Leave nothing to chance
The advent of auto-enrolment in the UK has heightened focus on the need for robust risk management systems for DC plans. Gail Moss looks at a range of approaches
As auto-enrolment continues apace in the UK, the role of risk management
in default funds from providers cannot be overstated, especially since these funds are aimed at the huge numbers of defined contribution (DC) plan members who often expect to leave even the most basic investment decisions up to others.
“Without a proper risk strategy, what could go wrong is that pension plans don’t deliver returns in line with members’ expectations,” says Jane Ridgley, product director, workplace savings, auto-enrolment provider Legal & General. “That could mean they panic and pull out of a pension plan, missing out not only on the opportunity to make tax-free savings, but also on their employer’s contributions.”
Here we outline different approaches to risk management in a DC context.
1. NEST in class
For the UK National Employment Savings Trust (NEST), there is also another reason to develop a successful risk strategy – its target market has lower earnings and savings than those of people currently saving for a pension, so their risk capacity is much lower.
NEST realised it had to start early on in setting up its strategy. “Implementing a risk management system is an expensive task in terms of money and resources,” says Rudyard Ekindi, director of investment research, NEST. “So we engaged with our peers in the UK and overseas who had launched their DC schemes without a risk management capability, only realising its importance years later. The delay in setting up a system was very costly for them.”
NEST also recognised that it was important to make risk management integral to the default scheme as a whole, rather than merely bolted on.
And the support function also had to be fully staffed: “Using sophisticated governance systems means you need senior analysts to get full value, as you have to understand what to do with the information,” says Ekindi.
NEST has around 16 investment staff to cover the full range of investment functions. Ekindi says it might look rather large while assets are low, but stresses that this was necessary for a robust set-up. However, the team will grow only marginally as the assets grow exponentially.
But Ekindi says it is not enough to protect against risk – a pension fund needs to protect against the right risks.
For example, NEST uses a dynamic – not mechanistic – lifestyling approach to protect a member’s investment against falling stock markets when the member approaches retirement.
“It’s a lottery as to whether you retire just after a market crash or not,” says Ekindi. “The traditional mechanistic lifestyling model is wrong, because it focuses on only one risk and doesn’t take timing factors into account, such as market conditions. We think there are a number of risks which need identifying carefully, and we chose to include this mistiming risk in our risk management strategy.”
In 2011, after a thorough search, NEST engaged MSCI to provide a risk management system. Its BarraOne risk management platform has allowed NEST to identify which risks it was exposed to, which of those risks it did not wish to be exposed to, and the size of those risks.
This has led to the design and implementation of a dynamic risk budgeting approach, with strategic reporting supported by risk-adjusted performance (RAP) indicators.
“The most important thing in building a portfolio with different percentages in different asset classes is to find out the swing in valuations on a one, two, three-year time horizon,” says Kurt Winkelmann, managing director and head of risk and analytical research, MSCI.
“That’s important because it gives you some sense as to whether you can deliver on the commitments made to beneficiaries.”
Once MSCI had been appointed, NEST spent six months with its dedicated team refining and implementing the system.
MSCI was able to use its experience in other DC markets such as Australia and the US to enhance the techniques it suggested to NEST.
The resulting RAP framework means that NEST can look for rewarded risk factors, while limiting exposure to less rewarded ones.
Importance of default
The NEST default fund is a series of dated funds matching each member’s anticipated date of retirement. Asset allocation is carried out by NEST itself.
The timeline for each fund consists of a foundation phase up to age 29, followed by a growth phase until age 55, then consolidation until 65.
It is during this last phase that derisking takes place.
In contrast to traditional schemes, which normally run a linear derisking programme over 10 years, NEST derisks in tranches during the consolidation phase, making adjustments according to two factors.
“Most important is what the state of capital markets is at that time,” says Ekindi. “If there’s a lot of pressure on bonds, say, we delay derisking. But we could bring it forward if, say, the member’s pot is well ahead of their long-term expectations.”
The other factor is how members are going to take their savings when they retire.
Where a member plans to use a proportion of their pot to buy an annuity, that proportion must be hedged against annuity prices 10 years before retirement. Where the member plans to take cash, that amount is invested in cash.
To make the right decisions, NEST depends on information from risk management consultants, custodians, and other third parties.
“However, estimates vary according to the different methodologies companies use, so we have to carry out reconciliations between these figures,” says Ekindi.
Pension fund trustees should also be aware that the quality of information varies according to the asset class.
“When you’re looking at traditional asset classes, you get consistency of information,”
Ekindi continues. “But less traditional and less liquid assets require more adjustments to the figures. Trustees need diversification, but that leads to more hidden costs.”
The risk-management system is also being fine-tuned to deal with potential crisis scenarios. But Ekindi warns that there is still potential for mistakes.
