What's wrong with this critique of indexing?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
Lauretta
Posts: 615
Joined: Wed Jul 05, 2017 6:27 am

What's wrong with this critique of indexing?

Post by Lauretta » Sat Apr 28, 2018 10:48 am

Can you please point out at which point the following reasoning is wrong?
1) If one believes in the EMH and in common asset pricing models, then one accepts the idea that all information about firms in the stock market is baked into their current prices, so that for every stock you should expect the same risk adjusted return.
2) If this is the case, then my understanding is that the expected risk adjusted return for Apple should be the same as that for any other stock, say for News Corporation (if this weren't so their stock prices would instantly be adjusted - since markets are efficient - till that condition holds).
3) If both stocks (Apple and News Corporation) have all information baked into their price, and have the same risk adjusted returns, the rational thing for me to do is to make the same bet on each of them. The only reason why I should bet more on one stock than on the other, is if I had reason to believe that it would be more likely to outperform (and according to point 2) this is not true)
4) Yet by indexing, for every 100$ invested in the S&P500, I invest more than 3.5$ in Apple and less than one cent on News Corporation. How can this make sense? (i.e. what justifies this much bigger bet on Apple, just because it's bigger and more money is invested in it by other people?).
When everyone is thinking the same, no one is thinking at all

quantAndHold
Posts: 1701
Joined: Thu Sep 17, 2015 10:39 pm

Re: What's wrong with this critique of indexing?

Post by quantAndHold » Sat Apr 28, 2018 10:58 am

In a theoretical world, it would work as well to equal weight the stocks. And indeed, backtesting has shown that it would have worked somewhat better in the past to equal weight, because the very biggest companies tend to do worse than the rest over the long run. But, costs and taxes matter. Holding the portfolio at equal weight requires a lot of trading to keep the weighting equal. It also requires selling winners, which is tax inefficient. Cap weighting is buy and hold, which is more efficient, both in terms of trading costs and taxes.

User avatar
Lauretta
Posts: 615
Joined: Wed Jul 05, 2017 6:27 am

Re: What's wrong with this critique of indexing?

Post by Lauretta » Sat Apr 28, 2018 11:05 am

quantAndHold wrote:
Sat Apr 28, 2018 10:58 am
In a theoretical world, it would work as well to equal weight the stocks. And indeed, backtesting has shown that it would have worked somewhat better in the past to equal weight, because the very biggest companies tend to do worse than the rest over the long run. But, costs and taxes matter. Holding the portfolio at equal weight requires a lot of trading to keep the weighting equal. It also requires selling winners, which is tax inefficient. Cap weighting is buy and hold, which is more efficient, both in terms of trading costs and taxes.
yes I totally agree with you on the question of costs; I myself use index funds for individual countries partly for this reason, and partly because I didn't think much about it when I started investing (but I deviate from market cap weighting when it comes to allocating to different geographical regions).
But I was wondering whether there was anything wrong in the reasoning above: i.e. to me it's obvious that market cap weighting means having larger bets on bigger firms (or, as far as geographical regions are concerned, on the US if you allocate globally) and it doesn't seem to be consistent with the idea that each security (no matter how large or small) has all information baked into its price and so should have the same risk adjusted returns, so that I don't see any thoeretical reason for favouring larger stocks just because other investors put more money into them.
When everyone is thinking the same, no one is thinking at all

johan851
Posts: 92
Joined: Sun Apr 17, 2011 12:59 pm

Re: What's wrong with this critique of indexing?

Post by johan851 » Sat Apr 28, 2018 11:15 am

I've wondered about equal weighting as well. The buy and hold point above is a good one. Another reason would be minimizing market impact. If, say, 10% of the market is held through index funds, and it's all equal weight instead of cap weight, the result will be massive over-investment in a bunch of tiny companies and a resulting skew in their prices.

Cap weighting also produces some other effects. It's why holding an S&P 500 or large cap index is a reasonable substitute for a total market index.

User avatar
tadamsmar
Posts: 7627
Joined: Mon May 07, 2007 12:33 pm

Re: What's wrong with this critique of indexing?

Post by tadamsmar » Sat Apr 28, 2018 11:15 am

Lauretta wrote:
Sat Apr 28, 2018 10:48 am
Can you please point out at which point the following reasoning is wrong?
1) If one believes in the EMH and in common asset pricing models, then one accepts the idea that all information about firms in the stock market is baked into their current prices, so that for every stock you should expect the same risk adjusted return.
2) If this is the case, then my understanding is that the expected risk adjusted return for Apple should be the same as that for any other stock, say for News Corporation (if this weren't so their stock prices would instantly be adjusted - since markets are efficient - till that condition holds).
3) If both stocks (Apple and News Corporation) have all information baked into their price, and have the same risk adjusted returns, the rational thing for me to do is to make the same bet on each of them. The only reason why I should bet more on one stock than on the other, is if I had reason to believe that it would be more likely to outperform (and according to point 2) this is not true)
4) Yet by indexing, for every 100$ invested in the S&P500, I invest more than 3.5$ in Apple and less than one cent on News Corporation. How can this make sense? (i.e. what justifies this much bigger bet on Apple, just because it's bigger and more money is invested in it by other people?).
2) says all stocks are equal
3) If one is better, then we should buy more of that one. But since Apple and News Corp are equal, then we don't care about whether we own more of one or the other. There is no rational reason to own equal amounts. This assumes the the risk-adjusted return of stocks is the only consideration.

But risk-adjusted return is not the only consideration.

The reason to own multiple stocks is because stocks move somewhat independently with some covariance. So you get an overall lower risk-adjusted return for buying multiple stocks. The reason you would not own them in equal amounts has to do with the co-variance and variance of stocks.

According to the EMH, the index represents the rational ratio of stocks that we should buy.

Also, owning the index cuts transaction costs relative maintaining some other ratio of stock ownership.

Another issue is that you need to keep return high enough. If the return is too low then you would have to buy on margin to increase your return.
Last edited by tadamsmar on Sat Apr 28, 2018 11:31 am, edited 4 times in total.

User avatar
SimpleGift
Posts: 2936
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: What's wrong with this critique of indexing?

Post by SimpleGift » Sat Apr 28, 2018 11:23 am

Lauretta wrote:
Sat Apr 28, 2018 10:48 am
Yet by indexing, for every 100$ invested in the S&P500, I invest more than 3.5$ in Apple and less than one cent on News Corporation. How can this make sense? (i.e. what justifies this much bigger bet on Apple, just because it's bigger and more money is invested in it by other people?).
If one assumes that stock prices are rationally determined as the discounted present value of all future cash flows, the market has determined that Apple's future cash flows vastly outweigh those of News Corporation, and thus Apple's much higher relative price.

In this context, "making the same bet on each of them" means buying each of them at their relative cap-weights — not at equal weights.
Cordially, Todd

Fclevz
Posts: 351
Joined: Fri Mar 30, 2007 11:28 am

Re: What's wrong with this critique of indexing?

Post by Fclevz » Sat Apr 28, 2018 11:25 am

Lauretta wrote:
Sat Apr 28, 2018 10:48 am
4) Yet by indexing, for every 100$ invested in the S&P500, I invest more than 3.5$ in Apple and less than one cent on News Corporation. How can this make sense? (i.e. what justifies this much bigger bet on Apple, just because it's bigger and more money is invested in it by other people?).
Think of it as buying a proportional amount of earnings. Apple makes a gazillion times more money than News Corp, and so for that reason justifies a larger allocation.

User avatar
Lauretta
Posts: 615
Joined: Wed Jul 05, 2017 6:27 am

Re: What's wrong with this critique of indexing?

Post by Lauretta » Sat Apr 28, 2018 11:33 am

SimpleGift wrote:
Sat Apr 28, 2018 11:23 am
Lauretta wrote:
Sat Apr 28, 2018 10:48 am
Yet by indexing, for every 100$ invested in the S&P500, I invest more than 3.5$ in Apple and less than one cent on News Corporation. How can this make sense? (i.e. what justifies this much bigger bet on Apple, just because it's bigger and more money is invested in it by other people?).
If one assumes that stock prices are rationally determined as the discounted present value of all future cash flows, the market has determined that Apple's future cash flows vastly outweigh those of News Corporation, and thus Apple's much higher relative price.

