Better growth? Or on how the value premium ends

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
Topic Author
backpacker
Posts: 1620
Joined: Mon Sep 22, 2014 2:17 pm

Better growth? Or on how the value premium ends

Post by backpacker » Wed Aug 17, 2016 8:34 am

Here's a puzzle about the value premium. In this thread from awhile back, I posted a scary looking chart. No one should conclude anything from a single chart, but it sure looks like the realized value premium is being squashed over time.

Image

If the value premium is in fact slowly shrinking, why is that? The most obvious idea is that investors are paying more for value stocks and less for growth stocks. In theory, whether investors are paying more for value and less for growth is something we can just check and, indeed, some people have. There's a chart from AQR showing that nothing has changed, that the spread between value and growth hasn't shrunk one bit. You can find that chart here along with some reasons for thinking it may or may not be misleading.

So we have a puzzle. Given realized returns, it sure looks like the value premium is shrinking. On the other hand, it looks like investors aren't paying more for value and less for growth. What gives?

It struck me yesterday while going for a walk that that the market could be getting better at identifying growth stocks. If it is, the value premium could go away without the spread shrinking. That would solve the problem.

I had always thought that shrinking the value premium without shrinking the value/growth spread would require the market to have companies with better growth prospects than in the past. But that's not needed at all. We don't need a market with better companies, we just need a market that's better at sorting the companies it already has into those that will have higher growth and those that will have lower growth. That's enough to shrink the value premium without shrinking the spread.

That the market is getting better at predicting growth isn't a ridiculous idea. It could be that markets have more information than they used to about companies or, more plausibly, markets are better at using the information they already have. The value premium could just be one more casualty of improved market efficiency.

Note that I'm not saying the value premium will end. I have no idea and, if Bogleheads has taught me anything, it's that I stink at predicting the future. But it's worth noticing if the value premium does end, it could end because the markets get better at predicting growth rather than because investors pay more for value and less for growth.

Update: I now have a more visual and clearer presentation of the basic idea later in the thread.
Last edited by backpacker on Sat Aug 20, 2016 5:57 pm, edited 1 time in total.

Elbowman
Posts: 507
Joined: Tue Apr 03, 2012 2:25 pm

Re: Better growth? Or on how the value premium ends

Post by Elbowman » Wed Aug 17, 2016 10:29 am

As I've mentioned before, this is why I like VBR. When Larry Swedroe submits some apocalyptic post about how real returns for TSM will be 2% over the next 20 years, and how we better have SCV and EM, I think "good thing I have VBR! With two thirds the loading of DFA, I'll still do fine if value actually saves the day." And when I see post like yours about how the value premium may shrink to nothing, I think "good thing I have VBR! With 850 companies for 0.08%, it would be a reasonable investment even if there were no size and value premia."

garlandwhizzer
Posts: 2480
Joined: Fri Aug 06, 2010 3:42 pm

Re: Better growth? Or on how the value premium ends

Post by garlandwhizzer » Wed Aug 17, 2016 12:43 pm

Very interesting post, backpacker, thanks for posting. Poses a good question: will increased efficiency of markets due to dominance by professionals doom the pursuit of alpha?

Garland Whizzer

larryswedroe
Posts: 15989
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: Better growth? Or on how the value premium ends

Post by larryswedroe » Wed Aug 17, 2016 2:12 pm

Totally different view
IMO classic case of failing to understand that ALL factors, beta included, have LONG periods of underperformance. The persistence of the value premium has been basically the same as the beta premium no matter what horizon you use. Now here's simple example, 2000-08 S&P underperformed Vanguard Intermediate Treasury by 115% if memory serves, while value premium has underperformed by 16% since 2008. So far less underperformance yet people question the value premium but not the beta premium.
This is IMO a very strong reason why the value premium will persist, impatience of most investors

I would note also that the value spread (value vs. growth in valuations) is about historically the average, so that is the best indicator we have of the forward premium

I would add this, value will tend to underperform in low growth and disinflation and that is the economic regime we have been in, so not a surprise. It's also why IMO it's important to diversify across many factors as all experience long periods of underperformance and why I favor a risk parity type strategy, which is what the "Larry Portfolio" is along lines of

I hope that is helpful
Larry

User avatar
Topic Author
backpacker
Posts: 1620
Joined: Mon Sep 22, 2014 2:17 pm

Re: Better growth? Or on how the value premium ends

Post by backpacker » Wed Aug 17, 2016 6:05 pm

larryswedroe wrote: I would note also that the value spread (value vs. growth in valuations) is about historically the average, so that is the best indicator we have of the forward premium.
The whole point of my post was that the value spread doesn't need to shrink for the expected value premium to go away. All you need is for the market to do a better job estimating future earnings.

User avatar
grap0013
Posts: 1884
Joined: Thu Mar 18, 2010 1:24 pm

Re: Better growth? Or on how the value premium ends

Post by grap0013 » Wed Aug 17, 2016 6:13 pm

Elbowman wrote:And when I see post like yours about how the value premium may shrink to nothing, I think "good thing I have VBR! With 850 companies for 0.08%, it would be a reasonable investment even if there were no size and value premia."
I read about asset allocation for about 5 months in 2010 and in October of that year is when I had your revelation above. Beta is a component of SCV and even when a pure beta strategy outperforms a SCV strategy you will still do well with the latter. Same holds true for stocks vs. bonds. You can cherry pick some time frames where bonds outperform but the outperformance is very small. Couple all that with we all are just dust in the wind and live extravagant lives compared to most of the rest of the world and one can take on an all SCV strategy for maximum diversification and sleep well at night.

Value is riskier than growth but "not really" if you catch my meaning. The chance of value outperformance is good and the magnitude of the potential benefits to the upside is higher than the magnitude of the downside. It's not symmetrical.
There are no guarantees, only probabilities.

User avatar
grap0013
Posts: 1884
Joined: Thu Mar 18, 2010 1:24 pm

Re: Better growth? Or on how the value premium ends

Post by grap0013 » Wed Aug 17, 2016 6:21 pm

backpacker wrote: The whole point of my post was that the value spread doesn't need to shrink for the expected value premium to go away. All you need is for the market to do a better job estimating future earnings.
That's asking a lot. I know a lot of people in the market that are not good with money. Plus value companies have more debt and variability in their earnings making future earnings estimates more difficult. Also, you have growth investors willing to pay a premium for lottery like effects that stimulate dopamine release in the prefrontal cortex. There is not a way to turn off that dopamine and it feels good.
There are no guarantees, only probabilities.

User avatar
RyeWhiskey
Posts: 864
Joined: Thu Jan 12, 2012 10:04 pm

Re: Better growth? Or on how the value premium ends

Post by RyeWhiskey » Wed Aug 17, 2016 6:29 pm

If Bogle were to post here I'm sure he'd show up with a Telltale Chart in this thread. :beer
This post was brought to you by Vanguard Total World Stock Index (VTWSX/VT).

User avatar
Topic Author
backpacker
Posts: 1620
Joined: Mon Sep 22, 2014 2:17 pm

Re: Better growth? Or on how the value premium ends

Post by backpacker » Wed Aug 17, 2016 6:43 pm

grap0013 wrote:
backpacker wrote: The whole point of my post was that the value spread doesn't need to shrink for the expected value premium to go away. All you need is for the market to do a better job estimating future earnings.
That's asking a lot. I know a lot of people in the market that are not good with money. Plus value companies have more debt and variability in their earnings making future earnings estimates more difficult. Also, you have growth investors willing to pay a premium for lottery like effects that stimulate dopamine release in the prefrontal cortex. There is not a way to turn off that dopamine and it feels good.
Could be, definitely. Mostly my "aha" moment was noticing that there is another possibility, not that this is how things really are. There's probably some clever way to check and see if the market is doing a better job predicting growth.

stlutz
Posts: 5429
Joined: Fri Jan 02, 2009 1:08 am

Re: Better growth? Or on how the value premium ends

Post by stlutz » Wed Aug 17, 2016 9:18 pm

I think this exchange is why I've become more skeptical of tilts of various types over this time. Backpacker posits a very legitimate and believable "story" as to how the value premium could go away and Larry replies with the same comments he makes in every thread.

Essentially what we get told us that one should only tilt if they are confident that they will never be convinced it is a bad idea. My thinking is just not that fixed.

Being open to being persuaded you are wrong probably makes you a better person. It might not make you a better investor however.

