The new profitability factor, US and int'l evidence

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.

Re: The new profitability factor, US and int'l evidence

Postby larryswedroe » Sat May 18, 2013 9:05 am

Swaption
The alpha definition has nothing to do with DFA or any firm. It's what left, the unexplained portion, from the regression. It might be skill for example, or it might be explained by another variable, not explained by the factors used in the regression.

I already provided several explanations of why this "anomaly" exists, both risk and mispricing.

As to whether it persists, if you agree with the risk story then you should expect it to persist. If you believe the behavioral story then you still have the simple pricing story, as long as profitability persists (profitability predicts growth) then it's simple math. You don't need any risk story or behavioral story. It's for same valuation you have higher profits and thus higher expected returns.

Best wishes
Larry
larryswedroe
 
Posts: 11201
Joined: 22 Feb 2007
Location: St Louis MO

Re: The new profitability factor, US and int'l evidence

Postby baw703916 » Sat May 18, 2013 7:30 pm

Isn't this saying that P/E should be the favored valuation metric?

1/PE = earnings/market = earnings/book * book/market = "profitability" * Fama-French value factor

Brad
Most of my posts assume no behavioral errors.
User avatar
baw703916
 
Posts: 5332
Joined: 1 Apr 2007
Location: Northern Virginia

Re: The new profitability factor, US and int'l evidence

Postby larryswedroe » Tue May 21, 2013 9:19 pm

Brad
No I would not draw that conclusion
If anything it would be price to CASH FLOW, not earnings. You want to move up the income statement avoiding things like accruals.

Also what works on paper may not work in practice if you have high turnover as result. FF chose BtM because it is more stable, resulting in lower turnover and higher tax efficiency.

But it does look like there is very persuasive evidence that combining a value and profitability metric adds significant value. Again suggest reading both the Novy Marx paper and the book Quantitative Value--a very good book I'll write some blog posts on, coming up

Best wishes
Larry
larryswedroe
 
Posts: 11201
Joined: 22 Feb 2007
Location: St Louis MO

Re: The new profitability factor, US and int'l evidence

Postby rmelvey » Tue May 21, 2013 9:51 pm

Brad,

I think I see what you are getting at. Let me know if I am wrong...

Holding P/E constant, a lower P/B implies a lower ROE. Holding P/E constant, a higher P/B implies a higher ROE. So, by looking for stocks with a good combination of low P/B and high ROE, you are essentially just getting low P/E stocks. P/E is the interaction of P/B and ROE.

This why I think Joel Greenblatt's "Magic Formula" is superior. It strictly looks at valuation of the firm as a going concern. It only looks at EV/EBIT, and EBIT/Operating Capital. Book value has nothing to do with his formula. The EV/EBIT indicates cheapness, and the EBIT/Operating Capital indicates quality (or organic growth potential).
User avatar
rmelvey
 
Posts: 402
Joined: 18 Sep 2010

Re: The new profitability factor, US and int'l evidence

Postby larryswedroe » Tue May 21, 2013 11:07 pm

rmelvey
Suggest you read Quantitative Value, which looks closely at Greenblatt's formula


Larry
larryswedroe
 
Posts: 11201
Joined: 22 Feb 2007
Location: St Louis MO

Re: The new profitability factor, US and int'l evidence

Postby rmelvey » Tue May 21, 2013 11:14 pm

I will add it to my reading list :happy Thanks Larry.
User avatar
rmelvey
 
Posts: 402
Joined: 18 Sep 2010

Re: The new profitability factor, US and int'l evidence

Postby larryswedroe » Wed May 22, 2013 8:30 am

rmelvey
Holding P/E constant, a lower P/B implies a lower ROE.


Don't think this is true.
Book value is historical long-term while p/e is current. It MIGHT be true over the long term but untrue today.

Larry
larryswedroe
 
Posts: 11201
Joined: 22 Feb 2007
Location: St Louis MO

Re: The new profitability factor, US and int'l evidence

Postby Hexdump » Wed May 22, 2013 9:19 am

Clearly_Irrational wrote:
OverTheHill wrote:First, if a person believes this [nonsense --admin LadyGeek] and follows it, then they apparently are guaranteed to profit better than anyone else on earth (even after paying fees). Second, if guaranteed isn't the right word, then something higher than 98.6% must be guaranteed. [Snarky comment removed by admin LadyGeek]


I think you're over-reacting and being excessively harsh. Of course Larry is talking his book, but that doesn't mean he's wrong. Frankly I'm rather interested in the findings and I'd like to see more data and evidence. Currently it's not applicable to my portfolio in a low cost way but I'd be open to the idea of making some tweaks if it turns out the research is solid. We know that the basic FF model explains somewhere in the 90%-ish range of results, but I'd be happy to embrace a model that explains more just as I was happy to switch to FF from CAPM. I'm not going to jump on the bandwagon until I see a lot more information but classifying it as marketing hype seems a bit too dismissive.

