Fundamental Indexing

[peter]

You mean VWO?

[tidal]

Morgan Stanley fundamentally analyzes fundamental indexing

In Conclusion
In short, the study worked out like this: during periods where value outperformed, the fundamentally weighted indexes beat market-cap benchmarks; during periods where growth outperformed, they trailed. The fundamental indexes may not be just value indexes, but they are certainly correlated with value-based outperformance.

[tidal]

Disputing the superiority of fundamental indexing

BOSTON — Fundamental indexing — touted as a “better mousetrap” alternative to capital-weighted market indexes — won’t necessarily catch more mice, according to Harvard professor Andre Perold.

In a draft paper, “Fundamentally Flawed Indexing,” Mr. Perold, the George Gund Professor of Finance and Banking at Harvard Business School in Boston, argues that a pillar of the fundamental-indexing sales pitch — that cap-weighted indexes deliver “inferior” returns by overweighting overvalued companies and underweighting undervalued companies — doesn’t stand up to scrutiny.

Marshaling the same models and two-stock portfolios used by proponents to make their case, his paper concludes that the idea “that capitalization weighting imposes an intrinsic drag on performance is false.”

Haven't seen the paper itself yet...

[Bylo Selhi]

Interview with Gus Sauter: Indexing fundamentals and fundamental indexing

[parvus]

bylo cited:

And I guess I’m scratching my head when I think, why would you use the number of employees to figure out how you should weight a stock?

That's not entirely fair; or to put it back to Sauter, why would you use a speculative take on a no-earner company with momentum gamblers behind it to construct your index?

The sheer effect of weighting by dividend yield would certainly give you a value bias. Using some of these other factors also produces a bit of a value and a small-cap bias.

And there's a problem in investing in companies that actually earn money and show it by paying the owners? Or should we all have been subprime borrowers and lenders? Are we talking about stocks that reflect real economic growth, or just the ones that the punters favour?

My greatest concern is I don’t think investors do realize what they’re getting. A lot of times, the way these are marketed is they beat the S&P 500.

Do they know what they're getting with the S&P 500? After seven years of a drawdown, guess what, the tech and telecomm stocks that drove the market bubble and the S&P 500 are still deadweights in arrears, return-wise.

And then finally, we create our own indexes for our own internal research, and we’ve compared the returns of those various indexes, our large-cap growth, our small-cap value, and so on, compared the returns over the last 40 years of our indexes versus these fundamental indexes.

Not quite the S&P 500 either.

We look at financial theory, and modern portfolio theory would not predict that small-cap value or any segment of the market would outperform. So you start wondering why, and I guess if you look at it in the broadest sense, some segment of the market had to outperform. If there’s a race, somebody’s going to win it.

And it so happened to be that the winner was small-cap value. Now, is there a reason for that? Quite honestly, it’s tough to point to a reason. Some of the researchers like Fama and French, who really got a lot of notoriety over their work identifying small-cap value, they propose that maybe you’re being compensated for taking on additional risk.

We’ve had a difficult time trying to identify the fact that there is additional risk. So it boils down to the fact that we don’t have a great deal of confidence saying that just because small-cap value has outperformed in the past means it will in the future. And you know, investors should just be aware that they’re taking that bet, that the past will repeat itself. We think it’s much more prudent to invest in the entire U.S. marketplace rather than a segment of it.

Modern portfolio theory or just CAPM — which is actually sub-optimal in mean-variance/MPT terms? Or is Gus simply confessing: "It is difficult to get a man to understand something when his job depends on not understanding it." After all, there's no guarantee that large-cap growth will perform in the future either. And BTW, which total market index? Why is MSCI better than Wilshire?

I mean, basically you’re getting beta; you’re getting the market rate of return, or a segment of the market.

Um hum. But maybe a better beta.

The interesting thing about a cap-weighted index is, it’s self-rebalancing. So you’re not going to incur transaction costs and potentially realize capital gains by continually rebalancing a fund. These fundamental indexes, since they are not cap-weighted, they need to be constantly rebalanced. And you have costs associated with that, and potentially capital gains that need to be paid.

Well, if you're talking S&P 500, how often is it rebalanced/restructured? And the Russell indices. How about a sampled index? How often are realized capital gains from rebalancing used to offset management fees? Would those capital gains have been more efficiently deployed in the shareholder's hands? Sorry, Gus, you're getting close to the financial pornography bus.

