LGIM's Graham Moles and John Roe look at the difference between cashflow matching and cashflow aware investing and discuss the role equities can play as part of the CDI spectrum.
The panellists
Chairman: Jonathan Stapleton, editor of Professional Pensions
Graham Moles, head of portfolio solutions at Legal & General Investment Management
John Roe, head of multi-asset funds at Legal & General Investment Management
The discussion
Jonathan Stapleton: Cashflow-driven investing, or CDI, is much discussed but is described in many different ways. Is there a risk it becomes a PR gimmick rather than a truly new solution?
Graham Moles: There's definitely a risk of confusion. The industry enjoys using three-letter acronyms and buzzwords and I think if we don't explain it properly to clients we run the risk of them not being able to fully understand the benefits these sort of solutions provide.
In addition, there are two categories of cashflow-driven solutions we're talking about here. The one most of the industry is focusing on is for schemes that are able to look through the mark-to-market risk and lock in returns for the longer term. They do that by matching asset cashflows with liability cashflows. This involves investing in assets such as corporate and government bonds, and lining up those cashflows with the liability payments.
Of course, there's another very important cashflow issue that people are focusing on: many of them are becoming cashflow negative and they have more outflows than inflows. Working out the way to pay those cashflows and not damage the scheme's position is very important too.
We describe these two sorts of CDI in different ways. The first is cashflow matching, the second is cashflow aware, which is more about looking across a wider spectrum of approaches and asset classes to get the cashflows you need.
To a degree, the decision whether you're cashflow aware or cashflow matching relates to how well funded your scheme is in general. Better funded schemes will be more focused on cashflow matching; less well funded schemes will look more towards cashflow aware, although it's not completely black and white.
John Roe: Yes, you get these other factors. So, for example, if you've got a very strong covenant, you can choose to run more risk, even if you're in a better funded position, because you've agreed that with your sponsor and have a very strong covenant. For smaller schemes, however, you haven't got that many lives - and you have a lot more uncertainty about your longevity risk, so you don't know as much about what you're matching, because you could get lucky or unlucky with your little group of pensioners. So the bigger the scheme, the easier it is to have more certainty and less random variation in the individual lives. Both of those are pretty important when it comes to which one you choose.
Second, there is the element around market pricing. Our team is more focused on market dynamics. A lot of people would argue inflation-linked government bonds in the UK are expensive, or corporate bond credit spreads are quite tight - that is a relative value consideration compared with assets that might give more attractive prospective returns. Right now we would argue equities are more attractive than credit, for example. And, for schemes that can afford to mismatch their assets and liabilities more, but are still worried about the outflows, cashflow aware might be more sensible than cashflow matching.
Jonathan Stapleton: Graham, does there need to be a mindset shift in the industry, given the challenges we're currently facing?
Graham Moles: Absolutely. There is almost a two-pronged issue here. Even with traditional investing, the approach now is very complicated. Trustees are trying to weigh up value at risk (VAR), volatility and funding levels on multiple bases, and are really trying to work out the best strategy for paying all their pensions. But the metrics they use don't directly tally back to that so we've been working on a longer-term approach that focuses on how successful has any given strategy been in paying all the pensions. We call it our expected proportion of benefits met, or EPBM.
A really good example of this - especially when you're looking at a longer-term investment approach, such as cashflow matching - is if you take a corporate bond, what you really care about as a scheme is does that bond default or not. The shorter-term mark-to-market benefits, or the diversification, is arguably less important, so it's imperative you have a framework looking at what happens to that bond when it matures or defaults on the journey, rather than just a one-year risk metric.
John Roe: They are very different metrics. If you actually sat down with a trustee and asked what they care about, they'd say paying pensions. And yet the whole industry has congregated around these short-term measures and, while I completely understand that in terms of making sure you and your sponsor can bear the mark-to-market risk, ultimately markets are far more volatile than cashflows.
So if you're not careful, you could get sucked into this short-
termism and lose focus on what you really care about, which is paying these pensions. I really believe schemes that can afford to should start to see VAR and the shorter-term measures as a constraint but be more focused on this longer-term EPBM type of measure as to what they're really trying to achieve.
If you took a new trustee and explained one approach or the other, they would find it far more intuitive to be focused on paying pensions than looking at movements in VAR, which is not necessarily helping anybody in achieving long-term outcomes.
Graham Moles: Also, with a longer-term framework, you can really start to factor in some of the other risks. John alluded to covenant risk - if you're looking at the one-year number, the chance of your sponsor defaulting is almost zero, but over the long term, about a third of pension schemes will default over the next 20 years, based on historical numbers.
So how do you capture that in a one-year risk number? It's almost impossible. Longevity is another big long-term risk, but over the short term you can't even see it. So if you're looking out across the entire length of a scheme's life in terms of paying out pensions, you can capture those risks.