In this context, "making the same bet on each of them" means buying each of them at their relative cap-weights — not at equal weights.
ok, I see your point, though intuitively it seems that a market cap weighted allocation is more vulnerable to the eventuality that something specific might go wrong to a large firm (think of BP (which made up a large proportion of the FTSE100) and Deepwater Horizon)
When everyone is thinking the same, no one is thinking at all

User avatar
Lauretta
Posts: 615
Joined: Wed Jul 05, 2017 6:27 am

Re: What's wrong with this critique of indexing?

Post by Lauretta » Sat Apr 28, 2018 11:43 am

tadamsmar wrote:
Sat Apr 28, 2018 11:15 am


The reason to own multiple stocks is because stocks move somewhat independently with some covariance. So you get an overall lower risk-adjusted return for buying multiple stocks. The reason you would not own them in equal amounts has to do with the co-variance and variance of stocks.

According to the EMH, the index represents the rational ratio of stocks that we should buy.
Thanks for your feedback. Does this mean that empirically it is found that the covariance between stocks is so adjusted that you actually get the highest risk adjusted returns for a portfolio consisting of a market cap weighted index? Intuitively it seems strange since with the current S&p500 you have a large proportion of large tech stocks which should have quite a high covariance. This would seem to be reduced in a portolio in which all stocks of the S&P500 are equally weighted(?)
When everyone is thinking the same, no one is thinking at all

User avatar
SimpleGift
Posts: 2936
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: What's wrong with this critique of indexing?

Post by SimpleGift » Sat Apr 28, 2018 11:55 am

Lauretta wrote:
Sat Apr 28, 2018 11:33 am
ok, I see your point, though intuitively it seems that a market cap weighted allocation is more vulnerable to the eventuality that something specific might go wrong to a large firm (think of BP (which made up a large proportion of the FTSE100) and Deepwater Horizon).
Honestly, the unsystematic risk (i.e., specific to a particular company) in most broad market indexes like Vanguard's Total U.S. Market, Total International, or Total World Stock Index doesn't cause me to lose any sleep:
However, for many of MSCI's individual country indexes, your point is well taken, since these are often much more concentrated in a few large companies. The solution: Don't invest in poorly-diversified individual country indexes — especially if broader market index funds or ETFs are readily available.
Last edited by SimpleGift on Sat Apr 28, 2018 11:57 am, edited 1 time in total.
Cordially, Todd

longinvest
Posts: 2792
Joined: Sat Aug 11, 2012 8:44 am

Re: What's wrong with this critique of indexing?

Post by longinvest » Sat Apr 28, 2018 11:56 am

Lauretta,
Lauretta wrote:
Sat Apr 28, 2018 10:48 am
3) If both stocks (Apple and News Corporation) have all information baked into their price, and have the same risk adjusted returns, the rational thing for me to do is to make the same bet on each of them. The only reason why I should bet more on one stock than on the other, is if I had reason to believe that it would be more likely to outperform (and according to point 2) this is not true)
As of 03/31/2018, Vanguard Total Stock Market Index Fund (VTSMX) had (among others) the following two holdings:

Code: Select all

Holding Shares Market Value
Apple Inc.               112,662,178   $18,902,460,225
Urban One Inc. Class A           950            $1,900
In total, VTSMX held 3637 different holdings totaling $672.4 billion.

If VTSMX wanted to invest as much into Urban One Inc. Class A than into Apple Inc., it would have to invest ($672.4 billion / 3637) = $184,877,646 into each of these two holdings. That would be ($184,877,646 / $1,900 X 950) = 92,438,823 shares of Urban One Inc. Class A.

Using google, I found this. According to Urban One, Inc. Class A Common Stock 10-Q Aug. 8, 2017 1:48 PM | Seeking Alpha, there were 1,663,632 outstanding class A shares of Urban One Inc. on July 31, 2017.

I'm too lazy to find a more recent number, but the arithmetic is clear: VTSMX would need to own 55 times the total number Urban One Inc. Class A outstanding shares. That's impossible.

Even if VTSMX was restricted to owning no more than 100% of the outstanding shares, I don't think that I need to explain how that would be much riskier than owning (950 / 1,663,632) = 0.0571% (that's less than 1%) of outstanding shares.
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VLB/ZRR

User avatar
Lauretta
Posts: 615
Joined: Wed Jul 05, 2017 6:27 am

Re: What's wrong with this critique of indexing?

Post by Lauretta » Sat Apr 28, 2018 12:01 pm

SimpleGift wrote:
Sat Apr 28, 2018 11:55 am
Lauretta wrote:
Sat Apr 28, 2018 11:33 am
ok, I see your point, though intuitively it seems that a market cap weighted allocation is more vulnerable to the eventuality that something specific might go wrong to a large firm (think of BP (which made up a large proportion of the FTSE100) and Deepwater Horizon).
Honestly, the unsystematic risk (i.e., specific to a particular company) in most broad market indexes like Vanguard's Total U.S. Market, Total International, or Total World Stock Index doesn't cause me to lose any sleep:
However, for many of MSCI's individual country indexes, your point is well taken, since these are often much more concentrated in a few large companies. The solution: Don't invest in poorly-diversified individual country indexes — especially if broader market index funds or ETFs are readily available.
Thanks; good point, I agree :happy
When everyone is thinking the same, no one is thinking at all

golfCaddy
Posts: 400
Joined: Wed Jan 10, 2018 10:02 pm

Re: What's wrong with this critique of indexing?

Post by golfCaddy » Sat Apr 28, 2018 12:02 pm

The EMH doesn't state that all stocks have the same risk-adjusted returns. The EMH states an optimal risk-adjusted portfolio is to own all stocks in proportion to their market cap. An efficient market bakes into the price not just the expected returns and standard deviation of an individual stock, but the correlation of that stock with all the other potential stocks in a portfolio. The essence of MPT is you can't look at an investment asset in isolation.

User avatar
neurosphere
Posts: 2822
Joined: Sun Jan 17, 2010 1:55 pm

Re: What's wrong with this critique of indexing?

Post by neurosphere » Sat Apr 28, 2018 12:08 pm

longinvest wrote:
Sat Apr 28, 2018 11:56 am
I'm too lazy to find a more recent number, but the arithmetic is clear: VTSMX would need to own 55 times the total number Urban One Inc. Class A outstanding shares. That's impossible.

Even if VTSMX was restricted to owning no more than 100% of the outstanding shares, I don't think that I need to explain how that would be much riskier than owning (950 / 1,663,632) = 0.0571% (that's less than 1%) of outstanding shares.
I was about to post a similar example, but you beat me to it. But this was going to be my example (although longinvest's is much better and 'rational')...

To take the argument further, reductio ad absurdum, let's just suppose we weren't limited to publically traded stocks. And we could easily invest in every american business. An equal weighted American Business Index, if you will, maybe with 500,000 business. Now imagine I just started my hot dog vending business, just me and a street cart. Instantaneously there would be as much money invested in my hot dog business as the largest US company. And that's not logical. But even so, imagine on a certain day I report double profits from the prior period, and my stock doubles. Now, every mutual fund following my American Business Index will need to sell shares of my hot dog company and buy some shares in every other company in the index, in order to keep equal weights. Imagine the transaction costs. :D
If you have to ask "Is a Target Date fund right for me?", the answer is "Yes".

User avatar
Lauretta
Posts: 615
Joined: Wed Jul 05, 2017 6:27 am

Re: What's wrong with this critique of indexing?

Post by Lauretta » Sat Apr 28, 2018 12:15 pm

golfCaddy wrote:
Sat Apr 28, 2018 12:02 pm
The EMH doesn't state that all stocks have the same risk-adjusted returns.
my understanding is that CAPM allows the price of each individual stock to be determined using its beta, so that for each security you have a linear relationship between beta and expected return (with the y intercept on the return vs risk graph being the risk free rate).
When everyone is thinking the same, no one is thinking at all

pastel
Posts: 11
Joined: Mon Jan 08, 2018 7:48 pm

Re: What's wrong with this critique of indexing?