Okay guys, eviscerate me.

[Bylo Selhi]

why would you use a speculative take on a no-earner company with momentum gamblers behind it to construct your index?

How many of those are in the S&P 500 (or TSX 60 or EAFE)? What percentage of their respective indexes do such companies represent?

And there's a problem in investing in companies that actually earn money and show it by paying the owners? Or should we all have been subprime borrowers and lenders? Are we talking about stocks that reflect real economic growth, or just the ones that the punters favour?

No! No! No! So?

Do they know what they're getting with the S&P 500? After seven years of a drawdown, guess what, the tech and telecomm stocks that drove the market bubble and the S&P 500 are still deadweights in arrears, return-wise.

Again, so? No stock picker, not even Warren Buffett, picks winners much more often than losers. Why should a fundamental index creator be any different?

there's no guarantee that large-cap growth will perform in the future either.

Nope. (Not that I concede the S&P500 is a growth index...) The only guarantee with market-cap weighting is that you get whatever your chosen index has to offer in the future (minus the MER.) What guarantees does fundamental indexing offer (apart from substantially higher MERs, transactional costs and more taxes than cap-weighting)?

And BTW, which total market index? Why is MSCI better than Wilshire?

It's not as an index. The broader the index the less significant the differences. Compare the correlation between Russell 3000 vs. DJ/Wilshire 5000 vs. MSCI Total Market. However, the licensing costs to the fund sponsor may differ significantly. The index reconstitution rules may also minimize the frictional costs of implementation. Those sorts of differences are significant to be sure, but they're not, er, fundamentally significant for most investors.

Well, if you're talking S&P 500, how often is it rebalanced/restructured? And the Russell indices. How about a sampled index? How often are realized capital gains from rebalancing used to offset management fees?

Depends on the index, how broad it is and how it's rebalanced. Are you arguing that the same isn't also true of fundamental indexes?

Would those capital gains have been more efficiently deployed in the shareholder's hands?

How would the answer to that question differ for fundamental indexes?

[parvus]

parvus wrote:

why would you use a speculative take on a no-earner company with momentum gamblers behind it to construct your index?

How many of those are in the S&P 500 (or TSX 60 or EAFE)? What percentage of their respective indexes do such companies represent?
Quote:

And there's a problem in investing in companies that actually earn money and show it by paying the owners? Or should we all have been subprime borrowers and lenders? Are we talking about stocks that reflect real economic growth, or just the ones that the punters favour?

No! No! No! So?

Tough lessons of the S&P 500

Standard & Poor's has been criticized by some Wall Street pros for allowing tech shares to become such a big part of the index. But the firm says it had no choice. As investors swarmed into tech issues, they boosted their importance in the market, obligating S&P to add more to the index, said Howard Silverblatt, senior analyst at the firm.

"The S&P 500 emulates the market, good, bad or indifferent," he said.

Rediscovering Dividends

It has taken over two years for most investors to realize that the robust markets of recent years were an expensive illusion. Dishonest accounting of earnings, a manipulated new issues market, and aggressive stock buybacks helped create the impression that stock price increases were inevitable and unlimited. Throughout this entire crazy period one thing that remained relatively constant was the cash dividends paid out by companies. While reported earnings boomed for the 500 companies in the S&P 500 Index during 1999 and 2000, the level of dividends paid out by the 500 companies remained the same as in 1997. Dividends are often a reliable indicator of the sustainable, real earnings stream at a company. It is interesting that after rising sharply in 1998 and 1999, the S&P 500 index is now back to its 1997 level.

Index inclusion criteria seem a little flexible, and S&P admits they are not fully rules-based, and there are five criteria, not just market cap (free float, liquidity, *profitability* and representativeness).

