Asness - The Small-Firm Effect Is Real, and It’s Spectacular

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
matjen
Posts: 1965
Joined: Sun Nov 20, 2011 11:30 pm

Asness - The Small-Firm Effect Is Real, and It’s Spectacular

Post by matjen » Fri Jan 23, 2015 12:57 pm

Another Cliff's Perspective for the forum to digest.

https://www.aqr.com/cliffs-perspective/ ... pectacular

"Does size matter? Certainly in financial markets, many researchers have questioned whether it does. In a new paper, we resurrect the size premium and restore it to its proper place alongside such stalwarts as value and momentum in terms of its efficacy and robustness. After so many years of intensive research questioning the efficacy of the size effect, that it can be emphatically revived may seem a bit far-fetched. So, how can it be that size does indeed still matter? I’ll get to that, but first a little background."
A man is rich in proportion to the number of things he can afford to let alone.

richard
Posts: 7961
Joined: Tue Feb 20, 2007 3:38 pm
Contact:

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by richard » Fri Jan 23, 2015 1:12 pm

He isn't really saying the small-firm effect is real. He's essentially pushing the new quality factor: "In its normal form the small firm effect is, on a host of dimensions, weak to possibly nonexistent. Once adjusted for quality exposure it is real and spectacular."

countmein
Posts: 404
Joined: Fri Dec 06, 2013 9:10 pm

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by countmein » Fri Jan 23, 2015 2:50 pm

richard wrote:He isn't really saying the small-firm effect is real. He's essentially pushing the new quality factor: "In its normal form the small firm effect is, on a host of dimensions, weak to possibly nonexistent. Once adjusted for quality exposure it is real and spectacular."
I agree that his conclusion is off-kilter. It's not exactly a small cap effect, rather that QMJ has a size gradient (as does value when sorted on P/B). Or that SMB has a QMJ gradient. But, does the SMB premium go all the way to negative at the junkier side of the QMJ spectrum? If so, couldn't you just as well say that there's a BMS (large cap) factor, inversely following a QMJ gradient?

User avatar
Robert T
Posts: 2532
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by Robert T » Fri Jan 23, 2015 6:19 pm

.
Since the inception of the Bridgeway fund BRSIX 7/31/1997 - 12/31/2014 - annualized return

11.8% = CRSP10 (Benchmark for BRSIX)
11.5% = Brirdgeway Ultra-small Co. Mkt (BRSIX)
10.0% = DFA US Microcap (DFSCX)
8.6% = Vanguard Small Cap Index (NAESX)
6.4% = Vanguard 500 (VFINX)

http://quotes.morningstar.com/chart/fun ... A%5B%5D%7D
.

ralph124cf
Posts: 2028
Joined: Tue Apr 01, 2014 11:41 am

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by ralph124cf » Sat Jan 24, 2015 10:59 am

The paper does not explain in any way how to implement this strategy, or how to distinguish quality from junk. How could one build an index fund of small quality? This sounds like straight stock picking.

Ralph

jdilla1107
Posts: 601
Joined: Sun Jun 24, 2012 8:31 pm

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by jdilla1107 » Sat Jan 24, 2015 12:08 pm

Robert T wrote:.
Since the inception of the Bridgeway fund BRSIX 7/31/1997 - 12/31/2014 - annualized return

11.8% = CRSP10 (Benchmark for BRSIX)
11.5% = Brirdgeway Ultra-small Co. Mkt (BRSIX)
10.0% = DFA US Microcap (DFSCX)
8.6% = Vanguard Small Cap Index (NAESX)
6.4% = Vanguard 500 (VFINX)

http://quotes.morningstar.com/chart/fun ... A%5B%5D%7D
.
Did you try any other time periods? 10 years looks pretty bad for BRSIX.

User avatar
nisiprius
Advisory Board
Posts: 36851
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: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by nisiprius » Sat Jan 24, 2015 1:59 pm

richard wrote:He isn't really saying the small-firm effect is real. He's essentially pushing the new quality factor: "In its normal form the small firm effect is, on a host of dimensions, weak to possibly nonexistent. Once adjusted for quality exposure it is real and spectacular."
Indeed. Now, this time let's stop the tape after the first sentence:
In its normal form the small firm effect is, on a host of dimensions, weak to possibly nonexistent.
What does this say about the reliability of what the "financial science" community has been telling us since 1981? That community was talking about the small firm effect "in its normal form." So "everyone" used to say that the small-firm effect "in its normal form" was real. In fact, spectacular might be a fair characterization: in 1983, Rob Arnott wrote:
A growing body of evidence exists that small-capitalization stocks significantly outperform large-capitalization stocks. This effect is so strong and so consistent that even advocates of the Efficient Market Hypothesis have found no refutation of this effect.
Since then there have some various discoveries, including the finding of methodological flaws in the data series in which the effect was discovered.

What people ought to be saying is "What the financial 'science' community used to tell us has turned out to be wrong." But nobody is saying this. It is magician's misdirection. We will tell you something different now, give it the same name as something we used to say, and claim that we used to say is still true. For a while, the claim was that the small-firm effect was not very powerful by itself, but become powerful when combined with value. I am not quite sure what Asness is saying about small value, are you? Since he is so insistent about the important of quality, I assume that he feels that the effect value is also weak, but I can't tell.

The lesson I take from this is that it is it is hard to know how far to rely on what we are told by the financial 'science' community.
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.

Caduceus
Posts: 1728
Joined: Mon Sep 17, 2012 1:47 am

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by Caduceus » Sat Jan 24, 2015 2:33 pm

If one could distinguish small junk from small quality, wouldn't it be better to focus on the best ideas after researching the crap out of them instead of buying fractions of 182 companies?

Such coyness.

User avatar
nisiprius
Advisory Board
Posts: 36851
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: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by nisiprius » Sat Jan 24, 2015 4:17 pm

At a quick glance, it appears as if Asness, Frazzini and Pedersen do define how "quality" is to be measured,

http://www.aqr.com/library/working-pape ... minus-junk
Our quality measures are constructed as follows (details are in the appendix). We measure profitability by gross profits over assets (GPOA), return on equity (ROE), return on assets (ROA), cash flow over assets (CFOA), gross margin (GMAR), and the fraction of earnings composed of cash (i.e. low accruals, ACC). In order to put each measure on equal footing and combine them, each month we convert each variable into ranks and standardize to obtain a z-score.
I'm not sure whether a sophisticated slice-and-dice Boglehead could easily get the data and calculate their "quality" score.

and do publish something on performance of QMJ:

https://www.aqr.com/ex/dataset/DownloadDataSet

But this only tells us the performance of named portfolios, not the ticker symbols of the stocks falling into each decile.

If GPOA, ROE, ROA, CFOA, GMAR, ACC etc. are easily obtained data than it seems legitimate; if not, it's secret sauce. Perhaps the people in this forum who invest in individual stocks can tell me whether the "premium" stock screeners include "quality" scores for stocks already.

They say "However, the explanatory power of quality for prices is limited, presenting a puzzle for asset pricing." I think this is saying that they claim to have found the holy grail--a passively computed score that violates the EMH. "As a result, high quality firms exhibit high risk-adjusted returns." You can buy high-quality stocks almost as cheaply as low-quality stocks. If true, naturally one wonders how long this will persist now that they've told everybody about it.

Higher risk-adjusted returns and an EMH anomaly were exactly what Banz claimed in 1981 for "the firm size effect as normally defined:"
This study examines the empirical relationship between the return and the total market value of NYSE common stocks. It is found that smaller firms have had higher risk adjusted returns, on average, than larger firms. This ‘size effect’ has been in existence for at least forty years and is evidence that the capital asset pricing model is misspecified.
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.

