Meeting documents

Return to Agenda

ITEM PF6

PENSION FUND COMMITTEE – 26 AUGUST 2005

ASSET ALLOCATION

Report by the Head of Finance & Procurement & the Independent Financial Adviser

1. Asset allocation is the distribution of investments across categories of assets, such as equities, bonds, property and cash. Asset allocation affects both risk and return and is the most important component of a Pension Fund’s overall investment strategy. It has been estimated that up to 80% of investment performance can be attributed to asset allocation.
2. In March 2001 the Myners review of Institutional Investment was published and it covered the subject of asset allocation in some detail, describing it as "an under-resourced activity." Prior to the review the majority of pension funds operated a peer group benchmark whereby a manager’s investment performance was measured against a chosen peer group. The Oxfordshire Pension Fund’s benchmark was the CAPS median fund return, which was made up of over 1600 UK pension funds, predominantly in the private sector. The consequence of this arrangement was that Schroders and Deutsche Asset Management, who were Oxfordshire’s managers at this time with multi asset briefs, tended to structure their portfolios in line with the CAPS benchmark.
3. Myners was critical of the Peer Group arrangement arguing that the resulting "one size fits all" asset allocation was not necessarily the right one at the individual pension fund level. He further argued that strategic asset allocation should primarily be determined by an individual fund’s liability structure and that each fund should consider having its own customised benchmark.
4. In 2002 the Oxfordshire Pension Fund Committee decided to move to a customised benchmark. This process involved the then Actuary, Watson Wyatt, being appointed to carry out an asset liability study for the Oxfordshire Fund. This exercise culminated in a new asset allocation, which was considered to be a better match for the Fund’s liabilities. The main outcome of this exercise was a reduction in equities (UK and overseas) from 75% to 67% and an increase in bonds from 15% to 22%, together with a rise in alternative assets (property, private equity and hedge funds) from 6% to 11%.
5. Bonds are currently considered by many to be the best match for pensions in payment because they are perceived to match actuarial projections and their prices in the shorter term tend to be less volatile than equities. However, over the longer term equities have tended to outperform bonds so there needs to be an acceptable balance of risk and reward when constructing an asset allocation strategy. The agreed strategy will impact on employer contribution rates and council tax.
6. Following the asset liability exercise, which also coincided with a period of manager underperformance, the Pension Fund’s management arrangements were reviewed. The review culminated in the implementation of a more specialist management structure with the existing four managers being appointed in July 2003.
7. The asset composition of the strategic benchmark, together with the managers and the assets for which they are responsible, is shown in Annex 1 (download as .xls file).
8. In February 2005 the Pension Fund Committee agreed to a process for reviewing the strategic asset allocation, coupled with a report on a tactical asset allocation overlay strategy. It is felt that now is the right time to review the asset allocation, following the results of the 2004 actuarial valuation and changes that have occurred in the investment markets since the last review was undertaken. This report considers these issues.

The Strategic Asset Allocation Review Process

9. In accordance with best practice as set out in the Myners Review of Institutional Investment and the Review of the Actuaries Profession by Sir Derek Morris, discussions were held on the strategic asset allocation with the actuary and the investment managers, particularly with UBS, who was appointed the multi-asset manager partly because of their expertise in asset allocation. It was evident during the manager selection visits that few multi-asset managers had a sound asset allocation process and that none of them matched UBS. Officers and the Independent Financial Adviser also attended several seminars and conferences on asset allocation over the past year to keep abreast on the latest thinking. In addition, officers discussed asset allocation with Barings ING and Legal and General to obtain further independent views.

Findings of the Review

10. Annex 2, Table 1 (download as .doc file) compares the Pension Fund’s current liability profile with the one prevailing in 2001, which was used for the 2002 asset liability study. This shows the profile has only changed modestly between 2001 and 2004. The main development being a 4.6 percentage point decline in pensioners, balanced by rises in deferred pensioners and contributing employers, meaning the Fund is slightly less mature than in 2001.
11. The Actuary was consulted on his forecasted longer-term changes to the Fund’s liability profile. He reported "I would expect pensioner liabilities to remain close to the current levels of around 40% of the liabilities for some time to come unless there are significant changes in membership profiles, changes in the scheme liability profile or assumed rates of investment returns."
12. Since the last review was undertaken in 2002 a number of factors has changed and trends developed including the following:

• Most pension funds are now showing funding deficits compared to the years up to 2000, the peak of the equity market, when surpluses were a norm. Most local authority pension funds are now heavily in deficit, with Oxfordshire at a funding level of 65%. It was decided in the Funding Strategy Statement that this would be recovered over a period of 25 years.
• Alternative assets such as commercial property, private equity and hedge funds are looked on with increasing favour as they are regarded as having a lower correlation, to a varying extent, to equities.
• Equity markets have become much more global and there has been a gradual reduction by Pension Funds in their UK equity allocations and an increase in their overseas equities. The WM Local Authority statistics show that between 2001 and 2005 the proportion of UK equities held by local authority pension funds fell from 50% to 40%. Over the same period there was an increase in overseas equities from 21% to 29%. This trend is expected to continue in recognition that a more diversified equity portfolio produces a better risk reward trade off.
• The trend to higher bond weightings, particularly in corporate defined benefits schemes, has gathered momentum. The reasons for this trend are discussed in more depth later in this report.

