Manias and Mimesis
Applying René Girard’s mimetic theory to financial bubbles
Tobias Huber
huber.tobias.a@gmail.com
Byrne Hobart
bhobart@gmail.com
October 11, 2019
”I can calculate the motion of
heavenly bodies, but not the madness
of people.”
Isaac Newton
Abstract: While next-generation models in quantitative finance have illuminated
the origins of market bubbles and crashes by incorporating herding and imitation behavior, the underlying cause of imitation in financial markets remains elusive. Given that
imitation is at the core of bubbles, we need a deeper understanding of the phenomenon.
In this paper, we apply René Girard’s mimetic theory in a series of case studies of historical speculative manias, from the 1840s railway mania to the ICO boom and collapse,
and even to present-day mimesis-driven market distortions.
1 Introduction
Bubbles shouldn’t exist. They represent anomalies that are irreconcilable with deeply
enshrined beliefs in market efficiency and rationality. Yet, from tulips and dotcom stocks
to Beanie Babies1 and Bitcoin2 , they are recurrent phenomena (see Kindelberger 2005).
As bubbles and busts occur with increasing frequency and intensity, some more
heterodox approaches in empirical finance started to incorporate the ingredients, such
as herding behavior or positive feedback, which give rise to the infamous “madness of
crowds” and “popular delusions” (see Spyrou 2013). Agent-based-models simulate,
for example, how macro trends emerge from traders crowding into the same strategies
(often modeled as a competition between “fundamental” investors and so-called “noise”
traders). Herding-behavior is also embedded in a popular bubble/crash-detection model
1
2
See Bissonnette 2015
See Wheatley et al. 2019
1
Electronic copy available at: https://ssrn.com/abstract=3469465
in quantitative finance to identify unsustainable market regimes.3
But what’s the source of FOMO—this fleeting feeling that moves entire markets?
While some of these next-generation models in quantitative finance have illuminated
the origins of market crashes, the nature of the imitation still remains elusive. Now,
given that imitation is at the core of bubbles, we need a deeper understanding of the
“irrational exuberance” fuelling speculative manias. For this, we need to transcend the
physics of finance and turn to the metaphysics of markets.
2 Mimesis and Bubbles: A Girardian View on Markets
One highly-valuable and theoretically-powerful model of imitative contagion, which infects all domains of human existence, is found in the work of René Girard.4 In his
theory of imitation, which spans religion, culture, art, philosophy, and anthropology, the
source of violence and its social diffusion are explained by the theory of mimetic desire.
One of Girard’s core insights—validated, for example, by the rise of countless Instagram
influencers—is that our desire is mediated by the desire of others.
Now, the concrete conceptual structure of mimetic desire is laid out in his 1972 book
Violence and the Sacred. According to Girard, mimesis, which he analytically separates
from imitation, is deeply antagonistic in nature—it operates as the engine of social rivalry
and conflict. While imitation, on Girard’s model, refers to the positive effects of copying
someone else’s behavior, mimetic desire—which desires the other’s desire—opens up a
deeply violent dimension. Mimetic desire is contagious.
As Girard shows, it accumulates under the social surface until the mimetic tension
is violently discharged in a “sacrificial event.” As the mimetic rivalry intensifies, which
is driven by constant social imitation, it threatens to erupt in violence.
In ancient religions, this violence is collectively released through the sacrificial scapegoating of a victim, who is expelled by the community. After the sacrificial event, the
sacred victim—which later, often, gets persevered in myth—eases the mimetic tensions
and reconciles the threat of eternal violence.
Girard argues that this communal sacrifice, an act of all-against-one violence meant
to prevent all-against-all violence, is the origin of the sacred. Violent religious sacrifice,
or other ritual practices, are needed so that violence can be constrained. In other
words, the sacred, around which Girardian mimetic theory is organized, represents a
social mechanism through which violence is “self-exteriorized.” The sacred—which is
culturally encoded in social rules, systems of prohibitions and obligations, and religious
rituals—channels the socially subversive and destabilizing violent forces, so that violence
can more efficiently be contained. (See Figure 1).
3
In the so-called Log-Periodic Power Law Singularity model, imitation behavior generates competing
positive feedback loops of high-return expectations and negative feedback spirals of bearish sentiment
that express themselves in super-exponential price-growth, which, in turn, indicate the statistical existence of a bubble (see Sornette and Cauwels 2015).
4
Despite Peter Thiel‘s effort to popularize Girard’s theory by translating mimetic desire into economic
competition, he still remains one of the most underrated thinkers of the 20th century.
2
Electronic copy available at: https://ssrn.com/abstract=3469465
Figure 1: The mimetic cycle
So, if you belong to an archaic religious cult, the scapegoating is rather straightforward; you could simply sacrifice an animal or human to a God. Since time immemorial,
God, or some other externalized authority acted as the universal regulator of social behavior. What do you do in a modern, desacralized society? The modern assumption
is that we’ve moved past the old rituals, but Girardians view it as a regression: instead of obviating the need for ritual, we’ve recreated the primitive, ad hoc practice of
scapegoating from which these rituals evolved.
The frenetic build-up of bubbles and their violent collapse provides some of the purest
examples of the mimetic process—they crystallize fear, hope, hype, overconfidence, or
under-confidence. Markets perform a similar function to religions in another more literal
sense: the translation of future cash flows into a present asset price is just another
way of reconciling the demands of the eternal future to the here-and-now. In other
words, markets—these sublime machines that synthesize beliefs and aggregate them
into prices—instantiate a secularized version of the sacred.
Desire and imitation spread through markets like a contagion. Amplified by highfrequency systems and momentum-driven trading strategies, the extreme reflexivity endemic in markets accelerates the inflation of bubbles. It also makes them excitable to the
smallest of disturbances that can cascade into drawdowns, dislocations, or even abysmal
crashes. It is this tendency of channeling our mimetic desires and violently discharging
the accumulating mimetic tension that our hyper-globalized and interconnected markets
share with ancient and more archaic sacrificial systems. Instead of some ritualistic sacrifice, you get catastrophic busts. Instead of expelling a victim from the community, you
can scapegoat HFT algorithms or short-sellers.
Financial markets are designed to be impersonal, but they still indulge in ritual: just
think of IPOs and lockups. The IPO-process involves gathering a bunch of people in
a very traditional, temple-like building, and having them ring a special bell and at a
specified time. The ritual is purely symbolic—whether you fat-finger the bell early or
late, the NYSE opens at 9:30:00:0000.
Later, the lockup represents another sort of ritual: you’re richer than you’ve ever
3
Electronic copy available at: https://ssrn.com/abstract=3469465
Figure 2: Triangular desire
been, but you can’t convert your wealth into consumption until a suitable time has
passed. It’s the modern world’s answer to a medieval squire spending the day before
his knighting in a state of fasting and prayer. (And indeed, many founders and early
employees spend the six months after IPO praying that nothing major goes wrong.) In
Lyft’s case, the scapegoating process has already started after shares crashed post-IPO.5
Girard speaks frequently of “triangular desire” (see Figure 2). A love triangle looks
like a literary device—A and B both want C, which they can’t have, and this sets up
dramatic tension to give the writer a better story. But to Girard, the love triangle is
not an attempt to make a story more literary but to make it more realistic. All of our
desires are mediated by the desires of others: C is irrelevant, A’s model of B and B’s
model of A are the real locus of desire.
In markets, triangular desire can work in two directions: on the behavior of speculators and the behavior of entrepreneurs.
A speculator wants to be the person who bought Amazon at the IPO; the entrepreneur wants to be the person who took Facebook public at a twelve-digit market
capitalization. Speculators can trivially emulate their mimetic models through leverage,
i.e. through believing what the model believes, only more so. If you’re competing with
someone who put 100% of their money into Bitcoin, you’ll never catch up if you make
the same trade. But if you borrow to put 200% of your money into Bitcoin, you will—at
least so long as your mimetic model was worth copying but was insufficiently certain of
their own correctness.
An entrepreneur must go to greater lengths to copy their model: they actually need
to found a company. But it’s possible, and as bubbles advance we tend to see more companies following the exact template of exemplary predecessors. In the 80s, Microsoft’s
5
https://cnb.cx/2CXvVvp
4
Electronic copy available at: https://ssrn.com/abstract=3469465
benefits were legendary: stock options, free food, even free soda. In the 90s, Internet
companies compulsively topped one another with better fringe benefits, wilder parties,
and more generous options. Of course, these companies didn’t compete to be as revolutionary as Microsoft had been at the same stage of its corporate existence. That’s hard
to do—and hard for outside investors to notice. During bubbles, companies find it easier
to replicate the surface attributes and let pattern-matching do the rest.
One of the constraints Girard sees on runaway mimetic desire is social stratification.
Don Quixote is a story about triangular desire mediated by a literary character: Don
Quixote himself wants to emulate the knight Amadis of Gaul, “the only, the first, the
unique, the master and lord of all those who existed in the world. . . ” Sancho Panza can
copy Don Quixote’s desires, but only in a limited sense: much of what Quixote pursues is
off limits to Sancho, and Sancho’s desires—food, money—are beneath Quixote’s notice.
When Girard cites later works, like Dostoyevsky, Stendhal, and Proust, this social
structure has broken down:
The frivolous and seductive nobleman of the Louis XV era has been replaced
by the scowling and morose gentleman of the Restoration. This depressing
character lives on his property, he works hard, goes to bed early, and worst
of all, even manages to economize. What is the significance of such austere
morals? Is it really a return to the ”ancestral virtues”? This is what we
are told constantly in the bien-pensant journals but there is no need to believe it. This gloomy, sour-tempered, and totally negative kind of wisdom is
typically bourgeois. The aristocracy is trying to prove to the Others that it
has ”earned” its privileges; that is why it borrows its code of ethics from the
class which is competing for those same privileges. Mediated by its bourgeois audience, the nobility copies the bourgeoisie without even realizing it.
In Memoirs of a Tourist Stendhal remarks sardonically that the revolution has bequeathed to the French aristocracy the customs of democratic,
protestant Geneva.
