The Single Greatest Predictor of Future Stock Market Returns

8/15/2019 The Single Greatest Predictor of Future Stock Market Returns

http://slidepdf.com/reader/full/the-single-greatest-predictor-of-future-stock-market-returns 1/22

1

Philosophical Economics

The Single Greatest Predictor of Future Stock Market Returns

Posted on December 20, 2013

Consider the following chart, which shows the average investor portfolio allocation to equities

from January 1952 to December 21!"

#he metric in this chart ta$es no input from any variables traditionally associated with

valuation" earnings, boo$ values, profit margins, discount rates, etc% &t consists only of a simpleratio between two numbers that can easily be calculated in '()D% *et, as a predictor of future

stoc$ mar$et returns, it dramatically outperforms all other stoc$ mar$et valuation metrics

commonly cited%



2

&n this piece, &+m going to do five things% 'irst, &+m going to eplain, in very simple terms, the

accounting principles behind the metric% #he eplanation will include instructions -with ready.

made lin$s/ for how to graph the metric in '()D% 0econd, &+m going to discuss the dynamics

of asset supply, with a special focus on equities% #hird, &+m going to challenge the conventional

framewor$ for understanding the relationship between valuation and stoc$ mar$et

returns% 'ourth, &+m going to introduce a new framewor$, one that relates stoc$ mar$et returnsto equity asset supply% 'ifth, &+m going to present a scatterplot of the predictive performance

of the metric alongside other metrics, and discuss what the metric is currently forecasting for

%0% equity returns% &+m going to conclude by briefly touching on the question of whether or

not the current %0% stoc$ mar$et is overvalued%3

Accounting Principles: Cash, Bonds, Stocks

#o begin, let+s arbitrarily divide the universe of financial assets into three categories" -1/ cash,

-2/ bonds, and -!/ stoc$s% 4y cash3, & mean ban$ deposits and circulating currency% 4y

bonds3, & mean any certificate of obligation to repay borrowed cashcommercial paper, bills,

notes, bonds, etc% 4y stoc$s3 -or equity3/, & mean shares of ownership in a corporation -public

or private/% 6ote that these definitions are intentional simplifications%

'inancial mar$ets function on the following principle% 'or every unit of every financial asset

in eistence, some investor somewhere must willingly hold that unit in a portfolio at all

times% 4y investor3, & mean whoever owns wealth% #here are intermediarieshedge funds,

mutual funds, pension funds, financial advisors, etc%that help investors allocate wealth% 4ut

these entities are not the actual investorstheir clients are%

#he financial mar$et is the place where investors decidevia tradeswho will hold what units

of what assets% 6ote that cash, as an asset, is special in that respect% &t is the medium through

which trades occur% &nvestors can only switch from one stoc$ or bond to another stoc$ or bond

by going through cash% #he going rate of echange -bid or offered/ between a unit of an asset

and cash is the mar$et price of the asset%

7t the margin, if no investor can be found that wants to hold a given unit of a given asset at the

prevailing mar$et price, then the mar$et price will fall until a willing holder is found% 8ith

respect to shares of a stoc$ or bond, the application is straightforward% &f no one wants to hold

a given share at 1, then we try 95% 0till no ta$ers: #hen we try 9, then ;5, then ;,

and so on% 8e continue until some investor emerges that finds the share sufficiently attractive

to hold at the offered price% #he concept applies analogously to cashif no investor wants to

hold cash, then the price that is bid on everything else will rise until everything else becomes

so epensive and unattractive that some investor somewhere capitulates and agrees to holdcash instead% <easured in terms of other assets, the price of cash falls%



3

#he supply3 of an asset is the total mar$et value of it in eistencethe total number of

outstanding units times the mar$et price of each unit% =ut differently, supply is the amount of

the asset available to be held in investor portfoliosthe amount available for investors to

allocate their wealth into% &n aggregate, investors have to want to hold the total supply of each

asset in eistence in their portfolios% &f there is too much supply of a given asset relative to the

amount that investors want to hold in their portfolios, then the the mar$et price of the asset will fall, and therefore the supply will fall% &f there is too little supply of a given asset relative to

the amount that investors want to hold in their portfolios, then the mar$et price will rise, and

therefore the supply will rise% >bviously, since the mar$et price of cash is always unity, 1 for

1, its supply can only change in relative terms, relative to the supply of other assets%

The Aggregate Investor Allocation to Equities

6ow, suppose that we open up every investor+s portfolio and calculate, for each investor, his

percent allocation to stoc$s, bonds, and cash% <y portfolio might be allocated ;5? to stoc$s,

15? to bonds, ? to cash% *ours might be allocated 5? to stoc$s, 2? to bonds, !? to

cash% 7nd so on%

#he question we want to answer is this" what would the average of all of these investors+

portfolio allocations loo$ li$e, weighted by si@e: <ore specifically, what would the average

investor allocation to stoc$s be: 7nd how would that average compare to the averages of the

past: &t turns out that this question predicts the mar$et+s future long.term returns better than

any other classic valuation metrics to date developedprice to earnings -=A)/, price to boo$

