Skip to main content
Message Font: Serif | Sans-Serif
 
No. of Recommendations: 3
Robert Shiller of Case-Shiller Index and Irrational Exuberance acclaim has a
recent book titled:
"Narrative Economics: How Stories Go Viral And Drive Major Economic Events".
His thesis is that when a narrative/story/premise catches on across a large enough
segment of society, it can have a significant effect on people's behavior,
and this in turn can effect the performance of economies and markets.
His book does a good job of going through some real world historical examples.
The Florida land boom of the 1920's is a classic and very illustrative of the fact that
the behavior does not in any way have to be "Rational" in an economic sense to exert
such influence/impact.

One can speculate in thinking about some of the narratives out there today as to what
they might imply about our economic or political future, short or long term. Closer to home,
e.g., this Board, there is a growing narrative among financial advisors and asset managers that
Value now should be over-weighted in portfolios in favor of Growth. This is arising not just
from traditional Value managers, but from the broader class of advisors/managers.
The primary reason for this being that those who believe the market is rich/elevated can attribute this
to Growth stocks being the primary culprits, while Value stocks have lagged behind during this bull run.
Fearing a diminution in performance in Growth stocks, while wanting to remain bullish or,
at least fully invested, they believe Value stocks should once again offer a better opportunity than growth.
To the extent this narrative sustains itself, the money will go into these stocks, just as it flowed
into the Growth stocks while their narrative was in vogue.

So, this should bode well for BRK while the narrative persists.

Carl
Print the post Back To Top
No. of Recommendations: 31
I think there is a lot of merit to the idea of narratives driving the big money movements.
A "cousin" view is the reflexivity arguments of Mr Soros: usually fundamental business and economic results lead to price reactions, but sometimes the causality is reversed.
Prices can and do affect fundamentals.
The two notions are linked in many ways, but big price movements can certainly create their own narratives.

But I would quibble with the idea that "value is going to outperform soon" is a narrative that has much traction these days.
The problem with spotting the dominant narrative is that we're the fish trying to spot the water:
we're all swimming in it, so the thing that should be most obvious is not readily apparent.
It takes great presence of mind to spot what narratives are forming your own investment stance.

The narrative that I see winning over the universe (and dragging real world economics with it) is that you'd be a fool to invest in anything other than a low fee market tracker, now or ever.
Or, for those of a certain age, "Nobody ever got fired for buying SPY".
This narrative is so pervasive that one can be attacked for even suggesting it has limitations.

It's a very safe investment choice, but it does have that occasional decade that loses you money...negative real total return.
15% chance historically, before considering the effect of the valuation level at the start of the decade in question.
About 26% chance of a negative decade historically, starting on dates with trend earnings yield in the most expensive quartile.

Since I have the table in front of me, some interesting numbers:
average ten year forward real total return, based on starting trend earnings yield, for the S&P 500 and cap weighted predecessors from 1930--
Cheapest quartile of purchase dates: average 10.8%/year (min +3.2%, max 19.1%)
Second quartile: average 9.2%/year (min -1.4%, max 16.4%)
Third quartile: average 5.6%/year (min -3.9%, max 15.7%)
Most expensive quartile: average 2.0%/year (min -5.7%, max 9.1%)
Those figures use monthly start dates, so the extremes on the worst possible days are a bit wider than shown.
Needless to say, even if things are different this time, today definitely falls in the category of "most expensive quartile".
The rear view mirror has been good, though. Ten years ending December were 11.3%/year.
The market is 79% more expensive than it was then, based on trend earnings yield, so without multiple expansion the decade just ended would have been 5.0%/year.

Jim
Print the post Back To Top
No. of Recommendations: 2
The market is 79% more expensive than it was then, based on trend earnings yield, so without multiple expansion the decade just ended would have been 5.0%/year.

The fundamental problem with any of this repeated sentiment is that a view of "PE".
Just using one metric to explain "market is expensive" is incorrect.

Here are the PEs of some other companies with high PEs
FB: 33
AAPL: 25
AMZN: 94
NFLX: 94
GOOG: 31
MSFT: 33
Print the post Back To Top
No. of Recommendations: 21
The fundamental problem with any of this repeated sentiment is that a view of "PE".
Just using one metric to explain "market is expensive" is incorrect.
Here are the PEs of some other companies with high PEs...