Browser
Posts: 4857
Joined: Wed Sep 05, 2012 4:54 pm

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by Browser » Sat Jan 24, 2015 4:41 pm

Don't active managers try to buy "high quality stocks?" How has that turned out?
We don't know where we are, or where we're going -- but we're making good time.

User avatar
JoMoney
Posts: 6083
Joined: Tue Jul 23, 2013 5:31 am

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by JoMoney » Sat Jan 24, 2015 4:45 pm

jdilla1107 wrote:...
Did you try any other time periods? 10 years looks pretty bad for BRSIX.
From inception of DFSCX to 7/31/1997 you'd certainly get a different picture
Morningstar Link
"To achieve satisfactory investment results is easier than most people realize; to achieve superior results is harder than it looks." - Benjamin Graham

User avatar
nisiprius
Advisory Board
Posts: 36851
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: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by nisiprius » Sat Jan 24, 2015 5:05 pm

You know what else is missing? Numbers. He says the effect is "spectacular." OK, how much is that in dollars?

How many basis points of excess return does he say can be earned from a strategy using small-cap quality stocks, adjusted to have the same risk as a traditional portfolio?

Does he say anywhere?
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
Robert T
Posts: 2532
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by Robert T » Sat Jan 24, 2015 8:19 pm

.
CRSP Decile Annualized Return: 89 years to June 2014 - from Bridgeway annual report
From largest decile to smallest decile

.9.36% CRSP1
10.73%
11.04%
11.14%
11.68%
11.57%
11.73%
11.78%
11.73%
13.36% CRSP10

Seems to be some size effect - over the long-term - so far. BRSIX tracks CRSP10, has lagged by only 0.3% on an annualized basis since inception, although with large annual tracking error. Obviously no guarantees.

countmein
Posts: 404
Joined: Fri Dec 06, 2013 9:10 pm

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by countmein » Sat Jan 24, 2015 8:34 pm

I still think it's misleading for Asness to represent this as a resurrection of a size premium. Size simply amplifies the Value and Quality premia, for better or worse: thus low-scoring BtM and QMJ stocks fare worse in smaller stocks than in large. Small stocks have an underperforming side (low BtM, high QMJ) and an outperforming side (high BtM, high QMJ), and the two more or less cancel each other out, leaving little to no net size effect (edit- exception crsp10 perhaps per Robert's data above).

Here's a very quick and dirty analysis of French's Size x BtM x Profitability ("OP") dataset (1964 - present). The following numbers are the rounded % average annual returns, not compounded, for the four combinations of lowest/highest BtM/Profitability. Small half of the market in blue, large half in red.

You can see that the "small premium" goes negative on the bad end of BtM and OP, but goes way positive on the good end of BtM and OP. However, the number of low BtM, low OP firms is far greater than the number of high BtM, high OP firms.

Echoing Nisiprius's concerns, it seems bizarre to me that the 'financial scientists' didn't do every possible permutation of these simple data sorting exercises decades ago.

Image

TradingPlaces
Posts: 1245
Joined: Sun Nov 09, 2014 1:19 pm
Location: 30.286029, -97.530011

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by TradingPlaces » Sat Jan 24, 2015 9:57 pm

nisiprius wrote:At a quick glance, it appears as if Asness, Frazzini and Pedersen do define how "quality" is to be measured,

http://www.aqr.com/library/working-pape ... minus-junk
Well, none of those variables are secret. Here is how it goes:

1. Companies carefully compile their quarterly required info: (1) balance sheet, (2) cash-flow statement, (3) earnings statement.

2. This gets published to SEC EDGAR, and the whole world gets to see it at once,

3. Third party providers like Reuters or CapitalIQ immediately pick it up and within hours disseminate, via their standard data subscription (which costs money) to their clients, a structure summary of that report. E.g., all those variables, or something required to compute those variables will be there.

E.g,.

ROA = (some form of earnings, like EBIT, or EBITDA, or Net Earnings) / Current or Lagged Assets,

ROA = (some form of earnings) / Current or Lagged Equity.

What Asness is saying is that it does not matter what you use precisely.

Another thing he is saying is that you don't have to use the FAST versions of these variables. By fast, I mean: if UPS reported on Friday, January 23rd, and the numbers they reported are Q4-2014, then you could define ROA as:

(a) Earnings_Q4_2014 / Assets_Q4_2014 (Fast), or
(b) Avg (Earnings for last 4 or 8 or 20 quarters as of Q4 2014) / Avg (Assets for the same amount of quarters as of Q4 2014).

Also, he us using SLOW versions of the variables, and argues that quality does not change over 1 or 2 quarters. I.e., once quality company, it stays as a quality company for years.

I guess this prevents regular rebalances.

Implementing this strategy for a retail investor is next to impossible.
Last edited by TradingPlaces on Sat Jan 24, 2015 10:06 pm, edited 1 time in total.

TradingPlaces
Posts: 1245
Joined: Sun Nov 09, 2014 1:19 pm
Location: 30.286029, -97.530011

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by TradingPlaces » Sat Jan 24, 2015 10:05 pm

While Asness is very much on the same side as French and Fama, it is in his best interest to introduce complicated factors. E.g., size factor, or value factor, are so easy, that one can get ETFs tracking it. At the same time, once these factors are easily tracked, the outperformance disappears.

So Asness publishes research showing complicated factors that can only be implemented (tracked) by professional organizations. Rest assured, they also have methods to calculate superior versions of these factors.

The point of publishing this information is two-fold:

- there is a puritan, academic desire to educate everyone, and occasionally take bragging rights on discovering something, and proving some others wrong,

- marketing (see, we are discovering new things, so you should invest with us),

- severe self-interest. So, you publish a very real, but weak effect, like those factors. At the same time, you have better versions in house. Once the idea gets adopted and disseminated, essentially, trading your better factors is front-running all those who are taking your factors, and implementing themselves.

Company reports better than expected returns: BAM, stock jumps. How does the stock jump? Well, sophisticated investors get the EDGAR numbers, crunch their models, and determine how much the new price should be. It used to take days for the price discover to occur. Now, most of the price discovery occurs in the post-4m, and pre-9:30 am session. Of course, longer effects kick in, and price continues to adjust for a day or two, but the returns during that period are more modest (and in some cases due to secondary effects).

Leeraar
Posts: 4109
Joined: Tue Dec 10, 2013 8:41 pm
Location: Nowhere

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by Leeraar » Sun Jan 25, 2015 1:43 am

petfd:

People for the Ethical Treatment of Financial Data.

L.
You can get what you want, or you can just get old. (Billy Joel, "Vienna")

Caduceus
Posts: 1728
Joined: Mon Sep 17, 2012 1:47 am

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by Caduceus » Sun Jan 25, 2015 2:15 am

This is the problem with "quantitative" thinking rather than accounting-based valuation models. The numbers or ratios that you are getting are accounting artifacts, so the idea that they are comparable across 182 firms without some kind of severe qualification is ludicrous.

This is not an issue of whether you are averaging the data out over one year or many years. One firm might be electing a particularly aggressive depreciation schedule in an attempt to defer taxes. Two otherwise comparable firms might be electing different inventory cost methods, so that the firm using LIFO ends up with a lower estimated book value than the firm with FIFO during an inflationary period.