-=A4/, price to sales -=A0/, C7=), q.ratio, <ar$et Cap to BD=, 'ed <odel, etc%

#o answer the question, we need to $now two things" -1/ the total amount of stoc$s that

investors in aggregate are holding, and -2/ the total amount of cash and bonds that investors

in aggregate are holding% <athematically, the total amount of stoc$s that investors are holding

divided by the total amount of everything -stoc$s plus bonds and cash/ that they are holding

ust is the average investor allocation to stoc$s%

6ow, to calculate the total quantity of cash and bonds in investor portfolios, we might thin$

that we can ust sum the total quantity of cash and bonds in eistence outrightthe total

amount floating around the economy% 7fter all, these securities have to be held by

investors% 4ut this approach won+t wor$% #he reason is that a large portion of the bonds in

eistence are actually held by ban$s, not by investors% #his fact etends to the central ban$ -the

'ederal (eserve/, which presently owns an unusually large quantity of bonds%

'ortunately, there+s a convenient way to get around the problem% (ecall that when the 'ederal(eserve buys bonds -treasury, <40, etc%/, it doesn+t add any net financial assets to investor



4

portfolios% (ather, it ta$es bonds out of investor portfolios, and puts newly created cash into

investor portfolios% &t changes the cash.bond mix of the assets that investors holdbut not the

total amount%

&t turns out that private ban$s do essentially the same thing when they buy assets% #hey ta$e

the assets out of the hands of investors, and put their own liabilitiesin the form of their own bonds or deposits -cash/into investor hands% -#hey can also fund purchases with equity sales,

but the equity component of a ban$s balance sheet is small enough to ignore%/

#he entities that create net new financial assets -that investors can hold/ are not ban$s, which

are ust intermediaries, but rather real economic borrowers% #he universe of real economic

borrowers consists of five categories" ouseholds, 6on.'inancial Corporations, 0tate and

Eocal Bovernments, the 'ederal Bovernment, and the (est of the 8orld% 8hen these entities

borrow directly from investors, the investors get new bonds to hold% 8hen the entities borrow

from ban$s, the investors get new cash to hold% #hat+s because when a ban$ ma$es a loan, the

money supply epands% #he loan creates a new deposit that didn+t previously eistsome

investor must now hold that deposit in his portfolio of assets%

&t follows, then, that if we want to get an estimate of the total amount of bonds and cash that

investors are holding at any given time, all we have to do is sum the total outstanding liabilities

of each of the five categories of real economic borrowers% #hose liabilities either translate into

cash that an investor somewhere is holding -if the entity too$ a loan from a ban$, which

epands the money supply/, or they translate into a bond that an investor somewhere is holding

-if the entity borrowed directly from the investor/% 6ote that the average bond trades close to

par -with some above, and some below/, so, in aggregate, the value of the liabilities

approimates the total mar$et value of the bonds%

4an$s don+t generally hold stoc$s% 0o to estimate the total amount of stoc$s in investor

portfolios, what we need to $now is the total mar$et value of all stoc$s in eistence% 8e end up

with the following equation"

&nvestor 7llocation to 0toc$s -7verage/ F <ar$et Galue of 7ll 0toc$s A -<ar$et Galue of 7ll

0toc$s H #otal Eiabilities of 7ll (eal )conomic 4orrowers/

8e can get all of the information in this equation from the 'low of 'unds report% #he

information is also conveniently available in '()D Braph% 7 lin$ to the calculated metric is

provided1 , and to a separately downloadable version of each series2%

1 http://research.stlouisfed.org/fred2/graph/?g=qis

2 http://research.stlouisfed.org/fred2/graph/?g=qab



5

6ow, the (est of the 8orld creates an interesting complication% =arts of our portfolios are

composed of stoc$s, bonds and cash denominated in foreign currencies -which do not show up

in these series and are not being counted, though they should be/% 4ut in the same way, some

parts of the portfolios of individuals in other countries are composed of stoc$s, bonds and cash

denominated in our currency -which do show up in these seriesand are being wrongly

counted, given that our goal is to $now our own allocations as domestic investors/% 7s anestimation, it wor$s to assume that the two cancel each other out%

The Unique Dna!ics o" Equit Asset Suppl

#he supply of cash and bonds that investors in an economy must hold perpetually increases

with the economy+s growth% #he cash and bonds in investor portfolios are literally made from3

the liabilities that real economic borrowers ta$e on to fund investmentthe fuel of growth%

#he following chart shows the annual growth of the total liabilities of all real economic

borrowerswhich, again, is the total supply of cash and bonds in investor portfoliosfrom 1952

to present% #he growth rate has ranged anywhere from around 5? per year to around 15? per

year% (ight now, it+s at the low end of the spectrum%

#rivially, if the aggregate investor is going to maintain a constant portfolio allocation to

equities, the supply of equities must grow commensurately with the supply of cash and



6

bonds% (ecall that investors, in aggregate, have to hold all of these assets at all times% &t follows

mathematically that the ratios of the total supplies outstanding must equal the ratios inside the

average3 investor+s portfolio%

#he supply of equities can increase in one of two ways" through the issuance of new shares, or

through price increases, i%e%, increases in the level of the stoc$ mar$et% #he chart below showsthe corporate sector+s net issuance of new equity, as a percentage of total mar$et value, bac$ to