FUNDAMENTAL ANALYSIS
The profitability criteria are four quarters of positive net income on an operating basis. Sometimes, Standard & Poor’s will include a company that would be profitable except for a loss due to a merger or acquisition. A recent example of this is JDS Uniphase (JDSU), which was added in July 2000.
<snip>
Finally, for the S&P 500 there are no capitalization restrictions. The guiding principle for inclusion in the S&P 500 is leading companies in leading US industries. Generally, companies are over $4 billion, although Standard & Poor’s sometimes adds companies below this range. For example, Visteon (VC) — a spin-off from Ford Motor Company — was added to the S&P 500 in June 2000 despite its market cap of only $1.6 billion because its sales were similar to those of S&P 500 companies and the company is considered to be a leader in its industry. In reality, most companies added to the S&P 500 are much bigger than $4 billion. For example, technology company JDS Uniphase entered the S&P 500 with a market cap of over $90 billion. Indeed, as of August 31, 2000, the average market cap for the S&P 500 was $26.7 billion, compared to $2.4 billion for the S&P MidCap 400 and $668 million for the S&P SmallCap 600.

I suspect we won't ever agree, er, fundamentally, on this. But I enjoy the opportunity to push my thinking and justify it, to figure out what's wrong and what's right. I wrote my lines as a kind of advocatus diabolis to Gus, to get my own thoughts going.

Even though I found the interview fascinating — thanks for linking it, BTW — I thought Gus was a tad more one-sided than he should have been, e.g., in not noting why number of employees was considered and rejected. He could have said that Arnott denies his indexes have a value and small-cap bias, but analysts think he does, and so does Vanguard's research; he could also have touched on the problem of constructing a really robust, low-turnover index. (Though that might be giving away trade secrets. )

And he could have mentioned that the FTSE RAFI indexes rebalance annually, just like Russell. Surely, in the sampling methodology he uses, there is rebalancing too. He's not saying, is he, that mid-cap value indexes are never rebalanced? Or ETFs based on them never paid a distribution? In effect, it's a level playing field: Vanguard's mid-cap value ETF faces the same problems as RAFI, albeit with a lower MER, which is certainly a factor to consider.

[gyrfalcon]

Even though I found the interview fascinating — thanks for linking it, BTW — I thought Gus was a tad more one-sided than he should have been, e.g.,

"Interviewer: I want to pick up on the idea of bets. At Vanguard, you manage a number of active quantitative strategies, which try to capture an index’s risk characteristics, while also providing outperformance through select bets. How is that strategy different from a fundamental index fund?

Gus Sauter: There are two components to a return. One is the market itself, and we call that beta. The second component of return is the manager’s capability, the portfolio manager. Do they have skill? And to the extent they do, we call that alpha. And their skill provides a return above and beyond what the market will give us. So they’re taking advantage of other investors in the marketplace.

I would note that most managers actually have negative alpha. Most managers do underperform in the marketplace. But what we’re trying to capture in our active quant work is we’re trying to provide the market rate of return. We’re trying to provide beta. And we’re also trying to provide that alpha, you know, some margin of return above and beyond the market rate of return."
------------------------------------------------------------------------------------------------------------------

What If Active Works?

http://www.indexuniverse.com/index.php? ... 4&issue=27

"But then a funny thing happened: The data got in the way. I compiled a list of all of Vanguard's active equity or mixed equity/fixed-income funds with at least a ten-year track record, and compared that ten-year performance to the most relevant benchmark (see "A Note on Methodology" section that follows). Of the 20 active funds, 14 beat their benchmark." ...

"But that's not what the data show. The data show that Vanguard's active funds beat their unmanaged benchmarks, and beat them cleanly, over long periods of time." ...

"While there are surely other reasons for the historical outperformance of Vanguard's active funds, the major sources are clear:

1) Low costs, which give them less distance to make up on their benchmarks each year.
2) Incentive payments tied to long-term performance (three to five years).
3) Strong restrictions on style drift.
4) Multiple managers that act to diversify fund strategy.

Whatever it is, it's working: Vanguard is running a lot of actively managed money, and it is running it well-particularly in the international and sector markets.

"It's funny," says Brennan. "Vanguard is well known for our index funds, and we do a great job with them. But if you combine all the assets, we actually have more actively run money than indexing money."

With the way they're beating their benchmarks, I'm not surprised."
-------------------------------------------------------------------------------------------------

[Taggart]

Investment News (U.S.)

Fundamentals-based index fund has goods, struggles for attention

Monday, June 18, 2007

“It was very novel for the patent office,” Mr. Sullivan said, “So novel, that it was hard to explain.”

That’s why it surprises Mr. Sullivan when Research Affiliates LLC chairman Robert Arnott, whose methodology debuted in 2004, often is credited with pioneering fundamentally based indexes. Pasadena, Calif.-based Research Affiliates’ fundamental-indexing methodology is used in managing more than $9 billion in assets worldwide.