Qualitative factors are equally important. I just went through the accounting statements of a small-cap firm with such poor corporate governance and so many related-party transactions that I would not touch it with a ten-foot pole even if its numbers were accurate. I suspect the majority shareholder might be passing inventory through a related company and purchasing that inventory at a higher-than-normal price than would be justified in a true arms-length transaction. So do you invest even if all the ratios are nice?

Selecting stocks based on quantitative data, and fetishizing ratios and figures like they are natural quantities rather than accounting quantities defies logic.

richard
Posts: 7961
Joined: Tue Feb 20, 2007 3:38 pm
Contact:

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by richard » Sun Jan 25, 2015 6:33 am

TradingPlaces wrote:<>
1. Companies carefully compile their quarterly required info: (1) balance sheet, (2) cash-flow statement, (3) earnings statement.

2. This gets published to SEC EDGAR, and the whole world gets to see it at once,

3. Third party providers like Reuters or CapitalIQ immediately pick it up and within hours disseminate, via their standard data subscription (which costs money) to their clients, a structure summary of that report. E.g., all those variables, or something required to compute those variables will be there.<>
Sophisticated traders pick up the information immediately. They don't wait for the summary services. In fact, there were recent stories that some were getting filings seconds before others and were trading on the information. For example, http://www.wsj.com/articles/fast-trader ... 1414539997

richard
Posts: 7961
Joined: Tue Feb 20, 2007 3:38 pm
Contact:

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by richard » Sun Jan 25, 2015 6:43 am

nisiprius wrote:<>What does this say about the reliability of what the "financial science" community has been telling us since 1981? That community was talking about the small firm effect "in its normal form." So "everyone" used to say that the small-firm effect "in its normal form" was real. In fact, spectacular might be a fair characterization: in 1983, Rob Arnott wrote:
A growing body of evidence exists that small-capitalization stocks significantly outperform large-capitalization stocks. This effect is so strong and so consistent that even advocates of the Efficient Market Hypothesis have found no refutation of this effect.
Since then there have some various discoveries, including the finding of methodological flaws in the data series in which the effect was discovered.

What people ought to be saying is "What the financial 'science' community used to tell us has turned out to be wrong." But nobody is saying this. It is magician's misdirection. We will tell you something different now, give it the same name as something we used to say, and claim that we used to say is still true. For a while, the claim was that the small-firm effect was not very powerful by itself, but become powerful when combined with value. I am not quite sure what Asness is saying about small value, are you? Since he is so insistent about the important of quality, I assume that he feels that the effect value is also weak, but I can't tell.

The lesson I take from this is that it is it is hard to know how far to rely on what we are told by the financial 'science' community.
That's the issue. For many years, we've been told there's a small company premium and given risk and cost of capital rationales for this premium and that it holds up out of sample. Now we're being told, in effect, "oops".

This again raises the issue of data mining. A lot of finance research seems to consist of searching for interesting results and then coming up with rationales for the results. This is not how it's supposed to work. You're supposed to have a hypothesis and then test it.

garlandwhizzer
Posts: 2013
Joined: Fri Aug 06, 2010 3:42 pm

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by garlandwhizzer » Sun Jan 25, 2015 2:22 pm

It is a bit confusing for me at least. At first there was a small factor which was easily measured by market cap alone. Then there was a question about its existence apart from value, such that most or all of the value premium existed in small. The small premium in itself had become supposedly statistically insignificant. At least SCV was consistent although there were multiple ways to define the value part (PB, PE, PS, PD) and therefore it was a bit more slippery to capture. Sometimes one of these value parameters seemed to capture it better, sometimes another. In addition the value premium was later found to sometimes be in conflict with other more recently discovered premiums like momentum and quality. That added to the explanation about why it was sometimes so hard to capture the value premium.

Now we are told that actually in contrast to what we were recently told the small premium does clearly exist but only in small quality, not small junk. This poses an interesting question. One presumes that small junk has deeper value than small quality if the market has any rationality at all. Hence small cap junk would be expected to have a greater SCV premium than SCQ. Most recently we've been told that SCQ is real and SCJ should be avoided. We used to be told that SC growth, not SCJ, was the thing to avoid. Presumably SCG is the opposite of SCJ. So it appears at least superficially that the former black hole of investing (SCG) has morphed into its opposite (SCJ).

I for one am confused by the ever changing factor zoo as it is has been presented up to now. Existing factors seem to come into style and go out of style, sometimes even disappear unless they are combined with yet another factor. In addition new factors emerge such as momentum, quality, profitability, low volatility, and illiquidity. Often they conflict with each other such that sophisticated portfolio analysis is necessary to keep them from canceling each other out and also protect them from negative alpha which can likewise cancel them out. Rather than getting simpler over time factor investing seems to get more complicated and require more and more skill to hit the moving target. Another observation I will make is that almost all factors were much stronger and more significant before they were discovered than they are now that the overwhelming majority of market participants are full aware of them.

In spite of my confusion and skepticism, I cannot defeat my ever present desire to outperform the market and hence I tilt modestly. However I am not at all certain that a tilted portfolio will outperform a simple cap weighted portfolio in after cost risk adjusted returns going forward from here.

Garland Whizzer

User avatar
JoMoney
Posts: 6083
Joined: Tue Jul 23, 2013 5:31 am

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by JoMoney » Sun Jan 25, 2015 3:16 pm

garlandwhizzer wrote:...I for one am confused by the ever changing factor zoo as it is has been presented up to now. Existing factors seem to come into style and go out of style, sometimes even disappear unless they are combined with yet another factor. In addition new factors emerge such as momentum, quality, profitability, low volatility, and illiquidity. Often they conflict with each other such that sophisticated portfolio analysis is necessary to keep them from canceling each other out and also protect them from negative alpha which can likewise cancel them out. Rather than getting simpler over time factor investing seems to get more complicated and require more and more skill to hit the moving target. Another observation I will make is that almost all factors were much stronger and more significant before they were discovered than they are now that the overwhelming majority of market participants are full aware of them. ...
Apparently there are over 330 "factors" discussed in the academic literature. In addition to the "Why most financial research is likely false..."
The paper below claims they find there is multi-dimensionality to returns, and the popular "factors" are diminished further when measured controlling for the impact of other "factors" some of which have a much higher significance than the popular ones often discussed on this board...

The Remarkable Multidimensionality in the Cross-Section of Expected U.S. Stock Returns
http://papers.ssrn.com/sol3/papers.cfm? ... id=2262374
"To achieve satisfactory investment results is easier than most people realize; to achieve superior results is harder than it looks." - Benjamin Graham

Pizzasteve510
Posts: 635
Joined: Sun Jul 27, 2014 3:32 pm

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by Pizzasteve510 » Sun Jan 25, 2015 3:21 pm

Deleted
Last edited by Pizzasteve510 on Sat Feb 21, 2015 11:51 am, edited 1 time in total.