User avatar
grap0013
Posts: 1884
Joined: Thu Mar 18, 2010 1:24 pm

Re: Better growth? Or on how the value premium ends

Post by grap0013 » Thu Aug 18, 2016 8:26 am

stlutz wrote:I think this exchange is why I've become more skeptical of tilts of various types over this time. Backpacker posits a very legitimate and believable "story" as to how the value premium could go away and Larry replies with the same comments he makes in every thread.

Essentially what we get told us that one should only tilt if they are confident that they will never be convinced it is a bad idea. My thinking is just not that fixed.

Being open to being persuaded you are wrong probably makes you a better person. It might not make you a better investor however.
^+1 Wise words. Take home point: stay the course no matter what strategy you pick. Tinkering will decrease your returns. The best asset allocation for an investor is the one that he/she can stick too.

I can even pick the black hole of investing small cap growth and mix it 60:40 with total bond and make a lot of money as long as I stay the course: https://www.portfoliovisualizer.com/bac ... alBond1=40
There are no guarantees, only probabilities.

larryswedroe
Posts: 15989
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: Better growth? Or on how the value premium ends

Post by larryswedroe » Thu Aug 18, 2016 8:45 am

backpacker
So you are saying the market was inefficient in estimating earnings and or the investors will stop favoring growth companies for the behavioral reasons they did in past. First, there is still the issue of value stocks being a clear risk story, with dozens of papers showing clear risk explanations including higher leverage, both operating and financial, and more volatility of earnings. So even if you believe people can forecast more accurately there should always still be an ex ante value premium unless you think risk and expected returns should be unrelated.

Continuing, here's simple explanation that would I think show your logic is faulty. The research does show that earnings growth of both value and growth stocks tends to RTM faster than market expects. So let's play this out. So now the market will do a better job and show lower future growth of earnings for growth stocks and faster for value. So if you now expect lower growth of earnings for growth stocks then you would pay a lower price for them and if you expect higher growth of earnings of value stocks then you should be willing to pay a higher price for them and thus the value to growth spread would narrow. Not remain the same.

Second,human behavior doesn't change, at least there is lots of evidence that it's highly persistent. Which is why behavioral anomalies like momentum have persistent for decades after their publications. And because of limits to arbitrage it's much easier to correct undervaluation but not overvaluation. And that is part of the value premium, the underperformance of overvalued growth stocks (especially small growth stocks with high investment and low profitability).

So IMO it's still always about valuations and relative valuations

And yes one should have strong convictions as to why one should believe a premium should persist because all will go through long periods of underperformance.

Finally I would add if you think earnings forecasts will improve you should also expect a lower equity risk premium going forward as there will be more certainty about future earnings and thus less surprises.

And Rye Whisky, as I pointed out many times the famous Tell Tale Chart should be infamous because it doesn't show what it claims to show and Bogle has been told this. It shows that ACTIVE VALUE managers did not outperform ACTIVE GROWTH managers. And the reason is that both style drift. He did not show that value stocks did not outperform growth stocks, which of course he cannot do that. Why he chose to do what he did I cannot explain as it doesn't make his case. Anyone checking the Ken French data base which has long be publicly available can show that what returns have been. This myth needs to die.

Hope that is helpful
Larry
Last edited by larryswedroe on Thu Aug 18, 2016 9:04 am, edited 1 time in total.

User avatar
Topic Author
backpacker
Posts: 1620
Joined: Mon Sep 22, 2014 2:17 pm

Re: Better growth? Or on how the value premium ends

Post by backpacker » Thu Aug 18, 2016 8:58 am

larryswedroe wrote: If you think earnings forecasts will improve you should also expect a lower equity risk premium going forward as there will be more certainty about future earnings and thus less surprises.
There are absolute earnings forecasts (company A is likely to earn x over the next five years) and relative earnings forecasts (company A is likely to earn twice as much as company B over the next five years). The market could get better at the latter without getting any better at the former. In that case, the value premium would go away without having any effect on the equity premium.

Or in plain English: The market could get better at comparing companies without getting better at predicting the next market crisis.

larryswedroe
Posts: 15989
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: Better growth? Or on how the value premium ends

Post by larryswedroe » Thu Aug 18, 2016 9:01 am

Backpacker, I guess anything is possible. But that seems like highly unlikely proposition. But any improvement in earnings forecasting ability lowers uncertainty and thus should lower the risk premium of all securities, so IMO that idea doesn't hold up.

And on the main issue
The research does show that earnings growth of both value and growth stocks tends to RTM faster than market expects. So let's play this out. So now the market will do a better job and show lower future growth of earnings for growth stocks and faster for value. So if you now expect lower growth of earnings for growth stocks then you would pay a lower price for them and if you expect higher growth of earnings of value stocks then you should be willing to pay a higher price for them and thus the value to growth spread would narrow. Not remain the same.
In other words if market becomes better at forecasting the spreads would narrow, and that would lower the value premium. And then of course you have the risk story which should ensure that it never entirely disappears ex-ante.

Larry

User avatar
nedsaid
Posts: 12668
Joined: Fri Nov 23, 2012 12:33 pm

Re: Better growth? Or on how the value premium ends

Post by nedsaid » Thu Aug 18, 2016 9:10 am

You have read it here first. The equivalent of Business Week's "Death of Equities" cover story just before one of the biggest bull markets in history. Here on Bogleheads, we are witnessing the "Death of Value" and the "Death of International Investing." At the same time, threads pop up that eight years of dividend chasing are not enough and that investors need to chase them even harder. Threads also want to keep chasing REITs despite very high valuations, recent hot performance, and the same eight years of dividend chasing. If those are not contrary indicators, I don't know what are.

I have noticed that small stocks are doing well in 2016 and small value a bit better than that. Plain old boring Large Value is doing very well this year despite last rites being said over the casket. Even my Large Value International fund is in the black while my other International Stock funds are in the red.

I don't know, Backpacker. You should have posted this about nine months ago. To my untrained eye, Value looks awfully good in 2016. The corpse is not only twitching but appears to actually be alive and well. I can't help but notice that the casket is empty!
A fool and his money are good for business.

User avatar
nedsaid
Posts: 12668
Joined: Fri Nov 23, 2012 12:33 pm

Re: Better growth? Or on how the value premium ends

Post by nedsaid » Thu Aug 18, 2016 9:17 am

larryswedroe wrote:
Second,human behavior doesn't change, at least there is lots of evidence that it's highly persistent. Which is why behavioral anomalies like momentum have persistent for decades after their publications. And because of limits to arbitrage it's much easier to correct undervaluation but not overvaluation. And that is part of the value premium, the underperformance of overvalued growth stocks (especially small growth stocks with high investment and low profitability).
Despite what the quants say, this is about human nature and human behavior. These behavioral anomalies haven't gone away even with the coming of the robots. Since humans program the robots, I would expect not only that these anomalies will persist but in certain circumstances will be amplified.
A fool and his money are good for business.

User avatar
nedsaid
Posts: 12668
Joined: Fri Nov 23, 2012 12:33 pm

Re: Better growth? Or on how the value premium ends

Post by nedsaid » Thu Aug 18, 2016 9:18 am

Oh, and I should bring up that Value "died" in the late 1990's too.
A fool and his money are good for business.

Dulocracy
Posts: 2303
Joined: Wed Feb 27, 2013 1:03 pm
Location: Atlanta, GA

Re: Better growth? Or on how the value premium ends

Post by Dulocracy » Thu Aug 18, 2016 9:28 am

I think it is highly unlikely that with people studying the market for long periods of time, all of a sudden they are going to start doing a better job of figuring it out. The market is like a slippery fish. As soon as you think you have it in one spot, it wiggles. You can try to hold it in one spot, but you will lose it. Alternatively, you can wiggle with it, but it winds up in a slightly different spot. Ok, bad analogy, but I think you get what I am saying. The fish wiggled before, but it will likely wiggle in a different way next time. It is not less slippery.

As far as Larry saying "the same thing," he would not be very credible if he was not consistent. He says the same thing, but in different ways. He also posts about how new studies or new information lends itself to back up the "same thing." In debate, that is staying on message. In science, that is showing consistent results. Either way, I do not fault him for what should be considered a good thing.