Post by pastel » Sat Apr 28, 2018 12:17 pm

Let me frame this differently - the goal of an index fund is to match its index. So, an S&P 500 index fund seeks to match the returns of the S&P 500. To do this, the fund must weight its holdings by market cap. This is because each stock's contribution to the overall index is based on its market cap, or the percentage of the index it represents. Equal weights would no longer match the overall index because the proportions would be wrong.

It's important to note that risk adjusted returns are not the same as absolute returns. Risk adjusted returns are 'how much return should I get for each unit of risk.' While two companies may have the same 'risk adjusted returns,' they have can have two different levels of risk and thus have different absolute returns.

User avatar
Lauretta
Posts: 615
Joined: Wed Jul 05, 2017 6:27 am

Re: What's wrong with this critique of indexing?

Post by Lauretta » Sat Apr 28, 2018 12:22 pm

longinvest wrote:
Sat Apr 28, 2018 11:56 am
Lauretta,
Lauretta wrote:
Sat Apr 28, 2018 10:48 am
3) If both stocks (Apple and News Corporation) have all information baked into their price, and have the same risk adjusted returns, the rational thing for me to do is to make the same bet on each of them. The only reason why I should bet more on one stock than on the other, is if I had reason to believe that it would be more likely to outperform (and according to point 2) this is not true)
As of 03/31/2018, Vanguard Total Stock Market Index Fund (VTSMX) had (among others) the following two holdings:

Code: Select all

Holding Shares Market Value
Apple Inc.               112,662,178   $18,902,460,225
Urban One Inc. Class A           950            $1,900
In total, VTSMX held 3637 different holdings totaling $672.4 billion.

If VTSMX wanted to invest as much into Urban One Inc. Class A than into Apple Inc., it would have to invest ($672.4 billion / 3637) = $184,877,646 into each of these two holdings. That would be ($184,877,646 / $1,900 X 950) = 92,438,823 shares of Urban One Inc. Class A.

Using google, I found this. According to Urban One, Inc. Class A Common Stock 10-Q Aug. 8, 2017 1:48 PM | Seeking Alpha, there were 1,663,632 outstanding class A shares of Urban One Inc. on July 31, 2017.

I'm too lazy to find a more recent number, but the arithmetic is clear: VTSMX would need to own 55 times the total number Urban One Inc. Class A outstanding shares. That's impossible.

Even if VTSMX was restricted to owning no more than 100% of the outstanding shares, I don't think that I need to explain how that would be much riskier than owning (950 / 1,663,632) = 0.0571% (that's less than 1%) of outstanding shares.
thanks for the feedback. Yes of course I understand that for reason of scalability equal weighting is not feasible on a large scale; that's why I was framing the question from the point of view of an individual small investors - on a large scale it cannot be realised.
However, from a theoretical point of view I am not sure that I understand why owning 100% of outstanding shares of Urban One Inc is in itself more dangereous than holding a small fraction of outstanding shares of Apple for the same amount of money.
I mean, if Urban One goes bust you loose all your money, but then if Apple goes bust you also lose all your money, even though you were holding only a small fraction of Apple.
When everyone is thinking the same, no one is thinking at all

ThrustVectoring
Posts: 378
Joined: Wed Jul 12, 2017 2:51 pm

Re: What's wrong with this critique of indexing?

Post by ThrustVectoring » Sat Apr 28, 2018 12:26 pm

Let's use a toy model to illustrate why cap-weighted indexing is a good choice.

Suppose the widget industry has two companies in it: company A is worth $8B, and company B worth $2B. Suppose that there's a big federal widget contract, worth $2B in market cap, and there's a 50/50 chance each company gets the contract (and the $1B in market cap is priced in already). Furthermore for simplicity, suppose each company has 100M shares, so share price is $80 for A and $20 for B

If company A wins the contract, company A becomes worth $90 and company B worth $10. If the other company wins the contract, A becomes worth $70 and B becomes worth $30.

With a cap weighting, you buy equal numbers of shares of A and B, and whoever wins the contract, you wind up with $100 per share. More generally, cap-weighting is indifferent to the outcomes of zero-sum competition between firms.

With equal weighting, you'd have 4 shares of B for every share of A. If A wins the contract, you gain $10 from A and lose $40 from B, for a net of $30. Likewise if B wins the contract, you gain $40 from B and lose $10 from A, gaining a net of $30.

Equal-weight indexes are not indifferent to the result of competition. They win out when smaller companies usurp established players, and lose out when big companies outperform and solidify their position even further.

User avatar
nisiprius
Advisory Board
Posts: 35735
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: What's wrong with this critique of indexing?

Post by nisiprius » Sat Apr 28, 2018 12:27 pm

(ThrustVectoring beat me to it while I was composing a psot. Here's the same point made in a different way).

Let's make a set of wildly unrealistic simplifying assumptions. Suppose there are only two stocks in the stock market, Apple and News Corporation. Let's say their market capitalizations are $1000 billion and $10 billion. Let's suppose that we are talking about such a short period of time, or anyway just talking, about a closed system: a situation in which no money enters or leaves the market--anyone who sells a stock immediately buys the other stock. So, for purposes of discussion, there is $1,010 billion in the market.

A closed system also implies no dividends--the only transactions that are occurring are speculative.

Let's say Lauretta has $10,000 each invested in Apple and News Corporation, total $20,000.
Let's say I have $19,802 in Apple and $198 in News Corporation, total $20,000.

Let's say a big institutional investor sells $5 billion worth of News Corporation. Since it's a closed system, he must buy $5 billion worth of Apple. The only way this can happen is if prices adjust in such a way that the market cap is now $1,005 billion for Apple and $5 billon for News Corporation, total $1,010 as before.

So, the price of Apple increases by +0.5% while the price of News Corporation drops by -50%.

So, Lauretta's holdings now have a market value of $10,050 + $5,000 = $15,050.

My holdings now have a market value of:
--for Apple, $19,802 + 0.5% = $19,901;
--for News Corporation, $198 - 50% = $99; total $19,901 + $99 = $20,000.

The point is not that I had more. That's a result of the story I chose to tell, naturally I made it come out that way.

It wold have gone the other way. If the investor had sold Apple. Let's say, $10 billion worth of Apple. Apple would have gone down 1%, News Corporation would have doubled.

Lauretta's total would be $9,900 + $20,000 = $29,900 and mine would be $19,604 + $396 = $20,000.

Here's the point. I was a neutral, disinterested party. My total did not change regardless of which way the transaction went.

Lauretta was taking sides. By equal-weighting the two, she was implicitly taking the News Corporation side. She was on the side of any speculator buying News Corporation, and against any specular selling Apple. She lost of a speculator sold News Corporation, won if the speculator bought it.

I was not taking sides. I was equally on both sides of the transaction, measured in dollars. I did not need to think about speculators; they had no effect on me, speculation cancelled out..

Despite the "equal weight," Lauretta was effectively concentrated in News Corporation.

Although I held 100 times as much Apple stock as News Corporation, in dollar value, I was not concentrated in Apple. It did not help me in the first example, when the big institutional investor bought Apple. It did not hurt me when the investor sold Apple.

In short, I was more diversified, Lauretta was more concentrated.
Last edited by nisiprius on Sat Apr 28, 2018 12:39 pm, edited 1 time in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

User avatar
whodidntante
Posts: 3327
Joined: Thu Jan 21, 2016 11:11 pm

Re: What's wrong with this critique of indexing?

Post by whodidntante » Sat Apr 28, 2018 12:34 pm

Meb Faber said that just about anything is better than cap weighting in one of his podcasts. If you add some sensible qualifiers like you keep all costs low, including taxes and trading costs, and you remain broadly diversified, that might be correct. Slavishly equal weighting probably isn't ideal due to friction from internal trading, though maybe passable in ETF form due to the tax advantages. I'm using quantitative weighting for half of my portfolio so I guess we will see which half outperforms over my lifetime.

golfCaddy
Posts: 400
Joined: Wed Jan 10, 2018 10:02 pm

Re: What's wrong with this critique of indexing?