User avatar
nisiprius
Advisory Board
Posts: 36851
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: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by nisiprius » Sun Jan 25, 2015 3:39 pm

Another question: a) Are there enough high-quality small-cap stocks to go around? b) Can this be used as a long-only strategy, or is this a long-the-high-quality, short-the-low-quality deal?
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
JoMoney
Posts: 6083
Joined: Tue Jul 23, 2013 5:31 am

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by JoMoney » Sun Jan 25, 2015 3:57 pm

... and what is the economic risk/return story for "Quality"? I thought these "factors" were supposed to explain risk/returns within an Efficient Market Hypothesis paradigm. If the expectation is that these factors are behavioral why would you not expect the market to behave differently when these strategies are prominent ?
"To achieve satisfactory investment results is easier than most people realize; to achieve superior results is harder than it looks." - Benjamin Graham

User avatar
nedsaid
Posts: 10485
Joined: Fri Nov 23, 2012 12:33 pm

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by nedsaid » Sun Jan 25, 2015 3:58 pm

richard wrote:He isn't really saying the small-firm effect is real. He's essentially pushing the new quality factor: "In its normal form the small firm effect is, on a host of dimensions, weak to possibly nonexistent. Once adjusted for quality exposure it is real and spectacular."
I read the Asness article and my reaction is "huh?" I thought that where the value premium showed up was in the small-cap area of the market and not so much in the large cap area. Now we are being told that the small-cap premium is due to "quality". So wouldn't this mean that we should tilt our portfolios towards small cap growth? We would find less quality in the value section rather than in the growth section. I define quality in part by consistently increasing growth in earnings. I also think of a strong balance sheet as being a good indicator of quality.

I get what Dimensional Funds are doing with screening small value stocks for not only value characteristics but also to get the momentum factor from negative to neutral. I can see here what they are doing is screening out the "value traps" that is companies that are undervalued because in essence they can't be fixed. But it would seem to me that "quality" would get you into the growth area. Haven't we been told that small-cap growth is the black hole of investing?

I suppose what this boils down to is screening out the "lottery" stocks on both the value side and on the growth side. That is picking companies that have good prospects going forward. So I can see what Asness is getting at but I am beginning to wonder if the academics are overthinking things.
A fool and his money are good for business.

User avatar
nedsaid
Posts: 10485
Joined: Fri Nov 23, 2012 12:33 pm

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by nedsaid » Sun Jan 25, 2015 4:08 pm

JoMoney wrote:
garlandwhizzer wrote:...I for one am confused by the ever changing factor zoo as it is has been presented up to now. Existing factors seem to come into style and go out of style, sometimes even disappear unless they are combined with yet another factor. In addition new factors emerge such as momentum, quality, profitability, low volatility, and illiquidity. Often they conflict with each other such that sophisticated portfolio analysis is necessary to keep them from canceling each other out and also protect them from negative alpha which can likewise cancel them out. Rather than getting simpler over time factor investing seems to get more complicated and require more and more skill to hit the moving target. Another observation I will make is that almost all factors were much stronger and more significant before they were discovered than they are now that the overwhelming majority of market participants are full aware of them. ...
Apparently there are over 330 "factors" discussed in the academic literature. In addition to the "Why most financial research is likely false..."
The paper below claims they find there is multi-dimensionality to returns, and the popular "factors" are diminished further when measured controlling for the impact of other "factors" some of which have a much higher significance than the popular ones often discussed on this board...

The Remarkable Multidimensionality in the Cross-Section of Expected U.S. Stock Returns
http://papers.ssrn.com/sol3/papers.cfm? ... id=2262374
I suppose at the one extreme, one could pick the Total US Stock Market Index. Or do a very sophisticated screen to screen for all 330 factors and pick only one stock! Once we arrived at the promised land of index investing, the academics now have assured us that factor investing will beat the indexes. The problem is that as the academics apply more and more factors to "passive" investing the more that "passive" investing looks more like the hated "active" investing. Factors have been known for years and active funds have tried to capitalize. So what is old is new again, active management but with lower expense ratios and better trading strategies.
A fool and his money are good for business.

User avatar
JoMoney
Posts: 6083
Joined: Tue Jul 23, 2013 5:31 am

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by JoMoney » Sun Jan 25, 2015 4:21 pm

nedsaid wrote:... Once we arrived at the promised land of index investing, the academics now have assured us that factor investing will beat the indexes. The problem is that as the academics apply more and more factors to "passive" investing the more that "passive" investing looks more like the hated "active" investing. Factors have been known for years and active funds have tried to capitalize. So what is old is new again, active management but with lower expense ratios and better trading strategies.
I don't know that the academics promise you'll beat the index, it seems to be the mutual fund marketing people pushing the idea that you'll beat other funds using these strategies. The academics like Fama seem to push the idea that these factors represent a way to measure risk, such that it's not really "beating", it's just assuming different risks based on a different way of measuring it... I don't think I buy it though.
"To achieve satisfactory investment results is easier than most people realize; to achieve superior results is harder than it looks." - Benjamin Graham

Leeraar
Posts: 4109
Joined: Tue Dec 10, 2013 8:41 pm
Location: Nowhere

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by Leeraar » Sun Jan 25, 2015 4:44 pm

JoMoney wrote:
nedsaid wrote:... Once we arrived at the promised land of index investing, the academics now have assured us that factor investing will beat the indexes. The problem is that as the academics apply more and more factors to "passive" investing the more that "passive" investing looks more like the hated "active" investing. Factors have been known for years and active funds have tried to capitalize. So what is old is new again, active management but with lower expense ratios and better trading strategies.
I don't know that the academics promise you'll beat the index, it seems to be the mutual fund marketing people pushing the idea that you'll beat other funds using these strategies. The academics like Fama seem to push the idea that these factors represent a way to measure risk, such that it's not really "beating", it's just assuming different risks based on a different way of measuring it... I don't think I buy it though.
Really? And DFA does not exist?

The search for a new secret sauce continues. I think this is very much related to this other current active thread,

viewtopic.php?p=2346191#p2346191

It's high time to look at behavioral investing, not from the point of view of the foibles of individual investors, but from how the financial advice industry exploits their clients.

A couple of years ago, Rick Ferri posted a list of reasons why you should have a financial or investment advisor. That list was remarkable, because it did not include better returns, and was mostly about protecting clients from their own behavioral mistakes.

Excuse me, I need to now sign up for three free dinners with presentations that promise guaranteed market returns with no risk.

L.
You can get what you want, or you can just get old. (Billy Joel, "Vienna")

User avatar
nedsaid
Posts: 10485
Joined: Fri Nov 23, 2012 12:33 pm

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by nedsaid » Sun Jan 25, 2015 6:33 pm

The whole point of factor investing is to try to beat the market or perhaps achieve market return with less risk. I see factors as the "why" the market does what it does. In other words, what causes performance? Larry Swedroe has posted at length about the factors and that we should see diversification not so much as diversification across asset classes but across factors. I think the diversifying across factors argument makes sense. All the things I have tried as an investor have been an attempt to increase returns, increase diversification (whatever that means), and hopefully decrease volatility. I am not adding complexity for complexity's sake.

So if factors weren't about beating the market, DFA would not exist. They largely built their business on the small and value factors.

The problem is that the number of factors keep increasing. I wonder that if you try for more than two or three at the most at a time that all you achieve is a cancelling out effect. The other problem with all of this is that factors are essentially quantifying human behavior. We can very accurately measure the effects of human behavior in the past but that knowledge has very limited power to predict the future.

I believe in factors to the extent that they reflect human nature and human behavior. The danger of looking and looking at data is that you will see what you want to see. It is difficult to be truly objective. We should look at all of this with a bit of skepticism.
A fool and his money are good for business.

User avatar
Random Musings
Posts: 5213
Joined: Thu Feb 22, 2007 4:24 pm
Location: Pennsylvania

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by Random Musings » Sun Jan 25, 2015 8:47 pm

Terri Hatcher said this year's ago. Still holds true.

