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: 605
Joined: Wed Jul 05, 2017 6:27 am

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

Post by Lauretta » Sun Apr 29, 2018 1:01 pm

nisiprius wrote:
Sun Apr 29, 2018 12:26 pm

In other words, all of the competing tilt, factor-based, fundamental indexing, smart beta strategies and products will beat the market, automatically, just by not being market-cap weighted, because market cap weighting is actually the worst of all. The whole contemporary zoo of factor-based, fundamentally indexed, smart beta products are all better than cap-weighted index funds. I imagine the Research Associates people, not speaking for academic publication, would say "of course our weighting system is the best." However, in a general sense, it is in the interest of Research Associates to make the case that investors should begin their consideration of mutual funds by screening out all cap-weighted products.
In more recent times RA have warned against factor investing:
https://www.researchaffiliates.com/en_u ... wrong.html

Of course what RA write is done also with the aim of promoting their products. But this doesn't mean that what academics write is more trustworthy or 'disinterested'.
When everyone is thinking the same, no one is thinking at all

gmaynardkrebs
Posts: 404
Joined: Sun Feb 10, 2008 11:48 am

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

Post by gmaynardkrebs » Sun Apr 29, 2018 1:15 pm

nisiprius wrote:
Sun Apr 29, 2018 12:19 pm

If we try the same thing, looking at portfolios consisting of a mix of the VFINX (S&P 500) and MGC, the Vanguard Mega Cap Index ETF, it is really a joke since the two asset points almost lie on top of each other, but, nevertheless, the result is the same: the MGC has just a hair higher return and just a hair lower standard deviation, and the two have literally 1.00 correlation with each other (down to rounding error), and accordingly the optimum combination is 0% S&P 500, 100% mega caps.
I'd be curious to run a data set for a period before the widespread adoption of passive index investing to see what the correlation was then. A correlation of 1 is astonishing, and I wonder if the rise of the passive index funds has anything to do with it.

User avatar
Christine_NM
Posts: 2625
Joined: Tue Feb 20, 2007 1:13 am
Location: New Mexico

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

Post by Christine_NM » Sun Apr 29, 2018 1:29 pm

Lauretta wrote:
Sun Apr 29, 2018 12:26 pm
golfCaddy wrote:
Sun Apr 29, 2018 12:15 pm

How do you define risk? In Fama's view, risk is multidimensional and defined as a weighting on one of the five factors.
profitability is one of the factors in the five factors model: why are profitable firms more risky?
It's the combination of smallness, low profitability, and high investment that makes for a higher risk profile according to what I can make of the five-factor model. Beware, this is a vast oversimplification. Swedroe had an in-English article about this in etf.com: http://www.etf.com/sections/index-inves ... nopaging=1


But even CMA (profitability-investment factor) has its winners, like Amazon investing revenue in itself in the early days. To exclude it because of CMA would have been a huge mistake, in my view a fatal error for this model.
17% cash 47% stock 36% bond. Retired, w/d rate 2.85%

lazyday
Posts: 3296
Joined: Wed Mar 14, 2007 10:27 pm

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

Post by lazyday » Sun Apr 29, 2018 1:39 pm

Christine_NM wrote:
Sun Apr 29, 2018 1:29 pm
It's the combination of smallness, low profitability, and high investment that makes for a higher risk profile according to what I can make of the five-factor model. Beware, this is a vast oversimplification.
I believe that high profitability and low investment have meant higher returns in backtests.

If low profitability gave high returns, then I might believe the risk story.

User avatar
Christine_NM
Posts: 2625
Joined: Tue Feb 20, 2007 1:13 am
Location: New Mexico

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

Post by Christine_NM » Sun Apr 29, 2018 1:51 pm

lazyday wrote:
Sun Apr 29, 2018 1:39 pm
Christine_NM wrote:
Sun Apr 29, 2018 1:29 pm
It's the combination of smallness, low profitability, and high investment that makes for a higher risk profile according to what I can make of the five-factor model. Beware, this is a vast oversimplification.
I believe that high profitability and low investment have meant higher returns in backtests.

If low profitability gave high returns, then I might believe the risk story.
small, low profit + high investment = lower average return. (Perhaps I should not have equated that with higher risk. I am not sure of 5-factor jargon yet.)
17% cash 47% stock 36% bond. Retired, w/d rate 2.85%

User avatar
nisiprius
Advisory Board
Posts: 35718
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 » Mon Apr 30, 2018 2:01 pm

Lauretta wrote:
Sun Apr 29, 2018 1:01 pm
...In more recent times RA have warned against factor investing:
How can "Smart Beta" Go Horribly Wrong?

Of course what RA write is done also with the aim of promoting their products. But this doesn't mean that what academics write is more trustworthy or 'disinterested'.
Notice the interesting inconsistency. Jason Hsu isn't an author of the this article, but he's mentioned eight times, and he's a partner in the firm of which author Rob Arnott is founder and chairman, so it is unlikely that Arnott doesn't know about his work.