13. Officers and the Independent Financial Adviser considered a number of different strategic asset allocation strategies and the main ones of these are set out in Annex 3 (download as .xls file). UBS ran the different strategies through their own portfolio computer model in order to analyse their long-term risk and reward characteristics. The results of this exercise are also shown in Annex 3 (download as .xls file) and include the longer-term investment return and risk projections. The sharpe ratio, which is explained on Annex 3 (download as .xls file) , attempts to show the most efficient portfolio in terms of both its risk and return characteristics.
14. The most appropriate asset allocation for Oxfordshire is dependent on members risk appetite and whether the emphasis should be on a lower risk/lower potential return or higher risk/higher potential return objective. If a stable, or even modestly declining, employer contribution rate to give greater certainty in the level of council tax, is the main objective then this would indicate a higher level of investment in bonds. Portfolio model A in Annex 3 (download as .xls file) has the largest holding in bonds at 36%. This is the portfolio with the projected least risk but also the lowest longer-term investment return. Annex 2 (download as .doc file) shows that the actuary is projecting that the deficit can be eliminated in 25 years if the longer-term investment return is 6.2%, all other factors being equal. Therefore, portfolio A would comfortably achieve this objective with its projected return of 8.68%. However, this strategy would tend to lock in the funding deficit for a longer period.
15. If on the other hand the main objective is to reduce the deficit, sooner rather than later, then a higher level of equity investment should in the long term provide rising real returns. However, equities tend to be far more volatile than bonds in the shorter term giving less certainty on the employer contribution rate and so the level of council tax. The current trend towards liability driven investment (i.e. just focusing on a fund’s liability profile) with an emphasis on bond investment does not normally over the longer term produce as high real returns as equity investment.
16. The trend to a higher proportion of bonds in pension funds, particularly defined benefit funds is driven to a considerable extent by the corporate sector, particularly quoted companies. Pension funds and funding deficits now feature prominently in company balance sheets with the result that corporate treasurers can only take a three to five year view of funding the pension fund and this drives them to higher bond investment even though it locks in deficits. Furthermore, the large closure of final salary schemes in the corporate sector is generating an increase in their maturity profiles, which encourages a higher bond weighting.
17. In contrast local authority funds are less mature and in paragraph 11 it was reported that the Actuary does not foresee changes to this position over the longer term. Furthermore, Oxfordshire, along with many other local authorities, has decided to reduce the funding deficit over 25 years and this would facilitate a higher weighting in equities, but with shorter term volatility and so a shorter term risk to the employer contribution rate and level of council tax. However, for long term investors inflation is the main risk. The historical returns of equities, bonds and cash over the long term, together with the probability of equities under-performing bonds and cash over varying time periods are shown in Annex 4 (download as .doc file). The estimated future returns of asset classes by UBS are shown in Annex 5 (download as .doc file). These returns were compared to Hewitt Bacon and Woodrow’s Investment Consultancy Arm’s own returns and appear to be fair.
18. The Independent Adviser’s favoured asset allocation is shown in Annex 3 (download as .xls file) column C. This model produces the best projected investment return but has the second highest risk characteristics. The sharpe ratios were very similar for all four portfolios, which implies that the risk reward trade offs are of a similar nature. The changes proposed by the Independent Financial Adviser are not a radical departure from the existing strategic asset allocation in the customised benchmark, but they do represent a move against the current trend by recommending a reduction in the bond weighting from 22% to 16%.
19. There would also be an increase in alternative assets from 11% to 17%. The Myners Review recommended that pension funds should consider the attributes and diversification benefits of alternative assets. Alternative assets not only offer the potential of good real returns but are also lowly correlated to equities and bonds and therefore help to reduce the Fund’s overall volatility.
20. There would be no change in the Fund’s 67% overall equity weighting but it is proposed that there would be a more even split between UK equities and overseas equities, in order to achieve more diversification and thereby achieve a potentially better risk and reward trade off. Hedging back to sterling can eliminate any concerns with currency risk within the overseas portfolio. The most effective means of achieving a more even split on equities is covered in paragraph 22.
21. The reason for recommending the asset allocation model C is that the Pension Fund is able to take a 25 year view (the recovery period for the deficit) and the forecasted higher returns, albeit at the expense of shorter term volatility, should eliminate the deficit sooner rather than later. The sharpe ratio of 0.40 for asset allocation model C (see Annex 3 (download as .xls file)) is similar to those of the other portfolios and provides a reasonable balance between risk and reward on a 25-year view.