As far as the value premium ending: it may. I recognized that when I started with a value tilt. My plan all along was to start with a heavy value tilt and to tilt less and less as I age (and need a less risky risk profile). Also, for the value premium to show up can take decades, so a value tilt really is not as appropriate for someone 10 years from retirement as it is for someone 40 years from retirement. Nonetheless, even if the evidence is clear that it is going away, I would stay the course. As grap0013 correctly pointed out: the worst thing you can do is keep messing with your asset allocation.
I'm not a financial professional. Post is info only & not legal advice. No attorney-client relationship exists with reader. Scrutinize my ideas as if you spoke with a guy at a bar. I may be wrong.

dabblingeconomist
Posts: 115
Joined: Wed Jul 13, 2016 8:42 pm

Re: Better growth? Or on how the value premium ends

Post by dabblingeconomist » Thu Aug 18, 2016 9:35 am

larryswedroe wrote:First, there is still the issue of value stocks being a clear risk story, with dozens of papers showing clear risk explanations including higher leverage, both operating and financial, and more volatility of earnings. So even if you believe people can forecast more accurately there should always still be an ex ante value premium unless you think risk and expected returns should be unrelated.
To the extent that it's market risk, it shows up in beta. To the extent that it's idiosyncratic risk, I can see a good argument for why this should have less impact on expected return over time: historically, even until quite recently, most stocks were held by individuals, who could not diversify fully and would want a premium on individually riskier stocks. (It could be a systematic risk other than market risk - indeed, this is the classical story - but I don't really believe that.)
larryswedroe wrote:Continuing, here's simple explanation that would I think show your logic is faulty. The research does show that earnings growth of both value and growth stocks tends to RTM faster than market expects. So let's play this out. So now the market will do a better job and show lower future growth of earnings for growth stocks and faster for value. So if you now expect lower growth of earnings for growth stocks then you would pay a lower price for them and if you expect higher growth of earnings of value stocks then you should be willing to pay a higher price fro them and thus the value to growth spread would narrow. Not remain the same.
No, I think his logic is still potentially valid. Suppose that previously, investors tried to sort stocks into low-growth and high-growth buckets, resulting in a big value to growth spread, but - unbeknownst to them - they made a lot of errors in this categorization. With this kind of error-prone, overconfident sorting, a healthy value return premium emerges.

Now, suppose that investors are still trying to sort stocks into low-growth and high-growth buckets, but they have gotten much better at doing so. In other words, they're finally able to justify their old overconfidence. The value to growth spread will still be just as large, but now it will do a better job of predicting earnings growth, meaning that the value premium will disappear going forward.

One can think of many other stories as well. Maybe typical investors are just as overconfident as ever, but they are now increasingly counteracted by a cadre of smart investors who are trying to exploit the value premium. In isolation, you'd think that this would lower the value to growth spread, which hasn't happened (at least not that much) - but a lot of other things could be going on at the same time. Maybe with modern technology, book value is an increasingly inaccurate measure. In that case, the value to growth spread should be increasing (because book value is all over the place relative to ideal market value) - and maybe overconfident investors are still trying to push it up even higher than that, but now they're counteracted by smarter investors, resulting in a stable spread.

One could go on and on. I don't know if any of these stories are definitely right, just that it's possible we will see a shrinking or disappearing value premium, and simple quantitative measures like the value/growth spread won't warn us about it (at least in many scenarios).
larryswedroe wrote:And because of limits to arbitrage it's much easier to correct undervaluation but not overvaluation. And that is part of the value premium, the underperformance of overvalued growth stocks (especially small growth stocks with high investment and low profitability).
If limits to arbitrage prevent other market actors from exploiting something, how are Boglehead investors like us able to exploit it?

User avatar
Topic Author
backpacker
Posts: 1620
Joined: Mon Sep 22, 2014 2:17 pm

Re: Better growth? Or on how the value premium ends

Post by backpacker » Thu Aug 18, 2016 9:48 am

dabblingeconomist wrote:
larryswedroe wrote:Continuing, here's simple explanation that would I think show your logic is faulty. The research does show that earnings growth of both value and growth stocks tends to RTM faster than market expects. So let's play this out. So now the market will do a better job and show lower future growth of earnings for growth stocks and faster for value. So if you now expect lower growth of earnings for growth stocks then you would pay a lower price for them and if you expect higher growth of earnings of value stocks then you should be willing to pay a higher price fro them and thus the value to growth spread would narrow. Not remain the same.
No, I think his logic is still potentially valid. Suppose that previously, investors tried to sort stocks into low-growth and high-growth buckets, resulting in a big value to growth spread, but - unbeknownst to them - they made a lot of errors in this categorization. With this kind of error-prone, overconfident sorting, a healthy value return premium emerges.

Now, suppose that investors are still trying to sort stocks into low-growth and high-growth buckets, but they have gotten much better at doing so. In other words, they're finally able to justify their old overconfidence. The value to growth spread will still be just as large, but now it will do a better job of predicting earnings growth, meaning that the value premium will disappear going forward.
That's it exactly! I wrote a response to Larry, but this is better than what I wrote, so I'll just nod in agreement.

User avatar
Topic Author
backpacker
Posts: 1620
Joined: Mon Sep 22, 2014 2:17 pm

Re: Better growth? Or on how the value premium ends

Post by backpacker » Thu Aug 18, 2016 10:07 am

nedsaid wrote:To my untrained eye, Value looks awfully good in 2016. The corpse is not only twitching but appears to actually be alive and well. I can't help but notice that the casket is empty!
Ahhh! Count Valula has risen and is stalking through the countryside! :D
nedsaid wrote:Oh, and I should bring up that Value "died" in the late 1990's too.
Maybe this time, "this time is different" is different? In the late 90s, we were in a growth bubble. Value lost big to growth. For the last ten years or so, value hasn't lost to growth. Its just been flat.
larryswedroe wrote:So if you now expect lower growth of earnings for growth stocks then you would pay a lower price for them and if you expect higher growth of earnings of value stocks then you should be willing to pay a higher price fro them and thus the value to growth spread would narrow. Not remain the same.
Here's a simple example that may help.

Say the market has growth companies and value companies. Every company had the same $100 in earnings last year (to keep things simple). Growth companies have a PE of 30 and value companies a PE of 10. There are no other companies.

Alpha Corp has a PE of 30 and Beta Corp has a PE of 10. The market thinks that Alpha Corp will grow faster but, in fact, is wrong about this. Beta Corp is the company that will in fact grow faster.

Fortunately, a bunch of clever analysts realize this, and so the price of Alpha is bid down and the price of Beta bid up and, in fact, their prices swap. Beta acquires a PE of 30 and Alpha a PE of 10. Since the market is right, this improves the expected future performance of growth relative to value. That the average PE of growth stocks and of value stocks remains constant throughout is just trivial math.
Last edited by backpacker on Thu Aug 18, 2016 11:36 am, edited 2 times in total.

larryswedroe
Posts: 15989
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: Better growth? Or on how the value premium ends

Post by larryswedroe » Thu Aug 18, 2016 10:17 am

Dabble
On first point, if that was correct then we would still have only a single factor CAPM model and would not need another factors. But other factors are required as they add explanatory power.

And for backpacker .
And the rest I already address, don't agree. The logic is as I stated, nothing more I can say, the spread must narrow. If they forecast better then the growth companies will have lower future earnings forecasted and and thus must have lower valuations and vice versa. This is simple logic. If you forecast lower earnings you will be willing to pay a lower multiple.

And yes the value spread could fall and likely has fallen since publication and for various other reasons. In fact we haircut all premiums 25% in our MCS and expected returns based on the literature showing declines. But the literature also shows that premiums that are risk based or have high limits to arb preventing corrections they persist. And obviously the risk stories are still there so in those cases the premiums should also persist. Remember the ERP has also come way down for some of the same reasons the other premiums have fallen.

As to Bogleheads "exploiting" the knowledge I've written 14 books on the subject with the latest hopefully out in about 3 months, Your Complete Guide to Factor Based Investing. But short answer is to
a) avoid the "lottery tickets" and the funds that own these type stocks
b) where have strong belief that a factor and it's premium should persist based on it's persistence, pervasiveness, robustness, intuitiveness and implementability, and it not being subsumed by other factors (not explained by other factors) then should allocate capital to that factor.

Hope that helps
Larry

larryswedroe
Posts: 15989
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: Better growth? Or on how the value premium ends

Post by larryswedroe » Thu Aug 18, 2016 10:55 am

backpacker sorry but that doesn't work. You cannot have it that way. You're trying way to hard to make a story work because you want it to work. At least that is what I'm thinking.The growth companies in your example are not really growth companies by your own definition as they should have been value companies.

The research does show that growth companies in fact do have faster growth of earnings, just that their forecasted growth rates of abnormal rates of growth are TOO HIGH (not that they are lower than value companies) and the reverse true of value companies. So the market in your example will correct than and lower the rate of growth in earnings but they will still be higher than for value companies. So the spreads would narrow but growth stocks would then have lower valuations but still higher than that of value companies. The spread then narrows and the premium shrinks.