Now, in the 2006 article, Cap-Weighted Portfolios are Suboptimal Portfolios, Hsu says that cap-weighting is uniquely bad. You don't need to do anything particular, to beat cap-weighting, you just have to use... well, let's quote his words: "However, portfolios constructed from weights, which do not depend on prices, do not exhibit the same underperformance observed for cap-weighted portfolios."

Yet "Smart Beta" portfolios are an example of "portfolios constructed from weights which do not depend on," well, let's say, "prices alone." So if "Smart Beta" portfolios can go horribly wrong, this (seems) inconsistent with the (apparent) claim that any departure from cap-weighting is better than cap-weighting.

A possible explanation might be found in a later paper mentioned in the article, a paper by Arnott andHsu and two others... The Surprising Alpha From Malkiel’s Monkey and Upside-Down Strategies. As I read it, this paper is saying much less intriguing than "cap-weighting is the worst strategy." They say
"value and small-cap exposures are naturally occurring portfolio characteristics, unless an investor constructs a portfolio to have a positive relationship between price and portfolio weights.
In other words, they say, most non-cap-weighted portfolios turn out to have small-cap value tilts, intentionally or not. (Malkiel's monkeys throwing darts at the stock listings incorporate a small-cap tilt because they are statistically doing equal weighting; the listing for Apple is printed in the same-sized type as the listing for News Corporation.) So instead of "anything is better than cap-weighting," we are just back to the old "small-cap value tilts are better than cap-weighting." They dodge the obvious by saying "We do not attempt to comment on the interesting debate regarding the nature of value and small-cap premiums," and also do not answer the question I always have: whether they are talking about superior risk-adjusted return, or just about a premium which might well be no more than reward commensurate with extra risk.

Anyway, we are not left with "cap-weighting is worst." We're left with "small-cap value is best, cap-weighting is worse, and" (I suppose) "whatever strategies result in large-cap or growth tilts, intentional or not, are worst." Same-old same-old argument that's been going on for decades.

And meanwhile large-cap growth funds, indexed or active, continue merrily on their way, with plenty of people investing in them with seemingly not-terrible results.
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
neurosphere
Posts: 2821
Joined: Sun Jan 17, 2010 1:55 pm

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

Post by neurosphere » Tue May 01, 2018 10:34 am

nisiprius wrote:
Mon Apr 30, 2018 2:01 pm
Anyway, we are not left with "cap-weighting is worst." We're left with "small-cap value is best, cap-weighting is worse, and" (I suppose) "whatever strategies result in large-cap or growth tilts, intentional or not, are worst." Same-old same-old argument that's been going on for decades.
The entire post from which this quote is taken is marvelous. :beer
If you have to ask "Is a Target Date fund right for me?", the answer is "Yes".

User avatar
Lauretta
Posts: 605
Joined: Wed Jul 05, 2017 6:27 am

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

Post by Lauretta » Tue May 01, 2018 10:49 am

nisiprius wrote:
Mon Apr 30, 2018 2:01 pm

And meanwhile large-cap growth funds, indexed or active, continue merrily on their way, with plenty of people investing in them with seemingly not-terrible results.
yes the results of large growth funds are good. It's just that studies have shown that historically, over long periods of time, those of small cap value have been better.
In the last 10 years growth has outperformed and indeed there have always been stretches of time when this was the case - people argue that this is precisely the reason why the value premium will persist, because you have to stick for a long time with an underperforming strategy and not many people are prepared to do it. Will value outperform in future? Nobody knows. But I think it makes more sense to base one's opinions on the results over the last 100 yrs, rather than on those of the last 10 yrs...
When everyone is thinking the same, no one is thinking at all

Alpha3
Posts: 5
Joined: Tue Apr 24, 2018 3:35 am

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

Post by Alpha3 » Wed May 16, 2018 8:11 am

Lauretta wrote:
Sat Apr 28, 2018 10:48 am
Can you please point out at which point the following reasoning is wrong?
This has been a very interesting thread to read. Here's my attempt to answer the original riddle:
In my mind, yours is not a critique to indexing, it is a critique to the particular indexing approach of the S&P500, which is cap weighting. And there's nothing wrong in the critique. As far as I can tell, allocation on the basis of market cap has been shown to deliver weaker performances. In other words, as per your last point, investing more in Apple than News corp isn't the best decision. That said, this doesn't prove indexing is wrong though, just that indexing on the basis of cap is far from optimal.

But I completely hear you. In fact, I had a very similar question (which I posted last week, coincidentally). I thought: if the S&P500 is widely regarded as the most efficient way to track the market, and the S&P is weighted on capitalization, surely the same principle can be applied to the problem of asset allocation. To my surprise, I found multiple papers with evidence suggesting this was a poor performing strategy.