I think you've kind of missed the point.
I'm not saying a high P/E means any particular firm is expensive.

Current earnings for a given company or industry don't tell a very reliable story.
Neither do smoothed earnings for a given company...things can change a lot.
Aggregate earnings for the entire broad corporate market are also problematic because of cyclical issues.

But smoothed aggregate earnings are both extremely meaningful, and (historically so far) extremely predictive.
Though one or two companies might have prospects of low earnings now and great earnings later, meaning they're cheap while looking expensive on earnings yields, the same can't be true of the economy as a whole.
The typical middle-of-the-pack company is only going to make so much money in a typical year, and that amount isn't going to rise very much more than GDP.
If some firms become wildly profitable, you can bet that some others are making very little, since aggregate profit as a share of GDP can rise only so far. (pretty far maybe, but not without limit).
As a result, the trend line through aggregate real earnings is a very stiff one.
It trends pretty darned well, once you take out the squiggles from the business cycle.
A decade of super-duper real profit growth might be 4% instead of 2%, not the sort of thing to drive multiples through the roof justifiably.

So, yes, it's only one metric.
But it's a really good one---for the broad market.

Note, this is not to say that a given valuation level means a given market move any time soon.
Maybe things will stay expensive for an arbitrarily long time, or get more expensive--who knows?
But it DOES tell you for sure that the broad US market is expensive.
As in,
(1) You're not getting very much in the way of earnings for your investing dollar these days (cyclically adjusted or not), and
(2) There are no prospects of aggregate corporate earnings growing fast enough to mean you'll get good earnings yield any time soon on today's purchase price.
The trend amount of earnings for the entire broad market just doesn't rise quickly over time, so a hope for a huge upswing would be delusional.
For one company or several, sure it can happen, but not an entire diversified economy in aggregate.

Of course, it goes without saying that in the long run only profits support value.
A high P/E makes absolute sense for a firm that isn't making profit now...but only based on the prospect of its making even more profit later.
You can change the time frame you're considering, but not what you consider.
By extension, the aggregate value of the broad market loosely tracks, and is ultimately capped by, a function of its future aggregate profits.

Jim
Print the post Back To Top
No. of Recommendations: 2
This is a bit outside the topic of the original post, but I’ve been wondering lately about the power of narratives translating to different actual economic outcomes.

The economy is driven by consumption and investment, but consumption and investment are driven by consumer and business confidence. What if narratives and confidence levels are more important drivers of actual economic outcomes than actual economic policies?

Imagine you have President and Administration X coming into office with a fantastic understanding of economic policy. His/her cabinet is loaded with some the best economic advisors in the world. They make all the correct “rational decisions” when it comes to setting the country’s economic priorities and policy.

However, the majority of consumers/business owners actually despise/fear President X. They have zero confidence in him/her. They think his/her policies are actually going to destroy jobs and the economy. So what do they do? Business owners slow or stop investing, they halt plans to expand capacity, build factories, hire more people, etc. As a result, the economy actually gets worse.

Now imagine 8 years later, President and Administration Y is elected to replace President and Administration X. President/Admin Y has near-zero understanding of economic policy. His/her cabinet is loaded with some the worst economic advisors in the world. They make all the wrong choices when it comes to setting the country’s economic priorities and policy.

However, consumers/business owners love and admire President/Admin Y. They have tons of confidence in him/her. They think his/her policies are going to create an economic boom not seen in a generation. So what do they do? Business owners invest rapidly, they take out loans to expand their capacity, build factories, hire more people, etc. As a result, the economy actually gets better.

Which administration is likely to lead to better economic outcomes? Which is more powerful: false narratives that inspire confidence or true narratives that may reduce confidence?
Print the post Back To Top
No. of Recommendations: 6
Which administration is likely to lead to better economic outcomes?

I'll take a stab at it:
Shorter term: #2. Animal spirits matter. Narratives drive real economic actions.
Longer term: #1. Eventually, sound policies matter more.
The Argentine government comes to mind.

In my more cynical moments I think that election cycles are designed by politicians specifically to land on the borderline between the two : )

Jim
Print the post Back To Top
No. of Recommendations: 0
I sense Carl embracing his bubble & JohnC trying to rationalize his away.

such fun
Print the post Back To Top
No. of Recommendations: 3
I think you've kind of missed the point.
I'm not saying a high P/E means any particular firm is expensive.
So, yes, it's only one metric.
But it's a really good one---for the broad market.

