Joseph Mariathasan assesses who is best placed to assume asset allocation decision-making for pension funds

Recent publicity surrounding multi-asset target return funds is another sign that asset allocation is finally getting the attention that it deserves. However, it is not clear that there are enough sources of real and independent expertise to satisfy marketplace requirements, even if it has not yet realised it.

Richard Graham, head of UK institutional business at Barings, explains: "We expect a very different investment environment over the next ten years where asset allocation is again the main driver of returns and portfolio diversification is key to delivering real returns and reducing risk."

Perhaps the most critical allocation decision investment boards face is allocating responsibility and accountability for asset allocation.

Guy Fraser-Sampson in his recent book Multi Asset Class Investment Strategy declares that it "may seem strange, given the consistent success in recent years of the Yale Model, as described in David Swensen's best-selling book (Pioneering Portfolio Management, 2001), that a multi
asset class approach to investment strategy for institutions such as pension funds should still be a controversial subject, and require justification."

Many would argue that the investment industry is merely going full circle after meandering down a route that, with hindsight, turned out to be both illogical and disastrous. The fundamental point driving the asset allocation decision debate that has to be borne in mind is that the "risk" any stakeholder is concerned about is the failure to meet his own objectives. As the objectives of the participants in the asset allocation decision tree are clearly different, their perceptions of risk are different.

Controversy arises when one or more stakeholders judge their economic interests are not being met by strategies being adopted or suggested by other players. Even though each player may be acting in a rational manner to control their individual risk perception. On the other hand, controversy can be potentially turned into collaborative behaviour when each participant can see evidence that its own objectives can be fulfilled through group actions.

Prior to the mid-1990s, the majority of mandates in the UK were for a balanced approach, which enabled fund managers to use their own discretion in asset allocation. The intention was that their own business objectives were clearly in line with the objectives of their clients. However, once the average asset allocation of the total universe of pension schemes started to be published, fund managers leapt onto these statistics as a method for controlling their own business risk by ensuring their own allocations did not deviate too much from the average.

This of course, had no real relevance for individual schemes and led to absurdities as managers remained in markets like Japan because they believed last year's statistics indicated everyone else had a certain exposure, leading to slow changes in asset allocation despite large market movements and mis-valuations.

One reaction to this was for
consultants to move to specialist mandates, through core and satellite approaches. Specialist mandates require specific narrow benchmarks for performance comparisons which essentially moved the problem from the asset allocation level to the
stock level as fund managers again controlled business risk by keeping deviations from an index and tracking errors within tightly constrained boundaries.

This again led to absurdities such as schemes with a 50% allocation to UK equities and 50% overseas in March 2000, when Vodafone was over 13% of the FTSE 100. There were fund managers who believed they were controlling risk by reducing their weighting in Vodafone to, say, 10% of their UK portfolio. From the manager's viewpoint, they were at risk because they were underweight Vodafone. From the scheme's viewpoint they were at risk in being so heavily exposed to Vodafone irrespective of index weighting and many would have had more exposure to a single stock than they would have had to whole asset classes such as private equity. Despite this, they were led to believe that investing amounts as small as 5% in a range of alternative assets would have been a risky strategy.

The problem that still exists in the current arrangements for fund management is that institutional pension schemes find it difficult and perhaps almost impossible to obtain objective asset allocation advice that is not geared to selling a specific range of products from a fund manager. Few fund managers have the expertise to move away from individual product silos to give asset allocation advice. Fewer still have found a way to make money from giving independent advice. The business model that works is to sell in-house multiple strategy funds in the guise of asset allocation, but this can hardly be regarded as objective and necessarily in the best interests of a scheme.

Investment consultants in contrast, may be independent, but they lack the market knowledge to undertake market driven asset allocation decisions. To undertake asset allocation objectively requires a very high level of skills and experience.

Typically this rules out investment boards, whose function is policy, governance and oversight. It arguably also rules out investment consultants, whose skill set has a theoretical bent, and whose market insights tend to be derivative, gleaned from investment managers.

Some fund managers may possess the skills, but lack the objectivity or the incentive structures to recommend best of breed third party managers.

The problem that arises is that the mismatches between involvement and responsibilities on the one hand, and skills, experience and resources on the other, end up damaging the wealth of the fund's beneficiaries.

A key issue for long-term asset allocation decision-making is another mismatch, that between agents and principals.

All of the players in the asset allocation decision-making value chain - investment managers, investment consultants and trustees - are agents acting for principals, the ultimate beneficiaries. As a result, they are all exposed to forms of career risk that create a mis-alignment of interests with beneficiaries. The most dramatic instance being horizon mismatch.

While a pension scheme may have a time horizon of 30 years or longer, no agents have horizons even half as long. This gives rise to shorter-term risks of large relative underperformance which all institutions have to grapple with. A few large US institutional investors stand out as role models. The Yale and Harvard Endowments have been consistently in the top percentile over 20 years by being dramatically different. At times Yale has had a 60% exposure to private equity, but both Yale and Harvard are famed for their forays into a diverse range of alternative asset classes. Both are large enough to support a professional investment staff, and both
have relatively few principal-agency frictions.