John Roe: We saw this with one very big scheme that we are lucky enough to work with, which found 80% of its risk over the very long term was in its covenant. And it is completely undiversifiable, because it is your covenant. How can schemes spend all this time on the other risks without allowing for the covenant? Allowing for it leads to the necessity to take more risk, because your sponsor may not be there, coupled with ultimately what proportion of benefits get met if you're forced into a buyout. That could happen, or you're forced into the PPF or a similar workout solution. What proportion of benefits get eroded in those circumstances? So it really is quite a fundamental shift for the industry.
Jonathan Stapleton: Stepping backwards, cashflow negativity can complicate matters for trustees. Why have schemes so suddenly gone cashflow negative?
John Roe: It's a combination of two main factors. First of all, bond yields in the UK have gone down a lot, they're very low. And therefore, the cost of providing pensions has gone up a lot for defined benefit (DB) schemes. So the transfer values they can offer the members to move out of the scheme look more attractive because they're higher multiples for the individual. This, coupled with pension freedoms, which have made it easier for people to move out of a DB scheme, have meant there has been a sudden pickup in the outflows from transfers. And that's the new part of the jigsaw.
Graham Moles: In addition to the short-term dynamics that John just mentioned, there's also a longer term trend. So with schemes being closed, they're no longer receiving future accrual contributions. In the past, schemes would be able to meet pension outflows with future accrual contributions and therefore their strategy would be largely untouched. So that's one factor.
Another factor is that funding levels have gradually started to improve again and so deficit contributions are no longer being paid as much, or certainly not for as long as they were in the past. So again, that will increase cashflow negativity.
And the final one, because these schemes are closed, there are just simply more and more pensioners, on a relative basis. So the outflows are getting bigger and bigger, relative to the scheme's assets.
Jonathan Stapleton: Given all these issues, Graham, do you think all schemes should be shifting towards cashflow matching?
Graham Moles: Yes - and I'm glad you said cashflow matching and not cashflow-driven investment! So cashflow matching - I think most schemes should start to move towards that because as their funding level improves, it is a way of generating cashflows naturally. And what does that mean? It means you don't need to sell assets, which you otherwise might have to do in a downturn. It also starts reducing the governance burden on schemes because, if you have naturally generated cashflows, you don't need to spend so much time working out where to get them from.
In addition, they're the sort of asset classes that insurers buy. So because of the regulation they are subject to, insurers have to buy contractual cashflows, they have to do cashflow matching. If a scheme is investing in similar strategies, its buyout volatility will be reduced. And in some cases those assets will actually be so attractive to insurers, they'll be willing to take them "in-specie" without causing any sales.
That's why we think schemes will move towards cashflow matching, but there are a couple of elements of caution here. Imagine a scheme that's fully funded on a gilts flat basis with no longevity risk. A traditional way of looking at the strategy may be to invest the scheme entirely in gilts to reduce the risk. If the sponsor goes bust before the last pension is paid, you're going to have to go to a buyout. And if the buyout price is more expensive, your pensions will take a haircut. So you may wish you had taken a bit more growth or risk in that scenario.
Another element of caution to take is that, as John mentioned earlier, spreads are quite tight at the moment. Some schemes are looking to move into more complicated assets, like infrastructure debt, as a way of cashflow matching but with extra yield. These investments come with their own complexities. If you are invested in something illiquid, by definition you can't sell it. So if you have transfer values and need to realise some assets you weren't expecting to, you can't sell that illiquid asset down unless you incur a large cost. Further, although insurers love those sort of assets, they're quite particular. So if you don't get exactly the one an insurer wants, you'll have to sell if you want to go to a buyout with it.
Saying this, the one thing that will definitely be the case is schemes will increasingly focus on their cashflows. Being able to pay the cashflows out and having cashflow awareness is something we think all schemes should keep in mind.
Jonathan Stapleton: It seems many schemes will fit into this cashflow aware category. What is that exactly?
John Roe: It's basically accepting you cannot de-risk - doing so is almost definite death because you haven't got enough money and if your assets don't grow you're either going to have to ask for an increase in contributions or you're going to run out of money. We just don't have enough assets; that's the starting point. You need assets to grow and you can show that, if a scheme is underfunded, the most important thing is its asset allocation. If you don't take enough risk or diversify then you can leave yourself with some pretty ugly outcomes.
But if you're going to own growth assets, you want to be able to own them for long enough to get paid for taking that risk. And if you're forced into selling those assets to meet cashflows, then the whole problem gets more complicated. Equities may do well over ten years then go down over a year and you're selling them for less than you bought them, making your situation even worse.