My personal interpretation is that beta and small are risk stories while value and momentum are behavioral. Profitability still needs explaining (risk or behavioral) and I'd want some kind of confirmation that it's unlikely to be arbitraged away before I started using it. Currently I don't use momentum directly, but I feel I like I get some exposure through the use of rebalancing bands.


Clearly_Irrational: I too am interesting in this as well as this discussion.
OverTheHill: You've made your point, [rude comment removed by admin LadyGeek]
Larry: Please ignore OverTheHill
User avatar
Hexdump
 
Posts: 1377
Joined: 10 Oct 2008
Location: Houston, Texas

Re: The new profitability factor, US and int'l evidence

Postby Hexdump » Wed May 22, 2013 9:28 am

larryswedroe wrote:clearly irrational
Of course Larry is talking his book


What I don't understand is why is anyone so surprised that with all the manpower and computer power and profit incentive that we would not eventually come up with a better model, just as FF did.

I hope that is helpful

Best wishes
Larry


It is helpful Larry, and what surprises me is that given all that firepower, it took so long.
User avatar
Hexdump
 
Posts: 1377
Joined: 10 Oct 2008
Location: Houston, Texas

Re: The new profitability factor, US and int'l evidence

Postby larryswedroe » Wed May 22, 2013 11:11 am

hexdump
I agree. The key was to move UP the income statement to gross profitability--not using earnings.
Larry
larryswedroe
 
Posts: 11201
Joined: 22 Feb 2007
Location: St Louis MO

Re: The new profitability factor, US and int'l evidence

Postby edge » Wed May 22, 2013 1:14 pm

Larry,

Thank you for your patient responses. I own many DFA funds (EM Value, Intl Vector, Global RE, Target Value), it seems like these additional factors will be factored into the fund management. IIRC MOM already was.
edge
 
Posts: 1799
Joined: 19 Feb 2007
Location: Great Falls VA

Re: The new profitability factor, US and int'l evidence

Postby rmelvey » Wed May 22, 2013 2:28 pm

larryswedroe wrote:rmelvey
Holding P/E constant, a lower P/B implies a lower ROE.


Don't think this is true.
Book value is historical long-term while p/e is current. It MIGHT be true over the long term but untrue today.

Larry


Hi Larry,

I am talking about the algebra of the metrics when looking at a single firm.

ROE = E/B.

So P/E = (B/E) * (P/B)

Fundamentally, P/E ratio already reconciles the desire for high ROE and low P/B. It is the perfect distillation of the two metrics. So building a screener or ranking system that looks for both low P/B and high ROE, is kind of silly because P/E already perfectly does that. I know that the researchers moved higher up in the income statement, but the logic still stands true.

I think looking for low P/E and high ROE is more internally consistent (or using similar metrics higher up on the income statement). With this type of screener you are strictly valuing the firm as a going concern, caring about flows accuring to you. So minimal capital backing up the firm implies that it would take little capital to continue to grow the firm, making it attractive to you as an investor.
User avatar
rmelvey
 
Posts: 402
Joined: 18 Sep 2010

Re: The new profitability factor, US and int'l evidence

Postby larryswedroe » Wed May 22, 2013 2:38 pm

rmelvey
What I can tell you is that if you screen for value first and then GROSS profitability you get signficantly higher returns than if you just screen for p/E,
If it was just p/e we would not be talking about a profitability factor.
Larry
larryswedroe
 
Posts: 11201
Joined: 22 Feb 2007
Location: St Louis MO

Re: The new profitability factor, US and int'l evidence

Postby rmelvey » Wed May 22, 2013 2:52 pm

larryswedroe wrote:rmelvey
What I can tell you is that if you screen for value first and then GROSS profitability you get signficantly higher returns than if you just screen for p/E,
If it was just p/e we would not be talking about a profitability factor.
Larry


Well, it would be interesting then to see a screen of just P/GP. Did my earlier logic make sense though?
User avatar
rmelvey
 
Posts: 402
Joined: 18 Sep 2010

Re: The new profitability factor, US and int'l evidence

Postby larryswedroe » Wed May 22, 2013 5:35 pm

rmelvey
What I can say is that some very smart people have looked at this and they (many independently) concluded that you need multiple screens--Larry
larryswedroe
 
Posts: 11201
Joined: 22 Feb 2007
Location: St Louis MO

Re: The new profitability factor, US and int'l evidence

Postby Clearly_Irrational » Wed May 22, 2013 5:46 pm

larryswedroe wrote:hexdump
I agree. The key was to move UP the income statement to gross profitability--not using earnings.
Larry


That doesn't surprise me, earnings are usually "managed".
User avatar
Clearly_Irrational
 
Posts: 2197
Joined: 13 Oct 2011

Re: The new profitability factor, US and int'l evidence

Postby AndroAsc » Wed May 22, 2013 7:09 pm

AQR and DFA is only accessible through advisors and require a hefty minimum of ~$1MIL correct? So for the rest of us, momentum and profitability factors are out of reach?