They give clear examples that in-house management is one way that can, if structured well, provide a good solution to the competing objectives with in the asset allocation decisions of large funds.

Staff are employed by the entity guaranteeing the pension promise or having ownership of the assets, with a focus on investment performance to meet liabilities, rather than beating arbitrary benchmarks to attract more business. They are able to place much greater emphasis on asset allocation while retaining the ability to outsource investment mandates to specialists when required for reasons of expertise or cost-effectiveness.

Internal fund managers have, in theory at least, the huge advantages of stability and security, essential to give confidence to take decisions that may be radically different from a peer group. The best example of this was probably George Ross Gooby, the Imperial Tobacco pension fund manager during the 1950s. He took a large bet in moving into UK equities from the more traditional bonds and thereby introduced the cult of the equity into institutional investments. He did this without reference to other schemes and without performance comparisons against arbitrary benchmarks.

In today's world, such a stance would be difficult to maintain even for an in-house scheme and impossible for a third party manager.

Is there a case for resurrecting in-house schemes?

There is certainly a case for structures that enable fund managers to have a closer alignment of interests with beneficiaries and sponsors and in-house schemes would have this advantage. But this also relies on maintaining a long term trust between sponsor and internal managers.

Indeed, in the case of Imperial Tobacco, the in-house managers took a strong value stance during the TMT boom that led to significant underperformance. The trustees reacted by shutting the investment team in 1999 and outsourcing. So, despite achieving returns that in absolute terms were probably attractive with hindsight, the lack of trust towards the in-house team led to events that left no-one better off.

External fund managers managing only part of a portfolio are not in a position to take on asset allocation decisions; investment advisers are reluctant to take responsibility for investment decisions they could be sued for rather than giving advice that need not be acted on, even if in practise, many trustees would prefer following that advice blindly. Adopting an approach of in-house schemes concentrating on a core expertise and outsourcing specialist funds externally would appear to make pragmatic sense. The schemes could consider a hierarchy of activities.

The most important would be to focus on utilising a risk budgeting approach to asset allocation. Developing in-house expertise in particular market segments may be appropriate.

Clearly, incorporating a wider range of asset classes would require expertise in new areas that in-house schemes may find unrealistic to acquire. Many schemes may also be too small to justify separate management and it is interesting to look at the example of the Dutch based industry schemes as a possible way forward even for
the UK.

A clear strategy with a much wider range of asset classes, would involve outsourcing many specialist mandates, and there may also be a case for outsourcing a core index or enhanced index portfolio to those firms that can manage such portfolios at low cost. Large funds could justify developing expertise in constructing portfolios of hedge and private equity funds and thereby avoid the double set of costs entailed by seeking a fund of funds route.

In addition, most schemes are not able to develop internal teams that are of sufficient calibre to take on the role of asset allocation and will turn naturally to investment consultants.

Yet the temptation of many traditional investment consultants towards asset allocation has been to use global capitalisation weighted indices as the benchmarks wherever possible.

This is a throwback to a belief that markets are efficient. Despite all the evidence that many players in the marketplace, such as central bank foreign reserve managers, are driven by non-profit maximising objectives.

Indeed, with regard to the US Treasury market, research by Francis and Veronica Warnock at the University of Virginia estimated in 2006 that "had there been no official flows into the US government bonds over the past year, the 10-year Treasury yield would currently be 90 basis points higher." As they point out, the results are "eminently plausible given that foreigners currently hold more than half of the US Treasury bond market".

To undertake asset allocation effectively does require a framework for valuation that seeks to identify long term relatively undervalued assets and disinvest or even go short of asset classes that appear significantly overvalued, which means a complete rejection of the market efficiency theorems and the use of capitalisation weighted market indices.

This approach would be focused on absolute returns rather than relative, and on long-term performance rather than short term. There appears to be a greater acceptance and interest in managing investments with absolute return or liability driven benchmarks from entities such as family offices and private client fund managers where there is a much closer relationship between the ultimate beneficiaries and the managers than in the institutional pension fund market. Endowments with their own investment departments such as Yale have also led the way in adopting very successful asset allocations that are radically different from the average stance based on a capitalisation weighted approach concentrating mainly on equities and bonds.

One model that could address the issues discussed in this article is that of a strategic partnership between (large) client and fund managers, where all parties share research and are expected to be mutually critical.

Another approach could rely on the new generation of fiduciary providers that seek to combine asset allocation with manager selection, taking into account the current market conditions for, say, value versus growth in the case of equity managers.

It will be interesting to see whether they win significant market share or whether traditional investment consultants adapt their offerings in a probably revolutionary rather than evolutionary change to their own structures and offerings.

One thing that is clear is that the whole area of decision making and governance in asset allocation needs to be brought into the open and a real debate started on who is best placed to give advice, who should take responsibility, and how should the various parties in the asset allocation debate get fairly remunerated for giving advice.

1 Warnock, Francis E. and Warnock, Veronica Cacdac, "International Capital Flows and U.S. Interest Rates" (September 2006). FRB International Finance Discussion Paper No. 840
Warnock, Francis E. and Warnock, Veronica Cacdac, "" (September 2006). FRB International Finance Discussion Paper No. 840