Indices and benchmarks: Clarifying the difference
EDHEC's Noël Amenc clears up some of the differences between reference indices and custom benchmarks.
The words 'index' and 'benchmark' are often used indiscriminately in practice even though they are, a priori, very different concepts. A reference index is a portfolio that should represent the performance of a given segment of the market, so the focus is on representativeness. A custom benchmark is a portfolio that should represent the fair reward expected in exchange for risk exposures that an investor is willing to accept, so the focus is on efficiency. For most investors, this distinction may be semantic, but it clearly leads to different approaches to passive investment.
For example, an index that is constructed differently to a cap-weighted index will always be considered a substitute for the latter, so it seems normal investors would expect this new reference index to have the same level of transparency, and perhaps the same level of popularity, as the previous one. In the end, what determines the success of a new reference index will be as much its financial characteristics as its 'popularity', not only with investors but also consultants.
Naturally, implementation of a new form of reference index is not risk-free. All rebalancing schemes, with the notable exception of cap-weighting or equal-weighting, assume a certain level of out-of-sample stability in the structures that led to the in-sample estimation of the parameters. Whether one tries to reduce the dimensions of a variance-covariance matrix with a factor model, or uses accounting attributes to define the size of a company and de facto its position in an index, or creates a link between the risk and return of a stock, all of these methodological choices are more or less relevant depending on the period chosen. That is why we have always considered that the evaluation of an alternative weighting scheme for an index can only be carried out over a long period.
Nonetheless, a track record, however long it might be, does not suffice. Model risk should also be documented, and it is always important to study both the optimality conditions of the indices proposed and the dependence on risk factors, and therefore on market conditions, over the long run. This serious approach to the performance of alternative forms of indices will probably lead investors to diversify the alternative forms of investment. As an example, it is interesting to observe that minimum volatility and efficient indices do not have the same outperformance in relation to cap-weighted indices in different market conditions.
Ultimately, we observe that while the marketing of these new approaches highlights the fact these indices are substitutes for traditional indices, the same promoters spend their time comparing performance with the cap-weighted equivalents. This comparison shows clearly that cap-weighted indices remain the final reference for investors because they represent the average of the market, the summing up of supply and demand. Even though everybody agrees they are inefficient, nobody is willing to take the risk of forgetting about them or moving too far away from them because the investment industry is a peer group industry. How will a CIO be able to justify the underperformance of his or her new non-cap-weighted index to the trustees, pensioners and consultants for whom the natural choice is that of the market?
A custom benchmark is a response to this difficulty. Its aim is not necessarily to replace an index as the final reference for equity allocation, but instead to allow for the use of a relative risk budget that any asset manager or investor could accept with regard to the average performance of the market represented by cap-weighted indices. This tracking error budget was traditionally used by active managers who were making active decisions (stock, sector or style picking) and at the same time attempting to optimise their portfolio under constraint with regard to the cap-weighted index.
To offer an alternative to these active managers, a customised benchmark provides optimal usage of the relative risk budget based on efficient diversification methods that respect the tracking error constraint in a very robust way. Investors will ultimately have the choice of using their tracking error budget either by seeking alpha - and logically creating competition between traditional active managers and hedge funds - or, using the beta set, they could implement competition between the cost of diversification under constraint provided by active managers and investment in systematic forms of diversification. Since the objective in using the latter relates to the implementation of a passive investment strategy, the cost of the new forms of systematic beta is quite low, and definitely represents significant progress in investors' search for the best value for money in their investments.
The goal of the custom benchmark is not to serve as an external reference for the investment, but to be a genuine representation of the investor's inter- or intra-class allocation choices. Ultimately, it is not so much the 'popularity' of a benchmark that will lead to its success, but its customisation and appropriateness to reflect the investor's strategic benchmark choices in terms of risk and reward. Since it is not being used as an external reference, a custom benchmark will be judged less on its transparency or its relative simplicity than on its capacity to enable investors to achieve their diversification objectives, notably with regard to an external reference represented by a cap-weighted index.
Noël Amenc is a professor of finance at EDHEC Business School and a director at EDHEC-Risk Institute