195%

7s we see in the chart, the corporate sector is inherently averse to the issuance of new

equity% )ach year, it adds very little additional supply, on net% &n various periods since the

early 19;s, it+s actually been a net destroyer of equity supplyta$ing supply off the mar$et

through acquisitions and buybac$s%

7s we eplained in our earlier piece on earningless bull mar$ets!, because the corporate sector

does not issue sufficient amounts of new equity each year to $eep up with the continually

increasing supply of cash and bonds, stoc$ prices have to rise over the long.term% &f they don+t,

stoc$s will become a smaller and smaller percentage of the aggregate investor portfolio% nless

investors, on average, want stoc$s to be a smaller component of their portfoliosbecause, for

eample, they increasingly prefer to hold other assetsthis outcome will not be allowed% 0toc$

3 http://philosophicaleconomics.ordpress.com/2!13/!"/2!/earningless#bull#mar$ets#h%#do#the%#happen/



&

prices will get pushed up on the growing relative scarcity until the aggregate equity allocation

preference is satisfied%

#aluation: Challenging the Conventional Understanding

#he total return of an equity security depends on two factors" -1/ the change in price frompurchase to sale, and -2/ the dividends paid in the interim% Dividends matter, but price is

$ing% &t drives total return%

<any investors don+t li$e the fact that price drives total return% &f price drives total return, it

follows that total return is a function of the shifting sentiment, preferences and epectations of

other peoplethose who ma$e up the mar$et and vote3 on what the price will be% &nvestors

don+t want their returns to be subect to the arbitrary vote3 of other people, and so they pretend

that as stoc$ mar$et speculators they are actually genuine businessmen who buy3 and own3

companies to hold forever% #hey tell themselves that their returns will somehow emerge

directly from the cash flows of the underlying businesses, regardless of what the mar$et decides

to do with price%

#his point of view ignores the fact that it ta$es decades to recoup an equity investment via

dividends, the only cash flows that are ever are actually paid out to buy.and.hold investors% #o

claim a return on a stoc$ in any other contet, an investor needs someone to sell it to% #he price

that other people are wiling to pay is therefore importantsupremely important% (ather than

resist this fact puristically, our responsibility as investors is to accept it and wor$ within it, by

understanding the behavioral propensities of our fellow mar$et participants, and getting in

front of emerging trends in how they choose to allocate their wealth%

>nce we agree that price is $ing, the net question is" how is price determined in a

mar$et: Galue mavens tend to thin$ that price is determined through the rational3 application

of normative valuation principles, such as #he stoc$ mar$et+s =A) ratio should be 15, plus or

minus a few points% &f interest rates are low, add a few points% &f they are high, ta$e a few points

off%3 >n this view, when the actual =A) ratio is above the appropriate value, disciplinedinvestors sell% 8hen it+s below that value, they buy% #hrough their buying and selling, the price

moves to where it should3 be, to fair value3, given the earnings% )very so often, emotions

disrupt the process, but as with everything, they eventually pass, and the process ta$es hold

again% Galue mavens loo$ for these disruptions as an opportunity to capture ecess return%

#here+s certainly some truth to this view, but it doesn+t give the whole story% ltimately, the

price of equity is determined in the same way that the price of everything is determinedvia

the forces of supply and demand% 'or any given stoc$ -or for the space of stoc$s in aggregate/,price is always and everywhere produced by the coming together of those that don+t own the



"

stoc$ and want to allocate their wealth into it, and those that do own the stoc$ and want to

allocate their wealth out of it% &f there is a different supply sought by the first group than offered

by the second, the price will shift until the imbalance equali@es%

6ow, there+s absolutely nothing that says that this process has to equilibriate at any specific

valuation% istory confirms that it can equilibriate at a wide range of different valuations% 'orperspective, the average value of the =A) ratio for the %0% stoc$ mar$et going bac$ to 1;I1 is

$%&%'% 4ut the standard deviation of that average is a whopping (&), more than 5? of the

mean% >ne standard deviation in each direction is worth *)+ in total return, or $+ per year

over 1 years%

#he same is true of the popular 0hiller C7=)% &ts long.term average is $%&+'but with a

standard deviation of -&%, again almost 5? of the mean% >ver the last 1 years, its value has

stretched from as low as %, to as high as )'a difference of .'' in total return% 6ote that

the periods in which it too$ on depressed values were hardly brief% &t spent the entire decade

of the 19s at bargain basement levels, frequently falling into single digitsthis in an

environment where interest rates were pinned at @ero%

7gain, it+s all up to the allocatorsthey decide how much of their wealth they are going to

allocate into stoc$s, how much eposure they are going to ta$e on% #heir preferencesor rather,

their efforts to put those preferences in place, by buying and sellingset the price% Galuation is

a byproduct of this process, not a rule that it has to follow% &n the 19s, investors decided, for

whatever reasonmemories of the Depression, a 8orld 8ar that the country might have lost,

price controls, high inflationthat they didn+t want large stoc$ mar$et eposure% #he fact that

bond yields were meager did little to alter this preference% 7nd so valuations stayed etremely

depressed% 8hen a vibrant, prosperous peacetime economy emerged in the 195s, this

preference obviously changed, and the biggest bull mar$et in history ensued%

4uy.and.hold is painted as the informed, responsible, pro.7merican thing to do with a

portfolio% 4ut, in terms of financial stability, it can actually be a very destructive

behavior% Consider the classic buy.and.hold allocation recommendation" -' to stoc$s, )' to bonds -or cash/% 8hat rule says that there has to be a sufficient supply of equity, at a fair3

or reasonable3 valuation, for everyone to be able to allocate their portfolios in this

ratio: #here is no rule%

&f everyone were to ump on the buy.and.hold bandwagon, and decide to allocate KA, but

equities were not already K? of total financial assets, then they would necessarily become

