Advanced Pension Finance based on an Active and Dynamic LDI Concept
Dr. Gerhard Scheuenstuhl
Managing Director
risklab germany GmbH
LDI stands for “Liability Driven Investments” and it generally refers to an investment strategy that is aligned with the liabilities. LDI topics are therefore primarily discussed in the context of pension investing with its long-term retirement provisions, and in connection with the life insurance business holding rather complex liability structures. The reasoning of aligning asset investing with liability obligations is therefore not new in these businesses. Known as Asset-Liability-Management it already has a long tradition. New, however, is the framework for some of these institutions considering changes in accounting and regulatory rules. The main drivers for the recent attention that LDI has attracted are the deficits that many pension funds have accumulated over the last few years, the growing orientation of international accounting rules on market values for assets and liabilities (see IAS 19 or FRS 17) and regulatory changes towards the introduction of minimum funding levels or riskbased solvency capital requirements (see current discussion about Solvency II). These trends also underscore a growing demand for new solutions to overcome the shortcomings of purely asset-driven investment strategies.
In practice we find different LDI approaches to face these challenges. Figure 1 shows three approaches to how LDI is perceived in the market:
Figure 1: LDI-Approaches based on different investment philosophies.
Passive immunisation: One increasingly popular strategy to face current challenges is liability matching. It embodies a pure immunisation strategy with respect to liabilities.
Active static approach: In order to enhance the performance of assets (e.g. to reduce contributions) and to hedge against uncertainties in long-term liabilities, it is necessary to include asset classes into the strategic asset allocation that are not perfectly correlated with liabilities, e.g. stocks or hedge funds. Depending on the investor’s risk preference he will choose a (static) strategic asset allocation (SAA) which will be either, for example, a balanced investment portfolio with moderate risks relative to liabilities (“Liability Balanced-Portfolio”) or a more aggressive investment portfolio with higher risks and higher return expectations (e.g. the “Liability Opportunity-Portfolio”). Active portfolio mandates aim to achieve an additional, uncorrelated return contribution enhancing the out-performance further.
Active dynamic approach: Considering accounting and regulatory requirements, a portfolio which promises a higher return in the long run can only be chosen when shortterm risks are controlled. In fact, in many cases it has, for example, to be ensured that the funding level does not fall below a certain threshold level over time or regulatory stress tests need to be passed on a quarterly basis. Dynamic LDI strategies provide a solution for this problem: depending on the available risk budget (e.g. reflected by a predefined funding level), the asset allocation is dynamically adjusted over time. The DSP-LDI strategy developed by risklab provides a tool to hedge liability risks by holding a portfolio with matching character in times when risk budgets are low (e.g. funding levels are low) and in times when risk budgets are high, it provides return opportunities by holding a portfolio with liability opportunity character. The DSP-LDI strategy includes not only pro-cyclic but also anti-cyclic elements and combines the return advantages of an aggressive strategic asset allocation with the risk advantages of a liability-matching strategy.
For most pension investors the goal of outperforming liabilities in the mid- to long-term, while actively controlling and managing financial, regulatory and balance sheet risks in the short-term is quite intuitive. The goal indicates, however, a rather complex decision situation which requires a fully integrated asset-liability model being capable of assessing the economic, regulatory and accounting impact resulting from a chosen dynamic pension investment policy – over a long-term investment horizon and under fund-specific conditional investment actions over time. The integrated asset-liability modelling fully reflects the concept of LDI and offers an extended holistic investment approach. Figure 2 illustrates the various intertwining components of a multi-period simulation framework needed to provide realistic projections of investment and accounting results.
Figure 2: Integrated Asset-Liability Framework for Multi-Period Projection of Future Funding Situation
A systematic process to derive an optimal LDI strategy starts with the explicit consideration of the investor’s individual goals, risk constraints, and the regulatory and accounting framework. They build the basis for judging the feasibility and quality of investment solutions.
A thorough assessment of the stochastic characteristics of the underlying liabilities and their sensitivities to all relevant risk factors is essential. These risk factors, like inflation, the term structure of interest rates, and the growth rates of salaries need to be modelled individually and consistently. Based on a set of different scenarios of possible future economic environments, a consistent stochastic description of liabilities will be given. This stochastic modelling of liability behaviour goes far beyond traditional actuarial modelling with its deterministic (expected) cash flows. The subsequent identification of a liability benchmark serves as a first orientation of the characteristics needed on the asset side. The value and stochastic nature of liabilities depend on the valuation methodology, which is set by regulation or given by the accounting rules. So, for example, in the context of the local accounting system the characteristics of the stochastic liabilities could also be reduced to a simple target return level describing the average growth rate of (book) liabilities. Under such conditions we can set this specific return requirement as the target level for the asset side. Then the LDI problem is reduced to an “Absolute Return” investment task which can, for example, be solved applying intelligent dynamic asset strategies focusing on asset return and asset risk profiles.
On the asset side, consistent modelling can be achieved by applying the same risk factor scenarios as were used for the generation of the liability scenarios. To derive an optimal risk profile relative to liabilities, the universe of assets should be rather broad and comprise various building blocks with different return profiles reaching beyond traditional equity and fixed securities including alternative assets, derivatives, and alpha from active management.
In a multi-period investment situation it is also important to consider path-dependent portfolio adjustments over time, which are the result of reactions to changes in the market environment or on the liability side. Dynamic LDI strategies, such as those drafted above, provide a rational and ex ante designed risk management solution.
The traditional approach of “Surplus Optimisation” aims to optimise the asset portfolio relative to the liabilities, where the relation is typically expressed in terms of a funding level or the surplus of assets over liabilities. With this the basic idea of LDI is addressed rather explicitly. In combination with a stochastic modelling of assets, a strategic portfolio allocation is optimised with respect to the investor’s preferences on “Surplus Return” and “Surplus Risk” regarding the behaviour of the funding level as the relevant decision figure. Interactions of assets and liabilities, scenario-specific cash flows over time, and especially conditional dynamic investment behaviour are explicitly taken into consideration within this optimisation model.
The integrated modelling approach enables us to identify a range of surplus-efficient investment portfolios as illustrated in Figure 3. The portfolios differ from each other in their composition and hence with respect to company-specific target values and risk constraints.
Figure 3: Dynamic portfolio positioning along surplus-efficient frontier within an integrated asset-liability modelling framework.
Choosing the “Liability Matching-Portfolio” reflects the attitude of a strongly risk-averse investor. It embodies an immunisation strategy with respect to the liabilities in order to avoid any financial risk in the pension trust. In fact, liability matching in an asset-liability context is equivalent to investing in risk-free assets in an asset-only context. This view is often expressed in corporate finance texts following the argument that in order to maximise shareholder value, a company should assume only risks related to their operative core business and not in connection with pension investing because the required risk capital to cover financial risks can be used more efficiently in the operative part. Depending on the desired hedging intensity, different strategies and related products are available including, for example, pooled LDI funds to enable duration matching and the construction of a direct “cash flow matching” portfolio. Long-term and inflation-linked bonds are likely chosen as risk minimising vehicles for the asset allocation. In this context LDI is called “Long Duration Investment Strategy”. Unfortunately, the possibilities to construct an efficient hedge to net off mortality risk are rather limited. Only a few and not very liquid “Longevity Bonds” are offered in the market and in most cases they embody a high basic risk related to the individual mortality exposure. Thus perfect liability matching is not achievable under practical conditions.
A reasonable approximation of a low mismatch risk exposure under realistic conditions is what Figure 3 shows as the “Liability Defensive-Portfolio”. This portfolio exhibits only a small amount of tracking error relative to liabilities. Investors who want to limit liability mismatch risks but also intend to generate additional investment return through their asset portfolio in order to reduce, for example, the financing costs of his promised obligations or to provide the necessary services to be successful in a competitive business environment might prefer the “Liability Balanced-Portfolio”. Depending on individual risk preferences and risk capabilities, a more aggressive investment portfolio (e.g. the “Liability Opportunity-Portfolio”) with higher risks but higher return expectations might be suitable.
Finally, the DSP-LDI dynamic strategy will shift the asset allocation conditionally on the available risk budget (e.g. reflected by a predefined funding level) along the efficient frontier depicted in Figure 3, with different surplus-efficient portfolios being chosen over time. This dynamic risk management provides protection against negative market movements relative to liabilities and thus a solution to the conflicting dilemma of long-term pension investment goals versus short-term (regulatory) risk requirements.
The optimal LDI strategy is selected on the basis of the fund’s individual goals and the (quantitative) key figures, which are used to manage the business. On the risk side these are, for example, ruin probabilities, the risk of a deterioration of the funding level, the risk of excessive additional contributions or the danger of violating regulatory requirements. Simple risk measures like the volatility of the asset portfolio are not sufficient to cover these aspects. On the return side the fund wants to maximise investment returns to improve security and benefit level to its members. A return goal could also be that the sponsor company wants to minimise the net financing cost of its given obligations, considering the side effects of the investment policy on its profit and loss statements, balance sheet ratios or liquidity status. Such effects are shown realistically by scenario projections of key decision variables over a multi-period investment horizon.
The advantage of an active and dynamic LDI-solution is illustrated in Figure 4: By choosing the DSP-LDI strategy funding costs of a pension plan can be reduced in this example substantially compared to a “Liability Matching” solution, while at the same time the underfunding risk is significantly lower and still within feasible risk limits than a return-comparable static SAA.
Source: Own Calculations. The pension fund is a closed final salary plan. Only retirees are left (“no active members“). Initial DBOs of EUR 540 mn according to IAS 19. The static “Liability-Balanced Portfolio“ consists of 41% bonds, 39% stocks, and 20% alternative asset classes.
Figure 4: Illustration of DSP-LDI effect: Reduce funding costs of a pension plan while explicitly controlling under-funding risk.
The practical implementation of an active-LDI investment solution fits well into already existing structures using active portfolio management in sub-funds combined with a dynamic risk overlay component. In a typical setup, the strategic positioning (example being the Liability-Balanced-Portfolio), is implemented in several active and/or passive sub-funds using fixed-income, equity or alternatives mandates employing different and diversified asset return sources. Dynamic risk management is done within a separate overlay fund not conflicting the sub-fund managers. In total, this generates an LDI solution with a clientspecific risk-return profile.
Figure 5: Solution set-up: Active-LDI investment concept using active portfolio management in sub-funds combined with a dynamic risk overlay component
Summing up our suggestions on pension finance based on an active and dynamic LDI concept we find that under the changed accounting and regulation perspective, an integrated asset liability optimisation approach tends to become the necessary framework for comprehensive decision-making. This is especially relevant in the rather complex world of pension investing with pronounced changes in (book and market) values and many interdependencies between assets and liabilities. Moreover, dynamic LDI strategies, like the DSP-LDI strategy, are an efficient and flexible tool to control market value risks over time and bring the long-term investment perspective practically in line with short term risk controlling.