I disagree.

More firms are investing and improving their assets particularly in a low interest rate environment for an extended period of time in lieu of increasing earnings.
Collectively their PE may be high but I would not say that they are "more expensive".
Their assets are worth much more where the growth is.
Amazon and Google are good proxies for this.

Comparing PE is not conclusive.
Print the post Back To Top
No. of Recommendations: 22
Comparing PE is not conclusive.

Well, no. Comparing PE is not conclusive...for individual firms.
However, the smoothed earnings yield (P/E) for the broad market *is* pretty darned conclusive.
That's simply because the long run trend of aggregate earnings is pretty darned predictable.

Unlike the situation for a single company, a buyer of the broad market knows how much cyclically adjusted earning power they're
getting for the money invested (not much), and know it isn't going to grow fast enough to make today's price look like a deal.
Aggregate earnings don't, and can't, rise very quickly through time.
A typical decade has shown a rise of smoothed real earnings of about 2.3%/year, standard deviation of only 1.6%. Most recent figure 3.1%, pretty normal.
Just imagine trying to forecast the earnings for Netflix or Amazon ten years in advance with that degree of precision.
That's why the current trend earnings yield for the broad market can give you actual conclusions, while the same figure for one firm can't.

The trajectory of future earnings is really hard to estimate for a given company, but for the broad market not so much.
What's hard to estimate is what multiple Mr Market will offer on that trend line of earnings in any given year.
We know for sure that things are expensive, both relative to history and in absolute terms.
(the trend earnings yield is under 3% on my figures, not exactly a huge return even before you compare it to the 16% on offer in summer 1982)
We don't know anything for sure about how long that will last.

More firms are investing and improving their assets particularly in a low interest rate environment for an extended period of time in lieu of increasing earnings.

True enough.
But you can only increase assets from earnings, so it's just another way of saying the same thing.

Value comes from current assets, and the trajectory of future owner earnings.
For some valuable firms earnings can arrive far into the future, but that wheeze doesn't work for the broad market.

Jim
Print the post Back To Top
No. of Recommendations: 1
"But I would quibble with the idea that "value is going to outperform soon" is a narrative that has much traction these days."

Jim-You are missing my point. Of course indexing is dominant today.
I was referring to a growing narrative among "active" investors i.e., non-indexers,
putting more weight on Value.
I am not crazy about the business media channels,but when I do my
treadmill workout I listen to the Bloomberg channel.
One of the costs of listening to the "news" part of the channel is having to listen
to the talking heads with market opinions. Many of these are active managers,
and they have been increasingly talking about putting higher weight on Value since
last fall. This means they are putting money there. If they continue with this view,
more money will continue to flow in to Value, relative to most of this bull cycle.
It does not mean indexing into the S&P 500 is going away. All I am saying is these active managers are putting more weight, and thus money, into Value. And, probably some indexers are doing it as well, i.e., via Value index funds, for the same reason. Thus, BRK should benefit.
Carl
Print the post Back To Top
No. of Recommendations: 4
Many of these are active managers, and they have been increasingly talking about putting higher weight on Value since last fall. This means they are putting money there.

Yes, I see that too.

Yet...
Even if that were the dominant theme, and turned into real action, I just wonder whether it's meaningful any more in the grand scheme of things.
The active management industry is fading so fast, to me it seems that they aren't what moves the needle on anything these days.
Collectively, they're like one biggish investment club. Do their movements really matter any more?

Prices are determined by who is willing to buy and/or trade, and today I perceive that as index trackers, pension and institutional funds (see above), and a very small number of really rich folks.
The index trackers and closet index trackers can't move individual pries relative to peers, only asset class prices.
One wonders what is moving the prices of individual stocks.
It's unlikely to be people like me, buying individual stocks for a personal account.

To get an idea of the index industry size these days, more than half the volume on the French market is now in the daily close interval.
That's when trackers "true up" their positions, to ensure they're tracking perfectly.

Jim
Print the post Back To Top
No. of Recommendations: 0
Well, no. Comparing PE is not conclusive...for individual firms.
However, the smoothed earnings yield (P/E) for the broad market *is* pretty darned conclusive.
That's simply because the long run trend of aggregate earnings is pretty darned predictable.