Mr. Sullivan initially decided not to promote the fund until it had a five-year track record. He then postponed promotion until he received a patent.

[Norbert Schlenker]

Links shamelessly cribbed from Bogleheads.

Arnott presentation

Asness presentation

An Overwrought Orthodoxy – Arnott

The Value of Fundamental Indexing - Asness

Letter to the editor from Asness

[parvus]

Apparently Gus and Rob thrashed it out here. (I haven't had time yet to listen to it.)

[NormR]

Links shamelessly cribbed from Bogleheads.

Arnott presentation

Asness presentation

An Overwrought Orthodoxy – Arnott

The Value of Fundamental Indexing - Asness

Letter to the editor from Asness

A fine set of links, Asness hit Arnott hard.

Here's an Arnott Lecture.

[sweedy]

Links shamelessly cribbed from Bogleheads.

Arnott presentation

Asness presentation

An Overwrought Orthodoxy – Arnott

The Value of Fundamental Indexing - Asness

Letter to the editor from Asness

Unfortunately these links don't work anymore. Can anyone post the materials again?

[sweedy]

Just compared the performance of XIU with CRQ. From Feb. 22, 2006 (the inception date of CRQ) to Aug. 17, 2007, the holding period price return on CRQ is 15.97% while that of XIU is 13.85%. The annualized return of CRQ is 10.51% compared with XIU's 9.15%. XIU's distribution yield is about 0.4% higher than CRQ (1.9% versus 1.5%). So overall CRQ outperformed by 0.96% per year. Fundamental indexing worked in Canada during this short period.

Of course, it is uncertain whether CRQ can continue beating XIU. Two other concerns: 1. CRQ is not very liquid with small trading volume. 2. If Claymore quits in a few years CRQ investors will be left with significant capital gains.

Solution: I am going to buy some CRQ in my RRSP.

[sweedy]

I also noticed that many financial gurus in FWF chose not to comment about fundamental indexing. For example, pitz, what do you think?

[NormR]

I also noticed that many financial gurus in FWF chose not to comment about fundamental indexing. For example, pitz, what do you think?

Ok, so I'm not pitz.

Fundamental Indexing (TM) is cleverly packaged value investing. As such, you might also consider other value offerings such a DFA's funds, value ETFs, the O'Shaughnessy funds, even active value managers, etc.

[sweedy]

I also noticed that many financial gurus in FWF chose not to comment about fundamental indexing. For example, pitz, what do you think?

Ok, so I'm not pitz.

Fundamental Indexing (TM) is cleverly packaged value investing. As such, you might also consider other value offerings such a DFA's funds, value ETFs, the O'Shaughnessy funds, even active value managers, etc.

Thanks for your comment Norm. I remember Arnott and his colleagues tried to show that market value indexing overweight overvalued stocks and underweighted undervalued stocks while fundamental index does not. What they did not emphasize is that because the "undervalued stocks" tend to become value stocks, fundamental indexing will put more weight on value stocks. So I agree that fundmental indexing is value investing repackaged.

[DenisD]

So I agree that fundmental indexing is value investing repackaged.

I prefer to think of it as value investing with a growth component. So, it should outperform value investing when growth outperforms, like recently. It might underperform value investing during value manias. IIRC, a back test showed that it outperformed at least one value index.

Obviously, the main disadvantage is higher fees.

[NormR]

So I agree that fundmental indexing is value investing repackaged.

I prefer to think of it as value investing with a growth component. So, it should outperform value investing when growth outperforms, like recently. It might underperform value investing during value manias. IIRC, a back test showed that it outperformed at least one value index.

Obviously, the main disadvantage is higher fees.

The 'growth' bit is just marketing spin. Asness (in the now invisible paper) destroyed the argument. It's just a different way of value investing and not all that different.

[DenisD]

The 'growth' bit is just marketing spin. Asness (in the now invisible paper) destroyed the argument. It's just a different way of value investing and not all that different.

I'll take your word for it. But the fundamental indexes do contain growth companies. I imagine, but don't know, more than value indexes.

Here's my worthless anecdotal evidence.