K? of total financial assets% #he ecess bidding would not stop until they reached that level% &t

doesn+t matter that the associated price increase would cause the =A) ratio to rise to anobscenely high value% #he supply.demand dynamic would force it to go there%



'

6ow, in the real world, valuation concerns can and do push bac$ on the equity allocation

process% 4ut, outside of etremes, they don+t tend to push bac$ with very much force, at least

not on their own% Eet me now eplain some of the reasons why%

8e can divide asset allocators into two types" mechanical allocators, and active

allocators% <echanical allocators are individuals that adhere to a strict allocation formula,regardless of circumstance% #wo eamples would be buy.and.hold investors that are always

1? invested -or always KA invested, periodically rebalancing, etc%/, and 1LAretirement

investors that invest automatically in accordance with a pre.defined program% #hese asset

allocators follow their processes come rain or shine, therefore they cannot be relied upon to

push bac$ against valuation ecesses% #hough they are not the maority of the mar$et, they are

a significant part of ittheir presence ma$es a difference%

7ctive allocators, in contrast, dynamically alter their allocations so as to maimi@e their

returns% ow do they try to maimi@e their returns: 4y allocating their wealth into the assets

whose returns they consider to be the most attractive, adusted for ris$% &t+s a competitive

processthey choose among their options, based on their assessments of what those options

are li$ely to produce%

0ome might interpret this to mean that they loo$ at the earnings yields on stoc$s, the yields to

maturity on bonds, and the yield on cash, and then choose% Eet+s suppose that asset allocation

were this easyust find the asset class with the highest yield, ris$.adusted, and allocate into

it% 8e would still have to answer the question" what is the future yield -at the current price/ of

each asset class, adusted for ris$: #o answer this question with respect to cash is hardwe

have to estimate future short.term interest rates% 8ith respect to bonds, even harderwe have

to estimate credit ris$% 8ith respect to stoc$s, the hardest of allwe have to estimate forward

earnings%

#o estimate forward earnings for stoc$s, we have to answer difficult questions about the future"

8hat will the traectory of nominal growth be: ow will profit margins evolve: 8ho can

answer these questions with a significant degree of empirical confidence, enough to be acontrarian that consistently fights the mar$et+s trends: Gery few people, and therefore the

answers to the questions end up reducing to biased reflections of prevailing mood,

etrapolations of recent eperience% 8hen the mood is high, and when recent eperience has

been positive, investors embrace optimistic assessments of what the future holdstherefore,

equities loo$ cheap, attractive% #he mar$et gets the opposite of the valuation pushbac$ that it

needs% 8hen the mood is low, and when recent eperience has been negative, investors

embrace more pessimistic assessments of what the future holdstherefore, equities loo$

epensive, unattractive% 7gain, the mar$et gets the opposite of the valuation pushbac$ that itneeds%



1!

&t turns out that even if fundamental questions about the future yields of cash, bonds, and

equities were resolved, asset allocation still would not be as simple as choosing the security that

offers the highest yield -ris$.adusted/% #he goal, again, is to maimi@e return% (eturn is not

the same thing as yield%

Branted, if a security is held to maturity, or, in the case of equities, for an infinite period oftime, the return will mathematically converge on the yield -provided, in the case of equities,

that all earnings are eventually distributed as dividends/% 4ut who among us buys bonds to

hold to maturity, or stoc$s to hold forever: <ost investor time hori@ons are not on the order

of decades, centuries or infinity, but on the order of days, months and yearsa few days -the

time hori@on of a swing trader/, a few months -the epiration date on a portfolio manager+s

grace period with clients, at which point they will start leaving if things aren+t wor$ing/, or a

few years -long.term value investors playing with their own money/%

#o $now the return of a security on a daily, monthly, or yearly time hori@on, it+s not enough to

$now what the yields are% *ou need to $now how the price is going to change% Biven future

cash flows -which we+ll assume you+ve accurately estimated/, this requires $nowing what the

future valuation will be%

#o illustrate, suppose that the =A) multiple on stoc$s is 2, the 1 year bond yield is 2?, and

the rate on cash is ?% 0uppose further that the earnings of each of these securities are going

to remain constant% Can you say which security will offer the highest return over the net few

years: *ou might say stoc$sthe earnings yield is 5?, a healthy !? more than bonds% #he

premium3 between the two is meaningfully higher than the historical average% 4ut that doesn+t

tell you what the return of stoc$s will be% &f the investment mood sours ever so slightly over the

net few years, and the mar$et concludes that a =A) of 1I is more appropriate3 than a =A) of

2, the return will be negativema$ing stoc$s significantly less attractive than the other

available assets, despite the higher earnings yield%

#he only way that you can $now what the future valuation of stoc$s will beso as to estimate

future returnsis to apply some conception of what+s fair, appropriate, reasonable, normal% 4utthe range of what can be rationali@ed as fair, appropriate, reasonable, normal is etremely wide,

too wide to be useful, and far too wide to provide reliable pushbac$ against a supply.driven

mar$et advance% 7ny number that is chosen will li$ely be nothing more than a reflection of the

prevailing allocation preferencethe prevailing appetite to be in or out of the asset class, based

on primordial hunches3 for where things are headed, themselves ust manifestations of

recency bias% >nce again, the mar$et will not get the valuation pushbac$ that it needs%

ltimately, valuation is a learned perception, learned through a process of social andenvironmental reinforcement% #he part of it that is not learned is ust a crude manifestation of