If you take a collection of 100 IBMs and 100 AMZN, it would appear that IBMs are cheaper. Wouldn't it ?

Not sure how difficult it is for you to do this, but if you take FAAMNG out of the data set you have, does it still say that today's market is expensive* ?

* then we can blame it on TSLA
Print the post Back To Top
No. of Recommendations: 9
If you take a collection of 100 IBMs and 100 AMZN, it would appear that IBMs are cheaper. Wouldn't it ?

Sure.
But there aren't 100 IBM's.
When we look at the S&P 500 or Wilshire 5000 or whatever, we're looking at pretty much everything.
It's by definition average. It doesn't look like IBM, and it doesn't look like Amazon.

The whole point of looking at the entire market is that it includes a very large subset of everything going on in the economy.
That means that we know that aggregate revenues will more or less track GDP.
(rising exports do add a bit of a boost, but this is a surprisingly small factor through time)
Corporate profits are a finite subset of GDP, so we know they're going to be relatively tame as well. Couple of percent a year, give or take.
Profits are an unknowable fraction of revenues, but the range is bounded. It'll always probably be maybe 8% plus or minus 8%.
So, we get a modest range on top of a trend line with an almost-known slope...long run, pretty predictable.

Not sure how difficult it is for you to do this, but if you take FAAMNG out of the data set you have, does it still say that today's market is expensive* ?

Nothing I've said has any relevance to any one, or all, of the FAAMNGT group. Or any single company or industry.
It's purely a comment on the broad cap-weighed US equity market, as a whole.
Indeed, the whole point is that things that you can say for sure about the broad market (it's very expensive) don't apply with any certainty to individual stocks.
The rule for the whole doesn't necessarily apply to the specific, and the even the technique we used for the whole doesn't work when used with the specific.
We know for absolute certainty that the broad cap weight US market is very expensive.
You get very little earning power for your dollar, and that aggregate earning power will continue to grow quite slowly.
That says nothing useful about any specific firm.

if you take FAAMNG out of the data set you have, does it still say that today's market is expensive* ?

To take this question literally, the answer is "no". (though with less certainty).
They are hard to value, but there is no doubt that some of them have stupendous earning power.
On some metrics, the worst overvaluation is among the bread-and-butter middle-of-the-road boring firms.
They've run up too much debt, sales growth is poor, and profit margins are not rising any more.
It's things like the proverbial Bert and Eddie's screen door company you have to worry about.
(well, they're all large caps, so maybe the B&E Screen Door Warehouse Chain)

That's the main reason the Russell 2000 has been underperforming the S&P 500 for ages now; it seems some of the overvaluation is already wearing off among the small and boring.
(and the reason I used the R2K for hedging!)

Jim
Print the post Back To Top
No. of Recommendations: 2
"More firms are investing and improving their assets particularly in a low interest rate environment for an extended period of time in lieu of increasing earnings."

That is not how accounting works.
Print the post Back To Top
No. of Recommendations: 2
"The rear view mirror has been good, though. Ten years ending December were 11.3%/year.
The market is 79% more expensive than it was then, based on trend earnings yield, so without multiple expansion the decade just ended would have been 5.0%/year."

Thank you, Jim. It's very instructive to see the individual components of total return calculated like this: growth in fair value, dividend, multiple expansion (maybe even inflation and stock repurchases if desired). Could you please elaborate on "trend earnings yield?" It looks like you had to smooth earnings (or earnings yield) to get a useful measure of growth in fair value. I you had just used the ttm earnings as of 12/31/19 ($138.91) and 12/31/09($50.97) as a measure of fair value growth, you would have gotten a very high estimate of fair value growth of 10.5%/year. 5% including dividend seems more reasonable. Or did I misunderstand?
Print the post Back To Top
No. of Recommendations: 10
Could you please elaborate on "trend earnings yield?" It looks like you had to smooth earnings (or earnings yield) to get a useful measure of growth in fair value.
I you had just used the ttm earnings as of 12/31/19 ($138.91) and 12/31/09($50.97) as a measure of fair value growth, you would have gotten a very high estimate of fair value growth of 10.5%/year.
5% including dividend seems more reasonable. Or did I misunderstand?