Post by golfCaddy » Sat Apr 28, 2018 12:44 pm

Lauretta wrote:
Sat Apr 28, 2018 12:15 pm
golfCaddy wrote:
Sat Apr 28, 2018 12:02 pm
The EMH doesn't state that all stocks have the same risk-adjusted returns.
my understanding is that CAPM allows the price of each individual stock to be determined using its beta, so that for each security you have a linear relationship between beta and expected return (with the y intercept on the return vs risk graph being the risk free rate).
Don't equate the EMH with CAPM. Fama has a five factor model.

User avatar
nisiprius
Advisory Board
Posts: 35735
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: What's wrong with this critique of indexing?

Post by nisiprius » Sat Apr 28, 2018 12:46 pm

whodidntante wrote:
Sat Apr 28, 2018 12:34 pm
Meb Faber said that just about anything is better than cap weighting in one of his podcasts. If you add some sensible qualifiers like you keep all costs low, including taxes and trading costs, and you remain broadly diversified, that might be correct.
No, it can't possibly be correct. Not without a lot of rhetorical mischief and mental reservations about what, exactly, is meant.

The market itself is cap-weighted.

If we ask all the cap-weighted total market investors to take one step forward, we are left with all of the investors following other strategies.

That group of investors, as a group, is cap-weighted, too. If some of them are outperforming the market, the rest of them must be underperforming the market. Therefore it cannot be true that "just about anything" is better than cap-weighting, unless you play some clever rhetorical games. I mean, you could go the circular reasoning route and say that, by definition, if some strategy underperformed cap weighting it must not have been sensible and shouldn't count.

You have a two hundred quarts of soup with two thousand dumplings in them. Cap-weighted investors go in and serve themselves a hundred quarts of soup, evenly stirred. The composition of their bowl is a perfectly match in composition, by percentage, with the composition of the original pot of soup. The remaining people are left with a hundred quarts of soup and a thousand dumplings. They all go in with their specially-shaped ladles and their skillful, sneaky stirring technique. They all take out a quart of soup, some ladling near the bottom, some near the top. It cannot possibly be true that "just about anything" is better than mixing the soup thoroughly and ladling out a bowl that's just the same as the "total soup." It cannot possible be true that "just about" everyone else manages to get more than ten dumplings.
Last edited by nisiprius on Sat Apr 28, 2018 12:56 pm, edited 4 times in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

gmaynardkrebs
Posts: 412
Joined: Sun Feb 10, 2008 11:48 am

Re: What's wrong with this critique of indexing?

Post by gmaynardkrebs » Sat Apr 28, 2018 12:48 pm

golfCaddy wrote:
Sat Apr 28, 2018 12:02 pm
The EMH doesn't state that all stocks have the same risk-adjusted returns. The EMH states an optimal risk-adjusted portfolio is to own all stocks in proportion to their market cap. An efficient market bakes into the price not just the expected returns and standard deviation of an individual stock, but the correlation of that stock with all the other potential stocks in a portfolio. The essence of MPT is you can't look at an investment asset in isolation.
A lot to unpack here, but I'm stumbling on the first sentence. Instead of stocks, for simplicity's sake take two bonds that are exactly equal except that one, with zero default risk is certain to return $100 at maturity, and the other, with a known default risk of 50% is certain to return $50. Why would the investor care whether she buys 1 of the safe bond or two of the risky bond at 1/2 the price, which is how the EMH would price them? Risk is risk, and $100 is $100. It seems to me that stocks are simply a more complex variant of that analysis, and the theory of the EMH says that the market will take all of these complexities into account in the same manner as the two bonds.

User avatar
Lauretta
Posts: 615
Joined: Wed Jul 05, 2017 6:27 am

Re: What's wrong with this critique of indexing?

Post by Lauretta » Sat Apr 28, 2018 12:54 pm

nisiprius wrote:
Sat Apr 28, 2018 12:27 pm
(ThrustVectoring beat me to it while I was composing a psot. Here's the same point made in a different way).

Let's make a set of wildly unrealistic simplifying assumptions. Suppose there are only two stocks in the stock market, Apple and News Corporation. Let's say their market capitalizations are $1000 billion and $10 billion. Let's suppose that we are talking about such a short period of time, or anyway just talking, about a closed system: a situation in which no money enters or leaves the market--anyone who sells a stock immediately buys the other stock. So, for purposes of discussion, there is $1,010 billion in the market.

A closed system also implies no dividends--the only transactions that are occurring are speculative.

Let's say Lauretta has $10,000 each invested in Apple and News Corporation, total $20,000.
Let's say I have $19,802 in Apple and $198 in News Corporation, total $20,000.

Let's say a big institutional investor sells $5 billion worth of News Corporation. Since it's a closed system, he must buy $5 billion worth of Apple. The only way this can happen is if prices adjust in such a way that the market cap is now $1,005 billion for Apple and $5 billon for News Corporation, total $1,010 as before.

So, the price of Apple increases by +0.5% while the price of News Corporation drops by -50%.

So, Lauretta's holdings now have a market value of $10,050 + $5,000 = $15,050.

My holdings now have a market value of:
--for Apple, $19,802 + 0.5% = $19,901;
--for News Corporation, $198 - 50% = $99; total $19,901 + $99 = $20,000.

The point is not that I had more. That's a result of the story I chose to tell, naturally I made it come out that way.

It wold have gone the other way. If the investor had sold Apple. Let's say, $10 billion worth of Apple. Apple would have gone down 1%, News Corporation would have doubled.

Lauretta's total would be $9,900 + $20,000 = $29,900 and mine would be $19,604 + $396 = $20,000.

Here's the point. I was a neutral, disinterested party. My total did not change regardless of which way the transaction went.

Lauretta was taking sides. By equal-weighting the two, she was implicitly taking the News Corporation side. She was on the side of any speculator buying News Corporation, and against any specular selling Apple. She lost of a speculator sold News Corporation, won if the speculator bought it.

I was not taking sides. I was equally on both sides of the transaction, measured in dollars. I did not need to think about speculators; they had no effect on me, speculation cancelled out..

Despite the "equal weight," Lauretta was effectively concentrated in News Corporation.

Although I held 100 times as much Apple stock as News Corporation, in dollar value, I was not concentrated in Apple. It did not help me in the first example, when the big institutional investor bought Apple. It did not hurt me when the investor sold Apple.

In short, I was more diversified, Lauretta was more concentrated.
Thanks for taking the time to give a detailed example! :happy I am not sure I fully understand though, since if the big institutional investor in your example sells 5 billion $ worth of News Corporation, there must be someone who has bought it; therefore its market cap cannot go down from 10 to 5 billion after the sale (this would only occur if the investor had sold it for for a lower price (in this example according to the maths it would be at half price) so that 5 billion $ would be wiped out from the market cap of News Corporation, which is impossible because you then go on to say that he invests the 5 billions he earned from the transaction into Apple). So if someone else bought 5 billion $ of News Corp as you have implied (perhaps by selling Apple in advance :wink: ), News Corp is still worth 10 billion $.
I understand however the point that smaller firms are more vulnerable to price impact in big transactions.
When everyone is thinking the same, no one is thinking at all

pastel
Posts: 11
Joined: Mon Jan 08, 2018 7:48 pm

Re: What's wrong with this critique of indexing?

Post by pastel » Sat Apr 28, 2018 1:02 pm

gmaynardkrebs wrote:
Sat Apr 28, 2018 12:48 pm
golfCaddy wrote:
Sat Apr 28, 2018 12:02 pm
The EMH doesn't state that all stocks have the same risk-adjusted returns. The EMH states an optimal risk-adjusted portfolio is to own all stocks in proportion to their market cap. An efficient market bakes into the price not just the expected returns and standard deviation of an individual stock, but the correlation of that stock with all the other potential stocks in a portfolio. The essence of MPT is you can't look at an investment asset in isolation.
A lot to unpack here, but I'm stumbling on the first sentence. Instead of stocks, for simplicity's sake take two bonds that are exactly equal except that one, with zero default risk is certain to return $100 at maturity, and the other, with a known default risk of 50% is certain to return $50. Why would the investor care whether she buys 1 of the safe bond or two of the risky bond at 1/2 the price, which is how the EMH would price them? Risk is risk, and $100 is $100. It seems to me that stocks are simply a more complex variant of that analysis, and the theory of the EMH says that the market will take all of these complexities into account in the same manner as the two bonds.
If a bond has a default risk of 50%, then it cannot be said that a return of $50 is certain. Rather, the expected return is $50. These two bonds are different because one is guaranteed $100, while the second cannot be certain. You are compensated for this uncertainty (risk)

User avatar
nisiprius
Advisory Board
Posts: 35735
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: What's wrong with this critique of indexing?