Code: Select all

Fund Name                 Tick   2006  YTD    1M    3M   12M
iSh Rus 1000 Growth Index IWF     8.7   6.3  -1.4   0.3  18.7
iSh Rus 1000 Index        IWB    14.9   3.4  -3.2  -2.0  15.3
iSh Rus 1000 Value Index  IWD    21.7   0.2  -5.0  -4.4  11.6
iSh Rus 2000 Growth Index IWO    13.1   3.7  -5.1  -1.5  16.3
iSh Rus 2000 Index        IWM    17.8  -0.8  -7.0  -4.3  11.9
iSh Rus 2000 Value Index  IWN    22.8  -5.6  -8.6  -7.7   6.2
PowerShares US 1000       PRF    18.0   2.7  -4.5  -2.5  14.5
PowerShares US 1500       PRFZ      ~   1.5  -5.9  -3.3     ~


From ETFZone to 31/Jul.

Notice that value outperformed last year, growth outperformed this year. Value beat RAFI last year, RAFI beat value this year.

I'm sure you can come up with anecdotal evidence to the contrary.

[NormR]

Here's the Asness paper

[DenisD]

Here's the Asness paper

Thanks for posting that link, Norm. I vaguely remember reading the article a month or two ago.

Frankly, I don't see what all the fuss is about. First, he discusses what an index is and is not and whether or not FI is an index. And whether it is active or passive. Who cares! Then, we have the startling revelation that if everyone invested in FI, it would be the same as the cap-weighted index. Next, low and behold, stocks whose dividend yields match the market have the same weight in fundamental and cap-weighted indexes. So what! Finally, there's the old "it's nothing new" argument.

In conclusion, what does he say?

In addition, at the right investment management fee and given more information about the details, I am a potential fan of fundamentally constructed portfolios.
.
.
.
The approach is a clean, simple way to describe how to construct an active, value based portfolio.

I have to disagree with one thing he said.

Both efficient and inefficient market explanations for the success of value predict higher returns for Fundamental Indexes, so evidence that they have higher returns adds literally nothing to this ongoing discussion.

By "evidence" he means a back test. I think a back test is essential to any mechanical method of investing. The longer, the better.

Fundamental Indexing (TM) is cleverly packaged value investing. As such, you might also consider other value offerings such a DFA's funds, value ETFs, the O'Shaughnessy funds, even active value managers, etc.

I'm curious, Norm. O'Shaughnessy did a 40 plus year back test of his algorithms. Without the back test, would you still include the O'Shaughnessy funds in that sentence?

[NormR]

Fundamental Indexing (TM) is cleverly packaged value investing. As such, you might also consider other value offerings such a DFA's funds, value ETFs, the O'Shaughnessy funds, even active value managers, etc.

I'm curious, Norm. O'Shaughnessy did a 40 plus year back test of his algorithms. Without the back test, would you still include the O'Shaughnessy funds in that sentence?

Just for clarity, the aforementioned list is in no way a recommended list.

But yes, O'Shaughnessy value is basically a high dividend yield fund. One can simply categorize it by the method used. In this context, past performance or back testing isn't all that interesting.

It's funny to see various 'academics' claiming to have 'invented' value investing using another name. At least O'Shaughnessy didn't. But it strikes me that FF and Arnott have both tried and succeeded to different degrees.

[DenisD]

O'Shaughnessy value is basically a high dividend yield fund.

I assume you mean US Value. It's not so high yield any more since he started adding buybacks to yield a few years ago. Some of the new funds are partially high yield. But they use more than one strategy.

I guess we'll have to agree to disagree on the importance of back testing.

[NormR]

O'Shaughnessy value is basically a high dividend yield fund.

I assume you mean US Value. It's not so high yield any more since he started adding buybacks to yield a few years ago. Some of the new funds are partially high yield. But they use more than one strategy.

Yes, I was referring to the value slice. Using shareholder yield isn't a particularly novel value strategy.

I guess we'll have to agree to disagree on the importance of back testing.

Sure, we can if you like. But if you want to figure out a fund's style then it is far easier to look at the method used. In this regard, O'Shaughnessy's funds are easier than most because they provide the 'formula' used to pick stocks. Tweezing fund style out of past performance alone is difficult and requires a huge chunk of data.

If you are considering an investment, testing is more interesting. But one does run the danger of becoming overly confident and data mining.