Fundamental Indexing

[kyounge1956]

I'm reading a most interesting book called The Fundamental Index, by Robert Arnott, Jason Hsu and John West. I'm not even halfway through, but so far, the central idea of the book appears to be that weighting the stocks in an index by market capitalization puts too much in stocks which are overpriced in terms of fundamental measurements like sales, profits or dividend yield, and too little in those which are underpriced. When the prices of the stocks move to more closely reflect the fundamentals of each business, more is lost on the overpriced stocks than is gained on the underpriced ones, creating a performance drag in the index which can be overcome by using some other characteristic to weight the stocks.

Comments?

 

[MasterBlaster]

The market cap versus other measure index weighting metric argument has gone on for along time. The book you are reading , evidently, is a non-proponent of market cap weighting.

John Bogle (of Vanguard fame) and others have suggested that market cap weightings are really the best way to index. Bogle writes very eloquently about this topic.

The analysis I have seen, and I haven't looked at much, suggest that an equal-weight index would fare better in down markets. The market-cap index would fare better in rising markets.

 

[mmcc]

Exchange Traded Funds (ETF) Investments with WisdomTree provides other indexing approaches. Jeremy Siegel, who is associated with WisdomTree, also covered the cap-weighting issue in one of his books (The Future for Investors).

 

[kyounge1956]

The market cap versus other measure index weighting metric argument has gone on for along time. The book you are reading , evidently, is a non-proponent of market cap weighting.

Yes, or a proponent of non-market-cap weighting...however you like to put it.

John Bogle (of Vanguard fame) and others have suggested that market cap weightings are really the best way to index. Bogle writes very eloquently about this topic.

Which of his books specifically would I read to find out the reasoning behind this position?

The analysis I have seen, and I haven't looked at much, suggest that an equal-weight index would fare better in down markets. The market-cap index would fare better in rising markets.

I don't have the book with me here, but I think it also said something along those lines. However, the benefit of the non-cap-weighted index in a down market is greater than that of the cap weighted index in an up market, so on balance the non-cap-weighted outperformed the cap-weighted.

Actually, if I'm understanding the thesis, an index weighted by anything other than market cap would eliminate the linkage between price and weighting, and so (if the hypothesis is correct) should outperform a cap-weighted index made up of the same stocks. However, I think the authors are only going to present evidence in favor of an index weighted by certain company fundamental measures. They favor that sort of an index over other possibilities, such as equal weighting, because it has other points in its favor in addition to eliminating the price/weight link.

 

[MasterBlaster]

So on balance the non-cap-weighted outperformed the cap-weighted

Whenever you hear stuff like this, always ask under what timeframe ? Under what basket of stocks. What was the investment climate over that period. Do the equal basket weighting folks continually rebalance - or what. And so on. It makes a difference.

Which of his books specifically would I read to find out the reasoning behind this position?

You'll have to do a search on this yourself as I don't exactly recall. perhaps some of the other posters here can chime in.

As I understand things the issue isn't completely settled.

 

[kyounge1956]

Whenever you hear stuff like this, always ask under what timeframe ? Under what basket of stocks. What was the investment climate over that period. Do the equal basket weighting folks continually rebalance - or what. And so on. It makes a difference.

You'll have to do a search on this yourself as I don't exactly recall. perhaps some of the other posters here can chime in.

As I understand things the issue isn't completely settled.

The authors do not favor equal-weighting, at least partly because such an index would need more frequent rebalancing than other possibilities, with associated extra costs.

I'll have to wait until I get home to check the time period over which the fundamental-weighted index outperformed the cap-weighted one, but IIRC it was a long time span, which would have included both up and down market climates.

 

[Maurice]

I'm reading a most interesting book called The Fundamental Index, by Robert Arnott, Jason Hsu and John West. I'm not even halfway through, but so far, the central idea of the book appears to be that weighting the stocks in an index by market capitalization puts too much in stocks which are overpriced in terms of fundamental measurements like sales, profits or dividend yield, and too little in those which are underpriced. When the prices of the stocks move to more closely reflect the fundamentals of each business, more is lost on the overpriced stocks than is gained on the underpriced ones, creating a performance drag in the index which can be overcome by using some other characteristic to weight the stocks.

Comments?

Lots of people address this by tilting to small cap and tilting to value.

As indices though, the market cap approach is the way to go. Otherwise you end up with meaningless garbage like the DJIA.

 

[yakers]

What would a fundamental 'index' track? I would worry about fees as well, fees are the one element of investing that I can control.

 

[MasterBlaster]

You track the basket

Fundamental indexing refers to holding a basket of stocks. There are a variety of ways to weight the basket.

usually they mean you hold either:

a) An equal dollar amount of each stock in the basket
b) an equal number (or constant - ie unchanging number) of shares in each stock in the basket.