And don't forget the risk story. Value companies should always have higher expected returns due to lower valuations. The risk premium required when discounting future earnings is higher than used for growth stocks. That higher discount rate is the higher expected return. It must be the case or you don't believe they are more risky

Hope that helps
Larry

User avatar
bertilak
Posts: 6876
Joined: Tue Aug 02, 2011 5:23 pm
Location: East of the Pecos, West of the Mississippi

Re: Better growth? Or on how the value premium ends

Post by bertilak » Thu Aug 18, 2016 11:00 am

larryswedroe wrote:IMO it's important to diversify across many factors as all experience long periods of underperformance ...
A while back I made a post which included something like the following:
  • "There are MANY factors. TSM has them all. Why hunt them down individually?"
You corrected me by saying (as I understood you) one is not exposed to a factor unless one holds it in greater proportion than the market.

Perhaps my understanding of your correction needs further correction. Would not diversifying across many factors dilute the exposure to each and tend to neutralize the positive effect of factors? The more diversified, the closer to TSM? Maybe the answer is that some factors are better than others? Despite their long periods of under-performance?
May neither drought nor rain nor blizzard disturb the joy juice in your gizzard. -- Squire Omar Barker (aka S.O.B.), the Cowboy Poet

User avatar
Topic Author
backpacker
Posts: 1620
Joined: Mon Sep 22, 2014 2:17 pm

Re: Better growth? Or on how the value premium ends

Post by backpacker » Thu Aug 18, 2016 11:45 am

larryswedroe wrote:The growth companies in your example are not really growth companies by your own definition as they should have been value companies.
Eh? A growth company is one whose valuation is in the top half. A value company is one whose valuation is in the bottom half. There is no should in the definition of growth and value.
larryswedroe wrote:The research does show that growth companies in fact do have faster growth of earnings, just that their forecasted growth rates of abnormal rates of growth are TOO HIGH (not that they are lower than value companies) and the reverse true of value companies. So the market in your example will correct than and lower the rate of growth in earnings but they will still be higher than for value companies. So the spreads would narrow but growth stocks would then have lower valuations but still higher than that of value companies. The spread then narrows and the premium shrinks.
Try writing this out step by step using the companies from my example. Explain which company's price changes and why that changes the average PE ratio of growth stocks or the average PE ratio of value stocks. Be clear about what exactly you're denying about my example. Are you saying that no value stocks will ever have higher growth than any growth stocks? Presumably not. So what are you saying?
Last edited by backpacker on Thu Aug 18, 2016 1:08 pm, edited 1 time in total.

larryswedroe
Posts: 15989
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: Better growth? Or on how the value premium ends

Post by larryswedroe » Thu Aug 18, 2016 1:01 pm

Bertilak

First factors are long short portfolios so TSM has no exposure to any factor except beta. Example, the small stocks in the TSM give you positive exposure to the size factor but the large stocks give you an exactly offsetting NEGATIVE exposure to the same factor. And that's true for value, momentum, quality, or whatever. So the only way to gain exposure to a factor is to own more of it than the market does, to "tilt" the portfolio towards a factor. Long only funds have limited ability to do that because they cannot go short the other side (which typically has about half the premium).

IMO, fwiw, is as follows, TSM has all its eggs in one risk basket, beta. And, as I explain in the new book, the typical 60/40 portfolio with intermediate safe bonds has about 85% of its RISK (not assets) in BETA and the rest in TERM, or credit if using corporates. And it has no exposure to the other factors. On the other hand, a "Larry Portfolio" with much lower beta but highly tilted has more of a risk parity profile with more exposure to various factors, including more on TERM. That is a more diversified approach and important as we don't know which factors will outperform in future and when, and it has historically cut way down on tail risk besides delivering more efficient returns.

Hope that helps
Larry

lemonPepper
Posts: 89
Joined: Wed Feb 18, 2015 7:43 pm

Re: Better growth? Or on how the value premium ends

Post by lemonPepper » Thu Aug 18, 2016 1:06 pm

Say the market has growth companies and value companies. Every company had the same $100 in earnings last year (to keep things simple). Growth companies have a PE of 30 and value companies a PE of 10. There are no other companies.

Alpha Corp has a PE of 30 and Beta Corp has a PE of 10. The market thinks that Alpha Corp will grow faster but, in fact, is wrong about this. Beta Corp is the company that will in fact grow faster.

Fortunately, a bunch of clever analysts realize this, and so the price of Alpha is bid down and the price of Beta bid up and, in fact, their prices swap. Beta acquires a PE of 30 and Alpha a PE of 10. Since the market is right, this improves the expected future performance of growth relative to value. That the average PE of growth stocks and of value stocks remains constant throughout is just trivial math.
Growth and value are relative terms. When the PE reverses, the alpha is value and beta is growth. If you consistenly own value companies, you are good.

Just stick with it for 40 years. Don't chase performance by selling value companies in 1999 to buy growth companies. Easier said than done.

larryswedroe
Posts: 15989
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: Better growth? Or on how the value premium ends

Post by larryswedroe » Thu Aug 18, 2016 1:11 pm

backpacker
Sorry have full time job and cannot spend more time on this issue which IMO is very simple.

But with that said, I will try one last time. But don't have more time to spend on this. Of course I'm not saying that some value stocks won't become growth stocks and vice versa. But that isn't the issue.

The market has been correct in forecasting that growth stocks WILL have higher future rates of growth in earnings, on average, just too high. And the same for value stocks, they would have lower future growth in earnings, just too low. The reason is it turns out that the RTM of abnormal earnings occurs faster than the market seems to expect. Fama and French among others have paper on this. So if the market had forecasted with perfect foresight the growth companies would still have been growth companies with higher valuations and the value companies would still have been value companies but with higher valuations and the value premium would have been smaller, not gone.

The reason that it should never happen is quite simple, as I have explained, value companies are more risky and we cannot forecast with perfection. If we could there would be no risk there. But the value companies have higher volatility of their earnings, making them more difficult to forecast and that is one reason why they get higher discount rates which leads to lower valuations and higher expected returns. If you don't get that part there is nothing more I can say. If you disagree then you believe that the value story is purely behavioral (against all the evidence) and then you have to posit that the behavior will change so that humans no longer act as humans and give up their preferences for right tailed skewness etc.]

In other words better forecasting, even if you believe that, and don't know why you would since so much is unknowable, would just shrink the premium. Not make it disappear.

I hope that helps
Larry

dabblingeconomist
Posts: 115
Joined: Wed Jul 13, 2016 8:42 pm

Re: Better growth? Or on how the value premium ends

Post by dabblingeconomist » Thu Aug 18, 2016 1:15 pm

larryswedroe wrote:And the rest I already address, don't agree. The logic is as I stated, nothing more I can say, the spread must narrow. If they forecast better then the growth companies will have lower future earnings forecasted and and thus must have lower valuations and vice versa. This is simple logic. If you forecast lower earnings you will be willing to pay a lower multiple.
This is where I disagree. The problem is that the very definition of "growth" and "value" is endogenous to the market. Unless I'm mistaken (which I might be, since finance is not my field), these categories are defined according to market-to-book ratios.

If investors aren't very good, then they will do a bad job in predicting growth. The market-to-book ratios that result from their investment choices will also, therefore, do a bad job in predicting growth. The question is what happens if, over time, investors get smarter. Then market-to-book ratios will start to do a better job of predicting growth.

The danger in your argument, I think, is that it relies upon a statistical observation (the mediocre predictive power of market-to-book) that is dependent on investor behavior. If investor behavior changes, then this statistical observation will no longer hold, and neither will the value premium; and this need not imply a decline in spreads.

Maybe it's unreasonable to think that investors as a whole are becoming smarter in this way - but as I mentioned, there are other stories that could fit the same data points too. Maybe dispersion in the (ideal market price)/(book) ratio has risen due to fundamentals, and at the same time a growing set of smart investors like yourself (who don't pretend to have any elixir for predicting earnings growth) have begun to bet on value. These two opposing forces could be roughly offsetting each other in their effect on spreads, producing the small decline in spreads that we observe today, even as the latter set of investors has eaten away at most or all of the value premium.
larryswedroe wrote:On first point, if that was correct then we would still have only a single factor CAPM model and would not need another factors. But other factors are required as they add explanatory power.
But the reason for pricing the Fama-French value "risk factor" so aggressively seems fairly elusive. As an economist, I can understand why market risk would be priced with a premium, even if the equity premium seems a little too high. After all, it's correlated with all kinds of other systematic risks in the economy.

