Germany’s largest independent pension provider, the Munich-based BVK multi-employer fund that runs €40bn is assets for 12 different professional and occupational groups has taken the next steps in developing its approach to asset allocation by making risk management core to the process.
The introduction of the fund’s new risk budgeting tool came in time for the two-yearly programme of reviewing the strategic asset allocation for each of the schemes under its charge last year. “We were able to use the tool for this exercise,” says Daniel Just, CIO of the BVK.
The risk budgeting process has been extended since then to monitor on a continuous basis the risk levels with the portfolios run by the fund. “Our consultants Mercer helped us with this,” he adds.
The reason the fund decided to concentrate on risk budgeting was to make sure that it was using its risk allowance as efficiently as possible. “We find that this is a very appropriate tool,” he says, adding that by relying just on the regulatory stress tests the danger is that the fund won’t use its risk budgets effectively.
His colleague André Heimrich points out: “When we want to fulfil the stress test with a low risk budget, we have to lower our exposure to the more risky assets. As a result, to meet the stress test equities and alternatives must be reduced significantly and we have to invest more in fixed income which is (due to the low interest rates) lowering the risk budget further. This is like a spiral and at the end we found ourselves in a 100 % fixed income portfolio, because there is no risk budget left.”
He adds: “But this also means we cannot reach the target return we need for the fund when fixed income return rates are below 4% and we are totally depending on the development of the fixed income market. “
“Our approach was to not place such emphasis on the stress test and to find another approach that better fits in with BVK’s particular needs,” says Heimrich.
The exercise started by grouping all the different schemes in the BVK stable. “Some of our schemes run on a pay-as-you-go basis, but with significant buffers - perhaps seven years’ worth,” says Just. “Others of our schemes are fully funded and so covered by their assets, and some others are in-between.”
Then the fund analysed all the fixed income investments on a single platform, as well as on the master fund structure it uses where all the non-fixed income assets - other than real estate- are run within a series of Spezialfonds, which the different BVK schemes have access to.
On the liabilities side, the fund made forecasts of the future liability structure for the years ahead. “Then we did a Monte Carlo analysis producing 1,000 different future scenarios on the asset side over a five-year horizon.”
This enabled BVK to define risk budgets for every pension scheme at a 98% level of confidence. “Our next task was to allocate this risk budget most efficiently.” This meant feeding into the scenario generator for the different asset classes, the return expectations, the volatility, the correlations, as well as the variance and the co-variance, says Heimrich.
The BVK says it is very conservative about the forecast assumptions it uses. As Just puts it: “We do not want negative surprises from being over optimistic.” So in the case of hedge funds, where a 5% return was shown to be possible with an expected volatility of 7% (as against an average volatility of 4 % for the class). “The outcome of the asset allocation optimisation was favourable to including hedge funds.”
The asset type with most volatility happens to be currency, points out Heimrich. “Our calculations come out with a volatility of 40% and returns of 15%.”
The next step was to integrate this into the fund’s Master KAG fund structure, which allows the different pensions schemes the ability to make their asset allocation across the BVK’s range of seven Master-Spezialfonds, covering the following areas: equity euro LC, equity world LC, equity SC & MC, spreads, Euroland balanced, EM-debts and high yields and hedge-funds
The asset allocation strategy is worked out on an individual basis for each of the 12 schemes and they can make the choice of asset categories on this optimised basis from the choices available within the Spezialfonds. “We asked ‘what will be the optimal asset allocation of the master fund structure, having regard to the risk budget, so that this is neither overshot, nor under-utilised’.”
All through the year, the risk element is monitored, firstly to ensure that everything is on track and operating within the expected parameters. Just refers to this as the ‘green flag’ scenario. But it is always possible for a new scenario to develop. He points to one of the pensions schemes that a change on the liability side with an impact on the risk budget, enabling the scheme to take on more risk. “We can then revise the allocation to use that risk optimally.”
This exercise can show up when there is risk budget available that is not being used, for example. “So the question then becomes one of being able to use this if we have the right market expectations to do so,” says Heimrich.
There is also the yellow flag situation, where a market perhaps gets into difficulties and as a result the scheme sails close to its risk budget, says Just. “Then there can be the red flag situation, when we are truly under water.” He adds: “Over a 50-year horizon, we calculate on the basis that there will be one occasion when a pension scheme cannot fulfil its liabilities.”
The forecasts for returns and volatilities are on the market expectations - for the beta. “It is important that the alpha side is also included in the risk budgeting process,” says Just.
“We use the Mercer manager database to arrive at expected consistent alpha that our chosen managers can produce,” points out Heimrich. “We have forecasts from the database for alpha, for the volatility of alpha and for the correlation between the
“Inputting these into the scenario generator provides a consistent, possible set of asset allocations for all the pension schemes. If this is done for say 1,000 time under a Monte Carlo analysis for both beta and alpha, you obtain a return profile of a certain asset allocation,” says Just. This can help show up the ‘fat tails’ issue, where extreme events can result in very significant and unexpected losses.
So from the huge range of scenarios generated for the different combinations off asset proportions and combinations, under different assumptions, a selection has to be made as to those that best meet the scheme’s requirements.
When it comes to the actual asset allocation, each of the retirement funds within BVK will have two basic decisions, the answer will depend very much on its risk budgeting. The first relates to the proportion of fixed income, which accounts for 70% of total assets of the fund and is an internally managed portfolio. The other is the master fund structure comprising of a range of externally managed Spezialfonds accounts. Altogether, there are now 45 of these in the hands of 35 outside managers.
Each plan can buy-into which ever of this array of investment choices it deems appropriate in the proportions needed. “It is up to them to decide if they want to use the more risky ones, for example, a pure equity funds,” says Just. The trade could be more fixed income but fewer high risk investments or less fixed income and more risk-based assets but not at the higher risk end, he explains.
“In one particular plan, as a result of the optimisation and risk budgeting process, the proportions came down from 84% fixed income and 16% in the master funds portion, to ratios of round about 70 to 30,” points out Heimrich.
The BVK,which acts like a manager for the 12 integrated pension schemes (Versorgungswerke), is not regulated by the BaFin, the federal supervisor, but is supervised locally by the Bavarian state. “We have explained very carefully our processes and procedures to them and we use BaFin as a reference. But we are somewhat freer as a result,” says Just.
The overall result of the new approach, which undoubtedly have played a part in winning the fund the Best Fund in Germany accolade at the recent IPE Awards in Paris, have been that the different plans have widened their use of the different opportunities provided within the master fund structure. “This has increased from around 16% of total allocation to nearly 30%,” Heimrich points out. Also, there has been a trend to chose the absolute return type strategies, by moving to alternatives, with €1bn in hedge funds. But what is really surprising is the variation between the allocations of the different funds. “It can differ significantly,” says Just.
The BVK is not resting on its laurels but is looking where next to take its allocation offering. Commodities was recently examined, but it is an area that has its particular problems if you down the indexing route favoured by many. The fund’s first exposure is to be through two specialist funds of hedge funds investing in commodities to the tune of €300m. “Our approach here was to see that there was no correlation to the commodity market,” says Heimrich. The currency arena has also been invested in with an initial €200m commitment.