The present is a bizarre parody of even this social compression. What do we make of
an unshaven young man who wears ripped jeans and a freebie t-shirt, sleeps until noon,
doesn’t shave, and spends all day alone on a computer? Maybe he’s unemployed, and
sleeps on his friend’s couch; maybe he’s a tech billionaire. It could go either way. Even
the maitre d ’s of San Francisco’s finest restaurants probably can’t tell 100% of the time.
This dissolution of social barriers intensifies competition by increasing the number of
viable competitors. When there are no general obstacles to achieving one’s dreams, the
temptation is to attribute any failure to the agency of someone else.
Now, is it not blasphemous to exploit Girard’s mimetic theory for insights into such
profane things as markets? Is it a sin to use elegant mathematics to render the blood
spatter in a video game more realistic? Was it wrong to sell the “Salvator Mundi” for
half a billion dollars? There isn’t an obvious answer. Sometimes, pure science, art, and
philosophy need to serve practical purposes—sometimes a measure of the significance of
5
Electronic copy available at: https://ssrn.com/abstract=3469465
a theory is how far afield it can be applied. A better understanding of financial markets
is one of them.
So, Girard’s mimetic theory provides a template to better understand the nature of
speculative bubbles. In the following section, we will now turn to historical cases studies
of financial bubbles to identify the mimetic dynamics driving them. While many of the
bubbles that we analyze below resulted in spectacular crashes, the mimetic dynamic
fueling them doesn’t have to lead to systemic instability, but is often a necessary component in the process of technological innovation. Bubbles incubated, as we will show
below, many revolutionary technologies and radical ideas. From the Manhattan Project
and Apollo Program to Moore’s Law and 90’s techno-Utopianism, the mimetic desires
underlying these bubbles resulted often in self-reinforcing feedback loops that lead to excessive enthusiasm and commitment by true believers and early adopters, beyond what
would be rationalized by standard-cost benefit analyses.6
Now, as the case studies below show, bubbles are fueled by a mimetic desire that
copies the superficial features of the object of speculation, but fails to derive the underlying principle. A better understanding of the mimetic process diving bubbles can,
thus, be highly profitable as it can give you an edge over superficial imitation, which
fails to pattern-recognize the feature that makes the speculative object of mimetic desire
singular.
3 Mimetic Bubbles: Some Case Studies
Violence and the Sacred abounds with case studies, from the Iliad and Oedipus the
King to the rituals of the Dinka and Lovedu tribes.
Fortunately, the history of finance is roughly coextensive with all of history. The first
things worth writing down were the terms of deals—in what is perhaps a metaphor for
the destructive nature of non-market conflict, some of the oldest writing we have consists
of cuneiform tablets recording the terms of loans and forward commodities contracts.
These were written on clay, which could be reused, but when a city was razed the fires
would bake the clay, writing the contract from random-access memory to permanent
storage.7
Financial markets are particularly well-documented because the tradable nature of
obligations means that the exact terms need to be set forth explicitly. While absolute
information availability has of course increased since the days of cuneiform, legal and
cultural norms still put finance ahead of other fields: no government agency, nonprofit,
or religious organization has the equivalent of the 10-K/investor deck/quarterly earnings
call combo that makes investing such an enjoyable hobby for people who like to spend
literally all of their time reading.
This informational bounty gives us numerous case studies, from the British Railway
Bubble to the present—and, as we’ll argue, to an area of the market that will in the
6
See Huber 2017
The sack of Ebla in 2,300 BC is one of the rare times that auto-save successfully preserved important
business documents following a sudden loss of power. Not that it did them any good.
7
6
Electronic copy available at: https://ssrn.com/abstract=3469465
future be regarded as an enormous bubble.
3.1 The British Railway Mania: All Aboard the Hype Train
In terms of GDP invested, the British Railway Mania was one of the largest financial
bubbles in history. At its peak in 1847, railway investment was a staggering 7% of GDP;
the modern US equivalent would be an investment bubble that produced $1.4 trillion in
capital expenditures in a single year.
Investor expectations during the bubble were surprisingly quaint; the ambitious expected to earn somewhere in the neighborhood of 10% per year on their investment, with
no growth. While this may strike contemporary observers as a modest return, it was
well ahead of the 3% that investors could expect to receive from long-term government
bonds, or the 4% they could expect to receive from loans to companies. Unlike more
recent equity bubbles, the railway bubble was not predicated on endless growth—the
expectation was that railway companies would pay out all of their income each year.
The bubble of the late 1840s echoed a similar, albeit much more modest, investing
boom in the 1830s. That bubble, like many since, was launched by a sort of capstone
development as part of a long-term technological trend. Steam power had been used for
over a century, primarily as a more efficient way to pump water out of flooded mines.
The low marginal cost of energy inputs for these devices provided an indirect subsidy
for rapid improvements in engine technology: since coal was cheap, and getting cheaper,
it made sense to build newer and more powerful steam engines. By 1830, this allowed
the Liverpool and Manchester Railway to offer intercity steam passenger service.
The mechanism for creating new railways lent itself to rivalrous competition: a new
company would raise a small sum of money from early investors, trading on the prestige
of its backers and the expectation of high returns. The railway owners would petition
parliament for permission to build their railroad. Having secured permission, they would
call down more capital from their initial investors, and begin laying track.
Since any given city pair could generally support a single connection, there was fierce
competition to be the first to build a new railway. Since the initial investment merely
paid for surveys and lobbying, it was capital-efficient. And since railways used prominent
backers to persuade both investors and legislators to back their plans, founders would
often compete to get local grandees on board.
While the actual data tell a compelling story—£107m invested over a three-year
period in which annual GDP averaged £588m, a 50% run-up in share prices in three
years, followed by a greater than 60% decline over the next half-decade—the single best
summary is a short story, How We Got Up the Glenmutchkin Railway and How We Got
Out of It, by railway attorney William Atoun, which Andrew Odlyzko uses to illustrate
the bubble in his paper, Collective hallucinations and inefficient markets: The British
Railway Mania of the 1840s.
Glenmutchkin tells the story of a pair of affable punters who decide to cash in on the
railway mania by promoting a new railway venture between two obscure Scottish towns.
They acquire aristocratic backers (a local drunk who styles himself a laird), raise funds
from socially-responsible investors (a puritanical merchant who invests on the condition
7
Electronic copy available at: https://ssrn.com/abstract=3469465
that the railway not operate on Sundays), and rush their plans to parliament before a
competitor can get the rights. Their goal is not merely wealth, but status: the narrator
says “Work I abhorred with a detestation worthy of a scion of nobility.” Glenmutchin
is thus a story of dual mimesis: a promoter copying the means of New Money to live
within the means of Old Money.
Notably, to a contemporary reader it would be unclear which venture in particular
Aytoun was mocking, because all the failures followed the same pattern: they used the
healthy returns of projects like the Liverpool and Manchester Railway to justify ambitious traffic plans for railways connecting smaller locales, and tended to raise less money
than was actually needed to complete construction—high demand for labor and materials
meant that the shoddy copies were more expensive to produce than the valuable originals. “These were the first glorious days of general speculation,“ says Glenmutchkin’s
narrator, “Railroads were emerging from the hands of the greater into the fingers of the
lesser capitalists.”
Perhaps the most Girardian narrative of the Railway Bubble was the story of George
Hudson, “The Railway King.” Hudson formed the York and North Midland Railway in
1835, and cobbled together a larger system through aggressive mergers. Hudson used
dubious financial maneuvers—paying dividends out of capital before earning enough
to justify them, borrowing for working capital—and during a depression in 1849, his
companies were unable to meet their loan payments. Hudson fled to the Continent, and
spent his last years dodging creditors and the law.
To outside investors, Hudson embodied the optimism of the railway boom: he built
and acquired railways with gusto, helped consolidate the industry, and improved cooperation across different railway companies. Once the bubble popped, Hudson remained a
symbol, now of the malfeasance inherent in late-bubble financial machinations. Edward
Chancellor describes Hudson as a “scapegoat for the failed speculations,” a single person who could be blamed for the collective madness of the British investing class (see:
Chancellor 1999: 150-151).
Hudson was hardly the only perpetrator, though. Throughout the industry, promoters underestimated costs and overestimated demand, leading to returns well below the
10% investors expected. Railway shares, which were expected to be a safe and steady
investment, lost half their value in the years after the bubble.
This illustrates a common dynamic in the bubbles we study: the presence of leverage
(both financial and operating) allows operators to choose how much risk to take. The
truest believers in the bubble narrative take the biggest risks, and to the extent that
their thesis is correct, they reap corresponding rewards. A charismatic entrepreneur can
raise more funds, but in doing so becomes synonymous with the story they tell about
their investment. When the bubble bursts, the story is revealed as a lie, or at least an
exaggeration—which turns the greatest promoter into the greatest liar.
Another dynamic we see in bubbles is that bubbles make the world more legible. The
political scientist James C. Scott uses the term “legibility” to describe how states impose
uniform standards on people and locations in order to more effectively conscript and tax
them. A bubble does this as well: before the railway bubble, travel was uncertain, and
8
Electronic copy available at: https://ssrn.com/abstract=3469465
Figure 3: British railway share performance, 1830-1850. Source: Andrew Odlyzko,
Collective hallucinations and inefficient markets: The British Railway Mania of the
1840s
humans moved at the same speed as letters. In a horse-drawn world, you might expect
a guest to arrive at home, and know roughly what month they’d arrive. With the rise
of railway timetables, it was possible to know, down to the minute, when they’d reach
the station. Other bubbles re-map the world in other ways: the growth of global trade
led Parliament to offer a bounty for accurate measurements of longitude; the demands
of commerce literally added another dimension to the world’s coordinate plane.
In a credit bubble, legibility means a better understanding of the world as it is—
more certainty about which debtors are money-good and which are not, which forms
of collateral have what sort of value. In an equity bubble, legibility means a better
understanding of the world as it can be; some kinds of knowledge (who will be where, and
exactly when) are available only if the technology of the railway, and the accompanying
information technology and social technology, are invented, turned into products, and
scaled up.
The legibility-inducing nature of bubbles creates a rivalrous dynamic because it requires visionaries to have a specific view of the future, incompatible with the status
quo. A railway is not merely an investment with an expected return; it’s also a slap
in the face for older social norms. For someone who benefits from the status quo, the
new technology forces a difficult decision: accept higher returns, but risk one’s social
standing, or fight.
