Fundamental indexing

Blogger-at-large, Yours Truly, wandered over to a debate organized by the London Quant Group (LQC) last Tuesday, on the topic of fundamental indexing. In one corner, we had Dr Adam Olive of HSBC, arguing for fundamental indexing, primarily as an improvement over pure market-cap (market capitalization) based indices. In the other corner, we had Ed Fishwick, of market heavyweight Blackrock and current chairman of the LQG, arguing against fundamental indexing as a closet mechanical value strategy. As both companies are deeply involved in the ETF (Exchange Traded Fund) industry, amongst others, as Blackrock is the current owner of the iShares brand, and HSBC has its own suite of index tracking ETFs, the discussion had a potentially practical set of implications for both sides. The wine, courtesy of LQG, fuelled each side of the debate very effectively, both during and after the formalities.

So what’s the big idea?

Fundamental indexing was invented as an alternative to market cap based indices. It argues that using market capitalization to determine proportions of investment in particular assets breaks down if statistically noisy prices distort values. Instead of using pure market cap, fundamental indices use key accounting measures such as sales, book value, earnings, dividends, or cash flow to calculate the relative weight of a particular index member. Arnott says in Fundamental Indexation:

While price inefficiency could lead to the observed alpha, as capitalization weighting assuredly overweights the overvalued stocks and underweights undervalued stocks, the superior performance may also be attributable to superior mean-variance portfolio construction or to hidden risk factors (in an APT or Fama-French framework), none of which violates the assumption of price efficiency.

If stock prices diverge significantly from their intrinsic value, for example like Cisco (CSCO) did during the dot-com boom, a market-cap based index containing Cisco would significantly overweight Cisco-relative to its intrinsic value. In contrast, a fundamental based index, or more practically, an ETF that tracks it, would mechanically look at Cisco’s intrinsic accounting values like book value or earnings, and compare it to the index universe, in order to work out how much of Cisco should be held at any given moment. These indices are then periodically rebalanced, typically once a year after all of the previous year’s accounting data are published.

Below you can see the anamoly being discussed:

There is no clear “rational” financial justification for the dot-com boom. Because Cisco was strategically buying dot-com startups at the time, which it thought complemented its business strategy, much of the jump in the valuation of Cisco itself reflects the relative market value of these closely held shares. Once dot-com mania ended, the value of these shares went back to a more rational value, closer to the actual intrinsic value, using something like the dividend discount model. From the point of vie of the chart, if you “cut out” the period from 1998 to 2001, it would have looked like a natural, gradual growth of value.

Because Cisco is a both a large and a well-run company, at least that’s what my friends who work there say, it is naturally a candidate for many different US indices. A fundamental based index, which would have tracked for example the growth in revenues and net profits during this period would have held much less of Cisco at the peak of its crazy-talk valuation, thus shielding investors from the downward slide back to reality. A market cap based index would have held a large amount of Cisco, because it’s a large company with a high market value, subjecting investors to downside that, in retrospect, looks rather needless.

But the debate isn’t really about individual stocks, it’s about indices, or collections of stocks and how they interact, i.e. portfolios. It’s this relative structure which is important. In fact, the idea came about possibly as a marketing ploy to help sell fundamentally indexed ETFs. The seminal Arnoff paper in the Financial Analysts Journal, where Arnoff was editor, inspired some of the members of the board to resign in protest, according to conspiracy theory touting detractors. Fundamental accounting numbers are important because they help establish the relative weight to the rest of the index constituents. Even though they may not be perfect, they are better than market cap. Statistically this comes out roughly to about a 2% positive difference (215 basis points in the Arnott  study), compared to market cap indicies over40 years of backtests. This comes out relatively clearly.

Fundamental indexing as a girly man mechanical value strategy

In the clip below, Arnold proves Hanz und Franz to be girlie men. The same can be said of fundamental indexing when compared to a well executed value investing strategy.

