Liability Driven Investment

5. Monitoring such a structure: Main elements

[this page | pdf | references | back links | custom searches]

Return to Overview

Next page


5.1          There are three key elements to the above structure that might need monitoring (other than the usual monitoring that would be carried out even if no LDI approach was being adopted on the return seeking portfolio):


(a)    The (actively managed) underlying bond portfolio. This would be assessed as usual for the asset management product in question. For example, if it involved management of a credit portfolio against a market index then performance and risk measurement and attribution analyses versus the benchmark in question might be reported as per the asset manager’s/pension fund’s usual reporting cycle.


(b)   The (passive) swaps overlay. This might for simplicity also be reported upon to a similar frequency, although most attention would be focused on those occasions when the swap positions needed to be altered.


(c)    The effectiveness of the choice of swaps overlay structure in relation to the scheme’s liabilities. Various approximations will have been interposed between the precise liability model available from the actuary and the precise structure of the swap portfolio. The swap portfolio being ‘execution-only’ in nature, this element of the decision-making is actually one that lies with the trustees, albeit only after taking advice from other parties.


5.2          The key additional requirement is to construct some sort of liability benchmark (or index) that reflects in a market-orientated way the nature of the liabilities. Constructing such a benchmark may also directly guide the choice of swaps to hold within the overlay portfolio.


5.3          The most obvious way to proceed is first to develop some cash flow projections, differentiating between ones with different sorts of economic sensitivities (particularly those where the sensitivities have option-like characteristics, such as LPI). For example, the liability flows might be differentiated by year of projected payment into those that involve:


(a)    Fixed monetary sums, e.g. those arising from benefits not subject to any increases.


(b)   Fully RPI inflation-linked sums, e.g. benefits subject to full RPI linked increases.


(c)    Sums that increase on a year by year basis on some more complicated measure driven by inflation at that time, e.g. LPI-type increases in payment. For these sorts of liabilities, the expected outgo during a given future year can still be derived from a single expected amount at outset, together with the history of RPI increases since then. If different ceilings, say 2.5% and 5% pa caps, apply then these flows should in principle be differentiated, as swaps to match them exactly would also differ.


(d)   Cash flows governed by more complex increase formulae dependent on multi-year investment or economic conditions. At least in principle, benefits linked to LPI in deferment fit into this category. The big difference between these sorts of cash flows and the sorts referred to in (b) or (c) are that they in principle require multi-dimensional matrices to specify as they depend jointly on date of withdrawal, assumed date of retirement, assumed date of payment and (for those already deferred pensions at outset) on how large RPI increases were prior to the start of the projection relative to the caps and floors present in individual members’ benefits. As with (b) and (c) they also depend on RPI increases post the start date of the projection.


Contents | Prev | Next

Desktop view | Switch to Mobile