But why do investors care so much about the systematic performance of value vs. growth that they're willing to give a premium to value - and even if they do, why should I think that I should be any more exposed to that risk factor (whatever it is) than the typical investor is?

It seems like there is an enormous tension between the Fama view of the value premium (that it's the market's rational pricing of exposure to an associated risk factor) and the typical practitioner's view of the value premium (that it's mainly the result of behavioral investors who are overoptimistic about growth).

User avatar
bertilak
Posts: 6876
Joined: Tue Aug 02, 2011 5:23 pm
Location: East of the Pecos, West of the Mississippi

Re: Better growth? Or on how the value premium ends

Post by bertilak » Thu Aug 18, 2016 1:26 pm

larryswedroe wrote:Bertilak

First factors are long short portfolios so TSM has no exposure to any factor except beta. Example, the small stocks in the TSM give you positive exposure to the size factor but the large stocks give you an exactly offsetting NEGATIVE exposure to the same factor. And that's true for value, momentum, quality, or whatever. So the only way to gain exposure to a factor is to own more of it than the market does, to "tilt" the portfolio towards a factor. Long only funds have limited ability to do that because they cannot go short the other side (which typically has about half the premium).

IMO, fwiw, is as follows, TSM has all its eggs in one risk basket, beta. And, as I explain in the new book, the typical 60/40 portfolio with intermediate safe bonds has about 85% of its RISK (not assets) in BETA and the rest in TERM, or credit if using corporates. And it has no exposure to the other factors. On the other hand, a "Larry Portfolio" with much lower beta but highly tilted has more of a risk parity profile with more exposure to various factors, including more on TERM. That is a more diversified approach and important as we don't know which factors will outperform in future and when, and it has historically cut way down on tail risk besides delivering more efficient returns.

Hope that helps
Larry
I think I follow that, but what confuses me is the concept of diversify across many factors. It seems like, in the extreme, by tilting towards ALL of them you would end up with no tilt at all, just as you would if you tried to hold every stock X% above its market weight you end up with the same relative weighting you had before his exercise. If you started with no tilt you still have no tilt.

Now, many is not the same as all, but it seems we are talking a matter of degree. The closer many gets to all the less we at tilting.

Am I making some unwarranted assumption about cap-weighting?
May neither drought nor rain nor blizzard disturb the joy juice in your gizzard. -- Squire Omar Barker (aka S.O.B.), the Cowboy Poet

User avatar
nedsaid
Posts: 12668
Joined: Fri Nov 23, 2012 12:33 pm

Re: Better growth? Or on how the value premium ends

Post by nedsaid » Thu Aug 18, 2016 1:31 pm

backpacker wrote:
nedsaid wrote:To my untrained eye, Value looks awfully good in 2016. The corpse is not only twitching but appears to actually be alive and well. I can't help but notice that the casket is empty!
Ahhh! Count Valula has risen and is stalking through the countryside! :D

Nedsaid: Yes, Count Valula has risen at least year to date 2016. My thesis is that we had a growth trend in the market from about 2010-2015. It appears that phase is over and I would expect value to be in vogue for the next few years. But I am still keeping my growth investments as I could be wrong.
nedsaid wrote:Oh, and I should bring up that Value "died" in the late 1990's too.
Maybe this time, "this time is different" is different? In the late 90s, we were in a growth bubble. Value lost big to growth. For the last ten years or so, value hasn't lost to growth. Its just been flat.

Nedsaid: Value definitely lost to growth from about 2010 through 2015. I could see it in my own investments. Over the last decade, I suppose the two might be about equal. The 2008-2009 financial crisis was particularly hard on Value as dividends were cut and many Value stocks are financial stocks.
larryswedroe wrote:So if you now expect lower growth of earnings for growth stocks then you would pay a lower price for them and if you expect higher growth of earnings of value stocks then you should be willing to pay a higher price fro them and thus the value to growth spread would narrow. Not remain the same.
Here's a simple example that may help.

Say the market has growth companies and value companies. Every company had the same $100 in earnings last year (to keep things simple). Growth companies have a PE of 30 and value companies a PE of 10. There are no other companies.

Alpha Corp has a PE of 30 and Beta Corp has a PE of 10. The market thinks that Alpha Corp will grow faster but, in fact, is wrong about this. Beta Corp is the company that will in fact grow faster.

Fortunately, a bunch of clever analysts realize this, and so the price of Alpha is bid down and the price of Beta bid up and, in fact, their prices swap. Beta acquires a PE of 30 and Alpha a PE of 10. Since the market is right, this improves the expected future performance of growth relative to value. That the average PE of growth stocks and of value stocks remains constant throughout is just trivial math.

Nedsaid: In real life, the transition from Value to Growth doesn't happen all at once as in your example. It would be a more gradual move from one end of the spectrum to another. There was a time, probably during the 2008-2009 bear market that certain popular growth stocks started hitting value screens. Not sure that means that they all the sudden became Value stocks but certainly some Value managers started buying them. There is a certain amount of perception here too, not sure it can all be boiled down to financial ratios. A Growth stock would have to run into trouble for an extended period of time before investors started perceiving it as an outright Value stock.
A fool and his money are good for business.

NMJack
Posts: 836
Joined: Sun Feb 14, 2016 1:22 pm

Re: Better growth? Or on how the value premium ends

Post by NMJack » Thu Aug 18, 2016 1:52 pm

Interesting thread, although I'm still not sold on all this factor "stuff." I understand that with selective back testing, a (crappy) case can be made for many theories saying tilt to this or that and you'll beat the market. If it was that easy, wouldn't the active fund managers all be doing this?

I do find myself losing interest and asking the question "so why can't I just leave everything in TSM and go enjoy life?" :confused

User avatar
Topic Author
backpacker
Posts: 1620
Joined: Mon Sep 22, 2014 2:17 pm

Re: Better growth? Or on how the value premium ends

Post by backpacker » Thu Aug 18, 2016 2:01 pm

larryswedroe wrote:backpacker [...]
You're making this too complicated and getting lost in the details. Just think about it this way.
dabblingeconomist wrote: If investors aren't very good, then they will do a bad job in predicting growth. The market-to-book ratios that result from their investment choices will also, therefore, do a bad job in predicting growth. The question is what happens if, over time, investors get smarter. Then market-to-book ratios will start to do a better job of predicting growth.
Making the same point: The market makes growth predictions and the value spread is the "price" of those predictions. Investors have in the past paid too much for the predictions they are getting, thus generating the value premium.

If you're paying too much for a bottle of toothpaste, there are two options. You can either pay less for the same bottle of toothpaste or pay the same price for a larger bottle of toothpaste.

Likewise, if investors have been paying too much for growth predictions, there are two options. They can either pay less for the same predictions (i.e. shrink the value spread) or the market can provide better predictions for the same price (i.e. without shrinking the value spread).

User avatar
nedsaid
Posts: 12668
Joined: Fri Nov 23, 2012 12:33 pm

Re: Better growth? Or on how the value premium ends

Post by nedsaid » Thu Aug 18, 2016 2:28 pm

bertilak wrote:
larryswedroe wrote:Bertilak

First factors are long short portfolios so TSM has no exposure to any factor except beta. Example, the small stocks in the TSM give you positive exposure to the size factor but the large stocks give you an exactly offsetting NEGATIVE exposure to the same factor. And that's true for value, momentum, quality, or whatever. So the only way to gain exposure to a factor is to own more of it than the market does, to "tilt" the portfolio towards a factor. Long only funds have limited ability to do that because they cannot go short the other side (which typically has about half the premium).

IMO, fwiw, is as follows, TSM has all its eggs in one risk basket, beta. And, as I explain in the new book, the typical 60/40 portfolio with intermediate safe bonds has about 85% of its RISK (not assets) in BETA and the rest in TERM, or credit if using corporates. And it has no exposure to the other factors. On the other hand, a "Larry Portfolio" with much lower beta but highly tilted has more of a risk parity profile with more exposure to various factors, including more on TERM. That is a more diversified approach and important as we don't know which factors will outperform in future and when, and it has historically cut way down on tail risk besides delivering more efficient returns.

Hope that helps
Larry
I think I follow that, but what confuses me is the concept of diversify across many factors. It seems like, in the extreme, by tilting towards ALL of them you would end up with no tilt at all, just as you would if you tried to hold every stock X% above its market weight you end up with the same relative weighting you had before his exercise. If you started with no tilt you still have no tilt.

Now, many is not the same as all, but it seems we are talking a matter of degree. The closer many gets to all the less we at tilting.