The problem is that, on its own, fundamental indexing is only a mechanical strategy, that attempts to get at fundamental values of companies, taking accounting data at face value. I’ve spent enough time digging through the financial statements of public companies, to know how little they really mean. Reporting standards, while standardized, require a level of interpretation that computers would not be able to automate, certainly any time soon. There are many factors which are completely not standardized, such as off-balance sheet holdings, special purpose daughter companies, derivatives reporting. Accounting standards can vary rather significantly from country to country.  Even within one industry, it’s somewhat of an oversimplification to compare a shoe factory with a biotech startup, even though they both use financial statements. Financial statements are meant to provide a common framework to help initiate the thought process, but closing it off at that level is rather short sighted. As much as I like automated trading, I know enough about fundamental analysis to look at a balance sheet and realize how little it really says.

Fundamental indexing tries to emulate value investing, as in what Warren Buffet or Graham and Dodd promote, but fails. The idea, in value investing, is to find a stock that pays out regular and increasing dividends, i.e. a massive cash flow machine (cash cow), and then buy it when it’s really cheap. For example, directly after a market crash, there are typically many more value plays available. This model treats stocks like a better bond, because the dividends extend into infinity, and they are consistently increasing. In order to do this successfully and systematically, a value investor has to have a decent understanding of the business, i.e. look beyond just the main numbers. Buffet for example typically researches whole industries, prodding each company one by one, in order to find the best possible companies. Most importantly, unlike an index, value investors do not invest in a large number of stocks in order to diversify. Buffet labels this “di-worseification”, where holding a large number of relatively unattractive investments attempting to minimize risk, actually ends up significantly diluting the potential upside. If you know you are buying a great stock really cheaply, why would you want to own half of what’s traded on the exchange? The “leftovers” aren’t as attractive of a deal, almost by definition.

Again, It’s Not Really an Index

At the same time, fundamental indexing could arguably be a little self-delusionary, in claiming that it is indexing anything at all. An index is meant to passively track the performance of a representative group of stocks. Sometimes it tracks a sample, sometimes the whole universe. At any moment when there is an attempt to tweak that index, it stops really being an index. It becomes an attempt at doing better than the index. This is pretty much getting into the definition of an active strategy, i.e. actively trying to do better than an index. If you are trying to beat the index, then there is not much of a point pretending you are still following an index.

My beef with fundamental indexing isn’t that it can’t work in theory; it’s that I don’t like the current implementation used in practice. There could be a number of improvements to fundamental indexing the way it’s sold currently as ETFs. For one, the provider could  make smart restrictions to the universe based on market conditions. Buying everything blindly just for the sake of diversification seems suboptimal as a strategy. If you can restrict the universe of stocks based on specific fundamental criteria, that should do better than just having everything. The reason for this is, in my opinion, too much interest in outperforming an index, or relative performance, rather than just delivering good returns. Ultimately investors want decent returns and to not lose money, not just to outperform an index.  This could possibly be done with options-based portfolio insurance, for example, in order to improve the risk profile.

What’s the final verdict?

If it isn’t as good as an actively managed value strategy, it may still have a place in an investor’s portfolio, if it is cheaper than what it almost replicates. If it systematically earns 2% more than a market cap index, but charges significantly less than 2%, and also charges less than a successful value-oriented mutual or hedge fund, it would be an attractive product. It would potentially be attractive on the retail front in that case, particularly to investors that want to be “hands off”, who don’t believe active management will add very much alpha, and who don’t have much exposure to “value” profit drivers in their portfolio already. Fundamental indexing could definitely be still improved as a concept, automating more filtering and analysis for the average consumer, rather than forcing him to want everything. Greater thought into how these relative fundamental values, as proxies for intrinsic value, interact could improve the product even further.

To summarize, fundamental indexing on its own doesn’t:

  1. automate detailed financial accounting interpretation, because it realistically can’t do as good of a job
  2. help identify the best investments, just buys everything and then tweaks weights

Speaking for myself, this is why I would expect fundamental indexing to underperform a decent value investing approach to the markets. It would be possible to automate value strategy in greater depth, but it would have  to look different than fundamental indexing, as it looks today.

One Response to “Fundamental indexing”
Check out what others are saying...
  1. […] you need to have a sense of what it’s intrinsic value is, similar to the argument in fundamental indexing. As a result, the EMH on its own doesn’t really say much. The capital asset pricing model […]

What do you think? Leave a comment.

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s