FTSE RAFI Canadian Index ETF to launch 22Feb06

Discuss your favourite picks, broker, and trading or investment style.
YogiBear
Veteran Contributor
Veteran Contributor
Posts: 1482
Joined: 20 May 2006 20:13

Post by YogiBear »

Shakespeare wrote:The top ten weightings:

Code: Select all

Weighting
Royal Bank of Canada (RY) 5.52%
BCE Inc (BCE) 4.79%
Bank of Nova Scotia (BNS) 4.50%
Manulife Financial (MFC) 4.18%
EnCana (ECA) 4.01%
Toronto-Dominion Bank (TD) 3.88%
Bank of Montreal (BMO) 3.54%
CIBC (CM) 3.00%
Thomson Corp (TOC) 2.88%
Sun Life Financial (SLF) 2.87%
Fascinating- all this attention to a so-called new "Fundamental Index", yet of the top 10 listed above, 6 are in the top 10 of XIU- and 3 of the other 4 (BCE, CM, SLF) are just below the top 10. Dramatically new index- or dramatically effective marketing???

Code: Select all

XIU Top Holdings as of June 2, 2006
6.49%	 ROYAL BANK OF CANADA
6.36%	 MANULIFE FINANCIAL CORPORATION
5.32%	 ENCANA CORPORATION
4.84%	 BANK OF NOVA SCOTIA (THE)
4.70%	 TORONTO-DOMINION BANK (THE)
4.53%	 SUNCOR ENERGY INC.
3.53%	 CANADIAN NATURAL RESOURCES LIMITED
3.42%	 BANK OF MONTREAL
3.16%	 BARRICK GOLD CORPORATION
2.93%	 CANADIAN NATIONAL RAILWAY COMPANY
To be fair, the cause of this rather substantial overlap may have less to do with any fault with the new "Fundamental Index", and more to do with the rather thin Canadian market wrt companies that satisfy the "four factors":
George$ quoting Claymore Investments wrote:total cash dividends (five-year average of all regular and special distributions)
free cash flow (five-year average cash flow)
total sales (five-year average total sales)
book equity value (current period book equity value)
So, what's the point? A way to "index" Canadian large/ mid-caps with a similar gain/ loss profile (a correlation of 0.90 between this index and the MSCI Canada index was mentioned upthread) but a (presumably) lower volatility than the S&P/ TSX 60? Couldn't roughly the same thing be accomplished by mixing XIU with an allocation to T-Bills (absent tax considerations, of course)?

Just wondering ... :wink:
DenisD
Veteran Contributor
Veteran Contributor
Posts: 4081
Joined: 19 Feb 2005 01:24
Location: Calgary

Post by DenisD »

DenisD wrote:According to this, RAFI Canada outperformed MSCI Canada by 3.76% using 20 years of data (1984 through 2004) with less volatility.
YogiBear wrote:So, what's the point?
Should be obvious if you believe that 3.76% spread. :wink:
YogiBear
Veteran Contributor
Veteran Contributor
Posts: 1482
Joined: 20 May 2006 20:13

Post by YogiBear »

DenisD wrote:
DenisD wrote:According to this, RAFI Canada outperformed MSCI Canada by 3.76% using 20 years of data (1984 through 2004) with less volatility.
YogiBear wrote:So, what's the point?
Should be obvious if you believe that 3.76% spread. :wink:
Hhmm, so, it's a matter of belief, is it? There are 3 points that you are not considering.

First of all, I would like to know exactly where that spread comes from, assuming the validity of the figure (I can't access the site to see the methodology, as it seems to require registration?). Consider that over the past 20 odd years, the big financial companies (banks and insurance) have generally outperformed the Canadian market on a total return basis. Yet, what % of the total market capitalisation did those companies make up in 1984? I would bet you pounds to pennies that their aggregate portion was much smaller back then than is the case today.

So, of your 3.76%, what part represents the (likely) larger allocation by the RAFI index than the general one in 1984 to financial services firms, and the slow convergence in allocation over the years as these companies made up a growing share of the total market while, in aggregate, having higher than market returns? I would suggest a substantial part. Given the very substantial overlap between the "fundamental index" holdings of these companies and the constituents of the TSX 60 (in XIU), how can the RAFI index continue to obtain excess returns over the TSX index in the future from holding these companies?