So when you think about cashflow aware investing, you're saying, well, number one, I must take risk because if I don't then outcomes are poor. So, I'm going to take risk but I'm going to think how I am going to meet my cashflows in the first five, 10 or 15 years. And in doing so, I can then be more confident that I'm not going to be a forced seller of assets. You shouldn't get too caught up about this. I'm an actuary and we absolutely love matching cashflows to pence and the rest of it, but the reality is that this can be slightly over the top. Most path dependent problems, which is what you get when you sell assets at depressed prices, are a first order problem. So if you ignore cashflows entirely, then they can cause a huge problem. But if you've taken out the first order problem, and the cashflow need is approximately met by asset cashflows, then the problem is largely gone and you can move on to look at something else.
Pension fund risks are like whack-a-mole: you whack one and then another pops up. There's no point in continuing to whack the same mole over and over again. At some point you've got to say this is no longer that risky; where's my next mole, where is the next risk I should be focused on?
Allowing for the cashflows on growth assets - the coupons on bonds, the dividends on equities and potentially even the redemption payments on bonds - can make a big difference and can allow schemes to hold the underlying assets for longer and help remove some of the short-term complications and focus on what really matters over the long term, which is paying benefits to members.
It is quite a simple solution and takes out that first order problem for trustees.
Jonathan Stapleton: When many people talk about CDI they mean using assets with contractual cashflows and bonds in particular. You're talking about equities - how does that fit into CDI and being cashflow aware?
John Roe: It's all one spectrum. Just like people used to draw this imaginary line between corporate bonds and all other cashflows, now this imaginary line sometimes gets driven between a cashflow that's contractual and a cashflow that isn't. First of all, a risky bond cashflow won't necessarily materialise, particularly when you get to the higher end of the risk spectrum. If you're getting a 9% coupon on something at a time when government bonds are only giving you 2% or 3%, then you should be aware that you're taking risk. And if you're taking risk, those cashflows may ultimately not come through. It's all just one spectrum. As you get to riskier assets, you should just take a bigger haircut.
So if you own a government bond, then you're probably going to allow for all the cashflow, because if the UK government goes bust, we've got bigger problems - it is shotguns and tinned food in caves time. So let's assume that doesn't happen. Then you've got corporate bonds, you take a haircut for a few defaults. Then you get into the higher risk fixed income: bigger haircuts.
When you get to equities, you would expect dividends to grow over time because profits grow and payouts to shareholders grow. But even if you are conservative in the cashflows you take, it would be very unusual for dividends to go to zero on the whole of a global equity index. Again, if that happens, we've got much bigger problems. If you have got these cashflows on equities, why wouldn't you include them as it means that for a given amount of risk assets, you can get more cashflow and worry less about being a forced seller.
It is really about not creating imaginary boxes to put things in and seeing it all as one spectrum - and then mixing and matching from across the spectrum to get the right asset allocation while coping with these outflows.
Graham Moles: John is absolutely right. It's definitely a spectrum. As an example, there are a good number of schemes that have cashflow problems but haven't looked to take dividends from their equities - that's a really simple way of turning the taps on to deal with a cashflow problem. In reality, cashflow matching is just the extreme end of cashflow awareness. That's why we've come up with these two categories, to try to separate the two issues a little bit because there is a risk of real confusion.
Given that you've got covenant risk, and that you need to keep improving funding levels, you're going to need some growth from almost all clients. So having a growth strategy that is able to pay out cashflows is exceptionally useful, helps you with the governance burden, helps you with the cashflow problem, and yes, you drift towards cashflow matching as you get better funded and need more certainty.
John Roe: A key point is that because of the covenant risk and the fact the sponsor may not be there indefinitely, you cannot ignore pricing relative to buyout pricing. You have to bear this covenant in mind - and this will encourage schemes to push for cashflow matching, but I think clients are increasingly agreeing that you shouldn't do this regardless of asset prices. You've got to take into account the valuations of assets and the attractiveness of assets, as well as your ultimate goal, which is to pay your pensions and the importance that buyout might play in that.
Jonathan Stapleton: We're reaching the end of this debate; what are your key takeaways?
Graham Moles: The matching cashflow side is a really important part of a scheme's toolkit and will become increasingly useful as they get better funded. But it's not a silver bullet; they need to ensure they're looking at all the risks across the spectrum and holding enough risk to ultimately close that funding level gap.
John Roe: Over the long term, owning growth assets can help you close your funding gap. But the risk is the short term; you also need sufficient cashflows coming off your assets to get through the short term without too much drama, and allow trustees to focus on longer-term decision making.
Constantly worrying about where the next bit of money is coming from takes up time and can be a real distraction from what schemes are trying to achieve. Both sponsors and trustees should focus on making it more about the long term by taking the cashflow out of the equation, even with underfunded schemes.