RM
I figure the odds be fifty-fifty I just might have something to say. FZ

User avatar
zaboomafoozarg
Posts: 1966
Joined: Sun Jun 12, 2011 12:34 pm

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by zaboomafoozarg » Sun Jan 25, 2015 9:29 pm

Wait, is this a Seinfeld reference? :D

User avatar
Robert T
Posts: 2532
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by Robert T » Sun Jan 25, 2015 9:43 pm

.
On the Asness et al paper - It seems they are saying that if you factor in 'quality' then the size premium was stronger, but I see little difference in saying if you factor in 'value' then the size premium was stronger. I think size does okay on its own. However, it seems avoiding small 'junk' improved returns, as did avoiding small growth - so not a new concept.

For example - using the RAFI data from 1976 to 2012 (as presented in this thread - viewtopic.php?f=10&t=126899 ), together with the quality-junk data from AQR website (QmJ 2x3 portfolios).

Annualized return - 1976 to 2012

AQR Small "neutral" = 14.7%
AQR Small "quality" = 17.0%
AQR Large "neutral" = 11.0%
AQR Large "quality" = 12.0%
--------------------------------
Small-large "neutral" = 3.7%
Small-large "quality = 5.0%

RAFI Small Value = 18.5%
RAFI Large = 13.3%
----------------------
Small-large = 5.2%

The RAFI (value tilt) small minus large was similar in size to AQR quality small minus large. If I had to choose between the two - I would go with value. Still uncomfortable with wide variability in quality metrics (e.g. correlation between Novy-Marx quality and Asness et al quality is about 0.5).

Robert
.
Last edited by Robert T on Mon Jan 26, 2015 3:01 am, edited 1 time in total.

User avatar
Ketawa
Posts: 1942
Joined: Mon Aug 22, 2011 1:11 am
Location: DC

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by Ketawa » Mon Jan 26, 2015 1:19 am

nisiprius wrote:Another question: a) Are there enough high-quality small-cap stocks to go around? b) Can this be used as a long-only strategy, or is this a long-the-high-quality, short-the-low-quality deal?
I believe "quality" is basically the same thing as "profitability" used in the AQR core equity funds. If I had to guess without bothering to do any real research into the matter, "quality minus junk" (QMJ) might be used to differentiate Asness' research from the Novy-Marx research on "profitable minus unprofitable" (PMU), while profitability is used in the fund marketing materials. QMJ and PMU are different, as Robert T pointed out; they only have 0.5 correlation. Who knows which is better.

The core equity funds target value, momentum, and profitability in one fund. IIRC, the sorts place approximately 40% importance on value, 40% on momentum, and 20% on profitability. They are long-only funds. QSMLX, the AQR Small Cap Core Equity Fund, had 628 holdings on 9/30/14. I don't hold it, but I might soon. I do hold QICLX (International Core Equity) and QEELX (Emerging Core Equity).

User avatar
Robert T
Posts: 2532
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by Robert T » Mon Jan 26, 2015 5:28 am

.
Ketawa wrote:The core equity funds target value, momentum, and profitability in one fund. IIRC, the sorts place approximately 40% importance on value, 40% on momentum, and 20% on profitability. They are long-only funds. QSMLX, the AQR Small Cap Core Equity Fund, had 628 holdings on 9/30/14.
Out of interest I did some due diligence on the AQR Core funds. They certainly look interesting, but the challenge of combining multiple factors in one strategy is to overcome costs of implementation. Time will tell if this is possible.

For example – here are the one-year turnover rates for the respective funds

AQR Core Equity = 277%
AQR US Small Core Equity = 189%
AQR International = 204%
AQR EM = 154%

Turnover incurs costs, higher in smaller caps, and EM. The paper “A New Core Equity Paradigm” estimated average annual trade costs for US large and small caps, from 1980-2012, to be 1.7% and 2.4% respectively (for average turnovers of 136% and 102%, lower that the above listed turnover so far). Estimated average returns, net of trade costs, for the value/quality/momentum series were lower than for the value only series (as below - from the paper). While the latter does not have trade costs subtracted, arguably these are much smaller. The value only series over this time period has similar returns to the RAFI value series. My conclusion from the US simulated series is that, on an after trade costs basis, the value/momentum/quality returns were no higher than a value only strategy (perhaps lower, certainly after taxes), although volatility was lower with the combination of the three factors. While the RAFI series had similar return/volatility as the ‘value only’ simulated series on the US side, they lagged in international.

Will be interesting to see performance. Obviously no guarantees.

1980-2012 – average [not annualized] returns

Large caps
17.3%/16.6 = Value/Momentum/Profitability (before trade costs – 136% turnover)
15.6%/ - = Value/Momentum/Profitability (after notionally estimated trade costs)
16.7%/18.1 = Value Only (Multiple Metric Sort) (before trade costs)
--------
16.6%/18.7 = RAFI Fundamental Midcap Value

Small caps
20.7%/19.9 = Value/Momentum/Profitability (before trade costs – 102% turnover)
18.3%/ - = Value/Momentum/Profitability (after notionally estimated trade costs)
18.8%/21.7 = Value Only (Multiple Metric Sort) (before trade costs)
--------
19.2%/22.9 = RAFI Fundamental Small Value

1990-2012 – average [not annualized] returns

International
11.2%/15.7 = Value/Momentum/Profitability (before trade costs – 125% turnover)
10.3%/ - = Value/Momentum/Profitability (after notionally estimated trade costs)
11.9%/18.7 = Value Only (Multiple Metric Sort) (before trade costs)
--------------
9.4%/21.8 = FTSE RAFI Developed ex US
10.2%/23.6 = FTSE RAFI Developed ex US Mid-Small

Robert
.

ScottW
Posts: 154
Joined: Tue Mar 25, 2008 8:13 am

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by ScottW » Mon Jan 26, 2015 9:06 am

I'd like to resurrect a post that SmallHi made close to 7 years ago. For those who joined after he disappeared from the board, he was a champion of both small and value.

The post: http://www.bogleheads.org/forum/viewtop ... 79#p213179
There hasn't been a microcap benefit (relative to small or mid) for over 60 years. Consider the following annualized returns from 1948-2007:

CRSP 1-2 (Mega Caps) = +11.4%
CRSP 3-5 (Mid Caps) = +13.1%
CRSP 6-8 (Small Caps) = +13.3%
CRSP 9-10 (Micro Caps) = +13.1%

To make matters worse, the only "true" micro cap index (Russell Micro Cap Index) suffers terribly from return drag (its trailed the CRSP 9-10 by over 300 bps annually since its 2000 inception -- not including ETF fees), and BRSIX has fairly large tracking error and is pretty concentrated.

I think most investors who wish to capture the small cap premium are better off underweighting mega cap stocks than overweighting micro cap stocks...

just my 2 cents.
This is more or less the strategy I've been following for the past 6 years. I never jumped into any of the "new and improved" fundamental indexes because I don't trust any of the providers, and the asset base for many of the available ETFs is small enough that I don't have much confidence they won't close or merge with other funds at some point. I've stuck primarily with Vanguard, and since they don't have strong value or size tilts, I've simple de-emphasized mega-caps and gone heavy on their value funds (which for Vanguard, amounts to having mild value tilt).

CliffA
Posts: 7
Joined: Tue Dec 03, 2013 1:07 am

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by CliffA » Mon Jan 26, 2015 12:16 pm

Thanks all.

I’m going to try to respond to some of the above in aggregate and in no particular order. Please lease read this old and a bit dated disclaimer from a guest blog I did for Joe Nocera at the New York Times http://graphics8.nytimes.com/packages/p ... a-blog.pdf and act accordingly (assume I’m lying and insane, etc.). There are lots of good comments above and, as this is a first draft and we can incorporate them, I really appreciate the input and wanted to respond.