Post by nisiprius » Sat Apr 28, 2018 1:03 pm

Lauretta wrote:
Sat Apr 28, 2018 12:54 pm
...if the big institutional investor in your example sells 5 billion $ worth of News Corporation, there must be someone who has bought it; therefore its market cap cannot go down from 10 to 5 billion after the sale...
It occurs incrementally. The big institutional investor wants to sell $5 billion worth of News Corporation. He starts by trying to sell 1,000,000 shares at, say $20/share. He finds no buyers. A potential buyer says "I won't pay $20/share, how about $19." The seller says "OK." At this point, the last transaction was at $19/share, down from $20, and News Corporation has lost 5% of its value... evaporated, vanished, poof, gone to the Valhalla of destroyed value. Valuehalla? The total market cap of News Corporation has gone down from $10 billon to $9.5 billion because two people in an auction market changed their mind about what it was worth.
Last edited by nisiprius on Sat Apr 28, 2018 1:18 pm, edited 2 times in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

golfCaddy
Posts: 400
Joined: Wed Jan 10, 2018 10:02 pm

Re: What's wrong with this critique of indexing?

Post by golfCaddy » Sat Apr 28, 2018 1:08 pm

Lauretta wrote:
Sat Apr 28, 2018 12:15 pm
golfCaddy wrote:
Sat Apr 28, 2018 12:02 pm
The EMH doesn't state that all stocks have the same risk-adjusted returns.
my understanding is that CAPM allows the price of each individual stock to be determined using its beta, so that for each security you have a linear relationship between beta and expected return (with the y intercept on the return vs risk graph being the risk free rate).
Let's unpack this a little more. Beta is not the risk of an individual stock. It's the systematic risk of an individual stock. Even within the context of CAPM, all stocks don't have the same risk-adjusted returns. They have the same systematic risk-adjusted returns. Equal weighting is only the optimal strategy if you construct a portfolio with no unsystematic risk.

User avatar
nisiprius
Advisory Board
Posts: 35735
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: What's wrong with this critique of indexing?

Post by nisiprius » Sat Apr 28, 2018 1:16 pm

P.S. Lauretta, just for fun... somewhat tangential to your topic, but worth the effort... take a little while to look at the October, 2017 annual report, page 7, schedule of investments for the PowerShares Russell 2000 Equal Weight Portfolio (EQWS). It is holding a total of $23,137,392 worth of stock, so with about 2,000 constituents--they say 1,931--it ought to be holding... $23,137,392/1,931 = oh, roughly $12,000 worth of each.

Take a look.

Of course, the index it tracks might not be defined exactly the way you might expect.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

jalbert
Posts: 2905
Joined: Fri Apr 10, 2015 12:29 am

Re: What's wrong with this critique of indexing?

Post by jalbert » Sat Apr 28, 2018 1:20 pm

If EMH is true and Apple and News Corp have the same expected risk-adjusted return, it does not mean they have the same risk. Tilting away from market cap weighting to the equal weighting is a tilt on size. The EMH would imply this is taking more risk to get a higher expected return. If you don't wish to take more risk you could avoid the size tilt or reduce the percentage of equities and increase one's cash position to balance risk between these two choices.

When you compare a market cap weighted portfolio to an equal weighted portfolio diluted with cash so that the risk is leavened, the market cap portfolio is much cheaper to implement because there is no need to rebalance. The point of the index fund then is that if expected risk-adjusted return is the same, go with the portfolio that is cheaper to implement and adjust risk to taste by adjusting percentage of equities.
Risk is not a guarantor of return.

User avatar
Lauretta
Posts: 615
Joined: Wed Jul 05, 2017 6:27 am

Re: What's wrong with this critique of indexing?

Post by Lauretta » Sat Apr 28, 2018 1:21 pm

nisiprius wrote:
Sat Apr 28, 2018 1:03 pm
A potential buyer says "I won't pay $20/share, how about $19." The seller says "OK." At this point, the last transaction was at $19/share, down from $20, and News Corporation has lost 5% of its value... evaporated, vanished, poof, gone to the Valhalla of destroyed value. Valuehalla?
Right, so the Valkyrie (Valuekyrie?) seems to be the risk of price impact of big transactions, which for smaller companies is larger than for big firms, since the price of the latter should be less affected by high trading volumes.
When everyone is thinking the same, no one is thinking at all

jalbert
Posts: 2905
Joined: Fri Apr 10, 2015 12:29 am

Re: What's wrong with this critique of indexing?

Post by jalbert » Sat Apr 28, 2018 1:25 pm

nisiprius wrote:
Sat Apr 28, 2018 1:16 pm
P.S. Lauretta, just for fun... somewhat tangential to your topic, but worth the effort... take a little while to look at the October, 2017 annual report, page 7, schedule of investments for the PowerShares Russell 2000 Equal Weight Portfolio (EQWS). It is holding a total of $23,137,392 worth of stock, so with about 2,000 constituents--they say 1,931--it ought to be holding... $23,137,392/1,931 = oh, roughly $12,000 worth of each.

Take a look.

Of course, the index it tracks might not be defined exactly the way you might expect.
I suspect they have wide rebalancing bands to reduce transaction costs.
Risk is not a guarantor of return.

gmaynardkrebs
Posts: 412
Joined: Sun Feb 10, 2008 11:48 am

Re: What's wrong with this critique of indexing?

Post by gmaynardkrebs » Sat Apr 28, 2018 2:09 pm

pastel wrote:
Sat Apr 28, 2018 1:02 pm
gmaynardkrebs wrote:
Sat Apr 28, 2018 12:48 pm
golfCaddy wrote:
Sat Apr 28, 2018 12:02 pm
The EMH doesn't state that all stocks have the same risk-adjusted returns. The EMH states an optimal risk-adjusted portfolio is to own all stocks in proportion to their market cap. An efficient market bakes into the price not just the expected returns and standard deviation of an individual stock, but the correlation of that stock with all the other potential stocks in a portfolio. The essence of MPT is you can't look at an investment asset in isolation.
A lot to unpack here, but I'm stumbling on the first sentence. Instead of stocks, for simplicity's sake take two bonds that are exactly equal except that one, with zero default risk is certain to return $100 at maturity, and the other, with a known default risk of 50% is certain to return $50. Why would the investor care whether she buys 1 of the safe bond or two of the risky bond at 1/2 the price, which is how the EMH would price them? Risk is risk, and $100 is $100. It seems to me that stocks are simply a more complex variant of that analysis, and the theory of the EMH says that the market will take all of these complexities into account in the same manner as the two bonds.
If a bond has a default risk of 50%, then it cannot be said that a return of $50 is certain. Rather, the expected return is $50. These two bonds are different because one is guaranteed $100, while the second cannot be certain. You are compensated for this uncertainty (risk)
OK I see that. So then let’s say that they efficient market hypothesis would put the price at $40. I think that is Loretta‘s point – – that the correct price will reflect this risk in which case, assuming one believes in the efficient market hypothesis pricing, the investor would be indifferent as to which one she buys.

quantAndHold
Posts: 1701
Joined: Thu Sep 17, 2015 10:39 pm

Re: What's wrong with this critique of indexing?

Post by quantAndHold » Sat Apr 28, 2018 2:19 pm

nisiprius wrote:
Sat Apr 28, 2018 12:46 pm
whodidntante wrote:
Sat Apr 28, 2018 12:34 pm
Meb Faber said that just about anything is better than cap weighting in one of his podcasts. If you add some sensible qualifiers like you keep all costs low, including taxes and trading costs, and you remain broadly diversified, that might be correct.
No, it can't possibly be correct. Not without a lot of rhetorical mischief and mental reservations about what, exactly, is meant.
Meb doesn’t entirely believe in EMH. Meb comes from a belief that the market is *mostly* efficient, i.e. it’s very hard to beat the market, but not impossible. One of the inefficiencies that Meb found has to do with size. Returns of the very largest companies tend to be less than the rest of the market. Which is why S&P500 minus the top 1 works better than S&P500, equal weight works better, random selection of stocks works better, etc, etc, etc. Like whodidntante said, practically everything works better than pure market cap. Problems arise, however in implementation, usually related to either cost or scalability.

selters
Posts: 537
Joined: Thu Feb 27, 2014 9:26 am

Re: What's wrong with this critique of indexing?

