Risk budgeting by financial institutions is enjoying an unprecedented spell in the spotlight. Since the late 1990s pension funds too have been struggling to integrate risk budgeting techniques into their processes in response to concerns about the level of risk that is acceptable in the portfolio but also as a result of the introduction of new risk measurement and management tools.
Actually, risk budgeting is no new invention at all, it’s a modern application of the classical Markowitz portfolio theory, but within the triangle ‘return–risk–correlation’ there is now more emphasis on the last two factors.
Risk budgeting can be defined as a policy in which asset allocations are viewed in the context of each asset’s contribution to the total portfolio risk as well as the contribution to the total portfolio return. Combining risk and return is nothing new. What is new is that risk budgeting has become an effective management tool for planning and controlling active management decision-making.
The lack of a common standard has led to the development of a multitude of risk budgeting concepts that are tailored to the specific needs and characteristics of each individual institution.
In this article we will expand upon the allocation methods ABP has developed to manage the risks that are being taken in its portfolio. ABP is the pension fund set up for employees in government, education and defence in the Netherlands representing 2.2m participants. Total assets amount to around e150bn.
We distinguish the following steps on the road to an optimal composition of the overall portfolio:
o the determination of the maximum risk level the ABP pension plan can tolerate;
o the optimal allocation of the total risk budget to various active investment management decisions: strategic alpha allocation, tactical alpha allocation and operational alpha allocation;
q risk budgeting as a framework for planning and monitoring the overall investment process.
The first step involves the assessment of the ideal asset mix to meet the pension liabilities. At this stage the normal investment policy is determined. The focus is on a highly aggregated level of asset categories: equities, fixed income and currency exposures.
Probably this step is the most essential one, because ABP’s key mission is to safeguard its long-term commitments through exposure to a specific risk profile.
So the question arises what the composition is of the ideal mix – or norm portfolio as it’s called in the Netherlands. ALM is a quantitative model for deriving that ideal asset mix. ALM involves setting a specific asset allocation for the long term. The logic is that each pension fund has specific risk-return objectives and cashflow requirements. Once the unique attributes of ABP’s liability structure have been determined, as well as its policy regarding preferred funding ratios and contribution levels, one can assess the appropriate asset mix. In the case of the ABP pension fund this boils down to a ratio between equities and fixed income assets of 60/40, and currency exposures in which 100% of the foreign fixed income assets are hedged.
This norm portfolio contributes to a rather stable structure of balance sheet components (assets, liabilities and surplus). On average and in the long-term the real returns on assets and liabilities are equal, the excess return is zero, and the excess return volatility is at an acceptable level. At the same time the level of yearly pension contributions are moderate as well as the changes from year to year.
Next steps concern the issue of how to enhance normal investment returns without undermining the risk profile as determined in the norm portfolio. This is about active management. In this phase we distinguish three types of alpha generators ie, excess return generators, first strategic alpha, second global tactical alpha and finally operational alpha.
Chances for successful active management grow with diversification across multiple sources of return.
The preceding ALM phase dealt with the liquid, developed markets: the conventional investments in equities (US, Europe and Japan) and US and EMU Treasury bonds. It is easy to buy and easy to sell at minimal transaction costs and the portfolios can be managed in a rather passive way. Asset management fees can be maintained low.
One could be satisfied with an asset management formula like this, but one could be more ambitious and try to enhance investment returns by implementing active strategies. After all, for a pension fund with a duration of around 101 an increase of return of 50 basis points (bp) can be translated into a 5%-point lower yearly pension contribution.
Attractive opportunities are offered by so called strategic alpha generators:
o private equity (holdings in risk capital of non-listed companies);
o emerging markets equity investments;
o credit bonds and collateralised fixed income securities, including high yield and emerging markets debt;
o real estate;
o hedge funds (absolute return driven investment strategies);
o commodities (investing in unleveraged futures or swaps on a commodity index);
o index-linked bonds (protection against unforeseen inflation shocks).
