The liability-matching investment model for pension and insurance funds has been fashionable with regulators since the equity collapse in 2000/01 and the emergence of large, uncovered liabilities in many schemes. Typically, the idea translates into measuring future cash flows and then buying long-dated bonds - a practice that in the UK at least has had an unfortunate effect on bond prices and yields. But can liability-matching be successfully applied to real estate? Historically, the risk and return profile of real estate does not have a particularly good correlation with liabilities (although other variables, such as inflation, have to be considered).
Our questions included the following:
n Have you adopted the liability-matching model?
n Are you being compelled to do it by your government/regulator?
n If so, how are you achieving it?
n Do you have any issues with the length of real estate investments?
n Does it matter if your scheme is young or mature and will this affect liability matching dramatically over time?
n Will UK and German REITs make liability matching in real estate easier?
n Is liability matching or investment performance your top priority for real estate investment?
n Would you consider securitising real estate positions to achieve perfect bond characteristics?
officers’ pension fund
We didn’t set out to invest in real estate on a liability-matching basis. But the way the fund has worked from the beginning is that we’ve bought for term, we’ve looked at the security of the covenant, and we’ve gone for capital preservation. We’ve been in property since 1974 and we’re still there.
This may, in some ways, resemble a liability-matching approach but I don’t think that was our thought at the beginning. When we looked at this in some detail, around three or four years ago, we did what we call our “triangle of risk” and the base of the triangle is our portfolio of long-term, supranational bonds and property. It’s capital preserving, income generating, and essentially low management (although our manager, Wilky, will crucify me for saying that), let’s say it “seems” low management. Financially, it encompasses about £750m (e1.1bn), which is about 25% of the fund.
Technically, you can’t say that it’s going to be long-term liability matching, because you’re never sure how good your covenant is. You could be left with a void even on the best of covenants.
We started to move towards liability-driven investing in the early 2000s. We started to move because we weren’t happy with relative return benchmarks. We wanted to look at real money, real returns, moving away from “negative/
relative” to cash positive.
The frustration came from, for example, a UK equity manager, who returns 17% when the FTSE All-Share is 22%. So he’s five under on a relative basis – horror or horrors! But if he was on an RPI+6 target (retail price index) he would have doubled it! So where is the fault? Is it with the manager or is it with the trustee in not setting a proper target?
I’m more than happy to have managers working on an RPI+ basis because it’s a real return, a cash return. Trustees can understand it.
Although we’ve been moving in this liability-driven direction, it hasn’t caused us to change our philosophy greatly, in fact we’re glad we moved when we did as we were able to pick up some early-mover advantage.
The chase for long-dated bonds is a conundrum. I recently discovered that a lot of the buying has been from overseas investors although there has been a big demand from the UK. Yes, it’s in response to the pensions regulator, and, yes, it’s a response to the PPF (pension protection fund) levy. The more you have in non-risk assets the less your levy will be. But as a Dutch commentator recently pointed out, it’s a bit lemming-like, isn’t it? I’m worried that some funds have been buying these bonds because they’ve been told to in order to reduce their risk-based levy.
And in a few years’ time when markets turn they will be left with underperforming assets and capital losses, and on the regulatory basis then, will probably have to buy more long bonds to fund the deficit that they thought they’d plugged.
We’re currently not doing it. Obviously we’re monitoring the situation but we don’t need to do it at the moment, one of the reasons being we’re a multi-employer scheme, not one single sponsor has a major exposure to us.
Going forwards, on our real estate allocation my attitude is that “if it ain’t broke, don’t fix it”. Our strategy has worked well for 30 or 40 years and it will continue to work well. We’ve got 8% invested, I’d probably like to take
it up a couple of points, if the right properties turned up, but I’m not
going to chase stupid yields when I don’t have to.
Liability matching has always played a role in real estate investment and is probably even more significant since the collapse in equities in 2000/01.
When pension companies are looking at all asset classes they now start from their liability position. They see certain criteria for each class. The criteria for real estate does not depend on if it is quoted or unquoted. In the Swiss market and probably Germany too, the institutions are increasing their real estate allocations and in the main they’re not looking for increased returns - at least that’s not their prime goal.
The prime goal for pension companies in Switzerland is diversification - which is essentially risk reduction. This risk reduction is being done to match liabilities. So real estate is being viewed as a (relatively) safe instrument with certain returns.
I would say that the “term issue” in real estate is not a big obstacle to liability matching. When an institutional investor says, “I want to have 10% of my fund invested in real estate,” this has been based on a high-level asset-liability study. It’s not an issue as to where within the real estate sector the money is invested, such as direct/indirect, open-ended/closed ended or whatever.
And it’s not an issue about how long the money is going to be invested. Provided the 10% allocation is maintained it does not matter how long individual investments are being made for.
In fact you could argue that shorter terms allows the investment team to respond better to market conditions.
The asset liability study will also take into account whether the pension company is growing, remaining stable or contracting and this will also have an impact on asset allocations.
BLUE SKY GROUP
Most real estate departments are not equipped to tackle liability matching. I think we are one of the few that look at both sides of the equation from a pension fund point of view. In terms of whether it’s voluntary or compelled by the state, I think it depends on the fund’s policy. Here in the Netherlands if you are willing to provide a guaranteed pension that is secure in terms of wealth, liability matching is something you have to look at. But others are not promising that. They will of course try to do it, but it’s not part of their strategy.