The second point should be obvious, but it seems to bear repeating. The 3.76% spread you quote is excess returns in the past. I will not argue with the assertion that value investing has the potential to provide excess returns, but I will ask you to suggest, given the size of the financial services industry today, where the RAFI index will continue to find a positive 3.76% spread over the TSX 60 in the future? Remember, the excess returns can only come from 2 places: differences in the stocks that make up the two indexes, and/ or differential weights assigned to the same stocks between the two indexes. Look again at what was posted:
YogiBear wrote:... of the top 10 listed above [for the RAFI index], 6 are in the top 10 of XIU- and 3 of the other 4 (BCE, CM, SLF) are just below the top 10
And most of the percentage allocations are fairly similar! Aside from the obvious presence of Barrick Gold in the TSX 60 instead of Thomson Corp (which is scary :shock: ), where do you see potential, at least in the top holdings, for a 3.76% outperformance going forward???

As a third, and final point: value outperformance, even if we accept for the sake of discussion that it exists, is a tricky beast. It is not consistent, but (apparently) ebbs and flows like the tide. A specific investor, working within a specific investing time-frame, cannot simply expect that over X period, his/ her investment in a RAFI index product will return 3.76% above the general Canadian index, minus expenses. The return may, in fact, trail the general index, even over a fairly long time period. That is why I asked, if a high correlation with the general index is desired but with reduced volatility, whether it would be more appropriate to consider a simpler (and much lower cost) solution:
YogiBear wrote:So, what's the point? A way to "index" Canadian large/ mid-caps with a similar gain/ loss profile (a correlation of 0.90 between this index and the MSCI Canada index was mentioned upthread) but a (presumably) lower volatility than the S&P/ TSX 60? Couldn't roughly the same thing be accomplished by mixing XIU with an allocation to T-Bills (absent tax considerations, of course)?
N.B. that my discussion of the relative merits of the RAFI index approach in Canada is distinct from my views on the value (pun intended) of Graham-style value stock-picking. If one creates one's own portfolio of individual stocks, then overlap with the TXS 60, for example, is only an issue if you want it to be ...
George$
Veteran Contributor
Veteran Contributor
Posts: 2612
Joined: 18 Feb 2005 20:46
Location: Toronto

Post by George$ »

Yogi:

I haven't considered your question in any detail. But let me say that Robert Arnott is probably one of the most astute observers of the financial markets I know. I cannot imagine him data-mining and promoting a flaky issue.

Have you watched and read his presentation? It is well worth the hour methinks. A lot of interesting slides accompany it.
http://www.claymoreinvestments.ca/ETF.aspx
YogiBear wrote: ..... Consider that over the past 20 odd years, the big financial companies (banks and insurance) have generally outperformed the Canadian market on a total return basis. Yet, what % of the total market capitalisation did those companies make up in 1984? I would bet you pounds to pennies that their aggregate portion was much smaller back then than is the case today.
You raise the matter of financial institutions. He shows two charts for the US. Both indexes have similar financial profiles over time - but not so with the other sectors.
Two of Arnott's 45 slides ...
http://www.claymoreinvestments.ca/Docs/ ... de0085.htm
and
http://www.claymoreinvestments.ca/Docs/ ... de0085.htm
YogiBear
Veteran Contributor
Veteran Contributor
Posts: 1482
Joined: 20 May 2006 20:13

Post by YogiBear »

George$ wrote:Yogi:

I haven't considered your question in any detail. But let me say that Robert Arnott is probably one of the most astute observers of the financial markets I know. I cannot imagine him data-mining and promoting a flaky issue.