As the bubble grows in scope, the contradiction becomes more visible, leading to
a third common characteristic of bubbles: the bifurcation of viewpoints into totallypositive or totally-negative. A single railway that replaces mules with steam engines
on a single route is not a social crisis; a bubble that sends a meaningful fraction of
9
Electronic copy available at: https://ssrn.com/abstract=3469465
the economy into one sector, that creates a new rich class of people and turns any
town not affected by the bubble into a backwater—that forces people to take sides. As
the bubble grows, the easy opportunities vanish, so new investment requires ever more
extreme estimates of future returns—in other words, as the bubble grows, it becomes an
increasingly pure bet on a revolution.
This amplifies volatility at the peak, and makes a collapse more catastrophic. If
there’s a bubble company whose valuation is 2x that of a comparable legacy company,
a prudent legacy investor might be interested in buying into the bubble if prices drop a
little. But if the bubble asset’s valuation is 10x that of a comparable legacy company,
there is no natural buyer outside the bubble constituency. Pyramiding leverage can
transmute bank loans from a generic bet on credit into a bet on the bubble, but equity
valuations are driven by small differences of opinion among true believers.
This combination of leverage and bifurcation ultimately produces a crashing denouement; a decline in equity prices or an increase in defaults suddenly reveals that there is
no marginal buyer waiting in the wings, and prices plummet. The revolution can still
happen—England still has railroads, and other bubbles have also reworked the geography and culture of other countries despite ultimately leading to poor returns. But the
utopia envisioned by the most aggressive participants is never realized.
3.2 The 1920s: An Electrifying Boom
The boom of the 1920s was unprecedented, as perhaps the first large bubble celebrating
technology as such rather than a specific technological advance. The bubble combined
novel technology with new financing forms: prior to the bubble, public equity markets
primarily focused on railroads, with a few listed financial firms. Ever since, industrial
companies, retailers, and other non-finance and non-transportation equities have predominated.8
In many ways, this bubble relied on the infrastructure—and legibility—introduced
by prior bubbles. A consumer products company like RCA could only ship radios to
a national market because railroads had made such a national market accessible. And
stores could only offer these new goods on credit because the banking system had grown
large and robust enough to support consumer credit.
Telegraph networks had been bootstrapped by the demands of the railroad and financial system—most telegraph lines ran along railroads and were used to coordinate traffic,
while the first international telegraph line was originally used to transmit exchange rates.
(To this day, investors refer to the GBP/USD exchange rate as “the cable,” named after
the first submarine cable between London and New York. At scales large and small, information asymmetry exerts a gravitational pull dragging wealthy places together through
8
There was a push-and-pull effect, here. On the push side, more non-railroad businesses scaled up
starting in the 20s. On the pull side, the rise of passenger cars sapped passenger traffic. Due to a
confluence of events—the Great Depression increasing government power, World War Two subsidizing
the growth of motor vehicle factories, and postwar economic prosperity—car owners became a major
political constituency, and as a result the interstate highway system finally crippled passenger rail in the
US.
10
Electronic copy available at: https://ssrn.com/abstract=3469465
cultural exchange and communications. Even now, fashion trends—whether based on
fabric or finance—move from Menlo Park to Manhattan faster than they move from
Menlo Park to Sacramento.)
The 20s bubble was built on many technological innovations, but the two key ones
were internal combustion and electricity. Both led indirectly to the financial innovation of retained earnings. For a railroad, high retained earnings make little financial
sense. Capital expenditures either occur continuously (replacing engines as they wear
out, adding small branch lines) or in a large discrete burst (adding a new trunk route
connecting two major cities). This led to a model in which a company would distribute
essentially all of its retained earnings as a dividend each year, with an expected longterm growth rate of somewhere between zero (the naive model) and the growth rate in
population in the areas served.
For an industrial concern in a growth industry, distributing all earnings makes little
financial sense, because growth is more continuous. Prior to the 1920s, electric utilities
started retaining a large share of their earnings, and reinvesting the proceeds in additional capacity. A company that provided enough power to illuminate the rooms of the
rich one year might try to serve the middle class over the next decade, and everyone
some time after that.
Meanwhile, the availability of electricity provided an implicit subsidy for manufacturers, and, later, domestic appliances.
Electricity exerted a profound impact on manufacturing; it literally reshaped factories. Prior to electrification, factories were generally powered by a single line shaft, which
distributed mechanical power to every part of the factory. Electrification allowed a more
distributed approach—instead of a factory radiating out from a single power source,
power could be distributed directly to machines. Prior to electrification, factories were
generally multistory (decreasing the average physical distance to a single power source),
while post-electrification factories are generally single-story buildings spread out over a
wide area.
This physical change in the constraints of manufacturing led to a change in the
relationship between investors and companies—instead of investing in a single project
of approximately fixed scale, shareholders owned a company that could continuously
reinvest and incrementally expand. This was a sea change in how investors thought
about company returns. The earlier paradigm was that shareholders were simply the
most junior claimants on the returns of a set of fixed assets, like a railroad or factory.
A company would return most of its profits as dividends, and expand by issuing more
stock and bonds. Electrification enabled continuous, rather than discrete, reinvestment,
meaning that the default plan for an industrial company was indefinite growth.
As electricity usage scaled up, the relative advantages of electrification multiplied:
at a large scale, an electrified factory could outproduce a comparable steam-powered
one. At a smaller scale, a traditional factory that had excess capacity could face competition from a nimbler electrified competitor. More electrified factories subsidized more,
and more efficient, power generation, which pushed down the cost per kilowatt-hour.
This also provided an implicit subsidy for more efficient electrical equipment, further
11
Electronic copy available at: https://ssrn.com/abstract=3469465
accelerating the transition.
(One can see a similar dynamic today in computing-heavy businesses, as cloud computation turns servers from capex to opex while relentlessly lowering their cost.)
This leads to a meaningful change in the cause and magnitude of valuation changes.
In a zero-growth model, the net present value of a security is just the present cash
flows, discounted back at some rate that incorporates their variability. So an investor
who tolerates an 8% return for a given stream of cash flows would value $1m in annual
profits at $1m/0.08, or $12.5m. When a company can grow at a steady pace, the value
of the cash flows can be approximately modeled as (cash flow minus capital expenditures
for growth)/(discount rate - growth rate). So if the same company can reinvest half of its
cash flow and expect to grow at 5% annually, its net present value is $500k/(8%-5%), or
$16.7m. Crucially, as the growth rate approaches the discount rate, the price/earnings
multiple (i.e. the inverse of the discount rate minus the growth rate—33.3x in the case of
the growth company, 12.5x in the case of the static company), asymptotically approaches
infinity. If investors believe the true long-term growth rate is 6%, the price/earnings
multiple is 1(8% - 6%), or 50x. At 7% growth, 100x. At 8% growth or higher, the
theoretical valuation is infinite.
The body of theory behind these valuations was not actually developed until the
1930s. John Burr Williams published his Theory of Investment Value, which discusses
the net present value model, in 1938.
(Notably, Benjamin Graham, the “father of value investing,” also published his masterpiece, Security Analysis, in 1934, the midst of the Great Depression. A collapse in
asset prices causes a lot of harm, but it does give the busy investor enough free time to
sit down and invent a good theory about where all their money went.)
While the theory didn’t quite exist yet, that theory merely formalized facts about
investor behavior. Implicitly, investors were ordering valuations by growth rates—if
company X traded at 10 times earnings, and Y grew faster than X, Y was worth some
premium.
The internal combustion engine as a technology didn’t exert the same kind of influence on companies’ approaches to shareholder returns, but it did enhance the trend:
auto companies were voracious consumers of capital, and as Henry Ford realized, they
paid well enough to expand their own market. GM could and did pay a dividend, but
reinvested profits in continued growth—and more than doubled its annual production
from 1925 to 1928 as a result.
Mass media became a more important factor in the 1920s, first as a consequence of
underlying technology changes and later as a force directly intertwined with the bubble.
Increased urbanization grew the population that could subscribe to a daily newspaper,
while the abundance of new consumer products gave newspapers something worth selling,
and growth in consumer credit let them sell to buyers who could not, strictly speaking,
afford it.
The rise of mass media created a mimetic dynamic to market-price movements. It
was widely understood by contemporary traders that the markets were manipulated,
but to the average trader this implied an opportunity, not a risk: famous bulls and bears
12
Electronic copy available at: https://ssrn.com/abstract=3469465
would assemble pools to drive stocks up or down, publicizing them through rumors and
newspapers. While a modern reader might associate RCA with a name like Sarnoff, an
investor in the 1920s would be more likely to associate RCA with Mike Meehan, who
organized multiple pools to manipulate the stock higher.
Over the course of the 1920s, RCA advanced from a low of 1 12 in 1921 to a high
of 114 34 in 1921, making investors who bought and held (and sold at the right time)
rich. But the stock was quite volatile throughout this period, so an investor who timed
the market well—who bought on margin before the Meehan Pool pushed the stock up
another ten points, and sold before it drifted down—would have outperformed even this
impressive growth.
Early investor pools touted promising new developments at a company: new products, faster growth, or at least the promise of a higher dividend. Over time, though, the
pool itself became a catalyst: as stocks became less connected to individual catalysts and
more linked to the pool operators themselves. Rumors of the involvement of Meehan,
or of Jesse Livermore, the most famous of the era’s speculators, could single-handedly
move the stock.
Leverage introduces a strong temptation towards mimetic behavior, because a levered
investor compounds their net worth faster than an unlevered one, and given enough
leverage and enough outperformance, a small investor can catch up to a wealthier one.
In the 1920s, borrowing 50-60% of the value of a stock was considered conservative, and
investors could borrow as much as 90% of their purchase value. Over a long enough time
period, an investor who consistently borrowed 90% of their money and didn’t get wiped
out could catch up to an unlevered investor in the same asset, no matter how large the
initial gap.
As in other bubbles, escalating asset price appreciation led to escalating mimesis. Traders started out following individual companies, noting their technology-driven
growth. Over time, they followed fads rather than firms—after Lindberg’s flight across
the Atlantic, aviation-related stocks popped, including Seaboard Air Lines, a railroad.
