In early 2005, as we reviewed the results of the 2004 triennial actuarial valuation, what became immediately apparent was a need to review the London Pensions Fund Authority's (LPFA's) investment strategy. A track record of full funding since inception in 1990 was at an end. The oft-told story of falling global equities markets and bond yields as well as increased longevity assumptions all combined to push the Authority's funding level significantly into deficit for the first time.

This article explores LPFA's experience with how it chose to respond to the funding deficit and eventually to the development of a Liability Driven Investing (LDI) programme. I will touch on the reasons such a programme was selected, the issues we encountered along the way and the lessons learnt. And, when all is said and done, whether we believe LDI will eventually lead the scheme out of deficit into full funding.

LPFA is one of the largest Local Government Pension Schemes in the UK, managing a multi-employer defined benefit (DB) pension scheme with assets of £3.5bn (€5.1bn) across two pooled funds.

One of these pooled funds has extremely mature liabilities that are inflation-linked in nature and the historic allocation to UK index-linked gilts was regarded at the time as an appropriate matching strategy. However, this approach was ultimately unsatisfactory in achieving that objective. With the funding level falling to 91% at the 2004 actuarial valuation and particularly due to the extreme maturity of the liabilities and lack of formal sponsor, a fundamental decision was made by the trustees to strengthen the link between the assets and liabilities.

As a public sector pension fund subject to government legislation, LPFA was required to retain all aspects of its final benefit plan design. At the same time employers were demanding contributions stability with the trustees having a preference for an investment strategy that offered ‘no surprises' and aligned the investment and funding objectives. As well, fund members wanted assurances their pensions would be paid on time (this point is important as the fund's ‘closed' nature means significant cash outflows have to be funded each year).

Risk management had historically been incorporated in the asset-liability modelling (ALM) process but not extended to operational management of the investment strategy. Agreement on how long-term objectives and investment risk should be budgeted across stakeholders was to become an important factor in developing the investment strategy.

With the authority's ultimate objective being the payment of pension obligations it was important to ensure sufficient assets were available to meet these liabilities. From there it became a logical step to instal the liabilities as the fund's benchmark against which to measure the investment performance. If the fund could not exceed the return on liabilities (movement in the liabilities due to the cost of pension obligations accruing, interest on the liabilities and the impact of changes in the inflation and discount rates) it could not achieve full funding. This was a significant development, particularly in the UK, where the traditional approach has been an asset-only and index-driven framework across the pension fund peer-group.


n initially reviewing what was to become the liabilities-driven investment strategy we went back to first principles to better understand the fund's liabilities. The liabilities were deconstructed from an actuarially constructed aggregate number into a series of pension payments, or cash flows (Figure 1). A comprehensive picture of the liabilities' characteristics was built up given a set of inflation and interest rate assumptions and further developed to undertake sensitivity analysis for changes to these assumptions.

A long-term objective was set in consultation with the investment consultant: to outperform the liabilities by 1.5% per annum net over a 20-year period. It was a balance between the trustees' desire for downside protection of the funding level and the need to outperform the liabilities and eliminate the funding deficit. However, a 20-year period, whilst setting a defined end-date for achieving full funding, to some extent removes accountability from those responsible for the investment programme. Short-term volatility makes it difficult to make an annual judgement as to the expected long-term success of the investment strategy.

Implementation of the strategy (see figure 2) brought with it a range of complex operational issues that did not exist in the previous (long-only bond/equity) investment strategy. As UK Local Government Pension Scheme investment regulations are not prescriptive on direct ownership of swap contracts, LPFA chose to implement its LDI programme via a pooled fund structure. Without any experience in the types of products and strategies the fund was to implement, it was very much a learn-as-you-go approach and quite strongly guided by the investment consultant and investment managers. Nonetheless, in working through the transition to the LDI strategy a number of issues arose:

o Investment manager agreements: the agreements are as complex as the strategy in areas such as the investment universe/constraints imposed, benchmark and performance fee definitions and calculation and the type of the structure used;

o Implementation timing: given the level at which the swap market was trading and increasing demand/supply imbalance should the decision to implement be strategic or tactical?: given the level at which the swap market was trading and increasing demand/supply imbalance should the decision to implement be strategic or tactical?