Have you watched and read his presentation? It is well worth the hour methinks. A lot of interesting slides accompany it.
http://www.claymoreinvestments.ca/ETF.aspx
YogiBear wrote: ..... Consider that over the past 20 odd years, the big financial companies (banks and insurance) have generally outperformed the Canadian market on a total return basis. Yet, what % of the total market capitalisation did those companies make up in 1984? I would bet you pounds to pennies that their aggregate portion was much smaller back then than is the case today.
You raise the matter of financial institutions. He shows two charts for the US. Both indexes have similar financial profiles over time - but not so with the other sectors.
Two of Arnott's 45 slides ...
http://www.claymoreinvestments.ca/Docs/ ... de0085.htm
and
http://www.claymoreinvestments.ca/Docs/ ... de0085.htm

[emphasis added]
I'm sure that I was not as clear as I might/ should have been, so let me clarify my position: I was not, in either post upthread, taking issue with the whole idea of "fundamental indexes" in general, nor with Arnott's approach in particular. Neither was I offering any comments on the validity of the fundamental index approach to such markets as the US, UK, Japan, etc. My comments were specifically directed at the Canadian market, and more specifically at the apparent implementation of the approach by the RAFI product, as evidenced by the extensive overlap between its holdings and those of the broader (non-fundamental, cap-weighted) TSX 60 index:
YogiBear wrote:To be fair, the cause of this rather substantial overlap may have less to do with any fault with the new "Fundamental Index", and more to do with the rather thin Canadian market wrt companies that satisfy the "four factors":
George$ quoting Claymore Investments wrote:total cash dividends (five-year average of all regular and special distributions)
free cash flow (five-year average cash flow)
total sales (five-year average total sales)
book equity value (current period book equity value)
On a more specific point, the aggregate weight of financial companies within the Canadian cap-weighted index and the changes to that relative weight over time:
George$ wrote:You raise the matter of financial institutions. He shows two charts for the US. Both indexes have similar financial profiles over time - but not so with the other sectors. [emphasis added]
I'm sure that I don't need to remind anyone of the dramatic differences between the makeup and relative sector allocations of the Canadian and US markets. WADR the fact that the relative weight of the US financial services sector has not changed over time is not relevant to the Canadian market. I also do not know whether the US financial services sector has seen the same level of total return out-performance relative to the total market as has been very clearly the case in Canada.

I do not know how much work, if any, Arnott and colleagues have done in testing the application of their approach to a Canadian context. That was not the point of my criticisms. My focus was the empirical fact that the RAFI index etf was not substantially different from XIU, at least among the top 40 odd %, yet claims for a 3.76% performance spread were being thrown around like leaves in an Autumn wind. Where did that spread come from- and more importantly, where will it come from in the future?
User avatar
Bylo Selhi
Veteran Contributor
Veteran Contributor
Posts: 29494
Joined: 16 Feb 2005 10:36
Location: Waterloo, ON
Contact:

Post by Bylo Selhi »

YogiBear wrote:Where did that spread come from
Nortel et al?
and more importantly, where will it come from in the future?
Nortel et al? :twisted: :lol:
Sedulously eschew obfuscatory hyperverbosity and prolixity.
YogiBear
Veteran Contributor
Veteran Contributor
Posts: 1482
Joined: 20 May 2006 20:13

Post by YogiBear »

Bylo Selhi wrote:
YogiBear wrote:Where did that spread come from
Nortel et al?
Certainly, in part at least. But is that an argument for the RAFI product, or for the old XIC?!
Bylo Selhi wrote:
YogiBear wrote:and more importantly, where will it come from in the future?
Nortel et al? :twisted: :lol:
Oh yeah? Look again at the relative weights of the top holdings in RAFI and XIU. Anybody see another Nortel (other than Barrick, which I agree is ...)? They are both stuffed with the big banks and insurancecos. Some difference ... So, what about that 3.76%? :lol:

On a more serious note: I'm not a stock-picker (our portfolio is entirely indexed), but for those who have the interest, skills, and time to do so properly, ISTM that a "roll-your-own" portfolio of value stocks would make a lot more sense than this RAFI etf. One could adjust the sector allocations, cap sizes, dividend payouts, currency weights, etc to precisely suit one's needs.

OTOH, I am interested to see how fundamental indexing works over time going forward, particularly in larger, more diversified markets such as the US and UK. If it continues to work well, then ... :D
User avatar
Bylo Selhi
Veteran Contributor
Veteran Contributor
Posts: 29494
Joined: 16 Feb 2005 10:36
Location: Waterloo, ON
Contact:

Post by Bylo Selhi »

YogiBear wrote:
Bylo Selhi wrote:
YogiBear wrote:and more importantly, where will it come from in the future?
Nortel et al? :twisted: :lol:
Oh yeah? Look again at the relative weights of the top holdings in RAFI and XIU. Anybody see another Nortel (other than Barrick, which I agree is ...)? They are both stuffed with the big banks and insurancecos. Some difference ... So, what about that 3.76%? :lol:
I was alluding to the possibility of Nortel making a very big comeback :lol:
Sedulously eschew obfuscatory hyperverbosity and prolixity.
YogiBear
Veteran Contributor
Veteran Contributor
Posts: 1482
Joined: 20 May 2006 20:13

Post by YogiBear »

Bylo Selhi wrote:
YogiBear wrote:
Bylo Selhi wrote:Nortel et al? :twisted: :lol:
Oh yeah? Look again at the relative weights of the top holdings in RAFI and XIU. Anybody see another Nortel (other than Barrick, which I agree is ...)? They are both stuffed with the big banks and insurancecos. Some difference ... So, what about that 3.76%? :lol:
I was alluding to the possibility of Nortel making a very big comeback :lol:
Sorry- my mistake!

Can I take it that you have some nagging doubts about the certainty of a 3.76% performance gap going forward- or are you still hoping to recoup the $125 ACB of the Nortel shares you still hold???!!! :P

(joke, that was a joke ... :wink: )
User avatar
NormR
Veteran Contributor
Veteran Contributor
Posts: 5234
Joined: 18 Feb 2005 11:19
Contact:

Post by NormR »

George$ wrote:Just as there are S&P, NYSE, MSCI etc variations of cap-weighted indexes, - why can't there be variations of the non-cap weighted index with lower MERs?
There are several ... Monkey Business.

Arnott's index falls somewhere between market-cap weighted and equal-weighted indices. You can buy both in the US. You can also tilt to value by brewing your own (i.e buying a regular low-cost market-cap index plus a value index like IVE for 0.18%).

In Canada it seems that everything is a little more expensive. But the market is smaller and one could consider tilting by buying individual stocks (Unbundling Canadian ETFs). Such an approach can be considered for even smaller portfolios when implemented via a low-commission broker such as Interactive Brokers.
George$
Veteran Contributor
Veteran Contributor
Posts: 2612
Joined: 18 Feb 2005 20:46
Location: Toronto

Post by George$ »

YogiBear wrote: .... . My focus was the empirical fact that the RAFI index etf was not substantially different from XIU, at least among the top 40 odd %, yet claims for a 3.76% performance spread were being thrown around like leaves in an Autumn wind. Where did that spread come from- and more importantly, where will it come from in the future?
Yogi: I think Arnott discusses this for the US data in slides
26
http://www.claymoreinvestments.ca/Docs/ ... de0344.htm
note that the two lists are not that different (likewise in Canada)
27
http://www.claymoreinvestments.ca/Docs/ ... de0345.htm
here he disusses how the overvalued equities do become "Fallen Angels" etc
28
http://www.claymoreinvestments.ca/Docs/ ... de0346.htm

I cannot do justice to Arnott with these illustrations . You must listen to his full presention - via links at http://www.claymoreinvestments.ca/ETF.aspx

Why wouldn't similar reasons apply to the Canadiam market?

I don't think the 3.76% is a given or hardwired in any way. As he presents in his data, there will be times when his index falls behind the cap-weighted index. But over time his reasons seem compelling -- that one should expect better returns.

My question is - will the better returns be above the 0.65% MER hurdle? Dunno.
User avatar
Bylo Selhi
Veteran Contributor
Veteran Contributor
Posts: 29494
Joined: 16 Feb 2005 10:36
Location: Waterloo, ON
Contact:

Post by Bylo Selhi »

YogiBear wrote:Can I take it that you have some nagging doubts about the certainty of a 3.76% performance gap going forward- or are you still hoping to recoup the $125 ACB of the Nortel shares you still hold???!!! :P
Touché :lol:
Sedulously eschew obfuscatory hyperverbosity and prolixity.
YogiBear
Veteran Contributor
Veteran Contributor
Posts: 1482
Joined: 20 May 2006 20:13

Post by YogiBear »

George$ wrote:I don't think the 3.76% is a given or hardwired in any way. As he presents in his data, there will be times when his index falls behind the cap-weighted index. But over time his reasons seem compelling -- that one should expect better returns.

My question is - will the better returns be above the 0.65% MER hurdle? Dunno.
Neither do I.

The concern underlying my criticisms is that fundamental indexing seems to be so much the creation, or outgrowth might be a better term, of the US market experience. Nothing wrong with that if one is investing Stateside- but the Canadian market really is different, as we know (a much higher level of natural resource related cyclicals is one obvious current difference, among others!). With a relative paucity of large companies to begin with, and even fewer that would meet the criteria to be included in the RAFI thing, is it any wonder that there is substantial overlap with XIU?

To put it another way, if one is putting together an etf holding 60-100 names, let's say, and there are two pools from which to choose, one with about 120 companies (TSX 60 and mid-cap), the other with 500+ companies (this does not even include large-cap companies not in the S&P 500, or those in the midcap 400, etc), which fundamental index is more likely to look, and act over time, substantially different from the market-cap index? Since finding a (positive) difference is the whole point of the exercise, which market is more suited to a fundamental indexing approach?

As you said, the real point is that even if one assumes better returns over time from the RAFI etf than from XIU (for example), "will the better returns be above the 0.65% MER hurdle?" (To compare with XIU, one should say "above the (0.65% - 0.17% =) 0.48% MER hurdle"). While I have my doubts, I really don't know either. I could be wrong- wouldn't be the first time, won't be the last! :P But I'm not buying in ...
YogiBear
Veteran Contributor
Veteran Contributor
Posts: 1482
Joined: 20 May 2006 20:13

Post by YogiBear »

George$ wrote:I haven't considered your question in any detail. But let me say that Robert Arnott is probably one of the most astute observers of the financial markets I know. I cannot imagine him data-mining and promoting a flaky issue.
I forgot to link this article by William Bernstein from his website discussing the work of Arnott et al. I don't think it has been linked on this thread, but was perhaps elsewhere.

Interestingly, on the basis of an analysis of fundamental indexing using the Three-Factor model, Bernstein concludes in part that:
... at a rough approximation, slightly less than two-thirds of the "excess return" of the RAFI over the S&P [500] is due to naïve [value] factor exposure , and slightly more than one-third seems to be inherent to the technique. Unfortunately, this latter effect is not statistically significant, raising the issue of data mining. [emphasis added]
Bernstein summarizes his conclusions with two points that are relevant to our discussion here:
  • 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.
[emphasis added]
My sentiments exactly, wrt the implementation in Canada of this approach through the RAFI Canada etf.
User avatar
yielder
Veteran Contributor
Veteran Contributor
Posts: 4911
Joined: 16 Feb 2005 07:47
Location: Hastings, Ontario

Post by yielder »

YogiBear wrote:One could adjust the sector allocations, cap sizes, dividend payouts, currency weights, etc to precisely suit one's needs.
In theory, yes. In practice, no. You can do what you suggest or you can buy when individual stocks are cheap. It's not often that you can do both unless it's a bear market where eveyrthing is being sold down.
YogiBear
Veteran Contributor
Veteran Contributor
Posts: 1482
Joined: 20 May 2006 20:13

Post by YogiBear »

yielder wrote:
YogiBear wrote:One could adjust the sector allocations, cap sizes, dividend payouts, currency weights, etc to precisely suit one's needs.
In theory, yes. In practice, no. You can do what you suggest or you can buy when individual stocks are cheap. It's not often that you can do both unless it's a bear market where eveyrthing is being sold down.
My comments from upthread concerning adjustable "roll-your-own" fundamental index portfolios that you quote above, and my question(s) for you below, come out of a basic belief that was summarized very well:
Upthread, Norbert Schlenker wrote:Being an advocate of disintermediation, may I suggest that big portfolios just roll their own? [emphasis added]
In my case, being an indexer, I rely upon intermediation in the form of index funds and etfs, because they provide vast diversification at a (relatively) low cost. I also make every effort to reduce the financial drag of that intermediation as much as I can by using the lowest cost instruments available to me.

My first concern about the RAFI etf is that relatively high intermediation costs are being charged for a product that seems remarkably similar to XIU, which is much cheaper. Second, unlike some others here, I have doubts about whether there will be enough, or any, future out-performance to justify that excess cost. Thus my suggestion that anyone who likes the concept of fundamental indexing might think about "rolling-your-own" (as long as trading costs are controlled, as set out by Norbert Schlenker upthread), with adjustments as desired.

I believe that your investment approach more or less follows the concept of putting together a portfolio of individual stocks that meet certain criteria for income, etc.? If so, in your experience, does the gap between theory and practice that you identify argue in favour of using instruments such as the RAFI etf or XDV, or the more numerous US value/ dividend-oriented etfs and funds, instead of individual stock purchases in some cases? If so, what would those cases be? In other words, when (if ever) is pre-packaged intermediation a better option for value oriented (or dividend-income oriented) investors, in light of the difficulties you identify in creating an optimal "roll-your-own" value stock portfolio at an acceptable price?
User avatar
Shakespeare
Veteran Contributor
Veteran Contributor
Posts: 23396
Joined: 15 Feb 2005 23:25
Location: Calgary, AB

Post by Shakespeare »

In other words, when (if ever) is pre-packaged intermediation a better option for value oriented (or dividend-income oriented) investors
When the market is large, the work of screening is substantial, and the sub-index offers reasonable diversication at low cost. None of these points apply to the TSX.
Sic transit gloria mundi. Tuesday is usually worse. - Robert A. Heinlein, Starman Jones
User avatar
Norbert Schlenker
Veteran Contributor
Veteran Contributor
Posts: 7960
Joined: 16 Feb 2005 09:56
Location: An Argument Surrounded By Water
Contact:

Post by Norbert Schlenker »

Yogi, you might also be interested in reading this thread.
Nothing can protect people who want to buy the Brooklyn Bridge.
YogiBear
Veteran Contributor
Veteran Contributor
Posts: 1482
Joined: 20 May 2006 20:13

Post by YogiBear »

Shakespeare wrote:
In other words, when (if ever) is pre-packaged intermediation a better option for value oriented (or dividend-income oriented) investors
When the market is large, the work of screening is substantial, and the sub-index offers reasonable diversication at low cost. None of these points apply to the TSX.
Agreed.
Norbert Schlenker wrote:Yogi, you might also be interested in reading this thread.
Thank you, Norbert, I had followed that thread last Winter, but forgot about it recently. I see that was where Bylo had already referenced the Bernstein article in which he dissects fundamental indexing using the three-factor model (and finds some possible evidence of data mining on Mr. Arnott's part, as I quoted upthread).

If I may be permitted a "final" rant on this subject: I have found, both in this thread and the one cited by Norbert, a tendancy to uncritically transfer the data and expectations concerning fundamental indexing in the US propounded by Arnott and his supporters to the Canadian context. This does not, of course, mean that RAFI will not work here- but it does mean that, in the absence of specifically Canadian data and testing, the case for fundamental indexing in Canada has not yet been made.

Before the famous 3.76% performance bonus is trotted out again, let me be clear what I mean by "specifically Canadian data and testing". I mean the sort of rigourous, third-party testing that was done by William Bernstein in the article referred to above: factor performance attribution using the three-factor model on Canadian data, including an assessment of the extent to which the RAFI "outperformance" is simply the result of inherent exposure to the Fama-French factors of size (i.e., smaller companies) and value, independent of any specific virtues of RAFI. This assessment would also include a determination of the statistical significance of any results, to allow for conclusions to be drawn concerning the possibility of data mining.

Once this sort of testing has been carried out and reviewed, then, and only then, will we have any real idea about the validity or not of the RAFI approach in Canada. Until then, any claims for fundamental indexing here are simply based on faith or hope (or theory). So, does anyone know of such testing? Have Arnott and colleagues carried any out for Canada? Claymore? Anyone?! :shock: :P
DenisD
Veteran Contributor
Veteran Contributor
Posts: 4081
Joined: 19 Feb 2005 01:24
Location: Calgary

Post by DenisD »

YogiBear wrote:The concern underlying my criticisms is that fundamental indexing seems to be so much the creation, or outgrowth might be a better term, of the US market experience.
Outperformed cap-weighted indexes in 22 other countries.
YogiBear
Veteran Contributor
Veteran Contributor
Posts: 1482
Joined: 20 May 2006 20:13

Post by YogiBear »

DenisD wrote:
YogiBear wrote:The concern underlying my criticisms is that fundamental indexing seems to be so much the creation, or outgrowth might be a better term, of the US market experience.
Outperformed cap-weighted indexes in 22 other countries.
Good- if you are investing in one of those 22 countries. But is Canada one of those 22? And have the results that indicated "[o]utperformed cap-weighted indexes" been subjected to the sort of analysis carrried out by Bernstein in the article cited upthread, where even in the US he concludes:
William Bernstein wrote:Thus, at a rough approximation, slightly less than two-thirds of the "excess return" of the RAFI over the S&P is due to naïve factor exposure [i.e., because RAFI is largely composed of stocks that fall within the "Value" universe, there is an automatic value-related return benefit that has nothing to do with RAFI as such- any value index would benefit], and slightly more than one-third seems to be inherent to the technique. Unfortunately, this latter effect is not statistically significant, raising the issue of data mining. [emphasis added]
I assume that you understand the import of that quote: even for the US, Bernstein finds that there is no statistically significant evidence that the RAFI in and of itself has added any value at all over the S&P 500, other than that added by the 2 Fama-French factors (size and value), which can be accessed through any small-cap or value fund- or get them both at the same time with VBR!

Unless there is evidence out there that specifically addresses these issues, the conclusion for the US seems pretty clear (never mind Canada, where no one has yet advanced significant evidence at all)- this dawg don't bark ... :shock: :P
George$
Veteran Contributor
Veteran Contributor
Posts: 2612
Joined: 18 Feb 2005 20:46
Location: Toronto

Post by George$ »

YogiBear wrote: ...If I may be permitted a "final" rant on this subject: I have found, both in this thread and the one cited by Norbert, a tendancy to uncritically transfer the data and expectations concerning fundamental indexing in the US propounded by Arnott and his supporters to the Canadian context. This does not, of course, mean that RAFI will not work here- but it does mean that, in the absence of specifically Canadian data and testing, the case for fundamental indexing in Canada has not yet been made.

Before the famous 3.76% performance bonus is trotted out again, let me be clear what I mean by "specifically Canadian data and testing". ...
Yogi;
I think I agree with you on the major issues but I'm wondering if you may not be overstating it. Yes to "statistically significant" but realizing it in finance is difficult and not always a exact science. And too often we get fooled by randomness via data mining in any of its many cloaks.

Both Bernstein and Arnott are sharp and honest (in my view) and I read them both. I welcome one being critical of the other and I try not to assume either uncritically.

I'm not ready to dismiss "fundamental indexing". Why am I interested in this topic? Answer: Arnott's past reputation and the reasoning he gives in his one hour web cast seemed reasonable and compelling. I'm inclined to think his reasoning is sound and has merit. But at the same time I don't give much reality (for going forward) to the 3.76% Canadian outperformance from the past.

Yes the Canadian market is different in many ways from the US but surely there are also common underlying issues that are the same. And if Arnott's logic has merit in the US, then it may also have the same merit here.

My two cents in the morning before I have my cup of coffee for the day. :roll:
DenisD
Veteran Contributor
Veteran Contributor
Posts: 4081
Joined: 19 Feb 2005 01:24
Location: Calgary

Post by DenisD »

YogiBear wrote:I assume that you understand the import of that quote: even for the US, Bernstein finds that there is no statistically significant evidence that the RAFI in and of itself has added any value at all over the S&P 500, other than that added by the 2 Fama-French factors (size and value), which can be accessed through any small-cap or value fund- or get them both at the same time with VBR!
I confess I've forgotten most of my limited knowledge of statistics.

Unfortunately, Canadians have no cheap small-cap value ETF like VBR. Are you saying that, if Bernstein's results were the same in Canada, we could just buy CRQ and somehow, magically, get small-cap value exposure as a bonus?
User avatar
NormR
Veteran Contributor
Veteran Contributor
Posts: 5234
Joined: 18 Feb 2005 11:19
Contact:

Post by NormR »

George$ wrote:But at the same time I don't give much reality (for going forward) to the 3.76% Canadian outperformance from the past.
It seems to me that the more interesting question is, "Is Arnott's new index fund likely to do better than a regular index fund (pick a value fund if you like) by more than the all-in fee difference?"
YogiBear
Veteran Contributor
Veteran Contributor
Posts: 1482
Joined: 20 May 2006 20:13

Post by YogiBear »

George$ wrote:I think I agree with you on the major issues but I'm wondering if you may not be overstating it. Yes to "statistically significant" but realizing it in finance is difficult and not always a exact science. And too often we get fooled by randomness via data mining in any of its many cloaks.