And at the peak, they chased other investors. When a famous plunger like Mike Meehan
was rumored to be buying RCA, the important detail was Meehan, not RCA.
Individual investors were not the only market participants to recognize the power
of leverage to magnify returns. Among the largest growth companies in the 1920s were
electric utilities.
The most iconic of the electric utility entrepreneurs was Samuel Insull. A British
immigrant, Insull moved to the US in his early 20s to work as Thomas Edison’s assistant.
When Edison refused to make him president of Edison Electric (the predecessor to
today’s General Electric), Insull took control of Chicago Edison company in 1887, where
he pioneered demand-based metered pricing and rapidly expanded his company.
Over time, the Insull complex, renamed Commonwealth Edison, expanded through
both internal reinvestment and aggressive acquisitions. At its peak, the company had
assets of over $500m backed by $27m of equity through a pyramid of holding companies.
Insull had the great misfortune of being featured on the cover of Time the month
after the crash of 1929. When his subsidiaries had trouble rolling over debt, the liquidity
13
Electronic copy available at: https://ssrn.com/abstract=3469465
Figure 4: The Dow, from bubble to crash. Source: Macrotrends.net
crisis quickly worked its way up the holding company structure, causing Commonwealth
Edison to collapse. Insull was tried for fraud, and, though acquitted, died in poverty,
with 84 cents in his pocket and $14 million in debt.
The rise and collapse of his empire transformed Insull from man to symbol: he
began his career as a sideman to the celebrity entrepreneur Edison, became a symbol
of corporate success thanks to laudatory media coverage and his wide investor base,
and eventually came to symbolize the excessive power of the plutocracy—when Franklin
Roosevelt mentioned him in a speech, it was not alongside other executives. Roosevelt
railed against “the Insull monstrosity,” condemning “the Ishmaels and the Insulls,”
turning Insull himself from a mortal to a mythic figure.9
3.3 Conglomerates: The Best and the Brightest
The 1960s saw the rise of a new corporate form, organized by a new kind of elite.
American regulation saw a step-function change following the disastrous collapse in the
1930s (even today, the ‘34 Act and ‘40 Act are frequently cited; American securities
regulation largely consists of riffs on New Dealers’ ideas). Large trusts collapsed under
9
FDR’s speech captures something important about the pace of technological change. To justify
regulation of public utilities, he cites the four hundred-year-old English legal tradition of regulating ferry
prices due to their essential and monopolistic nature. Insull’s reward for aggressive risk-taking was that,
in a single generation, he helped make his industry so valuable it became too important for someone like
him to be entrusted with it.
14
Electronic copy available at: https://ssrn.com/abstract=3469465
their own financial weight, while newly-invigorated antitrust laws made such companies
untenable.
At the same time, the business of business turned into an academic discipline. Equity
investors, enticed by the market’s healthy returns in the 1950s and increasingly forgetful
of the 1930s, flocked to the markets, investing both directly and through the new vehicle
of mutual funds. Mutual funds were both a cause and a symptom of a general decline
in economic inequality.
While captains of industry through the 1920s had been wealthy and powerful, and
speculators had aspired towards similar wealth and power, the decrease in inequality
following the Depression, the New Deal, and the financially repressive war years made the
middle class a larger factor in financial markets. And while many individuals purchased
stocks for their own account, institutions became an increasingly important factor: total
institutional ownership of US equities was under 5% in 1940, under 8% in 1950, but over
20% by the early 1970s.
Stocks were increasingly owned by institutions, and the practice of business was,
increasingly, institutionalized. In 1949, there were a total of around 3,000 MBAs. The
653 who graduated from Harvard that year were among history’s luckiest people; within
25 years, nearly half held the title Chairman, CEO, and COO—and, in a statistic that
tells us a bit about inflation and a lot about income inequality, 20% of them were
millionaires. The Depression and the Second World War had created a demographic
pinch in business, and especially in finance. A weak economy presented few opportunities
for promotion, and few opportunities to climb the ladder by switching jobs; the war
also took young people out of the corporate rat-race for half a decade. This led to a
boardroom population dominated by people whose careers had peaked in the 1920s and
whose most formative experience was the 1930s.
When younger investors finally started running major companies and serious money
in the 1950s, it led to a massive change in corporate and investment strategy.
The trend towards institutionalizing both the practice of business and the practice
of investing, coupled with stricter antitrust enforcement, led to the creation of conglomerates. Aggressive antitrust enforcement meant that companies had difficulty acquiring
direct competitors, suppliers, or customers, which limited the natural opportunities for
growth through acquisition. However, there was ample institutional capital available to
any company that could realistically promise steady growth.
In response, company managers started acquiring businesses far removed from their
core operations. At first, they could simply bottom-feed; there were many “leftover”
companies that had survived the Depression and the war years, and were trading at
cheap valuations. A conglomerate could take advantage of the valuation arbitrage, by
offering its high-multiple shares for the low-multiple shares of acquirees.
However, this led to a problem: over time, most of the small, forgotten public companies got snapped up by larger companies, or had their valuations buoyed by tenacious
value investors. So the conglomerateurs turned to larger companies, which put them in
conflict with the elites.
Leasco is an instructive case. Leasco was founded by Saul Steinberg in 1961, when
15
Electronic copy available at: https://ssrn.com/abstract=3469465
he was 22 years old. Initially, the company took advantage of two simple arbitrages.
First: IBM preferred to lease, rather than sell, computers, and depreciated their
computers aggressively. By buying IBM computers, leasing them at slightly lower rates,
but depreciating them much more slowly, Leasco could report higher profits, at least
in an accounting sense. While Leasco’s profits near the end of a lease would be lower
(either the hardware would be obsolete, or Leasco would be depreciating an asset that
IBM would have already depreciated to zero), short-term profits would give Leasco access
to capital, which it could use to grow its assets. As long as growth was high enough,
most of Leasco’s assets would be young enough to report high profits.
Second: Leasco benefited from a change in tax laws enacted in 1962, which gave
companies a 7% tax credit on capital expenses.
Both of these subsidized Leasco’s growth, but they also forced it to continue growing
to maintain its valuation.
By 1968, Leasco needed an alternative source of growth. They found one in conglomeratizing: Leasco made an unsolicited offer to buy Reliance Insurance, a 150-year-old
company ten times Leasco’s size. Computer leasing and insurance offer minimal synergies (insurance might theoretically be a cheap source of capital for a leasing company,
but state insurance regulators take a dim view of insurers using their float to juice the
returns of related parties). However, Leasco could make a financial argument: their high
valuation gave them immense buying power, while Reliance’s slow growth depressed its
own valuation.
The offer succeeded—although Reliance was able to negotiate a better price by,
among other things, claiming that it was negotiating to sell itself to a different leasing
company.
A year later, Steinberg attempted to acquire the much larger Chemical Bank, in a
deal that was scotched when Chemical called in favors with journalists and legislators to
kill the offer. To Steinberg, the Wharton-educated outsider with an enormous bankroll
provided by giddy investors, this was a simple commercial transaction. He had valuable
stock, and was offering to exchange it for Chemical’s cheaper shares. To Chemical, the
whole deal was anathema; a speculator and accounting con-artist who optimized his
operations down to the decimal point but was not, in the end, producing any real value.
Leasco had its ammunition: its stock had appreciated faster than any other member
of the Fortune 500 over the previous five years. Chemical Bank had its own, though:
its board included the chairman of AT&T, the finance chairman of U.S. Steel, the presidents of du Pont and IBM, a member of Texaco’s executive committee, and the former
president of the NYSE. It was a veritable Who’s Who of the American business establishment (Leasco compiled their dossier on Chemical through the literal Who’s Who,
which did not have a single Leasco board member listed.)10
Chemical also had some financial resources of its own, and hinted that Leasco’s
ability to make a compelling offer was contingent on a high stock price. In the three
weeks after the press started reporting on a possible Leasco offer for Chemical, Leasco’s
stock dropped from $140 to under $100.
10
All this is from John Brooks’ entertaining history of the 60’s stock market, The Go-Go Years.
16
Electronic copy available at: https://ssrn.com/abstract=3469465
Chemical also considered another option: mimicking Reliance. They looked into
acquiring another insurance company, similar to Reliance, to create an antitrust obstacle
to the acquisition. A large, staid company faced a threat from an outsider, and in
addition to attacking directly they attempted to mold their business into the form of
the outsider’s. As Girard says in Violence and the Sacred, “Both parties in this tragic
dialogue have recourse to the same tactics, use the same weapons, and strive for the
same goal: destruction of the adversary.”
Leasco had grown by following the business blueprint of an earlier age: a single
charismatic founder, high leverage, high-wire acquisitions, aggressive use of government
subsidies and loose accounting standards. Chemical was the result of exactly the same
sorts of behavior—generations earlier. When Steinberg used his Robber Baron-style tactics against an established company, whose directors largely hailed from Robber Barons’
creations, they fought back, and won. Leasco had done an end-run around the establishment, offering Big Blue’s hardware on terms Big Blue wasn’t interested in matching,
but it eventually reached a size where the optimal move was to join rather than compete with the establishment. But establishments stay established through resistance to
joiners, and Chemical played its role to perfection.
In retrospect, we can see that Chemical’s defenses were both costly and pointless.
Their aggressive manipulation of politicians, the media, and perhaps the stock market
fended off one attacker, but Chemical was ultimately sold to another financial services
conglomerate decades later. (That company, Chase, is now run by a CEO who had
cut his teeth at an acquisitive financial-services conglomerate that pyramided up small
acquisitions until it could acquire an insurance company, and then merged that insurance
company—successfully—with a larger New York bank.) “The targets are individuals,”
says Girard, but “it is the institutions that receive the blows.”
After the deal fell apart, Steinberg—who had been profiled by Forbes the year before
as the richest American under thirty in history—lamented that “I always knew there
was an establishment—I just used to think I was a part of it.” Perhaps he was, and
his non-membership was an illusion held by the Chemical board, or a trick they pulled
on their defenders. All parties involved were rich, power-hungry, and willing to bend
the rules to get what they wanted. As Girard quotes Stendhal saying, “The pettier the
social difference, the more affectation it produces.”
The Steinberg cycle—fighting the establishment, attempting to converge on it, and finally getting rejected by it—happened in parallel with James Ling’s Ling-Temco-Vought.