11

the behavioral bias of anchoringudging the attractiveness of a price -or a ratio/ by comparing

it to the price -or ratio/ that one is accustomed3 to seeing% &ronically, it is anchoring, not

valuation discipline3, that $eeps the mar$et from doing cra@y, bubbly things% =eople don+t li$e

to pay higher prices tomorrow than they could have paid today, or sell for lower prices today

than they could have sold for yesterday% #hat+s true regardless of what any valuation metric

says%

#o illustrate, suppose that you spend a significant amount of time in an environment where the

average valuation is 25 times -or more/% *ou acclimati@e to that valuation, it becomes your

anchor, what you are used to seeing% 7ll of the pundits3 that you watch on #G tell you that it+s

normal% 7ll of your friends, your fellow investors, say that it+s normal% <ost importantly,

whenever you+ve bought at or below that valuation, it+s worked the mar$et has rewarded you

with a positive outcome% 7nd so you+re comfortable buying at that valuation%

>bviously, in such an environment, you will come to perceive 25 times earnings as a perfectly

appropriate3 price for the mar$eta fair3 multiple% &f given an opportunity to buy the mar$et

at a lower pricefor eample, 2 timesyour reward circuitry will fire off, creating an appetite

to loc$ in the bargain3, ump on the big gains3 that it is offering%

4ut now switch the =A) in the eample from 25 to 15% 0uddenly, the same =A) of 2 will ma$e

you feel li$e you+re overreaching, eposing yourself to danger, buying too high% Bee, what if

the =A) falls bac$ to 15, where it usually is, what &+m used to seeing&+ll lose 25? in one moveM &

can+t afford that%3

4ecause valuation is a learned perception, driven by anchoring and by social and

environmental feedbac$, it tends to follow the mar$et% 7s valuations rise in a bull mar$et, prior

anchors wear off, and people get accustomed to higher valuationsover time, the valuations

stop feeling high3ma$ing room for them to go even higher% #heir perceived appropriateness

gets reinforcedsocially, in the mar$et discussion, and environmentally, through the incredibly

powerful feedbac$ of actually making money% &n a long, slogging bear mar$et, the opposite

occurs% )verything gets driven downwards%

#o illustrate, consider the eample of the most recent cycle% #here was a time, before the crisis,

when we tal$ed about trailing =A)s of 1; or 19 times earnings as reasonablemaybe even a

bargain, relative to the bubble that we had previously come out of% &f we thought the trailing

=A) was too high in the summer of 2I, we were told to ignore itbecause the mar$et was still

very cheap on forward estimates -themselves ust a reflection of the optimism/%

#hen, we had the Breat (ecession, a massive, negatively.reinforcing series of economic eventscompletely unrelated to stoc$ mar$et valuation, mind youthat shattered everyone+s equity



12

world view% 8e suddenly found ourselves seriously debating whether 1 times trailing

recessionary earnings was appropriate% 8e were supposedly in a new normal3, which, we

feared, implied structurally lower valuations%

#he crisis eventually abated, and the economy entered a recovery% 4ut people still had to wor$

off their fear conditioning and their anchoring% 7s the mar$et rose, the discussion shifted to whether 12 times was appropriate% #hen, 1 times% 6ow, 1I times% #he anchor, the goalpost,

has continued to move with the mar$et% =retty soon, the discussion will come full circle again,

and we will be as$ing ourselves whether 1; or 19 times is appropriate% 7nd, if the cycle isn+t

cut short by eternalities, as it was the last time, we may one day find ourselves discussing the

appropriateness of 25 timeswhich has been debated before in mar$et history -a few years

before the mar$et proceeded to go to /%

#he drivers of this recurring pattern are obvious" not some innate sense3 of fair value3, but

anchoring and the social.environmental reinforcement of the mar$et cycle itself% #he

perception of valuation is not capable of creating persistent, reliable resistance to the mar$et

cycle because it is an evolving function of the mar$et cycle% 7t etremes, it can push bac$but

it can+t push bac$ when it falls within the very wide range of what can be rationali@ed, which is

where it usually falls, and where it is now%

ltimately, we should be s$eptical of claims that investors are innately hardwired to act in a

certain way in response to any specific concept, argument, or data pointwhether it be

valuation3, or anything else% &nvestors elicit behavioral responses to these types of

informational inputs, but the responses are not innate% #hey are learned from the environment

through a process of conditioning and reinforcement%

&nvestors attend to concepts, arguments, and data points, etc% as a means to an endthe end of

predicting what the return will be, which, in practice, means predicting where the price is

headed% &f investors already have a hunch for where the price is headedwhich they often do

they will choose to embrace whatever concepts, arguments, data points, etc% fit that hunchor

they+ll ust ignore the mumbo umbo3 altogether, and go with their feel%3 0imilarly, if they are ust using the constructs to save face in social debateto avoid having to admit to themselves

and to others that they are wrongthey will ump on whatever concepts, arguments, data

points, etc% show that they are right -and ignore everything else/%

8hen investors don’t have a hunch for where prices are headed, and are genuinely trying to

use concepts, arguments, data points, etc% to assess what to do, the ensuing assessment ends

up being something inherently insecure, subect to constant feedbac$ and molding from the

mar$etresponsive to the result , and ditched when no longer wor$ing%



13

*ou might confidently thin$, for eample, that good obs number3 means the recovery is

pic$ing up steam3, and that you should increase your equity ris$but if the mar$et starts

consistently telling you that this is wrong, by its actual result , you will be affected by the

feedbac$% *ou may eventually find yourself pulled to function in accordance with the opposite

rule, that good is bad%3 3<aybe & shouldn+t rush to increase my eposure here% <aybe these

good obs numbers will lead the 'ed to tightenmaybe that+s why the mar$et is sellingoff% >ops%3