First, to be clear, there are two things we've put out. One is a separate blog entry by me summarizing the new paper, and then the new paper itself. And yes, the blog entry title was a Seinfeld reference. It was also an innuendo laced blog title referencing a separately innuendo laced paper title. I'm both ashamed and proud of that rare double... But, clearly I’m getting too old, as I thought more people would remember Seinfeld! The blog title just sounds too arrogant without knowing that reference and I’m sorry if some took it that way. I sometimes stupidly forget the timing of these things. Seinfeld ended in May of 1998. No LTCM, no Russian debt crisis, no tech bubble, no real estate bubble and crash, no GFC, no quantitative easing. Damn the world was young...

Moving on, as some noted, yes, we have not articulated a long only product based on this work. Frankly, we're not sure it will work after costs (it may - I really mean that we’re not sure yet!). We’re working on that issue right now. But there is a place for pure research even if a successful after trading cost product proves difficult (again, we are not saying that either, still working on it!). A factor model improvement can be useful for understanding how markets work whether an after cost long only product is great or not. Stay tuned.

But, one thing I have to point out - Ironically some seem to imply the paper is just marketing a product, but others critique us for not proving (or even arguing) that such a product is feasible. Either critique is fair, though I'd argue with both (pure research is valuable, and presenting a possible product is fair if you're honest about it). But it's not fair to critique us for presenting a product and for not presenting a product! But, in a general sense (as one entry mentioned above) I’d readily admit to this nefarious plan: We do think if we say things people find smart and informative, over time, it will help our business. Of course the rub is if we say the opposite... Yes, it’s an evil plot but now that we’ve been found out the Earth is safe once more :)

Some above looked at small cap investing over 1997-2014. 17-18 years is a lifetime to live through in investing but sadly is still way too short to test for factor performance. Though I will note our paper calls the post-1999 time period the “resurrection” for the pure (no adjustment for quality) size effect, so at least our stuff jibes with small alone working over this period (though it’s still worked better if less exposed to junk even over this “resurrection” period - at least in our long-short factor construction the same act of adding QMJ to the right-hand-side shows small to be short QMJ (i.e., short quality long junk) and the intercept goes up over 1997-2014 when QMJ is added). Also, without getting bogged down, remember too that 1997-2014 is still (with some wild ups and downs!) a pretty decent period for the market as a whole, and regular straight small investing still has a market beta above one, so adjusting for this small looks ok over this period but not quite as great as the raw numbers themselves (remember, if all you get out of small is a higher beta, it's not value added, you could've gotten it just with a bit more regular old equities).

Yes to something said a few times above, indeed we are not resurrecting a pure size effect using only size to sort stocks. We are arguing that marginal to quality (or controlling for quality) there is a strong and consistent (for factor returns) size-return relation and we think that’s a pretty important stylized fact for future theory to have to explain, but the size-return relation is only strong marginal to quality not alone (in fact that's our point!). Admittedly the blog title didn’t explain that but I think it's very clear in context (and it is directly implied in the paper title).

I don't get all the discussion of "secret sauce." This has more moving parts than a single factor or simpler tilt but is all disciplined rules based stuff. It might be more “complex” sauce, and indeed may be difficult to implement with simple screening software at home, but that’s a matter of effort and time not possibility. Reading a couple of our and others' papers you can replicate everything we’ve done. There is no judgment/skill on the fly here (whether you like that or not) only in the construction of the rules and tests.

A few question our quality composites from our QMJ paper (e.g., noting our PROFIT composite is not super highly correlated with Novy-Marx PROFITABLITY, something around 0.50). This is why we also test whether the Fama-French 5-factor model, and a few other quality proxies not from QMJ, show analogous results to adjusting for QMJ (they do). Of course we favor our own version of quality (or else we’d change it!) but our findings are not at all dependent on them.

The bottom line: over long histories, and most short-ones, small has beaten large, but not by a lot, and has been held back from a statistical and economically large victory by being very exposed to “junk vs. quality." If you control for that, all-else-equal small beats large with much more consistency (and comparable to some other famous factors). Whether this is simply an important result for theory and academics to explain, or indeed leads to a better way to invest directly in small caps, still remains to be seen (the practicality question). I hope me telling you that “we don’t yet know if we or you will actually make money from this but we think it’s pretty darn interesting on its own” is honest enough!

Again, thanks for the comments, this kind of input will make the next version of the paper better.

- Cliff Asness

exeunt
Posts: 907
Joined: Tue Jan 08, 2008 6:54 pm

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by exeunt » Mon Jan 26, 2015 1:18 pm

This is awesome. I love how Cliff will comment almost anywhere.

Is it fair to describe the revived small-firm effect as a quality non-linearity effect? A small quality firm that has similar metrics as a large quality firm probably has a lot more runway to invest incremental capital at above-market rate of returns and this potential likely isn't fully appreciated by the market. I'm thinking Wal-Mart during the 70s and 80s.

CliffA
Posts: 7
Joined: Tue Dec 03, 2013 1:07 am

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by CliffA » Mon Jan 26, 2015 4:02 pm

Oh, and while I won't address a specific product in depth (I get yelled at for that!), some of the turnover numbers above seem very odd and way too large. They are affected by inflows, especially when the funds are smaller, and how a fund like ours equitizes inflows - they are not regular, steady state turnover in the normal sense. Nobody would deny (and our papers, as cited above by others, show) that giving momentum a bigger role will raise turnover, but we don't think it's close to as much as quoted above (and, obviously, we believe it will raise net returns, not just gross returns, over the long-term).

Also, not to preach, but those who say they will watch performance to make a decision had best be prepared to watch for a very long time (which is fine by me), as we don't learn a lot about factor premiums on any time scale you usually achieve by watching live! You have to look at very long histories, come to your own conclusions about gross premia going forward (do you believe those long histories apply, or more relevant, apply enough?) and what's achievable after trading costs, and your conclusions certainly don't have to agree with mine. But neither of us will prove anything by "watching" for a while - even if it (wrongly) feels like a long while, sometimes an eternity! Note, I still watch it every day, most of the day, and my mood is still affected by it. Thus I'm a time-scale hypocrite...

-- Cliff

CliffA
Posts: 7
Joined: Tue Dec 03, 2013 1:07 am

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by CliffA » Mon Jan 26, 2015 4:08 pm

exeunt wrote:This is awesome. I love how Cliff will comment almost anywhere.

Is it fair to describe the revived small-firm effect as a quality non-linearity effect? A small quality firm that has similar metrics as a large quality firm probably has a lot more runway to invest incremental capital at above-market rate of returns and this potential likely isn't fully appreciated by the market. I'm thinking Wal-Mart during the 70s and 80s.
Yes, my letter to the National Enquirer just got printed last month. Took them a while to admit that the whole Jennifer Aniston / Alien Abduction thing could be explained by the 5-factor model.

It may be non-linear, we have not really looked for that, interesting. But I'd describe it even more simply. If you just sort on size and look at the small stocks you get a disproportionate amount of junky companies (and/or the small-junky companies are even junkier than large-junky companies). Since junk underperforms quality this is a drag on small's performance. I wish it were more complicated!

exeunt
Posts: 907
Joined: Tue Jan 08, 2008 6:54 pm

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by exeunt » Mon Jan 26, 2015 4:24 pm

CliffA wrote:
exeunt wrote:....
Yes, my letter to the National Enquirer just got printed last month. Took them a while to admit that the whole Jennifer Aniston / Alien Abduction thing could be explained by the 5-factor model.