Ling founded an electrical contracting company, which grew big enough to list its stock
(but small enough that he distributed prospectuses at the Texas State Fair), and then
cobbled together a collection of electronics-related companies and aerospace firms, before going further afield (meatpacking) and aiming for larger targets (Jones & Laughlin
Steel).
To investors, Ling looked like a financial wizard, able to juggle complex offerings,
buybacks, spinoffs, and tracking stocks. The truth was more prosaic: Ling’s first few
companies were high-growth electronics companies, which commanded high P/E ratios.
He used his highly-valued stock to buy companies in slower-growing industries. If a
17
Electronic copy available at: https://ssrn.com/abstract=3469465
company earning $10m that trades at 20 times earnings uses stock to buy another
company earning $10m that trades at 10 times earnings, the combined company has
50% more shares outstanding but twice the earnings—so earnings per share rise by 33%.
This encouraged conglomerates to acquire ever-dowdier companies at ever-larger scale.
Ling’s downfall was more straightforwardly due to excess risk-taking and high leverage. Unlike Steinberg, he didn’t have a confrontation with the moneyed upper-class—
perhaps because, to Texans like Ling, the big rich were newly-rich enough that merely
having a high net worth was sufficient, and perhaps because while Steinberg went after
a local company, Ling’s biggest acquisition was Pittsburgh-based Jones & Laughlin.
Ling was ultimately brought down by intense regulatory scrutiny. Although his company was wildly, perhaps pointlessly diversified, the justice department accused him of
unspecified antitrust violations. The ensuing negative publicity prompted a boardroom
coup. Years later, D Magazine would describe the directors’ motivation in ousting Ling
thusly: “Ling would become a highly visible scapegoat to blame for the company’s problems.” When he was a highly-levered acquirer in a bull market, Ling was a genius who
could do no wrong, who made the cover of Time in 1969. A few years later, when
a weaker stock price crippled his acquisition campaign and revealed the shoddiness of
LTV’s underlying businesses, it was Ling again who was synonymous with the fate of
the entire company.
In finance, well-funded outsiders eventually have enough access to capital and desire
for growth that their only option is to merge, literally, with the establishment. Girard
argues that Dionysian festivals that culminate in a brutal and personal sacrifice often
begin with “the deliberate violation of established laws. . . the overall elimination of
differences. Family and social hierarchies are temporarily suppressed or inverted; children no longer respect their parents, servants their masters, vassals their lords. . . for
the duration of the festival unnatural acts and outrageous behavior are permitted, even
encouraged.”
Baby Boomers look back nostalgically at the 60s as a festival of Dionysian excess,
which ended with a spate of horrifically violent acts—but the same Dionysian process of
upending hierarchies, flattening distinctions, and ultimately scapegoating an individual
also occurred at the corner of Wall and Broad. The Dyonisian process reached its
endpoint in the 1980s, when the Maenads of private equity finally dismembered the last
of the conglomerates. Private equity acquirers set M&A records buying conglomerates
like Beatrice and RJR Nabisco, with the express purpose of selling off the pieces and
turning them into more focused companies.
3.4 Mimesis Dot Com
The dot-com bubble is worth considering in light of the 1960s electronics boom. In
some ways, the 90s were the product of a multi-decade Kulturkampf as the long-hairand-sandals crowd wrested control of the computing industry from the buzz-cuts-and-
18
Electronic copy available at: https://ssrn.com/abstract=3469465
Figure 5: Conglomerates vs the S&P 500. Source: Conglomerate Boom 2.0: A Stable
Platform? JHL Capital Group
19
Electronic copy available at: https://ssrn.com/abstract=3469465
button-down engineering establishment.11
The dot-com bubble was defined by mass mimesis, even down to the name. When
Jeff Bezos called his company “Amazon.com,” it was a triumphantly nerdy decision,
equivalent to Id Software naming themselves Doom.exe. Amazon chose the name to
differentiate itself from companies that were less focused on the Internet. (The “Amazon”
in the name was a deliberate choice, albeit a more prosaic one; the web was small enough
at the time that many directories were alphabetical, and “Am-” was enough to put
Amazon first.)
In the late 90s, the mimetic bubble strategy was to consider the success of Amazon
and start an online bookstore whose name ended in dot-com. barnesandnoble.com raised
$450m in May of 1999. When Books-a-Million announced an updated website, it sent
their stock from $3 to $47 in days (and then back to $10 in a few weeks). Neither was
much of a business. But the correct move was to do something that was as daring in
1998 as selling books online was in 1994. Something like competing with Yahoo and
Altavista in search.
Looking back at Amazon, the Platonic form of the business was not “let’s sell books
on the Internet.” It was something more like this: Internet use is growing rapidly, and
the Internet provides a new layer between consumers and producers—it literally reshapes
the supply chain, by allowing a store to have an infinite stock of products, to display real
reviews, and to remember every customer and recommend new things they will enjoy.
What industry is most amenable to being reshaped by this? It needs to be an industry
where there are a variety of products, all of which can be identified by a single name that
someone could type into a search bar. In addition, to have meaningful economics, it needs
to be an industry where supply is fragmented. Finally, it must be an industry where the
value-to-weight ratio makes it economical to ship products directly to people’s homes.
What product fits all these desiderata? Books. So books are the first thing Amazon sold,
but they’re not the thing Amazon existed to sell. From the beginning, Amazon’s plan
was “Books, and . . . ’ Amazon’s thesis can be modernized in another way. They found
a structured database: the listings of books available from wholesalers. They made a
web front-end, and harvested the value of easier access, multiplied by the growth of the
Internet population.
Google did the same thing: they found a less structured database, consisting of the
link graph.12
11
The counterculture fought this war on many fronts. It’s worth noting that LSD was invented by
a research chemist working for a large, established Swiss chemical conglomerate, and popularized by,
among others, the author of an elegiac novel about the destruction of the German elite in World War
I; “magic mushrooms” first came to America’s attention when a J. P. Morgan Vice President wrote
about them for Henry Luce’s Time. Psychedelic culture, like the computer, was invented by squares and
appropriated by hippies.
12
You can think of the link graph as a structured database if you think of it as an enormous table
made of the Cartesian product of all pages on the web, where every entry is True or False depending on
whether a page links to another page or not. It’s very sparse, since there were 75m pages in Google at
the time that it started, so there would be a 75m row by 75m table, the vast majority of whose entries
were “False”. The PageRank algorithm simplifies this from something that would be impossible to store
to something that could only be stored and processed by someone with access to the computing resources
20
Electronic copy available at: https://ssrn.com/abstract=3469465
Figure 6: The Nasdaq Composite, 1995-2003. Source: BigCharts.com
Extracting this fundamental lesson might be called either a Platonic Mimesis or
Principal Component Analysis Mimesis, depending on whether one’s background is in
liberal arts or hard science. We’ll use the less clunky one. Platonic Mimesis means
identifying the core thesis of Amazon, an idea so core it might not have been apparent
to Bezos that this was what Amazon has accomplished, or to Larry and Sergei that it
was what they had copied. By copying at a fundamental level, they differentiated at
every visible level.
Google is in a deep way a spiritual successor to Amazon. When Amazon was founded,
the salient feature of the Internet was its incredible growth rate: Internet use was growing
at a 2,300% annual pace in 1994. In 1999, this growth had created a problem: the growth
rate in garbage far exceeded the growth rate in quality. Once again, the question was:
how will the Internet reshape the supply chain? But in this case, it wasn’t the supply
chain of physical goods, but the supply chain of knowledge. How do we learn what we
need to know?
There’s a Platonic symmetry between the backgrounds of Amazon’s and Google’s
founders. Jeff Bezos worked at a hedge fund. Hedge funds have the rare distinction
of being an industry defined entirely by compensation structure: whatever you do, if it
involves public markets and you’re paid approximately 2% of assets and 20% of profits to
do it, you work at a hedge fund. Google’s founders were, of course, grad students. Their
job is meant to be as fully disconnected from financial considerations as possible—if
what you’re paid has zero connection to either the resources you have stewardship over
of—to pick a random example—Stanford University.
21
Electronic copy available at: https://ssrn.com/abstract=3469465
or how well you manage them, you’re probably an academic. The ideal life to a grad
student is a life of study and research, culminating in tenure; after working their way up
a hierarchy of work whose economic output is never measured, they reach their career
peak—tenure, an indefinite exemption from thinking about the job market.13
So it makes sense that Bezos would start out by thinking of businesses that already
existed, and ask how the Internet would change them. Half a decade later, Larry and
Sergei built a company whose product was seen as a loss leader by everyone in the industry, and railed against ads as a corrupting influence. Google did eventually introduce
ads, of course, and sold $116bn worth of them in 2018, making them the world’s single
largest ad company.14 Amazon was born as a business plan, but turned into a collection
of science projects; Google was born as a PhD thesis and, especially under the influence
of CFO Ruth Porat, turned into an exceptionally well-run business worthy of countless
B-school case studies. Google’s science projects are a line item on the P&L and fodder
for puff pieces. At Amazon, a single weird science project, Jeff Bezos’ mandate that
every part of Amazon should interface with every other part through APIs, led to AWS,
which is the majority of the company’s market value.
Superficially, then, the stories of Amazon and Google are in tension with one another.
The hyper-commercial business, founded by a financier, making products available to
the mass consumers versus the anti-commercial academic project making knowledge
available to everyone. At a higher level of abstraction, both companies were the result
of smart, driven people asking: what technological change is most misunderstood today,
and how can we build an institution that fills the gap between what is now possible and
what currently exists?
Google was incorporated roughly half a decade after Amazon. And roughly half a
decade after that, a young founder identified an even sparser graph that was also rapidly
moving online and could be represented as an HTML front-end to a database. The result
was Facebook. Even more so than other bubbles, the dot-com bubble was the triumph
of legibility.
3.5 Searching for Another Satoshi: The ICO Bubble
On January 8, 2009, the pseudonymous programmer Satoshi Nakamoto released, on
an obscure cryptography mailing list, the white paper and software of bitcoin—“a new
electronic cash system that’s fully peer-to-peer, with no trusted third party” (Nakamoto
2008).