Am I making some unwarranted assumption about cap-weighting?
Bertilak, my take is that diversification across factors makes sense but I think that you don't want to try to invest in all of them. Most people here who tilt, tilt towards small value. So pretty much, they would invest across market, size (small), and value factors. Value tends to have low momentum and you are getting two different sets of stocks, Value and Momentum, and to a degree you would get canceling out. Also, it seems that the premiums for Momentum are smaller than for Size or Value. Momentum is more difficult for investors to capture as it is a higher turnover strategy and the premium is probably smaller.

What you don't want to do is be so diversified that you get a cancelling out effect on whatever excess return you were trying to achieve. To me, it seems like the most efficient way to do this would be market, size, value, and to a lesser degree momentum. Different factors do well at different times so investing across factors would smooth out your portfolio ride. The trick is to smooth out the ride without the canceling out effect. Thus it would seem to me that you would be tilting more to certain factors in comparison to others.

I invest across factors and some of this was done by accident before I learned all this stuff. Not sure if it has boosted returns but I have likely achieved volatility reduction.
A fool and his money are good for business.

larryswedroe
Posts: 15989
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: Better growth? Or on how the value premium ends

Post by larryswedroe » Thu Aug 18, 2016 3:37 pm

Dabbling
First the value effect holds just as much for other metrics, as it does for BTM and the premiums and t-stats are just as strong. And the explanatory power is high.

Second, there are many logical economic reasons for value stocks to have a risk premium that are well documented in the literature and it's why beta doesn't work alone. I've listed some of them here, value stocks tend to have more financial and operating leverage, and those are clearly riskier companies which should require a premium especially when companies exhibit both of them. They also have fewer sources of capital and more capital that cannot be cut back on during recessions, so they tend to do poorly in bad times, and stocks that do that should require large risk premiums.
I could go on and discuss monetary policy risk issues, but I've done that many times and listed the more than dozen papers that explain the risk story.

Now I happen to agree with the view that there is also a behavioral story, it's hard not to. But IMO it's clearly not black or white, it's some of both.

Hope that helps
larry

larryswedroe
Posts: 15989
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: Better growth? Or on how the value premium ends

Post by larryswedroe » Thu Aug 18, 2016 3:44 pm

Bertilak
Here's a simple example that might help. You can try this with the regression tool at portfolio visualizer.
Run the factor regression for a TSM portfolio and of course you get 1 on beta and 0 on size, value, mom, quality, and low vol if you want all those factors, or just use four or five. You can use Vanguard TSM fund. Then take their say intermediate Treasury fund. And then have say a 60/40 portfolio. You'll find a loading of beta of .6 (1 x 60%) and loading of about .11 on TERM (.27 x 40%). So those are your only two factor exposures.

Now try a portfolio that is only 40% stocks but all DFSVX. And now 60% the Vanguard bond fund.
Now your beta exposure is about .44 (1.1 x 40%), your size exposure is about .38 (.96x 40%), your value exposure is .25 (.62 x 40%) and your quality exposure it about 0.1 (.22 x 40%) and your TERM exposure is now higher at .16 (same .27 but now x 60%).

So you can see much more of a risk parity strategy. This is like the famous Ray Dalio all weather portfolio of Bridgewater Associates. Same concept, not having all your eggs in the beta basket.

Hope that helps
Larry

larryswedroe
Posts: 15989
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: Better growth? Or on how the value premium ends

Post by larryswedroe » Thu Aug 18, 2016 3:47 pm

backpacker, sorry but you are incorrect in your statements. ONLY PART of the value premium is due to poor forecasts. the rest is a risk premium. That's it. Nothing more to say. It's that simple. Correcting the errors of the past even if you believe that only NARROWS the premium, and it eliminates the behavioral part if you really believe investors will change.
It's not complicated at all. Just answer this question for yourself. Should stocks that have more volatile earnings, earnings that are thus more difficult to forecast, and have more financial and operating leverage and more irreversible capital require a higher discount when you value the expected earnings. Unless you say no, and I'd love to hear you defend that, then the discussion is over because the higher discount rate is your expected return. That's the risk story for value stocks.

BTW- in your own example in the early stages of the correction the value premium rises, as the overpriced growth stocks do poorly and the underpriced value stocks do well (this is exactly by the way how part of the value premium is created, by migration). But then once the growth stocks fall in value and leave the top 30% of p/b then aren't growth any more and the value stocks that aren't in the bottom 30% aren't value any more. When both are in the middle (called core) there is no impact on value premium. So in that example at least until they reach the other extreme in your example the value premium goes up, not down.

Larry

User avatar
Topic Author
backpacker
Posts: 1620
Joined: Mon Sep 22, 2014 2:17 pm

Re: Better growth? Or on how the value premium ends

Post by backpacker » Thu Aug 18, 2016 4:20 pm

larryswedroe wrote:backpacker, sorry but you are incorrect in your statements. ONLY PART of the value premium is due to poor forecasts. the rest is a risk premium.
I don't have a dog in this particular fight. Take whatever fraction of the premium is behavioral and ask, for that fraction, whether the value spread would have to shrink for it to go away.
larryswedroe wrote: It's not complicated at all. Just answer this question for yourself. Should stocks that have more volatile earnings, earnings that are thus more difficult to forecast, and have more financial and operating leverage and more irreversible capital require a higher discount when you value the expected earnings. Unless you say no, and I'd love to hear you defend that, then the discussion is over because the higher discount rate is your expected return. That's the risk story for value stocks.
Depends on whether the extra risk is diversifiable. If yes, then I would not expect a premium. If no, then I would.

The insight of MPT is that the cost of capital for a company is not an intrinsic feature. You have to know how those risks interact with other risks investors take.
larryswedroe wrote:BTW- in your own example in the early stages of the correction the value premium rises, as the overpriced growth stocks do poorly and the underpriced value stocks do well (this is exactly by the way how part of the value premium is created, by migration). But then once the growth stocks fall in value and leave the top 30% of p/b then aren't growth any more and the value stocks that aren't in the bottom 30% aren't value any more. When both are in the middle (called core) there is no impact on value premium. So in that example at least until they reach the other extreme in your example the value premium goes up, not down.
What you say here is compatible with what I said about my case. When investors identify an overpriced growth stock and an underpriced value stock, fixing that will result in value investors making money and growth investors losing money. None of this challenges the basic point that the value premium can shrink without shrinking the value spread.

larryswedroe
Posts: 15989
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: Better growth? Or on how the value premium ends

Post by larryswedroe » Thu Aug 18, 2016 4:40 pm

backpacker, sorry but your last point is just wrong on the math. In that very example the spreads shrink. And it's because the discount rates used are changing. IT's all about the discount rate used. If you don't get that there's nothing more I can say. I give up. Maybe it's my inability to explain it well.

To add I've said for very long time the value premium is part behavioral and part risk. The behavioral story will not disappear due to limits to arbitrage, unless investors change their preferences and I don't see why that will happen, possible but not likely. So the premium could easily shrink but not disappear.

Those extra risks are CLEARLY not diversifiable, you cannot diversify more leverage both operating and financial for example, and nor can you diversify away the other risks I've noted. That is why the literature shows value is a unique factor not explained by beta.

User avatar
Topic Author
backpacker
Posts: 1620
Joined: Mon Sep 22, 2014 2:17 pm

Re: Better growth? Or on how the value premium ends

Post by backpacker » Thu Aug 18, 2016 5:50 pm

larryswedroe wrote:backpacker, sorry but your last point is just wrong on the math. In that very example the spreads shrink.
Say it takes two hours for the prices of Alpha and Beta to adjust. Alpha moves down continuously from $300 to $100 and Beta moves up continuously from $100 to $300. Your point is that for the first hour, the value spread will shrink a bit, then go back up over the second hour and end up back where it started. Sure. I don't think anyone one cares about that.

The point is that the value spread is the same size after the errors are found and corrected as it was before the errors were found and correct but, after the correction, the forward looking value premium has shrunk. So you have a smaller premium without a smaller value spread.