Fundamental indexing is distinguished from market capitalization indexing (or just indexing by common reference). market capitalization indexing is where the basket is weighted by the total market cap of each stock in the index. Note that this type index is slanted towards the largest companies in the index.

 

[kyounge1956]

Fundamental indexing refers to holding a basket of stocks. There are a variety of ways to weight the basket.

usually they mean you hold either:

a) An equal dollar amount of each stock in the basket
b) an equal number (or constant - ie unchanging number) of shares in each stock in the basket.

What they are advocating in this book isn't either of the types of equal weighting you describe above. It is weighting the basket of stocks according to a five-year average of four characteristics of the business. I don't have them here, but they are things like sales, dividend yield etc.

Fundamental indexing is distinguished from market capitalization indexing (or just indexing by common reference). market capitalization indexing is where the basket is weighted by the total market cap of each stock in the index. Note that this type index is slanted towards the largest companies in the index.

Correct, and that is not necessarily a good thing. For example, weighting by market cap would have put more tech stocks with sky-high P/E ratios into the basket before the bursting of the tech bubble, precisely because their prices had gone up.

 

[MasterBlaster]

What they are advocating in this book isn't either of the types of equal weighting you describe above. It is weighting the basket of stocks according to a five-year average of four characteristics of the business. I don't have them here, but they are things like sales, dividend yield etc.

<caveot> I haven't looked into this issue much. However absent other information I do have an opinion. I would be in a skeptical show-me mode.

Sometimes, people fool themselves. They think that finding some past weighting of stocks that produced superior returns means that the same weighting will produce superior returns in the future. In reality all they did was find some weird combination of weights that made past results look good.

Are the weights based on anything other than data mining ? is there any theoretical basis for what they did ?

as a topic aside, In my opinion Bob Clyatt's larger than 4 % SWR that he advocates in his book "Work Less, Live More" falls in the data mining (ie- unrealistic) category

 

[kyounge1956]

<caveot> I haven't looked into this issue much. However absent other information I do have an opinion. I would be in a skeptical show-me mode.

Sometimes, people fool themselves. They think that finding some past weighting of stocks that produced superior returns means that the same weighting will produce superior returns in the future. In reality all they did was find some weird combination of weights that made past results look good.

Are the weights based on anything other than data mining ? is there any theoretical basis for what they did ?(snip)

I have the book in front of me now. The data go back to 1962 (as far back as five-year data was available, according to the book). Weighting by book value, cash flow, gross dividends or sales, or a composite of all four, outperformed cap-weighting over that period, and in most rolling five-year periods. As you mentioned above, the cap weighted index outperformed the other during strong bull markets—specifically during 58% of the periods when the market went up between 20 and 30% per year. In all markets that were rising more slowly or falling, the non-cap weighed index outperformed the cap-weighted in at least 80% of the rolling 5-year periods.

I haven't gotten to the chapter on the theoretical basis yet. If I understand the hypothesis, any weighting method other than market capitalization, even a nonsensical one like weighting the stocks by alphabetical order, should eliminate the performance drag associated with cap-weighting, though it wouldn't necessarily be a practical rule for creating a mutual fund. However, if sales, dividends, book value and cash flow are valid measures of a company's size (the book calls it an "economic footprint") then I don't think the results can be written off as due to data mining.

 

[kyounge1956]

What would a fundamental 'index' track? I would worry about fees as well, fees are the one element of investing that I can control.

It could track any market or segment of the market. It is just a different means of determining how much of each stock is included in the index. The authors say their large-company index has between 70 and 80% overlap with the S&P. You could use their method to put every stock in existence in order by "economic footprint" and then use the top X% of them or the bottom x# of stocks to make a large-company or small-company index. If bonds have "fundamentals" maybe even a bond index could be created along these lines.

I think there are actual index funds which are set up according to the ideas in this book but I haven't looked them up yet to see what their fees are.

 

[rayvt]

A cap-weighted index is automatically self-rebalancing. By definition, every stock in it is held in the right proportion.
Any other weighting is *not* self-rebalancing and must be periodically rebalanced by buying & selling stocks.
(Of course, when stocks move into or out of an index, they must be bought or sold, so there's always some trading no matter what weighting is used.)


"What they are advocating in this book isn't either of the types of equal weighting you describe above. It is weighting the basket of stocks according to a five-year average of four characteristics of the business."

IOW, this is not an index fund. It is just an actively managed fund that has some mechanical rules for what and when to trade.

 

[kyounge1956]

"What they are advocating in this book isn't either of the types of equal weighting you describe above. It is weighting the basket of stocks according to a five-year average of four characteristics of the business."

IOW, this is not an index fund. It is just an actively managed fund that has some mechanical rules for what and when to trade.

