates on USS running in parallel, the first is a public debate, of which Mike’s comments are a part, the other is a private debate between employers and the USS. The tension created between the very public and very private natures of these conversations is evident in Josephine Cumbo’s article in the FT.
Sam Marsh , from Sheffield University’s Maths school, has collected a petition of 1600 signatories which has been sent to Bill Galvin , CEO of USS – you can read the covering letter from the link on this tweet.
— Sheffield UCU (@sheffielducu) September 18, 2017
I intend to post relevant material that appears in the public domain as a record of this discussion. The quality of the debate is of course high, since the participants are familiar with the subject. First Actuarial, of which I am of course a Director, continue to participate – advising the UCU. Derek Benstead, who comments below, is a colleague at First Actuarial and an author of the First Actuarial report mentioned by Mike.
Derek Benstead of First Actuarial comments on Mike Otsuka’s article.
I like this article very much. It seems to me that you have studied both sides of the argument then tried hard to present both sides in a fair and balanced way.
A few comments if I may (from the author of the First Actuarial report to UCU).
The 32.6% cost of future benefits is only temporary, it reduces over the first 10 years by 5.3%, everything else being equal. A benefit cut judged on the 32.6% would need to be 2/3 reinstated by the end of the 10 years, everything else being equal. Changing benefits at every valuation is not a sensible way to run a scheme. Benefit changes need to be much rarer than that.
You wrote “First Actuarial also calculates that contributions going forward for future promises will need to grow at a somewhat greater rate—by 1.85% per annum in real terms—to reach the break-even point of covering those new promises. Since this break-even rate is 36% higher than the break-even rate for past promises, this figure is corroborated by USS’s statement that the cost of providing future promises is now about 35% greater than the cost of providing past promises.”
The similarity of the 36% and 35% is coincidental, there isn’t a link between the two. The “breakeven” discount rate values the past benefits at the asset value and it values the future benefits at the contribution rate. The difference in breakeven rates for past and future means there are proportionately more assets relative to past benefits than there are contributions to meet future benefits. The increase in the USS’s costing of future service is due to their reduction of the discount rate in the costing from 2014 to 2017.
You wrote “Why match pensions liabilities with a small stream of income that has the same shape as the liabilities rather than fund them from a less well matched but large stream of income that should more than cover the liabilities, whatever the mismatch?”
You’re right to conclude this is the critical question. I could expand it a little:
“Why match pensions liabilities with a small stream of income that is definitely too little to pay the benefits and requires more contributions rather than fund them from a less well matched but large stream of income that should more than cover the liabilities, whatever the mismatch?”
The critical task is to make benefit payments more certain, not to make the investment returns more certain. I hope that people will understand from our cash flow analysis that the fears of the proponents of bonds and Liability Driven Investment are exaggerated.
The employers sponsoring the USS should be interested in business efficiency, the cheapest way of getting a job done properly. The alternatives of running down the USS and going to Defined Contribution or to the Teachers Pension Scheme (TPS) will result in higher contributions from employers than carrying on with a sensibly invested USS.
The original purpose of starting a trust pension scheme is to invest it productively to the benefit of the employer’s contributions and members’ benefits. To invest in low risk / low return assets will definitely put the employers’ contributions up, and/or members’ benefits down, making the USS a poor reward in comparison with the TPS. DC has poor pension outcomes at current annuity rates, making DC a poor reward in comparison with TPS, unless the employers put in higher contributions than they do to USS. Closing USS would result in a loss of contribution cash flow, triggering a need to invest in bonds to provide for net cash outflow, resulting in lower investment return and higher deficit contributions to the USS for employers, in addition to contributions to whatever new scheme they institute. The best way forward for the employers is to ensure the USS is run to meet employers’ and members’ needs.