These are all special asset categories, special in terms of complexity, illiquidity and lack of market efficiency. Relative to the conventional asset types they provide potentially higher returns or lower risks (see figure 1). In addition to that, some of them have so much diversifying power (commodities, real estate and index-linked bonds) that on a total portfolio level they allow for additional exposure to risk-bearing assets despite their relatively low expected returns.
With the help of a well considered optimisation over conventional and alternative assets, and by taking into account world wide market opportunities and limitations, one can achieve a remarkable improvement of the risk-return profile (see point S in figure 2).
The exposures in the present optimal strategic alpha portfolio in the case of ABP are shown in table 1).
Given the unique attractiveness of index-linked bonds for pension funds, it is most likely that these instruments would have taken a larger part of the whole if they were broadly available in EMU capital markets.
Tactical alpha allocation is one step removed from strategic alpha allocation. It attempts to improve investment returns by taking advantage of temporary imbalances between global asset markets. Against the background of the very liquid and efficient nature of these markets one must admit that this kind of forecasting has a relatively high skill hurdle. The opportunities for breadth – the number of independent bets that can be taken – are also limited.
Because of the danger of pitfalls in TAA, ABP prefers a cautious stance and acts only in extreme circumstances. So, this discretionary decision taking goes hand in hand with market movements and risk premiums which by far surpass normal volatility.
The operational alpha allocation is more decentralised within the asset management organisation. This source of excess return results from capitalising on the advantages in the areas of systematic analysis and active portfolio management, which is put into operation in front-office units we call asset platforms: equities, fixed income and alternative investments. Each platform’s benchmark is directly derived from the strategic alpha allocation, which means that a platform is also responsible for efficient implementation through effective trading in financial markets.
The real added value of active management in this stage emerges from:
o Sector selection;
o Style selection;
o Credit quality selection;
o Stock picking;
o Timing;
o Spread bargaining.
The full risk profile is determined by ALM and multiple active management decisions. Although the objective of active management is to supply incremental return, on balance these deviations from the norm portfolio should not result in an increase in total portfolio risk. This predefined limit to all active decision taking is called the total risk budget.
As shown above, by adding low correlation asset types and strategies we have managed to lower the actual portfolio risk, but TAA and operational alpha strategies will surely not be able to reduce portfolio risk any further. On the contrary, they will increase portfolio risk. So, the implementation of these strategies has to be managed in such a way that the total risk budget is not exceeded. The ultimate goal is to reach a portfolio that suits the total risk budget and maximizes expected return (see point Z in figure 2).
We have chosen to employ the total risk budget by setting tracking error allocations to the three platforms. Tracking error is the active risk a portfolio manager takes relative to his benchmark. It is a transparent and well known concept in the asset management business.
As the total room for risk taking is limited, the size of the allocated tracking errors will be determined by the expected effectiveness of portfolio management. This is measured by the so called information ratio (IR – see figure 3). IR is expressed as the portfolio manager’s ratio of excess return to tracking error. The question is how a portfolio manager can achieve a high information ratio? This capacity strongly depends on the factors ‘skills’ and ‘breadth’.2 It should be clear that active management decisions are to be made by prudent experts and that large fluctuations in returns should be reduced by employing different investment styles.
Critical for the tracking error allocations are the correlations between the active bets done by the asset platforms on the one hand, and the correlation between the expected excess return of a platform and the return of the strategic alpha portfolio on the other hand. Just because these correlations are relatively low (0–0.15), it is possible to make use of strong diversifying powers.
Generally speaking, one can state that a high IR, a low strategic alpha exposure and a low correlation with active bets elsewhere in the portfolio causes high tracking error allocations. But a low IR, a high strategic
alpha exposure and clearly positive correlations with other managers’ active bets will push the tracking error allocation down.
In practice, tracking error allocations are assigned as a target to be achieved by portfolio management.
Step by step the risk puzzle for pension funds is being solved. The recent developments in risk budgeting have turned Modern Portfolio Theory into a useful practical management tool. It encompasses the real big issues for pension funds: the management of surplus risk and return. Moreover, it encourages but also disciplines active management styles that are to add investment returns without adding total portfolio risk.
Jan van Roekel is executive vice president allocation of ABP Investments in the Netherlands