How so? Definition of an index fund (from Wikipedia): "a collective investment scheme (usually a mutual fund or exchange-traded fund) that aims to replicate the movements of an index of a specific financial market, or a set of rules of ownership that are held constant, regardless of market conditions." A fund which weights the stocks by something other than market capitalization is following "set rules of ownership that are held constant, regardless of market conditions"—how is it not an index fund? Like an equally-weighted fund, it does involve more trading than a cap-weighted fund (although nowhere near as much turnover as a typical actively managed fund), and that additional trading increases the fund's expenses. The question then becomes, does the extra return exceed the additional cost?

 

[rayvt]

I would argue that it's not an index fund because it doesn't track any of the generally accepted indexes. Rather, it follows a set of rules that they made up. On other investment sites I read, this is called "mechanical investing".

Not to say that it's good or bad, just that they are using the term "index fund" in a meaning that is different from what most people think is the meaning.

 

[Gone4Good]

The question then becomes, does the extra return exceed the additional cost?

The more fundamental question is whether the approach generates any excess return at all. I'm skeptical that some static algorithm which tracks a couple of variables is going to be able to generate market beating returns. It's pretty easy to develop a screen that will pick winning stocks in retrospect. Achieving the same results in real time is a completely different matter though.

Besides, aren't these measures of value the same ones that many portfolio managers use to consistently generate sub-par performance? Can we really expect better results if we automate and simplify the stock selection process and call it an index? I'm guessing "no".

 

[kyounge1956]

"What they are advocating in this book isn't either of the types of equal weighting you describe above. It is weighting the basket of stocks according to a five-year average of four characteristics of the business."

IOW, this is not an index fund. It is just an actively managed fund that has some mechanical rules for what and when to trade.

How so? Definition of an index fund (from Wikipedia): "a collective investment scheme (usually a mutual fund or exchange-traded fund) that aims to replicate the movements of an index of a specific financial market, or a set of rules of ownership that are held constant, regardless of market conditions." (snip)

I would argue that it's not an index fund because it doesn't track any of the generally accepted indexes. Rather, it follows a set of rules that they made up. On other investment sites I read, this is called "mechanical investing".

Not to say that it's good or bad, just that they are using the term "index fund" in a meaning that is different from what most people think is the meaning.

If your definition of index funds includes only those that track a pre-existing index, then I guess this isn't one, but neither is it "just an actively managed fund". I think "mechanical investing" is a reasonable way to describe it.

 

[kyounge1956]

The more fundamental question is whether the approach generates any excess return at all. I'm skeptical that some static algorithm which tracks a couple of variables is going to be able to generate market beating returns. It's pretty easy to develop a screen that will pick winning stocks in retrospect. Achieving the same results in real time is a completely different matter though.

It depends on whether the Efficient Market Hypothesis is correct or not. If 100% of stocks are correctly priced at all times, then there wouldn't be any excess returns. But if there are such things as overpriced stocks, cap-weighting will automatically put a larger and larger proportion of the index into the very stocks that will eventually collapse when the bubble bursts. If there is a way to formulate an index which avoids this problem, it should outperform an index which has that vulnerability, unless the cost of avoiding the problem is greater than the problem. What the fundamental index is claiming is not the ability to tell in advance which stocks are over-or underpriced, but to break the link between the amount of a stock in the index and its price, which results in the over-and under-pricing canceling each other out. The other way this wouldn't work is if the amounts of underpricing and overpricing in the market are not randomly distributed with respect to the "economic footprint" of the stock. That might work either to the advantage or disadvantage of non-cap weighting.

Besides, aren't these measures of value the same ones that many portfolio managers use to consistently generate sub-par performance?

No, I think these measures are more similar to those used to determine which stocks to include in a "value" index. I thought the sub-par results of active traders come not from the measures the fund managers use, but from lack of diversification via stock picking/market timing (i.e. they don't buy the whole market and don't hold the stocks for long period) and the associated trading costs.
A Fundamental index fund would have more trading cost than a cap-weighted fund, but much closer to the cap-weighted one than to a typical actively traded fund.

Can we really expect better results if we automate and simplify the stock selection process and call it an index? I'm guessing "no".

If we are automating and simplifying the elimination of a drag on performance which is inherent in the way we had been formulating indexes in the past, we can expect better results, regardless of what we call the resulting portfolio. If there is not really a drag on performance inherent in cap-weighting, or if the automated process is vulnerable to the same flaw, automating and simplifying won't do any good.

I still haven't seen a ticker symbol in the book for the sort of index they are talking about, so no way yet to look up and see if the theory works out in practice.

 

[MasterBlaster]

Kyounge1956:

I suggest you post these same questions (regarding index weighting) over at the boglehead forum.