13
Ironically, this monastic detachment from grubby commercial concerns is subsidized by college endowments, whose investments skew heavily to, wait for it, hedge funds. Further ironies abound. Bezos’
hedge fund was founded by a college professor, and mostly hired academics. And while Stanford is not
technically a hedge fund, it did end up with a healthy stake in Google.
14
This is one of history’s most extreme ideological reversals. To find anything comparable, you have to
look back to figures like Martin Luther, devout Catholic cleric who split the church in two, or Napoleon,
who joined a movement opposed to monarchial overreach and wound up being a dictator. While this
pattern is striking, it’s a pattern, not exactly an outlier: If you’re sufficiently committed to destroying
some powerful institution, you eventually end up forming an even more powerful institution, and since
that institution is subject to the same evolutionary pressures, it copies features of what it replaces.
22
Electronic copy available at: https://ssrn.com/abstract=3469465
What followed was one of the most intense bubbles in history. Over the past two
years—in a series of speculative boom and busts—bitcoin’s price super-exponentially
increased from zero to almost $20,000 at the peak of the so-called “crypto mania” in
early 2018. Bitcoin’s rise spawned one of the purest mimetic bubbles—the ICO mania
of 2017. Since it inception, thousands copies of Bitcoin—so-called altcoins—have been
released in Initial Coin Offerings, or ICOs.
While more than $20 billion have been raised through ICOs since 2017, most of
the projects have already failed. Some, in fact, did not launch anything other than a
plagiarized whitepaper and a working address to receive funds. The ICO-bubble was
driven by the mimetic desire to copy Nakamato’s invention of Bitcoin. While they all
published a white paper, they failed to replicate Bitcoin’s “immaculate conception.”
On January 12, 2009, Nakamoto sent the first Bitcoin transaction to Hal Finney.
When Nakamoto released his code and whitepaper—when Bitcoin’s price was $0—only
an obscure group of early adopters, which consisted mainly of cryptographers, cypherpunks, and developers, was interested in Bitcoin.
Bitcoin’s increasing price attracted new early technology adopters and ideologically
motivated entrepreneurs and investors, which started to invest in the Bitcoin infrastructure. The gradual build-out of the Bitcoin infrastructure, the first rudimentary
exchanges, for example, such as the infamous Japan-based Mt. Gox exchange, which
allowed the conversion from fiat currency into Bitcoin, attracted early retail investors.
The liquidity that these early speculators provided triggered an inflow of more capital
and attention—this resulted in 2013 in the first bitcoin bubble in which bitcoin’s price
peaked at $1,100. After the collapse of the bubble, which was followed by a two-year
bear market, regulated exchanges, such as GDAX, and OTC brokers, such as Cumberland Mining, started to enter the market. The inflow of retail investors, deepening
of liquidity, and the entrance of institutional investors gave in 2017 rise to the largest
bitcoin bubble yet when the price spiked to almost $20,000.
Bitcoin’s adoption can, thus, be characterized by a fractally repeating, exponentially
increasing series of bubbles. It is this sequences of speculative bubbles that bootstrapped
a new form of digital money, which had zero value, to a network that, at the peak of the
last bubble, was valued at more than $320 billion. These bubbles represent unsustainable
accelerating phases of price corrections and rebounds, which are driven by self-reinforcing
feedback loops of herding behavior and FOMO.
Each bitcoin hype-cycle and speculative bubble intensified the mimetic desire to buy
bitcoin. As bitcoin’s price spirals upward, more and more investors are drawn into the
bubble by their desire not to miss out on another price increase. While Bitcoin—the
protocol and protocol-native cryptocurrency—has properties that renders them clearly
superior to other cryptocurrencies or assets, the bitcoin bubbles are mimetic phenomena.
It seems that Nakamoto was aware of the mimetic dynamics that drive speculative
bubbles. In a forum post on the P2P Foundation website, dated February 18, 2009, he
described the mimcetic feedback loop embedded in Bitcoin: “As the number of users
grows, the value per coin increases. It has the potential for a positive feedback loop; as
users increase, the value goes up, which could attract more users to take advantage of
23
Electronic copy available at: https://ssrn.com/abstract=3469465
the increasing value.”
In some sense, Nakamoto seems also to have been aware of the Girardian mimetic
conflict, which, ultimately, results in scapegoating. On December 12, 2010, Nakamoto
published the last message on BitcoinTalk, when bitcoin was $0.20, and then, after a few
email exchanges, disappeared. While he mined an estimated 700,000 bitcoin—which are
currently worth around $8.2 billion—he never sold a single bitcoin. He seemed to have
sacrificed everything to Bitcoin and completely disappeared. Through his pseudonymity
and disappearance, he most likely avoided the scapegoating that would have occurred
were his identity known.
Satoshi is a unique figure in a Girardian sense. So little is known about him that it’s
effectively impossible to compete with him. You can’t be Satoshi-but-more-charismatic,
because nobody knows for sure that they’ve met him. You can’t be Satoshi-but-moreparanoid, because he still hasn’t been firmly identified. You can’t be Satoshi-but-moreforward-thinking, unless you start now and develop something that’s a bigger deal than
Bitcoin ten years from now, without getting caught. And all of this means that Bitcoin
is a perfect focal point for a sustained bubble. Essentially Satoshi has done in life what
Romulus did in legend; transform from a human being into an abstract and untouchable
symbol of the institution he founded.
While later ICO’s all superficially copied bitcoin, such as its code or published
whitepapers, it is precisely this “immaculate conception” of bitcoin that all later ICO
failed to replicate. Bitcoin—its pseudonymous creator and launch on an obscure mailing list, its technology and incentive design—represent a singular event that cannot be
repeated or copied.
The altcoin bubble recalls other historical bubbles. In the 1840s, railroad promoters
would claim that various members of the English nobility had invested in their enterprises, whether or not this was strictly true. In 2017, it was de rigueur for new crytpocurrency startups to claim that their advisors included famous cypherpunks and investors,
even if these advisors were not involved or were—in the case of Hal Finney—deceased.
Price appreciation attracts mimesis, and a crescendo of copies is often enough to halt
price appreciation. With Bitcoin trading at just above half of its peak value, it’s both
too late to call it a failure and too early to call it a success. But as the ICO bubble and
collapse shows, there was more to Bitcoin than just a whitepaper, some source code, and
a rabid fanbase. Bitcoin combined two features that no previous digital currency had
matched and that no future one could: it was just good enough, and it was first.
3.6 The Factor Investing Bubble: Mimesis Made Manifest
The peak of every bubble feels like the end of history: asset prices may be volatile (in
the case of equity bubbles) or stable (in credit), but either way they only make sense in
light of some step-function change in the way the world works.15
15
It’s useful to distinguish between equity and credit bubbles. While they both involve higher asset
prices, the fundamental theses are generally reversed. An equity bubble is a bet that the future will be
radically different, that the upside from new ventures is unbounded. A credit bubble is the inverse, a bet
that the future will be like the past, only more so. In an equity bubble, since investors are extrapolating
24
Electronic copy available at: https://ssrn.com/abstract=3469465
Figure 7: Total Cryptocurrency Market Capitalization, 2016-2019. Source: CoinMarketCap.com
What, then, should we make of factor investing? Does it imply that the history of
discretionary investing has an unknown start date but a definite end point?
In his Advances in Financial Machine Learning, Marcos Lopez de Prado says:
Discretionary portfolio managers (PMs) make investment decisions that do
not follow a particular theory or rationale (if there were one, they would be
systematic PMs). They consume raw news and analyses, but mostly rely on
their judgment or intuition. They may rationalize those decisions based on
some story, but there is always a story for every decision. Because nobody
fully understands the logic behind their bets, investment firms ask them to
work independently from one another, in silos, to ensure diversification. If
you have ever attended a meeting of discretionary PMs, you probably noticed
how long and aimless they can be. Each attendee seems obsessed about one
particular piece of anecdotal information, and giant argumentative leaps are
made without fact-based, empirical evidence.
While this is more than a little rude—he probably won’t get invited to any more
meetings with discretionary PMs—there’s a grain of truth.
Quantitative investors do come up with theories, and these theories can be empirically
verified. Or rather, they can collect data, and out-of-sample data can match a theory or
refute it. In a sense, the discretionary investor has one theory, a theory so fundamental
it’s left unstated: that changes in fundamentals drive changes in asset prices.
A quant may have many observations—that momentum persists, that positive carry
predicts higher returns, that small-caps outperform large-caps and that a price-to-book
ratios mean revert over time. But in another sense, the quant has many theories but no
underlying theory. After all, if there are reasons for these observations, the reasons can
apply differently in different cases, i.e., they can be applied in a discretionary way.
far into the uncertain future, volatility tends to be high—if you go from thinking the Internet will
multiply the economy’s value tenfold to thinking it will double it, you’re still an optimist, but your price
target just cratered by 80%. Credit positions only get more valuable as the consensus view of the future
narrows, so they tend to involve lower volatility. This taxonomy is imperfect—it implies that stat arb
is a credit bubble while new-issue junk bonds in the early 80s were an equity bubble—but it’s a good
distinction to keep in mind.
25
Electronic copy available at: https://ssrn.com/abstract=3469465
Price-to-book may be mean-reverting because investors overweight the saliency of
changes in earnings. Or price-to-book may be mean-reverting because it reflects the
fact that companies with low returns on equity tend to be more vulnerable to extreme
downside moves. (Warren Buffett first invested in Berkshire-Hathaway when it was a
textile manufacturer selling at below tangible book value. When the textile operations
were liquidated in 1985, looms Berkshire had bought for $5,000 four years before were
sold for $26 apiece, below the cost of removing them.)
Momentum might work because it identifies cases where anchoring bias prevents
investors from appreciating important news. Or it could work because it systematizes
fear-of-missing-out, creating a sort of Faux MO that’s quicker than the real thing.
Systematic investing matches Rene Girard’s description of snobbery in Proust:
The snob is also an imitator. He slavishly copies the person whose birth,
fortune, or stylishness he envies... The snob does not dare trust his own
judgment, he desires only objects desired by others. That is why he is the
slave of the fashionable.