Recommended change to Alliance Bernstein’s Investment Process

22. The most effective way of achieving a more even split on equities, discussed in paragraph 20, would be by changing the method by which Alliance Bernstein manages global equities. Currently Alliance Bernstein manages global equities, including the UK, within an Oxfordshire customised Global Diversified Value platform, which has provided extremely good investment performance. Annex 1 (download as .xls file) provides a breakdown of the current global equity benchmark, which has a 36.4% weighting in the UK and a 63.6% weighting overseas.
23. However, it is proposed that the Fund moves into Alliance Bernstein’s standard Global Value product. This portfolio is more concentrated, with 100-140 stock holdings against 150-200 and so it has a higher risk profile than the existing arrangement but it does have the potential to produce a higher return. However, by switching into this portfolio it is also likely to provide the desired split, at the total fund level, between UK and overseas equities. The Global Value product would be constrained by market capitalisation rather than the existing customised global equity benchmark set out in Annex 1 (download as .xls file). Currently under a market capitalisation basis the global equity portfolio weighting in the UK would only be 10% compared to 36.4%. This would have the effect of producing a more even split in equities at the total fund level, though this would fluctuate depending on changes to Alliance Bernstein’s investment strategy. As a consequence of this proposed change to the recommended strategic asset allocation (column C in Annex 3 (download as .xls file)) the percentages for UK and overseas equities respectively are shown in brackets since they will fluctuate.
24. Annex 6 (download as .doc file) compares the two different Alliance Bernstein investment products. The management fees for the Global Value product are higher but to date the extra fees would have been dwarfed by the excess investment returns achieved.

Rebalancing the Portfolio

25. The recommended changes to the strategic asset allocation will necessitate a rebalancing of the existing portfolios. However, it is not envisaged that this will produce major changes because the proposed asset allocation is not too far adrift from the current actual asset allocation (see Annex 3 (download as .xls file)). Furthermore, the Fund has a large in house cash balance at the present time which should help avoid, or at least minimise, any disturbance to the managers’ portfolios. In February 2005 the Pension Fund Committee agreed that the strategic benchmark should be rebalanced every 12 months.

Tactical Asset Allocation

26. While the proposed long term strategic asset allocation of the Fund is structured with the main objective of eliminating the Pension Fund’s deficit sooner rather than later, it does not take any account of the valuation currently placed on assets in the market. Thus, equities could be overvalued absolutely, as in March 2000, or relative to bonds, or visa versa. A tactical asset allocation overlay strategy would seek to take advantage of perceived market misvaluations on a short-term basis by super-imposing a different asset allocation for the whole fund, whilst leaving the underlying long-term strategic asset allocation in place and the managers’ underlying portfolios undisturbed.
27. UBS is a natural choice to operate a tactical asset allocation overlay strategy for the Pension Fund as they are already manager of the multi-asset section of the portfolio and operate asset allocation within that portfolio. Officers received a presentation from another manager who also operates a tactical asset allocation overlay strategy, but using that manager would add a fifth manager to the Pension Fund Management team and the company might operate a strategy at variance to that of UBS, the Fund’s existing multi-asset manager.
28. The tactical asset allocation (TAA) overlay strategy operated by UBS is operated through a separate fund. The Oxfordshire Pension Fund would contribute around 3% of the total Fund’s assets into the UBS TAA fund. The risk reward target would determine the exact percentage. The TAA fund operates through derivatives (futures in the equity and bond markets, forwards in the currency markets etc), which are both quicker and cheaper than buying or selling actual securities. The TAA fund operates within tight control ranges to limit the downside risk. In fact operating a tactical asset allocation overlay strategy reduces the overall risk level of the total Pension Fund, while it raises the expected return. It will be necessary to rebalance periodically the TAA Fund back to its agreed percentage weighting (around 3%).
29. The details of fees for operating the UBS TAA Fund are shown in Annex 7 (download as .doc file). The fees have been compared to those operated by another manager who offers a similar TAA product. The latter’s fee would be more expensive than UBS, particularly as UBS would not charge a fee for operating tactical asset allocation over the multi asset portfolio they already manage for the Pension Fund and also their rates are marginally lower than those of the other manager. It is not considered necessary that tactical asset allocation should be put out to tender, as the Fund would be merely using UBS Services across a broader area of the overall Pension Fund portfolio.

RECOMMENDATIONS

30. The Committee is RECOMMENDED to:

1. approve the new strategic asset allocation in Column C in Annex 3 (download as .xls file) and to rebalance the existing portfolios to reflect this;
2. approve that the portfolio managed by Alliance Bernstein should be moved, phased if necessary, from their Global Diversified Value platform to their Global Value platform;
3. approve that UBS be appointed to manage the tactical asset allocation strategy and that officers negotiate with UBS the exact percentage (around 3%) of the total Pension Fund assets to be contributed to the tactical asset allocation fund.

SUE SCANE
Head of Procurement & Finance

A F BUSHELL
Independent Financial Adviser

Background Papers: Actuarial Valuation Financial Assumptions, UBS Global Asset Management investment strategy papers, Hewitt Bacon and Woodrow’s long-term investment risk and return assumptions.

Contact Officer: Tony Wheeler, Pension Fund Investments Manager (01865) 815287

August 2005

Return to TOP