It may be non-linear, we have not really looked for that, interesting. But I'd describe it even more simply. If you just sort on size and look at the small stocks you get a disproportionate amount of junky companies (and/or the small-junky companies are even junkier than large-junky companies). Since junk underperforms quality this is a drag on small's performance. I wish it were more complicated!
My favorite was when you bashed DeLong in the comments section of one of his Project Syndicate columns.

The reason I came up with the quality non-linearity story is because I'm struggling to understand why size itself would be a source of excess return that's seemingly unrelated to numerous plausible proxies for risk. If size itself is rewarded, I would expect to see it everywhere, as with value and momentum. Is there a size effect in bonds? Commodities? Currencies? The failure for size to naturally appear out of different data sets suggests to me the effect you've identified is logically related to quality stocks and not size itself.

Leeraar
Posts: 4109
Joined: Tue Dec 10, 2013 8:41 pm
Location: Nowhere

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by Leeraar » Mon Jan 26, 2015 4:45 pm

CliffA wrote:Oh, and while I won't address a specific product in depth (I get yelled at for that!), some of the turnover numbers above seem very odd and way too large. They are affected by inflows, especially when the funds are smaller, and how a fund like ours equitizes inflows - they are not regular, steady state turnover in the normal sense. Nobody would deny (and our papers, as cited above by others, show) that giving momentum a bigger role will raise turnover, but we don't think it's close to as much as quoted above (and, obviously, we believe it will raise net returns, not just gross returns, over the long-term).

Also, not to preach, but those who say they will watch performance to make a decision had best be prepared to watch for a very long time (which is fine by me), as we don't learn a lot about factor premiums on any time scale you usually achieve by watching live! You have to look at very long histories, come to your own conclusions about gross premia going forward (do you believe those long histories apply, or more relevant, apply enough?) and what's achievable after trading costs, and your conclusions certainly don't have to agree with mine. But neither of us will prove anything by "watching" for a while - even if it (wrongly) feels like a long while, sometimes an eternity! Note, I still watch it every day, most of the day, and my mood is still affected by it. Thus I'm a time-scale hypocrite...

-- Cliff
Did I just learn that I don't know something I thought I knew?

I assumed "turnover" is independent of inflows and outflows. No?

And by the way, I don't accept the idea that if the factor does not show up, it's because I didn't wait long enough.

L.
You can get what you want, or you can just get old. (Billy Joel, "Vienna")

CliffA
Posts: 7
Joined: Tue Dec 03, 2013 1:07 am

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by CliffA » Mon Jan 26, 2015 5:59 pm

Leeraar wrote:
CliffA wrote:Oh, and while I won't address a specific product in depth (I get yelled at for that!), some of the turnover numbers above seem very odd and way too large. They are affected by inflows, especially when the funds are smaller, and how a fund like ours equitizes inflows - they are not regular, steady state turnover in the normal sense. Nobody would deny (and our papers, as cited above by others, show) that giving momentum a bigger role will raise turnover, but we don't think it's close to as much as quoted above (and, obviously, we believe it will raise net returns, not just gross returns, over the long-term).

Also, not to preach, but those who say they will watch performance to make a decision had best be prepared to watch for a very long time (which is fine by me), as we don't learn a lot about factor premiums on any time scale you usually achieve by watching live! You have to look at very long histories, come to your own conclusions about gross premia going forward (do you believe those long histories apply, or more relevant, apply enough?) and what's achievable after trading costs, and your conclusions certainly don't have to agree with mine. But neither of us will prove anything by "watching" for a while - even if it (wrongly) feels like a long while, sometimes an eternity! Note, I still watch it every day, most of the day, and my mood is still affected by it. Thus I'm a time-scale hypocrite...

-- Cliff
Did I just learn that I don't know something I thought I knew?

I assumed "turnover" is independent of inflows and outflows. No?

And by the way, I don't accept the idea that if the factor does not show up, it's because I didn't wait long enough.

L.
Alas, sadly neither your acceptance, nor for that matter mine, changes the math by the tiniest fraction :)

User avatar
nedsaid
Posts: 10485
Joined: Fri Nov 23, 2012 12:33 pm

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by nedsaid » Mon Jan 26, 2015 6:38 pm

To Cliff Asness,

Thanks for posting on this forum. It is fun to know that investment experts actually read the stuff that rank amateurs like us post. It is also good to know that you have a great sense of humor.

Don't feel too bad about getting beat up with foam rubber hammers on this forum. The Bogleheads are a pretty opinionated bunch and are "equal opportunity bashers." Larry Swedroe gets bashed here too but he gives as good as he gets. I have learned so much from his many posts on this forum. I have to admit I laughed through your post, humor does a good job of defusing critics.

Part of what you experienced is the frustration from investors. I went to a Paul Merriman seminar a few years ago presented by Tom Cock, a financial media guy in the Seattle, Washington area. It was a great couple of seminars and I learned a whole lot. As Tony Robbins would say, "I was excited and ready to take massive action." Later on, the Vanguard Small Cap Value EFT that I bought was roundly criticized for not having good enough small or value characteristics. Then I learned from Morningstar that the Micro-Cap Index ETF underperformed even the small cap indexes because of front running by traders, the expense ratio, and trading costs. It was like, "Oh man!! Now they tell me!"

The Bogleheads, if anything are a conscientious group and are pretty good at following instructions. So a bunch of us followed the recommendations and did our small-cap tilting without the DFA products or the DFA advisors. I am sure that I am not the only one expressing frustration that the investment products that I bought to capture the small and value premiums are not good enough. It gets even more frustrating when you find out the small premium might be all about quality and not value.

When Larry Swedroe comes off his Boglehead sabbatical, I would be very interested in his comments about your blog and white paper. I know he very much respects your work.

Again, thanks for posting. It is nice to know that people in the investment community actually read our stuff.

Best wishes,

Ned
Last edited by nedsaid on Mon Jan 26, 2015 6:50 pm, edited 1 time in total.
A fool and his money are good for business.

Leeraar
Posts: 4109
Joined: Tue Dec 10, 2013 8:41 pm
Location: Nowhere

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by Leeraar » Mon Jan 26, 2015 6:40 pm

CliffA wrote:
Leeraar wrote:
CliffA wrote:Oh, and while I won't address a specific product in depth (I get yelled at for that!), some of the turnover numbers above seem very odd and way too large. They are affected by inflows, especially when the funds are smaller, and how a fund like ours equitizes inflows - they are not regular, steady state turnover in the normal sense. Nobody would deny (and our papers, as cited above by others, show) that giving momentum a bigger role will raise turnover, but we don't think it's close to as much as quoted above (and, obviously, we believe it will raise net returns, not just gross returns, over the long-term).

Also, not to preach, but those who say they will watch performance to make a decision had best be prepared to watch for a very long time (which is fine by me), as we don't learn a lot about factor premiums on any time scale you usually achieve by watching live! You have to look at very long histories, come to your own conclusions about gross premia going forward (do you believe those long histories apply, or more relevant, apply enough?) and what's achievable after trading costs, and your conclusions certainly don't have to agree with mine. But neither of us will prove anything by "watching" for a while - even if it (wrongly) feels like a long while, sometimes an eternity! Note, I still watch it every day, most of the day, and my mood is still affected by it. Thus I'm a time-scale hypocrite...

-- Cliff
Did I just learn that I don't know something I thought I knew?

I assumed "turnover" is independent of inflows and outflows. No?

And by the way, I don't accept the idea that if the factor does not show up, it's because I didn't wait long enough.

