Indices - August 29, 2008

Beauty is in the eye of the beholder - interview with Rob Arnott

In this month's interview, Rob Arnott, Chairman and Founder of Research Affiliates, discusses the merits of fundamental indexing and addresses some of the questions raised by the EDHEC Risk and Asset Management Research Centre's recent position paper on the subject.

Rob Arnott

You have recently accepted an invitation to join the EDHEC Risk and Asset Management Research Centre’s International Advisory Board. Could you tell us the reasons behind your decision to join the board?

Rob Arnott: EDHEC is doing some of the best new work in the European Academic community on modern finance questions. At many institutions, there’s a tendency to be “married” to the core assumptions of modern finance. The Efficient Market Hypothesis, CAPM, Fama-French, and so forth, are seen as fact, not theory. In some cases, this is even to an extent that challenges the legitimacy of data that contradicts the theory. In my view, theory is not fact … at best it’s an idealized approximation of the real world, a guide to how finance and markets should work in a perfect world. I like the fact that EDHEC is willing to challenge some of the core precepts of modern finance, which is (in my view) the only way to improve theory and to make theory ever more relevant to the practitioner community.

Researchers from EDHEC presented a paper on fundamental indices at the EFM Symposium in Nice in April, an event at which you were the keynote speaker. This paper, even though it recognises the usefulness of fundamental indices and their superiority over cap-weighted indices, wonders whether these indices might not simply be the equivalent of a Value strategy. Moreover, in EDHEC’s research, fundamental indices do not appear to outperform equal-weighted indices. What are your views on these points?

Rob Arnott: All along this Fundamental Index journey of the last four years, I have been honoured by the calibre of individuals and institutions that have reviewed our work. The attention and conclusions – whether they agree with our interpretations of our results or not – are a testament to the theoretical and practical applications of the idea.

I am very accustomed to the value critique. In fact, my colleagues and I noted the Fundamental Index produces only a skinny alpha net of the Fama French factors in our original research. But I think that misses the point. What’s wrong with a liquid, low turnover, economically representative, and massively scalable portfolio that produces significant return above cap weighted benchmarks? Few investors would be happy with positive Fama French alpha but negative excess returns versus popular benchmarks like the S & P 500. To my knowledge, the Fundamental Indexes are the only commercially available index to offer a value “tilt” relative to the capitalization-weighted markets, while avoiding large negative Fama-French alphas. Even the EDHEC paper found that our methodology was one of the few to deliver a statistically significant Fama-French-Carhart alpha!

The other interesting sidebar is the manner in which our industry assumes that the cap-weighted index is the “style neutral” portfolio. Relative to itself, of course it is; but, relative to the real-world economy, it most assuredly is not. A growth stock selling at four times the market multiple – for growth that may or may not materialize – gets four times the weight? Cap weighting structurally favours growth stocks relative to their current economic scale. And unless “trees grow to the sky”, cap weighted investors shouldn’t expect to experience any performance advantage for this growth bias! Fundamental indexing doesn’t structurally allocate more to growth, relative to a company’s economic footprint. It eliminates future expectations and simply weights each company by how large it is in the economy. This critique is a matter of our frame of reference: viewed with a cap weighted lens, a Fundamental Index portfolio will naturally have a value bias.

As for equal weighting, theoretically it should produce the same return as the Fundamental Index (as would a random weight or “dartboard” portfolio.) In fact, I would assert equal weighing should produce a higher return than the Fundamental Index due to its massive small cap bias – remember you are putting the same weight to the smallest company as the largest. However, this doesn’t happen. The Fundamental Index has beaten the S & P 500 Equal Weight Index by about 80 basis points annually since the 1989 inception of SPEWI.

Regardless, both should result in portfolio weights that are uncorrelated with pricing error and, in so doing, avoid the return drag from cap-weighting. That said, the equal weighted portfolio fixes the one flaw in cap-weighting but pops a few leaks of its own namely higher turnover, scant capacity, poor economic representation and higher transaction costs. The beauty of the Fundamental Index is it preserves the many benefits of passive investing but eliminates the one glaring negative: assuredly overweighting the overvalued and underweighting the undervalued.

At the end of the day, are fundamental indices not simply a way of accessing active systematic stock picking with Value criteria through low-cost vehicles such as ETFs?

Rob Arnott: On this one, beauty is in the eye of the beholder. “Active systematic stock picking” to me involves making a relative value determination on a stock or group of stocks. Is this security undervalued or overvalued? Is it a “buy”, a “sell” or a “hold”? The Fundamental Index makes no such judgement – it’s just asking, “how big is this company?” We don’t assert that a company with a higher relative weight in the Fundamental Index is necessarily “undervalued,” in much the same way that an equal weighted indexer isn’t claiming that portfolio’s weights are “right”. If that was the case, every company over a 0.20% weight in the S & P 500 would be “overvalued”! Neither the Fundamental Index nor equal weighing are stock picking in the traditional sense. Their outperformance comes from decoupling errors in portfolio weights (relative to the unknowable “fair value weight”) from errors in the price (relative to that self-same “fair value”).

Now there are those that claim any deviation from cap-weighting is a stock picking strategy. Fine. That’s a semantics issue, not an investments issue. The real question is whether it benefits investors. We know cap weighting can often lead to inefficient exposures that drag down returns and exposes investors to a heavy growth bias. Up until the Fundamental Index, investors wishing to counter these effects had to go down the active management path, or maybe hire value indexers with their stark negative Fama-French alphas. Now the benefits of indexing – low turnover, costs well below active managers, massive diversification, economic representation and significant capacity – can be had without the inefficient by-products of cap weighting. That is a very exciting development.

About Rob Arnott

Robert Arnott is Chairman and Founder of Research Affiliates. He has served as editor of the Financial Analysts Journal, and has published more than 100 articles in the FAJ, the Journal of Portfolio Management, the Harvard Business Review, and other respected journals. He has been a frequent contributor to leading financial journals and books and he has received five Graham and Dodd Scrolls and Awards, awarded annually by CFA Institute for best article of the year. He also received two Bernstein-Fabozzi/Jacobs-Levy awards from the Journal of Portfolio Management and Institutional Investor magazine.

Rob is the former chairman of First Quadrant, LP, where he developed quantitative asset management products. He also served as global equity strategist at Salomon Brothers (now part of Citigroup), the president of TSA Capital Management (now part of Analytic), and as a vice president at The Boston Company (now PanAgora).

Rob Arnott graduated summa cum laude from the University of California, Santa Barbara, in economics, applied mathematics, and computer science in 1977.

URL for this document:

Hyperlinks in this document: