Equal weighting or market cap weighting?

[AustenNut]

I read this post (http://www.fool.com/investing/general/2013/01/06/the-most-important-investment-chart-of-the-last-de.aspx#lastVisibleParagraph) and wanted a Boglehead perspective on it. It's talking about the S&P 500 and discusses whether it is better for everything to be equally weighted (0.2% holding of each of the 500 companies) or if they should be weighted by market weight.

Now I don't have any S&P funds (I invest in funds that try to replicate the whole market), but I was curious about what more knowledgeable people than I think. And if you think one method is better than the other, would this also apply to other types of index funds?

[NYBoglehead]

Market Cap weighted, no doubt about it. The smaller companies do not make anywhere near the revenue that the larger companies do. You wouldn't want the smallest company in the S&P 500 to represent 0.2% of your holdings when it is not representing 0.2% of the earnings.

[madbrain]

Look at this 5-year chart of EWI (S&P 500 equal weighted) vs S&P500.

http://finance.yahoo.com/echarts?s=ewi# ... undefined; [link fixed --admin LadyGeek]

[LFKB]

Look at this 5-year chart of EWI (S&P 500 equal weighted) vs S&P500.

http://finance.yahoo.com/echarts?s=ewi# ... undefined; [link fixed --admin LadyGeek];

I don't think your link is working properly but if I run the chart you described it shows S&P 500 equal weight is up 25% over the last 5 years whereas S&P is up 3.5% in that same time frame. Is that what you were going for?

[Elbowman]

In "The Four Pillars Of Investing" Bernstein talks about Bogle's first index fund, where he was forced by his employer to make it equal weighted instead of cap weighted. The transaction costs required to maintain equal weighting were so high that the attempt was a disaster.

The index may look good, but can it be efficiently implemented?

[pascalwager]

EWI seems to be an Italy index fund.

[madbrain]

Look at this 5-year chart of EWI (S&P 500 equal weighted) vs S&P500.

http://finance.yahoo.com/echarts?s=ewi# ... undefined; [link fixed --admin LadyGeek]

I don't think your link is working properly but if I run the chart you described it shows S&P 500 equal weight is up 25% over the last 5 years whereas S&P is up 3.5% in that same time frame. Is that what you were going for?

If you copy/paste the whole link it will show one being down 60% and the other at 0%.
Anyway, I got the symbol wrong.

[madbrain]

EWI seems to be an Italy index fund.

LOL, thank you for pointing out my error. I am blaming it on giving up soda as part of my new year's resolutions.
And I should have looked at a total return chart, not price chart.

VADBX is one equally weighted S&P500 fund.

Looks like over the very long term, since 1987, the total return is about the same for both versions.

http://quote.morningstar.com/fund/chart ... %2C0%22%7D

[brianbooth]

Is the chart taking dividends into account? If it's just looking at stock prices, then it should be no surprise that the megacap companies that pay out a ton of cash have lagged.

[FinancialDave]

RSP Is actually the equally weighted index fund that I have had in one of my accounts for the last 5 years -- it really looks very much like the Vanguard small-cap ETF:

It's chart looks much better:

http://finance.yahoo.com/echarts?s=ewi# ... undefined; [link fixed --admin LadyGeek]


The yahoo site won't keep the compare, so you have to click on COMPARE tab and add the SP500, then re-draw the chart.

2year chart looks right on top of each other, but over 5 years the RSP seems to be up an extra 20%.


fd

[Call_Me_Op]

Equal weighting is likely to give you a greater return because it is in-effect tilted toward the smaller-caps.

[Rodc]

Equal weighting is likely to give you a greater return because it is in-effect tilted toward the smaller-caps.

Or rather a tilt towards mid-caps.

This can be done more cheaply by simply holding either a mid-cap market weight or by adding a small cap to S&P 500 (or better by adding small to TSM)

To the OP: you might try a search on the site. Many past threads on the topic.

[magician]

I'm surprised that nobody's yet mentioned a third common weighting approach: price-weighting.

The Dow-Jones Industrial Average and the Nikkei 225 are price-weighted indices.

The S&P 500, the NASDAQ, and the Wilshire 5000 are value-weighted (market-cap-weighted) indices.

The Value Line and the Financial Times Ordinary Share indices are equal-weighted (or unweighted).

Each has its advantages and disadvantages; each has its biases.

[Rodc]

I'm surprised that nobody's yet mentioned a third common weighting approach: price-weighting.

The Dow-Jones Industrial Average and the Nikkei 225 are price-weighted indices.

The S&P 500, the NASDAQ, and the Wilshire 5000 are value-weighted (market-cap-weighted) indices.

The Value Line and the Financial Times Ordinary Share indices are equal-weighted (or unweighted).

Each has its advantages and disadvantages; each has its biases.

Do any of the price weight indices get used to form mutual funds?

[Aptenodytes]

I'm surprised that nobody's yet mentioned a third common weighting approach: price-weighting.

The Dow-Jones Industrial Average and the Nikkei 225 are price-weighted indices.

The S&P 500, the NASDAQ, and the Wilshire 5000 are value-weighted (market-cap-weighted) indices.

The Value Line and the Financial Times Ordinary Share indices are equal-weighted (or unweighted).

Each has its advantages and disadvantages; each has its biases.

I knew the Dow was price weighted but I never thought there was an advantage. What is it?

[nisiprius]

First off, that article is garbage. His authority, Arnott, makes the plausible argument that it might make "more sense for indexes to weight companies by their economic footprint -- things like revenue, earnings, and dividends." That shouldn't be presented as support for an equal-weighted index, in which companies with small revenues, earnings and dividends are given equal weight with companies that have huge revenues, earnings, and dividends! It is an argument for a tweaked, adjusted, tuned, or modified version of a cap-weighted index--such as those which Mr. Arnott's firm provides

Market-cap weighting has a number of properties that make it special--not necessarily "best," but special. It is not an arbitrary choice.

1) Market-cap weighting is the weighting that cancels out the effect of speculative trading and leaves you with the returns generated by the stock-issuing businesses. If someone makes a huge speculative sale of Millepore stock and buys the same dollar value of, let's say, Zimmer stock, the price of Millespore stock moves down and the price of Zimmer stock moves up. Under appropriate assumptions, if you are holding a cap-weighted total stock market portfolio, this does not affect the dollar value of your holdings at all. However, anyone who is holding any other portfolio will either have more weight in Millespore or more weight in Zimmer, and the value of their holding will change.

2) Market-cap weighting incurs lower trading expenses for a mutual fund, because it automatically stays in balance. There is no need to buy or sell shares merely to match the index weighting.

3) In the words of Jeremy Siegel--let me note upfront that he says the stated assumptions are not true and that "fundamentally-weighted" indexes are better--but, nevertheless:

Capitalization-weighted indexes... under certain assumptions give investors the "best" tradeoff between risk and return. That means that for any given risk level, these capitalization-weighted portfolios give the highest returns, and for any given return, these portfolios give the lowest risk. This property is called mean-variance efficiency.

Equal weighting is one of an infinite number of non-market-cap weightings. It's not clear what special properties it's supposed to have. I cannot for the life of me see some obvious reason why I'd want to own the same dollar value of Zimmer stock and Exxon Mobil stock. It just seems goofy to me. Can anyone give any reason other than "it outperformed the cap-weighted S&P?"

Like many weightings, it overweights small-caps--obviously--so it does the things that overweighting small-caps does. Some people, perhaps many in this forum, think that's beneficial. But why equal weight, rather than, say, cap-weighted within the small-cap category? To the unaided eyeball, what, exactly, has RSP done that Vanguard Small-Cap Index fund didn't do more of? And that's the plain-Jane Vanguard fund the connoisseurs sneer at. The gourmet funds like DFA Small-Cap did even more.



With the non-market-cap weightings, one wants to understand what the theory is supposed to be. Otherwise, it just comes down to looking at past performance.

[Rick Ferri]

There is no free lunch. As others have pointed out, the reason S&P 500 equal weight (EW) has outperformed cap weight (CW) is because mid-cap stocks outperformed large cap stocks over the period being measured. That being said, if EW is your method for buying mid-cap stocks, then it's fine.

Rick Ferri

[stlutz]

Equal weighting is one of an infinite number of non-market-cap weightings. It's not clear what special properties it's supposed to have. I cannot for the life of me see some obvious reason why I'd want to own the same dollar value of Zimmer stock and Exxon Mobil stock. It just seems goofy to me. Can anyone give any reason other than "it outperformed the cap-weighted S&P?"

Just for the purposes of discussion...

Exxon Mobil and Zimmer would share many of the same risks in that their performance is subject to the good/poor decisions of management as well as to the changing marketplace. XOM just experiences those on a much larger scale. Why is Exxon's management so much better than Zimmer's that you should invest 40x as much money in it?

From a philosophical perspective, equal weight is a way to say that you're not making a bet on any one company. Furthermore you're not making any bets on factor mispricings--growth vs. value, large vs. small etc. You're just buying a basket of stocks and accepting the returns. Of course, for practicality purposes, you have to limit yourself to companies that are large enough to be tradeable/investable at a low cost, so you'll lose some purity there.

The first index fund actually was equal weighted. Market-cap weighting "won" out for reasons of practicality--back in the 70s the costs of maintaining equal-weighting were much higher than today. So, if the index fund was just invented in say, 2001, history could have worked out much differently.

All of that said, I don't own any equal-weighted funds nor do I intend to. As Rick and nisiprius have noted, if you want more Midcap exposure, you can do that more cheaply by adding a midcap fund than by switching to RSP.

[pascalwager]

X-Ray shows RSP: 51% large, 48% mid, and 1% small. So the near 10-year outperformance may/must be due to midcap outperformance.

[FinancialDave]

To the unaided eyeball, what, exactly, has RSP done that Vanguard Small-Cap Index fund didn't do more of? And that's the plain-Jane Vanguard fund the connoisseurs sneer at. The gourmet funds like DFA Small-Cap did even more.



With the non-market-cap weightings, one wants to understand what the theory is supposed to be. Otherwise, it just comes down to looking at past performance.

First I'm not sure comparing a small blend fund to a large blend fund is a valid reason to suggest that cap-weighting outperforms equal weighting -- seems to me the only true comparison is to compare two funds that have exactly the same stocks in them, only at a different weighting (RSP & SP500) -- if your question is "is equal weighting better than cap weighting" -- otherwise why don't I just compare say NAESX to VFINX and put all small cap stocks in my portfolio.

If you want to compare the metrics of NAESX & RSP, the trend chart doesn't tell the whole story. On a risk adjusted basis, using Morningstar 3 year Beta & 3 year return data, RSP has a risk adjusted return of 10.88, while NAESX has a risk adjusted return of 10.83 - hardly a difference to write home about.

While I am no expert on the theory of equally weighted funds, I seem to remember it this way;

No one knows which stocks are going to perform better, so why give the advantage to the larger-cap variety stocks, other than for the convenience of very little rebalance expense.

As you suggest it does give a small-cap slant to any equally weighted fund -- which is by the way NOT the same as comparing a small-cap fund to a large-cap equally weighted fund, since of course these two funds have COMPLETELY DIFFERENT stocks in them, and thus are comparing different slices of the market.

fd

[Jebediah]

I would think equal weight might have an advantage due to internal rebalancing bonus

[Barry Barnitz]

Hi:

Jack Treynor (one of the developers of CAPM theory) provides a portfolio theorist's examination of market indifferent valuation (which includes equal weighting) in his paper, Why Market-Valuation-Indifferent Indexing Works, Financial Analysts Journal, vol. 61, n°5, September/October 2005.

link to summary
link to paper

regards,

[umfundi]

Equal weighting is likely to give you a greater return because it is in-effect tilted toward the smaller-caps.

I have long thought that throwing darts at newspaper stock listings is, in effect, equal weighting. Which is one explanation why throwing darts might beat active management: The tilt is different.

Keith

[hafius500]

Hi:

Jack Treynor (one of the developers of CAPM theory) provides a portfolio theorist's examination of market indifferent valuation (which includes equal weighting) in his paper, Why Market-Valuation-Indifferent Indexing Works, Financial Analysts Journal, vol. 61, n°5, September/October 2005.

link to summary
link to paper

regards,

But the very same website (EDHEC) summarized academic papers that (mathematically) refuted Teynor's and RAFI's claims (Search for Perold or Kaplan). You don't need high-school maths to understand Perold's proof. Kaplan adds another proof that comes to the same conclusion. He adds conditions under which fundamental strategies might work.
Who would need an active manager if a simple trading strategy like equal-weighting let us outperform by underweighting overvalued securities? These strategies overweight small- and value stocks and such stocks can be overvalued.

Fundamental Indexing: Breakthrough or Old Idea in New Marketing Garb?
Authors: Michael Edesess Source: Advisor Perspectives, Date: August 2008

...its promoters have introduced this new form of index weighting as an alternative to capitalisation weighting, which was said to overweight overpriced stocks and underweight underpriced stocks. Several authors, including Arnott and Hsu (2008), Hsu (2006) and Treynor (2005), have attempted to mathematically prove this over-weighting/under-weighting argument. Other authors, including Kaplan (2008) and Perold (2007), have criticised these proofs,...The advocates of fundamental indexing claim that a market-cap-weighted portfolio will be overpriced on average, though some of the stocks in it may be overpriced and some underpriced. Using mathematical arguments, the author of the present article proves that these claims are false. He also proposes an intuitive explanation of his formal proof...As a result, the market price of stocks can be considered an unbiased estimate of their fair value. This conclusion is extended to a market-cap-weighted index portfolio which is the average market cap of all stocks. So, the author concludes that the average mispricing in a market-cap-weighted portfolio is expected to be zero. The formal proof is provided in the appendix of the article...

Fundamentally Flawed Indexing - Authors: André F. Perold, Source: Financial Analysts Journal, vol. 63, n°6, Date: November/December 2007

For these authors, market-cap weighting over-weights overvalued stocks and under-weights undervalued stocks... In this article Perold calls into question the assertion that market-cap weighting is an inferior strategy. He develops a formal model to support his theory, but also explains it using a simple example with only two stocks. He demonstrates that, assuming that the two stocks are randomly overvalued or undervalued, cap-weighted and equally-weighted portfolios based on these two stocks will have the same expected return....Indeed, stocks are randomly mispriced and it is not possible to know if a stock is overvalued or undervalued. As a result, cap-weighting does not systematically over-weight overvalued stocks and thus cause a drop in performance

Why Fundamental Indexation Might – or Might Not – Work, Authors: Paul D. Kaplan, Source: Financial Analysts Journal, vol. 64, n°1, Date: January/February 2008

He also notes that Hsu’s conclusion that fundamental-weighted indices lead to higher expected returns than capitalisation-weighted indices has no theoretical underpinning despite Hsu’s arguments to the contrary... He also notes that as there is no theoretical foundation for deciding whether the inequality he derives holds or not, there is no theoretical foundation for fundamental indexation. In addition, he points out that fundamental indices are value-biased Finally, Kaplan suggests using a combination of both weighting systems to capture the information contained in both these systems.indices...

Links to abstracts and papers: Perold, Kaplan

Fundamental Indexing: A Verbal Optical Illusion
Michael Edesess, Ph.D.,January 13, 2009

This argument depends on frequent repetition of the following statement or statements like it: “Capitalization-weighted indexes overweight overpriced stocks and underweight underpriced stocks.”...The claim is wrong, as shown by using fundamental indexers’ own assumptions. They assume that...But if the average pricing error of the whole stock market is zero then so is the average pricing error of a market portfolio

This article has a linked to the quoted article by Edesess..

Response to Rob Arnott’s Defense of Fundamental Indexing, Michael Edesess, PhD, Partner and Chief Investment Officer, Fair Advisors, February 10, 2009

They go much further by intimating that it is mathematically provable that a fundamentally-weighted index must outperform a capitalization-weighted index. These intimations of mathematical certainties have mostly been made verbally (though erroneous proofs using mathematical notation have been attempted). When the verbal
claims are analyzed, they do not stand up to scrutiny.
Does a cap-weighted index portfolio hold more in overpriced stocks?
One of the claims made by Arnott is that “If we have a cap-weighted portfolio, we know most of our money is in companies that are above fair value.”i
In the January 17, 2009, issue of Advisor Perspectives, I showed that we know no such thing, by presenting a simple counterexample...

[Call_Me_Op]

Equal weighting is likely to give you a greater return because it is in-effect tilted toward the smaller-caps.

Or rather a tilt towards mid-caps.

Please re-read my statement carefully. I said "tilted toward smaller caps."

[Rodc]

Equal weighting is likely to give you a greater return because it is in-effect tilted toward the smaller-caps.

Or rather a tilt towards mid-caps.

Please re-read my statement carefully. I said "tilted toward smaller caps."

I did read it carefully.

If one wants to be more specific it is a tilt towards mid cap because equal weight has no small caps. So I'm not saying you are wrong at all, in fact I am agreeing with you, just trying to be a little more precise.

[Rodc]

First I'm not sure comparing a small blend fund to a large blend fund is a valid reason to suggest that cap-weighting outperforms equal weighting -- seems to me the only true comparison is to compare two funds that have exactly the same stocks in them, only at a different weighting (RSP & SP500) -- if your question is "is equal weighting better than cap weighting" -- otherwise why don't I just compare say NAESX to VFINX and put all small cap stocks in my portfolio.

Fama-French show (to the extend anything in investing is ever proven) that performance is driven by taking on "risk" factors, namely Market risk, Value company risk and Small company risk. This is the basic three factor model (there are extensions as well).

So any two well diversified no cost portfolios are expected to have (very similar) the same basic return behavior if they have the same risk exposure even if they do not have the same stocks. So in practice you want the one with the lowest costs.

One can of course disagree with this theory, and many do. But this theory is behind the many posts regarding equal weight simply being a higher cost method of boosting the small/mid size company exposure of a portfolio. No need to have the precisely the same companies in the portfolio.

[Joe S.]

Equal weighting is likely to give you a greater return because it is in-effect tilted toward the smaller-caps.

Or rather a tilt towards mid-caps.

Please re-read my statement carefully. I said "tilted toward smaller caps."

I did read it carefully.

If one wants to be more specific it is a tilt towards mid cap because equal weight has no small caps. So I'm not saying you are wrong at all, in fact I am agreeing with you, just trying to be a little more precise.

As I love to quibble over nothing, I prefer

Equal weighting is likely to give you a greater return because it is in-effect tilted toward the smaller-caps.

This statement has the advantage that it can be generalized to many indexes, not just the S&P 500.

[Rodc]

Equal weighting is likely to give you a greater return because it is in-effect tilted toward the smaller-caps.

Or rather a tilt towards mid-caps.

Please re-read my statement carefully. I said "tilted toward smaller caps."

I did read it carefully.

If one wants to be more specific it is a tilt towards mid cap because equal weight has no small caps. So I'm not saying you are wrong at all, in fact I am agreeing with you, just trying to be a little more precise.

As I love to quibble over nothing, I prefer

Equal weighting is likely to give you a greater return because it is in-effect tilted toward the smaller-caps.

This statement has the advantage that it can be generalized to many indexes, not just the S&P 500.

Bold added.

Then you might like the full quote:

Or rather a tilt towards mid-caps.

This can be done more cheaply by simply holding either a mid-cap market weight or by adding a small cap to S&P 500 (or better by adding small to TSM)

The full quote also has the advantage that it shows I was agreeing with Call_Me_OP.

[Joe S.]

As I love to quibble over nothing..

Bold added.

Then you might like the full quote:

Or rather a tilt towards mid-caps.

This can be done more cheaply by simply holding either a mid-cap market weight or by adding a small cap to S&P 500 (or better by adding small to TSM)

The full quote also has the advantage that it shows I was agreeing with Call_Me_OP.

Next time I quibble I will research the whole thread more carefully.

[Aptenodytes]

Now that we know that mid caps are smaller than large caps, can someone fill me in on what benefits come from weighting an index by price, as the Dow Jones Industrial Average does?

[nisiprius]

Now that we know that mid caps are smaller than large caps, can someone fill me in on what benefits come from weighting an index by price, as the Dow Jones Industrial Average does?

If we go back to the original purpose of the Dow Jones Industrial Average, it was to be used in combination with the Dow Jones Railroad Average as part of some technical-analysis market-timing "Dow Theory" thing. It wasn't intended to be a measure of the total market, and it wasn't intended to be an investment portfolio.

If I had to guess, I'd guess that the benefit of weighting the index by price was ease of computation. Dow started calculating indexes in 1884, and the DJIA was introduced in 1896. Heaven knows how he did the calculation, but the Comptometer would have been new technology; it was your archetypic "adding machine."

WIkipedia says it "it could also do subtractions, multiplication and division." But it doesn't say how; from pictures of it, it certainly wasn't a one-button operation. I'm guessing you did it by repeatedly adding shifted versions of the number, i.e. to multiply 2718 by 714 perhaps you'd key in 2718, pull the lever four times, key in 27180, pull the lever once, key in 271800, pull the lever seven times.

The modern Monroe rotary calculator we are, I am sure, all familiar with , which did multiply and divide, wasn't to be invented until 1912.

In short, to try to retrofit a modern rationale onto the construction of the Dow Jones Industrial Average is like asking how a Model T Ford planetary gear system is better than an automatic transmission.

[Aptenodytes]

If I had to guess, I'd guess that the benefit of weighting the index by price was ease of computation. Dow started calculating indexes in 1884, and the DJIA was introduced in 1896.

Thanks -- that makes sense as a historical explanation. But in terms of what benefits price weighting has today, I haven't seen any yet. I will conclude that there are none, and the the Dow Jones Industrial Average has no purpose other than drawing attention to itself, which is why its continued dominance in the media is grating to me.

[Rodc]

If I had to guess, I'd guess that the benefit of weighting the index by price was ease of computation.

That is my understanding.

I don't know that there is anyone who suggests that price weighting should be used or is used in making portfolios (mutual funds, etc.) Seems like a red herring in this thread. But maybe someone somewhere uses price weighting. Must be fun when some company does a stock split...

Nisi, enjoyed the picture.

[umfundi]

... has no purpose other than drawing attention to itself, which is why its continued dominance in the media is grating to me.

I thought that was the Kardashians.

Keith

[Rodc]

If I had to guess, I'd guess that the benefit of weighting the index by price was ease of computation. Dow started calculating indexes in 1884, and the DJIA was introduced in 1896.

Thanks -- that makes sense as a historical explanation. But in terms of what benefits price weighting has today, I haven't seen any yet. I will conclude that there are none, and the the Dow Jones Industrial Average has no purpose other than drawing attention to itself, which is why its continued dominance in the media is grating to me.

Everyone has heard of it, and 99% of the population has no real idea of what it is so they don't know or care that it is a goofy measure.

That said, it does somehow correlate not too badly with the stock market as a whole. But personally I wish they focused on TSM, or at least S&P 500 (and included dividends).

[nisiprius]

Jack Treynor (one of the developers of CAPM theory) provides a portfolio theorist's examination of market indifferent valuation (which includes equal weighting) in his paper, Why Market-Valuation-Indifferent Indexing Works, Financial Analysts Journal, vol. 61, n°5, September/October 2005.

link to summary
link to paper

Objection. That paper concerns indexing approaches that differ from cap-weighting, and don't tie themselves to market valuation, but it does not make a case for equal weighting. In fact he sets it up only to knock it down. He calls it "an extreme example," uses the word "alas," complains about it having a small-cap bias, and spends the rest of the paper finding ways to remove that bias. I THINK the gist is to find something analogous to nonparametric statistics, using things like rank-ordering, to build portfolios that are actually not necessarily very different from cap-weighted portfolios... not sure.

[FinancialDave]

This thread is getting just totally off-base to what the OP had originally asked:
paraphrase -- "Is it better to Cap-Weight, or Equal Weight the SP500."

To study that the ONLY way to do it is to put exactly the same 500 stocks in two portfolios and run whatever analysis you want on it -- NOTHING else (IMHO) answers the question, at least for me.

In the end the question is really UNANSWERABLE, for any time in the future and TOTALLY ANSWERABLE for any historical period. The reason for this is that no one knows if the larger cap companies of the index are going to perform better than the smaller ones -- and in my opinion I have no reason to believe that company 500 on the list has any less of an advantage to turn a 10% profit next year than company #1 on the list --- this is the whole theory in a nutshell to say "if I don't know which of the companies will return more why don't I just weight them all equally!"

But sure, IF I knew for a fact that company #1 was more likely to turn a profit then I would overweight that company.

fd

[Random Musings]

I'd rather just small-value load.

Cheaper to do that the equal weight approach.

RM

[Aptenodytes]

This thread is getting just totally off-base to what the OP had originally asked:
paraphrase -- "Is it better to Cap-Weight, or Equal Weight the SP500."

To study that the ONLY way to do it is to put exactly the same 500 stocks in two portfolios and run whatever analysis you want on it -- NOTHING else (IMHO) answers the question, at least for me.

In the end the question is really UNANSWERABLE, for any time in the future and TOTALLY ANSWERABLE for any historical period. The reason for this is that no one knows if the larger cap companies of the index are going to perform better than the smaller ones -- and in my opinion I have no reason to believe that company 500 on the list has any less of an advantage to turn a 10% profit next year than company #1 on the list --- this is the whole theory in a nutshell to say "if I don't know which of the companies will return more why don't I just weight them all equally!"

But sure, IF I knew for a fact that company #1 was more likely to turn a profit then I would overweight that company.

fd

It got off base only after the question was perfectly well answered, so no harm done. Build your equity portfolio with index funds that are weighted by market cap. I think that consensus was pretty clear.

Will mid-caps always outperform large-caps? That's a different question, and my summary of the collective answer doesn't prejudge the answer. It just says, if you want to place a bet on mid-caps, buy a mid-cap index fund that is market-cap weighted. Buying an equal-weighted S&P500 fund would be an illogical approach to making that bet.

[Barry Barnitz]

HI:

Here are the five year annualized total returns for S&P indexes as of 1/7/2013:

S&P 500 TR 2.91%
S&P 500 EW TR 6.41%

S&P 400 TR 6.90%
S&P 400 EW TR 9.59%

S&P 600 TR 6.84%
S&P 600 EW TR 9.12%

LInks to the S&P index pages (which provided the returns data) are in the wiki :Equal Weighted Indices - Bogleheads

regards

[AustenNut]

OP here. I just wanted to thank everyone for sharing their insight. As a novice I will read ideas elsewhere that sound really interesting, but it's good to know there are helpful, knowledgeable people here who can provide greater depth and perspective on an issue. Many thanks for what you do.

[nisiprius]

Barry, as alway I appreciate your data and the Wiki article. From which I learn that the index was only launched in 2003. I don't know how many years of data it would take to make me feel that there's enough data to conclude anything, but five years or even eight ain't it. Anything can outperform anything for eight years.

It appears as if CRSP calculates an index

1000501 CRSP Equal-Weighted Index of the S&P 500 Universe

I wonder if it goes back to 1926 and how they handle the 1957 transition? Perhaps S&P 500 Universe means the 90-stock average before 1957. There's also

1000001 CRSP NYSE Equal-Weighted Market Index

Anyway... as far as I know it takes a fair amount of money to get access to those data sets, but I wonder if anyone has access to them or knows of some handy 80-year growth charts for these indices?

[nisiprius]

Equal weighting is one of an infinite number of non-market-cap weightings. It's not clear what special properties it's supposed to have. I cannot for the life of me see some obvious reason why I'd want to own the same dollar value of Zimmer stock and Exxon Mobil stock. It just seems goofy to me. Can anyone give any reason other than "it outperformed the cap-weighted S&P?"

Why is Exxon's management so much better than Zimmer's that you should invest 40x as much money in it?

I gave my rationales and none of them has anything to do with management. The big rationale is what I will call "inertia." This is my own theorizing, derived no doubt from things I read and half-understood. I keep waiting to get smacked down on this by someone who actually knows financial economics... but anyway.

Suppose that XOM is forty times as big as Zimmer, and suppose that you have a period of time during which absolutely nothing happens, there is no new fundamental information, and suppose some big trader makes a purely speculative move, sells a huge block of XOM and buys Zimmer. The price of XOM falls, the price of Zimmer rises. Since we assume there's no information out there, the rest of the market assumes that the total value of all extant shares of XOM and Zimmer haven't changed. This is a little tricky to word. Suppose we have two blocks of XOM and Zimmer stock that are both initially priced at $1,000,000 total. It seems to me that the effect of the trade will be that the price of the block of Zimmer will rise forty times as much as the price of the block of XOM falls.

If I don't want my portfolio to fluctuate in response to this purely speculative trade, I need to hold exactly forty times as many dollars' worth of XOM as I do of Zimmer.

If I hold them in any other ratio, then I am "siding" with one side or the other of the speculative trade. If I am overweighting XOM, my portfolio value falls. If I am overweighting Zimmer, it rises. If I am cap-weighting them, it stays the same.

Assuming that isn't totally bogus, that is a rationale for cap-weighting that singles it out as having a unique, and possibly desirable property. And it has nothing at all to do with any judgements about the skill of XOM or Zimmer's management.

What is the rationale, specifically, for equal weighting, that singles it out as having a unique and possibly desirable property? Under what set of assumptions is it an optimum weighting? Is there anything beyond "naïve diversification" and past performance?

What is a reason to suppose equal weighting would be better than ticker-symbol weighting, a method I have just invented in which we take the average value of the ASCII code values of the letters in the ticker symbol (A = 65, B = 66, ... Z = 90)?

[Barry Barnitz]

Hi Nisi:

The longest data sequence for an equal weighted index I have been able to find is Wilshire's equal weighted 5000 index. The Wilshire index calculator provides annual return data for the two indexes over the 1991-2012 period (Wilshire 5000 and Wilshire EW 5000 ). This data sequence is included in the appendix table in the wiki page.

regards,

[Rodc]

Suppose we have two blocks of XOM and Zimmer stock that are both initially priced at $1,000,000 total. It seems to me that the effect of the trade will be that the price of the block of Zimmer will rise forty times as much as the price of the block of XOM falls.

nisi, I'm lost as to why you think that would happen. BTW: I'm not an expert and not making any attempt to smack anyone.

As long as the buy was of modest size relative to the total capitalization or total outstanding shares of Zimmer why would there be much of any change to Zimmer's stock price? And if the buy were big enough to drive up Zimmer's price, who would make such a trade in the first place?

All S&P 500 companies are in the general scheme of things rather huge. I don't think any one sale is going to cause much of a price change.

I mean, this is not like equal weighting Exxon and a local restaurant chain with 6 (or even 100) outlets.

I understand this is just a thought experiment, but I don't understand it.

[Rodc]

Also, basic least squares math says that if you want to mean optimize a portfolio you weight based on mean, standard deviation and correlation. In real life you can't estimate those things worth a darn, but if you could that is what you would do.

Out of 500 large companies:

Do we really think that largest companies in general have higher mean returns?

Do we really think that largest companies in general have lower standard deviation?

Do we really think that largest companies in general are less correlated one to another?

And even if we do, do we think these things are higher and lower to the degree that suggest the very wide range of weights seen in a cap-weighted index fund?

Instead if I think one large company is likely to have about the same mean, the same standard deviation, the same correlation one to another, then you should equal weight. (If you think the reverse, then weight the smaller large companies higher than the largest).

This of course ignores costs. And costs are very important.

I do think that some of the special attributes like low cost to implement and low cap gains are very nice to have in a fund.

Like nisi's post, this is meant to be more of a thought experiment than practical. So, please check the "Standard deviation is not risk" posts at the door.

[steve r]

Hi Nisi:

The longest data sequence for an equal weighted index I have been able to find is Wilshire's equal weighted 5000 index. The Wilshire index calculator provides annual return data for the two indexes over the 1991-2012 period (Wilshire 5000 and Wilshire EW 5000 ). This data sequence is included in the appendix table in the wiki page.

regards,

Nisi - the Wishire data goes back to 1971. Average montly returns are much higher with EW. This of course is a heavy tilt in favor of small.
One dollar invested in the EW index in January 1971 is worth 866 today (1.5 percent per month average, NOT CAGR).

Playing with the data a bit more.

[Rodc]

Hi Nisi:

The longest data sequence for an equal weighted index I have been able to find is Wilshire's equal weighted 5000 index. The Wilshire index calculator provides annual return data for the two indexes over the 1991-2012 period (Wilshire 5000 and Wilshire EW 5000 ). This data sequence is included in the appendix table in the wiki page.

regards,

Nisi - the Wishire data goes back to 1971. Average montly returns are much higher with EW. This of course is a heavy tilt in favor of small.
One dollar invested in the EW index in January 1971 is worth 866 today (1.5 percent per month average, NOT CAGR).

Playing with the data a bit more.

Other than with a tiny fund this cannot be done in the real world: simply not enough micro cap shares available to purchase and costs would likely be prohibitive. So, other than as an academic study, I'm not sure EW5000 data are useful.

EW S&P 500 has the advantage that is can be done in the real world.

[Barry Barnitz]

Here are current EW index returns as of 1/9/2013. The Russell and S&P EW indexes are investable (Guggenheim ETFS). The Wilshire EW index is not currently investable. Also Wilshire only has data for this series as of 12/31/2012.

Code: Select all

                     3 yr   5 yr   10yr
Russell 1000      11.00%   3.40%   7.20%
Russell 1000 EW   12.90%   8.20%  12.60%
Russell 2000      12.40%   5.80%   9.70%
Russell 2000 EW   10.80%   5.70%  11.00%
         
S&P  500          10.79%    3.00%    n/a
S&P  500 EW       12.47%    5.00%    n/a
S&P  400          13.64%    7.36%    n/a
S&P  400 EW       13.62%   10.11%    n/a
S&P  600          14.24%    7.24%    n/a
S&P  600 EW      14.11%    9.60%     n/a
         
Wilshire 5000     11.15%   2.03%   7.85%
Wilshire 5000 EW   8.55%   5.52%  14.08%

LInks (which provided the returns data) are in the wiki :Equal Weighted Indices - Bogleheads. The page now has a more complete table (1971-2012) series of returns for the Wilshire 5000 EW index.

regards,