Yes, I used smoothed real earnings.
Like the E10 used by CAPE, just a little smoother.
Using a single year of earnings (instantaneous P/E) doesn't give a usefully predictive model, as earnings are too volatile.
They're all but useless as a valuation metric.

Note, CAPE is itself giving slightly odd results right now.
It says the market is 34% more expensive than its average since 1990 based on the E10 earnings yield, which a is startlingly low figure.
But it's good to remember the theory behind it: average a decade of earnings so you get some good years and some bad years, in an attempt to estimate a normal year.
(technically it estimates what a normal year would be now, minus five years of normal growth, because of the age of the figures in the average)
But there hasn't been a bad year in the last 10 years, so CAPE is overestimating earnings and underestimating
overvaluation at the moment compared to its usual behaviour, since it's merely averaging a bunch of bull market years.
(as an aside, ignore those who say that you SHOULD ignore the credit crunch profit drop.
It was far from the largest earnings recession on record, and the average is *supposed* to include an earnings recession)
I find a 17 year WMA works a little better for smoothing the real earnings, for that and other reasons.
It has the advantage of a bit more history being considered, and no "cliff edge" when the oldest year drops out of the average.
The method I use is actually the average of four smoothing methods, including WMA17 and traditional E10.
Since the four have different lags and earnings grow over time, I scale each of them separately to give the best match to current earnings before I average them.
The earnings level it comes up with is an attempt to estimate what you'd see right now if this year were neither unusually good nor unusually bad.
At year end 2019, measured in year-end 2019 dollars, it estimates the "normal" earning power of the S&P 500 at about $109.

That's up a remarkable 3.66% from a year earlier, five year rate of change 3.20%/year.
Earnings have been so strong lately, and for so long now, that even the heavily smoothed trend lines are bending upwards.
It would not be prudent to extrapolate those slopes.
Real profit growth is not going to exceed real sales growth by several percent a year forever.
For comparison, real (smoothed) earnings per S&P 500 point grew 1.89%/year in the 40 years 1975-2005, and the prior 40 years 2.36%/year.
Historically, "a little over 2%/year" is normal for real earnings growth of the cap weighted broad US market.
If that rate seems low, don't panic.
That's not the same as market total returns, which also include dividends--you get the earnings, not just the earnings growth rate.
You get your salary, not just the annual percentage raise.

Jim
Print the post Back To Top
No. of Recommendations: 3
Jim,

I'm a big fan of your posts, and you've taught us all a lot on this Board over the years.

I interpret your comments as suggesting 3.6% or 3.2% real EPS growth rates are too high to expect going forward, and the future will look more like the historic norm of "a little over 2%/year." If I misinterpreted that, please do let mew know.

I am REALLY struggling to understand why it makes sense to assume future real EPS growth for the S&P 500 will be more similar to the two periods you mentioned, 1975-2005 & 1935-1975, rather than what we've seen since 2005.

First, I’m more or less in agreement with your normalized S&P 500 EPS figure of $109. That comes out to a Return on Equity of 11.8% ish which is on the low side of the past 10-20 years, but may be reasonable looking forward given where we are in the current business cycle. Leverage ratios are also pretty normal right now, so that ROE figure looks like a reasonable long-term base to start from.

Now let's see if we can take that starting point, layer on a few assumptions, and see what kind of normalized EPS growth we might expect.

The S&P 500 pays out roughly 37% of it’s earnings as dividends and plows back (retains) the remaining 63%. So if we take our 11.8% normalized ROE figure X 63% plow-back ratio, we’d end up with a sustainable (nominal) EPS growth rate of about 7.4% for the S&P 500. A reasonable estimate for future real EPS growth rate would therefore seem to be roughly 7.4% minus whatever your inflation assumption is. If you assume say 2% inflation, that gives you an estimated (whopping) 5.4% real EPS growth rate going forward, which is way WAY higher than we’ve seen in the past. It's also much higher than even the 3.6% real growth we've seen recently.

I double-checked the S&P 500 real EPS growth rates you provided (1.89%/year in the 40 years 1975-2005 and the prior 40 years 2.36%/year) and they are indeed correct.

How then can it be that normalized real EPS growth rates are RADICALLY higher in recent times vs yesteryear?