“Trustees can spend a lot of time thinking about the impact of a particular event on the scheme. If only one event happens, that’s OK. But when a scheme gets into trouble, it’s usually because it’s hit by a perfect storm of several events at once.”
Risk management should therefore include assessing the chances of such events happening together, rather than just evaluating the impact of each event on its own.
2. Drawdown constraints
“The main problem in terms of risk management at the global portfolio level is that it is supposedly done by the choice of strategic asset allocation,” says Adina Grigoriu, chief executive, Active Asset Allocation (AAAiC). “And the way strategic asset allocation has been determined in recent years hasn’t changed much, with allocation results relying entirely on risk, return and correlation hypotheses. But return forecasts are too difficult to get right on a consistent basis, and this can be very costly in terms of risk management.”
Instead of trying to forecast returns, Grigoriu says there should be a focus on managing risk. And once risk has been properly defined according to the specific situation of the investor, it is easy to define the allocation that respects the risk constraints at a certain point in time. This allocation will need to be adjusted over time to take into account the fact that the world is changing, the risk perception varies and that the market moves constantly.
Furthermore, as portfolio values increase, it makes sense to try to protect higher and higher values and preserve the performance achieved by the portfolio so far.
“So you need to add some drawdown constraints – ie, stipulate the maximum amount which investors can tolerate losing,” she says. “This is usually 10 to 15%, but can be much lower for some investors.”
AAAiC uses advanced stochastic methods to create scenarios for how assets might behave in the future, then chooses the parameters for the asset allocation models, such that investor’s risk constraints are met whatever happens.
“But the asset allocation will need to be changed as market conditions change,” says Grigoriu.
Whereas in lifestyling strategies the allocation only varies with time, factoring in drawdown constraints – say 15% – means that exposure to the riskiest assets can be temporarily reduced, sometimes drastically, as market conditions worsen. But it also means that the potential for taking risks, thus generating performance, remains throughout the portfolio’s lifetime.
“There is not much downside to our approach, except that it can be seen as very complex compared with a lifestyling strategy, where the allocation is pre-defined and often independent of market conditions,” says Grigoriu. “But it does cut off downside risk and generates more positive returns.”
Legal & General’s key risk management tool is diversification. Its default fund, based on Legal & General’s Multi-Asset Fund, includes 20 asset classes and 6,500 underlying investments. The fund also targets volatility of two-thirds that of global equities over the long term, so it is less risky than a pure equity fund.
Ridgley says: “The fund was developed purely for the pensions market, not for any other investment sector, and has achieved strong returns in decent markets. Our auto-enrolment opt-out rate is currently under 10%.”
Default funds are analysed by a governance committee independent of Legal & General’s workplace savings division, and Legal & General Investment Management. Employers who operate strong governance within their business can request a different default solution for their membership demographic. An independent investment consultant also reviews the investment strategy for the Mastertrust product. Ultimately, the trustee board reviews whether the portfolio has behaved as expected.
“Sponsoring employers of DC schemes need to be comfortable that what they are being offered suits their scheme’s demographic,” says Ridgley.
She says it is important for them to set a clear brief to their auto-enrolment providers:
“They have to make them look at the scheme from the member’s point of view.”
4. NOW let’s look at charges
For auto-enrolment provider NOW Pensions, a key requirement of a good default fund is keeping costs down.
“Default funds need to deliver three things – cost-efficient returns, highest possible returns and low downside risk,” says Morten Nilsson, chief executive, NOW Pensions. “Yet if you take the best-performing funds and apply the heaviest charges, they become the worst performers.”
Last October’s Pensions Institute report on auto-enrolment, Caveat Venditor, found that thousands of UK employees in DC schemes pay annual charges of up to 3%, six times the charges of up to 0.5% imposed by modern multi-employer schemes.
“Unless older high-charging schemes are abolished, their use for auto-enrolment will lead to the UK pensions market facing a mis-selling scandal on an unprecedented scale,” the report warned.
NOW Pensions’ auto-enrolment product costs members £0.3-1.5 per month for administration, plus 0.3% in annual management charges.
NOW Pension’s risk management modelling starts by fixing the overall risk level required in the portfolio. Once set, the percentage of the portfolio placed in each of five asset classes – equities, rates, credit, inflation and commodities – is determined.
“The portfolio is not dominated by equities, so we are not overly sensitive to the way equity markets perform,” says Nilsson. “These five asset classes perform differently in different economic scenarios, so there should be an overall stable return whatever happens.”
He says: “We also use derivatives to allow us to react even quicker if we see something we like or don’t like. We use active beta strategies, but we’re not stockpickers.”
Using the experience of its parent ATP, NOW Pensions also uses DB techniques in order to hedge future liabilities, protecting members’ pension values even further.