Discretionary investors tend to quantify their views, and many of them form their
opinions through what we would consider quantitative processes today. Benjamin Graham data-mined his way to the conclusions in Security Analysis. Michael Milken originally touted high-yield bonds because an NBER study showed that low-rated bonds
persistently outperformed investment-grade credits. Philip Fisher, the dean of the qualitative investment school, relentlessly quantified his own efforts (in his growth-investing
opus, Common Stocks, Uncommon Profits, he notes that the average return from ideas
pitched to him by other investors slightly exceeded his average return from ideas he
found himself).
So many of the great qualitative investors might be categorized as a discretionary
layer on a quantitative signal. Conversely, there’s an underlying theory that explains
quants’ behavior: by backtesting different signals, they’re implicitly reverse-engineering
the thought processes of successful investors.
This is not how they frame it, or how they think about it. If you pressed them, they
would say that some signals produce value because investors are irrational—but every
successful investor is necessarily the more rational counterparty to a given trade, so this
is another way of saying that quants are copying what successful investors have done.
In some cases, this is explicit: quantitative investors have produced factor analyses
of various famous investors’ returns (Buffett, as it turns out, is long the low-vol and
quality factors.) To the extent that a backtest observes a law of nature, one can describe
a discretionary investor as an amateur quant for whom luck and a concentrated portfolio
more than compensated for their primitive wetware computing setup. But to the extent
that backtests reveal things that turned out to be true, a different suggestion presents
itself: good investors identify good investments ex ante; quants determine what these
investments had in common ex post.16
16
You can reinterpret Warren Buffett and Peter Lynch’s careers as a series of macro bets, primarily on
the fact that the US would relatively de-industrialize, but continue to grow consumption, which would
26
Electronic copy available at: https://ssrn.com/abstract=3469465
The quantitative viewpoint is the financial equivalent of strict materialism; discretionary strategies imply some sort of transcendence. Sometimes this is in the sense
that the idiosyncratic returns of a given asset can be meaningfully predicted. In other
cases, discretionary investors are recapitulating Keynes’ critique of econometrics: that
inferences from historical data assume a stationary distribution, but the motivations of
economic actors are intrinsically nonstationary. As with strict materialism, it’s metaphysically unsound; the premises have to be assumed, but determining their validity is
outside the scope of the system.
Quants, in other words, engage in abstract and accidental mimesis. They mine historical data to find out what sorts of investments good investors would have made,
and they build systems to automatically copy these abstracted, hypothetical investors.
They’ve systematized the approach that retail investors used ad hoc when they discovered in the early 2000s that Peter Lynch filed the occasional 13G on a microcap stock.
But, unlike these retail investors, the algorithms don’t read the Journal and see Lynch
warning them not to ride his coattails.17
Since quants are mining data for signals that predict outperformance, there are three
possible endgames:
1. We reach an equilibrium, where fundamental analysis produces some excess returns, and quants are able to harvest a portion of it;
2. As quantitative strategies get more nuanced, and incorporate more data, they
become de facto fundamental strategies, driven less by backwards-looking models
than by forward-looking interpretations of fundamentals;
3. They overshoot: instead of quants reading 100% of the signal represented by raw
data, they read more signal than actually exists. In the short term, this leads
to outperformance: higher asset allocation to quantitative strategies means the
same trades are made in greater size. But in the long term, these trades push
prices further and further out of equilibrium. Eventually, companies either issue
or buy back stock to take advantage of these mispricings, which causes quantitative
strategies to underperform, lose assets, and explosively unwind.
There are many quantitative signals—you can count the big four (value, momentum, carry, short-volatility), or you can enumerate millions.18 But the single biggest
quant signal is the invisible one: the idea that a basket of assets that share similar
characteristics can be traded as a single unit.
Quants who buy the highest-carry decile of the market, and short the lowest-carry
decile, accept that they’re facing adverse selection. Some of the stocks they buy will cut
their dividends, some of the ones they short will get taken private. But they can measure
the average performance of the basket, and with sufficient diversification the individual
lead to a larger financial sector and more returns for branded products. But Lynch and Buffett didn’t
start with a macro thesis and then rush to implement it; the thesis is something you can extract from
asking what ties their various trades together.
17
https://www.wsj.com/articles/SB109779511066146071
18
https://bloom.bg/2NUWbgz
27
Electronic copy available at: https://ssrn.com/abstract=3469465
deviations within the basket will largely net out. Short one stock that doubles, and
you’re broke; short two hundred stocks, and the cost of a double is measured in basis
points.
In a multi-factor model, the return of a given security is the sum of risk free return
plus the returns of various factors weighted by the security’s exposure to those factors,
plus a random error represented by epsilon. The temptation is to use diversification to
make this epsilon equivalent to its use in calculus, as a number so small it’s arbitrarily
close to zero.
Embedded in this practice is the single biggest signal every quant trades on: that the
cost of active security selection exceeds the cost of adverse selection from poor security
selection.
This is the element of factor-based investing that can be subject to bubble dynamics.
While it’s possible that active management was overpriced (and fee compression across
the industry certainly lends credence to this thesis), it’s impossible to be overpriced at
any price.
Quantitative investors are not, in fact, the biggest bettors in this trade: quants
represent around 20% of the $3 trillion hedge fund industry. According to Morningstar,
passive strategies represented almost $4 trillion of US equities at year-end 2018.
All of the investors in these strategies are crowded into a single trade: the betagainst-stock-selection trade. When these assets are a small share of the market, it’s
irrelevant, just as a handful of individual investors replicating a major investor’s holdings
won’t make a difference. But as money flows from active funds to passive funds, it
produces a toxic dynamic: historically, active investors have outperformed before fees
and underperformed after fees. This means that their underlying portfolios outperform
the broader market, and the set of stocks under-owned by active investors underperform.
If assets shift from active to passive, this requires a net purchase of stocks that are
likely to underperform, and a net sale of stocks that are likely to outperform. If we
imagine this as a factor—call it the “stocks pitched at idea dinners” factor—being long
the factor was a tailwind to returns through most of history, and has become a large
and growing headwind as passive funds have grown in importance.
This means active investors face a painful dilemma: they can either mitigate their
underperformance by cutting fees and hugging close to the index, or they can attempt
to make their portfolios as un-market-like as possible. There are examples of both;
some equity managers move into private company investing, or return outside money
to run more concentrated portfolios with their own assets. But others try to keep
up with the market by behaving as much like a passive fund as possible: zero cash
allocation, approximately the same sector biases as their benchmark indices, with either
a tilt towards or away from riskier companies depending on their near-term view of the
market.
Ultimately, the near-term result of this trend is a steepening of the power-law distribution of returns. The very best investors have an infinitely stupid but infinitely
well-funded counterparty to trade against. Everyone else sees diminishing relative returns, and thus diminishing compensation.
28
Electronic copy available at: https://ssrn.com/abstract=3469465
Figure 8: Total Value of Exchange-Traded Smart Beta Products. Source: Morningstar
To most asset managers, the best risk-adjusted decision is to run a passive portfolio
that charges like an active fund. But this leads to further problems down the road: as
fees get compressed, managers lose the revenue necessary to support deep research, and
they have less room to hire. This leads to a demographic pinch: the industry shrinks
headcount slowly through attrition, but doesn’t acquire new talent. Just as in the period
from the 1930s through the 1950s, we can expect the active management industry to
slowly age, and to have an accordingly lower risk tolerance.
To a passive investor or a quantitative investor, this looks like vindication. Systematic
investors start with the assumption that active security selection is insufficiently rigorous,
and see stock-pickers underperform. Passive investors believe that management fees are
too high, and they see those fees drop.
But the implicit mimesis of passive and systematic investing means that these investors are slavishly copying the exact people they’re mocking—more and more money
is following investors less and less worthy of emulation. It’s the fate of passive and systematic investors to end up like Ted Haggard’s congregation listening to a sermon on the
importance of family values, or Bernie Sanders supporters applauding a millionaire who
rails against billionaires: faithfully absorbing a message from someone who no longer
believes it.
29
Electronic copy available at: https://ssrn.com/abstract=3469465
4 A Mimetic Model of Bubbles: Catalyst, Differentiation, Bifurcation,
Crisis, Resolution
Every bubble begins with something real, which inspires entirely too much fakery. It
begins with Da Vinci but gets dominated by Peter Paul Biros19 , because while genius is
rare, the demand of the credulous will always be met by a healthy supply of fraud.
In each of the case studies we’ve looked at, the bubble began with a fundamental
change in the economy, often a profitable one. Railroads were efficient as a means of
conveyance and a generator of dividends, as the 1830s boomlet showed. Electricity did
in fact change the world, and lower dividend payouts did, too, by putting more power in
the hands of entrepreneurs relative to financiers.20 The conglomerateurs were broadly
right in 1960: American management was too old, too hidebound, unaware of better
management practices, and generally under-levered. You wouldn’t be reading this if the
dot-com bubble didn’t have something to it. In the case of Bitcoin, it’s too early to say
for sure, but 99.9% of its price appreciation has occurred subsequent to the first Slashdot
thread calling it a bubble in 2012. Factor-based investing has given investors a better
way to think about performance and returns, albeit one that behaves differently once
acted upon.
The beginning of every bubble is a series of triumphs over skeptics. In fact, in most
cases the peak of a bubble is a peak of absolute hype, while the start of a bubble is
the peak of hype relative to reality. Forming a company is hard, and the more novel
the company the harder it is. Broadly speaking, it’s a poor risk-adjusted decision,
and requires some bad decision-making to get going. (Be suspicious of founders who
seem mentally well-adjusted; they’re probably sociopathic mercenaries cynically riding
a trend. Their startups are a good risk-adjusted bet, at least for them.)
The cycle of a startup might look like this:
1. Someone has an obviously insane idea, like selling Basic to computer hobbyists in
1975 or selling books on the Internet in 1994.
2. They wildly oversell this idea to investors and early employees—it’s quite common
for startups produce sales projections in their early stages, but rare for them to
hit their initial numbers.
3. The combination of investor money and talent makes the idea a little less crazy,
albeit still ambitious.
4. On the basis of this track record, the founder raises more money and recruits more
talent.