Bogleheads :: Index

They have some very smart people over there that can (perhaps) discuss this with more rigor.

Let us know what the consensus is for the best way to index a portfolio.

 

[Gone4Good]

If there is a way to formulate an index which avoids this problem, it should outperform an index which has that vulnerability

But that is exactly the problem. Nobody has demonstrated an ability to do this.

In order for the fundamental index to outperform the traditional index the stocks that it overweights need to outperform the stocks that it underweights. That is exactly what every active manager on the planet is trying to do. They haven't been successful.

 

[MasterBlaster]

William J. Bernstein is one of those indexing guru's notably of the book "The Four Pillers of Investing" etc.

Here is (one of) Bernstein's take on fundamental indexing as described in this linked article:

Fundamental Indexing

<from the article>
In summary, we draw two conclusions:

• Fundamental indexing is a promising technique, but its advantage over more conventional cap-weighted value-oriented schemes, to the extent that it exists at all, is relatively small. Attempts should be made to confirm this work within a multifactor framework both abroad and with pre-Compustat U.S. data.

• Even assuming that fundamental indexation produces returns in excess of its factor exposure, caution should be used in the practical application of this methodology. Differences in the expenses, fees, and transactional costs incurred in the design and execution of real-world portfolios can easily overwhelm the relatively small marginal benefits of any one value-oriented approach. The prospective shareholder needs to consider not only the selection paradigm used, but just who is executing it.

 

[kyounge1956]

But that is exactly the problem. Nobody has demonstrated an ability to do this.

In order for the fundamental index to outperform the traditional index the stocks that it overweights need to outperform the stocks that it underweights. That is exactly what every active manager on the planet is trying to do. They haven't been successful.

I don't think it's exactly the same. Active managers, in order to succeed, must determine in advance, which stocks will over-perform and which will under-perform, or when the market will change from bull to bear. Fundamental weighting doesn't need to know which stocks are which or predict market moves, it just removes cap-weighting's guarantee that a fund's exposure to an overpriced stock will increase and its exposure to an underpriced stock will decrease. If I understand it correctly, the idea is that the pricing errors are randomly scattered through the total stock market, and thus the use of a weighting scheme unrelated to price means that the fund will contain equal amounts of overpriced stocks with increasing weight, overpriced stocks with decreasing weight, underpriced stocks with increasing weight and underpriced stocks with decreasing weight, and the errors will cancel each other out. If the pricing errors are not randomly distributed, at some point the fundamental index will stop outperforming the cap-weighted one.

 

[kyounge1956]

William J. Bernstein is one of those indexing guru's notably of the book "The Four Pillers of Investing" etc.

Here is (one of) Bernstein's take on fundamental indexing as described in this linked article:

Fundamental Indexing

<from the article>
In summary, we draw two conclusions:

• Fundamental indexing is a promising technique, but its advantage over more conventional cap-weighted value-oriented schemes, to the extent that it exists at all, is relatively small. Attempts should be made to confirm this work within a multifactor framework both abroad and with pre-Compustat U.S. data.

• Even assuming that fundamental indexation produces returns in excess of its factor exposure, caution should be used in the practical application of this methodology. Differences in the expenses, fees, and transactional costs incurred in the design and execution of real-world portfolios can easily overwhelm the relatively small marginal benefits of any one value-oriented approach. The prospective shareholder needs to consider not only the selection paradigm used, but just who is executing it.

more reading required. I don't know what the Three Factors are, or alpha ratios measure, or the meaning of R-squared. I don't understand what the article is saying. >sigh<

 

[ESRBob]

Good thread,
By "factor exposure" Bernstein means that certain factors (e.g. smallness or book-value ratio, a way of classifying something on a value-growth continuum) have been shown through repeated historical research to generate excess returns which most researchers agree accrue to the additional risks such companies face. (Small stocks are perceived to be riskier possibly because they can get cut off from credit during a credit crunch or value stocks are perceived to be risker because they are (often) companies whose stock prices are beat up because something has gone wrong with the business)

So what Bernstein is saying (and I agree) is that any outperformance by these fundamentally-ranked portfolios may only be the expected return for the additional risk they are taking on by having more small or value in their mix than the S&P

This may be just what you want, but it isn't a magic bullet way to earn a free lunch, extra return for no additional risk.

Also, as someone said in this thread, people have been putting similar portfolios together for years without Arnott's indexes by using small and value-tilts.

One trivia footnote --I worked for Arnott years ago (1982,83) when he was a portfolio manager at the Boston Company and I was a programmer pulling together the screens through the historical CRSP and Compustat data to look for things that would produce excess returns... I can assure you people have been looking for ways to mine that data to find a magic formula for a really long time. If anything gets found, the advantage tends to get traded away in a hurry. Made me into a lifelong slicer/dicer indexer...