Maybe your idea is that if this were happening, we should be able to detect little "dips" in the spread during the process of correcting errors. I strongly doubt that given the size of individual pricing errors relative to the size of the overall market and the length of time over which they would be found and corrected.

larryswedroe
Posts: 15989
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: Better growth? Or on how the value premium ends

Post by larryswedroe » Thu Aug 18, 2016 7:11 pm

backerpacker
Your example is just wrong. Not how it works. In your example, the first case the value company got the higher FUTURE returns. And now in the second case the value company will get the higher FUTURE returns. The spreads are the same and the expected returns are the same going forward. Otherwise you are saying the risks of the two companies are the same. But that's not true in your own example as one is now growing faster and the other slower and delivered a negative shock. In your example you assume the same discount rates are applied to the expected cash flows. But they aren't. The market is pricing their risks very differently. That is what the research shows, the value companies are riskier. Now if you don't agree, read the research and explain where it's wrong. Explain as I said how companies with higher leverage, higher operating leverage, more irreversible capital and more volatile earnings are not riskier. Like I said, that's rhetorical. If you don't get that then I'm sorry,nothing more can say. So I really am done here. Maybe someone else can do a better job of explaining it.
Larry

User avatar
Topic Author
backpacker
Posts: 1620
Joined: Mon Sep 22, 2014 2:17 pm

Re: Better growth? Or on how the value premium ends

Post by backpacker » Thu Aug 18, 2016 9:36 pm

larryswedroe wrote: In your example you assume the same discount rates are applied to the expected cash flows. But they aren't.
Larry Swedroe: It sometimes feels like you just skim posts for keywords, then repeat vaguely related talking points. You're better than this. Is this how you interact with clients?

I used a single discount rate to keep the example simple. Reworking it with different discount rates is trivial.

Say that Alpha and Beta have the same past earnings of $10 and same expected future earnings BUT it turns out that Beta is a RISKY company that gets a higher discount rate. So Alpha is worth $200 and Beta $100. Growth stocks are (arbitrarily) stock with PE greater than 15 and value stocks those with a PE of less than 15. So Alpha is a growth stock and Beta a value stock.

The market then realizes that it was wrong about its growth estimates and correctly predicts that Alpha will have half the growth that was previously thought and Beta double the growth. So Beta goes to $200 and Alpha to $100, since it is still using RISKY discount rates for Beta and not for Alpha.

We get the same result. The future realized value premium shrinks, because the market corrects an error, and the value spread remains the same.

User avatar
nedsaid
Posts: 12668
Joined: Fri Nov 23, 2012 12:33 pm

Count Valula rises again

Post by nedsaid » Thu Aug 18, 2016 11:32 pm

Just for fun, let's look at the year to date performance of funds that I own personally to see how Value is doing in 2016.

US Total Stock Market Index +8.54%
Large Cap Plain Old Boring Value Fund +12.29% (active)
Vanguard Value Index +9.14%
Income and Growth +10.69% (active)

Hmmm. My Large Cap Plain Old Boring Value Fund is active and it is doing great in 2016! I benchmark this fund against Vanguard Value Index. You can see that the Value Index is beating US Total Stock Market by 60 basis points or 0.60%. So the value effect for 2016 isn't as great as I thought, but that Large Cap Plain Old Boring Value Fund is having a great year.

I also own Income and Growth, it is a quantitative fund that tries to replicate the S&P 500 with less risk and higher yield. It is a Value Fund and you can see that it is also beating US Total Stock Market and also beating the Vanguard Value Index.

I own all the funds above except for Vanguard Value Index.

International Index +2.83%
International Large Cap Value +10.50% (active)

It appears that Value is winning on the International part of the equity markets too.

Vanguard Small Cap Index +10.22%
Vanguard Small Cap Value ETF +12.33%

Just for fun:
International Mid/Small-Cap Blend +10.13% (active)
International Mid/Small-Cap Growth +3.13% (active)

Certainly, 2016 so far is a Mid/Small-Cap year both here in the US and Internationally. Value is doing well but not beating the indexes by as much as I thought.

So Count Valula isn't dead yet, showing some signs of life for 2016!
A fool and his money are good for business.

Waba
Posts: 377
Joined: Wed Feb 26, 2014 9:19 pm

Re: Better growth? Or on how the value premium ends

Post by Waba » Fri Aug 19, 2016 12:06 am

backpacker wrote:
larryswedroe wrote: In your example you assume the same discount rates are applied to the expected cash flows. But they aren't.
Larry Swedroe: It sometimes feels like you just skim posts for keywords, then repeat vaguely related talking points. You're better than this. Is this how you interact with clients?

I used a single discount rate to keep the example simple. Reworking it with different discount rates is trivial.

Say that Alpha and Beta have the same past earnings of $10 and same expected future earnings BUT it turns out that Beta is a RISKY company that gets a higher discount rate. So Alpha is worth $200 and Beta $100. Growth stocks are (arbitrarily) stock with PE greater than 15 and value stocks those with a PE of less than 15. So Alpha is a growth stock and Beta a value stock.

The market then realizes that it was wrong about its growth estimates and correctly predicts that Alpha will have half the growth that was previously thought and Beta double the growth. So Beta goes to $200 and Alpha to $100, since it is still using RISKY discount rates for Beta and not for Alpha.

We get the same result. The future realized value premium shrinks, because the market corrects an error, and the value spread remains the same.
So you say that Beta goes to $200 with expected earnings of $20, a PE of 10 (which makes it a value stock per your own definition) and presumably a future worth of $220, and Alpha goes to $100 with expected earnings of $5, a PE of 20 (which makes it a growth stock) and presumably a future worth of $105? Going 1x long Beta and 2x Short Alpha will then net me $10 in value premium. Count me in.

larryswedroe
Posts: 15989
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: Better growth? Or on how the value premium ends

Post by larryswedroe » Fri Aug 19, 2016 7:25 am

Backpacker


You keep making up contrived examples which are not valid. In your example a value stock should be a growth stock and vice versa. And you continue to ignore the risks. The value premium cannot disappear if there is risk. And the there is strong evidence as I said that it is risk and unique non-diversifiable So that alone should end the discussion unless you disagree, and as I said then explain why companies with clearly riskier characteristics are not more risky
Now back to your example.
As I stated the market is not that inefficient.
So let's try this one more time.
Growth stocks are in fact growth stocks, they have higher growth of earnings. It's just that the evidence shows the market overestimates their ability to grow. So as example, let's say the market forecasts growth of earnings of growth stocks of 7% and it turns out they only grow 6% and value companies instead of growing at 1% grow at 2%. So now the market wakes up and makes better forecasts. And gets it correct. The valuations of growth stocks comes down and the valuations of value companies goes up and the spread narrows. It cannot stay the same if forecasts improve.

And as I said, the return on investments is simply a function of the discount rate used. That's it. Your idea or statement about CAPM is false understanding. So since growth companies have more forecastable earnings (with less error) and the earnings flow of value companies is riskier (less forecastable) then it must be the case that the discount rate used for value stocks is higher and thus the expected returns are higher and that is reflected in the valuation spreads. Improving forecasts doesn't change that, just reduces the errors. Not makes them disappear. You will always apply a higher discount rate to riskier assets. And the common examples of simple risk characteristics of value companies are there and well documented. These are simple facts which is why I am having problem with the discussion.

Note none of this makes value better than growth, just riskier and higher expected returns as compensation, UNLESS you also believe the behavioral story, and in that case then value is at least partly a free lunch.

Larry

User avatar
nedsaid
Posts: 12668
Joined: Fri Nov 23, 2012 12:33 pm

Re: Better growth? Or on how the value premium ends

Post by nedsaid » Fri Aug 19, 2016 9:55 am

backpacker wrote:[
Say that Alpha and Beta have the same past earnings of $10 and same expected future earnings BUT it turns out that Beta is a RISKY company that gets a higher discount rate. So Alpha is worth $200 and Beta $100. Growth stocks are (arbitrarily) stock with PE greater than 15 and value stocks those with a PE of less than 15. So Alpha is a growth stock and Beta a value stock.
Hi Backpacker, interesting discussion. When determining whether a company is value or growth, you have to look beyond P/E ratios. Here is why. High P/E ratios can indicate investor optimism and high earnings growth or it can just indicate depressed earnings. That is the flaw in your analysis.

The financial media used to talk a lot about cyclical stocks, stocks that are very sensitive to economic cycles. Autos, homebuilding, forest products, mining would all be good examples of cyclical stocks. There is an old saying that you buy these when P/E's are high and you sell them when P/E's are low. When P/E's are high, this reflects both a depressed economy and depressed earnings. At some point, the stock price quits falling because the market knows that the economy will improve in the future and earnings will improve with it. During booming times, P/E's get to be low because the market perceives an earnings peak and knows that earnings won't grow much from there. The market knows the economy will cool someday.

So in practice, certain economically sensitive stocks could have high P/E ratios and yet be considered a value stock. Also, what happens if you have no "E" at all? P/E is an important metric but one needs to look at other metrics as well.
A fool and his money are good for business.

