Implementing alpha-beta separation
IPE asked three pension services - in Denmark, the Netherlands and Sweden - the same question: ‘What does alpha-beta separation mean to you?' Here are their answers:
Erik Valtonen, chief investment officer at the third Swedish national pension fund AP3, which has AUM of around SEK227bn (€24bn)
In 2007, alpha-beta separation began to have a significant impact on our portfolio structure, although in previous years we had different types of unfunded, internal tactical asset allocation (TAA) mandates that could be viewed as part of alpha-beta separation. For 2008, we want to take the remaining steps to fully implement alpha-beta separation.
For us alpha is something that has to do with skill while beta is exposure to a risk premium. Traditional betas are related to equity and fixed income. Beta is generally much easier and cheaper to put in place although for alternative beta you will also need to have special skills.
Alpha is difficult to find, which is why we have several people looking at various alpha sources. In 2007, we started a new, external programme in global TAA, a new alpha source for us. We have also started looking into other types of alpha sources.
We haven't yet divided the organisation into alpha and beta teams although in our management framework we do the thinking and budgeting of alpha and beta separately.
One of the benefits of separation is that it has improved transparency in the portfolio. It is also much more flexible to do the asset allocation via a beta overlay strategy, compared with the traditional method of changing exposure by moving the monies between different portfolios.
However, the downside is that separation leads to a rather complex portfolio structure, and with an increased use of derivatives it may become an operational challenge. So pension funds would need to be of a certain size to rise to the challenges it brings, such as risk management and legal needs and more derivative transactions.
Cost issues can also arise when you use pre-packaged solutions to get the alpha delivered because they may include hidden fees. So you may end up paying a lot for such a structure.
You will also have to bear in mind that there is no guarantee of finding alpha.
But when you think of alpha and beta as separate return sources you realise that the traditional way of delivering those as a package is not optimal. And the strategy is not all about alpha. It is about beta too, because you can address the issue of the optimal beta, which is not necessarily a market cap-weighted index. In an alpha-beta separated portfolio strategy you are set to reach the best combination of alpha and beta sources, as well as a good design of the alpha and beta components.
Alpha-beta separation seems to be on everybody's lips. Although not many investors undertake full-scale alpha-beta separation yet, my impression is that people have started to recognise its benefits.
Bjarne Graven Larsen, chief investment officer at the Danish Labour Market Supplementary Pension, ATP, which has AUM of DKK420bn (€56.4bn)
Before we do alpha-beta separation we undertake a separation of our investments into a hedge and investment portfolio. That means that the interest rate risk on all our liabilities is hedged in a hedge portfolio that does not consume any of our liquidity.
Our liquidity is invested in the investment portfolio, whose purpose it is to maximise the real value of future pensions. We do not use benchmarks to do that. Instead it is simply about achieving a high enough absolute return to keep up with longevity and inflation.
And for that reason in January 2006 we started separating the investment portfolio into two sub-portfolios - the alpha and the beta portfolio.
The beta is an absolute return portfolio where we have separated the investments into five different risk classes: equity, interest rates, credit, inflation and commodities. The first process is deciding how to allocate risk to the five baskets. We try to minimise the risk of big losses and at the same time make sure that we can fulfil the purpose of our pensions.
Second, we decide about the implementation strategy for the beta portfolio, for example whether to invest in private or listed equity, small or large cap. We have a different implementation strategy in each basket so our beta portfolio is not a passive index-linked portfolio, rather it is a highly active one. The reason why we call it beta is because it is the area where we get some market risk premiums.
Alpha, on the other hand, is not about taking market risk premiums, it is about trying to be market-risk neutral and having skilled people who can add value on top of the beta portfolio. It basically uses the same techniques of a multi-strategy hedge fund. We have nine or 10 different strategies or teams, most of which are market-neutral meaning they do not systematically take long or short positions in equity, for example.
One of the big advantages of the beta portfolio is that by having an absolute return focus, you are not just trying to do what the markets and benchmarks do. In other words, there is no more benchmark slavery but a freedom to allocate to where you think you get the best risk-adjusted returns.
In the alpha portfolio the focus is on risk taking in the areas where you have skills and the ability to decide not to take positions in other parts of the market where you do not have skills.
I think separation has reduced our financial risk. But due to new instruments and more derivatives, the operational side has become more complex. However, we have tried to improve our operational set-up to avoid the increase in operational risk.
I have not seen anyone else in the Danish market implementing alpha-beta separation on a similar scale to us because we are still in the early stages of this strategy.
Philip Menco, CEO at De Eendragt Pensioen NV, which has AUM of around €1bn
e have a sizeable portable alpha mandate, which is typically part of alpha-beta separation. Some people talk about alpha-beta separation but few actually put it in place, which makes us one of the more advanced players. In particular, there is interest in portable alpha, which is the more traditional idea of alpha-beta separation. But it is scarce and difficult to find, and when you do find it, fees are sometimes extreme.
We transited from an enhanced indexing, or alpha, equities strategy, to an alternative beta strategy for equities, the so-called anti-benchmark strategy of Lehman Brothers, in January.
We made the choice between several 130/30 MSCI-based strategies, where the fund managers aim for alpha, and two beta strategies that focus on a totally different composition of the index. After taking everything into consideration we decided that the latter would be a more consistent way to outperform the market.
It meant the adoption of a totally different investment strategy that uses a different, non-regular index. Nonetheless, we still use the MSCI as a reference base, but just to enable us to make a comparison to see if the beta strategy is superior.
The main driver behind this strategy move is the expectation of higher returns and lower risk than in the traditional MSCI route. The assumption is based on back tests and confirmed by one-and-a-half years of real data.
We had been searching for a fairly systematic type of investment process and arrived at a list of five alpha managers that run against a well-known benchmark portfolio and try to beat it, and two beta managers. We chose Lehman Brothers Asset Management in Paris.
The major risk is that any manager you select could fail to deliver, in particular when you choose something completely different to the rest. But we have high confidence in this manager.
Because we changed from a pension fund into an insurance company in 2006 the culture at De Eendragt is different from that of most pension funds. We are still comparable to a pension fund and everyone knows very well what we are doing and what are the risks. Risk is not always negative - it could also give you an extra reward.
If you are a very large pension fund you can put these strategies into place. If you are smaller you can still arrange them but the problem is then to find a manager because you need a lot of knowledge and understanding of the product.
Whether alpha-beta separation will take off in the Dutch market depends on the quality of the existing alpha managers, their fees and what returns are available in the more traditional approaches.
Our plan is now to monitor the money and the asset manager and watch the way the strategy develops for us but that is generally what we do with every asset manager.