Post by selters » Sat Apr 28, 2018 2:44 pm

Lauretta wrote:
Sat Apr 28, 2018 10:48 am
Can you please point out at which point the following reasoning is wrong?
1) If one believes in the EMH and in common asset pricing models, then one accepts the idea that all information about firms in the stock market is baked into their current prices, so that for every stock you should expect the same risk adjusted return.
2) If this is the case, then my understanding is that the expected risk adjusted return for Apple should be the same as that for any other stock, say for News Corporation (if this weren't so their stock prices would instantly be adjusted - since markets are efficient - till that condition holds).
3) If both stocks (Apple and News Corporation) have all information baked into their price, and have the same risk adjusted returns, the rational thing for me to do is to make the same bet on each of them. The only reason why I should bet more on one stock than on the other, is if I had reason to believe that it would be more likely to outperform (and according to point 2) this is not true)
4) Yet by indexing, for every 100$ invested in the S&P500, I invest more than 3.5$ in Apple and less than one cent on News Corporation. How can this make sense? (i.e. what justifies this much bigger bet on Apple, just because it's bigger and more money is invested in it by other people?).
I think this is a great question. I've tried to ask this question before, for example in a thread I called "Does Market Cap Weighting Follow Logically from the Efficient Markets Hypothesis?" several years ago. It got a couple of dozen replies, but to be honest, I did not get it after that either. Maybe I wasn't able to ask the question as elegantly as you. Or even good enough to get my point across to the rest of the Bogleheads. The common reply that market cap weighting is how the global market has collectively allocated capital, and therefore you should, too, does not make sense to me. Hopefully I'll be enlightened when I've read the replies to this thread.

selters
Posts: 537
Joined: Thu Feb 27, 2014 9:26 am

Re: What's wrong with this critique of indexing?

Post by selters » Sat Apr 28, 2018 3:14 pm

jalbert wrote:
Sat Apr 28, 2018 1:25 pm
nisiprius wrote:
Sat Apr 28, 2018 1:16 pm
P.S. Lauretta, just for fun... somewhat tangential to your topic, but worth the effort... take a little while to look at the October, 2017 annual report, page 7, schedule of investments for the PowerShares Russell 2000 Equal Weight Portfolio (EQWS). It is holding a total of $23,137,392 worth of stock, so with about 2,000 constituents--they say 1,931--it ought to be holding... $23,137,392/1,931 = oh, roughly $12,000 worth of each.

Take a look.

Of course, the index it tracks might not be defined exactly the way you might expect.
I suspect they have wide rebalancing bands to reduce transaction costs.
And between each rebalancing, there will have been huge winners and terrible losers. Stock A and B, both with an allocation of $100 after the semi-annual rebalancing, will likely have drifted to $200 and $50 after six months.

drk
Posts: 491
Joined: Mon Jul 24, 2017 10:33 pm
Location: Seattle

Re: What's wrong with this critique of indexing?

Post by drk » Sat Apr 28, 2018 3:36 pm

quantAndHold wrote:
Sat Apr 28, 2018 2:19 pm
nisiprius wrote:
Sat Apr 28, 2018 12:46 pm
whodidntante wrote:
Sat Apr 28, 2018 12:34 pm
Meb Faber said that just about anything is better than cap weighting in one of his podcasts. If you add some sensible qualifiers like you keep all costs low, including taxes and trading costs, and you remain broadly diversified, that might be correct.
No, it can't possibly be correct. Not without a lot of rhetorical mischief and mental reservations about what, exactly, is meant.
Meb doesn’t entirely believe in EMH. Meb comes from a belief that the market is *mostly* efficient, i.e. it’s very hard to beat the market, but not impossible. One of the inefficiencies that Meb found has to do with size. Returns of the very largest companies tend to be less than the rest of the market. Which is why S&P500 minus the top 1 works better than S&P500, equal weight works better, random selection of stocks works better, etc, etc, etc. Like whodidntante said, practically everything works better than pure market cap. Problems arise, however in implementation, usually related to either cost or scalability.
If I may be glib: market-cap weighting is the worst form of indexing, except for all the others.

buylowbuyhigh
Posts: 23
Joined: Thu Dec 21, 2017 8:34 am

Re: What's wrong with this critique of indexing?

Post by buylowbuyhigh » Sat Apr 28, 2018 3:37 pm

Lauretta wrote:
Sat Apr 28, 2018 10:48 am
3) If both stocks (Apple and News Corporation) have all information baked into their price, and have the same risk adjusted returns, the rational thing for me to do is to make the same bet on each of them.
I like to think that this exactly what a cap weighted index fund does: it should hold x% of Apple shares and x% of News Corp shares.

Why not own future dividends of the whole stock market without preferring one company or other? If Apple is believed to produce multiple times the dividends of some other company, own multiple times the value in Apple vs the other company. Then let the market decide how valuable the future dividends are on the company, sector and country level. Personally I want to own x% of every company (the free-float part of course) on earth, measured by market cap (shouldn't this be also x% of the floating shares by definition?). Don't care about size, valuation, sector or domicile. Currently I still omit for example micro cap in frontier markets due to costs however :mrgreen:

Wonder if this view is presented somewhere. To me it's the most intuitive reason to cap weigh.

User avatar
Lauretta
Posts: 615
Joined: Wed Jul 05, 2017 6:27 am

Re: What's wrong with this critique of indexing?

Post by Lauretta » Sat Apr 28, 2018 3:47 pm

buylowbuyhigh wrote:
Sat Apr 28, 2018 3:37 pm


Why not own future dividends of the whole stock market without preferring one company or other? If Apple is believed to produce multiple times the dividends of some other company, own multiple times the value in Apple vs the other company.
The trouble is that if something unexpected happens to a single very large company like Apple (like it happened to big stocks like BP with Deepwater Horizon or with France Telecom) you will be much more damaged than if you held an equal weight portfolio (though as note above by SimpleGift, even Apple does not make up such a large proportion of the US stock market to make it a serious problem for US investors)
Last edited by Lauretta on Sat Apr 28, 2018 3:53 pm, edited 1 time in total.
When everyone is thinking the same, no one is thinking at all

User avatar
tadamsmar
Posts: 7627
Joined: Mon May 07, 2007 12:33 pm

Re: What's wrong with this critique of indexing?

Post by tadamsmar » Sat Apr 28, 2018 3:50 pm

Lauretta wrote:
Sat Apr 28, 2018 11:43 am
tadamsmar wrote:
Sat Apr 28, 2018 11:15 am


The reason to own multiple stocks is because stocks move somewhat independently with some covariance. So you get an overall lower risk-adjusted return for buying multiple stocks. The reason you would not own them in equal amounts has to do with the co-variance and variance of stocks.

According to the EMH, the index represents the rational ratio of stocks that we should buy.
Thanks for your feedback. Does this mean that empirically it is found that the covariance between stocks is so adjusted that you actually get the highest risk adjusted returns for a portfolio consisting of a market cap weighted index? Intuitively it seems strange since with the current S&p500 you have a large proportion of large tech stocks which should have quite a high covariance. This would seem to be reduced in a portolio in which all stocks of the S&P500 are equally weighted(?)
It's not supported or rejected on empirical grounds. That's very hard to do.

By EMH reasoning, if there was convincing evidence that the mix was wrong, then, by definition, that which is convincing would convince investors and the mix would adjust in accordance with the evidence

User avatar
Lauretta
Posts: 615
Joined: Wed Jul 05, 2017 6:27 am

Re: What's wrong with this critique of indexing?