FYI, ETFs-of-the-day that pop up with momentum and profitability in the title do not count as desirable investment products.
AndroAsc
 
Posts: 880
Joined: 21 Nov 2009

Re: The new profitability factor, US and int'l evidence

Postby pcsrini » Wed May 22, 2013 8:33 pm

CXOAdvisory summary on Quantitative Value:

http://www.cxoadvisory.com/22933/fundam ... ive-value/
pcsrini
 
Posts: 54
Joined: 24 Jan 2011

Re: The new profitability factor, US and int'l evidence

Postby baw703916 » Sun May 26, 2013 3:29 pm

rmelvey wrote:
larryswedroe wrote:rmelvey
Holding P/E constant, a lower P/B implies a lower ROE.


Don't think this is true.
Book value is historical long-term while p/e is current. It MIGHT be true over the long term but untrue today.

Larry


Hi Larry,

I am talking about the algebra of the metrics when looking at a single firm.

ROE = E/B.

So P/E = (B/E) * (P/B)

Fundamentally, P/E ratio already reconciles the desire for high ROE and low P/B. It is the perfect distillation of the two metrics. So building a screener or ranking system that looks for both low P/B and high ROE, is kind of silly because P/E already perfectly does that. I know that the researchers moved higher up in the income statement, but the logic still stands true.

I think looking for low P/E and high ROE is more internally consistent (or using similar metrics higher up on the income statement). With this type of screener you are strictly valuing the firm as a going concern, caring about flows accuring to you. So minimal capital backing up the firm implies that it would take little capital to continue to grow the firm, making it attractive to you as an investor.


rmelvey and Larry,

The mathematical identity is indeed what I was referring to. Larry, thanks for pointing out that "profitability" tends to refer more to gross earnings than bottom line (even though Charlie Munger likes to call EBITDA "bull**** earnings"). If profitability and BtM are distinct factors which correlate with returns, then perhaps P/E is best viewed as a hybrid screen which captures some combination of the two.

Brad
Last edited by baw703916 on Sun May 26, 2013 3:43 pm, edited 1 time in total.
Most of my posts assume no behavioral errors.
User avatar
baw703916
 
Posts: 5332
Joined: 1 Apr 2007
Location: Northern Virginia

Re: The new profitability factor, US and int'l evidence

Postby rmelvey » Sun May 26, 2013 3:40 pm

baw703916 wrote:
rmelvey wrote:
larryswedroe wrote:rmelvey
Holding P/E constant, a lower P/B implies a lower ROE.


Don't think this is true.
Book value is historical long-term while p/e is current. It MIGHT be true over the long term but untrue today.

Larry


Hi Larry,

I am talking about the algebra of the metrics when looking at a single firm.

ROE = E/B.

So P/E = (B/E) * (P/B)

Fundamentally, P/E ratio already reconciles the desire for high ROE and low P/B. It is the perfect distillation of the two metrics. So building a screener or ranking system that looks for both low P/B and high ROE, is kind of silly because P/E already perfectly does that. I know that the researchers moved higher up in the income statement, but the logic still stands true.

I think looking for low P/E and high ROE is more internally consistent (or using similar metrics higher up on the income statement). With this type of screener you are strictly valuing the firm as a going concern, caring about flows accuring to you. So minimal capital backing up the firm implies that it would take little capital to continue to grow the firm, making it attractive to you as an investor.


rmelvey and Larry,

The mathematical identity is indeed what I was referring to. Larry, thanks for pointing out that "profitability" tends to refer more to gross earnings than bottom line (even though Charlie Munger likes to call EBITDA "bull**** earnings"). If profitability and BtM are distinct factors which correlate with returns, the perhaps P/E is best viewed as a hybrid screen which captures some combination of the two.

Brad


Now that I think more about, separating the two factors does make sense for attribution analysis (how much of returns came from cheapness, and how much from quality). However, recognizing that the two factors combine to basically become P/E is still useful IMO.
User avatar
rmelvey
 
Posts: 402
Joined: 18 Sep 2010

Previous

Return to Investing - Theory, News & General

Who is online

Users browsing this forum: ajcp, dipsylala, GTi, InvestorNewb, jp3051, linakin, OutOfCyan, RetiredInNH, Scooter57, serbeer, sosodan, Steady, Will do good and 91 guests