What are the arguments that are today inciting a large number of European institutional investors to implement a core-satellite investment process?

Noël Amenc: Let's be open about it! The success of the core-satellite approach is firstly due to the savings made on management fees. In fact, we can distinguish a core made up of passive management vehicles (index funds, ETFs, etc.) with low management fees, from one, or several, very active satellites that are made up of funds with a strong tracking error, or even funds with no constraint whatsoever on managing relative risk with regard to a benchmark (hedge funds, for example).

Investors and their consultants consider it absurd to pay active management fees for funds or mandates whose tracking error constraint, in relation to their benchmark, actually renders them 95% passive. It is therefore better to divide a portfolio into two different parts and, for example, ensure that a tracking error constraint of 5% is respected by investing 75% of the assets in the core using supports with very low management fees, being passive, and 25% in the satellite, i.e. in very active funds that will allow themselves a tracking error of 20%. This passive core - active satellite division generally allows management fees to be reduced by more than a third.

But is this approach not challenged by the criticism aimed at benchmarked management since the bear market of 2000, 2001 and 2002? Could we not envisage totally active management of the core?

Noël Amenc: Benchmarks have been unfairly taken to task as, most of the time, they are put on a par with indices. However, these are two different concepts: while an index aims to represent a market or a particular risk (style, sector, etc.), a benchmark, on the other hand, tries to be representative of a management strategy, i.e. of the risks taken by the investor, and nothing obliges the latter to use the ones available in the market and thus "stick" to indices.

Choosing not to move away from a market index is to deprive oneself of the principal source of improvement of the risk/return combination - strategic allocation. A benchmark is the product of this allocation and not an index-related inevitability.

Finally, not using a benchmark means renouncing any long-term risk strategy. It is hard to imagine investors with long-term liabilities venturing into a pure "bottom up" or opportunistic approach to their investment.

According to you, therefore, an index cannot be a good benchmark?

Noël Amenc: I would say that it is not their vocation. Indices were created to represent risk; this is their main merit, but also their main drawback. They provide investors with a consensual view of the risks that they wish to take and, in most cases, are indeed representative of that risk. However, the purpose in building them is definitely not to make them optimal in terms of diversification. It is therefore easy to demonstrate that an optimal combination of a similar series of sector or style indices always outperforms the global market index from which they were derived in terms of risk-adjusted return over the long term. In the same vein, it is better to invest within the Euro zone using domestic indices rather than a single global, representative index.

An index is a straightforward and efficient support for constructing a benchmark, not a substitute for allocation.

A benchmark as a combination of ETFs for example?

Noël Amenc: Why not? The investor, the consultant or the manager can then concentrate on the main challenge, which is managing risk and risk diversification, to construct long-term performance.

It is also one of the main advantages of the core-satellite approach. It enables the management of beta in the core to be distinguished from that of alpha in the satellite, thereby facilitating the use of specific expertise.

In order to implement a core representing his main assets, it is likely that the investor would want to turn to large institutions who would guarantee delivery of the "normal" returns associated with his allocation choices.

In the case of the satellite(s), the investor will perhaps be less interested in the reputation or the financial soundness of the supplier than in their ability to create alpha, rewarding the tracking error risk generated by the satellite in an optimal way. The objective assigned to providers of the satellite is no longer to secure "normal" returns but to create "abnormal" returns or alpha.

The satellite, as an alpha-providing asset group, is therefore intended to accommodate investments in hedge funds?

Noël Amenc: There's room for hedge funds everywhere!

When constructing optimal and efficient benchmarks, hedge funds, due to their diversification qualities not only in terms of correlation but also in terms of co-skewness (managing the skewness of the return distribution) or co-kurtosis (decreasing extreme risks), must be included in the core part of the portfolio.

These investments then need to be secured. This security requirement is what is making "investable" indices successful today. Managed from a "managed accounts" platform, they enable operational risk to be very strongly limited by replacing the risk of investing in "exotic" boutiques (i.e. hedge funds) with the reliability of the control procedures and financial soundness of the manager of the platform in question.

On the other hand, when using hedge funds as a substitute for traditional active management in a satellite, since they reward the active management risk better and therefore the tracking error with regard to the global benchmark, the funds selected must be those that generate the best alpha possible. The investor therefore accepts the risk, by looking for the best ones ex-ante, of finding the worst ones ex-post! This uncertainty is, as a matter of fact, what makes funds of hedge funds successful and interesting.

In the study conducted on the use of EuroMTS trackers, you introduce the concept of dynamic core-satellite management. Can you tell us more about this innovation?

Noël Amenc: Tracking error is a bit like cholesterol: there is good cholesterol, when you move away from the benchmark and outperform it, and bad cholesterol when you move away from the benchmark and the latter "ends up doing better" than the active manager.

When the investor allocates his investment between the core and the satellite, he does not generally know ex-ante whether he will obtain good or bad tracking error. The idea of dynamic core-satellite allocation is to integrate the distinction between good and bad tracking error in the apportionment of the core-satellite of the investments and therefore allow optimisation of the size of the satellite.

Indeed, we have adapted the well-known cushion method of the constant proportion portfolio insurance technique (CPPI) to the core-satellite approach in order to implement a positive tracking error "guarantee". Thus, when the satellite outperforms the benchmark, we increase investment in the first to the detriment of the second, whereas when the satellite underperforms the benchmark, we reduce its share in the portfolio in order to avoid falling below the "guaranteed" tracking error value.

We have tested this approach in the context of both stock and bond investment management; for unchanged performance of the investment supports included in the core and the satellite, it allows the portfolio's performance to be improved almost "mechanically" and to a significant degree, since it can exceed, by means of diversified management, 200 annual basis points.