Post by Lauretta » Sat Apr 28, 2018 3:50 pm

selters wrote:
Sat Apr 28, 2018 2:44 pm
Lauretta wrote:
Sat Apr 28, 2018 10:48 am
Can you please point out at which point the following reasoning is wrong?
1) If one believes in the EMH and in common asset pricing models, then one accepts the idea that all information about firms in the stock market is baked into their current prices, so that for every stock you should expect the same risk adjusted return.
2) If this is the case, then my understanding is that the expected risk adjusted return for Apple should be the same as that for any other stock, say for News Corporation (if this weren't so their stock prices would instantly be adjusted - since markets are efficient - till that condition holds).
3) If both stocks (Apple and News Corporation) have all information baked into their price, and have the same risk adjusted returns, the rational thing for me to do is to make the same bet on each of them. The only reason why I should bet more on one stock than on the other, is if I had reason to believe that it would be more likely to outperform (and according to point 2) this is not true)
4) Yet by indexing, for every 100$ invested in the S&P500, I invest more than 3.5$ in Apple and less than one cent on News Corporation. How can this make sense? (i.e. what justifies this much bigger bet on Apple, just because it's bigger and more money is invested in it by other people?).
I think this is a great question. I've tried to ask this question before, for example in a thread I called "Does Market Cap Weighting Follow Logically from the Efficient Markets Hypothesis?" several years ago. It got a couple of dozen replies, but to be honest, I did not get it after that either. Maybe I wasn't able to ask the question as elegantly as you. Or even good enough to get my point across to the rest of the Bogleheads. The common reply that market cap weighting is how the global market has collectively allocated capital, and therefore you should, too, does not make sense to me. Hopefully I'll be enlightened when I've read the replies to this thread.
thanks for pointing this out. That was an interesting thread, I like the fact that they pointed out the discrepancy between the elegant academic theories and the empirical facts.
When everyone is thinking the same, no one is thinking at all

Thesaints
Posts: 1390
Joined: Tue Jun 20, 2017 12:25 am

Re: What's wrong with this critique of indexing?

Post by Thesaints » Sat Apr 28, 2018 3:55 pm

Lauretta wrote:
Sat Apr 28, 2018 10:48 am
3) If both stocks (Apple and News Corporation) have all information baked into their price, and have the same risk adjusted returns, the rational thing for me to do is to make the same bet on each of them.
Why ?
The only reason why I should bet more on one stock than on the other, is if I had reason to believe that it would be more likely to outperform (and according to point 2) this is not true)
That's not correct. Their risk adjusted return might be the same, but not their return.
4) Yet by indexing, for every 100$ invested in the S&P500, I invest more than 3.5$ in Apple and less than one cent on News Corporation. How can this make sense? (i.e. what justifies this much bigger bet on Apple, just because it's bigger and more money is invested in it by other people?).
Capitalization weighted indexes have nothing to do with expected returns.
If you have an ice cream parlors index, assuming their risk is the same, naturally a parlor with 300 locations would be more valuable than one with only 3.

User avatar
nisiprius
Advisory Board
Posts: 35735
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: What's wrong with this critique of indexing?

Post by nisiprius » Sat Apr 28, 2018 4:00 pm

quantAndHold wrote:
Sat Apr 28, 2018 2:19 pm
nisiprius wrote:
Sat Apr 28, 2018 12:46 pm
whodidntante wrote:
Sat Apr 28, 2018 12:34 pm
Meb Faber said that just about anything is better than cap weighting in one of his podcasts. If you add some sensible qualifiers like you keep all costs low, including taxes and trading costs, and you remain broadly diversified, that might be correct.
No, it can't possibly be correct. Not without a lot of rhetorical mischief and mental reservations about what, exactly, is meant.
Meb doesn’t entirely believe in EMH. Meb comes from a belief that the market is *mostly* efficient, i.e. it’s very hard to beat the market, but not impossible. One of the inefficiencies that Meb found has to do with size. Returns of the very largest companies tend to be less than the rest of the market. Which is why S&P500 minus the top 1 works better than S&P500, equal weight works better, random selection of stocks works better, etc, etc, etc. Like whodidntante said, practically everything works better than pure market cap. Problems arise, however in implementation, usually related to either cost or scalability.
quantandhold, it is impossible.
  • It doesn't have anything to do with the EMH.
  • It doesn't have anything to do with implementation.
  • It doesn't have anything to do with cost.
  • It doesn't have anything to do with cost or scalability.
It has to do with arithmetic. If the pot of soup has an average of 10 dumplings per ladle, and I stir the soup vigorously so that my ladleful contains an accurate duplicate of the composition of the whole pot, I get 10 dumplings (on average). It is not possible that "practically everything else" will get more than ten dumplings (on average).

Unless you hide something huge under the word "practically."

Issues of implementation, cost, and scalability make things worse, of course.

What might be happening is publication bias. Nobody bothers to talk or write about things they've tried that have, in the past, worked worse than pure market cap. It could be true that "practically everything people publish articles about had, at the time of publication of the article, worked better than pure market cap." However, "everything people publish articles about" is a big qualification, and "had, at the time publication of the article" is another big qualification.
Last edited by nisiprius on Sat Apr 28, 2018 4:11 pm, edited 6 times in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

Thesaints
Posts: 1390
Joined: Tue Jun 20, 2017 12:25 am

Re: What's wrong with this critique of indexing?

Post by Thesaints » Sat Apr 28, 2018 4:04 pm

quantAndHold wrote:
Sat Apr 28, 2018 2:19 pm
random selection of stocks works better,
That's kind of hard to believe: "random selection of stocks" would replicate the cap weighted index.

buylowbuyhigh
Posts: 23
Joined: Thu Dec 21, 2017 8:34 am

Re: What's wrong with this critique of indexing?

Post by buylowbuyhigh » Sat Apr 28, 2018 4:05 pm

Lauretta wrote:
Sat Apr 28, 2018 3:47 pm
buylowbuyhigh wrote:
Sat Apr 28, 2018 3:37 pm


Why not own future dividends of the whole stock market without preferring one company or other? If Apple is believed to produce multiple times the dividends of some other company, own multiple times the value in Apple vs the other company.
The trouble is that if something unexpected happens to a single very large company like Apple (like it happened to big stocks like BP with Deepwater Horizon or with France Telecom) you will be much more damaged than if you held an equal weight portfolio (though as note above by SimpleGift, even Apple does not make up such a large proportion of the US stock market to make it a serious problem for US investors)
Don't know what happened to France Telecom, but BP is still in existence isn't it? What if Deepwater made up 100% of some small company and you invested in that company as much in currency as say in Apple? Couldn't this single-company risk also be priced in by the market so that Apple is cheaper than it otherwise would be? (Honest question, don't know the answer.)

golfCaddy
Posts: 400
Joined: Wed Jan 10, 2018 10:02 pm

Re: What's wrong with this critique of indexing?

Post by golfCaddy » Sat Apr 28, 2018 4:17 pm

gmaynardkrebs wrote:
Sat Apr 28, 2018 2:09 pm
OK I see that. So then let’s say that they efficient market hypothesis would put the price at $40. I think that is Loretta‘s point – – that the correct price will reflect this risk in which case, assuming one believes in the efficient market hypothesis pricing, the investor would be indifferent as to which one she buys.
I think your implicit assumption is that the two bonds have a correlation of 1. That's not a good assumption.

gmaynardkrebs
Posts: 412
Joined: Sun Feb 10, 2008 11:48 am

Re: What's wrong with this critique of indexing?

Post by gmaynardkrebs » Sat Apr 28, 2018 4:36 pm

golfCaddy wrote:
Sat Apr 28, 2018 4:17 pm
gmaynardkrebs wrote:
Sat Apr 28, 2018 2:09 pm
OK I see that. So then let’s say that they efficient market hypothesis would put the price at $40. I think that is Loretta‘s point – – that the correct price will reflect this risk in which case, assuming one believes in the efficient market hypothesis pricing, the investor would be indifferent as to which one she buys.
I think your implicit assumption is that the two bonds have a correlation of 1. That's not a good assumption.
Why would they not?

drk
Posts: 491
Joined: Mon Jul 24, 2017 10:33 pm
Location: Seattle

Re: What's wrong with this critique of indexing?