A few possibilities come to mind:
• Significantly lower effective US corporate tax rates today compared to yesteryear
• Significantly lower dividend payout ratios, [which mean increased % plowback ratio (retained earnings) which all else equal, leads to higher EPS growth rates] compared to yesteryear.
• Significantly lower inflation rates compared to yesteryear.

This suggests a combination of higher nominal EPS growth rates AND lower inflation rates leading to much higher real EPS growth rates.

Returning to your expectation of future real EPS growth rates for the broad US Index being closer to the 2+% historic average rather than the 3.6% recent average (and certainly WAY lower than the 5.4% I came up with)...how do we get there?

I would have to assume real EPS growth rates that low (<3.5%) will require some mix of the following:
• Normalized nominal returns on equity being much lower than 11.8% going forward. (maybe due to higher future effective corporate tax rates or something else?)
• Significantly higher dividend payout ratios and lower earnings plowback (retention) ratios reduce future nominal EPS growth rates.
• Future inflation is way higher than 2% and dragging down the real EPS growth rate.
• Some other factors that I missed?

Jim, I appreciate your engagement on this topic because it's critically important. A 2% real EPS growth and a 5% real EPS growth lead to radically different valuations if we're looking out over 25+ years.

Please let us know why we should NOT assume real EPS growth rates will continue to be higher (3-6%) going forward for the broad US Index than they have been historically (2-3%)

Thanks,
John

https://commons.wikimedia.org/wiki/File:US_Effective_Corpora...
https://d2wsh2n0xua73e.cloudfront.net/wp-content/uploads/201...
http://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/...
https://www.yardeni.com/pub/spxratios.pdf
https://www.investopedia.com/terms/s/sustainablegrowthrate.a...
Print the post Back To Top
No. of Recommendations: 0
John, a 4 word response may suffice for your roughly 600 word question:

Regression To The Mean
Print the post Back To Top
No. of Recommendations: 2
I am REALLY struggling to understand why it makes sense to assume future real EPS growth for the S&P 500
will be more similar to the two periods you mentioned, 1975-2005 & 1935-1975, rather than what we've seen since 2005.


Simple.
The recent upswing in the rate of real profit growth is transient statistical noise, due to the big recent "kinda cyclical" upswing in net margins.
That's ultimately just a transient thing, because net margins can't grow faster than sales without bound.
Even if they stay high for a long time, we'll see that the rise was a one time thing, not a new higher growth rate.

Real aggregate profits over time won't grow any faster than real aggregate sales, which won't grow any faster than real GDP, which seems to grow at around 2%, give or take.

Net margins have high decades and low ones, but the range is ultimately bounded.
Thing of aggregate net profits as mainly tracking aggregate sales, which are relatively smooth, plus or minus squiggles for profitability variation through time.


I would have to assume real EPS growth rates that low (<3.5%) will require some mix of the following:
• Normalized nominal returns on equity being much lower than 11.8% going forward. (maybe due to higher future effective corporate tax rates or something else?)
• Significantly higher dividend payout ratios and lower earnings plowback (retention) ratios reduce future nominal EPS growth rates.
• Future inflation is way higher than 2% and dragging down the real EPS growth rate.
• Some other factors that I missed?


Not really.
All it would require is sales growing at around the same rate they have been, and net margins NOT continuing to rise to ever higher levels.
(recent numbers are the highest in decades, about 1950 depending which precise figures you use)
S&P 500 real sales have risen at 2.0%/year since late 2002.
S&P 500 real sales have risen at 2.0%/year since mid 2012.
Only 1.27%/year since the end of 1999.
So, if net margins flat line at this high level and sales continue to grow at around that 2% rate, profits will grow at that rate.
If net margins mean revert at all, profits will grow more slowly while that happens. Potentially a lot more slowly.
It's not rare to see a decade of no growth in real (smoothed) earnings per S&P 500 point.

Separately, you can't compare the real EPS growth figure to the ROE directly.
There are a few extra moving parts.

Please let us know why we should NOT assume real EPS growth rates will continue to be higher (3-6%) going forward for the broad US Index than they have been historically (2-3%)

In short, it simply can't happen unless either net margins keep rising and rising, up beyond 100% of sales, or sales growth picks up well beyond historical norms.
Which would require GDP growth to pick up to almost unheard of levels.
Sales growth has in fact been pretty good just lately, but that seems to be mostly recovery from a Loooong stretch of doldrums.
The pre-credit-crunch real sales level wasn't exceeded till March 2018.
So the rates over long baselines are probably a better guide, as above.
Which makes sense...they're close to real GDP growth figures.