7dmittedly, when enough people grab onto and act on concepts, arguments, data points, etc%,

they can become powerful forces that drive mar$et outcomes, especially when they have a basis

in reality that gives them credibility and forces people to believe them% #he refleivity of price

confirmation increases their allure and persuasiveness, which causes more people to latch onto

them, which fuels further price changes, therefore more price confirmation, and so on in a

feedbac$ loop that continues until reality pushes bac$%

owever, it+s hard for valuation to pic$ up steam in this way, because unli$e other themes that

might move mar$ets, it has no obective basisit+s a personal opinion, easy to dismiss% &t

represents a resistance to what the mar$et itself is doingand is therefore already on the road

to being disconfirmed simply by the fact that it is being raised% &f valuations are too high, then

why are we where we at them: 8hy aren+t we falling: 7bsent some $ind of confirmation or

feedbac$, the theme can+t go anywhere%

>utside of cyclical downturns in which profits themselves plunge, valuation never enters the

discussion as a surprise, an insult3, but rather is only introduced gently, gradually, as the

mar$et advancesusually by those who are not part of the advance% <ar$et participants

therefore have time to acclimati@e to it as a theme% &t can+t produce the $ind of shoc$ and

surprise that would catch people offsides and provo$e mounting, refleively self.fulfilling

reactions% #hat+s why it usually ta$es a recessionor some $ind of noious catalystto unwind

a valuation ecess% #he ecess alone can+t correct itself%

&n the tech bull mar$et, the overvaluation theme had a hard time pushing bac$ even when inde=A)s were in the !s and s% =eople tal$ed about overvaluation, they worried about it% 4ut

then they $ept watching the price go upso what do you do: *ou don+t tell the mar$et that it+s

wrong, you trust your environment, you go with the flow% &n the end, it too$ a tight 'ed, a

recession with falling earnings, a slew of corporate ban$ruptcies and scandals, unfamiliar

accounting changes that led to further earnings plunges, a terrorist attac$, a war in the <iddle

)ast, and so on, to finally get the mar$et moving reliably in the downward direction, so that the

valuation ecess could be corrected%





15

7s equity investors, we tal$ a lot about asset allocation% &t+s essentially the most important

aspect of portfolio managementhow we+re allocated within the space of individual stoc$s and

bonds, and across the space of assets in general% &+m ;5? equity, 15? cashAbonds% *ou+re 5?

equity, 5? cashAbonds% Joe over there is 1? equity, ? cashAbonds, etc%

8hat+s funny is that we never thin$ to as$" how is it possible for all of us to get to within areasonable range of these preferred allocations at the same time: 7fter all, we+re trading a

limited supply of things amongst each other% #he answer, of course, is that the supply, properly

understood, automatically shifts to meet our allocation preferences via the changes in price

that we cause when we try to put those preferences in placethat is, when we buy and sell at

the margin% &n bull mar$ets, we frequently find ourselves searching for opportunities to put

our allocation preferences in placeour equity eposures are rarely as high as we would li$e

them to be%

& therefore propose a new way of framing equity total returns% #a$e the previous equation, and

substitute 7ggregate &nvestor 7llocation to 0toc$s3 and &ncrease in 0upply of Cash and

4onds3 for 3=A) <ultiple Change3 and )=0 Browth%3 8e then have,

-1/ #otal (eturn F =rice (eturn H Dividend (eturn

-2/ =rice (eturn F =rice (eturn from Change in 7ggregate &nvestor 7llocation to 0toc$s H =rice

(eturn from &ncrease in Cash.4ond 0upply -(eali@ed if 7ggregate &nvestor 7llocation to 0toc$s

8ere to 0tay Constant/ H Dividend (eturns

Combining -1/ and -2/,

-!/ =rice (eturn F =rice (eturn from Change in 7ggregate &nvestor 7llocation to 0toc$s H =rice

(eturn from &ncrease in Cash.4ond 0upply -(eali@ed if 7ggregate &nvestor 7llocation to 0toc$s

8ere to 0tay Constant/ H Dividend (eturn

&n the previous way of thin$ing, the earnings grow normally as the economy grows% &f themultiple stays the same, the price has to risethis price rise produces a return% 8hen the

multiple increases alongside the process, the return is boosted% 8hen it decreases, the return

is attenuated% #he multiple is said to be mean.reverting, and therefore when you buy at a low

multiple, you tend to get higher returns -because of the boost of subsequent multiple

epansion/, and when you buy at a high multiple, you tend to get lower returns -because of the

drag of subsequent multiple contraction/%

&n this new way of thin$ing, the supply of cash and bonds grows normally as the economygrows% &f the preferred allocation to stoc$s stays the same, the price has to rise -that is the only



