Pension funds should be more concerned about investment managers’ asset allocation skill, but there’s no easy way to measure it, says Joseph Mariathasan

T he phrase ‘you can only manage what you can measure’ is often noted when assessing fund manager performance through quarterly measurements of relative performance against an agreed equity or bond index. The corollary to this observation, however, is that what cannot be measured easily is often not managed at all, even if it is much more important.

This is the case with asset allocation, which lacks a universally accepted methodology for assessing performance. Because of this, little in the way of resources and little attention is devoted to it, even though it is of far more significance to a pension fund than whether a manager outperforms or underperforms an index by amounts in a year that are a fraction of the daily market movements that were experienced in October.

It is rather like the story of a pension fund trustee walking down a dark street who sees a man scrambling around on his knees under a lamppost. The trustee asks the man what he is doing and he replies: “I am an actuary and I dropped my wallet in the bushes so I am looking for it.” On asking why on earth he is searching for it under a streetlight if he had lost the wallet elsewhere, the trustee is told: “There is more light here.”

As Robert Hayes, managing director of the strategic advice and solutions team at BlackRock notes: “Consultants’ advice is often not measured very closely but managers are measured very closely and probably too frequently.”

Whoever gives asset allocation advice, it is difficult to assess and therefore is often not measured at all. In contrast, forcing managers to stick closely to benchmarks that have no relevance to a pension fund’s liabilities but merely provide a simple way of measuring a manager’s skills is rather like scrambling around under a lamppost purely because there is more light. As Hayes argues, when it comes to asset allocation, “conscious moderate competence is better than unconscious incompetence.”

The problem that the present institutional arrangements pose for investment 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, and fewer still have found a way to make money from giving independent advice.

A 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. As Erik van Dijk, managing director of Compendeon points out: “If you talk to any big house, they would all claim to do asset allocation. But fund management firms usually create a long/short global macro structure. In the end, everyone needs asset allocation but if you want long only, people are not so interested as there are no direct fees.” 

Gap in the market

Not surprisingly, investment consultants are seeing a gap in the marketplace and have been attempting to develop their asset allocation expertise to provide a value-added service that can generate extra fees for themselves. As one manager at a major UK pension fund argues: “The larger actuarial firms are all owned by US parents who would be saying to them that their European subsidiaries advise on five times the asset base of the US, but the US makes twice as much money.”

Asset allocation services are one way of redressing the balance. But while they might be independent, the mainstream traditional firms have lacked the market knowledge to undertake market-driven asset allocation decisions. The issue is, can you adopt asset allocation strategies that span the continuum between the very long term and the very short term?

As Hayes points out, the belief in a long-run premium that you could capture by holding assets regardless of the prices at the start is a sub-optimal strategy. “Should we have the same weighting now when the FTSE is close to 4,000 as we had during the period 2001-2004 when the FTSE was 7,000? How much should we have in credit when credit spreads are 300 compared to the amounts we had when spreads were 50? Having the same weights cannot be the right answer.”

Van Dijk says only a few investment managers provide a strong philosophical framework for asset allocationHe mentions Goldman Sachs, BlackRock, Mellon, Wells Fargo and Bridgewater. New players include a number of consultants such as his own Compendeon and  Cardano.

Perhaps the most signficiant development has been the emergence of the fiduciary model in the Netherlands, which is now being exported elsewhere. In principle, it can be attractive for many, particularly smaller, pension schemes which are, in effect, outsourcing virtually all of the management and operational decision making for the pension scheme. It ensures that the thinking behind areas such as asset liability modelling, asset allocation and manager selection are undertaken in a holistic manner.

But there are also criticisms. As the UK pension manager argues: “The fiduciary model introduces lots of conflicts of interest. The managers are trusted advisers yet want the business themselves.”

BlackRock offers a fiduciary product but, as Hayes admits, 70% of the funds are managed internally with typically 30% being allocated to external managers, although this can vary from client to client.

As the UK pension manager continues: “People have spent the last 10 years unbundling but are now rebundling with the fiduciary model. It has not got a lot of traction in the UK, as trustees want to run the funds themselves. They want to control the assets.”

But asset allocation advice can come in a number of guises. For BlackRock it spans not only the fiduciary service that links to the liabilities, but also includes a pure global tactical asset allocation (GTAA) overlay, absolute return strategies and also, according to Hayes, a pure asset allocation service with a flat fee unconnected to other services, although BlackRock at present has just a single client in this. Hayes admits: “We don’t expect many to pay for asset allocation on a standalone basis. The reason is, how do you measure it? What constitutes success?” 

Philosophical approaches

“A skilful financial economist can make a sound case for investing pension assets in either equities, bonds or any asset class in between based on corporate finance theory,” according to Moshe Milevsky and Mike Orszag, co-editors of Journal of Pension Economics and Finance.

Understanding the philosophical approach behind asset allocation is a critical issue and ties in closely with measurement of performance and ultimately on what is a fair and effective means of remuneration. Pension fund trustees are left facing a plethora of conflicting advice and themes. Terms such as ‘liability driven investment’ can mean many different things and imply contradictory strategies. At one extreme, LDI can mean hedging a snapshot view of the pension fund liabilities with a portfolio of nominal and index-linked bonds and swaps. This is a very expensive proposition particularly when assumptions on the liability side change, leaving the so-called matching asset portfolio with unhedged risks.

At the other extreme, it could be argued that approximate matching using inflation-hedging assets are all that is required. But this can lead to large weightings in equities, which might not be very different from a portfolio devised by a person who had never heard of the term LDI.

The traditional model of asset allocation based on first undertaking an asset liability management study, then adopting a fixed asset allocation for the next three years based on it, breaks down because of the lack of market valuation and input and also as Van Dijk argues, because it is based on very large building blocks.

Many investment consultants make crude categorisations of different assets as either ‘return seeking’ or ‘liability matching’. This has become a prevalent idea in Holland, says Hayes, with all fixed income assets, for example, automatically lumped into liability matching. But as he points out, there is a continuum of characteristic rather than a ‘barbell’ and many assets do not fit into a model-driven approach. He says: “High yield, for example, is an asset class with years of feast and years of famine. The average risk and return looks OK but you actually want to either have a zero amount,  or else a very large amount.”

Tactical asset allocation models provide more asset classes and the issue is how to incorporate the medium-term TAA-type analysis into a longer-term asset allocation philosophy. Ideas such as mean reversion, espoused by firms such as GMO, certainly strike a chord. As the UK pension manager admits, trustees understand it. And mean reversion models give information signals to make decisions even if you may not get the timing right.

Van Dijk says that mean reversion should be part of any TAA system, on behavioural specialist grounds, and for general quantitative and fundamental reasons. The problem is that while there may be plenty of academic evidence to show that all asset classes revert to a long-run mean valuation, that process might be so slow that it lies beyond an individual’s or even institution’s time horizon.

The greatest challenge in asset allocation may be to actually have an open and informed debate as to what are the alternative philosophical approaches that pension funds can choose from, and what are the implications in terms of service providers, measurement of performance, selection of investment managers and the use of alternative as well as mainstream asset classes.