User avatar
Topic Author
backpacker
Posts: 1620
Joined: Mon Sep 22, 2014 2:17 pm

Re: Better growth? Or on how the value premium ends

Post by backpacker » Fri Aug 19, 2016 3:12 pm

nedsaid wrote:Just for fun, let's look at the year to date performance of funds that I own personally to see how Value is doing in 2016.
Interesting stuff! Thanks for this nedsaid. Sometimes it's good to stop staring at the theories in your head and take a good look at the actual world.
nedsaid wrote:
backpacker wrote:[
Say that Alpha and Beta have the same past earnings of $10 and same expected future earnings BUT it turns out that Beta is a RISKY company that gets a higher discount rate. So Alpha is worth $200 and Beta $100. Growth stocks are (arbitrarily) stock with PE greater than 15 and value stocks those with a PE of less than 15. So Alpha is a growth stock and Beta a value stock.
Hi Backpacker, interesting discussion. When determining whether a company is value or growth, you have to look beyond P/E ratios. Here is why. High P/E ratios can indicate investor optimism and high earnings growth or it can just indicate depressed earnings. That is the flaw in your analysis.
Hmmm..yes. The problem with trying to figure this stuff out is that you can't put the whole economy in your head at once, so you have to simplify. You have to cut things out until you have a small enough "world" to fit in your brain. But then there is always the question of whether you have just simplified away something important.

Setting aside the variable discount rates Larry likes to talk about, PE ratios are basically just a prediction of future growth compared to last year's growth (or whatever). But as you point out, future growth could be high relative to last year because last year was strangely low. Or it could be because the future looks especially good.
nedsaid wrote: So in practice, certain economically sensitive stocks could have high P/E ratios and yet be considered a value stock. Also, what happens if you have no "E" at all? P/E is an important metric but one needs to look at other metrics as well.
Maybe that brings us back the value of multiple sorts? You have a company whose earnings have just fallen off a cliff but should bounce back, so PE will be weirdly low. But book value should stay pretty steady, so we can see that they're a value company looking at that.

User avatar
Topic Author
backpacker
Posts: 1620
Joined: Mon Sep 22, 2014 2:17 pm

Re: Better growth? Or on how the value premium ends

Post by backpacker » Fri Aug 19, 2016 3:15 pm

Waba wrote:
backpacker wrote:
Say that Alpha and Beta have the same past earnings of $10 and same expected future earnings BUT it turns out that Beta is a RISKY company that gets a higher discount rate. So Alpha is worth $200 and Beta $100. Growth stocks are (arbitrarily) stock with PE greater than 15 and value stocks those with a PE of less than 15. So Alpha is a growth stock and Beta a value stock.

The market then realizes that it was wrong about its growth estimates and correctly predicts that Alpha will have half the growth that was previously thought and Beta double the growth. So Beta goes to $200 and Alpha to $100, since it is still using RISKY discount rates for Beta and not for Alpha.

We get the same result. The future realized value premium shrinks, because the market corrects an error, and the value spread remains the same.
So you say that Beta goes to $200 with expected earnings of $20, a PE of 10 (which makes it a value stock per your own definition) and presumably a future worth of $220, and Alpha goes to $100 with expected earnings of $5, a PE of 20 (which makes it a growth stock) and presumably a future worth of $105? Going 1x long Beta and 2x Short Alpha will then net me $10 in value premium. Count me in.
Right. As the errors are eliminated, value investors will make money. I agree with that. The question is, has that already happened? If it has, how would we know? One usual answer is that you compare the average PE of growth stocks to the average PE of value stocks. If the errors have been eliminated, the difference will have shrunk. I'm not so sure that's true though.

Basically, I don't have any idea whether the premium is dead, but I'm skeptical that we can use value spreads to confirm that it's alive.

User avatar
Topic Author
backpacker
Posts: 1620
Joined: Mon Sep 22, 2014 2:17 pm

Re: Better growth? Or on how the value premium ends

Post by backpacker » Fri Aug 19, 2016 3:56 pm

larryswedroe wrote: And you continue to ignore the risks. [...] Note none of this makes value better than growth, just riskier and higher expected returns as compensation, UNLESS you also believe the behavioral story, and in that case then value is at least partly a free lunch.
Take the correct expected future earnings for stocks. Apply whatever risk-based discount rates you like. This will give us what we can call the risk-adjusted "fair value" of the stocks. You think that value stocks have a bit of a free lunch, even if they aren't all free lunch. So we need to get a free lunch from somewhere. The way to do that is to introduce pricing errors. But what kind or errors? And where? Can we introduce them without increasing the spread relative to fair value? Because if we can introduce them without increasing the spread, we can eliminate said free lunch without decreasing the spread.

Mathematically, the answer is trivial. We can. The real question is whether there is a natural and systematic way to introduce pricing errors that don't increase the spread. I'm not sure I know the answer to that yet.
larryswedroe wrote: Growth stocks are in fact growth stocks, they have higher growth of earnings. It's just that the evidence shows the market overestimates their ability to grow. So as example, let's say the market forecasts growth of earnings of growth stocks of 7% and it turns out they only grow 6% and value companies instead of growing at 1% grow at 2%. So now the market wakes up and makes better forecasts. And gets it correct. The valuations of growth stocks comes down and the valuations of value companies goes up and the spread narrows. It cannot stay the same if forecasts improve.
Sure. But that could happen because five of the growth stocks grow 7% but there was one bad apple that grew 1%. And there were five value stocks that grew 1% but one golden apple that grew 7%. If the market figures out how to find the rotten apple and the golden apple ahead of time, it will put the golden apple in the growth basket and the rotten apple in the value basket. The value premium goes away but the spread stays put.

Growth stocks here really are growth stocks. They on average grow 5% faster than value stocks. And it's true that on average the market overestimates their ability to grow by 1%. And it's true that the market on average underestimates the ability of value to grow by 1%.

Now, maybe you think that the market couldn't be this inefficient. Sure. Maybe that's true. But there are lots of growth rates between 1% and 7% too. The market just needs to find a few 2% stocks that are really 1% stocks, a few 5% stocks that are really 6% stocks, and so on and so forth. All you need is the right permutation swapping stocks form the different baskets. The result eliminates the behavioral value premium but leaves the spread intact.

The right answer to these question depends to some extend on how much of the behavioral value premium is driven by a few migrants. If most of it is driven by a few migrants, the market can eliminate the premium by plucking out the migrants ahead of time. No need to bid up the average price of value or down the average price of growth.

User avatar
nedsaid
Posts: 12668
Joined: Fri Nov 23, 2012 12:33 pm

Re: Better growth? Or on how the value premium ends

Post by nedsaid » Fri Aug 19, 2016 5:30 pm

Backpacker, looking at this from 40,000 feet it appears that the value premium for Large Stocks is shrinking. You can also see that the Value premium is smaller for Large Stocks than for Small Stocks. From what I have been reading, it seems that the Value premium for Small Stocks is unchanged.

What accounts for this in my opinion is that Wall Street research mainly focuses on the Large Caps and trading volume for the market is mostly here. Over time, trading volume has risen and probably made the markets more efficient. Since the lion's share of the trading volume has gone to Large Cap, you would expect most of the market efficiencies to be there too.

My best guess is that Value wins over time because of human nature, human behavior, and human emotion. Most investors are not patient and tend to want to chase what has recently performed well. It just seems that investors gravitate towards the popular growth stocks. It also makes sense that Value stocks are riskier because of lower quality balance sheets and more volatile earnings. Wall Street likes steady and predictable. What Larry says about overshooting on earnings estimates for Growth Stocks make sense too. That is also an element of human nature. Large Growth stocks seem more like a sure thing and are perceived as safer. But "safe" isn't always safe.

My "Four Horsemen of Underperformance" of GE, Pfizer, AIG, and Microsoft show what happens when investors get too optimistic and bid stock prices too high. Though I bought these stocks at "bargain" prices, there was still too much expectations built in. In another thread, I did off the top of my head estimates of my annual average return from these stocks over the time I owned them. GE was less than 1%, Pfizer 3%-4%, AIG (a big minus here), and Microsoft at maybe 8%. Microsoft was flat for the first 7 years that I owned it. So this makes Larry Swedroe's point.

What is weird is that Value stocks have the reputation of being riskier but when you look at the actual standard deviations, Value most often is less volatile than Growth.

Pretty much, it is easier to beat low expectations than high expectations. In other words, the bar is set higher for Growth than it is for Value.

This isn't rocket science. You just think through how markets work and how humans behave and this seems pretty intuitive at least to me.
A fool and his money are good for business.

Post Reply