16

way for the supply of stoc$s to $eep up with the rising supply of cash and bondsrecall that the

corporate sector is not issuing sufficient new shares of equity to help out/% #hat price rise

produces a return% 8hen the preferred allocation to equities increases alongside this process,

it boosts the return -price has to rise to $eep the supply equal to the rising portfolio

demand/% 8hen the preferred allocation to equities falls, it subtracts from the return -price

has to fall to $eep the supply equal to the falling portfolio demand/%

6ow, instead of saying that the =A) multiple is mean.reverting, we say that, for a given set of

environmental contingenciese%g%, history, culture, demographics, etc%the equity allocation

preference is mean reverting% &t rises in epansionary parts of the cycle, as people become more

optimistic about the future and more eager to maimi@e what they see as attractive returns

-Lelly, we believe in this bull mar$et, we+re fully invested, our clients are fully invested%3

something you hear frequently on C64C these days/, and it falls in contractionary parts of the

cycle, as people become less optimistic about the future and more concerned about protecting

themselves from losses -<aria, we+re cautious here, we+ve raised cash, we want to see signs of

stabili@ation before we deploy it%3something that you heard frequently on C64C in +; and

early +9/%

&f you buy in periods where the investor allocation to equities is low, you will get the dividend

return plus the price return necessary to $eep the portfolio equity allocation constant in the

presence of a rising supply of cash and bonds, plus the price return that will occur when equity

allocation preferences return to more normal levels% *ou will get in front of the equity supply

squee@e of the net bull mar$et, when ris$ appetite and the associated desire to be invested in

equities recovers% #hus your return will be higher than normal% #his is what happened to

investors in the 19;s%

&f you buy in periods where the investor allocation to equities is high, you will get the dividend

return plus the price return necessary to $eep the portfolio equity allocation constant in the

presence of a rising supply of cash and bonds, but then you will have to subtract the negative

price return that will occur when equity allocation preferences fall bac$ to more normal

levels% #his is what happened to investors in the 21.2! bear mar$et%

#his way of thin$ing about stoc$ mar$et returns accounts for relevant supply.demand

dynamics that pure valuation models leave out% #hat may be one of the reasons why it better

correlates with actual historical outcomes than pure valuation models%

&t can eplain, for eample, the earningless bull mar$et of the 19;s% nbe$nownst to many,

earnings were not rising in the 19;s bull mar$et% #hey actually fell slightly over the period

4 http://philosophicaleconomics.ordpress.com/2!13/!"/2!/earningless#bull#mar$ets#h%#do#the%#happen/



1&

which is unusual% 4ut prices didn+t carethey s$yroc$eted% #he =A) ratio ended up rising well

above 2, despite interest rates near 1?a valuation disparity never before seen in

history% Galuation purists can+t eplain this movethey have to postulate that the common

sense3 rules of valuation were temporarily suspended in favor of investor cra@iness%

4ut if we loo$ at what investor allocations were bac$ then, we will see that investors werealready dramatically underinvested in equities% &f prices hadn+t risen, if investors had instead

respected the rules of valuation3 and refrained from ac$ing up the =A) multiple, the etreme

underallocation to equities would have had to have grown even more etreme% &t would have

had to have fallen from a record low of *% to an absurd $+ -see blue line in the chart below,

which shows how the allocation would have evolved if the =A) multiple had not

risen/% >bviously, investors were not about to cut their equity allocations in half in the middle

of a healthy, vibrant, inflation.free economic epansiona period when things were clearly on

the up% 7nd so the multiple eploded%

6ow, recogni@e that this framewor$ leaves plenty of room to ac$nowledge the relevance of

classical valuation considerations% Disparities in valuationbetween equities and their own

history -the valuation levels investors are anchored to, accustomed to seeing, that they considerto be normal3/ and between equities and other asset classes -bonds and cash/can certainly

cause investors to want to change their allocations and eposures, especially when the

disparities are significant and can+t be dismissed or rationali@ed away% &f such a change unfolds,

prices will rise or fall accordingly% 4ut if such a change doesn+t unfold, then prices are not going

to respond% 6or should3 they%

&n a way, the metric already offers a rough estimation of classical valuation% &f the average

investor allocation to equities is abnormally low, then prices are probably abnormally lowthe

mar$et+s probably cheap% Ei$ewise, if the average investor allocation to equities is abnormally

high, then prices are probably abnormally highthe mar$et+s probably epensive% 7nd so an



1"

investor that is value.sensitive can still use the metric as a way of assessing the mar$et

opportunity%

Co!paring the 4etrics on Per"or!ance

#he following chart is a scatterplot of the new metric% #he y.ais is 1 year 0=O total return,

the .ais is the average investor equity allocation% #he solid red line is the current value of the

metric% 6ote the ecellent fit%

(ight now, at its current value, the metric suggests a future 1 year nominal total return for

equities of around -% istorically, whenever the mar$et was at the current level, the low end

of the return was a tad less than %, and the high end was around 5%

#he following charts show scatterplots of the other metrics% &t+s not even worth speculating on

what returns they are suggesting right now, because the fits are atrocious, especially in the

current valuation range% #he )quity P.ratio, for eample, puts the mar$et+s current future 1

year returns anywhere from as low as * to as high as 5% #he 0hiller C7=) puts the returns

anywhere from as low as ' to as high as $'with an ironic bias to the upside% <ar$et Cap

to BD= puts returns anywhere from 6+ to * -which is why it has become fashionable among

bears/%



1'



2!