Post by drk » Sat Apr 28, 2018 4:49 pm

Lauretta wrote:
Sat Apr 28, 2018 3:47 pm
The trouble is that if something unexpected happens to a single very large company like Apple (like it happened to big stocks like BP with Deepwater Horizon or with France Telecom) you will be much more damaged than if you held an equal weight portfolio (though as note above by SimpleGift, even Apple does not make up such a large proportion of the US stock market to make it a serious problem for US investors)
The issue is that the risk of a catastrophic event is higher for a smaller company than for a larger one. If there is an enduring small-cap premium, that's why it exists: you're taking more risk by owning companies that are less resilient to adversity. By holding an equal-weighted portfolio, you're amplifying your exposure to those idiosyncratic risks.

From peak (4/20/2010) to trough (6/30/2010), BP lost about 55% of its market cap. Let's game-plan two scenarios:
  • In a market-cap-weighted portfolio, Apple's losing 55% of its $824 billion market cap would have about 50 times the impact that a 55% drop in Snap's $17 billion market cap would have
  • In an equal-weighted portfolio, Snap's shedding $9 billion in capitalization would have the same impact as a $450 billion drop for Apple
Logically, a $9 billion drop in capitalization is way, way more likely than a $450 billion one. In fact, it happened to Apple just on Friday, as it dropped 1.16% from $833 billion to $824 billion. To be honest, I didn't notice, but I sure would have if I held an equal-weighted portfolio and Snap dropped $9 billion. Size fosters resiliency and robustness, and I'm happy to be able to take advantage of that.

Granted, the counter-argument is that small companies are way more likely to double in size than are giant ones. Unfortunately, in an equal-weighted portfolio, you still don't get to just pick the few winners that see that happen. Instead, you end up taking the much bigger pack of losers along with their locally catastrophic but globally minuscule losses.

Please forgive me for letting this ramble a bit.

User avatar
Lauretta
Posts: 615
Joined: Wed Jul 05, 2017 6:27 am

Re: What's wrong with this critique of indexing?

Post by Lauretta » Sat Apr 28, 2018 5:14 pm

drk wrote:
Sat Apr 28, 2018 4:49 pm
Lauretta wrote:
Sat Apr 28, 2018 3:47 pm
The trouble is that if something unexpected happens to a single very large company like Apple (like it happened to big stocks like BP with Deepwater Horizon or with France Telecom) you will be much more damaged than if you held an equal weight portfolio (though as note above by SimpleGift, even Apple does not make up such a large proportion of the US stock market to make it a serious problem for US investors)
The issue is that the risk of a catastrophic event is higher for a smaller company than for a larger one. If there is an enduring small-cap premium, that's why it exists: you're taking more risk by owning companies that are less resilient to adversity. By holding an equal-weighted portfolio, you're amplifying your exposure to those idiosyncratic risks.

From peak (4/20/2010) to trough (6/30/2010), BP lost about 55% of its market cap. Let's game-plan two scenarios:
  • In a market-cap-weighted portfolio, Apple's losing 55% of its $824 billion market cap would have about 50 times the impact that a 55% drop in Snap's $17 billion market cap would have
  • In an equal-weighted portfolio, Snap's shedding $9 billion in capitalization would have the same impact as a $450 billion drop for Apple
Logically, a $9 billion drop in capitalization is way, way more likely than a $450 billion one. In fact, it happened to Apple just on Friday, as it dropped 1.16% from $833 billion to $824 billion. To be honest, I didn't notice, but I sure would have if I held an equal-weighted portfolio and Snap dropped $9 billion. Size fosters resiliency and robustness, and I'm happy to be able to take advantage of that.

Granted, the counter-argument is that small companies are way more likely to double in size than are giant ones. Unfortunately, in an equal-weighted portfolio, you still don't get to just pick the few winners that see that happen. Instead, you end up taking the much bigger pack of losers along with their locally catastrophic but globally minuscule losses.

Please forgive me for letting this ramble a bit.
Thanks for your feedback :happy This makes sense; but in an equal weighted portfolio in order to get the same damage as you'd get in a market cap weighted portfolio where Apple lost 55% (as BP did) you'd have to get a a big drop in many companies - since in an equal weighted portfolio none of the firms has a big overall impact as they each make up only a small fraction of it. So even though each small company is less resilient than Apple, a single black swan event happening to Apple in a market cap weighted portfolio would only be matched if you had many companies suffering concomitantly a loss in an equal weighted porfolio. Which scenario is more likely? I wouldn't know how to quantify it... :confused
When everyone is thinking the same, no one is thinking at all

pastel
Posts: 11
Joined: Mon Jan 08, 2018 7:48 pm

Re: What's wrong with this critique of indexing?

Post by pastel » Sat Apr 28, 2018 5:38 pm

Lauretta wrote:
Sat Apr 28, 2018 5:14 pm
drk wrote:
Sat Apr 28, 2018 4:49 pm
Lauretta wrote:
Sat Apr 28, 2018 3:47 pm
The trouble is that if something unexpected happens to a single very large company like Apple (like it happened to big stocks like BP with Deepwater Horizon or with France Telecom) you will be much more damaged than if you held an equal weight portfolio (though as note above by SimpleGift, even Apple does not make up such a large proportion of the US stock market to make it a serious problem for US investors)
The issue is that the risk of a catastrophic event is higher for a smaller company than for a larger one. If there is an enduring small-cap premium, that's why it exists: you're taking more risk by owning companies that are less resilient to adversity. By holding an equal-weighted portfolio, you're amplifying your exposure to those idiosyncratic risks.

From peak (4/20/2010) to trough (6/30/2010), BP lost about 55% of its market cap. Let's game-plan two scenarios:
  • In a market-cap-weighted portfolio, Apple's losing 55% of its $824 billion market cap would have about 50 times the impact that a 55% drop in Snap's $17 billion market cap would have
  • In an equal-weighted portfolio, Snap's shedding $9 billion in capitalization would have the same impact as a $450 billion drop for Apple
Logically, a $9 billion drop in capitalization is way, way more likely than a $450 billion one. In fact, it happened to Apple just on Friday, as it dropped 1.16% from $833 billion to $824 billion. To be honest, I didn't notice, but I sure would have if I held an equal-weighted portfolio and Snap dropped $9 billion. Size fosters resiliency and robustness, and I'm happy to be able to take advantage of that.

Granted, the counter-argument is that small companies are way more likely to double in size than are giant ones. Unfortunately, in an equal-weighted portfolio, you still don't get to just pick the few winners that see that happen. Instead, you end up taking the much bigger pack of losers along with their locally catastrophic but globally minuscule losses.

Please forgive me for letting this ramble a bit.
Thanks for your feedback :happy This makes sense; but in an equal weighted portfolio in order to get the same damage as you'd get in a market cap weighted portfolio where Apple lost 55% (as BP did) you'd have to get a a big drop in many companies - since in an equal weighted portfolio none of the firms has a big overall impact as they each make up only a small fraction of it. So even though each small company is less resilient than Apple, a single black swan event happening to Apple in a market cap weighted portfolio would only be matched if you had many companies suffering concomitantly a loss in an equal weighted porfolio. Which scenario is more likely? I wouldn't know how to quantify it... :confused
A typical total market index fund seeks to match the returns of the market as a whole. Indexes are market cap weighted because how much a firm contributes to the overall market and economy is reflected by its economic output, which in turn is reflected in its market cap.

By equal weighting each firm, you are no longer seeking to match the market as a whole. Rather, it is an implicit bet on smaller firms. There is nothing wrong per se with this approach, but you are no longer indexing in the traditional sense. It would be similar to allocating more to small/mid cap firms and less to large cap.

User avatar
Lauretta
Posts: 615
Joined: Wed Jul 05, 2017 6:27 am

Re: What's wrong with this critique of indexing?

Post by Lauretta » Sat Apr 28, 2018 5:49 pm

pastel wrote:
Sat Apr 28, 2018 5:38 pm

By equal weighting each firm, you are no longer seeking to match the market as a whole. Rather, it is an implicit bet on smaller firms.
As I mentioned in my OP since each firm, large or small, is priced efficiently, I don't see why I am betting on smaller firms if I equally weight all companies. It rather seems to me that I bet on large firms if I do market weighting. If I am equally confident that the price of a small firm incorporates all available information, just like the price of Apple does, it makes more sense to bet equal amounts on the 2.
When everyone is thinking the same, no one is thinking at all

Post Reply