Whatever principle is making cap weighting to be a bad heuristic for asset allocation, surely it must also be working at the equity allocation level.

Interestingly, I found several studies suggesting that 1/N (equal weighting) was possibly the best ex-ante porfolio. It is the one Markowitz ended up choosing, in spite of comming up with MPT!!! :) I recommend the following paper on the topic:
https://pdfs.semanticscholar.org/094d/2 ... 06953e.pdf

This also compares a global market asset allocation on cap weight, vs several other heuristics, showing that cap weighting isn't preferrable: http://www.sr-sv.com/global-market-port ... rformance/

More evidence: http://www.macroresilience.com/2010/07/ ... llocation/

In short, I think you can conclude there's nothing wrong with the principles you call upon, and equal size investing (as you were suggesting) has been shown to deliver stronger performance than cap weighting.

PS: Thinking about this has made me wonder whether the size factor in the French-Fama models is actually introduced by the very wide adoption of S&P500 as a benchmark. By general allocating on the basis of weight, small companies are penalized. As their cap increases however, allocators increasingly have to add more weight, which in turn drives prices up. This made me wonder... if at some point the S&P500 lost its current glamour, and was replaced by 1/N indexes, perhaps the size factor would all together disappear and be fully arbitraged out. I'm still pondering on this :)

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 » Thu May 17, 2018 6:42 am

Alpha3 wrote:
Wed May 16, 2018 8:11 am
Interestingly, I found several studies suggesting that 1/N (equal weighting) was possibly the best ex-ante porfolio. It is the one Markowitz ended up choosing, in spite of comming up with MPT!!! :)
To be specific, Markowitz used 1/2 (equal weighting) of stocks and bonds, nothing more complex than that:

http://www.mymoneyblog.com/harry-markow ... folio.html

But he was no more wedded to equal weighting than he was to MPT:

"Say you were 65, and invested $1 million, with 60 percent in stocks and 40 percent in bonds," he said. "It became $800,000, and you are not happy, but you lived to invest another day."

http://articles.chicagotribune.com/2010 ... -investing (<- This is a good read)

He seemed to choose practical heuristics that were rough approximations of MPT.

PS: That last article indicates that Markowitz did a lot of things. Early in his life, was 50/50, he rebalanced only with new money , in 2007 he decided the market was risky and sold a bunch of stock, he did not get back in till the the Fed started increasing interest rates, he made a bad real estate investment, he is big into municipal bonds in his late 80s.

But, there is not a enough quantitative info in that article to indicate that he would have done worse if he had just earnestly invested according to MPT principles as best he could for his whole life and engaged in systematic rebalancing and chose an appropriate level of risk based on age. His 2007 move (not staying the course) seems good, but slavishly rebalancing an MPT portfolio once a year though that recession (staying the course) was good too.

User avatar
unclescrooge
Posts: 2091
Joined: Thu Jun 07, 2012 7:00 pm

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

Post by unclescrooge » Thu May 17, 2018 8:58 am

nisiprius wrote:
Sat Apr 28, 2018 9:30 pm
(Sorry, I didn't see that ThriftyPhd had beaten me to it. He makes exactly the same PortfolioVisualizer comparison I show below, except he chose RSP as Portfolio 1 and SPY as portfolio 2.)
clown wrote:
Sat Apr 28, 2018 7:53 pm
...Therefore, in my view, the most important thing among two strategies is how much money is made and how much risk is taken. Being a skeptic, I prefer to run my own numbers. The following spreadsheet confirms the outperformance of equal-weight with only marginal increase in risk (if volatility is your definition of risk, a definition with which some investors disagree).
Why did you choose 2008 as your starting point instead of using all available data? If I go into PortfolioVisualizer and compare SPY and RSP, and use all available data, this is what I'm seeing.
Portfolio 1, blue, is SPY;
portfolio 2, red, is RSP.
  • Yes, RSP had higher return.
  • Yes, it had higher risk.
  • No, it didn't have "only marginal increase in risk."
  • No, it didn't just have an increase as measured by standard deviation, but by every measure calculated--higher standard deviation, worse worst year, and worse drawdown.
  • No, it was not superior if you looked at risk-adjusted return. If you look at risk-adjusted return the difference goes away. The Sharpe ratio was exactly the same to within roundoff error; the Sortino ratio was only a hair higher.
(Jan 2004 - Mar 2018)

Source

Image

RSP has the lure of novelty and something different, but I think it's just a sloppy and imprecise way of applying a modest small-cap tilt (within the universe of the S&P 500).
Given portfolio 2 has the same Sharpe ratio, and tiny bit better sortino, wouldn't you want to prefer that portfolio?

Wouldn't it tilt more towards mid-cap?

Post Reply