&n our earlier piece5

, we pointed out that the classic 0hiller C7=) wrongly labeled the <arch2! mar$et as significantly overvalued, and the <arch 29 mar$et as barely below fair value

-an epic, inecusable blunder/% 8e pointed out that one advantage of the pro.forma C7=),

which tried to eliminate accounting inconsistencies, was that it correctly identified the mar$et+s

attractive valuation in these periods% &t called <arch 2! a decent value, and <arch 29 a

screaming buy%

&t turns out that li$e the pro.forma C7=), this metric also called 2! and 29 correctly% &t

signaled the <arch 2! mar$et as a reasonable buy, and the <arch 29 mar$et as a

screaming buy, on par with levels seen at the secular low of the last bear mar$et, 19;2%

5 http://philosophicaleconomics.ordpress.com/2!13/12/13/shiller/



21

A /ote on 78vervaluation9

#here+s a raging debate right now between bulls and bears over whether the %0% stoc$ mar$et

is presently overvalued% #he debate rages on because the term is poorly defined% 8hat,

precisely, does it mean to say that something is overvalued3:

8hen we say that the stoc$ mar$et is overvalued3, we might mean that it+s currently valued

more epensively than it typically has been in the past% >ver its history, the %0% stoc$ mar$et

has offered, on average, some epected total returnsay ( to $'% 4ut now it+s priced for

% or - -using our metric/% 0o it+s overvalued%3

'air enough, bulls shouldn+t disagree% #here are tons of reasons why the present stoc$ mar$et

is unli$ely to produce the ( to $' returns that it has produced, on average, throughout

history% >n almost every relevant measure, it+s starting out from a higher.than.average level%

#he more important question, however, is this" why should the stoc$ mar$et offer investors

the average historical return right now: &f, over the net 1 years, bonds are offering

investors *&(, and cash is offering them less than $, why should stoc$s be priced to offer

them ( to $':

ow would that even be sustainable: &f equities were offering an ;? to 1? return, we would

all choose to allocate the bul$ of our portfolios into them, rather than languish in the Q&(=*

nothingness of bonds and cash% #here obviously isn+t enough equity supply for all of us to

allocate in that way, and so the price would get pushed up, and the epected return pulled

downvery quic$ly%

6ow, it+s a mista$e, obviously, to ma$e an assessment of valuation based strictly on a

comparison between the yields of stoc$s and bonds, as the 'ed <odel suggests we do% #he yield

of an equity security, again, is not the same as its return% *ou can buy the mar$et at !! times

earningsa !? earnings yieldbut your return over the net 1 years isn+t going to be !?% &t

will probably be ? -or less/, as the mar$et contracts from the obscene valuation at which you bought it% &f you were to try to ustify the stoc$ mar$et+s price by comparing its !? yield to the

1 year bond yield at 1?, touting the healthy ris$ premium -2?greater than the historical

average/, you would obviously be ma$ing a huge mista$e% #he real ris$ premium on your stoc$

investment would be negativeyou would end up with a loss%

4ut if you properly estimate long.term equity returns using other methodsfor eample, the

method &+ve proposed, which puts the future return for the stoc$ mar$et at 5? to K?then it

ma$es perfect sense to assess the appropriateness3 of the current valuation through a processof comparison with the investment alternatives% &n the current case, the alternatives of cash



22

and bonds are offering much less than 5? to K?so there+s a decent ris$ premium in place for

equities% #he mar$et is not overvalued3it doesn+t belong3 at a lower valuation% #o the

contrary, it+s priced where it should be, given the alternatives% &nvestors have done their obs

properly, leaving no easy arbitrages to eploit%

6ow, if bears want to argue that it+s unwise to loc$ in 5? to K? equity returns right now -oreven !? or ?/, because the mar$et cycle will eventually produce selloffs in which greater

returns are made available, my response would be" who said anything about loc$ing anything

in: Eet+s time the mar$etas bears seem to want to do% &+m all for that approach%

4ut timing the mar$et doesn+t mean boycotting it until it hands you, on a silver platter, the

high returns that you+re demanding% 7fter all, there+s an ecellent chance that it won+t hand

them to youthere+s no reason it has to% Beneral societal progressparticularly in the area of

economic policyma$ingreduce the odds that it will% (ather, timing the mar$et means

monitoring for the types of processes that tend to cause mar$ets to sell offcapturing equity

returns except when there are signs of those processes emerging% Galuation3at least in the

range that we+re currently atis not one of the processes that cause mar$ets to sell off -or, for

that matter, that stop mar$ets from selling off/% 0o stop worrying about it%

4ig selloffs usually occur in association with recessions% #hat+s where mar$et timers ma$e their

moneyby anticipating turns in the business cycle% 7 hint to bears" if you+re calling for a

recession right now, in this monetary environment, you+re doing it wrong%

http"AAphilosophicaleconomics%wordpress%comA21!A12A2Athe.single.greatest.predictor.of.

future.stoc$.mar$et.returnsA

Documents

The Single Greatest Predictor of Future Stock Market Returns