[This article initially appeared on IndexUniverse.eu.]

It’s approaching five years since Robert Arnott, Jason Hsu and Philip Moore published their famous paper on fundamental indexation. An intense debate over the most appropriate method of equity index construction followed, much of which can be followed in the articles of IndexUniverse.com and various publications in the Journal of Indexes.

Over the last 18 months, as equity investors were in survival mode amid free-falling share prices, the index debate moved to the background. But it hasn’t gone away, and a number of recent academic and practitioner initiatives show that it may be coming to the fore again. Why, and to what extent are European ETF investors venturing into “noncap-weighted” index territory?

To recap, capitalisation weighting allocates stocks a weighting in an index according to the constituent companies’ market value (calculated as the share price multiplied by the number of shares in issue). Critics of this methodology argue that it automatically overweights overvalued companies and underweights undervalued ones—a stance that appears empirically hard to disprove, given the prevalence and reoccurrence of market bubbles and busts, but one that also requires the rejection of the central tenets of modern financial theory, notably the efficient markets hypothesis.

Competing index weighting methodologies are varied, and include equal weighting, dividend weighting and weighting according to some quantitative model. Such models include the “RAFI” methodology, pioneered by Arnott’s Research Affiliates, which weights companies according to four fundamental valuation metrics; and also quite different, quantitatively driven portfolio selection techniques, such as Lyxor ETF’s Wise Quantitative Strategy, which follows a model developed by Societe Generale’s index team in Paris.

It’s important to note that many capitalisation-weighted indices are modified in some way, whether by adjustment for companies’ free-float (the proportion of the issued share capital that is available for trading) or by caps on individual stock or sector weightings. Therefore many such benchmarks already differ substantially from what might be called pure capitalisation weighting.

European Overview

On the face of it, capitalisation weighting appears dominant among European investors, both those investing passively, and those managing active funds to outperform a benchmark. The vast majority of them have up to now preferred large, familiar index names, often European country indices (all capitalisation-weighted) and pan-European or international country or regional indices (similarly compiled, with the exception of a few anomalously structured indices like the Dow Jones Industrial Index and the Nikkei 225 in Japan, both of which are arithmetic averages).

The recently released EDHEC 2009 ETF investor survey revealed that 72% of investors use broad market equity ETFs (which are predominantly capitalisation-weighted) as core holdings, with much smaller percentages using style ETFs (28%) – a category that includes ETFs with alternative weighting methodologies, but also a variety of themed ETFs, ethical, environmental, and so on), and sector ETFs (12%) for the same purpose.

Furthermore, assets allocated to the main noncap-weighted ETF categories have shrunk over the last year. Deutsche Bank’s ETF weekly liquidity trends report for 3 June 2009 showed that total assets in European RAFI ETFs had declined to €129 million from €251 million a year earlier, assets in dividend-weighted ETFs had fallen from €2.3 billion to just over €2 billion over the same period, and the €33 million in assets in the SPA MarketGrader ETF range (which followed a methodology that was both equally weighted and quantitatively-driven) in June 2008 had disappeared, following the compulsory redemption of these funds in March this year. Although assets in ETFs described as “style-other” in Deutsche Bank’s report rose from €385 million to €663 million over the same period (June 2008 to June 2009), this reflects several new product launches, with the number of funds in this catch-all category rising from 21 to 41.

Commenting on noncap-weighted indices, and specifically on the fundamental indexation model as proposed by Research Affiliates, Nizam Hamid, head of sales strategy at iShares in London, told IndexUniverse.eu that “there is a deep theoretical flaw in assuming that market-cap-weighted indices are defunct. Effectively, asset pricing models take into account asset prices, and market-cap-weighted indices therefore represent the opportunity set of potential portfolios that can be invested in. At a broad institutional level, investors cannot invest on the basis of fundamental weights—if they did, there would be significant distortions.”