Liability matching is not new, of course; we saw it in the past when most of the funds had their assets in fixed income, but given the current situation of most pension funds worldwide it’s something they are looking into quite seriously; given the ageing of the population it’s a topic that’s going to hit them hard in the future.
I don’t think you should look purely at real estate in terms of liability matching. You have to look across all asset classes. You will obviously have a large portion in fixed income and then you’ve got the equity side, and then you can diversify into other asset classes, real estate being one of them. And you need liquidity, you can’t put everything in fixed income and real estate.
In terms of duration the fixed income allocation is obviously a good way to handle liability matching. Duration is a big issue with equities and real estate and even academic studies on the subject appear to contradict each other. And if its unclear in equities it will hardly be well known for real estate. In the UK, for example, you still have long leases, but it isn’t really guaranteed that your tenant will last the whole period. And in Europe lease structures can be much shorter.
For our own schemes you could say they are liability-matched to a certain extent and we are currently in the midst of a couple of asset-liability studies. I think in the Netherlands you will see a model emerge where the fixed income portfolio secures the duration risk, mitigating changes in interest rate, and then the equity and real estate portfolios will be the classes that have to add extra returns to make sure you can meet your liabilities in the long run.
Geert de Nekker
I think that the liability-matching concept can be applied to real estate allocations very well.
Especially in the Dutch situation a lot of the liabilities are linked to inflation and real estate is one of the top assets to also be linked to inflation. So it’s a very good hedge and that’s why I think real estate has such a good position in Dutch investment pools.
I don’t think it’s necessary to go as far as securitizing real estate investments to achieve the asset/liability match. I think real estate is a good enough match in itself, held either directly or indirectly through funds and so on.
You can, of course, also hold a proportion of your portfolio in commercial-mortgage backed securities (CMBS) although this would be regarded as another “alternative” investment that would sit alongside the real estate holding.
The investment characteristics are slightly different but the diversification that this offers is to be welcomed.
In the Netherlands asset/liability-matching has become very important. There are a few new rules in place,
we refer to them as the NFTK, and these rules force the investors to have more focus on the asset/liability-match.
I would say that the liability-matching environment has made investors either equally enthusiastic or more enthusiastic to invest in real estate. Consequently, the proportion of real estate in portfolios will either remain the same or will increase.
For years now real estate has been a very popular asset class in the Netherlands and this appeal is undimmed.
I don’t think that shortening holding periods are a big problem for liability-matching using real estate. I should stress that Dutch investors are also focused on long-term bonds - and that is also driven by the asset/liability situation - but equity and real estate
are preferred investments because
the bonds don’t give you a good performance. Hence the “hybrid” thinking we are now doing. We are being squeezed by the duration of the bonds and by their low yields so equities and real estate are relied on to improve the overall investment performance.
I don’t know if the German and British REIT legislation, when it finally arrives, will affect attitudes to real estate and liability-matching. But I do believe real estate will become more popular among investors, especially in the German situation but probably also in the UK.
I do believe real estate can play a role in asset liability matching, but the problem is that it’s so difficult to find the right buildings. Everyone is looking for the same type of thing and it’s very competitive. Therefore you cannot really push into the real estate market within a short time.
Real estate can help, of course, but to make a strategy out of it is difficult. The other thing is that most investors don’t have a buy-and-hold strategy any more. Even at BVK, we sell buildings after a while to realise profits. It’s an active management approach and therefore it’s hard to say what is the right duration to have on the real estate side. What we are doing at the moment is looking for opportunities in Germany. I think the German market is quite attractive, a lot of foreign investors are pushing into it, but we are also selling in Germany, so it is a buy and sell strategy.
We have just finished our strategic asset allocation and we have a strong resolution to invest more into pan-European real estate. This will be done indirectly through funds.
We have three managers at the moment, Generali, TMW and Invesco. All three do a very good job for us: at the moment we have several new buildings going through the due-diligence process and we want to push ahead with this process.
In Germany a lot of insurance companies are trying to do asset liability matching. But we don’t, because our goal as a state owned pension plan is to provide the first pillar of the pension system – the basic pension. So when you are a medical doctor in Bavaria you are not obliged to pay into the state system but you must do so into ours. When you manage the basic pension system your goal is to provide a pension which can follow, more or less, the salaries for medical doctors at the moment.
So if you would do duration-matching we would fix the asset side but our duration on the liability side would be, for example, in the case of pharmacists 24 years. So you would never be able to match this duration, and if you did, when yields are so low, what would the situation be in ten years’ time? Nobody knows what the inflation rate will be then. So we need to follow the inflation rates in a flexible but stable way.
If we were to have very short durations, our results would be very volatile. If we had a very long duration then it would be too inflexible to follow the inflation rate.
Therefore we don’t think that duration-matching is the best way forward.
Some insurance companies use it because they have a different goal – the nominal contract – which is something quite different. So we are quite sceptical on the matching subject. We adjust a little bit, but it is not a real matching.
We are looking for investment performance, but of the risk-adjusted variety, that is really important. We had a very good 2005, looking across all the asset classes we had an overall performance of over 6.5%. So that was OK, but was we are doing at the moment is more and more diversification. Apart from the international real estate porfolio we will invest quite a lot in hedge funds, commodities, and high yielding emerging market equities.
At the moment 13 new mandates have to be awarded. So we’ll be having a lot of beauty parades in the forthcoming weeks.
Overall, our alternative investments this year will be worth in the region
We’re not raising our equity portion substantially, we’re looking for total return products.