: given the level at which the swap market was trading and increasing demand/supply imbalance should the decision to implement be strategic or tactical?

o Ongoing monitoring and reporting: with the strategies managed within a pooled fund structure what type of reporting do the trustees require and to what extent is this available, ie, unit price valuation or details of all holdings within the pooled fund?

o Accounting: how are the products and securities to be accounted for in the accounts, including the impact of any leverage? How are the portfolios valued if they are generally pooled structures, or do the manager valuations become the default?

o Custody: is the custodian able to handle the complexity of LDI mandates;

o Performance measurement: What are the capabilities and experience of the provider in measuring the return of the liabilities? Is there access to the liabilities cashflows and discount curve? Are the investment manager's valuations being used?

When initially reviewing the LDI market 12 months ago it was a sector at an embryonic stage in the UK. There was considerable emphasis on the use of cashflow matching and in particular the ‘matching' properties. The key focus was on those risks that could be hedged; the economic risks, inflation and interest rate, without reference to those risk exposures that remained, such as:

o Deficit: a proportion of the liabilities cannot be hedged due to their being insufficient assets so the fund remains fully exposed to the economic risks;

o Longevity: longevity assumptions are increasing and will adversely impact the liabilities valuation;

o Alpha: the cashflow matching mandates have two components, matching and alpha generation. To the extent the alpha generation component does not achieve its target the deficit will not be recovered in the required time period;

o Equity: to increase investment performance there may be an allocation to equities, which has a high degree of volatility relative to the liabilities;

o Actuarial assumptions: to the extent different assumptions are likely to be applied at the next actuarial review the liabilities valuation and the matching strategy will be affected;

o Legislation: the liabilities will be impacted by changes to UK local government scheme regulations;

o Basis risk: The actuarial valuation has hitherto used a gilt-based discount rate, yet most LDI programmes are implemented using swaps rather than physical gilts for a variety of reasons. To the extent that the swap and gilt curves do not move in line there will be a mismatch. This may lead to the situation that the managers outperform their benchmark but the fund underperforms the return on the liabilities;

o Counterparty risk: Although the LDI programme is invested via a pooled fund structure, rather than the direct holding of swap contracts, LPFA remains exposed to the risk that the other (counter)party to the swap contract defaults.

The swaps market is generally more liquid and provides better matching opportunities across the yield curve, allowing bespoke solutions to be developed, compared to the physical gilts market. The use of mandates with a combination of matching and active components may not be the most efficient and effective approach to achieving the fund's investment objective. There are two aspects to this. Firstly, is it more efficient, and does it offer the fund greater control, if the total matching programme is with a single manager? Secondly, what are the practical implications, and cost, of unwinding the matching component if one of the investment managers is not achieving its target outperformance from the actively managed component?

It was important the trustees took on comprehensive advice before approving the strategy and were not wholly reliant on the investment consultant or investment managers. There is no monopoly on information and there will always be more than one ‘optimal' investment strategy.

It was important for the trustees to be actively involved in the manager selection process and setting in place a formal, rigorous, due diligence process prior to putting in place the agreements. A reporting and monitoring process should be agreed at this time. Due to the complexity and bespoke nature of the strategy it is even more important that the trustees understand how the fund's assets are being managed and to ensure this is consistent with what was signed up for. And not to forget there is likely to be a significant increase in management fees if the LDI assets are sourced from internally managed assets or an externally managed fixed interest fund.

Strong relationships should be developed with the investment managers and regular meetings held to discuss all aspects of the mandate, both investment-related and operational. At the end of the day the key question is whether this is explained to the trustees in such a way that is easily understood?

There are many salutary lessons to be learned. Firstly, it is important to focus on your own fund's objectives, not those of your peers. This may require breaking long-held habits but it is the first step to achieving the objectives set for the fund. Another way of looking at it is to say the appropriate trade-off between reward and risk for an optimal outcome is fund specific.

The current investment environment and the range of structures and products are complex and it has been a steep learning curve for all parties. But let's be clear; the investment managers and investment consultants offering LDI solutions do not have all the answers. The upside of getting a bespoke solution is that there is no template answer. A review of LPFA's LDI programme after its initial 12 months would regard it as a successful decision.

Paul Kessell is a member of the investment team at London Pensions Fund Authority, a £3.5bn (€5.3bn) UK local authority pension fund