Paul Lohrey, chief investment office of Vanguard in Europe, made a similar observation in an article published recently in the Euromoney ETFs & Indices Handbook 2009. Lohrey states that “the primary objective of most indexes is to accurately measure the performance of a given market or market segment.” He then goes on to write that “non-traditional indexes, such as equal-weighted indexes, fundamental-weighted indexes and other predominantly non-market capitalisation-weighted indexes do not share this objective.”

So two of the biggest names in the indexing industry set their stall out against alternative methodologies to capitalisation weighting, something that is reflected by the predominance of funds tracking more traditional indices in both firms’ product ranges (although both offer dividend-weighted ETFs).

However, Chris Sutton, senior investment consultant at Watson Wyatt in the UK and former head of iShares in Europe, argues in the same Euromoney publication that indices based on alternative weighting schemes may produce superior risk-adjusted returns, even after allowing for the potentially higher turnover of such approaches. Calling noncapitalisation-weighted methodologies the “beta prime” family, Watson Wyatt subdivides them into four categories: equally weighted, factor tilted, fundamentally weighted and risk-weighted. Sutton argues in his study that both the fundamentally weighted and risk-weighted methodologies can provide significant improvements in investment efficiency, whereas capitalisation weighting can produce a performance drag.

In his presentation at the recent EDHEC conference in Paris, Lionel Martellini, professor at the French business school, criticised the efficiency of capitalisation weighting, arguing that such indices can be far less diversified than the headline number of stocks suggests, due to concentrations in larger companies. Martellini, however, argued that, when adjusted for what he called “value bias”, the RAFI fundamental indices do not outperform their cap-weighted counterparts, and he suggested that equity index compilation according to risk weighting would be more efficient.

Tobam, a Paris- and London-based asset management firm whose leading staff formerly worked as the quantitative asset management team at Lehman Brothers, already manages around US$ 1 billion in assets under such a risk-weighted methodology, and is investigating the possibility of launching ETFs, one of the firm’s principals, Michael Gran, told IndexUniverse.eu. Tobam’s methodology, which is called “anti-benchmark”, selects stocks so as to maximise the diversification benefits to the portfolio. It does this not by using actual historical correlations between stocks, which Tobam’s team accept are unstable, but by using the hierarchy of correlations, which the firm’s researchers say is more likely to remain invariant over time.

Meanwhile the providers of ETFs compiled according to the more established noncapitalisation-weighted methodologies (such as RAFI) remain confident that their funds will gain traction in Europe. Tim Mitchell, head of specialist fund sales at Invesco PowerShares (one of the three European issuers of FTSE RAFI ETFs, together with Lyxor and Sweden’s Xact) told IndexUniverse.eu that the poor performance of fundamentally indexed funds in 2008, which resulted from an overweight position in financials, had largely been corrected this year, so that on a three-year view, the RAFI funds have outperformed their capitalisation-weighted competitors in most markets.

PowerShares has recently cut the annual total expense ratios on its FTSE RAFI ETFs from 0.75% to between 0.39% and 0.5%, something that the firm expects to result in further inflows. Mitchell, however, conceded that the generally poor performance of funds following quantitative equity models during the bear market had put some investors off, particularly when the equity selection methodology was opaque and difficult to understand.

Jacob Sternius, acting CEO at Stockholm-based Xact ETF, which runs two FTSE RAFI ETFs, explained to IndexUniverse.eu that some of the Scandinavian institutions that are the main investors in these funds withdrew their funds at the beginning of last year, causing a drop in assets under management. Sternius said he saw the FTSE RAFI ETFs as a suitable long-term investment for those with a contrarian bias, since they have a value tilt, but added that his firm saw benefits in both this and the more traditional method of index weighting.

All in all it appears that, while the poor performance of many quantitatively driven funds during the equity bear market has dealt a blow to those opposing capitalisation weighting, the debate over the most appropriate way to invest by index has not gone away and, if anything, it is attracting more and more participants. While the arguments over equity index construction are sometimes abstruse, this area of the market remains one of the most interesting for investors to follow.