Both Bernstein and Arnott are sharp and honest (in my view) and I read them both. I welcome one being critical of the other and I try not to assume either uncritically.

I'm not ready to dismiss "fundamental indexing". Why am I interested in this topic? Answer: Arnott's past reputation and the reasoning he gives in his one hour web cast seemed reasonable and compelling. I'm inclined to think his reasoning is sound and has merit. But at the same time I don't give much reality (for going forward) to the 3.76% Canadian outperformance from the past.

Yes the Canadian market is different in many ways from the US but surely there are also common underlying issues that are the same. And if Arnott's logic has merit in the US, then it may also have the same merit here.
I certainly did overstate the case in my last post in response to DenisD: to once again repeat examples of "outperformance" without looking into how that excess return was earned is meaningless. If I say that my small cap value fund outperformed the S&P 500 over the past 20 years, well, so what?! If we believe Fama-French, that is what we would expect on average. OTOH, Bernstein is not so categorical as to say that fundamental indexing is useless; he only says that its value-added in the US appears "relatively small" and has yet to be statistically proven.

You are absolutely correct, of course, that using statistical evidence, which necessarily is always from the past, to provide guidance for the future is a process that one should undertake with great care. Yet those tools do provide insight, so at least they should be used in a way that allows others to criticize methodology and conclusions. Have you read the Bernstein piece, in particular where he says:
William Bernstein wrote:While noting that three-factor regression of their indexes had "exposure to the value factor and, to a lesser extent, the size factor," as well as an "estimated alpha of –0.1 percent" (presumably per year), [Arnott, Hsu, and Moore] softpedal the possibility that a large part of the excess return of their fundamental indexes came from exposure to the two "supplemental" Fama-French factors, while nodding implicitly to it by observing that other value indexes do even worse, with alphas in the –1.5% range. [emphasis added]
The impression that comes across (although he does not say it, of course) by the use of "softpedal" is that Arnott et al are aware of potential problems with their claims, yet are down-playing those problems. So, we have Bernstein doing the best that we can do in finance, subjecting claims to independent statistical testing, and finding them wanting. The ball is in Arnott's court- has he answered those criticisms directly, has he adduced new evidence, etc.?

A concern of mine is related to something you wrote upthread (very insightful too, if I may say :D ):
George$ upthread wrote:I'm wondering how the 0.65% MER gets divided. Any informed thoughts?

It seems to me that IF this RAFI index takes root and becomes "big" - Arnott will become one very wealthy man.
Even just 0.10% of $50 billion is $50 million a year, year after year after year.
I first decided to post in this thread precisely because I felt that there was quite a bit of enthusiasm for fundamental indexing (which I shared at first), without enough critical questioning of the "but how does it work exactly" kind. I suspect, meaning absolutely no disrespect for anyone (posters here or Mr. Arnott and colleagues) that there was a certain sense that this guy is a reigning genius, so if he says it works, it must. But I referred in my first post upthread to "dramatically effective marketing", and that was the feeling one got- fundamental indexing was (and perhaps still is) riding a tidal wave of publicity leading inexorably to product launches. If Mr. Arnott and colleagues do get rich as you wonder, good for them- in our system, they have succeeded. :D But is that outcome the same as advancing finance theory? Not necessarily. So it is important to get past the hype and look at the issues critically- and to his credit Arnott has invited such debate by putting his theory out in public in great detail for all to analyze. It is the objective debate, and not the marketing, that will lead to greater understanding of indexing methodologies.

As far as Canada is concerned, you summarized the situation quite well:
George$ wrote:Yes the Canadian market is different in many ways from the US but surely there are also common underlying issues that are the same. And if Arnott's logic has merit in the US, then it may also have the same merit here.
The question which is still up in the air, of course, is whether fundamental indexing even "has merit in the US" in fact, as opposed to theory. More importantly for us, AFAIK there has not been any serious testing of the theory in the Canadian context at all, so we are all equally in the dark here! :shock:
Post Reply