Jim
Print the post Back To Top
No. of Recommendations: 0
"Yes, I used smoothed real earnings..."

Thank you for elaborating, Jim.
Print the post Back To Top
No. of Recommendations: 1
Thanks Jim, that helped a lot.

I was going to ask about whether a large portion S&P 500 sales growth coming from outside the US might bend/break the rule about real S&P 500 sales growth not being able to be higher than real US GDP growth, but I doubt the impact of that would be enough to make any meaningful difference.

Therefore I must concede that you're most likely correct about future real EPS growth being significantly worse going forward.

Thanks for changing my view.
Print the post Back To Top
No. of Recommendations: 1
Stocks have returned 6.5% real (Siegel's constant) yet GDP and aggregate real sales have not risen by that for a long time (since the 1960s). The way to reconcile these is greater retention ratio, expanded P/Es, higher margins. The first two obviously cannot rise indefinitely. But this is the era of high fixed costs (of building websites, apps, cloud infrastructure) and close-to-zero variable costs. So I don't see why profits would not grow a lot faster than sales.
Print the post Back To Top
No. of Recommendations: 16
Stocks have returned 6.5% real (Siegel's constant) yet GDP and aggregate real sales have not risen
by that for a long time (since the 1960s).
The way to reconcile these is greater retention ratio, expanded P/Es, higher margins.
... So I don't see why profits would not grow a lot faster than sales.



I think you have to think a bit more clearly about how the market works.
The rate of return on stocks is not expected to be, and is not, the rate of rise of the value of the market index.
Each year you get the earnings on your holdings, not the rate of increase of the earnings of the holdings.
Over the very long run, real broad market earnings rise something near 2%/year, not 6.5%/year.

Historically speaking, Siegel's constant of around 6.5% real return comes from the sum of numbers like 1.5-2.5%/year real earnings growth for the index plus 4-4.5% current dividend yield.
If you're not getting dividends similar to that, you're not going to get real total returns near Siegel's constant.

The value of the index tracks the real cyclically adjusted earnings of the index, which typically rises around 2% a year, give or take.
Absent changes in valuation multiples, in any given holding period your real total return equals the rise in the index value plus your dividend yield.

Dividend yield are under 2%/year now, so even with no change in market valuation multiples you should expect maybe 4%/year in real total returns.
And less, if valuation multiples drop a bit in the direction of historical norms.

The bigger point:
Profits can't rise as a percentage of sales over time, except in the "up" portion of cyclical variation.
Even a thousand years from now, profits will not exceed 100% of sales. Probably never more than 20%, at a guess.
As real GDP generally rises only in the vicinity of 2%/year, that's the best net profits do as well, over the long run.
Hey, be a huge optimist, say 3%.

If real earnings are rising at <<3% a year and dividends are getting you <<2% a year, you have no reason at all to expect a real total return from the broad market of more than <<5%/year.
Anything better than that is likely transient and cyclical multiple expansion.
We've seen a lot of that lately.
And it might be much worse in a plausible stretch of multiple compression.

If you want to reverse engineer reasoning to support an expectation of higher returns, you need a sustained and much higher rate of real GDP growth than ever seen, and/or ever rising net margins beyond their current 70 year highs.
Those aren't likely, so real total returns for the broad US market over 4-5% aren't likely.
I think anything over real 3%/year would be a wonderful result for the next 5-10 years.

Jim
Print the post Back To Top
No. of Recommendations: 0
I think you have to think a bit more clearly about how the market works.
The rate of return on stocks is not expected to be, and is not, the rate of rise of the value of the market index.
Each year you get the earnings on your holdings, not the rate of increase of the earnings of the holdings.
Over the very long run, real broad market earnings rise something near 2%/year, not 6.5%/year.


Sorry, I don't follow.
If the market index rises by 10% in a year (ignoring dividends for a minute) isn't that the rate of return on the stock prices? What else is it?
Similarly, if P/E is constant and ignoring buybacks, isn't the percentage rise in the stock prices the same as the percentage rise in earnings i.e. earnings growth?