L.
Alas, sadly neither your acceptance, nor for that matter mine, changes the math by the tiniest fraction :)
I agree, the "math" is exact. It's the theory that is missing.

L.
You can get what you want, or you can just get old. (Billy Joel, "Vienna")

User avatar
Ketawa
Posts: 1942
Joined: Mon Aug 22, 2011 1:11 am
Location: DC

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by Ketawa » Mon Jan 26, 2015 6:55 pm

Leeraar wrote:
CliffA wrote:Oh, and while I won't address a specific product in depth (I get yelled at for that!), some of the turnover numbers above seem very odd and way too large. They are affected by inflows, especially when the funds are smaller, and how a fund like ours equitizes inflows - they are not regular, steady state turnover in the normal sense. Nobody would deny (and our papers, as cited above by others, show) that giving momentum a bigger role will raise turnover, but we don't think it's close to as much as quoted above (and, obviously, we believe it will raise net returns, not just gross returns, over the long-term).
Did I just learn that I don't know something I thought I knew?

I assumed "turnover" is independent of inflows and outflows. No?
The turnover numbers come from the September 2014 Annual Report, pages 172-175.

I also assumed that inflows and outflows would not affect the reported turnover. Here is how the SEC says to calculate turnover. It appears to be based on total assets, so inflows and outflows would have a large effect. Paragraph 13.(a).4.(d) here: http://www.sec.gov/about/forms/formn-1a.pdf
(d) Calculate the Portfolio Turnover Rate as follows:
(i) Divide the lesser of amounts of purchases or sales of portfolio securities for the fiscal year by the monthly average of the value of the portfolio securities owned by the Fund during the fiscal year. Calculate the monthly average by totaling the values of portfolio securities as of the beginning and end of the first month of the fiscal year and as of the end of each of the succeeding 11 months and dividing the sum by 13.
(ii) Exclude from both the numerator and the denominator amounts relating to all securities, including options, whose maturities or expiration dates at the time of acquisition were one year or less. Include all long-term securities, including long-term U.S. Government securities. Purchases include any cash paid upon the conversion of one portfolio security into another and the cost of rights or warrants. Sales include net proceeds of the sale of rights and warrants and net proceeds of portfolio securities that have been called or for which payment has been made through redemption or maturity.
(iii) If the Fund acquired the assets of another investment company or of a personal holding company in exchange for its own shares during the fiscal year in a purchase‑of‑assets transaction, exclude the value of securities acquired from purchases and securities sold from sales to realign the Fund’s portfolio. Adjust the denominator of the portfolio turnover computation to reflect these excluded purchases and sales and disclose them in a footnote.
(iv) Include in purchases and sales any short sales that the Fund intends to maintain for more than one year and put and call options with expiration dates more than one year from the date of acquisition. Include proceeds from a short sale in the value of the portfolio securities sold during the period; include the cost of covering a short sale in the value of portfolio securities purchased during the period. Include premiums paid to purchase options in the value of portfolio securities purchased during the reporting period; include premiums received from the sale of options in the value of the portfolio securities sold during the period.

Leeraar
Posts: 4109
Joined: Tue Dec 10, 2013 8:41 pm
Location: Nowhere

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by Leeraar » Mon Jan 26, 2015 7:06 pm

Ketawa wrote:
Leeraar wrote:
CliffA wrote:Oh, and while I won't address a specific product in depth (I get yelled at for that!), some of the turnover numbers above seem very odd and way too large. They are affected by inflows, especially when the funds are smaller, and how a fund like ours equitizes inflows - they are not regular, steady state turnover in the normal sense. Nobody would deny (and our papers, as cited above by others, show) that giving momentum a bigger role will raise turnover, but we don't think it's close to as much as quoted above (and, obviously, we believe it will raise net returns, not just gross returns, over the long-term).
Did I just learn that I don't know something I thought I knew?

I assumed "turnover" is independent of inflows and outflows. No?
The turnover numbers come from the September 2014 Annual Report, pages 172-175.

I also assumed that inflows and outflows would not affect the reported turnover. Here is how the SEC says to calculate turnover. It appears to be based on total assets, so inflows and outflows would have a large effect. Paragraph 13.(a).4.(d) here: http://www.sec.gov/about/forms/formn-1a.pdf
(d) Calculate the Portfolio Turnover Rate as follows:
(i) Divide the lesser of amounts of purchases or sales of portfolio securities for the fiscal year by the monthly average of the value of the portfolio securities owned by the Fund during the fiscal year. Calculate the monthly average by totaling the values of portfolio securities as of the beginning and end of the first month of the fiscal year and as of the end of each of the succeeding 11 months and dividing the sum by 13.
(ii) Exclude from both the numerator and the denominator amounts relating to all securities, including options, whose maturities or expiration dates at the time of acquisition were one year or less. Include all long-term securities, including long-term U.S. Government securities. Purchases include any cash paid upon the conversion of one portfolio security into another and the cost of rights or warrants. Sales include net proceeds of the sale of rights and warrants and net proceeds of portfolio securities that have been called or for which payment has been made through redemption or maturity.
(iii) If the Fund acquired the assets of another investment company or of a personal holding company in exchange for its own shares during the fiscal year in a purchase‑of‑assets transaction, exclude the value of securities acquired from purchases and securities sold from sales to realign the Fund’s portfolio. Adjust the denominator of the portfolio turnover computation to reflect these excluded purchases and sales and disclose them in a footnote.
(iv) Include in purchases and sales any short sales that the Fund intends to maintain for more than one year and put and call options with expiration dates more than one year from the date of acquisition. Include proceeds from a short sale in the value of the portfolio securities sold during the period; include the cost of covering a short sale in the value of portfolio securities purchased during the period. Include premiums paid to purchase options in the value of portfolio securities purchased during the reporting period; include premiums received from the sale of options in the value of the portfolio securities sold during the period.
Thank you. I had assumed that "turnover" was a simple measure of churning the account. I will have to study what you sent. Always, something new to learn!

Thank you,

L.
You can get what you want, or you can just get old. (Billy Joel, "Vienna")

User avatar
matjen
Posts: 1965
Joined: Sun Nov 20, 2011 11:30 pm

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by matjen » Mon Jan 26, 2015 8:39 pm

Cliff thank you for chiming in and rounding out your blog and paper. Greatly appreciated!
A man is rich in proportion to the number of things he can afford to let alone.

CliffA
Posts: 7
Joined: Tue Dec 03, 2013 1:07 am

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by CliffA » Mon Jan 26, 2015 8:46 pm

matjen wrote:Cliff thank you for chiming in and rounding out your blog and paper. Greatly appreciated!
Like I can help myself :)

Thanks all.

-- C

User avatar
matjen
Posts: 1965
Joined: Sun Nov 20, 2011 11:30 pm

Re: Asness - The Small-Firm Effect Is Real, and It’s Spectac

Post by matjen » Mon Jan 26, 2015 8:52 pm

CliffA wrote:
matjen wrote:Cliff thank you for chiming in and rounding out your blog and paper. Greatly appreciated!
Like I can help myself :)

Thanks all.

-- C
That's what I admire about you. You are a scrappy fellow who will defend his opinion to anyone...even us retail investors. Well done! I know you may shy away from a particular product but I would invite you or one of your peeps (Antti is always welcome!) to chime in on our lengthy QSPIX thread. Many are watching it and a few have invested...happily so I might add. :sharebeer

viewtopic.php?f=10&t=151111
A man is rich in proportion to the number of things he can afford to let alone.

Post Reply