5. Over time, either reality converges on these wild expectations, as the expectations
get toned down and reality catches up—or exaggeration becomes the only possible
19
https://www.newyorker.com/magazine/2010/07/12/the-mark-of-a-masterpiece
In countries where business is more reliant on debt than equity, and on bank loans rather than bonds,
financiers have more power. And since they’re more concentrated, it’s easier for government to exercise
power over them. When regulators are prudent, that can be a useful way to coordinate different parts
of a complex supply chain, but when they’re corrupt, or incompetent but not corrupt, it’s a disaster.
20
30
Electronic copy available at: https://ssrn.com/abstract=3469465
way to keep the company alive, and founders ratchet up self- and other-deception
until it reaches the breaking point.
At each stage, the future looks both brighter and more certain—at least for true
believers. Over time, this leads to bifurcation, both among companies and investors:
companies can choose to “buy in,” as Barnes & Noble did with barnesandnoble.com,
or as Chemical’s board did when they considered diversifying into insurance lest Saul
Steinberg’s insurance company diversify them out of a job. Other companies turn their
backs on the bubble—they don’t say “Not yet,” they just say “No.”
In financial markets, this process leads to steady reinforcement of both theses. When
Netflix was a small player in electronics, a Blockbuster investor could ignore them. As
Netflix grew, the population of Blockbuster investors evaporatively cooled, as anyone
who worried about Netflix sold out. Meanwhile, an investor interested in the “movie
rentals” business could have looked at Netflix and Blockbuster as comparable companies
in 2004; one a bit more expensive and faster-growing than the other, but both in roughly
the same business. A few years later, this was untenable: there was no chance that a
Blockbuster investor would switch to Netflix if Netflix dropped 20% on poor earnings; to
own Blockbuster was to think that Netflix was overpriced by an order of magnitude. To
the Blockbuster investor, owning Netflix is literally incomprehensible, and vice-versa.
As Girard notes, quoting Proust:
Of such a nature is hatred which compounds from the lives of our enemies a
fiction which is wholly false. Instead of thinking of them as ordinary human
beings knowing ordinary human happiness and occasionally exposed to the
sorrows which afflict all mankind and ought to arouse in us a feeling of kindly
sympathy, we attribute to them an attitude of arrogant self-satisfaction which
pours oil upon the flames of our anger. For hatred transfigures individuals
no less than does desire and like desire sets us thirsting for human blood.
On the other hand since it can find satisfaction only in the destruction of
the supposed self-satisfaction which so irritates us, we imagine that selfsatisfaction, see it, believe it to be in a perpetual process of disintegration.
No more than love does hatred follow the dictates of reason, but goes through
life with eyes fixed on an unconquerable hope.
This places bubble investors and bubble skeptics in separate worlds. It also introduces
artificial stability into the price of legacy companies and artificial volatility into the price
of challengers. If no reasonable person could hold a view that would justify owning
both the incumbent and the challenger, then there’s a smaller natural constituency
of buyers when the challenger has a tough quarter. On the incumbent side, a nearly
opposite dynamic holds: at a macro scale, their investors are betting on mean reversion;
if Blockbuster loses share to Netflix one quarter, it’s not evidence to a Blockbuster owner
that such share losses will accelerate—so they tend to buy the dip, in dip after dip.
This bimodal investor population explains why growth companies tend to be more
volatile as they peak in value, while value companies tend to drift along quietly until
31
Electronic copy available at: https://ssrn.com/abstract=3469465
they suddenly plummet to zero. These extreme price movements are not just caused
by exogenous fundamental factors; they’re caused by investors’ recognition that they’re
increasingly alone in their views. As Girard says, “A clear view of the inner workings
indicates a crisis in the system; it is a sign of disintegration.”
A clear-eyed view of the bubble cycle—an enterprise that starts out fabulously overhyped, and gradually grows into the hype—would lead to collapse, so participants rely
on self-reinforcing collective delusions to maintain their esprit.
Perhaps the best summary of this bifurcated view, in which a narrative of success necessarily describes the hubristic self-destruction of competitors, comes from Neal
Stephenson’s Cryptonomicon, published in 1999 at the peak of the dotcom bubble. In
Cryptonomicon, a newly-formed startup presents their business plan:
Phase 1: After taking vows of celibacy and abstinence and foregoing all of our
material possessions for homespun robes, we (viz. appended resumes) will
move into a modest complex of scavenged refrigerator boxes in the central
Gobi Desert, where real estate is so cheap that we are actually being paid to
occupy it, thereby enhancing shareholder value even before we have actually
done anything. On a daily ration consisting of a handful of uncooked rice
and a ladleful of water, we will [begin to do stuff].
Phase 2, 3, 4, . . . , n - 1: We will [do more stuff, steadily enhancing
shareholder value in the process] unless [the earth is struck by an asteroid a
thousand miles in diameter, in which case certain assumptions will have to
be readjusted; refer to Spreadsheets 397-413].
Phase n: before the ink on our Nobel Prize certificates is dry, we will confiscate the property of our competitors, including anyone foolish enough to
have invested in their pathetic companies. We will sell all of these people
into slavery. All proceeds will be redistributed among our shareholders, who
will hardly notice, since Spreadsheet 265 demonstrates that, by this time,
the company will be larger than the British Empire at its zenith.
Stephenson is exaggerating. (Barely. The British Empire was roughly a quarter of
world GDP in 1870, or around $130bn in 1990 international dollars, which is within
spitting distance of the annual revenues of several major technology companies. The
British Empire did not deliver the same levels of free cash flow conversion as a typical
large software company.)
Conveniently among investors, loyalty tends not to last past delisting. There may
be a few Blockbuster partisans out there who still rail against Reed Hastings and all he
represents, but for the most part, investors take their lumps and move on. A corporation
is immortal so long as it’s either generating cash flow or on life support (whether from
operations or from investors), but dies immediately when the support is withdrawn.
Once it dies, its constituency naturally disperses, dramatically reducing the odds of
cyclical rivalry.
(There are cases of this, like Dodgeball’s founders building the competing Foursquare,
Liberty Medical’s founder forming Liberator Medical with the same business model, and
32
Electronic copy available at: https://ssrn.com/abstract=3469465
some serial entrepreneurs are that way because they’re serial failures as operators but
serial successes as fundraisers, but it’s very much the exception.)
Financial bubbles are, in one sense, the most benign expression of mimesis. Money
changes hands, but the world does tend to improve, either through over-investment in
technology (in equity bubbles) or through a more detailed mapping of the status quo
(in credit bubbles). If Lenin had been alive in 1997, he might have assaulted the capitalist establishment by launching an over-funded B2C play, writing his business plan in
Starbucks rather than composing manifestos at Café Central. Cynicism, manipulativeness, narcissistic charisma, sociopathic risk tolerance—all of these can be channeled into
merely wasteful behaviors rather than catastrophically destructive ones.
Adjacent to financial bubbles are the technology bubbles that are funded by governments, rather than by private actors. Both the Manhattan Project and the Apollo
Program were enormous technical risks requiring a single-digit percentage of GDP to
be invested in an uncertain outcome. One result of these bubbles was coordination—
building a tenth of the Apollo Program would have been a foolish investment for a private
or public actor, but given a government commitment to completing the entire project,
it made sense for individual participants to build out smaller components. The Apollo
Program required advances in rocketry, communications, computation, and simulation;
the project wouldn’t have been viable if any of these had failed, but the existence of the
project provided the initial shove necessary to get the technological ball rolling. To take
an extreme example, when Voyager was launched, it had the ability to encode extremely
succinct messages. The technology to decode them did not exist, but the team believed
(correctly) that this technology would be developed in time.
A clear-eyed view of financial bubbles collapses their originality but reveals their
internal structure. Crypto-currency promoters use the same strategies as railway entrepreneurs, passive investors apply the same logic as conglomerateurs, everyone assiduously seeks to be the most original-looking copy of a model they don’t fully understand.
When we stand outside the bubble, we can see it as not merely a financial phenomenon,
but a sociological and even literary one. Shakespeare, the Brothers Grimm, St. Paul,
and Walt Disney all understood this: the best stories to tell are the ones that have
been told over and over again, and the stories of competitive mimesis and of deification
followed by scapegoating are the oldest and best of all.
5 References
Bisonnette, Zac. The Great Beanie Baby Bubble: Mass Delusion and the Dark Side of
Cute. Portfolio, 2016.
Brooks, John. The Go-Go Years: The Drama and Crashing Finale of Wall Streets
Bullish 60s. John Wiley & Sons, 1999.
Chancellor, Edward. Devil Take the Hindmost: A History of Financial Speculation.
Penguin Group, 1999.
Girard, Réne. Violence and the Sacred. Continuum, 2005.
33
Electronic copy available at: https://ssrn.com/abstract=3469465
Hobart, Byrne, “Good Bubble, Bad Bubble http://www.byrnehobart.com/blog/goodbubble-bad-bubble/,” 2010
Huber, Tobias A. “Bubbles As Innovation Accelerators https://medium.com/newco/innovativeexuberance-ad75ee39f4c,” 2017
Kindelberger, Charles P., and Robert Z. Aliber. Manias, Panics and Crashes: A
History of Financial Crisis. Palgrave Macmillan, 2005.
Odlyzko, Andrew. “Collective Hallucinations and Inefficient Markets: The British
Railway Mania of the 1840s.” University of Minnesota, 2010.
Nakamoto, Satoshi. ”Bitcoin: A Peer-to-Peer Electronic Cash System.” 2008.
Sornette, Didier and Peter Cauwels. “Financial Bubbles: Mechanism, Diagnostic
and State of the World. Review of Behavioral Economics. 2015; 2(3): 279–305.
Spyrou, Spyros. ”Herding in Financial Markets: A Review of the Literature.” Review
of Behavioral Finance. 5 (2): 175 - 194 , 2013.
Stephenson, Neal. Cryptonomicon. Random House, 1999.
Thiel, Peter A. and Blake Masters. Zero To One: Notes on Startups, or How to
Build the Future. Broadway Business, 2014.
Wheatley Spencer, Didier Sornette, Tobias Huber, Max Reppen and Robert Gantner.
“Are Bitcoin bubbles predictable? Combining a generalized Metcalfe’s Law and the LogPeriodic Power Law Singularity model.” Royal Society Open Science. 2019. 6: 180538.
34
Electronic copy available at: https://ssrn.com/abstract=3469465