Look-through earnings are not mine, they are for the company to dispose of as they see fit. My return comprises of dividends, multiple expansion, and earnings growth. I believe the late Jack Bogle agrees.

The bigger point:
Profits can't rise as a percentage of sales over time, except in the "up" portion of cyclical variation.
Even a thousand years from now, profits will not exceed 100% of sales. Probably never more than 20%, at a guess.
As real GDP generally rises only in the vicinity of 2%/year, that's the best net profits do as well, over the long run.
Hey, be a huge optimist, say 3%.


No doubt. But a company with a 5% profit margin and no variable costs, will double its stock price if the sales rise by 5% and that 100% is the return that stockholders will experience.
($100 sales - $95 cost = $5 profit; $105 sales - $95 cost = $10 profit; constant P/E).
Print the post Back To Top
No. of Recommendations: 0
If you want to reverse engineer reasoning to support an expectation of higher returns, you need a sustained and much higher rate of real GDP growth than ever seen, and/or ever rising net margins beyond their current 70 year highs.
Those aren't likely, so real total returns for the broad US market over 4-5% aren't likely.
I think anything over real 3%/year would be a wonderful result for the next 5-10 years.



Wait a minute...I thought that as long as central banks keep lowering interest rates, that we can count on the same kind of returns as we've seen since 2009, forever and ever.
Print the post Back To Top
No. of Recommendations: 7
Over the very long run, real broad market earnings rise something near 2%/year, not 6.5%/year.
...
Sorry, I don't follow.
If the market index rises by 10% in a year (ignoring dividends for a minute) isn't that the rate of return on the stock prices? What else is it?
Similarly, if P/E is constant and ignoring buybacks, isn't the percentage rise in the stock prices the same as the percentage rise in earnings i.e. earnings growth?



If the index rose at 10%/year, you'd be right.
But it doesn't.

First, remember we're talking about real returns, so don't forget to take out inflation.

Over the last many years, we have seen the broad US equity index rise for a combination of these reasons:
(1) tracking the rising smoothed earnings power of its constituents, best thought of as rising real sales per index point;
(2) a boost from rising net margins in recent years for reasons other than tax cuts (higher profit per dollar of sales lately);
(3) a one time boost for the discontinuous rise in value from the tax cuts at end 2017; and
(4) getting more expensive.

#1 is around 2%/year, give or take. It tracks real GDP, basically, plus or minus very small factors for exports and imports, changing share of GDP from public companies, etc.
Those can be big theoretical holes in GDP as a yardstick, but surprisingly small factors when measured.
#2 has brought that rate up to 3-4%/year lately, but that will halt. It may or may not become a drag factor, impossible to know.
Net margins may stay at these 70 year highs, or may fall back somewhat towards old levels, but they won't rise by half again.
One can not assume a perpetually falling labour share of income.
#3 I estimate this as a one time rise in value of 10%.
That is, I'm estimating a dollar of US real sales post 2017 is permanently worth 10% more in net after tax profit than a dollar of sales before that.
But feel free to plug in your own figure if you have a better guess. But it's not a trend.
#4 is the rest of the index return lately. Simply rising valuation multiples: higher prices for the same sized stream of future earnings.
Like #2, multiples might go up a bit more, or stay the same, or fall somewhat in the direction of historical norms, but can't rise forever.
I certainly expect that current valuation levels are above whatever the average turns out to be in the next 15 years.

In short, looking to the future, the only component which you can be assured will continue is #1.
The rest can't be extrapolated.
To the extent they are cyclical in nature, you may end up seeing them as reductions to #1 rather than increases as they have been recently.

It's true that companies reinvest profits. Buybacks count in that.
But, it is very important in any discussion to avoid double counting the reinvestment.
That money comes from the earnings, which you've already counted and whose trend you're already tracking.
For example, remember that a dollar spent on buybacks reduces the market cap of a firm.
If done at a fair value, it leaves the value of each remaining share unchanged, but there are fewer of them afterwards.
And there are offsetting destructions of profits, like all buybacks done at above fair value, gifts of shares to employees, overpaying for acquisitions of firms outside the index, and so on.
It is possible for sales per index point to rise faster than aggregate corporate sales of the same firms over a long period, but, to make a long story short, they don't.

Jim
Print the post Back To Top